August 2-4, 1983

CONF-830 882—Vol.1

DE84 002315 DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents tbat its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government 01 any agency thereof. Volume 1 Proceedings of the Second Symposium on Containment of Underground Nuclear Explosions Kirtland AFB, Albuquerque, NM August 2-4, 1983

Compiled by Clifford W. 01 sen

Printed by Field Command/Defense Nuclear Agency Kirtland Air Force Base, N.M.

NOTICE PORTIONS OF THIS REPORT ARE ttlEGSBLE. it has been reproduced from the best available copy to permit the broadest Lawrence Uvermore National Laboratory DISTRIBUTE OF 7^ !::-^-;j « „ LIST OF ATTENDEES JTAINMENT OF UNDERGROUND NUCLEAR EXPLOSIONS Allen, Robert T. Pac Tech Hague, Richard W. NVOO App, Frederick N. LANL Hatch, Melton A. EG&G Aut, Warren FCDNA Hawkins, Ward L. LANL Ballantine, George FCDNA Hearst, Joseph R. LLNL Barthel, James R. S-Cubed Hier, James A. FCDNA Basinger, Delia M. LLNL Higginbotham, Michael L- S-Cubed - Batra, Ravi LANL Higgins, Gary H. LLNL Behne, Joseph C. LLNL Hill, George EG&G Bel low,Jr., Bernard W. EG&G Hill, Leslie R. SNLA Bloemker, Carl F. FCDNA Holmes, Robert C. Tech Reps Breeze, Stephen P. SNLA Holzer, Alfred LLNL 3roderick, Nicholas EG&G Homuth, Emi1 F. LANL Brown lee, Robert R. LANL House, Jack W. LANL Brown, Roger J. FCDNA Howard, Lowell FCDNA Brown, William T. SNLA Howard, Nancy W. LLNL Broyles, Carter D. SNLA Hudson, Billy C LLNL Bruesch, Grant T. F&S Hunsaker, Travis D. EG&G Burkhard, Norman R. LLNL Jenkins, Evan C. USGS Carlson, Richard C. LLNL Jones, Eric M. LANL Carroll, Roderick D. USGS Jorgenson, Eric D. EG&G Case, Clinton M. DRI Kalinowski, John A. EG&G Chavez, Patrick F. Kaye, Marie A. LANL Carothers, James E. SNLA Cheney, James A. LLNL Keller Carl E. FCDNA Clark, Stephen R. LANL Keller, Charles F. LANL Cogbill, Allen H. EG&G Kernpka, Steven N. SNLA Coleman, Ronald R. LANL Krier Donathon J. LANL Cooley, Craig H. EG&G Kunkle, Thomas D. LANL Cook, C. Wayne Terra Tek LaComb, Joseph W. FCDNA Covington, Harry R. SNLA LeDelfe, Carol M. LANL Coyle, Philip E. USGS Larson, -Donald B. LLNL Cramer, James D. LLNL Lie, Kyoon-Haeng S-Cubed Oavey, William R. SEA Lowry, William E. LLNL Davies, William J. SNLA Lundberg, Anders W. LLNL Dismukes, Charles R. F&S Malik, John S. LANL Diven, Benjamin C. S-Cubed Margolin, Leonard G. LANL Dockery, Holly A. LANL Marusak, Nancy L. LANL Dolce, Sanders R. LANL Mathews, Mark A. LANL Donithan, James D. SNLA McKague, H. Lawrence LLNL Douglass, Carl W. LLNL-N McKinnis, William B. LLNL-N Drellack, Sigmund L. F&S McNairy,Jr., Ray M. LLNL Duff, Russell E. F&S Meadows, Wayne R. LANL Ferderber, Lawrence J. S-Cubed Mehl, Clarence R. SNLA Florence, Alexander L. LLNL-N Miller, Lori R. PI Freeman, Kenneth S. SRI Montgomery, Stephen T. SNLA Fogel, Martin B. FCDNA Moran, Bill LLNL Gaffney, Edward S. Pac Tech Morrison, Frank A. LLNL Geil, Robert G. LANL Morrison, Jtrry N. F&S Glenn, H. David LLNL Mudra, Paul J. NVOO Glenn, Lewis A. LLNL Navarro, Richard NVOO Gonzales, Jose L. LLNL Nilson, Robert H. S-Cubed Griffiths, Steward K. F&S O'Brien, Margaret B. F&S Grinnell, Stephen E. SNLA Oliver, Ronald D. LANL Gulick, Charles W. LLNL Olsen, Clifford W. LLNL SNLA Olwin, Richard B. LANL LIST OF ATTENDEES (Continued) Patch, Dan F. Pac Tech pawloski, Gayle A. LLNL Penland, Charlene K. LLNL Peterson, Edward W. S-Cubed Piwinskii Alf J. LLNL Planner, Harry H. LANL Plimpton, James D. SNLA Proffer, Wi1liam J. S-Cubed Rambo, John T. LLNL Reed, Leonard L. S-Cubed Reed, Melvin E. EG&G Richardson, William W. LLNL Richter, Jerome M. EG&G Rimer, Norton S-Cubed Ristvet, Bryon FCDNA Ryerson, Fredrick J. LLNL Saunder, B. Roy LANL Schmidt, Casey S. LLNL-N Shannon, Spencer S. LANL Sharps, Joseph A. USGS Shroba., Ralph R. USGS Smith, Barham W. LANL Smith, Carl W. SNLA Smith, Redo H. Terra Tek Spataro, Samuel J. LLNL Stubbs, Theodore r. EG&G Summa, William J. FCDNA Tatom, Jerry f1. LANL Terhune, Robert W. LLNL Thomsen, Jeffrey M. PI Toman, John LLNL Tootle, Jeannette M. LLNL Townsend, Dean R. F&S Travis, Bryan J. LANL Twenhofel, William S. SNLA Van de Werken, Martha G. USGS Wagoner, Jeffrey L. LLNL Wallace, Oelmont EG&G Waller, Nannette L. FCDNA Warren, Richard G. LANL Wheeler, Vernon E. LLNL Wiehe, John W, S-Cubed Wohletz, Kenneth H. LANL Yarrington, Paul SNLA Young, Kathryn M. LLNL VOLUME 1 Table of Contents Paqe HOW WE GOT WHERE WE ARE, or Containment, The Reluctant Science; James E. Carothers 1 Cements for Large Diameter Nevada Test Site Drill Holes; Grant T. Bruesch 17 LLNL Stemming Advances: W. E. Lowry, A. L. Lundberg and L. I. Starrh 49 Quantitative Determination of Minerals in Nevada Test Site Samples by X-ray Diffraction; Gayle A. Pawloski . 75 Containment We'll Logging Program; Ronald D. Oliver . 97 A Possible New Acoustic Logging Probe for NTS; Ravi Batra 103 Spectral Gamma-Ray Logging for Clay Content; Huang Long-ji and Joseph R. Hearst 113 Lithology and Log Comparison of Some Boreholes in Southern Yucca Flat; Mark A. Mathews and Carol M. LaDelfe 121 Ash-Flow Tuff Distribution and Fault Patterns as Indicators of Rotation of Late-Tertiary Regional Extension, Nevada Test Site; Holly Dockery Ander 155 Subsurface Scarps on the Paleozoic Surface at the Nevada Test Site- Faults or Topography; H. L. McKague (Abstract) 175 Mapping of Paleozoic Structure in North-Central Area 7 of the Nevada Test Site - A Geophysical Case History;'Fredrick N. App, Wayne R. Meadows, and Allen H. Cogbill . 177 Determination of Subsurface Geological Structure with Borehole Gravimetry; S. R. Clark and J. R. Hearst 205 Geochemical Similarities Between Volcanic Units at Yucca Mountain and Pahute Mesa: Evidence for a Common Magmatic Origin for Volcanic Sequences that Flank the Timber Mountain Caldera; R. G. Warren . . . 213 Soil Properties as Age Indicators for Quaternary Surficial Deposits and Faults in the Nevada Test Site Area, Southern Nevada; R. R. Shorha (Abstract) 245 Influence of Geologic Structure on Alluvial Sedimentation in Northwestern Yucca Flat, Nye County, Nevada; J. L. Wagoner 247 Geologic Investigations of Drill Hole Sloughing Problems, Nevada Test Site; S. L. Drellack, Jr., W. J. Davies, J. L. Gonzales, and W. L. Hawkins 271

VI VOLUME 1 Table of Contents (Continued) Page LLNL Site Selection Procedures; C. W. Olsen 293 Thermodynamics of Hydrogen Generation; E. S. Gaffney 305 A Chemical Investigation of Glasses Produced by the kAINIER Underground Nuclear Explosion; A. J. Piwinskii, F. J. Ryerson and W. F. Beiriger 323 Calculation of Early .ime Subsidence Phenomena (A Calculational Study of the CARPETBAG Event); J. T. Rambo, D. E. Burton, F. A. Morrison, Jr., and R. W. Terhune 369 The PG-2 Photogrammetric Plotter: A Rapid and Accurate Means of Mapping Surface Effects Produced by Subsurface Nuclear Testing at the Nevada Test Site, Nevada; Martha Garcia Van ae Wcrken 393 Containment Analysis for the QUESO Nuclear Event; H. D. Glenn, T. F. Stubbs, J. A. Kalinowski, and E. C Woodward 411 Author Index ..,...., 431 VOLUME 2 Table of Contents e Spherical Wave Particle Velocities in Geologic Materials from Laboratory Experiments; J. C. Cizek and A. L. Florence 1 Finite Difference Simulations of Particle Velocity Records from Small-Scale Explosive Tests; Norton Rimer, K. Lie, and J. T. Cherry 41 Containment Failure Criteria ana Their Validation Using Small-Scale Experiments; Martin b. Fogel 71 Residual Stress Fields - Results from High Explosive Field Tests; Carl W. Smith 87 Results of a Laboratory Test Program in Support of Containment Modeling Efforts for Nevada Test Site Materials; C. H. Cooley, R. H. Smith, J. F. Schatz, and J. LaComb 109 Analysis of Near Field Ground Motion from Nuclear Detonations in High porosity Media; Robert W. Terhune and Myron Heus'inkveld . . 123 A New Technique for Modeling Fracture and Spall; Barham W. Smith and Leonard G. Margolin ...... 165 The Development of Containment Models with Application tv High Explosive Events in Tuff; W. T. Brown and P. F. Chavez (Abstract) 177

vii VOLUME 2 Table of Contents (Continued) Page Laboratory Hydrcfracture Experiments for Containment Investigations; J. C. Cizek and A. L. Florence 179 In-situ Steam Fracture Experiments; P. L. Lagus, E. W. Peterson and H. E. Wu 197 Comparison of the KRAK Model with Experimental Data; Bryan J. Travis 231 Extent of Gas Fracturing Around an Explosively Driven Cavity; R. H. Nilson 253 Numerical Analysis of Hydraulically-Driven Fractures: Comparisons with Similarity Solutions and Experimental Results; Stewart K. Griffiths, Carl W. Smith, and Robert H. Nilson 281 LOS Flow Reduction; J. R. Barthel 307 Efforts to Reduce Pipe Flow: The LS-6 Experiment; L. R. E. Miller, J. M. Thomsen, and R. J. Funstori 335 Containment Science on a Centrifuge; E. S. aaffney and ,1. A. Cheney 365 Model Testing of the Low Yield Concept; C. W. Gulick 379 Calculated Loading on a Fast Acting Closure; S. T. Montgomery ... 401 HURON LANDING Post-Shot Report; R. E. Duff, J. LaComb, and R. Bass 413

VTIl VOLUME 3 (Confidential-Restricted Data) Table of Contents Page Decoupled Nuclear Test Design - Experience and a Novel Approach; Dan F. Patch and J. Eddie Welch 1 Calculation of High Surface Velocity Due to Focusing in the TYBO Event; John T. Rambo and Jon B. Bryan 35 The Effect of a Helical Insert on Detonation Induced Flow in a Line-of-Sight Pipe; C. Wayne Cook and Robert C. Bass 53 Maximum Credible Yields as a Source of Surprise, or A Possible Lcophole in the Logic of Containment; Robert R. Brownlee 59 Zero Expected Release Operational Model; Gary H. Higgins 63 Containment Phenomenology: Differences Between High- and Low-Yield Events in Yucca Flat; B. C. Hudson 73 A New Look at Experience with CO2 in the Work Point Region; Nancy W. Howard 83 Empirical Ground Motion Predictions for Use in Containment Evaluation; Vernon E. Wheeler ... 103 Review of Surface Effects from Underground Nuclear Explosions Near the Yucca Fault, Nevada Test Site; Ward L. Hawkins ... 115 Design Support for Low Yield Testing; R. T. Allen, Dan Patch and Eddie Welch 139 Development of a Low Yield Test Bed - FAC and MIDNIGHT ZEPHYR; Leslie R. Hill and William J. Summa 187 The Yield of BANDICOOT; R. R. Brownlee, T. D. Kunkle, and E. M, Jones I97 Surface Detection of Radioactive Gases from Underground Testing; R. G. Geil 207

VOLUME 4 (Secret-Restricted Data - Sigma 1) Table of Contents Page Blast Containment in the High Energy Density Facility; L. A. Glenn 1 PROCEEDINGS OF THE SECOND SYMPOSIUM ON CONTAINMENT OF UNDERGROUND NUCLEAR EXPLOSIONS

This is Volume 1 of a set of four volumes of the Proceedings of the Second Symposium on the Containment of Underground Nuclear Explosions held in Albuquerque, New Mexico on 2-4 August 1983. Volumes 1 and 2 contain unclassified work; Volume 3 is Confidential - Restricted Data; Volume 4 is Secret - Restricted Data, Sigma 1. Volume 1 contains a Table of Contents for all four volumes, an author index, ana a list of attendees. The meeting was hosted by Field Command/Defense Nuclear Agency. They are to be thanked for being the host as we'll as for printing these proceedings. The participation of all presenters and attendees was most appreciated.

The Program Committee Les Hill, SNLA Nancy Howard, LLNL Carl Keller, DNA Cliff Olsen, LLNL Barry Smith, LANL HOW WE GOT WHERE WE ARE or Containment, The Reluctant Science

Steps Along the Way

Invited Paper by James E. Carothers, Chairman Containment Evaluation Panel ITHHin Ci J1H" i^

4 a »

<••>' S * ' i'

.., _ j s

h \ u ; '.;•'! f J • r had [w-

of rhr r?;; r^"!i;:?'fd ^ 'S w n'.iriaf.ion ^'«>h;ne at thi- AEC's expanse, Despite the precautionary •measu^s. the ARC insisted that **r«di»tion had not re»ehrd

FOOGY WEATIfEB. The blast, set off in ggy r, was ieft Sn Nevada Utah snd «av* CaJfiforn- s4 as f«r as 400 miles away, * *ood SIR tefore breakfas* HOW WE GOT WHERE WE ARE or Containment, the Reluctant Science

During the history of nuclear testing each event has had to be justified by somebody or some group of people, broadly defined, usually, as the sponsoring organization, but usually a person. A nuclear detonation with its expenditures of money and nuclear material, its attendant possible domestic or international consequences, is not a trivial happening and each is approved specifically by the President. Sometimes the justification is straightforward and I am sure it sometimes seems routine. After all, there is a certain amount of money, a certain number of shot slots, and what shot is to be done? Other times, as in the case of the Cannikin event, there is a vary broad national and international debate about doing the event at all. However, it usually comes down to the fact that somebody, somewhere, thinks, "That's my shot. Mine." There are some things which must be done to accomplish the purpose of the shot, but they are very simple in the mind of the person who thinks of it as "my shot." He wants a certain kind of device, he wants to put a few detectors here and there, and he wants to have a long string that's attached to some kind of trigger, and he wants to pull the string. The bomb goes off, he gets his data, and that's it. It's really pretty simple, and that's really all you need for the purposes of the event, in the mind of the person who justified the event.

Now please note that I did not say anything about weather patterns, fallout: patterns, Pike model, towers, tunnels, emplacement holes, ground motion, Pz surfaces, phase of the moon, radiation monitor?, or any number of other things that are completely irrelevent to the purposes of the person who wants to do the shot. Nuclear devices really don't have to be put in holes in thfe ground, they really don't have to be hung on balloons, they really don't hav« to be put in tunnels, and other esoteric places.

It's not necessarily easy to justify the need for a nuclear detonation, but it can be done by some people, usually designers. There are a lot of other people who would like to do some kind of experiment on the device, but they have absolutely zero chance of justifying a nuclear detonation for their purposes. So they come around, with the proposal, "Gee, if you'd only let me look at a little piece of the device — you're not going to use all of the energy — it would really be nice." So we get to add-ons. Nobody likes add-ons, except the people that are involved in them. They complicate the experiment, they cost extra money, they have the possibility of compromising the primary purpose of the event, and so the person whose "my shot" it is doesn't like them. The Test Director doesn't like them because they're like fleas and each one has to be scratched. They multiply rapidly, they suck the financial blood, and they cause a lot of irriatation. They are basically the poor, who are always with us, and while out of some form of basic humanity we might be willing to give them some small handout, we certainly don't think they deserve the same amenities as do the rich; namely, the person whose "my shot" it is. TTSF Exposed to Rays Test; All We T „ WASHINGTON.' March 11 •— (i.A'inr: bu* m f<' effects have; ? r —The Atomic Energy Cor*. mission is&ici toi-gbt that VAC:\

"were 5ubj radiation''' tiur; ". rv- atomic test :n :Vi-- M. .'!.''. -: fands but all these c\;•••.-.•-.-j ;-- "reported we!!-" The commission anr.o,:- "• •.::: March 1 that 'iho fn-st of :% •• of nuclear tests had >•'.?.:->•.; .: the Pacific provtr.i: ^ :;,..:::•:: NONE BUaNEI>. The commission ar:v.." • •• -ment today said:

- routine atomic test 1114!^ >iar- Flyer in W shall Islafcds. t4*'fnty-f-sght United States pernonHel ;ind Admits Theft I S3€ residents were trannported from nei|fhbormg: at-f;!!-, to Kwmjalein Islmnl according to plan« as a •»»•

**The hidivkluaK -\»

And, I am sorry to tell you, after the two previous speakers have pointed out to you how important you are, that you are an add-on too. You cost time and money and potentially interfere with the schedule and with the data. But, nonetheless, it does appear that you're always going to be there.

With that as an introduction, I'd like to talk about how we got where we are, how it came to be that you are here, and how you got to be so necessary to the nuclear device program that it looks as though you will always be here, add-on or not.

If we are to think about how we got here, we need to back to very early times, the beginnings, and with the advantage of knowing some of the results, to see what some of the significant places were. Where was the beginning? It's hard to find beginnings, and each of us might a different place to start. It's somewhat like trying to find a small rivulet somewhere that later is a large river. Is it this one that's the beginning, or is it that one? Perhaps for us, one of the beginnings occurred when Becquerel laid a peice of pitchblend on a piece of photosensitive material and found that it left an image. Maybe part of the beginning was when Roentgen used these unknown radiations to take the first X-ray picture, showing that these radiations penetrated human tissue and were absorbed more in some parts than in others. Where ever we start, the path leads to the Curies, to people like Lawrence and others who worked with ingenious mechanisms to probe the invisible core of the elements, and it certainly leads to those who first observed the fission process. Things were going along, these beginnings, in a rather mild, academic, scholarly way. There were these people, poking around in their laboratories with things that most people didn't concern themselves with, no'r were most people much interested in.

Then, lying across everything, there was the absolutely cataclysmic catastrophy of an enormous war, which changed the path we are trying to follow from an orderly, scholarly remote thing into a central fact in the lives of the people of the world. Certainly here, in this great war, the route we are following can clearly be found. I need to talk about that war just a little bit to place what happened in context, because it was very, very significant in determining the course of how we got to where we are.

This war was probably the greatest catastrophy which has ever been visited upon the human race. It covered the earth. It involved essentially every nation and hundreds and hundreds of millions of people. Tens of millions of people died. They died in a lot of different ways, but the end result was the same. They died. They died on the battlefield, in airplanes, in ships. Bombs fell and cities burned and villages burned and the people died in their homes and in the streets. People died a lot of ways; they fought in the snow an while. ??:•

2 rj. /A wf

•;•• r

» f <• I. f (»

• i Sr'.o-n

?s,, , o.» .,.. ; Oi.lt. f,

i iot th

•d by tin

BRAVO Event February 28, 1954 and froze there, they drowned in the sea, they died on beaches, in the desert, in the fields. They died defending things, and they died attacking things. Many of them slowly starved to death, or died of disease. Some of them had nice funerals, and some them funerals that were not so nice, and the cemetaries grew and they filled, and sometimes the bodies became just piles of bodies. Death was very public and everywhere, and grief was still a solitary thing, but it was everywhere too. And that went on for a long time. It went on for years, and then, one day, there was a flash of light, and then another one, and the war ended. Just like that. To the people of the world it was almost like a miracle. A hundred thousand people in Japan joined the other millions of the dead, but just about everyone else thanked whatever God they thanked and said, "Maybe now I will live. Maybe now my children won't starve. Maybe my son will come home - or my husband - or my father. Maybe I can think about having a life to live." There was a great outpouring of - happiness is not the right word - relief, perhaps.

The people who had caused this to happen were looked on almost with awe by the people of the United States. These miracle workers had caused this to be so. They were no longer simple fuzzy headed folks working in dusty laboratories. They were imbued by people with the role of those with an arcane knowledge. The knew something that other people didn't know, and that, in some sense, other people couldn't know. They held a secret. They knew, and they could deal with this awesome force, and they could tell you about it, except they couldn't because it was a secret. It was a secret partly because you couldn't understand anyway, but it was a secret mostly because if a country knew it, it was invincible, and if it didn't, it was helpless. Therefore, it was very important that those who might wish you harm should not learn your secret. And therefore, only a selected few should be privy to the secret, and the rest should rely on what they were told by those who knew.

Some countries began to work very hard to have such weapons for themselves, so they would not be helpless before those who knew-the secret. And, developments continued in this country. Less than a year after the end of the war there were experiments to see what the effects of the atomic bomb were on naval vessels. We happened to have a number of captured ships, so some were t?ken to Bikini so effects could be studied. Even af this early time there were a few people who were concerned about effects beyond those that were predicted or planned. Anatol Schneiderov, in an article in the journal Science, spoke of fractures in the earth's crust and possible changes in the world's weather from dust thrown into the air. And some people listened and worried, particularly some of the seamen who were to support the experiments, and their families. But no one spoke to their concerns other than to cancel shore leaves before the ship sailed. And the Crossroads experiments were done.

Not long after that, some three years, the Soviets detonated an atomic bomb of their own. That bothered people, because that meant they too knew the secret, and we hadn't told them. So, how could they have learned it? They must have stolen it - obviously they must have stolen our secret. Th?t kind of thinking colored the social fabric of this country for some time. If you keep insisting it's a secret and un-understandable, the only way someone can find out how to make an atomic device is to steal the secret. It wasn't explained as the application of relatively simple concepts in physics which many people "OJCM UJ Cl UJ could work out, particularly after they knew that the process worked. So, since they had stolen some of the secret from as, that meant we should say even less about it, because who knew who the thief was or where the thieves were?

Then, another war started, in Korea. The experiments that were done at Enewetak and Bikini took a lot of military support, and were a long way away. Experiments someplace closer to hand would be much more convenient and much less of a drain on resources needed to support the war in Korea, and so people began to think of a continental test area,. la was on December.18, 1950, two days after he had declared a National Emergency due to the Korean fighting, that President Truman approved a recommendation that what is now the Nevada Test Site be established. About a month later, in January of 1951 the first airdrop of Operation Ranger occured in Nevada.

The people who planned and conducted Operation Ranger knew, of course, that radioactive material would be generated and would possibly produce exposure to people in the general peculation. They had to think about what might happen and what they would do under certain circumstances, based on certain assumptions. Unfortunately, these assumptions couldn't be publically discussed, because that would or could lead to information which was secret. However, that information is now declassified and one of the assumptions that was made was the following: "It is as.sumed that any member of the general public may receive external exposure of up to 25r without danger." That was based on the idea that people received that much exposure in some medical x-rays and it did not appear to hurt them.

The Ranger series was all airbursts, and passed uneventfully, except for a snowstorm which deposited radioactivity in Rochester, New York. Fccept for two 1.2 kt surface shots at the end of 1951, all the detonations of Buster- Jangle were airbursts. Tumbler-Snapper, in 1952, marked the beginning of events of several kilotons fired on towers. In 1953, most of the events of Upshot-Knothole were tower shots, yields began to increase, and there began to be difficulties. SIMON was fired, some cars off site were contaminated and had to be washed down. There were a number of deaths in sheep herds adjacent to the Test Site and there were people who believed fallout from the shots caused these deaths. After shot HARRY, people in St. George, Utah, were told to stay indoors for a few hours to decrease their exposure from the fallout on the community. There was a rainout of activity in Troy, New York. Enough people began to question what was happening to cause a couple of Congressional inquiries to the AEC, but there wasn't anyone to tell the people what was happening because public discussion by knowledgeable people would give away, probably, some secrets to our enemies. Who some years before had detonated their own atomic device. By not discussing these things we presumably kept the Soviets, and certainly the people of the United States, in ignorance of what was happening.

There was a Commission investigation of the sheep deaths, and I would like to quote now from a document, DOE/ES-0005, titled "The United States Nuclear Weapons Program: a Summary History."

"The Commission immediately launched an investigation of.the sheep deaths and a full-scale review of Nevada testing. Commission scientists, 'fully aware that the future of continental testing might hang on the results,1 • :/V

LLJCM

<—•Ol o

10 concluded that fallout did not cause the sheep deaths. They remained silent on whether it might have been a contributing factor and in pubiic announcements the Commission 'glossed over the fact that scientific opinion on the question was not unanimous.1"

In 1954, anGther problem arose. The Castle series began with a 15 megaton shot called BRAVO. Unfortunately, the fallout from BRAVO took a path which resulted in the exposure of 236 Marshalese and 31 weather service personnel who were on downwind atolls, the crew of the Japanese fishing vessel Lucky Dragon, arid incidently, the ships of the task force. These incidents received considerable attention in other countries, and the exposures to the crew of the Lucky Dragon caused considerable furor in Japan. Prime Minister Nehru of India called for a rest moratorium. Fallout had become a matter of world concern.

The committee that reviewed the sheep deaths had made, in their report, various suggestions for ways to reduce fallout. In those days we, all of us, tended to subscribe to the catchy slogan, "The Solution to Pollution is Dilution." If you had a plant which was spewing out nasty things into a river, and things were getting unpleasant in the river, what you really needed was a bigger river. That way there would be more dilution and the problem would go away. Using that philosophy, if you're going to shoot off atomic devices, and they pollute things, the thing to do was to dilute the radioactivity more. One of the things you could do was to make the towers taller, and if that wasn't enough, you could put them up with balloons. There wasn't the concept of trying to control at the source.

Public concern continued to grow. The National Academy of Science undertook a radiation study at the request of the AEC. The JCAE and the Senate Armed Services Committee held hearings on fallout hazards. The Federation of American Scientists proposed that a United Nations commission assess the hazards of testing. Finally, in 1955, Commissioner Libby of the AEC actually released scientific data on fallout,

By 1957, some shots began to be done underground. Why? Well, if you look at the type of events they were, it's easy to believe that the self interest of the people doing shots led to this development. They were safety shots of zero yield, or a few tons at most, and they would tend to contaminate the Test Site. If you put them in a hole in the ground, plutonium wouldn't blow around and make operational problems. There was, however, the RANIER event at a yield of 1.7 kt, which was fired in a tunnel, and was contained, and which pointed in the direction of doing something with radioactive debris other than trying to dilute it. There were a few other underground events in 1958, and then there came the moratorium.

When testing was resumed in 1961, shots were done underground, except for those that weren't. We didn't know a great deal about the containment of events underground, and the first eleven we did either leaked or vented some fraction of the fission products. That didn't seem to be so bad when you considered the more than 100 megatons that the Soviets and the U.S. had or were in the process of detonating in the atmosphere. If you vented ten percent of a two kiloton shot, that seemed trivial compared to twenty or

11 ; • /

SSX

12 thirty megatons in the atmosphere.

Then in 1963 came the Partial Nuclear Test Ban Treaty, and with it came the TEP and categories. The TEP was supposed to "review data associated with proposed nuclear tests or demonstrations in terms of past experience, probable violations of the treaty provisions and current detection techniques." Presumably if someone invented a new detection technique that was ten times as sensitive as the one currently in use, you would try to release one tenth as much as you had previously thought acceptable. Category A, as it was then, said "An underground test which, on the basis of experience, should not vent significantly. It must be understood that even in this category unforseen conditions may develop which result in the release of detectable levels of radioactivity at the border." That means New York, or New Orleans, or Los Angeles, if the wind is blowing in that direction. The Planning Directive for 1964 said, "The emplacement and firing of devices will be designed to result in containment- in all cases where this requirement is not inconsistent with the technical objectives." So you certainly ought to contain all of the shots as long as it didn't bother you too much to do that.

We went on that way for a while, but there was one group of people who began to develop more rigorous standards for themselves. Like the people who first put safety shots in holes, this group began to apprectiate the fact that it was in their own best interests to try to understand containment. They worked in tunnels, they put in elaborate and expensive hardware, and they began to get tired of losing their experiments when the tunnel blew out in one way or another. DNA and Sandia people, associated with these tunnel events, began to think, perhaps not iihat they should understand containment, but ce. tainly that they should understand how to protect the experiments they wanted to do. Suddenly they were not an add-on anymore. They became an integral part of "my shot," and they have done a significant and impressive body of scientific work in the field of containment — they really don't like losing all their data.

The other thing tbit was happening was that the outside world was changing, but the rest of us were going along without much concern or awareness of that. In fact, we were shooting cratering shots up until the end of 1968. Then one day there was another fairly impressive cloud that appeared after a test. It was called BANEBERRY, and we were surprised when people said, "Stop what you're doing." We weren't allowed to shoot anymore? Just because we leaked a litle radioactivity, which we could certainly show you and convince you by logic and science wasn't enough to bother anybody? But nobody really wanted to listen to that, which seemed strange to us, as scientists. They seemed to be operating on some other basis, which they felt was sufficient for them, Six months later we were allowed to resume testing, but we were told flatly that we were not to leak anymore. And we haven't. Why not? What did we learn during those six months that made such a difference? That you ought to plug up the cables that often leaked, that you ought to put the stemming in a little differently, that you ought to pay a little more attention to where you located the emplacement hole? We didn't learn anything in those six months, except that it was important to stop leaking radioactivity. The implication to me is that we could have stopped a long time before. And maybe we would have if anybody had been listening to people who didn't understand and hadn't been told and whose concerns were usually dismissed as unfounded, and if anybody had realized t'uat the enormous bank of trust and respect

13 .'. -iiti'r it in t<> 1. .i|. 'I.! '! •.til! fll.-CM C

!• .if. :(,•• • n-i«<"! '••uki-mia thi»t

• •< 'l r.. i in-!..* ihi nurfncc f>fO«fi»iJj« 'Miti f..: n-it'iVMiriK !he

BANCBERRY Event December 1970

14 that had been reposed in the nuclear scientists after World War II had been spent. It seems to me that the people don't trust you> or me, anymore to tell them the truth, or Co deal properly with what they consider their interests to be if it inconveniences us, or to tell them what will or won't hurt them. I suggest that we will stay stopped from leaking things, regardless of all the logic and science which we might put forward to say, "Well, a little radiation won't hurt you. Trust me." They don't trust us.

To those of you who wish to know and to understand and to insure that we can deliver what we promise, it pains me to say, as I said in the beginning, that you tend to be add-ons, and that what militates most against the advancement of your science is your success. Containment is a science reluctantly supported, and I think only supported because we have been dragged to where we are. However, I think it would be wise for those who take success in a limited number of shots as a reason to question the need for any further understanding to remember the words of Albert Einstein, which went something like this, "No number of experiments can prove me right, and a single experiment can prove me wrong."

15 CEMENTS FOR LARGE DIAMETER NEVADA TEST SITE DRILL HOLES

Grant T. Bruesch Fenix & Scisson, Inc.

Emplacement hole casing cements have been counted on to help the stemming material and gas seal plugs divert the upward migration of post shot gases into the formation. This paper is a review of the material properties of cements currently in use and the properties of several potential candidates for new uses. Poten- tial candidates for surface hole lining without steel casing include: formed concrete, precast concrete, and shotcrete.

I. INTRODUCTION

Cementing of the large diameter drill holes at the Novaila Test Site (MTS) has been considered important to containment since the U-3cy hole, Pike, was used in 1964. Many different slurry types have been used during the 20 years that Fenix & Scisson, Inc. (F&S) has worked as A-E for drilling for the AEC, ERDA, and DOE. Various light weight flurries were used on Pahute Mesa during the 1960's. Emplacement holes were fully cased from working point ~o surface until 1969.

Since Baneberry, in December 1970, a description of the cementing of each large diameter emplacement hole has been presented by an F&S representative to che Contain- ment •..valuation Panel (CEP) for review. After changing over <:.~) partially cased drill holes in 1971, the cementing oi. surface casing and intermediate casing strings

17 changed very little until the end of 1980. Oil field type equipment and slurries were used on all jobs. Since the end of 1980, some surface casing jobs and one intermediate casing job used a cement and sand blend delivered and poured by standard transit mix trucks (Redi-Mix).

Since new variations in cements are being discussed, it is appropriate to review where we have been and to look at some of the cement candidates for future use.

The Laboratories will, of course, review and approve the characteristics of any new cement candidate, together with the formation characteristics at that specific site, before it is presented to the CEP.

II. CEMENT TYPES

A. Current Standard Cements

The types of cements listed here have been in common use since 1970 for cementing the steel casing annulus or for openhole plug-bark and drill-out operations for hole stabilization. Neither the many different slurries used during the 1960's when holes were being cased from bottom to top, nor the rock matching grouts of Water Ways Experiment Station (WES) design will be reviewed. Primary slurries used since 1970 include:

18 2 API Class "A" (common Portland Type I). "A" with CaCI- (to accelerate setting time). "A" with Gel (Bentonite for lighter weight and economy). Poz-"An (Fly Ash for lighter weight and economy)• Common Gypsum Cement (Casting Plaster MP for rapid setting and expansion). "A" - Gypsum Cement (for rapid setting and expansion). "Redi-Mix"4 (Portland aype II and sand for economy).

The term nRedi-Mix" designates cement slurries delivered to the location in transit mix trucks.

B. Possible Candidates for NTS Use

There are several cement candidates for use without steel casing to protect the drill hole condition. All of these have been used elsewhere for other purposes and could be adapted for use in large diameter drill holes, if proven to produce good results and be economical.

At present the following are being considered:

1. With steel casing

a. Variations of current slurry types. b. Current mixes with a layer. or layers of noncementitious materials.

19 2. Without steel casing

a. Formed concrete (using collapsible or slip forms).

b. Precast concrete (concrete cylinders in place of steel casing).

c. Shotcrete (sprayed on concrete).

d. Chemical grouts (solidifying in formation) .

. Different cement slurries and non-setting mixtures are being considered for use in the emplacement hole stemming column. These are being discussed in another paper, "LLNL STEM- RING PLUG ADVANCES," by W. E. Lowry, A- L. o

Lundberg, and L. I. Starrh.

III. DEVELOPMENT OF NTS CEMENTING PRACTICES

A. Surface Casing Cementing Large diameter surface casing and the accoiripanying annulus cementing procedures were developed gradually. They were intended, (1) to keep the loose shallow alluvial materials from falling in on the drilling assembly, (2) to prevent the caving of the alluvium back to ground surface, and (3) to prevent thi; flow of circulating medium back to ground surface outside of the surface casing.

20 The unconsolidated alluvium in the upper part of Yucca Flat holes frequently extends to a depth of 100 feet. Surface casings of 40 feet or less were used, at first, with only the lower part of the casing solidly cemented in the hole. Remedial cementing jobs were frequently required in Yucca Flat. In spite of this, holes were lost. The surface casing length was increased to the depth of 120 feet which was the practical depth limit of the auger and bucket rigs. There was a significant reduction in hole loss due to caving and washing out around the surface casing.

Surface casing lengths of 10 feet to 60 feet were used for Pahute Mesa and Rainier Mesa, depending on the thickness of alluvium or friable tuff that could be drilled with an auger rig. During the 1960's different types of problems existed for Pahute Mesa and Rainier Mesa than for Yucca Flat. The well indurated tuff and rhyolite on the Mesas did not tend to cave, but open fractures allowed the low pressure air being used in direct 9 circulation to escape to the formation. If fractured tuff was near the surface, the low pressure air (3 psi to 5 psi) came to the ground surface outside the surface casing. After a heavy rain steady streams of bubbles could be seen at various locations around the rig. Currently, 60 feet of surface casing is preferred for most Pahute Mesa holes. This allows sufficient drill collars to be run in the hole at the start of the rotary rig drilling, to efficiently drill the hard zones.

21 A dual string reverse circulation system is being used currently for all large diameter drilling at NTS. This reduces the problem of washing out formation and cement from around the surface casing, but in Yucca Flat caving to the surface still occurs at some locations below 120 feet. At other locations less than 120 feet of surface casing would be raquired to stabilize the hole for drilling with the current

The casing and cements have at times helped the stemming and alluvium to inhibit the flow of radioactive gas to ground surface. Prior to Baneberry and the use of CTE plugs and gas blocks, cement plugs (usually gypsum cement) vere occasionally placed in the top of the casing of LASL emplacement holes after an event to check the flow of gas up through gravel stemming.

Since Baneberry, there have been a few occasions when radioactive gases have reached the bottom of the surface casing at approximately 120 feet without being detected at ground surface. Gas with radioactivity reached the bottom of the surface casing at Riola and was detected at ground surface outside the surface casing; it may have had one of its routes through a fracture in the cement outside the surface casing.

The surface casing and cement around the outside of this casing are considered by some to be significant containment features of the current stemming plan for tests in Yucca Flat and for Pahute Mesa tests that have alluvium or unfractured tuff around

22 the surface casing. Some of the physical properties of these cements will be shown in tables in Part III.

B. Stabilization of Zones Below Surface Casing Depth

Caving in the alluvium and the friable bedded tuff zones found below surface casing depth has been a 12 problem for many years= The change to dual string reverse circulation, standardization of approximately 120 feet as surface casing depth, and good cementing practices eliminated the washing out and initiation of caving in the unconsolidated material in the upper 100 feet of the alluvium. At greater depths, lenses of loose materials continued to cajse caving and hole loss. "Geologic Investigations Into Drill-Hole Sloughing Problems, Nevada Test Site," by S. L. Drellack, Jr., W. J. Davis, J. I.. GonzaleSf and W. L. Hawkins will address this problem.

In the past, these deeper zones have been stabilized, when possible, either by intermediate casing through the potential caving zone before a serious problem developed, or after a problem developed, by plugging back through the enlarged interval with cement and drilling back through the cement plug. A series of plug-back and drill-out operations have been required at times.1 4 The casing and/or cement used for stabilization have been considered significant for containment evaluation.

23 C Liner Cementing for Water Shut Off

During the period up through 1969, emplacement holes were cased and cemented from total depth to the surface using a variety of cement slurries. By 1971, an almost complete changeover was made to uncased holes below surface casing depth except for hole stabilization for water shut off.

Steel casing liners were used to line the hole from above the static water level to bslow the working point to provide a dry emplacement hole. The liner annulc.s was filled with cement.

IV. PHYSICAL CHARACTERISTICS

A. Plastic Data

The physical characteristics typically checked for MTS slurries without sand or aggregate are density and se^ time 01 surface samples. The slump is checked for slurries with sand and/or aggregate. Weight of dry materials, volume of water used, and density are monitored for quality control and quality assurance.

B. Hardened Cement Characteristics

Trie physical properties of hardened cement are not routinely checked for each job. Studies have been made of the physical properties o" NTS cement types. Halliburton Company cement cic.ta1 2 is shown is shown in Table 1. Data from recent specific WPS job oriented studies by WES at Vickfiburg, Missis- sippi and by Holmes & Narver (H&N) Material Test TABLE 1 STANDARD NTS CEMENTS DATA FROM HALLIBURTON CEMENTING TABLES PSl UNCONFINED PSl SHEAR BOND SLURRY TYPE BULK COMPRESSIVE STRENGTH TO CASING DENSITY .25 DAY 1 DAY 3 DAY 7 DAY 1 DAY 3 DAY 7 DAY

"A" 60°F 1.87 SOFT 330 1620 4250 80°F 1.87 60 1220 3200 6060

"A" 2% CACL 2 6O°F 1.87 190 1450 80°F 1.87 685 3125 "A" 2%GEL 60°F 1.76 SOFT 365 80°F 1.76 70 1090 110 100 200 50-50 POZ-"A" 2% GEL 60°F 1.70 SOFT 100 375 1325 8O°F 1.70 SOFT 350 880 1620 80 130 200 75-25 POZ-"A" 2% GEL 60°F 1.63 SOFT SOFT 120 80°F 1.63 SOFT 50 170 Lab at MTS are shown in Table 2. Data for permeability, tensile strength, and heat of hydration are limited. The following physical properties are especially important to some MTS uses:

Set Tima Unconfined Compres- Bulk Density sive Strength Permeability to Shear Bonding to Gas Casing Heat of Hydration Shear Bonding to Formation

C. Physical Characteristics Related to Use

1. Current Cements for Steel Surface Casing.

The surface casing cement is intended to hold the casing and shallow formation in place during the drilling, logging, side wall sampling, emplacement, and stemming operations. In some cases it is required to support a strongback with up to 600,000 jpounds of load plus the weight of the casing.

Shear bonding required to support a 697,500 pound load on ten feet of the cemented annulus of 98 inch I.D. - 3/4 inch wall casing would be 18.6 psi. This would include 600,000 pounds for the strongback and 97,500 pounds for 120 feet of 98 inch - 3/4 inch wall casing.

The material in the annular space between the outside of the casing and the formation has

26 TABLE 2 STANDARD NTS CEMENTS AND VARIATIONS *DATA FROM WES & H&N LAB TESTS

*#PSI UNCONFINED HEAT OF % EXPANSION SLURRY TYPE BULK COMPRESSIVE STRENGTH HYDRATION 3"x3"x10" MAX. INCREASE DENSITY .23 DAY 1 DAY 3 DAY 7 DAY 28 DAY 3 DAY 7 DAY TEMP. TEMP. WES LAB TESTS

75-25 "A" GYP. CEM. 1.81 1790 .020 .037 "REDI-MIX" 2.16 2450 2700 .001 -.001

H&.N LAB TESTS

"A11 2% CACL^ 1.81 5200 6300 212 140 -.016 -.027 75-25 POZ "A" 2% GEL, 1.63 >40 896 3000 4% CACL2

50-50 POZ "A" 2% GEL 1.70 150 133 61 "REDI-MIX" 260 109 37

'•VARIATIONS OF 29% NOTED BY HALLIBURTON IN 1 DAY TESTS ON API CLASS A ("A") CEMENTS FROM DirrERcNT MANUFACTURERS. ASTM C 109-11 NOTED THE MULTILABORATORY COEFFICIENT OF VARIATION WAS 7.3% AND THAT THE SINGLE LABORATORY COEFFICIENT OF VARIATION WAS 3.8%. •TESTS SAMPLES KEPT 73* F J3°AND AT ATMOSPHERIC PRESSURE. been relied on to be less permeable than the formation and to remain less permeable after a low yield test has caused stemming material to fall out up to the bottom of the surface casing.

A review of physical properties and experience indicates that the cements used have exceeded these requirements when intact, though there are significant variations in properties.

Current standard cements for surface casing strings are as follows:

"A" with 2% CaCl2 75-25 "A" with gypsum cement "Redi-Mix" 50-50 Poz-"A" plus 2% gel

The recent shear bonding with iron tests conducted by H&N, together with studies on shear bonding with casing and formation by G. W. Evans and L. G. Carter of Halliburton, published in 1962,' indicate that the shear bonding of cements used for the 120 feet surface casing greatly exceeds the required strength.

Holmes & Narver, Inc. data in Table 1 for "A" 2 percent gel and 50-50 Poz-"A" 2 percent gel slurries show one day bond strength to casing of 80-110 psi and seven day bond strength of 200 psi for both slurries.

28 The Halliburton studies were made on a variety of cements including "A", 50-50 Poz~"An plus 2 percent gel, and "A" plus 4-12 percent gel. They first did bonding studies using API class A Cement ("A") with casing of various finishes including new, new sandblasted, used wire brushed, used slightly rusty, and used rusted. The shear bonding to casing after 48 hours ranged from a low of 46 psi on new pipe, wet with water base drilling mud, to a high of 422 psi for "A" cement on rusted pipe.

Figure 1 shows the bonding of cement slurries to dry used pipe after curing for 24 hours at 80°F and atmospheric pressure. Both shear and hydraulic bonding were tested. The configuration used is shown in Figure 1A. The shear bonding ranged from a high of above 200 psi for Class A cement down to a low of 65 psi for Class A with 12 percent bentonite and 3 percent lignin retarder. The hydraulic bonding was higher ranging up near the compressive strength of the cements.

Figure 2 shows the test results of cement slurry bonding to Berea Sandstone and Indiana Limestone. Test samples were kept at 80° F and atmospheric pressure until set then pressured to 1600 psi for 24 hours. The best bonding is with sandstone which has a permeability of 200 millidarcies. It ranges from above 300 psi with Class A cement down to 90 psi on the Class A with 12 percent bentonite and 3 percent lignin retarder. The shear

29 o o o o" o

• STRENGT H z c> £ 1] VI L> i HI KE Y < i tl o i 8 r o LLJ q Q_ d f Q_ o I— fl I— z ±a. LLJ

LLJ r O —I i o i m j y LL O

| UJ i ! i O £ I— r—n ! t — - — H — — i a: n LU - — o o r- (Z o 1 I o — i— Q: J 1 CL ! i i CD 1 I

i I z Q to I/) z LU z LU LU O Lu 2 2 cc O LU LU LU < 2 LU LU LU 2 CQ j ca u N N o a. o ^ CL o CM CL CM

30 CEMENT SLURRY

FORCE CEMENT SLURRY \ \\\\\1L\\\W\ ' \ ' - • Y \ ' Y "CEMENT '' Y SLURRY

-MUD CAKE

FORMATION CORE

i\\l I \\l 1\\

PRESSURE —i ^—PRESSURE

SHEAR BOND TEST HYDRAULIC BOND TEST TO PIPE HYDRAULIC BOND TEST TO FORMATION

FIG. 1A-BONDING TESTS CONFIGURATION FlS

API CLASS A CEMENT

PO2ZOLAN X CEMENT 2% BENTONITE

POZZOLAN Y CEMENT 2% BENTONITE API CLASS A CEMENT 1 2% BENTONITE 3% LIGNIN RETARDER

API CLASS A CEMENT 1% LOW-FLUID-LOSS

KEY LATEX CEMENT SHEAR BOND COMPRESS IVE STRENGTH

HYDRAULIC BOND RESIN CEMENT

100 1,000 10.000 100,000 SHEAR BOND, COMPRESSIVE 1 HYDRAULIC BOND STRENGTH-PSI FIG. 2-BONDING PROPERTIES OF CEMENT TO FORMATION CEMENT NOT SQUEEZED-WALLS CLEANED G. W. EVANS 8. L.O. CART!"" Hi.LL'PUTCIJ OIL VVCLI. C!tS::"UT|t<3 COMPANY bonding to limestone, which has permeability on the order of one millidarcy, ranged from above 200 psi down to 60 psi. The hydraulic bonding was highest on a resin cement for both sandstone at 470 psi and limestone at 195 psi. The lcwest for limestone was 60 psi with a latex cement, and the lowest for sandstone was 150 psi with both of the pozzolan slurries and with the Class A 12 percent bentonite slurry.

One series of tests was made with mud cake on the cores, and another series was made squeeze cementing at 100 psi pressure. The bonding to formation with mud cake was on the order of 40 psi for a cement with 12 percent gel and approached 100 psi for the Class A slurry without gel. Wheii the cement was squggzed against dry cores until set, the shear bond to formation approached or exceeded the unconfined compressive strength.

The permeabilities of the cements used are much lower than the formation unless the cements are fractured.. An H&N Materials Test

Lab test on "A" 2 percent CaCl2 showed the cement samples permeability to gas was 25 millidarcies. Other tests of air dried samples have been higher, but all are in the millidarcy range, compared to tests of 2.5 to 13 darcies for aluvium and some bedded tuff. Surface casing cements that have been challanged by upward migration of radioactive gases have not appeared to furnish a faster route to ground surface than the alluvium unless it did so on Riola.

33 According to studies by the American Concrete Institute (ACI),1 8 prolonged proper curing reduces the permeability; rapid drying out increases it through the development of fine shrinkage cracks. The addition of fly ash or increasing cement-to-water ratio reduces permeability according to ACI studies.

The shear bonding and the permeability of the contact between the cements and the casing are satisfactory, according to a combination of laboratory tests and field observations. When shrinkage occurs, the cement tightens around the outside of the casing. The shear bonding and permeability of the contact with the formation also seem to be satisfactory. This contact is not a straight line because of the irregularity of the bore hole wall. The cement slurries without slump fill the small irregularities with fluid slurry before setting up.

Well blended slurries with proper amounts of water have been observed without megascopic cracking at the surface. Spots of poor cement with too much water have been noted to form small cracks within the cement and along the celler pipe-cement contact. For cements that shrink with age, very fine short cracks can be expected to eventually form in the weaker spots including those, if present, near the formation-cement interface. These fine short cracks would not be expected to form a continuous path through good quality cement slurry zones that cured slowly underground.

34 According to cement bonding toformation test stronger contacts would be expected for the very permeable Yucca Flat alluvium than for a densely welded tuff. A thicker cepent filter cake with reduced water content produces a stronger contact zone in the highly permeable intervals.

Some of the hardened cement types expand slightly if unconfined after the initial hardening and should not form shrinkage cracks. Good quality gypsum cement and 75-25 "A" gypsum cement are of the expanding type.

Good quality REECo Type II and sand slurries have some slump and would not be expected to completely fill all bore hole wall irregularities. The volume used and the field observations indicate that most of the irregularities are filled, and only short discontinuous paths for gas migration would be expected.

2. New Possibilities for Surface Hole Could Include:

a. Steel casing and the above cements, or minor variations, with layers of very fine grained dry materials or unset soft plug material.

b. Shorter surface casing.

c. Formed concrete.

d. Precast concrete casing with cement in the annulus.

35 e. Shotcrete.

It would be feasible to use some noncementitious material in part of the annular space for surface casing longer than 40 feet if the material was less permeable to gas and water than the formation and if it did not leave an open microannulus along the side of the casing for water or gas migration. Because of the known variations in field emplaced cements and the fact that the formation and/or cement breaks away from the bottom of the surface casing at times, it would be prudent to- cement a minimum of 40 feet of the annul us with a cement having lab tested shear bonding of 40 psi or greater.

Bentonite or fly ash would be possibilities for layers of low permeability dry materials in the annulus. Layers of wet materials would have to be short enough in height not to collapse the casing. The 3/4 inch wall thickness 98 inches internal diamater (I.D.) casing has a collapse factor of 26 psi. A safety factor of two would allow only 13 psi of hydrostatic head to be used. For a very low yield test, these materials should stay in place and remain less permeaole than the formation.

3. Hole Stabilization Cements

The primary purpose of this treatment is to keep the formation in place throughout the drilling, logging, emplacement, and stemming operations. The strength, flexibility, and permeability may be significant considerations for these materials.

36 if a steel intermediate casing is used, the standard cements have adequate physical characteristics for this interval. tfl-»= steel irt«armecliate casing without perforations does not allow diversion of gases to the formation above the bottom of the casing. It does furnish a solid container for stemming platform plugs or gas seal plugs.

Current standard cements for cementing the annul us of che intermediate casing are:

"A" with 2% CaCl2 75-25 "A" - Gypsum Cement "Redi-Mix" 8 Sk mix Portland Type II and future possibilities for cementing the casing swulus includei

Poz - "A" slurries Standard slurries with intermittent zones of dry clastic materials Concrete casing with cement in the annulus

For stabilization without steel casing, the previous practice of plugging back with cement through enlarged and caving intervals was successful in most attempts. It was very expensive in some cases due to the cost of rig time and materials. New methods will have to be more economical in order to be successful.

Historically, the "A" 2-3 percent CaCl2 plus nylon fibers seemed to be the most successful.

37 The nylon fibers aided the cement plug in hanging together during the drilling out process. The higher ratios of shear strength and tensile strength to compressive strength seemed to be helpful. Figure 3 portrays the stabilization of U-2eo emplacement hole. The upper zone from 1625 feet to 1660 feet was stablized before the lower zone from 1690 feet to 1740 feet caved. Some holes required as many as seven or eight plug-back and drill-out operations and more than 30,000 cubic feet of cement slurry for stabilization.

Future possibilities for stabilization without casing include:

Shotnrete Formed concrete Grouts

Shotcrete should be able to stabilize a zone that is only slightly enlarged but has potential for caving later during drilling, emplacement, or stemming operations. Opera- tional problems will need to be worked out before this method is economical. A field test at U-7bk demonstrated that strong concrete could be successfully sprayed on the walls of a large diameter hole down to a depth 19 of 110 feet. The before and after caliper log configurations in Figure 4 show the thickness or the shotcrete. This was a research project and the first downhole operation for the remote control application of shotcrete. With additional experience more shotcrete can be placed in the enlarged intervals and less

38 U-2EO HOLE STABILIZATION NEAT CEMENT .2V. CACL2 & 1/4 LB. NYLON FIBER / FT 3

80" 00" 100" 110" 120" 130" 140" 15501

CALIPER

15921 - 1600' PLUG NO. 1 USED 8500 FT^ DRILLED OUT 55O0 FT 3 LEAVING 3000 FT3 RUN 1

PLUG NO. 1 1650'

CALIPER IRUN 3 1700'

1725' - PLUG NO. 2 670O FT3 USED 3 DRILLED OUT 4900 FT LEAVING 1600 FT3 1750'

1785' -

1800' DRILLED TO 183V. HOLE CAVED. PLUG NO. 1. 1725' TO 15921. DRILLED OUT PLUG AND FILL TO 183 1', HOLE CAVED. SET PLUG NO. 2. 1785' TO 167CT. DRILLED OUT PLUG AND FILL.

FIGURE 3

39 F,S U-7BK

100" 1 10" 120" 130" 140"

10'

RUN NO. 1

50'

CALIPER RUN

100'

SHOTCRETE MIX POUNDS 8.5 SACK MIX WEIGHT VOLUME FT3 CEMENT TYPE II 799.0 4,077 SAND NO. 4 2,234.6 14,655 AGGREGATE 3/8" 729.6 4,641 WATER 175.6 2,815 AIR 810

TOTAL 26,998

FIGURE 4

40 on the bit size hcle. Nine panels of shot - crete were tested. Their seven day unconf inecl compressive .strengths ranged from 5898 psi to 7722 psi. Shotcrete should be solid enough to withstand trip danage from a drill string if care was taken when the drill collars, stabi- lizers, and bit were at shotcrete depth.

Another prospect is formed concrete which would be poured into collapsible or slip forms. The si;re'igths would be similar to shotcrete but wir.r. a smooth inner wall.

Both the she terete and formed concrete wuuld have to be placed above fluid level in a r.one that was stable at the time of placer-Tit. They would, in r.iost cases, reduce the siz of the hole.

One possibility for the replacement of :eel casing for any of its uses at surface, i ter- mediate, or liner depths would be precast concrete cylinder.1.. They have been product , and tested to failure at external hydrot:atic heads of 300 p:;.l to 1100 psi. These would be cemented in plaice like the steel casing. They could be run below fluid level. The strengths are adequate £ ir any of these uses, and they are lower in price than steel casing. Uncon- fined compressive strengths were consistently above 11,000 psi. Some operational procedures need to be formulated before this system can be fully evaluated. 3. Formation Stabilization With Mud Additives or Chemical Grouts

An ideal way to stabilize potential caving zones in the alluvium would be to have some additive in the drilling, mud that would seep

into these permeable intervals and set upf thus forming a wall of stronger formation. Several materials have been discussed as possibilities, and one has been tried with inconclusive results.

Surfseal, a paraffin wax based additive, was used in emplacement hole U-lOas to treat a caving interval just below the surface casing. The hole was successfully completed to a depth of 1200 feet. Laboratory tests on simulated alluvium and a review of nearby holes did not prove a significant improvement of formation stability using Surfseal. Nearby holes had caved to some degree and stabilized without treatment.

Sodium silicate grouts might be used to treat caving or potential caving zones at a cost that is comparable to or less expensive than cements. The stabilized formation would be expected to have unconfined compressive strengths on the order of 25 psi to 350 psi2 0 and be lower in permeability than the formation. It might be possible to develop its use as a mud additive. The treated hole permeability could be studied for comparison with stemming material and plug permeability. Other more expensive grouts are available that have been used to stabilize soils. One may be found in the future that will be easier and more economical to use. Figure 5 shows a possible method of spot treating with chemical grouts after drilling through a potential caving zone.

4. Liner Cements

Liner cements have been considered important primarily to hold the liner in place. The standard slurry for cementing these liners has

been "A" with 2-4 percent CaCl2 filling most of the annular space from the bottom of the liner up to near the top of the liner which typically extends up to 50 feet above the static water level.

Future changes may include filling half, or less, of the liner annular space and the use of weaker cements.

43 3 SPOT GROUTING UNDER © DRILLING FLUID

GROUT TUBING GROUND LEVEL

"NORMAL" ALLUVIUM

DRILLING FLUir

\.i POTENTIAL CAVING ZONE 7y>:'

"NORMAL" ALLUVIUM

FIGURE 5 V. REFERENCES

1. G. C. Mathis, "Evaluation of Materials and Techniques for Cementing Casing in Large Diametar Holes," Fenix & Scisson, Inc. NVO-38-17 (April 1969).

2. "Halliburton Cementing Tables," Halliburton Company, Duncan, OK (1979).

3. "Testing Gypsum Cements and Plasters," United States Gypsum, Chicago, IL, Bulletin No. TAC-206 (1983).

4. "Surface Conductor Casing Mixture REECo-1," Watenvays Experiment Station, vicksburg, MS, Submission No. 3 (May 1981),

5. F. J. Solaegui- "Collapsible Forms," REECo Design Drawings, Second Meeting on NTS Drill- ing Research by Test Group Directors Ad Hoc Subcommittee (TGDAHS) Las Vegas, Fenix & Scis- son, Inc. TES 3404 (January 1983).

6. R. Schalge, "Precast Concrete Cylindrical Liners," Fourth Meeting NTS Drilling Research TGDAHS, Mercury, NV, Fenix & Scisson, Inc. Report TES 3492 (March 1983).

7. F. E. Valencia, J. H. Pye, and C. D. Breeds, "Fast Automatic Shotcrete Technique," International Ground Support Systems, Inc, Denver, CO (1982). 8. W. E. Lowry, A. L. Lundberg, and L. I. Starrh, "LLNL Stemming Plug Advances," Second Contain- ment Symposium, Albuquerque, NM (August 1983).

9. R. J. Schoeppel, W. A. dandier, T. B. Dillinger, "Wall Stabilization of Large Diameter Drilled Holes," Fenix & Scisson, Inc., Report NVO-38-4 (1966).

10. O» C. Gilliland, "Air-Assist, Reverse-Circula- tion Systems," Fenix & Scisson, Inc., Report NVO-38-12 (1967).

11. E. C. Woodward, Jr., "Riola Release Report," Lawrence Livermore Laboratory Report In Press (1983).

12. G. T. Bruesch and J. R. Gesin, "Emplacement Hole Caving and Formation Stabilization Yucca Flat," Fenix & Seisin, Inc. Report (January 1971).

13. S. L. Drellack, Jr., W. J. Davies, J. L. Gonzales, Fenix & Scisson, and W. L. Hawkins, Los Alamos National Laboratories, "Geologic Investigation into D»-ill Hole Sloughing Problems," Nevada Test Site, Second Contain- ment Symposium, Albuquerque, NM (August 1983).

14. R. G. Miller, "Case Histories of Cement Plug- backs and Drillouts," Lawrence Livermore National Laboratory, TCDAHS Second T4eeting (January 1983).

46 15. "Telecommunication from E. - J. Sowder, LANL" (1981).

16. G. W. Evans and L. G. Carter, "Bonding Studies of Cementing Compositions to Pipe and Forma- tions," Halliburton Company, Duncan, OK; Pre- sented Southwest District, API Division of Production (March 1962).

17. D. F. Snoeberger, C. J. Morris, and G. A. Morris, "Field Measurements of Permeabilities In NTS Area 9," Lawrence Livermore National Laboratory Report UCID-1589 (August 1971).

18. A. M. Neville, "Hardened Concrete Physical and Mechanical Aspects," American Concrete Insti- tute and Iowa State University Press (1971).

19. T. J. Clark, "U-7bk Shotcrete Experiment," Fenix & Scission, Inc., DR-1665 (June 1983).

20. "SIROC Soil Stabilizer," Soiltech Department, Raymond International, Inc., NY (1983).

47 r

L'CRL- 8942] REV. I PREPRINT

LLML STEMMING PLUC ADVANCES

W. E. Lowry A. L. Lundberg L. I. Starrh Mechanical Engineering Department Lawrence Livernore National Laboratory Livernore, California

DNA 2nd Containment Symposium Albuquerque, New Mexico

July 29, 1983

I his is a preprint of a paper intended foi publication in a journal or proceedings, since chdn^es mat he made before publication, this preprint is made available v-Ub (he understanding Mat it will not he cited or reproduced without the permission of Me aulhoi ABSTRACT

In early 1982 the Lawrence Livermore National Laboratory (LLNL) conducted its first examination of in-place rigid Coal Tar Epoxy (CTE) stemming plug material. That evaluation, and subsequent downhole samplings, cast uncertainties on the emplaced strength characteristics of CTE plugs. A number of replacement plug materials were evaluated, and a two-component epoxy system was chosen for in-piace testing in the fall of 1982. Showing high shear, flexural, and compressive strengths, and requiring relatively simple fielding techniques, this material (denoted Two Part Epoxy, ur TPE) was specified for its first application on the MANTECA event in December 1982. This first emplacement proved successful, and was followed by TPE use on the CHEEDAM event in February 1983. Details of the structural and thermal evaluations of the epoxies and prospective cement-based materials are presented.

INTRODUCTION

The Lawrence Livermore National Laboratory uses epoxy/aggregate composites as plugs, or discrete high strength components in the column of material placed around and above the experimental package in an underground nuclear test. These plugs serve several functions, but the most significant roles are as stemming platforms and gas flow restrictions. The stemming platform is expected to support the granular material (stemming) above the plug if the stemming below falls away. As an impediment to gas flow, the plug is meant to provide a low permeability restriction in the stemming column to inhibit gas flow through the cable bundle and up the hole wall surface. Additionally, at times, LLNL uses a formation-coupling plug to secure the emplacement pipe to the hole wall. This forces a buckling failure below, preventing damage to experimental hardware above. Normally, this plug is not considered to have a significant role in containment.

A typical LLNL stemming design has two TPE epoxy/aggregate plugs (see Figure 1). One is located at the bottom of the surface casing, with approximately ten feet in the casing and three feet below. The portion of the plug below the casing is meant to provide a seal to resist gas

50 CABLES

100 " TPE PLUG TOPPED WITH LIQUID PLUG

200 - CABLE FANDUT CTYP)

300 - TPE STEMMING PLATFORM CSQMETIMES TOPPED WITH LIUUID PLUG)

400 - COARSE

500 -

600 -

700 - FINES

800 -

900 - CABLE GAS BLOCK CTYP)

1000 - MAGNETITE

PROBERTITE 1089 WORK POINT: 1070 ft

Figure 1. Typical current LLNL stemming plan.

51 flow outside the casing (the ultimate release path on the RIOLA event.) The section of the plug in the casing is to remain integral in the casing through slapdown acceleration loadings. The lower plug, or stemming platform, is expected to survive the initial ground shock loading and support the stemming above if either subsurface collapse or stemming fall removes the stemming below. The stemming platform is usually keyed into the formation with intentionally cut washouts.

Coal Tar Epoxy (CTE) has been used by LLNL for thirteen year? in a variety of formulations as stemming plugs. The fielding of these plugs required the artful proportioning and mixing of as many as six different components in concrete mix trucks and then dispensing the material either directly into the hole with the mixed aggregate or pumping the liquid through a pipe to the plug elevation and dumping the gravel into the hole simultaneously. While the CTE plugs in general appeared to perform as expected, their erratic thermal performance and occasional disfunction (e.g., evidence of plug slippage in the surface casing) caused concern that, because we rarely knew if a plug was challenged by radiation in the hole, these features may not be performing as well as desired.

The RIOLA leakage of 1981, and subsequent investigations into the cause of the release, indicated that the stemming platform in that emplacement hole had in fact failed and was no longer there.

RIOLA stimulated LLNL (and LANL) to carry out experiments to evaluate the characteristics of the CTE material as placed in a deep location., The intense scrutiny, including tests of recovered plugs, identified real problems with procedures, quality control, and quality assurance practices, and highlighted some fundamental problems and pitfalls in the basic material In addition there were growing concerns for health hazards with some components of CTE. LLNL at this time chose to replace CTE with an improved material in order to eliminate the problems that had been identified. A discussion of the tests done by LLNL on CTE is -included in Appendix A.

52 THE SEARCH FOR A NEW MATERIAL

A small effort had been directed toward a new material during the preceding couple of years, encouraged by the potential for lower cost and to reduce the health hazards mentioned above. One was a polyester concept that was judged to require a great deal of development work, particularly for deep-hole emplacements of large quantities. The other, which can be called the origin of the material now in use, focused on modern epoxy materials being brought into commercial use for coating applications. These seemed to hav^ some desirable features, such as a reduced health hazard, the capability to cure and bond in moist environments, and simplified fielding procedures.

The general requirements were easily defined at that time. Already established was a minimum strength goal which was clearly feasible in the epoxy-gravel materials. Field simplicity was a strong requirement and a lesser problem in health hazards was desired. The maximum temperature in curing was to be less than 150° F to prevent damage to downhole cables. Due to the observations of the recovered plug materials, it was decided at the start that the plastic and the gravel would be completely mixed before sending the materials into the hole. Emplacement by means of a small diameter pipe was seen as a desirable feature of the material.

Laboratory tests were done on small batches until a good candidate formulation was identified. At that point, plans were made to go to the field for emplacement of a test plug which was to be recovered and examined in detail. That material is the same formulation that LLNL now uses at NTS--TPE, for Two Fart Epoxy. It is known by the Celanese Corporation as RDX-60366 and RDX-60367, an epoxy and its hardener, respectively. Preparation in the field involves mixing equal volumes of each of the two components (a procedure that works very well due to its obvious simplicity), combining with sand and gravel, mixing and dispensing into the emplacement hole.

53 The details of the testing program are discussed in the following sections and in Appendix 3. As things have turned out only one of the early objectives was not achieved. Emplacement through a small, pipe is not currently feasible. In all other respects the material is working out well for field operations, and has instilled enough confidence that LLNL is using shorter and fewer plugs than in the period following RIOLA.

TWO PART EPOXY MATERIAL EVALUATION

Strength of the material at shot time is of major importance; this can be as soon as 30 hours after the last plug is completed. Factors to consider are the degree of cure, actual temperature, laboratory measurement of strength and the degradation of strength in emplacement. TPE requires 24 to 48 hours to reach the fully-cured condition when mixed with the gravel to form a plug composite, and the temperature at the contact with the surface casing is the dominant interest. The strength of the epoxy is strongly affected by elevated temperatures.

After the Celanese material was selected by LLNL, two test plugs were poured and retrieved for strength determination in September, 1982. The first was poured directly from the surface into an 88" diameter can, which was suspended at the 1000-ft. depth'in a 96" hole (U2fe). The second was emplaced in a 48" can at 600 ft. by pouring the epoxy/aggregate mixture through a 4.5" pipe to the plug elevation. The epoxy/aggregate ratio was the same as had bee~ used in CTE emplacements in the past, filling the voids in quartz gravel (approximately 40% void fraction with the epoxy.) In volume parts, then, the mix proportions were one part gravel to .4 parts epoxy (giving a total volume equal to the volume of the gravel.) The results of the structural evaluations on the retrieved material are shown in Table 1.

i'he first plug as it is being separated from the can is shown in Figure 2. AVG. STRENGTH STAND. DEV. TEST NO. SAMPLES (psi) (psi)

Flgxural 4 1001 92 (dumped into hole) Flexural 4 967 184 (Poured through pipe) Flexural 3 1455 163 (QA results) Compression 1 1334 — (. 0001/s strain rate) Compression 1 1276 (. 05/s strain rate) Compression 1 1972 — (. 1/s strain rate)

Table 1. Two Part Epoxy retrieved plug results.

Two issues are worth noting with the strength.test results. The first is that the downhole flexural strength is approximately 70% of that expected from the laboratory and quality assurance tests (the CiaJ T:r Epoxy downhole plug material proved to be only 25% to 30% of its QA ^ Length). The second point is that emplacement through the 4.5" pipo had no significant effect on the strength of the downhole material, and visual examination of the retrieved plugs gave no evidence of foreign material intrusion or layering. In general, for TPE and CTE plugs mixed completely at the surface and dumped into the hole, LLNL has no evidence that this emplacement technique grossly affects the homogeneity of the plugs.

Once the plug tests showed positive results, it was decided to proceed with the field emplacement of the new epoxy on the MANTECA event, to be stemmed in December, 1982. Plans were to place a deep plug (375' down) through a small pipe, about 4.5" diameter, then later place the plug in the surface casing by dropping it directly from the surface. 55 Figure 2. The first U2fe test plug being extracted.

56 Conducti si ty: Temp. (F) Cond. (Btu/hr-ft-F)

UnfillQd TPE: 86 . 125 138 . 123 188 . 126

TPE Fi .led w/Gravel: 86 . 849 135 .826 185 .794

Specific Heat (TPE Only): .592 e 176 F

SpGcific Heat of Ovgrton Sand: .200 1 176 F

Specific Hgat of Quartz Itravel: .205 i 176

Epoxy heat of reaction: 113 Btu/lb

Table 2. Thermal properties of TPE and aggregate.

As discussed above, the test of putting mixed material down through the pipe was successful. In the MANTtCA operation it did not work out at all. The esrly test was a small batch and was noticeably aerated by the mixing process. The MANTECA plans reverted to the straight dumping process as had been done on prior events with CTE, and as had been done on one of the TPE. tests. No problems were observed and this has become the LLNL method of placing TPE.

The testing and thermal property measurements that followed indicated that earlier attempts at measuring the temperature rise were probably not of sufficient size or of the appropriate configuration to create an adiabatic environment, < condition which the center of the plug represents. The thermal characteristics of the TPE and aggregate material are shown in Table 2. With this information, the temperature rise seen on the MANTECA plugs was in fact consistent with the properties of the components. 57 Q) i_ D a ai Q. CHEEDAM E Q)

6 7 8 9 10 n 12 13 14 15 Time (Hrs)

Figure 3. Temperature responses showing the influence of the aggregate changes after MANTECA. To reduce the temperature rise to a more acceptable level, a number of modified aggregate formulations were tested. The regular quartz gravel was supplemented with silica sand (Overton sand) in different ratios, and sand alone was t~~ted as an aggregate. The epoxy/sand formulation >"RS ruled out because, while it poured easily through the emplacement pipe, it exhibited unacceptable shrinkage during the cure process. The sand/gravel aggregate formulation that was chosen for further testing was a combination of 3 parts quartz gravel to 1 part sand, combined with .75 parts epoxy (to make a total volume equal to the volume of the gravel). With this formulation, the expected temperature rise was 51 F°, compared to the 90 F° expected (and measured) on the MANTECA formulation. No changes in the epoxy itself were made.

This fix was verified with a test plug pour in U9y24 in February, 1983, and subsequently fielded on CHEEDAM and later events. Figure 3 Illustrates the temperature history of the two epoxy/aggregate formulations, comparing the MANTECA and CHEEDAM temperature resporees. Emplacements since CHEEDAM have followed that general 50 F° temperature

rise. 58 Full Curg Strength of 3: 1: .75 Mix of Gravel. Sand, TPE: (5 day cure § ambient, 20 hrs. @ temp. )

Temp. CF) Strsnqth (psi) CompressivQ (Meas.) 80 1260 100 580 120 350 140 210 160 160

Direct shear: 80 630 (. 5 x compr. str. ) 100 490 120 175 140 105 160 80 Degree of Cure: Insulated compressive test samples indicated approximately 60% of full cure strength at 24 hrs. and 100% at 48 hours (referenced to Full cure at temp.)

Table 3. Mechanical properties of the TPE/aggregate mix as functions of temperature.

Pertinent structural properties for the modified epoxy/aggregate formulation are provided in Table 3. 3y adding sand to the quartz gravel, it has been possible to reduce substantially the epoxy required to bond the aggregate together, reducing the temperature rise by 33%, reducing costs of emplacement, and maintaining the high structural properties. This modification, however, has made the mixture sufficiently less liquid to rule out any possibility of emplacing it through a small diameter pipe to the plug elevation. This does not seem necessary in light of the negligible increase in strength seen in the U2fe test plug emplacement through a pipe, compared with the previous emplacement by dumping directly into the hole.

59 The expected variations in the void fractions and particle size distributions in the gravel and the sand has led us to a revised formulation that is nominally a little rich in liquid, with epoxy about 10% greater than the mix used in CHEEDAM. Volume proportions of gravel:sand:epoxy are now at 3:1:0.83 in current applications.

ADVANCED CONCEPTS IN STEMMING DESIGN

The new epoxy plug material development program has enlightened the LLNL Containment community with the realization that polymer plug materials are sensitive to environmental conditions, costly and difficult to emplace, hazardous to workers, and susceptible to strength reduction from their own heat of reaction. In the process of evaluating the new material and alternative emplacement concepts it was clear that there was virtually no experience outside the Nevada Test Site pertinent to the emplacement of epoxy/aggregate concrete in such massive concentrated volumes- While the Celanese epoxy was essentially derived from products used in the coatings industry, there was little information that could be extended to nuclear event applications.

In a parallel, but much smaller effort, LLNL was evaluating cements as possible plug materials. Combining the understanding of the polymer problems and some knowledge of cement materials, it was decided to wipe the slate clean and consider the ideal way to fill an emplacement hole for containment of a nuclear evc^t.

In its simplest form, tne emplacement hole would be completely filled with a material that would act like the local medium and have enough strength to remain in place through the various collapse scenarios. It would also be of low permeability, would not damage cables, be quick and inexpensive to emplace, and be a reliably performing material. While filling the emplacement hole entirely with one good material is a simple design, it would be too costly.

60 The two regions of concern in stemming design are the stemming directly above the experiment and the upper region of the emplacement hole. The region directly above the canister is of interest because it is in that range"during low yield events that a competent material will enhance the integrity of the residual stress field and maintain the containment cage. This has the effect of preventing early time gas flow up the stemming column. The upper region of the hole is a critical area in that it is the last defense should a subsurface collapse eliminate the lower portion of the stemming column. Consequently, it is desireable to have a containment feature in the upper hole that will remain effective throughout any of the collapse scenarios, acting as a gas blockage as well as a structural member.

LLNL is pursuing actively a stemming design concept that addresses directly both the prevention of early time gas flow and the integrity of the upper stemming column. The initial concept is shown in Figure 4. It consists of a feature extending from the top of the diagnostic canister out to a range of 1.75 to 2 cavity radii for yields up to perhaps 12 kt. It is sized large enough to accomodate uncertainties in yield and in the location of the region of high residual stress (estimated at 1.25 cavity radii), and is to have properties that allow it to enhance the stress field rather than provide the discontinuity inherent in low strength, highly compressible granular stemming material. The upper features are long (perhaps 40 to 50 ft.) fairly strong components designed to remain in place through ground shock and slapdown accelerations and provide multiple redundancy to subsurface collapse.

Most promising at this point is a concrete consisting of gynsum cement and coarse aggregate (tuff fragments from the NTS shaker plan). This combination has Several advantages:

1) Gypsum cement achieves full strength in less than one hour. 2) The gypsuni slurry is of very low viscosity, and flows easily around cables and downhole hardware.

61 CABLES

100 - ZZZ2Z

200 -

THICK CEMENT/AGGREGATE FEATURES PROVIDING MULTIPLE 300 - REDUNDANCY TO SUBSURFACE inn COLLAPSE PHENOMENA 400 -

500 -

600 - COARSE

700 -

800 -

900 - VERY THICK (BUT POSSIBLY STAGED) CEMENT/AGGREGATE FEATURE TO ENHANCE THE 1000 EFFECTIVENESS OF RESTOUAL STRESS CAGE

1089 WORK POINT: 1070 ft

Figure 4. The proposed stemming design using large cement features.

62 3) Gypsum expands linearly .3%, enhancing coupling to the formation. 4) The strength of the cement/aggregate concrete is reasonably high, in the range of 1000 to 1300 psi in laboratory tests. 5) The material has numerous fielding advantages, including low heat of hydration, relatively low cost, ease of emplacement, no significant health hazards, no strength sensitivity to exotherm temperatures and most importantly, a broad industrial experience base from which to draw.

Recent fielding tests by LLNL have evaluate? emplacement methods and equipment. The most nrobable technique will involve simultaneously pumping gypsum slurry through tubing to the plug location while dumping the aggregate material into the hole with conventional stemming equipment.

The rates of emplacement possible with this approach suggest that total stemming times could be reduced by as much as 60%. While the cost savings can be significant with this approach, the confidence gained by emplacing as much as 250 ft. of "plug" material (compared to the current designs utilizing 23 to 25 ft. of epoxy plug material) is substantial. The projected plans for the emplacement of this design include laboratory and analytical evaluations of the concept's effectiveness, emplacement of the upper features as an addition in a satellite hole, emplacement of the lower plug feature on events with appropriate diagnostics, and eventually fielding of the complete stemming design in the second quarter FY o4.

SUMMARY

The CTE evaluation program illuminated several problems with materials and techniques in plug emplacements. Development of the Two Part Epoxy system, with the modified aggregate, has provided a plug material that has shown good downhole strength, consistent thermal behavior (providing some indication of consistent emplacement), repeatable structural characteristics, and ease of fielding. The new material has instilled sufficient confidence that LLNL has been able to use fewer and shorter plugs than in the past. 63 However, with this greater understanding of epoxy characteristics, it is clear that there are many problems associated with the emplacement of polymer plugs in stemming columns, due primarily to their sensitivity to environmental conditions and mixing techniques. Coupled with these concerns is the uncertainty of the usefulness of a small number (2 or 3) of short discrete structural plugs in the stemming column when the predictability of subsurface collapse is imprecise. The new concepts in stemming design address these uncertainties with several large cement/aggregate features to provide multiple redundancy, and a feature in the lower region of the hole designed specifically to enhance the capability of the low yield events to essentially contain themselves in early times.

These concepts have the potential of providing increased confidence in the engineered containment fetures, while simplifying and shortening stemming operations.

ACKNOWLEDGEMENT

The authors and the Nuclear Test Engineering Division would like to acknowledge the extensive development work performed by Phil Fleming in the early stages of the new material evaluation program.

Work performed under the auspices of the U. S. Department of Energy by the Lawrence Livermore National Laboratory under contract No. W-7405-ENG-48.

I)IS( I UMKK

Ihis document was prepared as an accounl of work sponsored h\ an ajiencv of the I niicii Stales (.overnment Neither the I nited Slates (iovc r»ent nor (lie I ni\ersit\ of ( alifornia. nor an\ of their employees, makes aiw >iarr*.i(s. express or implied, or assumes an> li'cal liahilitv or responsiliili;-, .or the ac- eurac>. completeness, or usefulness of anv information, apparatus, product, or process disclosed, or represents that its use would not infrinjic privatch owned rights. Reference herein to ans specific commercial products, piocess, or service In trade name, trademark, manufacturer, or otherwise, does not nccessarih constitute or impl\ its endorsement, recommendation, or favoring h\ the 1 niU'd States (.nu'rnment or the I niicrsitv of ( aiifornia. Hie views and opinion's of authors expressed herein do not n*.cessaril\ stare or reflect those of the I mled States f ioverninenl thereof, and shall not he used for advertising or product endorsement purposes.

64 APPENDIX A: COAL TAR EPOXY EVALUATION

Three separate tests were carried out with the intent of evaluating the coal tar epoxy plug materials in place at depth. The investigations were planned to provide insights into the basic material, the emDlacement methods, or any other factor of the downhole environment.

The first test was dene in hole U4ac with fully mixed material (epoxy and gravel) poured to a depth of 788 feet into an instrumented container forming a plug 88" in diameter and 10 feet thick. The second and third tests were done in conjunction with the demonstration of Li.NL's Cable Downhole System. Because there was no nuclear device but it did have an array of cables, these tests had the potential of providing more information than just on material properties. Large samples were retrieved with long, conical slotted buckets, pulled out of the plugs before significant strength developed, but late enough that the material remnined in the buckets. In addition to visual observations of the material, a number of tests /ere done to evaluate mechanical properties, degree of cure, and possible contaminants. Samples were taken for QA purposes in the same manner as for the nuclear events to compare with the downhole samples. Numerical results are given in Table Al.

Observations from U4ac

Visual inspection showed that the material was well mixed with no indications of layering or gross contamination. Flexural tests and shear tests produced results wel1 below the values expected. Infra-red spectral analysis indicated that the epoxy had not reached a state of full cure. Review of all procedures and tests of the QA samples showed no deviations from normal event practices and experience.

65 Flexural Strength Direct Shear Sample (psi) (psi)

U4ac Downhole 284 103

U4ac QA 857 n/a

LLNL History 630 n/a

CDS Retrieved Mat'1. 262 n/a

Laboratory Snmples n/a 577

Table Al. Coal Tar Epoxy retrieved material properties.

Observations from the CDS Tests

One of these tests was done by fully mixing the material at the surface and dumping directly into the hole, while the other had liquid pumped to the plug location through tubing with the gravel added from the stemming conveyor. The fully mixed material remained mixed at depth, while the pumped plug exhibited severe layering of aggregate and epoxy. Structural tests again showed strength to be significantly degraded from expectations.

Other Observations

The tests done on CTE afforded opportunities to evaluate the procedures, equipment and QA methods outside of the schedule pressure of a nuclear event. In addition, the ongoing event series did allow some testing of samples made up in addition to the normal set required. Several things

66 became quite clear:

1) The CTE formulation appeared sensitive to environmental and fielding conditions: moisture, initial temperature, mixing orders of components, etc.

2) Mixing of the CTE components required overly precise control for the field conditions.

3) The QA procedures were not capable of identifying problems with mixing procedures and tended to mask small problems with the formulations.

These problems made it clear that it was time to find a new material.

APPENDIX B: TPE AND GYPSUM CEMENT TESTING TECHNIQUES It was decided early in the material testing program to consider the epoxy/gravel mixture a concrete and test it as such. Some effort was directed toward identifying a single standard test (preferably ASTM) that could reliably characterize the most important strength properties of the plug material (compressive, tensile, and shear strengths). The same tests were applied to the cement materials. Pictures of a number of these test specimens, both epoxy and cement, are in Figures Bl through B8.

A series of tests was conducted on the TPE/gravel formulation used on MANTECA (1 part gravel, .4 part TPE), to obtain enough data to determine if the ratios between the various tests were constant within normal data scatter. Data are presented in Table Bl for .samples at ambient temperature, tested at least 10 days after they were poured.

The tests used were as follows:

1. Cornpressive test ASTM C-39

67 AVG. STRENGTH STAND. DEV. TEST NO. SAMPLES (psi) (psi)

Compression i n t 2584 21

Direct Tensile 603 59

Split Tensile 5 565 8

Flexural 5 1353 89

Double Shear 5 1522 190

Punch Shear 3 1303 100

Pull-Out Shear 20 1200 60

Table Bl. TPE mechanical properties comparing test methods.

2. Split tensile ASTM C496 3. Direct tensile - ASTM D2936 4. Flexure ASTM C78 5. Double shear 2" x 2" x 10" Western Technology Inc. custom tooling 6. Pull-out shear ASTM C900-38T 7. Punch shear 3" hole with 2.5" D punch, 3" thick sample.

A total of 130 mechanical tests were made. The data presented is a representative sample of these tests. Based on this-testing program, the following conclusions can be made about TPE and pea gravel concrete:

1. The compressive strength of the material is probably the best all around measurement if only one type of test is to be made.

68 2. The tensile strength appears to be about .2 times the compressive strength.

3. The shear and flexural strength is about .5 times the compressive strength.

Similar tests were conducted on the gypsum cement, along with measurements of peak temperatures. It appeared that gypsum cement with or without filler rock could meet the LLNL temperature requirements. U S Gypsum W-60 or Halliburton "Cal Seal", two commercially available gypsum cements, were reported to have a heat of hydration of 47.8 Btu/lb. and a specific heat of .22 Btu/lb F° . With a mix ratio of 100 lb W-60 to 40 lbs of water the calculated temperature rise would be 76 F°. This would be the highest value that could be reached assuming a good mix of gypsum cement and water and no separation during emplacment. Any addition of rock, gravel or gypsum gravel would result in a lower temperature. Tests in 55-gallon insulated drums confirmed the temperature prediction. A number of measurements were made on gypsum cement with and without various fillers. Some general comments follow:

1) Tensile strength and energy absorbtion are greatly increased with the addition of about .5 lb. of fiberglass per 100 lb. gypsum cement.

2) The fiberglass strands (.5 inches long) are very difficult to blend because of wetting problems.

3) Unfilled gypsum cement typically reaches 2500 psi compressive strength in one hour, and increases in strength as the material dries.

4) Filling the gypsum cement with coarse fill reduces by about a 50% the compressive strength of the concrete, depending c;. the aggregate used.

69 V

H f

i'3

-' w'!

Figure Bl. TPE Compressive Samples

Figure B2. TPE Pullout Test Samples 70 Figure B3. TPE Punch Shear Test Samples

Figure B4. TPE Direct Shear Test Sample

71 Figure B5. TPE Direct Tensile Test Sample

Figure B6. Gypsum Cement Wall Shear Test Sample 72 Figure B7. Gypsum Cement with Fiberglass Split Tensile Test

73 Figure B8. Gypsum Cement Punch Shear Test.

74 UCRL-89414 PREPRINT

QUANTITATIVE DETERMINATION OF MINERALS IN NEVADA TEST SITE SAMPLES BY X-RAY DIFFRACTION

Gayle A. Pawloski

This paper was prepared for submittai to Second Symposium on Containment of Underground Nuclear Explosions, Kirtland AFB, Albuquerque, New Mexico, August 2-4, 1983.

July, 1983

This is a preprint of a paper intended for publication in a journal ur proceedings. Since changes may be made before publication, this preprint is made available with the understanding that it will not be cited or reproduced Hithout the permission of the author.

75 QUANTITATIVE DETERMINATION OF MINERALS IN NEVADA TEST SITE SAMPLES BY X-RAY DIFFRACTION PAWLOSKI, Gayle A., University of California, Lawrence Livermore National Laboratory, P. 0. Box 808, Livermore, California 94550 Abstract The external standard intensity ratio technique has been developed into a routine procedure for quantitatively determining mineralogic compositions of Nevada Test Site (NTS) samples by X-ray diffraction. This technique uses ratios of X-ray intensity peaks from the same run which eliminates many possible errors. Constants have been determined for each of thirteen minerals commonly found in NTS samples -- quartz, montmorillonite, illite, clinoptilolite, ^Hstobalite, feldspars, caicite, dolomite, hornblende, kaolinite. muscovite, biotite, and amorphous glass. Ratios of the highest intensity peak of each mineral to be quantified in the sample and the highest intensity peak of quartz are used to calculate sample composition. The technique has been tested on samples with three to eleven components representative of geologic environments at NTS, and is accurate to 7.0 wt% of the total sample. The minimum amount, of each of these minerals detectable by X-ray diffraction has also been determined. QUANTS is a computer code that calculates mineral contents and produces a report sheet. Constants for minerals in NTS samples other than those listed above can easily be determined, and added to QUANTS at any time. Introduction X-ray diffraction (XRD) is a quick and easy technique to identify and quantify mineral content of Nevada Test Site (NTS) samples. Sample preparation is quick, analysis equipment is automated, and run time for each sample is on the order of one hour. Rocks from NTS have a small range of mineralogic composition; there are probably no more than twenty minerals that have been found at NTS. These minerals include quartz, montmorillonite, illite, clinoptilolite, cristobalite, feldspars, caicite, dolomite, hornblende, kaolinite, muscovite, and biotite. Amorphous glass, which is commonly found at NTS, is not a mineral. However, for ease of discussion, it will be included in the list of minerals in this paper. Each sample typically contains five to ten minerals total. Because most of the minerals are found repeatedly in NTS samples, qualitative interpretation of results is straightforward and quickly accomplished. Quantification of the minerals is another story. For the past number of years, Lawrence Livermore National Laboratory has been reporting semi-quantitative results from X-ray diffraction analysis. Three minerals (montmorillonite, a clay; clinoptilolite, a zeolite; and caicite, a carbonate) have been quantified by reporting that the mineral content of a sample falls at or within ranges of tested values by comparing the X-ray intensities of the unknowns to previously run known compositions. All other minerals were reported as occuring weakly, moderately, or highly. These adjectives could not indicate quantity, since intensities from minerals widely differ, and no known values had been tested. For. example, 5 wt% quartz may give a "high" intensity peak, while it may require 50 wt% cristobalite to form such a "high" peak. It became obvious that there was a need to solve this confusing problem. Although containment concerns center on determination of clay and carbonate content, determination of total mineral content cf a sample is helpful in characterizing the geologic setting. It was necessary to develop a new powder X-ray diffraction procedure to routinely determine mineralogic composition of NTS samples.

76 X-ray Diffraction Principles This is a very brief review on XRD principles; for more thorough information the reader should check Klug and Alexander (1954). A monochromatic beam of radiation strikes the sample and the reflections collected by a counter as the scan progresses give information on d spacings (separation of parallel planes in the crystal lattice) and intensities (counts of photons) from single crystals or powdered samples. As the beam strikes the sample, some of it is absorbed and some is reflected. Some reflected beams reinforce each other (the path difference equals a whole number of wavelengths), and Bragg's Law describes the situation, as shown in Figure 1. At other angles 9 the beam is not reflected due to interference. Because all minerals have characteristic atomic structures, they can be identified by a unique X-ray diffraction pattern. For example, Figure 2A shows the XRD pattern of pure quartz. Reflections will always occur at certain 28 angles (or d spacings) and the relative intensities of the reflections will always be 'in certain ratios to each other. Figure 2B shows a sample that contains quartz with other minerals. Quartz can be identified because reflections from its atomic structure will always occur at the same 20 angles (or d spacing^) and the relative intensities of the reflections are again at the same ratios to each other, as in Figure 2A. No matter how many components are in a sample, if quartz is present it can be recognized by its characteristic XRD pattern. The same is true for all minerals.

The intensities of X-ray diffraction lines may differ due to several reasons -- polarization, Lorentz, temperature, atomic scattering, structure, multiplicity, absorption, and machine factors (Klug and Alexander, 1954). These differences in intensities can have a large effect on quantitative work. But because our X-ray scan is of short length, many of these factors can be assumed to be small and thusi require no correction. Choice of an efficient analysis method can reduce other factors. A theta compensating slit has been added to our X-ray unit. It functions as an automatic divergence slit and parallel (Soller) slit. The purpose of this slit is to keep a constant area of the sample irradiated throughout the scan, as opposed to a fixed div^. gence slit which allows different areas of irradiation as the scan progresses. There were two reasons for installing the theta compensating slit on our equipment! Since it is difficult to directly obtain comparable intensities due to a number of factors listed above, the next best step is to keep the area of the sample irradiated constant throughout the scan. The same X-ray intensity is always on the same amount of sample and comparison of intensities is easier. The second and more important reason was to give us better data at low angles. We need to identify and quantify clays, which occur at low angles. Figure 3 shows how difficult this could be before the theta compensating slit was installed. The same sample was run twice,/once in a fixed divergence slit mode, and the second time in a compensating slit mode. At low angles in the fixed mode (Figure 3A) the X-ray source radiates almost directly into the detector. Background is high at the beginning of the scan, and progressively drops off as the scan continues. Because of the X-ray source peak and the progressively different background values, intensity corrections may need to be done before interpretations can begin. A peak at about 5°20 is due to 15 wt% montmorillonite. It is difficult to identify this peak in Figure 3A. But in 3B, a scan utilizing the

77 compensating slit, background is relatively constant throughout the scan, and the montmorillonite peak is easier to identify. Installation of the theta compensating slit has made XRD analysis at low angles much easier to accomplish.

Although the theta compensating slit has made examination of NTS samples more direct and less time-consuming, the slit has placed a limitation on us. Widely used relative intensities published by the Joint Committee on Powder Diffraction Standards (JCPDS) cannot be utilized, since this data was obtained under fixed slit conditions. Fixed slit intensities are not directly comparable to those obtained under compensating slit conditions. Also, constants which will be described later in this paper cannot be directly used by others operating under fixed slit conditions. A compensating slit can also be run in a fixed slit mode -- we are not limited to one method of operation. It is important, though, to realize that the intensities collected under these two different operating conditions are two different types of data, and are not directly comparable.

Figure 4 shows a diagramatic equatorial view of an XRD unit. The sample is tangential to the focusing circle. The detector pivots around the sample during a scan. For parafocusing, the sample-d' tector distance must remain equal throughout the scan. To accomplish this, the sample rotates with one-half the angular velocity as the detector (the sample will be at angle 0 and the detector at angle 29 to the X-ray source). The compensating slit rotates with the sample in order to maintain the constant area of irradiation of the sample throughout the scan.

XRD Quantification Methods Why are intensities important? Qualitative determinations of minerals are accomplished by identifying peak locations in an XRD pattern, but minerals can be quantified by knowing integrated intensities, densities, and absorption properties of the sample: K.- x.

( j

I-jj = integrated intensity of the i^n line from component j Kj = constant dependent on the geometry of the diffractometer and the nature of component j XJ = weight fraction of component j Pj = density of component j UJ = mass absorption coefficient of component j 1%, = mass absorption coefficient of the matrix (material that would remain if component j were removed)

78 Chemical variation and changes in the number of components can change densities and mass absorption coefficients, making this equation difficult to solve. Frequently, u£ 1S not known, and Equation 2 is used:

K, xi Iii = J—4 (EQ. 2)

uT = total mass absorption including component j This equation may require empirical corrections because of microabsorption effects. If another component is added to the samplj uy drops out of the equation because the intensities are measured on the same sample, and the mass absorption coefficients are characteristic of the sample. Small errors resulting from microabsorption effects can be eliminated using this equation: I.- K, p- x, x, T113 IT-TT x = constant -i (Eq. 3) Ik1 K1 P-| x1 x] 1^1 = integrated intensity of the k^h line from component 1 Ki = constant dependent on the geometry of the diffractometer and the nature of component 1 PI = density of component 1 X] = weight fraction of component 1

Tnese equations were developed by Klug and Alexander (1954), and have been used and modified oy many researchers. There are a number of techniques which utilize these equations for quantification of components in a sample. The first measures the intensity of the unknown and compares it to a pure standard. This requires a pure standard to be run also. The mass absorption coefficients of the standard and the unknown must be known. The second technique involves spiking or diluting the sample after its initial run. The spiking method adds fixed amounts of the component of interest and records the intensity of a XRD line from this component at each amount. The amount of the original sample is extrapolated

79 from this data. The dilution method follows these same steps, except an amorphous dilutent is added. This method requires repeated scans and collection of data for each component of interest. It is also necessary to obtain exactly the same component as the one to b« quantified for the spiking method. The third technique is the internal standard method. A fixed amount of a standard material not found in the sample is acided. The intensity of a XRD line from the standard is ratioed against the intensity of a line from the unknown, and comDared to a calibration curve. The calibration curve is a graph of I-jj/Ifci plotted against Xj/x-j for samples containing known mixtures of components j and 1. At least three points are needed to define the linear calibration curve. The fourth technique is the standardless method. The sample is chemically or mechanically treated to reduce the amount of a component. This method can be used for crystalline components only. A quantitative determination of a component is made by comparing the intensities of XRD lines of the component before and after treatment. This method requires repeated scans for collecting data, and chemical or mechanical treatment of each component in the sample. The external standard is the fifth technique. Ratios of XRD lines of components in the sample are compared to ratios of lines from previously mixed and run standards. This is the method we use for semi-quantitative analysis of montmori1lonite, clinoptilolite, and calcite. Standards of known compos it inns must be made for each component in the sample. The sixth technique, matrix flushing (Chung, 1974) utilizes Reference Intensities (Imineral^corandum) published by the JCPDS, instead of an external standard. The concent is that mass absorption coefficients, which are difficult to measure, are flushed out of the equation, and a simple relationship then exists between intensity and concentration. Corundum is usually added as the flushing agenu, but a component found in the sample may be used. This method has proven successful for quantifying amorphous components.

All of these quantitative techniques have been rigorously tested and work well. Some of the techniques require- mass absorption coefficients, which ran he difficult to accurately determine. They must be determined for each mineral to be quantified. A slight change in chemical composition requires a new coefficient to be determined. It is known that NTS minerals hav? variable chemical compositions -- for example, both Ca and Na clays exist, and there are various feldspars from the solid solution series. This would require knowing many mass adsorption coefficients, and using the correct one in the equation. Other quantitative techniques- require multiple runs for each mineral. This is time-consuminq considering the number of samples we may be expected to handle. We would like to quantify glass, An amorphous material. Most of these techniques have only been tested on crystalline materials. Another problem is that approximately twenty minerals havg been found at NTS, although only five to ten minerals are commonly found in any one sample. We would like to accomplish the quantification of all minerals in one step. Some of the techniques have been applied to binary systems, while'others have successfully handled multi-component systems. But -- the maximum number of components accurately measured by other workers has been four. We need to choose a technique that allows chemical variation of the samples (requires no mass absorption coefficients), can be used routinely (does not require multiple runs for each mineral), and can handle five to ten (or more) minerals at one time. All of the equations included in this paper require that constants be determined. These must not be difficult to measure for the quantification technique we choose.

80 Goehner (1982) lias modified and tested, on a three component system, an external standard intensity ratio technique based on an external standard method formulated by Copeland and Bragg (1958). The weight fractions of each component to be quantified are obtained: x. I . -J- = K y! (Eq. 4) xl h K is a constant which can be obtained from a single standard of a one-to-one mixture of components i and j (Chung, 1974). x. I. L _i = K J •> i:i mixture -»• K = -— (Eq. 5) Xl 1 j K values are known for all components in the sample (Equation 5), the sample to be quantified is run once, intensities of specific diffraction lines are ratioed, and Equation 6 is used to calculate the weight fractions of all components in the sample, totaling 100 wt%. I \ - 1 (Eq. 6) n = number of components in sample x. = weiqht fraction of components Because intensities that are ratioed are obtained from one sample run of short duration [y lhr.), many XRD problems are. minimized (Goehner, 1982). These include changes in barometric pressure, machine factors (KeV and Ma drift), long term drift of the X-ray tube (aging), matching mass absorption coefficients, sample problems such as compactness of the powder and sample transparency, and alignment problems like sample positioning in the focusing circle and slight goniometer alignment changes. Several problems not minimized that the worker needs to be concerned with are particle statistics, preferred orientation, microabsorption, extinction, and obtaining a standard similar to the unknown. The first four of these remaining problems can be de-s-U with by careful experimental planning and sample preparation. The final one,, obtaminq a standard similar to the unknown, remains a oroblem. NTS samples vary in" compos it ion. Samples with chemical variation would require more than one standard arH~-pr.nper selection of the correct K constant at the appropriate occasion. It is har"d--fco,.gbtain a representative sample from NTS of most minerals in pure form to use as"sta-n-d.a.rds. There were initial hopes of obtaining "average" NTS samples for use as sl:an-d-sr.d_s., but this was impossible. We will have to make an assumDtion that all s'airp-l-es^.wjl 1 react in XRD analysis like the standards used -- the mon-cmori 1 lonite in a saTnpte- tested today and one tested in the future will give comparable XRD results as the montmori1lonite used as a standard. Experimental work completed after this assumption was made confirms that this is a valid assumption.

The external standard intensity ratio technique was chosen to quantify minerals in NTS samples. It requires no mass absorption coefficients, can be used routinely, has been tested on three component samples, and K constants can be identified by making one standard for each mineral. Many XRO problems are also minimized because we are using ratios of data gathered over relatively short scans. The standards have to be made with one common mineral

81 that will be found in every NTS sample. Quartz was chosen as the common mineral -- in the several hundred samples I have analyzed by XRD, quartz has always been present in some amount. If quartz is not present in a NTS sample, which seems unlikely, this technique will not work, and there are two alternatives -- use another technique, or infer a nominal amount of quartz is present, calculate mineral composition using Equation 6, and tnen recalculate these values knowing that quartz is not present in the sample. Experimental Process Our X-ray unit utilizes a Cu X-ray tube, and machine operating conditions are 45 KeV and 30 Ma. A theta compensating slit is the diverging slit, and there is a 1° receiving slit. The XRD scans cover 2-45°26, automatically stepping 0.04°26 every four seconds. Good counting statistics are obtained from the step size and counting time. Total scan duration is 73 minutes. The scan length obtains much useful data for the mineral suite in a reasonable amount of time. The highest intensity peak of each mineral in question falls within the scan length. Other peaks of tnese minerals also fall within this length, ensuring positive identification of all minerals present. Tnere are three experimental parts to obtaining a working XRD procedure for quantifying minerals in NTS samples. K constants have to be determined for each mineral expected in a NTS sample, the external standard intensity ratio technique must be tested and its accuracy determined, and the minimum amount of each mineral that can be detected by XRD must be determined, to place a lower limit on the quantification procedure. It was necessary to determine optimal particle size of the samples. For good XRD data, particle size must be small, within the range of 5-60 microns. Grinding and sieving samples can require a large amount of time. Also, several minerals tend to form smaller particles (clays and zeolites) while others are difficult to grind (quartz and calcite). It was imperative that the sample preparation process not selectively choose or eliminate any mineral. Samples of known compositions within various size fractions were analyzed by XRD to determine that all minerals were present. The data showed mineral contents as expected. A decision was made to constrain particle size to >35<45 microns. This size fraction would not require excessive sample preparation, and would yield good XRD data. Thirteen minerals are common in NTS samples. These are quartz, montmorillonite, illite, clinoptilolite, cristobalite, feldspars, calcite, dolomite, hornblende, kaolinite, muscovite, biotite, and amorphous glass. Because it was difficult to obtain pure "typical" or "average" minerals from NTS samples to use in determining K constants, the assumption was made that all samples will react in XRD analyses like the standards used, and pure minerals were obtained from Wards Natural Scientific Establishment, Inc. to use as standards. Standards were made of each mineral in a one-to-one weight ratio with quartz. A K constant was calculated for each mineral using the highest intensity peak of quartz (100 peak = d spacing of 3.34) and the highest intensity peak of each mineral, as in Equation 5. Each standard was analyzed nine times, and the K values were then averaged. These are shown in Table I. The weight fraction of each mineral was then calculated to determine the accuracy of the K constants, using Equation 4. There was a surprisingly large error in calculating weight fractions of all minerals except muscovite. It was obvious that these K constants were not accurate.

82 Tnese K constants are simply the slope of the linear calibration curve where the standard is composed of equal amounts of quartz and mineral and the weight fraction is equal to 1.0. It was decided to prepare and test more standards of different compositions to accurately determine the calibration curve for each mineral. The K constant (slope) could then be calculated by a least squares fit of the data. Standards were made at 89:11, 80:20, and 33:66 ratios of the mineral to quartz. Standards were also made at 25:75 and 20:80 of illite and muscovite, which required better control where data points were slightly erratic. Each calibration curve was determined by a minimum of eighteen points. The K constants from the slope of the calibration curves are also shown in Table I. These new K constants proved to be much more accurate when calculating weight ratios of standards, with the exception of muscovite. Muscovite contents are most accurately calculated using the mean of the intensity ratio of the 1:1 standards, and this mean will be the K constant for muscovite. For all other minerals, K constants are the slope of the calibration curve.

Sixteen samples of various contents from the list of thirteen minerals were prepared and analyzed by XRD to test the external standard intensity ratio tecnnique and determine its accuracy for quantifying minerals in NTS samples. Sample content ranged from three to eleven mineral components. With the exception of biotite and Cristobalite, which were in two and three samples respectively, each mineral was present in at least four different samples. Quartz was present in all. The samples were prepared to be representative of geologic environments found at the NTS. Altered samples were composed of predominately clays and zeolites, fresh samples were composed entirely of quartz, Cristobalite, and glass, and there were various intermediate compositions. Equation 6 was used to quantify each sample. The error in quantifying each mineral in a sample and the maximum error for each sample were tabulated and are shown in Table II. The distribution of error is shown in Figure 5. The average error is 0.0 wt%, which was expected knowing that calculated quantities could deviate above or below the known value. The standard deviation of this error is 2.89 wt%. The maximum error for this technique is ± 7.0 wt%. This maximum error occurs approximately one third of the time (5 out of 16) and is the most conservative method of reporting error in the technique. For this reason, the error in the external standard intensity ratio technique for quantifying minerals in NTS samples is ±7.0 wU. Tne final experimental phase was to determine the minimum amount of each mineral that could be detected by X-ray diffraction. This would place a lower limit on the quantification procedure based on what could actually be observed in the XRD data. Each XRD pattern contains a small amount of background. The analyst must be able to detect at least one peak (preferably two or three) from a minerals pattern above this background level to positively identify the presence of that mineral. The minimum amount of each of the minerals detectable by XRD is listed in Table III. With the exception of glass, all minerals can be accurately identified below 10 wt%, and some as low as 0.5 wt%. The minimum amount of glass detectable is 40 wt%. This is a much nigher amount than the other minerals. Glass is an amorphous material, and is identified by XRD as a wide hump extending from approximately 19 to 36°26. Small amounts of glass (less than 40 wt%) are difficult to distinguish from

83 normal or abnormal background levels. Litiiologic information and/or knowledge of the entire mineral composition of the sample will help indicate if much glass is present in a sample. For example, knowing that the sample came from a vitrophyre would give a nigh probability that the sample composition included glass, and ;.; highly altered sample would most likely contain clay and zeolite, with small amounts of glass, if any at all.

The minimum amounts of minerals detectable shown in Table III are the true minimum tnat can be observed in the XRO dat:. It is oossible that a content lower than these values may be calculated using Equation 6. It must be remembered that the calculated values are based on what is actually observed in tne XRD pattern '-nth an error of ± 7.0 wt%.

QUANTS QUANTS is a computer code written to calculate mineral contents based on Equation 6 ind oroduce a report sheet. K constants for each mineral are stored in tne code. An incut file consisting of drill hole identificatio'i, sample date, XRD date, and sample information must be made. The satmlo information includes deoth in feet (depth in meters is calculated), sample type (cuttings, sidow-j 1 1, or percussion gun), and ratios of the highest intensity neaK of eacn mineral present and the highest intensity peak of MUtirtz -- uie satin two peaks of the samnle that were used for calculating the K constant of the standard. If a mineral is not present in a sample, a value of CO must be entered. These values must occur in the input file in the same order tney occur in on the report sheets. The input -File and the XRD data file are stored for easy computer access. A report sneet is shown in Figure 6.

Routine XRD Procedure F)TS"s~aTnpTeT~s7Jbmitted for X-rey diffraction are crusned and sieved, the size fraction from >3b<45 microns is X-rayed under the machine conditions ">reviously specified, and the raw data is reduced by an analysis •"ode and displayed in a readable manner (Goehner, 1982). The analyst then identifies the minerals present. Glas<- is identified by subtracting out the background of tne scan, and determining if the subtracted data represents glass or background. Occasionally some peaks will be composed of more than one mineral and will need to be deconvoluted (namely illite, muscovite, and biotite). The nignest intensity peak of eacn mineral present and the highest intensity peak' of quartz (the same two peaks of the sample that were used to calculate the K constant of the standard) are ratioed, and an input file is made for QUANTS. The code is run, and a report sheet is issued. The input file and the XRD data file Are stored for future use.

Conclusions Th~e~e~x"ternal standard intensity ratio technique has been successfully tested to quantify minerals in ^TS sample. '< constants have been determined for thirteen minerals commonly found in NTS samples -- quartz, montmorillonite, illite, clinoptilolite, Cristobalite, feldspars, calcite, dolomite, glass, hornDlende, kaolinite, muscovite, and biotite. The quantification of these minerals is accurate to + 7.0 wt%. The minimum amount of eacn mineral t'iat can be detected by X-ray diffraction has also been determined. These are listed in Table III. Discussion TTTTs" possible that minerals other than those listed above may be found in NTS samples. In order to accurately quantify all minerals in a sample, a K constant for each mineral must be stored in QUANTS. New K constants can easily be determined, and then added to the code at any time. The establishment of this technique as a routine procedure for quantifying minerals is an accomplishment in several ways. First, it eliminates other methods that produced senri-quantiative results for selected minerals. We can now accurately quantify all minerals in NTS sample's"! Secondly, most analysts who utilize X-ray diffraction are attempting to identify and quantify chemical compounds. We have switched this emphasis to commonly found minerals. Finally, we have stretched the bounds of quantitative techniques. Most of these techniques have been developed, tested, and utilized on systems containing two to three components. The external standard intensity ratio technique has now beon tested and will routinely be used on systems contain' ,j up to eleven components from a list of thirteen commonly found minerals. There is a possibility that this technique can be expanded even further.

Acknowledgments I would like to than!: Bill Beiriger and Bern Qualheim for the large amount of time tiiey spend helping with sample preparation. Norm Burkhard was instrumental in writing the code QUANTS. I also appreciate the suggestions and editorial comments made by Larry McKague.

*This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract No. W-74O5-Eng-48.

I)IS( I \I\IH<

I his document was prepared as ;in iiccounl of stork sponsored l)> an ayi'tic) of (IK- I nili'd Stales (.miTiimv t. \ cither tht 1 nili'd Stales (.ournnunt nor the I ni*ersil\ of California nor n\ of thtir tinf)!o>i s. nukes an> press or implied, an\ k'j»al liahi!il> or responsihilil the ac- cur;ic\. coniplcU'tU'ss, or usr nliK'ss of an> information, apparalu iduct, or process disclosed, or represei ts dial hs use vsouid not infringe pr \ OVMU'd rights. Reference lu* in to an\ specific conmit-rciai.products, proc r scr l>\ trade nami1, trad mark, manufacturer, ur cithcruisc. dois n ess a i I > constitute or impt> i cndorst'menl, ricomiiK'iidnliiin. or favoring h\ tl t n led Sljiti's (internment r (hi- I nivmil) »r ( alifornij. 1'he vit-us and o|: 1 authors expressed h i do nor neccssarih slatt or ri'flccl (hose of ttv I n led Slates («o-ernnient thereof, and shall not he used for a dorsemeiil purposes. References

Chung, Frank H., Quantitative Interpretation of X-ray Diffraction Patterns of Mixtures. I. Matrix-Flushing Method for Quantitative Multicomponent Analysis, 1974, Journal of Applied Crystallography, Vol. 7. Copeland, L. E., and Bragg, Robert H., Quantitative X-ray Diffraction Analysis, 1958,-Analytical Chemistry, Vol. 30, No. 2. Goehner, Raymond P., X-ray Diffraction Quantitative Analysis Using Intensity Ratios and External Standards, 1982, Advances in X-ray Analysis, Vol. 24. Goehner, Raymond P., SPECPLOT User Manual: Revised 1/82 for DEC Operating Systems, General Electric Report No. 82CRD114, 1982. Klug, Harold P., and Alexander, Leroy E., X-ray Diffraction Procedures for Polycrystalline and Amorphous Materials, 1954.

86 List of figures and Tables

Figure 1 For a given d spacing and a given wavelength x, the various orders n of reflection occur only at precise values of angle , which satisfy Bragg's Law. Figure 2 Quartz has a unique XRD pattern (A) due to itr> characteristic atomic structure. This unique pattern allows identification of quartz in samples composed of quartz and other minerals (B). Figure 3 A fixed slit scan (A) shows progressively decreasing background. A montmoril lonite peak ('V5°26) is difficult to distinguish. A theta compensating slit scan (B) shows relatively constant background, and the montmorillonite peak is easier to identify. Figure 4 Diagrammatic equatorial view of an XRD unit.

Table I K constants for minerals commonly found in NTS samples, determined by two different methods. Table II Error in quantifying known sample contents using the external standard intensity ratio technique. Figure 5 Distribution of error in quantifying known sample contents by the external standard intensity ratio technique. Table III Minimum amount of mineral detectable using X-ray diffraction. Figure 6 The computer code QUANTS quantifies the mineralogic composition and produces a report sheet.

87 Bragg's Law

given wavelength

CO nA=2dsin9

various orders crystal of n reflections spacing occur only at precise angle e

Figure 1. For a given d spacing and a given wavelength A, the various orders n of reflection occur only at precise values of angle e, which satisfy Bragg's Law. v 1 6-21-83 DIF ALEX QSTAND 3TART ANGLE- 2.0C0 STEP SIZE- 0.040 TIME INTERVAL' 4.00SEC 0 . 4.5T 0 « 3.6 + 2.7-- 1.8" 0.9 +

0.0 •+• 0.20 1.36 2.52 3.68 4.84 6.00 2 THETA X 0.1 CD B 5 6-22-83 DIF ALEX Ul START ANGLE- 2.000 STEP SIZE- 0.040 TINE INTERVAL* 4.00SEC 0 1.4- 0 0 1.1" 1 0.8--

0.S--

0.3--

* »• • 1.36 2.52 3.68 4.84 6.00 2 THETA X 0.1 Figure 2. Quartz has a unique XRD pattern (A) due to its characteristic atomic structure. iuIrtflT Pf*tten? allows identification of quartz ir samples composed of quartz and other minerals (B). 4 6-24-83 DIF ALEX U3-FIXED £TART ANGLE- 2.000 STEP SIZE- 0.040 TIME INTERVAL- 4.00SEC

. l.0r 0 0 0.8"

0.6 " " '

0.4

0.2t

0.0 0.20 1.06 1.92 2.78 3.64 4.50 o B 2 THETA X 0.1 x 3 6-24-83 DIF ALEX U3-C0MP START ANGLE- 2.000 STEP SIZE- 0.040 TIME INTERVAL' 4.00SEC 0 < 1.0T 0 \ 0-8 + 0.6--

0.4--

0.0 1—- 0.E0 1.06 1.92 2.78 3.64 4.50 2 THETA X 0.1 Figure 3. A fixed slit scan (A) shows progressively decreasing background. A montmorillonite peak (v5°2e) is difficult to distinguish. A theta compensating slit scan (B) shows relatively constant background, and the montmorillonite peak is easier to identify. scintillating detector focusing circle

X-ray goniometer tube tower (goniometer circle)

compensating slit

Figure 4. Diagrammatic equatorial view of an XRD unit. I5r

10 o o

i I I I Illlllf II I • -7-6-5-4-3-2-1 0 I 2 3 4 5 6 7 % error Figure 5. Distribution of error in quantifying known sample contents by the external standard intensity ratio technique. XRD

CL1 N

/i 0. 0 . 1 1 . 0. s. . 7 0. 0. 0. 0. 0. 0. ) i . / * . 0. ,V . (5 0. IO! 7 '? 0. 0. n 5. 0. 0. 1 V'J . 1 ,? 0 . i -. •' o. /i 0. 6] 0. 0. 0. 0. 1 . '•: .<< 0. 0 . 0. 7. 0. i ., (' 5! I "•"'. 0. 5" . . 1 0. 0. 0. -1 c , 6] 0. 0. *'•' . 1 0. 0. 1 . 0. G ! 0. n6. 2. 1 L'. J 06.. 0. 0. 0. 0. ••) 6! 4' . n. 0. -;•. ^ 1 . 0. P. 0. 0. 0. 0. 0. "i, 0 7" p, /i i 0 0. 0 . 0. 0. 0. 0. V) . 0. i i •. 0. 0. 0. 0. 0. 0. i I''. • 1 . 13. 0. ir.. 0 0. 0. 0. 0. 0. 0. i •\, • ' i"l 0. 0. i. 1 • '. 0. i 0. 0. 0. 0. 0.

;•, 0. 0. 2. 1 •' 0 0. 0. 0. 0. 0. 0. r.. i , 0. o. 3. ! i'.' <>. u 0. 0. c. 0. 0. 0. ... . •' ' / . 0. 2. 9. 1) 0. fj. 0. 0. 0. 0. <'l .1 , 1 '••. 2. r'O. 0. 0. 1 . 0. 0. 0. '.;, *.' 'i . u. Cl'. 2. 20 . ^ 1 A c. 0. 0. 0, 0. 0. z> !"7. o. 23. y. 0 0. 0. 0. 0. 0. 0. 1 1 1" I •',. •' . \ ' ' . \ , 1 •'•. 1 li. o! 1 . 15. u. 56 . 0. 0. 0. 0. A J , !" '* 1 '•':. ? A . 24. 2. 8. t\ 0. C. •I , 0. 0. 0. 1 :'" i i . * . .;, .-' •? S2. n. 4. 2S. 11 '. 5 t). 2 0. i. 0. 0. c. 1 •: -1. A\ . 1 5 o! 1 . 39. n 0. 0. 0. 0. 0. 0. i . ',">.') . •3 ">. . 0. o'. 0. 0. 4. 0 0. 5 0. 0. 13. 0. 0. 0. 0. 0. 0. IP. 1 !9 0. 5 0. 0. 18. 0. 0.

ri K'T." . M"t T •.. or ' 21 M1NL£R'\LS ANALYST I F. PAW I _<-<$!<. QUAR •• QUA7TZ pen o = DOLC 111 TE OATEV JUNE 2,1974 MONT (?l '"•. • = GL XRD DAT I".:? L 9,1 9S3 - flCNTMO ^H LONITE : v° ~. • AF ILL I = !LL!TE H'"-. ^i! ~ ] \~ MDi C'\Ti£^ r\\-\: \.E TYPE CL 1 M fll ril. ! TE |C:.0L = K." "l\ i "-J i TE o CLIMOP CP?i > -• CP i § re.'.^,1 i"il hi'ISC - [\'j !^ A.' : TE r, - pi"T?i i I." - i Oi" C1.JN FELI3 JiOT = l~> 1I 1 i TE s - SI'IX-.WAI. L CAL C = CALC1T 7

Figure 6. The computer code QUANTS quantifies the tnineralogic composition and Droduces a report sheet. Mineral K constant 1:1 slope of (x) calibration curve Quartz 1.0000 1.0000 Montmori1lonite 23.8202 22.0412 II lite 50.6306 30.2904 Clinoptilolite 12.2582 9.7432 Cristobalite 1.4695 1.2940 Feldspars 1.3267 1.2774 Calcite 0.6790 0.6544 Dolomite 0.4901 0.3528 Glass 52.7867 36.8405 Hornblende 3.2370 2.7698 Kaolinite 10.5109 10.5970 Muscovite 2.5758 1.9180 Bioti te 0.3694 J. 0.4304 Table I. K constants for minerals commonly found in NTS samples, determined by two different methods.

94 Maximum Sample Q MO IL CC CR FS CA DO GL HO KA MU BO Error 1 +4 -2 +1 -3 +4 2 +4 +2 -4 -1 ±4

3 +2 -3 -1 -7 +5 +3 +3 -1 -7 4 +4 -2 -1 +2 0 -1 -2 +4

5 ••4 -4 +7 -3 +2 0 -2 -2 +2 -4 +7

6 +4 -3 0 0 +3 +1 -2 -3 +4

7 +6 -7 -1 +1 + 1 -7

8 +7 -2 +3 -2 -4 0 -3 + 7 9 -1 0 + 1 +4 + 1 0 -2 0 -3 +4

10 +6 -7 +3 -5 +3 -7

11 0 +4 -4 ±4 12 +2 -1 -1 -1 +2 13 0 -1 +4 -2 +4

14 0 +3 -2 +2 -2 +3 15 +3 -3 -2 0 0 -1 -1 +3 ±3

16 + 2 0 -2 _] ±2 1 i x = ().O wt* s = 2.89 n = 95

Q = Quartz FS = Feldspars HO = Hornblende MO = Montmorillonite CA = Calcite KA = Kaolinite IL = Illite DO = Dolomite MU = Muscovite CC = Clinoptilolite GL = Glass BO = Biotite CR = Cristobaiite

Table II. Error in quantifying known sample contents using the external standard intensity ratio technique.

95 Quartz 0.5 wt% Dolomite 0.5 wt%

Montmori1lonite 5.0 Glass 40.0

Illite 7.0 Hornblende 2.0

Clinoptilolite 5.0 Kaolinite 5.0

Cristobalite 1.0 Muscovite 3.0

^Feldspars 2.0 Biotite 5.0

Calcite 0.5

Taole III. Minimum amount of mineral detectable using X-ray diffraction.

96 LA-UR-B3-2110

Los Alamos National Laboratory is operated by the University of California for the United States Department of Energy under contract W-7405-ENG-36.

TITLE: CONTAINMENT WELL LOGGING PROGRAM

AUTHOR(S) Ronald D. Oliver

SUBMITTED TO: 2nd Containment Symposium, Albuquerque, NM August 2-4, 1983

By acceptance of this article, the publisher recognizes that the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or to allow others to do so. for U.S. Government purposes.

The Los Alamos National Laboratory requests that the publisher identify this article as work performed under the auspices of the U.S. Department of Energy

Los Alamos National Laboratory- Los Alamos, New Mexico 87545

FORM NO. 836 R< ST NO. 2629 5/81 97 CONTAINMENT WELL LOGGING PROGRAM

Ronald D. Oliver Geophysics Group Los Alamos National Laboratory

ABSTRACT The objective of the Los Alamos Containment Well Logging Program is to establish, compile, and distribute containment- related information. A description of this system will be presented. Fielding of developmental and established types of logging equipment is underway. Measurement capabilities range from borehole gravity and magnetics to color video and photography. A PDP 11/34 computer system is used to record, correct, plot, and analyze the log results. A healthy mix of theoretical predictions, modeling, and field measurements is being achieved. The computational effort includes seismic, gravity, magnetic, and radiation transport modeling. The logging system's design capabilities include routine logging, research and development logs, and the versatility to conduct borehole experimentation. The final system will provide a capability for advancing significantly the science of well logging for containment purposes and will provide an opportunity to effect savings in test-related operations costs.

INTRODUCTION The capability to predict and certify the total containment of an underground nuclear test is a testing requirement. The understanding of contained underground nuclear testing has evolved from intuitive ideas to experimentally backed computational models and Jjn situ geophysical measurements. One method of obtaining geophysical measurements has been borehole well logging. Los Alamos started an experimental program to augment the current Nevada Test Site (NTS) logging capabilities in 1980. The project was designed to give the Los Alamos containment community a state-of-the-art experimental and production logging capability. The major effort to date has been concentrated on system design and procurement. By the end of FY 85 the effort will have switched to 70% experimental support and production logging and 30% tool research and development.

WELL LOGGING SYSTEM The present well logging system consists of a logging truck, a Digital Equipment Corporation (DEC) PDP 11/34 computer system, and an operations building. The logging truck and draw works are fitted with 6000 ft of MC-40 logging cable. This 0.701-in.-diam cable has 16 isolated conductors and a television coax surrounded by a multistranded steel outer armor. The cable is terminated with a 3-l/2-in.-diam cablehead that is compatible with Lawrence Livermore National Laboratory and Los Alamos well logging equipment.

9b The logging truck instrumentation compartment contains the computer system and logging instrument panels. The computer system is a DEC POP 11/34 mainframe with: an HP 2623A graphics terminal and printer/plotter; a VT-100 terminal; two RLO2 disk drives; one dual floppy disc drive; one Versatec V-80 printer/plotter; and a microprocessor-based depth system. Figure 1 is a block diagram of the system hardware. The computer uses a standard DEC operating system, RSX11-M plus Version 4.0, and has "user friendly" log data acquisition playback and cross correlation programs. The logging tool instrument panels are interfaced to the computer through plug-in interconnect modules and generally remain in the logging van. Development work on the system is based in an operations building, currently in Area 3 at NTS. This building houses equipment for the development, repair, and maintenance of the logging system and logging tools.

LOGGING TOOLS UNDER DEVELOPMENT There are two classes of logging tools under evaluation and development— the industrial production logging equipment and the containment/volcanic specialized logging equipment. The industrial production logging equipment has been extensively used on site by the logging contractor. Our similar tools are being experimentally used to identify design changes and recommendations for modification to the logging contractors' equipment owned by the Department of Energy. The present industrial production logging equipment is limited to: - Borehole temperature log (Bell Petroleum System) - Fluid ID/water locator logs (Bell Petroleum System) - Drill collar locators (Bell Petroleum System) - Gamma ray logs (Bell Petroleum System/EG&G Geometries) - Gamma ray-neutron logs (Bell Petroleum System) - Gamma ray-epithermal neutron logs (Bell Petroleum System)

The addition of a gamma-gamma density log in the near future has been proposed. The containment/volcanic specialized logging equipment will provide logs with both proven and yet unproven containment certification value. These types of logging services in some cases are available from private logging contractors, but scheduling and experimental requirements have necessitated procurement. The experimental capabilities under development include: - Slimhole gravity log (LaCoste and Romberg) - Bighole gravity log (LaCoste and Romberg) - Orientated three axis flux-gate magnetometer logs (Humphrey, Inc.) - Proton precession magnetometer log (EG&G Geometries) - Magnetic susceptibility/conductivity log (Simplex Manufacturing Co., Inc. - Induced polarization log (Mount Sopris Instrument Co.) VERSATEC V-80 PRINTER/PLOTTER INSTRUMENT SERIAL INTERFACE PORTS (2 ea.) RL02 DEPTH SYSTEM 10 CHANNELS DISC DRIVES Z-80 BASED

PDP - 11/34 INSTRUMENT o PARALLEL INTERFACE o PORTS COMPUTER 4 CHANNELS

DUAL FLOPPY HP2623 GRAPHICS INSTRUMENT DISC DRIVE TERMINAL ANALOG INTERFACE PORTS VT-100 GRAPHICS 8 CHANNELS TERMINAL

Figure 1. NTS Well Logging Van Computer System. - Self potential/resistivity log (Mount Sopris Instrument Co.) - Spectral gamma log (EG&G Geometries) - Borehole color video and 35-mm photography log (Sea Hawk Marine, Inc.) - Bighole acoustic interval velocity log (Internal) - Bighole oriented skid (Internal)

We anticipate that these specialized logs will improve the capability to certify borehole: - Formation density - Formation impedance - Formation acoustic velocity - Formation clay content - Formation water content - Formation porosity - Lithologic unit identification - Fault location

PROPOSED LOGGING TOOLS Future development effort for containment-related research is proposed for the following types of well logging equipment: - J_n situ stress/strength logging system - In situ permeability logging system - Slimhole spectral gamma logging system ^post shot) - Gyro directional system - Bighole nuclear logging skid - Slimhole acoustic velocity log

RESULTS As the above equipment is fielded and the logging data is folded into the Los Alamos data base system, a new confidence level in containment data will be achieved. This will result from the experimental work that calibrates the new logging equipment to the NTS volcanic environment and the implementation of computer-aided, well log analysis techniques. The computer automation of the well log analyses has improved the well log interpretation quality and efficiency. The improved well logging capability enhances the capability to understand, predict, and certify the total containment of underground nu~lear explosive testing.

101 BIBLIOGRAPHY Adams, John A. S., and Paolog Gasparini, Gamma-Ray Spectrometry of Rocks, Elsevier Publishing Company, 1970. Banerjee, Subir J., Advances in Geophysics, Vol. 23, Publication No. 1036, School of Earth Sciences, Dept. of Geology and Geophysics, University of Minnesota, Minneapolis, MN. Daniels, Jeffrey J., Interpretation of Well Logs for Mineral Exploration and Evaluation, Short Course Notes, November 10-12, 1981, Denver, CO. Gamma-Ray Logging Workshop, Grand Junction, Colorado, February 17-19, 1981, Bendix Field Engineering Corporation and U,Z. Department of Energy. Hearst, J. R., and R. C. Carlson, Geophysical Logging Services Available to LLNL, Los Alamos National Laboratory Report No. UCID-16603, Rev. 1 (Distribution Limited), June 1980. Hearst, Joseph R., Richard C. Carlson, and John T. Rambo, K-Division Experi- mental Work, Part 1: Geophysics Research for Underground Explosion Tech- nology, Los Alamos National Laboratory Report No. 1970-UCRL-50855, April 17, 1970. Hilchie, Douglas W., Applied Openhole Log Interpretation, Dept. of Petroleum Engineering, Colorado School of Mines, Golden, CO, 1978. Marine Technology Society, San Diego Section, Remotely Operated Vehicles Conference and Exposition, March 14-17, 1983, San Diego, CA. Monterey Containment Symposium Proceedings, Monterey, California, August 26-28, 1981, sponsored by Los Alamos National Laboratory, Los Alamos, New Mexico, Report No. 87545-LA-9211-X, Vol. 1 Conference, issued February 1983. Nettleton, L. L., Monograph Series, No. 1, Elementary Gravity and Magnetics for Geologists and Seismologists.. Southwest Research Institute, Technical Report and Instruction Manual, SWRI Project No. 14-4924, San Antonio, TX, February 1979.

102 LA-UR-B3-2267

Los Alamos National Laboratory is operated by the University of California for the United States Department of Energy under contract W-7405-ENG-36.

TITLE: A POSSIBLE NEW AC0U3TIC LOGGING PROBE FOR NTS

AUTHOR(S): Ravi Batra

SUBMITTED TO: 2nd Containment Symposium, Albuquerque, August 2-4, 1983

By acceptance of this article, the publisher recognizes that the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form o1 this contribution, or 1o allow others to do so. for U.S. Government purposes The Los Alamos National Laboratory requests that the publisher identify this article as work performed under the auspices of the U.S. Department of Energy.

Los Alamos National Laboratory Los Alamos.New Mexico 87545

FORM NO 836 H4 ST NO 2629 5/81 103 A POSSIBLE NEW ACOUSTIC LOGGING PROBE FOR NTS

Ravi Batra Geophysics Group Los Alamos National Laboratory

ABSTRACT

The Lawrence Livermore National Laboratory Modular Borehole Seismic Reflection Tool was tested at the Nevada Test Site in hole Ue7i to determine its suitability as an acoustic logging device in air-filled holes. This tool was originally designed to detect gas-producing fracture zones in the Devonian shale. Limited use of the tool for its intended purpose indicated its possible adaptation for the nuclear testing program.

The probe is 11.4 cm (4.5 in.) in diameter, consists of one or more source and receiver transducers made of piezoceramic elements, and is capable of operating in wet or dry boreholes up to 25 cm (9-7/8 in.) in diameter. The probe incorporates effective vibration isolation between source and receiver modules and has hydraulically operated pistons on these modules to provide borehole clamping. As deployed in hole Ue7i, with three receiver modules separated from the transmitter module by distances of 1, 3, and 6 m, respectively, the tool stood about 11 m tall and weighed about 340 kg (750 1b).

Preliminary results from the experiment indicate that the receiving transducers discriminate against compressional P-waves when used as a borehole velocity measurement probe. P-waves could be picked irregularly with some degree of certainty in the Tertiary tuffs, but not at all in the alluvial section. Shear waves could generally be identified with confidence within the tuffs, but not in the alluvium where coupling of the transducers with the borehole wall was particularly difficult. Shear wave velocities determined between any combination of receiver transducers were generally precise to within a percent or two. The operation of the tool was tedious and involved delays of 5-10 min between readings to allow Tor extension and retraction of the pistons.

Several modifications are suggested to incorporate this tool as a velocity measurement probe.

INTRODUCTION

The Containment program at Los Alamos recognizes the need to measure in situ interval velocities in emplacement and/or exploratory holes at finer Tntervals than can be accurately determined from the up-hole geophone survey. This need stems from the requirement to determine sharply defined acoustic impedance boundaries and their effects on stress wave propagation, an important consideration in containment evaluation. Furthermore, acoustic (or seismic)

104 velocities constitute primary data in the generation of synthetic seismograms that are widely used at Los Alamos to aid the interpretation of seismic reflection surveys. Since several of Los Alamos' nuclear tests are conducted below the static water level, we have been investigating the acquisition of an acoustic logging tool that can acquire data in fluid-filled and dry holes.

Acoustic logging tools that can operate in wet and dry, air-filled holes are not commercially available. Within the country^ only" Southwest Research Institute (SRI), San Antonio, Texas, has previously built such a tool. SRI, in fact, has built three similar tools for use in geologic environments quite different than that prevailing at the Nevada Test Site (NTS), all for totally different applications: for the U.S. Bureau of Mines, Denver; for the U.S. Geological Survey (USGS), Denver; and for Lawrence Livermore National Laboratory (LLNL).

The Bureau of Mines, Denver, has used acoustic velocity measurements for both P- and S-waves to contribute substantially to improved geologic interpretation of coal stratigraphy and detection of subsurface voids. The tool built for LLNL in 1978 was designed to detect fractures around wellbores in the Devonian shale as part of the gas flow stimulation program. This LLNL tool has been underutilized over the years and had never been used in a downhole configuration. In cooperation with LLNL personnel, it was decided to determine its suitability as an acoustic logging device for deployment downhole in the NTS environment with the idea of adapting it for use in the weapons testing program.

TOOL DESCRIPTION

This acoustic logging system is of modular design and may be deployed downhole with one or more sources and a multiplicity of receivers at various separation distances. It can operate in dry or wet boreholes to an effective pressure depth of 500 psio (approximately 1100 ft of water) and over a temperature range of 0 to 85 C. The transmitter and receivers are piezoceramic crystal stacks contained within hydraulically operated pistons (Fig. 1). The probe operates by clamping these pistons against the borehole wall and pulsing the transmitter stack. The system utilizes solid-to-solid coupling to transmit the signal through the formation, and thus fluid is not required as REACTION PISTONS with other acoustic logging probes (Fig. 2). Opposing reaction pistons maintain the probe essentially centralized in the hole. The system incorporates effective (>120 db) vibration isolation between source and receiver modules using a non-metallic seismic isolator (-Fig. 3). Without this feature the "tool signals" would TRANSDUCER seriously mask the desired signals that propagate through the geologic Figure 1, Piezoceramic crystal stacks material. The surface control unit within hydraul ically operated (Fig. 4) is used to operate the pistons. downhole probe hydraulics, indicate

105 the probe status, and condition CABLEHEAD the signals received from the UPPER ASYMMETRIC downhole receiver modules. The BOWSPRING signal averager digitally sums the CENTRALIZER CONTROL MODULE selected number of receiver wave- REACTION PISTONS forms to be averaged (from 2 to 32 "MODULE COUPLING waveforms may be summed). The digitized sampling rate is also adjustable from 10,000 to 320,000 DETECTOR MODULE *3 1 m samples per second. The digital output of the signal averager may be recorded onto a suitable DETECTOR MODULE +2 recording device, in this case a 1 m digital tape recorder. However, the writing of each record onto magnetic tape had to be performed DETECTOR MODULE # manually. The control module directly below the top centralizer ACOUSTIC ISOLATOR-^ receives commands from the surface MODULE (FLEXIBLE) 2 m control unit and converts them to control functions for the other I transmitter and receiver modules. SOURCE MODULE The probe is 11.4 cm (4-1/2 in.) MODULE COUPLING in diameter and is designed to operate in holes up to 25 cm REACTION PISTONS (9-7/8 in.) in diameter. As de- LOWER ASYMMETRIC ployed in hole Ue7i, with three BOWSPRING v receiver modules separated from CENTRALIZER BOREHOLE 6-10 (in) the transmitter module by distances of 1, 3, and 6 m, respectively, the tool stood about 11 m tall and weighed about 340 kg. Figure 2. Physical -layout for the seismic velocity logging probe.

OPERATION

The tool was deployed in hole Ue7i in air only at 30 different stations in the depth ranqe from 1570 to 210 ft. The. logistics of using the tool were tedious. Assembly of the entire tool took Figure 3. Acoustic isolator module. almost a half-day. In operation, we generally averaged 10-15 min per station in the borehole for recording data from all three receivers. A considerable portion of this time was spent in delays between stations to allow for extension and retraction of the hydraulically operated pistons. In the extension mode, these pistons operated in sequence starting with the uppermost transducer, that is, the transducer nearest to the control module. The manually operated digital tape recorder exacerbated the delays in acquiring data. This implies that to log a typical NTS borehole about 2000 ft deep, logged every 10 ft, wo'jld take 40-50 h. Furthermore, the hydraulic system seemed to tire out toward the end

106 DISPLAY CABLE LINES PROBE TRANSMITTER RECEIVER SIGNAL AVERAGER XMTRMOOE SUMMATIONS SAMPLE RATE FILTERS O O m O O * ^ i * o o GAIN ••©• HOH PASS LOW PASS O O RCVR GAIN IdB) O O . * . O (UPHOLD O O XMTH POWER DISPLAY .0. •£>• -0 OFFSET ncvnaEtECT OUTPUT XMTR SEIECT o o o Q o o o o o o o o o

Figure 4. Surface control unit—front panel controls. of the experiment and became more sluggish. It was often difficult to obtain good coupling of the transducers with the borehole wall, even in the tuffs under apparently good borehole conditions. This was next to impossible in the alluvium. It is suggested that the problem may be too much, rather than too little, pressure being exerted by the pistons forcing the soft rock material to "give" somewhat and causing the clamping mechanism to sense a "void."

DATA ANAI YSIS AND RESULTS

Full waveform records were obtained from each of the three receivers at 30 depth stations in hole Ue7i. Signal gains were adjusted for each receiver to provide distortion-free recording without compromising first breaks. The digital output of the signal averager was transferred to magnetic tape. Polaroid oscilloscope photographs were also taken at each station. Fig. 5 is a typical display of the recorded waveforms from each of the three receivers. The first arrivals were thought to be P-waves; velocity analysis later suggested that these are possibly shear waves. These waveforms were amplified on a computer CRT display to see if P-waves could be identified at all in the earlier part of the record. They could be picked irregularly in the tuffs but not at all in the alluvial sections (Fig. 6). It was concluded that this tool discriminates against P-waves traveling along the axis of the borehole. This is because the sensitive axes of the receiver transducers are oriented normal to the borehole wall. The tool was originally designed to be sensitive to reflections within the formations rather than waves traveling along the axis of the borehole. Shear waves could be identified with confidence in the tuffs but not in the alluvium. As pointed out earlier, it was difficult to obtain good coupling, in the alluvium5 of the transducers with the borehole wall to insure good noise-free recording. The signal averager was used to improve signal-to-noise ratio for each receiver output. Depending on operator judgment, 8, 16, or 32 waveforms were summed. No significant improvement was apparent and it was not possible to eliminate the noise in the early part of the record. This noise is presumably feedover of electromagnetic interference from the high voltage transmitter. It may also be self-noise of the signal amp!ifier.

107 X toR3

D= 1,400 (ft)

S=160 (kHz) 1st ARRIVALS

D= 1,450 (ft) D = 1,400 (ft) S= 160 (kHz) S = 160 (kHz)

o 0C

X to Rn D= 1,480 (ft) JXXoJ*l S= 160 (kHz) to R,

0 200 400 600 800 1000 WAVEFORM SAMPLE POINTS AFTER XMTR PULSE 0 200 400 600 800 1000 WAVEFORM SAMPLE POINTS AFTER XMTR PULSE

L_ ,_, Tlll_ _,_. SAMPLE POINTS "I (INCREASED AMPLITUDE TO ILLUSTRATE BODY WAVE ARRIVALS) [TRAVEL TIME DELAY-SAMPL|NGRATE(S)| Figure 5, Display of recorded velocity measurement Figure 6. Display of recorded velocity measurement waveforms, Borehole Ue71. waveforms, Bordhole Ue71. Increased amplitude illustrates body wave results. Pulse propagation velocities were computed from onset of signal from various receiver combinations (between receivers 1 and 2, 2 and 3, and 1 and 3) for both P- and S-waves. For S-waves, the three computed values of velocity are generally in good agreement (Fig. 7). The scatter is somewhat larger for P-waves, but still fairly consistent for each depth station (Fig. 8). Shear velocities were also derived from later phases in the wave train by correlating first and second positive oeaks between receivers. These were compared with the velocities derived from the onset on the signal. The corresponding velocities are in excellent agreement (Fig. 9), suggesting preservation of the first few cycles of the signal waveform out to at least 6 m, that is, the separation of the furthest receiver from the source. Lack of apparent velo-ity dispersion suggests that welibore surface waves or guided "tube" waves are probably absent in the first few cycles. As mentioned earlier, because of constant adjustments to receiver gains, no attempt was made to study attenuation characteristics.

i 1 r 1 r

1800 1800 I I I 700 900 1100 1300 1500 1700 1000 1400 1800 2200 2600 3000 VELOCITY (m/s) VELOCITY (m/s) Figure 7. Velocity vs depth profile Figure 8. Velocity vs depth profile derived for shear waves, derived for P-waves.

109 600

1OOO

Q. LU a 1500 - V1>3 FROM ONSET OF SIGNAL

V1f3 FROM 1st POSITIVE PEAK

Vi,3 FROM" 2nd" pbsfflVE PEAK 2000 "500 700 900 1100 1300 1500 1700 VELOCITY (ms) Figure 9. Velocity comparisons for various phases between receivers 1 and 3.

To be effectively used as a velocity measurement probe at NTS, the following modifications to the tool would have to be made: t The receiver transducer crystals should be replaced with a piezoceramic stack oriented to respond preferentially to compressional waves. As an option, to detect shear waves as well, another set of crystals separately and selectively sensitive to shear waves could be included. This would provide two channels of data for each receiver. • The pumps to the hydraulic pistons are presently designed to operate sequentially from the topmost transducer. To reduce the probe clamp time, these motors could be modified to operate simultaneously. Also, higher speed gears could be installed to reduce individual clamp time for each transducer. • The dynamic range of the system could be improved by replacing the present analog-to-digital converter of 12-bit accuracy with one of 16-bit accuracy. The signal averager could be removed altogether. Summing of waveforms, if necessary, could be achieved with a microprocessor or computer before being transferred to magnetic tape. • There should be improved signal cabling and shielding in the probe modules to reduce electromagnetic feedover from the transmitter. Also, improvements in the signal amplifier to reduce self-noise to the lowest practical level are necessary. • There should be automatic entry into each digital record of all probe operating parameters such as probe and receiver gains, filter settings,

110 probe status (clamped or void), etc. The surface control unit should be redesigned with BCD readouts and the electronics modified to include an automatic gain-ranging device. All these modifications should speed up the data acquisition process. • A real-time caliper measurement should be incorporated for use with this tool to seek out the smoother portions of the borehole. 0 A big-hole logging skid for this tool should be designed.

SUMMARY 1. The source of the modular borehole probe can emit high frequency (5 kHz) acoustic pulses capable of being detected by receivers separated by a distance of at least 6 m from the source. 2. In their present orientation, the receiver transducers discriminate against the detection of compressional (P) waves. 3. Shear waves can be identified with confidence in the tuffs but not in the alluvium. 4. The early part of most records are marked by coherent noise that could not be eliminated by simple stacking in the signal averager. 5. For shear waves, pulse propagation velocities derived from the onset of the signal between any combination of receivers were s/ery consistent. Velocities derived for compressional waves, when detected, were reasonably consistent. 6. There was no interference from wellbore surface waves, at least in the first two or three cycles of the detected waveforms. 7. The operation of the tool was tedious and involved delays between readings to allow for extension and retraction of the hydraulically operated pistons. 8. It was often difficult to obtain good coupling of the transducers with the borehole wall, even in the tuffs and especially in the alluvium. 9. To be effectively used as a velocity logging device, several costly modifications would have to be incorporated.

Ill BIBLIOGRAPHY

Carroll, R. D., 1979, Results of Shear- and Compressional-Wave Studies in the Diablo Hawk TRCX-1 and TRCX-2 Drill Holes Using a Geophysical Logging Probe, Memorandum dated September 20, 1979, to Commander, Fieid Command, DNA, Kirtland AFB, New Mexico (unpublished).

Carroll, R. D., 1980, Results of Velocity Logging in V12n.l2 UG-1 with Dry- Hole and Wet-Hole Tools, Memorandum dated February 8, 1980, to Commander, Field Command, DNA, Kirtland AFB, New Mexico (unpublished).

Kirk, G. W., and F. X. Herzig, 1983, Demonstration of Borehole Acoustic Velocity Measurements at the Nevada Test Site, Final Technical Report, Southwest Research Institute, San Antonio, Texas, Project No. 14-7299-007.

Owen, T. E., S. A. Suhler, and J. H. King, 1975, Analysis of Borehole Seismic Velocity Tests—Nevada Test Center, Final Technical Report, Southwest Research Institute, San Antonio, Texas, Project No. 14-4095-002.

Snodgrass, J. J., 1982, A New Sonic Velocity-Logging Technique and Results in Near-Surface Sediments in Northeastern New Mexico, U.S. Dept. of the Interior, Bureau of Mines and Minerals Environmental Technology Program, Technical Progress Report 117.

Suhler, S. A., W. R. Peters, E. C. Schroeder, and T. E. Owen, 1979, Modular Borehole Seismic Reflection Probe, Final Technical Report and Instruction Manual, Southwest Research Institute, San Antonio, Texas, Project No. 14_4924.

112 SPECTRAL GAMMA-RAY LOGGING FOR CLAY CONTENT

Huang Long-ji; East China Petroleum Institute Dongying, Shandong, People's Republic of China

Joseph R Hearst; Lawrence Livermore National Laboratory Livermore, California, USA

Introduction If one is to predict the phenomenology of an underground nuclear test, one must be able to characterize the site. Geophysical logs are run and samples are taken from drill holes. Data are processed, and interpretations of the processed data are combined with previous data to make a site characterization. One of the many parameters that are examined is clay content. Both the amount of clay and the stratigraphic thickness of the clay-bearing rock will affect the strength of the medium in which the test is conducted. We are concerned with weight fractions greater than about 0.2 of the clay minerals montmori1lonite, illite, and kaolinite, all of which occur at the Nevada Test Site. Montmori1lonite is by far the most common, and therefore has the largest effect on the strength of the medium. To evaluate the containment of radioactive products from a test the types of clay must be identified, its amount quantified, and its stratigraphic thickness determined. It is desirable to obtain the clay content from wireline logs rather than from samples, both because the cost is less and because the log is a continuous record. In the logging industry, two methods are commonly used to estimate clay content: electrical resistivity and total count from a natural gamma-ray log. Pawloski (1981, 1983a) has shown that at NTS, although intervals with high clay content exhibit low resistivity, so do many intervals with low clay content, and consequently low resistivity is not a reliable indicator of high clay content. Therefore it is desirable to use an additional method to confirm high clay content. Montmorillonic clay, however, does not always contain large enough combined amounts'of radioelements (uranium, thorium, and potassium) to distinguish it from other rocks, especially NTS tuffs, and consequently is not amenable to detection by total gamma-ray count. It has been demonstrated however, (Alexeev, 1973) that

113 -2- thorium content is sometimes useful as a clay content indicator. We are therefore investigating the use of gamma-ray spectral data as an indicator of clay content. This is a progress report on that investigation. Silicic tuff such as that which occurs at the Nevada Test Site often contains more thorium and uranium than the sedimentary rocks that occur at NTS, such-as sandstone and carbonates. In an oxidizing environment, such as might occur above the water table, uranium is in the hexavalent state, and therefore can form soluble compounds, whereas thorium is relatively insoluble. Consequently if the region above the permanent water table is temporarily saturated, and the tuff is altered by circulating water, the uranium can be leached out and the thorium left behind. This will lead to a high thorium/uranium ratio in the alteration product (Fertl, 1979) which, in our case, is montmorillonite.

Experiments In the present study, over 100 samples including Paintbrush, Belted Range, and Indian Trail tuff, from .line drill holes at the Nevada Test Site were analyzed for U, Th, and K abundances at the LBL low background garcma- spectrometric facility. The samples, ranging in weight from 2 to 200 g, were assayed by a NaI(Tl)-based gamma-spectrometer system, in which measurement precision varied from HH 10% for the smallest samples to _+ 1% for samples weighing at least 80 g. The clay content of 43 of these samples from two drill holes was determined by Pawloski (1983b) using X-ray diffraction analysis. Twelve samples from the same two drill holes had been analyzed earlier by the U. S. Geological Survey (McArthur, 1976), permitting a check on clay content. A check on the the thorium content was provided by X-ray fluorescence analysis of twenty of the 43 samples. The XRF results agreed well with the gamma-ray spectroscopy data. The gamma-ray and XRF data and weight fraction montmorillonite are listed in Table 1 and plotted in Figures 1 and 2, together with the clay content determinations. The thorium data in Fig. 1 do not show a systematic trend with respect to clay content. The uranium data in Fig. 2 suggest an inverse correlation with clay content for clay content less than about 0.15. Thus, taken singly, radioelement contents are not definitive of clay content. When clay content is plotted against the Th/U ratio in Fig. 3, there is a general trend of increasing clay content with increasing Th/U ratio, and a

114 -3- very dramatic increase in clay content when the Th/U ratio is greater than about 9. The data from these two drill holes indicate that most samples whose weight fraction montniori 1 lonite is greater than about. 0.4 have Th/U ratios greater than 9, while rocks with lower clay content have ratios that range down to less than 2. This suggests that a general threshold be considered: tuff in the unsaturated zone at NTS with a \iery high (greater than 9) Th/U ratio may have clay content greater than 0.4. Thus high Th/U ratio determined by downhole gamma-ray spectral logs might indicate high clay content in tuff of the unsaturated zone at NTS. For unaltered tuff at NTS, the Th/U ratio is generally less than 10 (Vogel, 1983). Table 2 shows correlation coefficierits for U, Th, K and combinations of these with clay content for all 43 samples. Notice that the ratio Th/(U+K) correlates as well as Th/U. Figure 3 also shows data from six other samples whose clay content has been determined by Pawloski. These samples come from five other holes at NTS. Although the samples were too small to obtain Th/U ratios with an accuracy better than +20%, they agree well with the other data.

Future work Evidently, materials with high montmori1lonite content often exhibit both low resistivity and high Th/U ratio. Although neither of these is an unambiguous indicator of high clay content, if they are used in combination they may be more definitive. Therefore several tasks must be accomplished before logs can be used as reliable clay indicators at NTS: 1. More experiments like those described above must be performed on samples from more holes at NTS, to confirm the relationship between Th/U and clay content. It is likely that a more sophisticated algorithm, involving K as well as Th and U, can be found to obtain better values of clay content from gamma-ray spectral log data. 2. We need to be able to get high-quality gamma-ray spectral logs. 3. We must compare clay content, gamma-ray spectral data, and resistivity in the same holes to learn if the combination can improve identification.

Acknowledgement This work could not have been accomplished without the detailed advice and cooperation of H. Bowman, A. Smith, and H. Wollenberg. Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract No. W-7405-ENG-48. .,/

115 -4-

References: Alekseev, F. A. ; "Application of Nuclear Methods in Petroleum Geology", Nedra, Moscow, 1973 (in Russian) Fertl, W. H.; Gamrna Ray Spectral Data Assists in Complex Formation Evaluation Log Analyst Vol 20, No. 5 p. 3S 1979 McArthur, R. D. ; "The Subsurface Geology of the Lawrence Liverrnore Laboratory Portion of Area 4 at the Nevada Test Site" Lawrence Livermore Laboratory Report uCRL-52061, 1976 Pawloski, G.A.; Resistivity Measurements as an Indication of Clay and/or Zeolite Content in Tertiary Volcanic Tuffs and Quaternary Alluvium at the Nevada Test Site, Proc. Monterey Containment Symposium, 1981 Pawloski, G.A.; Personal Comiiiunication 1983a Pawloski, G. A.; Quantitative Determination of Minerals in Nevada Test Site Samples by X-rsy Diffraction, Proceedings Second Symposium on Containment of Underground Nuclear Explosions, Albuquerque, 1983 Vogel, To ; Personal Communication, 1983

116 -5- TABLE 1 Uranium, thorium, potassium, and montmori1lonite contents from 49 NTS samples

Hole Depth (ft) From Gamma-ray Spectroscopy From XRF From XRD U(ppm) Th(ppm) K(%) Th/U Th(ppm) Wt. Fr. Mont. Ue4AF 200 2.99 8.76 1.44 2.93 6.76 0. 300 3.06 8.16 1.22 2.67 7.60 0 475 3.03 14.03 1.37 4.63 12.39 .19 550 2.57 13.04 1.15 5.07 12.60 .34 650 2.05 17.42 2.18 8.51 18.14 .19 700 3.02 8.98 1.01 2. 7 7.14 .04 825 0.93 21.55 0.99 23.25 20.88 .82 850 1.25 14.35 0.92 11.41 12.07 .45 925 0.93 17.72 1.28 18.98 15.96 .42 950 1.67 20.69 1.49 12.33 .86 1000 0.89 15.53 0.40 17.38 17.76 .84 1025 1.76 18.25 1.17 10.36 .80 1050 1.78 20.29 1.73 11.40 20.17 .82 1100 2.46 19.18 2.29 7.80 20.12 .78 1175 1.77 23.70 1.18 13.41 23.26 .87 1220 1.28 16.06 1.77 12.51 15.94 .52 1240 2.03 18.55 4.38 9.13 20.92 .75 1300 2.97 19.68 3.10 6.63 .67 1320 3.25 18.23 3.49 5.61 20.13 .18 1340 2.07 18.24 2.05 8.82 .58 1360 2.39 12.01 2.29 5.02 11.60 .24 1400 1.08 17.90 1.27 16.63 18.63 .43 1440 3.93 10.98 1.13 2.79 10.55 0 1460 J.73 13.08 1.51 3.50 12.31 0 Ue4AC 175 4.19 10.67 2.07 2.55 0 300 3.83 14.50 2.38 3.78 .07 375 2.78 18.82 3.46 6.77 0 450 4.08 18.90 3.59 4.63 .04 600 4.38 18.50 3.75 4.22 0 700 6.92 20.05 3.80 2.90 0 780 5.08 19.09 2.79 3.75 .19 860 3.91 22.58 3.73 5.78 .13 920 4.69 21.78 3.d4 4.64 .04 980 4.10 6.56 3.16 1.60 .03 1100 3.09 16.63 3.01 5.38 .06 1240 3.05 17.88 2.91 5.85 .14 1380 2.02 15.32 1.78 7.57 .36 1440 1.92 16.03 2.02 8.36 .25 1500 5.03 15.05 1.20 2.99 .49--.70 1540 2.66 14.76 1.95 4.63 .12 1650 2.57 13.90 1.91 5.AT .31 Ue2C0 1782 6.11 44.66 2.84 7.30 .50--.75 U2ET 600 5.26 21.34 4.08 4.06 .1 U4AJ l?50 8.81 35.82 4.61 4.07 .25--.35 U4AK 1250 3.43 17.44 4.40 5.09 .15--.25 U4AK 1300 6.42 27.70 4.62 4.31 .15 U10CA 570 4.24 23.00 3.70 5.42 .35--.50 U19AC 1350 12.51 24.47 3.99 1.9b .1

117 -6-

TABLE 2 CORRELATION COEFFICIENTS R U Montmorillonite -.64 Th—Montmorillonite .46 K Montmori llonite -.36 Th/U---Montmorillonite .73 Th/(U+K)—-Montmorillonite .73

118 -7-

Fig. 1 Thorium vs. measured weight fraction montmori 1- lonite for 43 samples

.10 .20 30 40 50 60. .70 .80 .90 1.00 CLAY

Fig. 2 Uranium vs. meas' red weight fraction montmori1- lonite for 43 samples

10 .20 .30 .40 .50 .60 70 .60 .90 1.00 CLAY 119 Legend 20 x UE4AF + UE4AC A U4AK ID • U2ET JO . U4AJ o UIOCA a UE2CO • UI9AC

0 05 1.0 Fraction Montmor i I ion ite (measured)

Fig. 3 Thorium/uranium ratio vs. measured weight .^action montmoriilonite for 49 samples

120 LA-UR-S3-2263

Los Alamos National Laboratory is operated by the University of California for the United States Department of Energy under contract W-7405-EMH-36

TITLE: LITHOLOGY AND LOG COMPARISON OF SOME BOREHOLES IN SOUTHERN YUCCA FLAT

AUTHOR(S): Mark A. Mathews and Carol M. LaDelfe

SUBMITTED TO: 2nd Containment Symposium, Albuquerque, August 2-4, 1983

By acceptance of this article, the publisher recognizes that the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or lo allow others to do so. for U.S. Government purposes.

The Los Alamos National Laboratory requests that the publisher identify this article as work performed under the auspices of the U.S. Department of Energy.

Los Alamos National Laboratory Los Alamos.New Mexico 87545

FORM NO. 836 R4 121 ST. NO. 2629 5/81 LITHOLOGY AND LOG COMPARISON OF SOME BOREHOLES IN SOUTHERN YUCCA FLAT

Mark A. Mathews and Carol M. LaDelfe Geophysics Group Los Alamos National Laboratory

ABSTRACT

Logs acquired from five exploration boreholes and one emplacement hole in Southern Yucca Flat are compared and correlated amongst themselves and with the lithology from these boreholes. The exploration holes (Uelq, Ue7i, Ue7j, Ue7k, and Ue71) and one emplacement hole (U31a) were logged by Birdwell and Bendix Field Engineering Corporation (BFEC), although only one log was acquired by BFEC for the emplacement hole U31a. The logs compared are:

Birdwell Logs Bendix Logs

Gamma Ray Gamma Ray Density Spectral Gamma Caliper Caliper Resistivity Conductivity Epithermal Neutron (one) Thermal and Epithermal Neutron Magnetic Susceptibility

The logs have been corrected for borehole size variation and smoothed or filtered with a 50-point running average. These logs have been correlated with the lithologic intervals of their respective boreholes and show a reasonable correspondence between boreholes. The magnetic susceptibility log response was used to identify and correlate various units in the Timber Mountain tuffs and other tuffaceous lithology. Preliminary crossplots of density vs magnetic susceptibility of some units indicate a correlation that appears to correspond to magnetite content and magnetite grain size within these units. A thorium anomaly zone obtained from the spectral gamma log lies above the Ammonia Tanks member of the Timber Mountain tuffs in borehole Ue7i. This zone correlates to a low thermal neutron response (rare earth elements and neutron absorbers present) and a high epithermal neutron response (puzzling result?). Results of the other logging responses are correlated and compared to the geology and to each other.

I. INTRODUCTION

Numerous jn_ situ physical properties are required to adequately document the subsurface geology for containment purposes. A detailed knowledge of these physical properties is mandatory to ensure safety during and after the detonation of an underground nuclear event.

122 Wireline well log measurements acquired in exploratory (borehole diameter of 6 to 18 in.) and emplacement (borehole diameter of 48 to 120 in.) boreholes are used to resolve problems and make predictions for containment and engineering purposes. The problems to be solved by these measurements are: 1. Lithological identification of formations traversed by the drill; 2. Detection of certain minerals or materials (clay, water content, carbonate, etc.) either penecrated by the borehole or located close to it; 3. Subsurface correlations to assist with structure and isopach mapping and identification of discontinuities and structures (faults, unconformities, etc.); 4. Containment diagnostics or problems in the vicinity of the cavity region (strength of materials, tensile rarefaction energy, plasticity of surrounding rock, etc.); and 5. Drilling and emplacement operation difficulties. A variety of wireline well logs were acquired in Southern Yucca Flat from five exploration boreholes and one emplacement borehole using Birdwell and Bendix Field Engineering Corporation (BFEC) logging subcontractors. The genera, geological settings, the lithology column, and the quality of the acquireci wireline well logs are discussed in the first part of this paper. Then a comparison is made of similar logs obtained from the six boreholes by both logging subcontractors. These comparisons show differences because the logging probes of each subcontractor are slightly different and because the borehole environment" represents a dynamic system that is affected by drilling substances. Similar logs acquired at different dates respond to this changing dynamic system with different log responses.

The last part of this paper compares various log responses. Trends and linear relationships are analyzed between log responses and different tuff units. Materials, such as clay and rare earth element zones, were identified from log responses or descriptions, and their exact locations in the boreholes were determined from wireline well logs. These materials were generally described in the lithologic log and are items of interest for containment purposes.

II. GEOGRAPHIC AND GEOLOGIC SETTING The holes included in this study are located in west-central Area 7 and adjacent portions of northern Areas 1 and 3 (Fig. 1). The general lithologic logs of these boreholes are shown in Fig. 2. The Yucca fault and a buried, gravity-inferred fault lie along a north-south trend northward from the vicinity of U31a (Fig. 3). Quaternary alluvium ranges in thickness from 118 (Ue7j) to 389m (Uelq). Tertiary tuff units overlie Paleozoic limestones and dolomites. The surface of the Paleozoic rocks varies from 652 to 738 m below the ground surface and is offset by the Yucca fault east of U31a and the buried fault, west of U31a (Fig. 4). Ue71 intersects a branch of the Yucca fault, which offsets the Paleozoics and the overlying Tertiary tuff units (Fiq. 5). Depths to the bases of Cenozoic units and top of Paleozoic rocks in each hole are shown in Table I.

123 © UE7i © UE7£

SCALE METERS 0 2000

Fig. 1. Location map. UE1q U3la UG7k UE7j UE7i UE7I

. • QAL i." •. ,-j 0»L I. • « \*i • ;;. : :..•••

Fig. 2. LITHOLOGY

I" •••. ••,)

F/V»*1 ALLUVIUM Ln

'.'•'.'•:'•:'•'!, TERTIARY VOLCANICS

HP^ PALEOZOIC CARBONATES E 207,000 m £680,000 E 200,000m £685,000

-N847.000

-N258,000m

CT*

-N257,000 m -N843,000

I T Fig. 3. -Isopach map of U31a site and vicinity. U3la EAST-WEST CROSS SECTION

WEST Projected Projected EAST A U7aq U3la U7k 1200J

1000J

PAINTBRUSH TUFF

GROUSE CANYON

TUNNEL BEDS AND OLDER

-1000 -INFERRED FAULT TOPGALLANT YUCCA FAULT FAULT SYSTEM 1103.4m T.D. \ \ SYSTEM -1200 500 1000 SWL = Static Water Level

Meters Fig. 4. U31a east-west cross section. E680.000 E685.OOO

UE7I

-4000

1000- QohTv /toeki (undivldtd) -3000

05 -2000 O o 500-

• > -i UJ I'J -1000 UJ Pt Reck*

4// dtpthn In maltr* S.L.. .S.L.

Fig. 5. Generalized geologic cross section through the Ue71 site. TABLE I STRATIGRAPHY (Depth in Meters)

Hole Qal Tma Tmr Tp Tbg Pz (top)

Ue7i 125 ___ 215 276 305 721 Ue7j 118 148 270 365 390 702 Ue7k 1G2 239 293 401 424 738 Ue71 308 326 446 468 — 552 U31a 183 259 412 611 620 — Uelq 389 413 535 634 645 728

III. WIRELINE WELL LOGS COMPARED Logs run by Birdwell and BFEC are shown in Table II. Since the primary assignment for BFEC was to run spectral gamma, thermal and epithermal neutron, and magnetic susceptibility logs in these holes, the caliper logs were run for use in calculating corrections to the spectral gamma logs; BFEC electric and gamma ray logs were run in tandem with magnetic susceptibility and neutron logs. The quality of the Birdwell logs is fair to good. The repeat sections generally repeat very well and the log responses are generally free of high frequency jitter or noise. The only exceptions to this lack of noise are intervals in the Birdwell density logs and the Birdwell epithermal neutron log (tool problems and operator error). The quality of the BFEC logs is poor to good. Most of the logging data needs to be filtered to remove high frequency jitter or noise. It appears that the sensitivity is set too high in most cases (operator inexperience)."

TABLE II WIRELINE LOGS

Birdwell Logs Bendix Logs Gamma Ray Gamma Ray Density Spectral Gamma Caliper Caliper Resistivity Conductivity Epithermal Neutron (one) Thermal and Epithermal Neutron Magnetic Susceptibility

129 Also, part of the spectral gamma log results (borehole Uelq, Ue7j, and the upper 160 m of Ue71) are useless because the wrong size crystal had inadvertently been used (operator error). The magnetic susceptibility log had temperature drift problems (tool problem) and bad zero settings (operator error) in some intervals. Otherwise the quality of the data is reasonable.

IV. COMPARISON OF SIMILAR LOGS In this section, Birdwell and BFEC caliper logs and gamma ray logs are compared. Also, the Birdwell resistivity logs are compared with the BFEC resistance and conductivity logs. Reasonable comparisons in many cases are impossible because of tool differences or borehole changes. A. Caliper Logs vs Time Change of Boreholes Caliper logs are run by Birdwell during completion of each hole using a six-arm caliper tool (three independent arm-pairs). Therefore, the Birdwell caliper logs included herein (Fig 6) were run between November 10, 1980 (Uelq) and August 30, 1981 (Ue71). BFEC caliper logs (Fig. 7) were run in April 1982. Most differences between the two sets of logs are probably attributable to degradation of hole condition during the intervening time. The BFEC caliper log in Ue7i shows zones of noticeable shrinkage in hole diameter (to less than 15 cm from a bit-size of 22.2 cm) between the depths of 534 and 598 m. The areas of hole shrinkage alternate with some fairly significant enlargements (to at least 32.6 cm). Water level was at approximately 534 m when that log was run. Because no pure multi-arm caliper log was run at or near the same time as the BFEC spectral gamma/caliper, we cannot say whether the difference in the logs is due to tool type, orientation, time, or some of each (Hilchie, 1968). A Birdwell caliper log run (not shown), after the hole was deepened in April 1983, shows only the enlargements in the region from 534 to 598 m, but, again, this log was run immediately after redrilling and it indicates the general scouring and roughening of the hole by the deepening procedure. Spikes shown projecting to the left (into the hole) on the plotted logs (Figs. 6 and 7) are primarily the result of a problem with the plotting software.

B. Gamma Ray Logs vs Geology Birdwell ran gamma ray logs in Ue7i, Ue7j, Ue7k, and Ue71 (Fig. 8). Major changes in lithology with depth are easily discernible on all four logs. The Paleozoic limestone-dolomite response is significantly lower than that of the Tertiary sediments and tuffs immediately above. Timber Mountain tuffs have a consistently higher response than the Paintbrush or Grouse Canyon tuffs. In Ue7i and Ue7k, the Ammonia Tanks member of the Timber Mountain tuffs stands out distinctly, but in the Ue71 log there is no obvious division between the Ammonia Tanks and the Rainier Mesa members. There are several sets of minor signature responses on the logs through the older tuffs in Ue7i, Ue7j, and L)e7k that do not obviously appear on the Ue71 log and may have been lost or distorted in that hole due to faulting. BFEC gamma ray logs were run with both of the neutron logs in each hole (Fig. 9). Log characteristics are not as well defined as on the Birdwell logs because they appear to be much noisier. Litholoqic changes can be picked out if interface depths are already known. Minor signatures seen on the corresponding Birdwell logs cannot be picked out on the BFEC logs.

130 o

Hld3Q

131 c: o

8 8

132 cr, < Q

-„.».-.'.•••'-—•-'-

.'••'.• •-.'••'.•«

vl\

^:y

(<") HXd30

133 IK <

134 C. Resistivity and Conductivity Comparison Plots of the Birdwell resistivity logs are shown in Fig. 10. BFEC recorded resistance with the neutron logs and conductivity with the magnetic susceptibility log. The BFEC conductivity logs are shown in Fig. 11. Since the Birdwell 64-in.-normal resistivity log is run only in a wet-hole environment, we do not have these logs for the full depth of any of the holes. A conductivity log should be the inverse of a resistivity log within a scale factor. Where the two sets of logs overlap in depth, most attempts at comparison reveal very poor correlation. The BFEC conductivity logs appear to be influenced by temperature, show no significant features, and rarely display either the inverse character or the numerical inverse of the Birdwell logs. The conductivity log is insensitive to media with resistivities greater than 10 ohm-m (Scott et al., 1981), which includes most borehole environments at the NTS and therefore should be used only for monitoring system drift and noise.

V. INSPECTION AND COMPARISON OF VARIOUS LOG RESPONSES In this section the Birdwell resistivity logs and BFEC magnetic susceptibility logs are correlated with lithology and compared to each other. The magnetic susceptibility logs exhibit good characteristic curves in the older tuff units, such as the Tunnel Beds, and are easier to correlate than the resistivity logs. A trend between density and magnetic susceptibility is analyzed next with the help of crossp.lots. Linear relationships for the Rainier Mesa tuff unit and the Grouse Canyon tuff unit are given. A reasonable correlation between the potassium content from the spectral gamma log and the clay content described in the lithologic log is portrayed next. High potassium content correlates approximately two-thirds of the time with clay and argillized tuff in this preliminary study. Finally, a discussion and partial explanation is given in the last part of this section for an apparent correlation between an increased thorium content from the spectral gamma log, a decreased thermal neutron count rate, and an increased epithermal neutron count rate, which all occur at the same depth interval for borehole Ue7i. A. Resistivity and Magnetic Susceptibility Comparison Magnetic susceptibility logs were acquired in the five exploration holes and one emplacement hole by BFEC in September 1981. The magnetic susceptibility log and system has been discussed in the literature since 1952 (Broding, et al.). This logging system uses a ferrite-core inductor in one branch of a basic Maxwell bridge circuit to sense the effects of formation conductivity and magnetic susceptibility. The sensing coil is inductively coupled to the rocks surrounding the borehole and the recorded resistive and reactive signals represent, to a good approximation, the conductivity and magnetic susceptibility of these rocks. A complete description, along with the temperature compensation, is presented in Emilia e.t al. (1981) and Scott et al. (1981).

135 UE1q U3la UE7k UE7j UE7i UE7I

Fig. 10. RESISTIVITY BRDWEU. UE7k UE7J UE7I UE7I

Fig. 11. CONDUCTIVITY BENCHX

800 -101 2345-10 12345-2 4-2 0 2 4 6 8 10

137 The relationship between magnetic susceptibility and ferromagnetic mineral content varies with mineral assemblage and rock type and with grain size, shape, and orientation. There is a strong and sometimes nearly linear correlation between magnetic susceptibility and ferromagnetic mineral content (Rao, 1956; Negata, 1961; Kahn, 1962). The volcanic tuff uriits at the Nevada Test Site (NTS) have various amounts of magnetite, a ferromagnetic mineral. The magnetic susceptibility logs from the six boreholes are shown in Fig. 12. Large shifts in micro-cgs units (units defined in Broding, 1952) at shallow depths and the lack of instrument zero (negative micro-cgs) are caused by improper temperature regulation and operator error. Even with these problems, correlation between the boreholes (the amplitude is reduced on the emplacement hole U31a) is reasonably good in the tuff units. A pronounced anomaly between the Rainier Mesa tuff and the Paintbrush tuff units can be correlated amongst all the boreholes except borehole U31a, where the suspected anomaly is below the bottom of the log. These magnetic susceptibility logs exhibit distinctive and correlative characteristic curves in the Tunnel Bed tuffs also.

The resistivity logs routinely acquired at the NTS have been previously described (Carroll and Muller, 1973; Marusak, 1981). These logs are used for detailed stratigraphic correlation and structural interpretation. Resistivity logs were acquired in the five exploration holes and one emplacement hole by Sirdwell and are shown in Fig. 10. The resistivity logs correlate reasonably well among the various boreholes between the Rainier Mesa tuff and the Paintbrush tuff units. Correlation in the Tunnel Bed tuff units, below the Paintbrush tuff units,, is not as easy and is questionable in some places on these logs. The magnetic susceptibility probe works in both air-filled and fluid- filled boreholes with no change in sensitivity. This probe can be run continuously without stopping f-om the bottom to the top of the borehole without regard to the fluid level in the borehole. This operation cannot be done with the resistivity log. The resistivity log requires two separate and different probes (with different sensitivity and depth of penetration) to completely log a borehole that is partially filled with fluid. Also, a gap (due to probe design and operation) of approximately 60 ft exists between the results of these two probes in a typical borehole. B. Density Comparison and Correlation with Magnetic Susceptibility A borehole-compensated density log was acquired in all five of the exploration boreholes and an uncompensated density log was acquired in the emplacement hole* These logs were run by Birdwell and are shown in Fig. 13. High magnetic susceptibility and density anomalies occur on Ue71 at approximately 340 m (Rainier Mesa) and on Ue7j at approximately 180 m (Rainier Mesa). Crossplots of the magnetic susceptibility and density for the Rainier Mesa interval for boreholes Ue7j and Ue71 are shown in Figs. 14 and 15. Reasonable correlation is exhibited and a similar slope of the least squares fit of a straight line is shown. This implies a linear relationship between the magnetite content and the density in the Rainier Mesa unit as y = 697 x - 846 for Ue.7j, and y = 610 x - 981 for Ue71, where x is density in Mg/m3 and y is magnetic susceptibility in micro-cgs units.

133 UE1q U3la UE7k UE7j UE7i UE7I

i Fig. 12. MAGNETIC SUSCEPTIBILITY BENCHX

600- >r

800- 8 O Q O §° § §5 Si S! 2 8

MICRO-CGS UNITS B

HldBQ

140 UE7L DENSITY VS MRG. SUSCEPT.

in o

SLOPE - 610.11 INTERCEPT - -981.72

FIBS RESID - 0.086 RMS RESID - 0.122

1.2 2.4 MG/M3 RfilNlTR

Fig. 14. Crosspiot, UE7J DENSITY VS MRG. SUSCEPT,

o ' o a: o

SLOPE - 697.08 INTERCEPT - -846.17 S FIBS RESID - 0.097 RMS RESID - 0.125

1.5 2 2.5 MG/M3 RfllNlLR MESH

Fig. 15. Crosspiot. UE1Q DENSITY VS MflG. SUSCEPT,

ao:

SLOPE - 3804.31 INTERCEPT - -7158.16

0.034 RMS RES ID - 0.043

1.4 1.6 —1 2.2 2.4 2.6 GROUSE CANYON

Fig. 16. Crossplot. A crossplot of the magnetic susceptibility and the density tor the Grouse Canyon interval for borehole Uelq shows a \/ery good correlation and linear trend. This is exhibited in Fig. 16 where the least squares straight line fit gives y = 3804 x - 7158, where x is density in Mg/m^ and y is magnetic susceptibility in micro-cgs u..its. Th'is interval has a broad magnetic susceptibility span for a narrow density range. The magnetic susceptibility probe recorded negative micro-cgs units in several of the boreholes logged. This is due to a bad zero setting (operator error) and a poor temperature compensation (tool problem). Even with these problems, a trend between density and magnetic susceptibility is seen in several of the tuff units logged and is due primarily to the amount of magnetite present. Changes in magnetite grain size and orientation account for some of this correlation, and this characteristic is presently under investigation. C. Correlation of Potassium Content from the Spectral Gamma Log with Clay

A spectral gamma log was acquired in each of the five exploration boreholes. The BFEC spectral logging probe has two crystals, one of which is large (1-1/2 in. x 12 in.) and unfiltered, while the other is small (1 in. x 6 in.) and filtered with lead, cadmium, and copper. A large crystal is used for logging general lithology, and a small crystal is used for logging uranium- ore-grade zones. Problems were encountered in boreholes Uelq, Ue7j, and Ue71. The S/Tiall crystal was inadvertently used to log boreholes Uelq and Ue7j; the small crystal was also inadvertently used to log the upper 165 m of borehole Ue71. Count rates from this small crystal were extremely low (statistically unreliable) and calibration factors and spectral stripping coefficients that were applied to this data were not appropriate for these low count rates. The potassium (K), uranium (U), thorium (Th) content cannot be adequately resolved from the low count rates recorded using the small crystal. Therefore, only the K,U,Th contents from boreholes Ue7i, Ue7k, and below 160 m of Ue71 are reasonable and usable. The final spectral gamma log results (corrected for horehole size variation and borehole casing, and separated into K, U, Th content) are shown in Figs. 17 (K, potassium content), 18 (Th, thorium content), and 19 (U, uranium content). Potassium zones of 6% or greater for boreholes Ue7i, Ue7k, and Ue71 were used as comparison zones with the lithrlogic description of these boreholes. The object of this corriparison was to determine if high potassium correlated with clay content. Eighteen zones were compared. Twelve of these zones compared favorably with the lithologic description of non-swelling clay, common K feldspar, and argillized tuff. The other six zones had no lithologic description of clay, K feldspar, or argillized tuff and, consequently, no correlation with potassium content. These units may have been missed in describing and annotating the lithologic logs. It appears from this preliminary study that high potassium content obtained from the spectral gamma log correlates reasonably well with clay and argillized tuff and should be considered as a logging technique for identifying clay zones.

144 145 CO

s 8

8 a o

HldBO

146 147 D. Correlation of the Thermal, Epithermal, and Spectral Gamma Thorium Responses Thermal and epithermal neutron log results are shown in Figs. 20-23. These logs have been corrected for borehole size variation and a 50-point running average has been used to filter the data, as shown in Fig. 21 for the thermal neutron logs and in Fig. 23 for the epithermal neutron logs. Also, a correction was made to the neutron logs acquired in fluid and air-filled portions of each borehole so that the final log results would be equivalent. These calibration data were supplied by BFEC. The corrections worked better for the epithermal neutron logs (Figs. 72 and 23) than they did for the thermal neutron logs (Figs. 20 and 21), where gaps and offsets in data still exist (see Ue7i and Ue71). All data are plotted in counts per second because there was no calibration data to convert counts into water contents.

A thorium anomaly or increase exists at a depth of 140 m for borehole Ue7i, as shown in Fig. 18. At this same depth a decrease in counts per second can be seen for the thermal neutron log (Fig. 21). The water level occurs at a depth of approximately 540 m in the borehole. An explanation for this thorium increase and thermal neutron count decrease is that, from sample analysis, the increase in thorium is associated with an increase in rare earth elements. An increase in rare earth elements is also associated with an increase in thermal neutron absorbers, and these lower the thermal neutron count rate.

At this same depth, however, the epithermal neutron count rate increases (Fig. 23). This result is not understood at the present time. Water content remains constant through this zone. Further study is needed to resolve this discrepancy.

VII. CONCLUSIONS A variety of wireline well logs were obtained in Southern Yucca Flat from five exploration boreholes and one emplacement borehole using two logging subcontractors. The geologic environment that the boreholes traversed is alluvium, Tertiary volcanics, and Paleozoic limestone and dolomite. The quality of all logs is reasonable except for some of the spectral gamma logs (small detector used) and magnetic susceptibility logs (temperature fluctua- tions) acquired by BFEC. Also, the natural gamma logs acquired by BFEC are marginal because of the noisy behavior exhibited. Filtering of these gamma logs would have reduced the noise level and yielded usable data. Birdwell natural gamma logs below the water table exhibit a dead or very quiet response. Major changes in lithology are discernible on all these natural gamma logs.

Caliper logs showed variation in response from different caliper tools and the changing dynamic borehole system when logged on different dates. BFEC resistance and conductivity logs do not correlate with Birdwell resistivity. These loqging tools are not equivalent and appear incommensurable. The magnetic susceptibility log, despite its problems of zero reference and temperature drift, gave better and more detailed signature and trace

14b UE1q UE7k UE7i UE7i UE7I

Fig. 20. THERMAL NELITRON 400 CORRECTED FOR CALFER BENDIX

500-

700

800- o o o o o in o io •- •- -N W CPS UE1q UE7k UE7j UE7i UE7I

100- r i i

200-

30OT Fic. 21. THERMAL NEUTRON CORRECTED FOR CALPER 400-f BENDIX

LOG DATA FI.TERED WITH 50 POINT AVERAGE 5004

i\ ! 600-

f 700- 500 - 100 0 150 0 250 0 200 0

800- 8 g § s in o ID Ol

CPS UE1q U31a UE7k UE7j UE7i UE7I

Fig. 22. EPTTHERMAL NEUTRON CORRECTED FOR CAUPER

HRDWELL - U3la

BENOIX - OTHEPS

r- »- CU

° g 8 S 8 8 8OS8S8S

CPS UE1q UE7k UE7j UE7!

100*

200-

300" L

Fig. 23. EPITHERMAL NEUTRON fO 400-1 CORRECTED FOR CAUPER B0OX

LOG DATA FILTERED WITH SO POINT AVERAGE soot

! I

700- s s

800- o o o o o

*- -- CM C\J U7 O Hi O 1O

CPS response in the tuff units (especially the older tuff units) than the resistivity logs. The magnetic susceptibility can be operated continuously in a borehole that is partially fluid-filled. In other words, borehole fluid or lack of fluid has no effect on the response of the magnetic susceptibility log. This characteristic makes logging operationally easier than with the Birdwell resistivity logs. Magnetic susceptibility logs and density logs, when crossplotted for various tuff units, exhibit a linear or near-linear response. Magnetite content and/or magnetite grain size and orientation account for this correlation. It is recommended for these reasons that the magnetic susceptibility log be routinely acquired in boreholes at the NTS. The potassium channel of the spectral gamma log correlates reasonably well with clay and argi11ized tuff zones. The thorium channel of the spectral gamma log correlates with rare earth element zones. Also, because neutron absorbers are present in rare earth element zones, the thermal neutron log shows a decrease in count rates, as expected. Unexpectedly, the epithermal neutron log shows an increase in count rates for this same interval. We use the epithermal neutron log routinely to estimate water content for containment purposes. This unusual response means further study is needed on the epithermal neutron log. However, because of the potential of the spectral gamma log for locating clay zones with greater reliability than with the resistivity log, it is recommended that it be routinely acquired in NTS boreholes.

ACKNOWLEDGMENTS We thank Ron Oliver for assisting Bendix Field Engineering Corporation during the logging at Nevada Test Site.

153 BIBLIOGRAPHY Broding, R. A., Zimmerman, C. W., Somers, E. V., Wilhelm, E. S., and Stripling, A. A., 1952, Magnetic Well Logging, Geophysics, Vol. 17, pp. 1-26. Carroll, R. D., and Muller, D. C, 1973, Techniques for the Quantitative Determination of Alluvial Physical Properties from Geophysical Logs, Southern Yucca Flat, Nevada Test Site, USGS Open File Report, USGS-474-175, p. 113. Emilia, D. A., Allen, J. W., Chessmore, R. B., and Wilson, R. B., 1981, The DOE/Simplec Magnetic Susceptibility Logging System, Transactions: SPWLA Twenty-Second Annual Logging Symposium, Paper S. Hilchie, D. W., 1968, Caliper Logging—Theory and Practice, The Log Analyst, Vol. 9, January-February 1968, pp. 3-10. Khan, M. A., 1962, The Anisotropy of Magnetic Susceptibility of Some Igneous and Metamorphic Rocks, Journal Geophysical Research, Vol. 67, No. 7, pp.. 2873-2885. Marusak, N. L., 1981, Stratigraphic Interpretation Us ing Electric Logs, Pro- ceedings of the Monterey Containment Symposium, pp. 147-162. Negata, T., 1961, Rock Magnetism, Tokyo, Maruzen, 350 p. Rao, V. B., 1956, Magnetic Properties of Magnetite, Geophysics, Vol. 21, pp. 1100-1110. Scott, J. H., Seeley, R. L., and Barth, J. J., 1981, A Magnetic Susceptibility Wpll-Logging System for Mineral Exploration, Transactions: SPWLA Twenty- Second Annual Logging Symposium, Paper CC. Stratton, J. A., 1941, Electromagnetic Theory, McGraw-Hill, New York.

154 LA-UR r-83-22S0

lei Altmet Klimml Ubor«tO'y » op«r«1*s by iria Univtrtify Of C«"fOf"'»

TITLE ASH-FLOW TUFF DISTRIBUTION AND FAULT PATTERNS AS INDICATORS OF ROTATION OF LATE-TERTIARY REGIONAL EXTENSION, NEVADA TEST SITE

AUTHORtS) HOLLY DOCKERY ANDER

SUBMITTED TO 2ND CONTAINMENT SYMPOSIUM 2-4 AUGUST 1983 KIRTLAND AFB ALBUQUERQUE, NM

By icctpunc* of ihil »1idt ih« pubiithir r*cognit*i mti tht U S Gov«rnm«ni rtirm • nonaxclutiv*. roytUyfrM lietnM to puMni 0' raproouc* fitpublnhtd 'cm of thit comribulion. or lo allow oifici 10 do to lor US Ocwnmtm Purpo«»»

Tut Lot Alimos tanonti USo'iiory rtQuam thit lh» pgbh»h»t io»nnr, tnn anici* •) wo'k p»nOrm»a urvfl«' mt tutpicn o» lh» U.t Oapirtmtnt of E n»rpv

Los Alamos National Laboratory Los A!amos,New Mexico 87545

UIIU MO MM S'll 155 ASH-FLOW TUFF DISTRIBUTION AND FAULT PATTERNS AS INDICATORS OF ROTATION OF LATE-TERTIARY REGIONAL EXTENSION, NEVADA TEST SITE

Holly A. Dockery Earth and Space Sciences Division Los Alamos National Laboratory

ABSTRACT Isopach and structure contour maps generated for Yucca Flat as well as fault pattern analyses of the Nevada Test Site (NTS) can aid in more efficient site selection and site characterization necesary for containment. Furthermore, these geologic studies indicate that most of the alluvial deposition in Yucca Flat was controlled by north- trending faults responding to a regional extension direction oriented approximately 20° to 30° west of the N50°W direction observed today.

The Yucca Flat basin-forming Carpetbag and Yucca fault systems seem to be deflected at their southern ends into the northeast- trending Cane Spring and Mine Mountain fault systems. Left-lateral strike-slip displacement of ^1A km found on these northeasterly faults requires that most of the displacement on the combined fault systems occurred in an extension field oriented approximately N80°W. Fault movement in this extensional field postdates the Ammonia Tanks tuff (^11 m.y.) and was strongly active during deposition of some 110U meters of alluvium in Yucca Flat. Time of rotation of regional extension to the presently active N50°W direction is unknown; however, it occurred so recently that it has not greatly modified fault displacement patterns extant at the NTS

156 INTRODUCTION

The Nevada Test Site (NTS), Yucca Flat in particular, is characterized by a multitude of geologic data. Continuous coring, sidewall samples, and cuttings from drillholes provide a degree of three-dimensional control on distribution of rock units in Yucca Flat basin unmatched anywhere in the southern Cordillera. However, due to programmatic constraints, a detailed synthesis and analysis of all data has not been accomplished. The aim of this study has been to generate maps for both Los Alamos and Livermore use areas in Yucca Flat showing: 1) thicknesses of major Cenozoic rock units (isopach maps) and 2) the absolute elevation of a given horizon, such as the base of a rock unit (structure contour r.iaps). Other aspects of the study have included analysis of slip lineations formed on fault planes by fault movement (slickensides) and fault orientations.

The products of primary interest to Containment are the isopach and structure contour maps. These maps were constructed utilizing lithologic contacts rhosen from samples taken in Yucca Flat drillholes. After plotting either the thickness of a chosen major rock unit (such as Alluvium, Rainier Mesa Tuff, Paint- brush Tuff, Grouse Canyon Tuff, etc.) or elevation of the base of a unit with respect to sea level at each respective drillhole location, the data was contoured at an interval reasonable for each unit.

The distribution of ash flow tuffs which predominantly fill the basin was heavily controlled by the topography existing at the time of deposition, preferentially filling depressions and covering higher elevations with a thimer veneer. Orientations of large fault troughs or erosional valleys as well as horst blocks and paleogeographic highs can then be deduced from the thickness pattern of major units. Units of anomalous thickness can easily be located by comparing actual well data with that projected by the contour maps. The isopach maps can then be used to aid in recognition of major subsurface faults and their orientations as well as indicating general trends of thickness changes due to deposition or erosion. This will allow more accurate estimates of unit thickness at any given location.

Structure contour maps can be used in much the same way as isopach maps in determining subsurface fault trends by the linear disruption of contour lines. In addition they are helpful for constructing the areal dip of a unit and the elevation at which a unit should be encountered.

157 KM

FIGURE 1

Physiographic provinces of the western United States.

158 Thus these maps are helpful in site selection, site characterization, research and development of unused areas by indicating areas of geologic complexity as well as areas which are relatively uncomplicated.

REGIONAL SETTING

The NTS is located in the Basin and Range Province (Figure 1) which is characterized by normal faulting, thin crust, and high heat flow (Lachenbruch and Sass, 1977; Thompson and Burke 1974, Roy et al., 1968). Before 35 myBP, the Basin and Range was undergoing compression related to plate convergence resulting in thrust faulting, folding, and calc-alkalic volcanism (Eaton, 1979).

Extension forming the Basin and Range started around 35myBP (Cross and Pilgi r, 1978; Eaton, 1979). This extension has been attributed to either back-arc extension re iltiny from heating of North American asthenosphere forming diapirs of magma which rise and spread laterally (Scholz, et al., 1971; Snyder, et al., 1976) or to changes in absolute plate motions during Farallon-North American plate convergence (Cross and Pilger, 1978). The change from subduction to a transform boundary along the Pacific coast, which occurred around 25 myBP (Atwater, 1970; Pilger and Henyey, 1977), is thought by some workers to have caused much of the later deformation related to extension (Atwater, 1970; Christiansen and Lipman, 1972; Synder et al., 1976) as well as a change in volcanism to fundamentally basaltic (Christiansen and Lipman, 1972). However, problems with temporal and spatial correlation of transform faulting with extensional features have led some workers to believe other forces, such as continuing mantle diapirism may also be effecting the Basin and range (McKee, 1971; Noble 1972).

Paralleling the California-Nevada border between Reno, Nevada arid Las Vegas, Nevada, lies a zone where the north-northeast linear grain of the Basin and Range is disrupted. This physiographic sub-province, called the Walker Lane- Las Vegas Shear Zone (Figure 2), has been postulated to be an area of right-lateral shear (Stewart, et al., 1968). Certainly right-lateral shear has been well document- ated at the southern terminus of this zone in the Las Vegas valley (Fleck, 1970; Anderson, et al., 1972) and at the northern end near Lake Pyramid (Bell and Slemmons, 1979). Although the physiographic zone passes through western NTS, evidence for an extremely large amount of right-lateral slip in the region near NTS has not been reported.

159 FIGURE 2

Major geologic structures in southern California and southern Nevada.

160 The intent of work at NTS has been to attempt documentation of changes in stress directions and timing during the Cenozoic in order to better understand the mechanisms causing crustal extension in the Basin and Range. The synthesis of this work into a regional interpretation will be presented in a future publication.

FAULTING AT NTS

Three prominent fault trends are observed at NTS, northeast, north-south, and northwest. Short trace length faults of various orientations are also present.

The north-south faults are the present basin formers. The axis of Yucca Flat trends north-south. It is bounded to the easl by the Jangle Ridge-Paiute Ridge fault system and to the west by the Carpetbag and related faults (Figure 3). Most of the present day activity in the basin occurs along the Yucca Fault (Figure 3) which runs down the center of the valley. The age of these north-south faults is presently thought to be between 17.8 and 7 million years (Ekren, et al., 1968).

Major northeast trending faults include the Mine Mountain, the Cane Spring, and the Rock Valley (Figure 3), all of which exhibit large components of left-lateral slip. The eastern portions of the Cane Springs and the Mine Mountain faults appear to swing to the north into the north-south fault systems forming Yucca Flat. There is no evidence that these faults continue across or were offset by the north-south faults. For this reason it is conjectured that the northeast and north-south faults are all one continuous system. The Rock Valley Fault does not appear to change its trend. Ekren, et a1. (1968) believe the first northeasterly faults formed after 26.5 myBP.

Northwesterly trending faults are observed in the ranges. They have also been defined in the basin through borehole geology and gravity methods. A large northwest-trending gravity high found bisecting the basin (Carr, 1974; John Ferguson, S.M.U., personal communication, 1981) i.iay be related to this fault set. No Quaternary faults of this orientation are found. The age of initiation of these faults is thought to be the same as the northeast-trending faults (Ekren, et al., 1968), but they do not appear to be active in the present stress regime.

161 1. YUCCA FAULT

2. JANGLE RIDGE/PAIUTE RIDGE FAULT SYSTEM

3. MINE MOUNTAIN FAULT SYSTEM

4. CANE SPRING FAULT SYSTEM

5. ROCK VALLEY FAULT

6. CARPETBAG FAULT

CALICO HILLS ^ CENTER mhwm

QUATERNARY LAS VEGAS VALLEY" ALLUVIUM SHEAR ZONE FIGURE 3 I TERTIARY I VOLCANJCS MAP SHOWING TIMBER MOUNTAIN CALDERA AND PATTERN OF BASIN-RANGE FAULTS IN NEVADA TEST SITE AND VICINITY PRE-TERTIARY ROCKS Faults of short trace length are found throughout the area. Their orientations appear to be random. Such faults are contained in blocks bordered by major faults and probably are the result of a locally altered stress field controlled by the large block bounding faults. The small faults are thus only indirectly related to the regional stress f ield.-

The extension direction presently active at NTS and in the surrounding area is N50°W. This is derived from northeast-trending explosion related fractures, northeast trending cracks at Yucca Lake and, northwest-southeast enlargement of drillholes at NTS (Carr, 1974), from earthquake first motion studies (Al Rogers, U.S.G.S., personal communication 1981), and direct strain measurements (Obert, 1964). In the present stress field, the north-south faults should be undergoing right- oblique slip, the northeast faults almost purely dip-slip, and the northwest faults almost purely strike-slip (unless they are parallel to the extension direction, in which case there would be no movement on the northwest faults at all). However geologic evidence shows a ratio of 5:1 left-lateral strike slip to dip-slip movement on the Mine Mountain Fault (Orkild, 1968) occurring sometime after deposition of the Ammonia Tanks tuff (^11 myBP). Therefore it is apparent that in the past the stress field has differed from that observed at present. From knowledge of rock behavior in brittle deformation (Jaeger and Cook, 1969, pp 74-102), and also from the presence and orientation of "5" shaped faults or Riedel shears (Wilcox, et al., 1973) between major northeast-trending faults (Figure 4) the orientation of the extension direction at the time of formation of the northeast-trending faults was approximately N80°W.

SLICKENSIOE ANALYSIS

Slickensides are lineations gouged on fault surfaces at the time of movement. In a rough sense they are the line of intersection of the fault plane and the extension direction. Faults, cnce formed, can move in response to a variety of stress orientations. The result is that fault traces are relatively insensitive to stress; change. However slickensides are good indicators of fault motion and the single stress field in which it operated.

163 ai Z o 51 5 1 N -UJ \ \ 4 £ \i 5 1 u< (0 I D ta. 1 >u- < cc a. tu CJ D f C J UJ

#*• /^^

FIGURE 4

Schematic diagram of the proposed intersection of northeast-trending and north- south-trending faults and related Riedel shears.

164 Over 200 slip measurements were collected on major fault zones at NTS. The data were then analyzed via an iterative least squares fit program which determines the stress tensor active when slip occurred (Angelier, 1979 and M. L. Zoback, U.S.G.S., written communication, 1982). Approximately 70% of the data showed fault movement in a stress field with the least principal stress oriented N78°W.

THREE-DIMENSIONAL ANALYSIS OF YUCCA FLAT

Generalized isopach maps derived from detailed master maps are shown in Figures 5-9. The intent of this sequence is to show the development of the basin from early Cenozoic to present. Figure 5 shows the distribution of all tuffs underlying the Grouse Canyon (>13.8 myBP in age). These include the Tunnel Beds, Tub Springs, Yucca Flat, Fraction, Red Rock Valley, and other tuff units. Although a northwest trending depression existed, there is no evidence that any of the presently active faults (present day surface trace of faults shown on figure for reference) controlled deposition of these older units. Figure 6 of the Grouse Canyon Tuff shows possible activity on a fault with a similar trend to the Yucca Fault. This fault may or may not be related to the Yucca Fault. However, sparse data to the south and to the northeast does not allow good control on the fault's orientation. In any event, the maximum throw before or during deposition of the Grouse Canyon Tuff was only 30 meters, which is mininml compared to displacements averaging 1000 meters on major Basin and Range fault sets. The isopach of the Paintbrush Tuff (Figure 7) includes the Paintbrush Tuff, the Area 20 Tuffs, and the Wahmonie-Salyer sequence. The interpretation is more complex due to a northwest-trending isopach anomaly, which may be related to the northwest-trending gravity high observed by Carr (1974) and Ferguson (1982). But the major north- south-trending faults still do not appear to be active during this time. At 11 myBP the Rainer Mesa Tuff (Figure 8) also shows no depositional control exerted by north-south faults. The Alluvium isopach (Figure 9) shows a dramatic change in the character of the basin. Sometime after 11 myBP the Yucca and Carpetbag Faults became major basin-forming faults. If, as discussed previously, the Cane Springs Fault curves continuously into the Jar.gJe Ridge - Paiute Ridge fault set and the Mine Mountain Fault and its plays are continuous with the Yucca and Carpetbag faults, the following scenario can be constructed. In a N78°W extension field, north-south-trending faults would be favorably oriented to have a major component of dip-slip motion with a sm?ill element of right-lateral strike-slip. Carr (1974) has suggested that the north-south faults in the Yucca Flat area have right-lateral

165 ISOPACH OF TUFFS OLDER THAN THE GROUSE CANYON TUFF

CONTOUR INTERVAL 100 Meters

Scale 0 2000 meters

4-N810.000 N810,000-h E670.000 E700.000

FIGURE 5

Isopach of tuffs older than the Grouse Canyon Tuff in Yucca Flat

166 ISOPACH OF THE GROUSE CANYON TUFF

CONTOUR INTERVAL 10 Meters

Scale P 2000 meters

N820.000 -+- N820.000 -f- E670,000 FIGURE 6 E700,000

Isopach of Grouse Canyon Tuff in Yucca Flat

167 ISOPACH OF PAINTBRUSH TUFF

CONTOUR INTERVAL 50 Meters

__Scale 0 2000 meters

N815,000-f- £670,000 + N815,000 E700,000

FIGURE 7

Isopach of Paintbrush Tuff in Yucca Flat

168 ISOPACH OF RAINIER MESA TUFF N 1

CONTOUR INTERVAL 50 Meters

Scale 0 2000 meters

N807.000 + + N807,000 E670.000 E700,000

FIGURE 8

Isopach of Rainier Mesa Tuff in Yucca Fiat

169 ISOPACH OF ALLUVIUM

CONTOUR INTERVAL 200 Meters

Scale 0 2000 meters

N810,000-|- -f-N810,000 E667,000 E700.000

FIGURE 9

Isopach of Alluvium in Yucca Flat

170 displacement based on left-stepping en-echelon faults and surface cracks, indications of offset of Paleozoic structures, slickensides, and shape of the medial Yucca flat depression. On the inside of the proposed bends in these fault systems would exist a major space problem where the west side of the north-south fault is moving north and the north side of the northeast fault is moving west. An inordinate amount of subsidence must take place at the curve to accommodate the relative displacements. The evidence for this is seen on the southwestern end of Figure 9 where the alluvium reaches a maximum thickness of 1161 meters in well Ue6d. A similar thickness might be anticipated under Yucca Lake. The timing of fault movement influenced by the N78°W least principal stress may then be placed as syn-depositional with nnuch of alluvium. Change in extension direction from N78°W to N50°W then must have occurred in the Quaternary.

CONCLUSIONS Analysis of map units, fault patterns, and slickenside data show the presence of a minimum principle stress oriented N78°W operative some time before the present day N50°W extension. Offset of an 11 my old tuff along a NE trending left- lateral fault is evidence that the N78°W extension was active sometime after deposition of that unit".. The configuration of alluvial troughs in Yucca Flat suggests that the faulting in response to this stress field happened concurrent with deposition of much of the alluvium. Although the north-east faults are almost perpendicular to the N50°W least principal stress and therefore are very favorably oriented to displace in a dip-slip sense, very little dip-slip motion has occurred ^approximately 400 meters, Orkild 1968). Hence it may be postulated that the predominate Basin and Range faulting at NTS occurred in an older N78°W extension field. Rotation of stresses to a N50°W extension has occurred very recently and has evinced very littie response from existing faults.

171 REFERENCES

Anderson, R.E., C.R. Longwell, R.L. Armstrong, and R. F. Marvin, 1972, Signifi- cance of K-Ar ages of Tertiary rocks from the Lake Mead region, Nevada- Arizona, Geol. Soc. Amer. Bull., v. 83, p. 273-288.

Angeliw, Jacques, 1979, Determination of the mean principal directions of stresses for a given fault population, Tectonophysics, v. 56, p. T17-T26.

Atwater, Tanya, 1970, Implications of plate tectonics for the Cenozoic tectonic evolution of western North America, Geol. Soc. Amer. Bull., v. 81, p. 3513- 3536.

Bell, E.J and D.B. Sitstrimons, 1979, Recent crustal movements in the central Sierra Nevada-Walker Lane region of California-Nevada Part II, the Pyramid Lake right slip fault zone segment of the Walker Lane, Tectonophysics, v. 52, p. 571-583.

Carr, W.J., 1974, Summary of tectonic and structural evidence for stress orientation at the Nevada Test Site, United States Geological Survey Open File Report 74-176, 53 p.

Christiansen, R.L. and P.W. Lipman, 1972, Cenozoic volcanism and plate-tectonic evolution of the western United States, II, Late Cenozoic, Royal Soc. London Philos. Trans., v. 271, p. 249-284.

Cross, T.A. and R.H. Pilger, Jr., 1978, Constraints on absolute motion and plate interaction infer "3d from Cenozoic igneous activity in the western United States, Am. Jour. Sci., v. 278, p 865-920.

Eaton, G.P., 1979, Regional geophysics, Cenozoic tectonics, and geologic resources of the "asin and Range Province and adjoining regions, RMAG-UGA-1979 Basin and R^ge Symposium, p. ll-T^.

172 Ekren, E.B., C.L. Rogers, R.E. Anderson, and P.P. Orkild, 1968, Aqe of basin and range normal faults in Nevada Test Site and Nellis Air Force Range, Nevada, Geol. Soc. Amer. Memoir 110, p. 247-250.

Fleck, R.J., 1970, Age and possible origin of the Las Vegas Shear Zone, Clark and Nye Countie, Nevada, Geol. Soc. Amer. Abs. with Prog., v. 2, p. 333.

Ferguson, J., 1972, Gravity analysis of Yucca Flat, southern Nevada Methodist Univ., Ph.D. dissertation, p. 110.

Jaeger, J.C. and N.G.W. Cook, Fundamentals of Rock Mechanics, Chapman and Hall, Ltd., New York, New York,1969, 515 p.

Lachenbruch, A.H. and J.H. Sass, 1977, Heat flow in the United States and the thermal regime of the crust, Heacock. J.E., ed., The Earth's Crust, Am. Geophys. Union Monograph 20, p. 626-675.

McKee, E.H., 1971, Tertiary igneous chonology of the Great Basin of the western United States - Implications for tectonic models, Geol. Soc. American Bull., v, 82, p. 3497-3502.

Noble> D.C., 1972, Some observations on the Cenozoic volcano-tectonic evolution of the Great Basin, western United States, Earth Planetary Sci, Letters, v. 17, p. 142-150.

Obert. L., 1964, Operations Nougat and Storax, in situ stresses in rock, Rainier Mesa, Nevada Test Site, U.S. Bureau of Mines Report WT-1869, 95 p.

Orkild, P.P., 1968, Geologic map of the Mine Mountain Quadrangle, Nye County Nevada, U.S.G.S GQ Series Map GQ-746.

Pilger, R.H. Jr. and T.L. Henyey, 1977, Plate tectonic reconstructions and the Neogene tectonic and volcanic evolution of the California Borderland and Coast Ranges, Geol. Soc. America Abs, with Programs, v. 9, p. 482.

173 Roy. R.F., E.R. Decker, D.D. Blackwell, and Francis Birch, 1968, Heat flow in the United States, Jour. Geophys. Research; v. 73, p. 5207-5221

Scholz, C.H., M. Barazangi, and M.L. Soar, 1971, Late Cenozoic evolution of the Great Basin, western United States, as an ensialic interarc basin, Geol. Soc. Amer. Bull., v. 82, p. 2979-2990.

Thompson, G.A., and D.B. Burke, 1974, Regional geophysics of the Basin and Range Province, Annual Review of Earth and Planetary Sciences, v. 2, p. 213- 238.

Snyder, W.S., W.R. Dickenson, and M.L. Silbermari, 1976, Tectonic implications of space-time patterns of Cenozoic magTiatisrn in the western United States, Earth and Planetary Sci. Letters, v. 32, p. 91-106.

Stewart, J.H., J.P. Albers, and F.G- Poole, 1968, Summary of regional evidence for right-lateral for displacement in the western Great Basin, Geol. Soc. Amer. Bull., v. 79, p. 1407-1414.

Wilcox, R.E., J.P. Harding, and D.R. Seeley, 1973, Basic wrench tectonics, Amer. Assoc. Petroleum Geol. Bull., v. 57, p 74-96.

174 SUBSURFACE SCARPS ON THE PALEOZOIC SURFACE AT THE NEVADA TEST SITE—FAULTS OR TOPOGRAPHY? McKague, H. L., Lawrence Livermore National Laboratory, Liveimoie, CA 94550 Various geophysical surveys used to locate the Paleozoic-Ccnozoic surface at the NTS often define sudden linear changes in elevation. These are termed scarps. In the containment evaluation of drill holes, these scarps are always assurr.ed to be fault related. Such subsurface features may reflect: (1) the vertical component of movement along faults, \2) ptcexisting topography, or (3) several combinations of (1) and (2). Criteria are developed, which allow one to speculate on the origin of subsurface sf'nps. by using oiientation, length, differences in elevation across the scarp and the slope of the scarp surface. As siles are located nearer to the edges of Yucca Flat, and in new areas in the more cential part of the valley, correct identification of scarps will n-suit in sofer sites and a more efficient utilization of NTS real estate. LA-UR-83-2266

Lot *Hmoi Nition*' op«'it>0 tf tin Unw»'»'ty Dt CtMpmn tor m» unifg 5m»i P«pinm>nt of eo«'gy u"tft' cantr»c

TITLE MAPPING DF PALEOZOIC STRUCTURE IN NORTH-CENTRAL AREA 7 OF THE NEVADA TEST SITE - A GEOPHYSICAL CASE HISTORY

AUTHOfi(S) Frederick N. App, Wayne R. Meadows and Allen H. Cogbill

SUBMITTED T-O 2nd Containment Symposium Z-h August 1983 Kirtland AFH Albuquerque, NM

%i »cc«pttnct o' (An iiidt in* publn^c fK9gnii«i me th« \j S Gpvcntrwni rcttmi t nonaielutivr rpyiHy-frM hetntt lo puDii»i pr rip'oduc* tt>*puDH«h»d lorm o< thii coniiitolion or to (Hot* om«rt lo do to. fo' US Ooviirnmanl purpoits TM Lot Alamoi Ntiiomi Ltborttorv 'Muntt ihil Ih* pytiiihc Wcniify mil t^icit •> wo'k p«rform*d under m« (utplea* pi lh« U 5 0*p>nm*m o'En*'O>

Los Alamos National Laboratory Los Alamos,New Mexico 87545

»STi'J HO UtU •T MO MMS'll 177 MAPPING OF PALEOZOIC STRUCTURE IN NORTH-CENTRAL AREA 7 OF THE NEVADA TEST SITE - A GEOPHYSICAL CASE HISTORY

F. IM. App, W. R. Meadows, and A. H. Cogbill

ABSTRACT

Results of surface-to-borehole (half-refraction) seismic surveys are presented for an area within the Department of Energy's Nevada Test Site (NTS), where a previously conducted reflection survey of potentially higher resolution had been only marginally successful. The goal of such surveys is to map the shape and configuration of tne Paleozoic (Pz) surface, which ranges in depth from 400 m to over 700 rn in the approximately 3 km by 1.8 km area studied. Knowledge of Pz structure is required for proper siting of device emplacement holes for the underground nuclear test program.

In the half-refraction technique, geophone packages are placed in exploratory boreholes at or near the upper surface of the high velocity refractor to be mapped, in this case the Pz surface. Surface sources distributed along lines radiating out from the well-head are used to generate seismic energy. We have used computer, two-dimensional ray tracing techniques to interpret structure from the seismic wave arrival times recorded from the individual surface sources. The ray trace modeling allows evaluation and comparison of extremely complicated subsurface configurations. Its chief limitation is that it is restricted to two dimensions.

17& Surveys in the area are incomplete, but from data processed and analyzed so far, it is clear that tiie method is effective for mapping the Pz. A large, previously unmapped Pz ridge has been identified in the central part of the study area. Major structure has been mapped near uie working point of a previous underground test. Through careful use of available Pz control, it has been possible to decrease to acceptable levels intarpretational errors due to our incomplete knowledge Ci Pz velocities. As a reb It of our half-refraction work, the studied area can be more properly and safely utilized for future underground testing.

INTRODUCTION

The half-refraction technique for seismic exploration gradually has been gaining acceptance as an effective method for mapping Paleozoic (Pz) surface at the Nevada Site (NTS) since its initial trial in the fall of 197(3 in Area 10 . In the spring of 1981, the technique was successfully applied in mapping a portion of Area 4 in the vicinity of hole Ue4c . Refraction techniques are being applied primarily because reflection methods have been only marginally successFul- at the NTS. Reflection surveys appear better suited for mapping the shallow ash-flow tuffs than the Pz, although some recent reflection work by LLNL does show some promise for mapping the Pz .

In the half-refraction technique, geophone packages are placod in exploratory holes at or near the upper surface of the high velocity unit that is to be mapped. Surface sources distributed along lines radiating out from the well-head are used to generate seismic energy. Figure 1 is a cartoon depicting a single source-receiver configuration. We use computer, two-dimensional (2-0) ray tracing techniques to interpret structure from seismic wave arrival times recorded from the individual surface sources. The ray, trace modeling allows evaluation and comparison of extremely complicated subsurface configurations. Its chief limitation

179

Fig. 1. The hdlf-refraction technique-the high velocity basement rock provides a minimum path for seismic waves generated at the surface and recorded at a geophone empleced near the contact of the low and high velocity materials. is that it is restricted to two dimensions. Figure 2a shows the ray paths calculated for a complex structure with varying velocities. Figure 2b is a travel time plot far the calculated ray paths.

SURVEY DESCRIPTION

Figire 3 shows the locations of half-refraction profile lines recorded :.n north-central Area 7. Thus far, surveys have been conducted out of holes Ue7h and Ue7i. We hope to conduct a survey out of LJe7j in the near future. Table 1 gives some facts pertinent to the L)e7h and Ue7i survey lines. Our intent is to 'map the Pz in a roughly 3 km by 1.8 km area. Pz tags in the area indicate that the Pz ranges in depth from about 400 m to over 700 m. Our goals are to 1) map the area in sufficient detail such that real estate usage for underground testing is optimized, 2) to re-evaluate the structural settings of previously conducted events in the area with an eye toward updating the previous experience base, and 3] to learn ipore about the strengths and limitations of the method prior to applying it tn other areas.

Drill hole Ue7h was drilled to a total depth of B38 m in December, 1981. It was later deepened to 780 m. The 601 m depth to the Pz rock was unexpectedly large and suggested the presence of significant faulting in the area. It was because of the suspected complexities, and tne fact that an already-drilled emplacement hole (U7bn) was only 25 m away, that Ue7h was chosen for the first of the proposed surveys. The survey was conducted during the winter and early spring of 1982. The survey was started with a single compone.it (vertical) phonr, clamp ad into the hole at 605 m depth, 4 m below the tuff/Pz contact. After completion of lines "3W", "NE", "WWW", we decided to obtain a 3-component (one vertical, two horizontal) phone for the remainder of the survey. We reo^cupied selected stations on the first three lines using the 3- "component phone and used it for all of lines "N", "S", "NW", and "ESE". The. on-site Birdwell g&<_.phone survey data acquisition equipment and personnel were used for the survey. The seismic source was the single

181 Ue7i Ue7h U7bh (projected} L i NE IE FROM UE7

1200 0 250 500 750 1000 1250 1500 1750 2000 DISTANCE ALONG PROF ILE

Fig. 2a. Calculated ray paths for line Ue7i-1E.

LINE IE FROV UE7 .800

Je7h tie .700 L 7 fa

.600 -

oo 00°

.500 IP 5P °° c ooocg0"**^00 n •

D calculated .400 O rassurad

,30C 1 1 i i 250 500 750 1000 1250 1500 1750 2000 DISTANCE (M)

Fig. 2b. Calculated and measured travel times along line Ue7i-1E.

182 Reversed line Ue7h coverage Ue9cn Ue7i coverage\

Ue7l

Ue7be /

•Ue7bi

Fig. 3. Locations of half-refraction survey lines. TABLE I. Description of Lines

Distance to Approximate Farthest Bearing From Station Recording Line Station (mj Weil-Head Interval (m) Dates (s)

Ue7h-NE 604 N 60 E 16.75 12/81, 2/82 Ue7h-SW 1492 S 60 W 16.76 12/31, 2/82 Ue7h-WNW 1913 N 79 W 16.76 12/81, 2/82 Ue7h-N 661 i\i 6 E 15.76 3/82 Ue7h-S 765 s 4 W 16.76 3/82 Ue7h-NW 831 N 59 W 16.76 3/82

i—i Ue7h-SE 1160* J 74 E 16.76 2/82, 3/82

Ue7i-1E 1946 S 82 E 16.76 4/33 Ue7i-lW 1373 N 77 W 16.76 4/83 Ue7i-2E 1715 N 7£ C. 19.05 4/83 Ue7i-3E 1851 S 37 E 19.05 4/33 L)e7i-3W 1143 N 41 W 19.05 4/83 Ue7i-4E 1372 N 46 E 19.05 4/83 Ue7i-4W 952 S 47 W 19.05

*Near station offset 299 rn SE of Ue7h wellhead.

184 Eirdwell Y600 vibrator. Data processing, which consisted primarily of providing AGC and True Amplitude record ssctions, was provided by Birdwell's Houston Processing Center.

Drill hole Ue7i was drilled in February, 1981 and tagged the Pz at 721 m. In Aoril, 1983, it was deepened to 884 m. After deepening Ue7i, a survey similar to the one at Ue7h was conducted with a three component phone clamped in the hole at 747 rn. The same Birdwell crew acquired the data; however the data were processed at the Seismograph Service Corporation (parent company of Birdwell) processing center in Tulsa.

Figure 4 is a typical record section of processed data. It is from lines Oe7i-1E and 0e7i-lW and consists of individual seismograrns arranged in correct spatial relationship to one another; i. e., in the order that they were recorded along the line.

VELOCITY CONSIDERATIONS

Because we are using a seismic technique, any interpretation is sensitive to the velocity model employed. Overburden velocities, defined here as the velocities of all materials overlying the Pz, are well known in the study area because of the numerous holes with geophone surveys in the area. Figure 5 is a comparison of overburden velocities, presented as time-depth curves, for holes in the irrmediate vicinity of the half- refraction survey lines. To our main depths of interest (400 m and below), the seismic travel times do not vary by more than 10 ms between holns. This means that interpretational errors associated with uncertainties in overburden velocity are relatively small, even under the assumption of a single overburden velocity function at all locations. In practice, we allow horizontal velocity variations to fit the locally known velocities to further reduce this error. Variations in travel times due to tne major structures in the area range from tens to hundreds of milliseconds. Our major concern lies with the Pz velocities. There are much fewer velocity data available for Pz rocks than for tuff and alluvium, and what data exist suggest considerable variations from one

185 L)e7h Line Ue7i-lw Ue7i Line Ue7i-1E

LJ •«•

Fig. 4. Processed record section for lines Ue7i-1E and Ue7i-lW. Z8T

TRAVEL TIME IN MSEC K) to 04 o Cn O cn o o o O O O o O o

1 < CD CD a o ri- al

aCD "D a CD cn c

«c2n

CD zr

CD area to another. So far, we have determined that uncertainties associated with poorly known Pz velocities can be made acceptably small if lines cross known Pz control points. The line having the best control is the one connecting holes Ue7h and Ue7i. This profile has reversed coverage. Line Ue7h-WNW extends west from L)e7h through Ue7i; line Ue7i- 1E extends to the east -From Ue7i through L)e7h. An interval velocity of 4064 m/s was determined from a geophone survey over the uppermost 33 m of Pz at Ue7h. Similarily, the uppermost 172 m of Pz at Ue7i yielded an interval velocity of 4916 m/s. The higher interval velocity at Ue7i is not necessarily the result of deeper sampling away from the interface or to improved accuracy provided by the larger interval. Independent "3-D" velocity logs over the same intervals also indicate higher velocities near the Pz surface at Ue7i than at Ue7h.

Whichever velocity function is used in the interpretation, it must allow agreenent between calculated and measured arrival times for the ray that crosses the tuff/Pz contact at a known Pz control point along the profile. This is a valuable constraint that greatly narrows the choice of velocity functions available to the interpreter.

Because line Ue7h-WNW and Ue7i-1E cover the feame basement between holes Ue7h and Ue7i, the structural interpretation between the two holes must fit the data from both of those lines. In Figure G the first arrival times are plotted for the two lines. The open circles represent the measured data; the open squares represent the calculated times of first arrivals if the Pz surface between, the two holes is a simple, flat homocline. Obviously, it is not a homocline. A structural interpretation along the survey line between WG7h and Ue7i is shown in Figure 7a. Figure 7b and 7c are the fits of calculated times to the measured values for lines Ue7i-1E and Ue7h-WNW respectively. The model fits the data very well. In this model, the Pz velocity immediately below the tuff/Pz contact was assumed to be about 4000 m/r» everywhere along the profile. This value is consistent with'that measured at Ue7h, but not at Ue7i, A vertical velocity gradient of 5.7 /s was assumed to exist at both LJe7h and Ue7i, until the velocity reached 6096 rn/s, below which depth a constant 6096 m/s velocity was assumed. The vertical

188 —*= DISTANCE ALONG PROFILE WEST FROM HOLE UE7H (MJ

1750 750 50 25 1500 1250 1000 ° ° ° .800 —i

.700 «£.

(SE C .600 UJ o o 1 o°

.500 • a 0°° RAVE L 1

.400 D calculated o measured

.300 _L _L _L i 0 250 500 750 1000 1250 1500 1750 2000

DISTANCE ALONG PROFILE EAST FROM HOLE UE7I (M

Fig. 6. Measured first arrivals along lines Ue7i-1E and Ue7h-WNW conpared to calculated first arrivals for a simple, flat homocline. Ue7h

1200 0 250 500 750 1000 1250 1500 DISTANCE ALONG PROFILE F:g. 7a. Best fit structural model for assumed vertical velocity gradient in Pz. LINf It FROM UE7i _ .800

700 Ue7h tie

a calculated

o nBasurad

300 2bU 500 750 1000 1250 1500 1750 2000 DISTANCE (M) Fig. ~5. Calculated and measured travel ti™s along line Ue7i-1E. LINE WNW FROM UC7H .800

Ue7i tie .700

£5 600 v>

.500

.400 a calculated • nawurad .300 250 500 750 1000 1250 1500 1750 2000 DISTANCE (M) Fig. 7c. Calculated and measured travel tim»s along line Ue7h-WNW.

190 gradient within the Pz ridge is a function of the height of the ridge, having a lower value of about 3.1 /s beneath its peak. The underlying assumption here is that Pz velocity at some depth reaches a constant value, regardless of the overlying Pz terrane, but that velocities near the Pz surface are uniformly relatively low and are independent of overlying overburden thickness. The contour plot of Figure 8 shows the velocity distribution used in the model. Note in Figures 7b and 7c that the assumed velocity model results in good ties of calculated arrival times to data at the Pz control points. A stronger vertical gradient beneath the ridge would result in a mis-tie. We could use a stronger gradient if we were to assume even lower Pz surface velocities,, but the available data do not permit such velocities. Figure 9a is the structure that results if we assume a Pz velocity that is everywhere constant at 4500 m/s (no velocity gradients]. Figures 9b and 9c are the travel-time fits. For computational convenience, the structural model of Figure 9a is described by a series of straight lines rather than a curved line such as is used for Figure 7a. The ridgs structure that results from the constant velocity assumption is about 50 m lower in elevation than the one resulting from the vertical gradient assumption. In both interpretations it is a sizable hill in the Pz. In order to effect ties to the Pz tags, we are constrained, under the constant velocity assumption, to use a velocity at or near 4500 m/s. One can argue that the two interpretations just presented are limiting cases, not in the purely mathematical sense, but in the practical sense of what we believe the range of possibilities in velocities to be. • Dne can postulate other perverse situations such as negative velocity gradients. These may occur locally, but we do not believe they exist to such an extent that they would have a significant impact on the interpretation.

Figure 10a is an interpretation using all measured velocities as additional constraints. For this case, we use the measured velocity of 4950 m/s from the Pz surface down to 884 m at Us7i. At Ue7h we continue to use the measured velocity of about 40QQ m/s near the Pz surface. It is assumed that the vertical gradient at Ue7h is such that velocities of 4950 m/s are attained at a depth of 884 m. Below 88'4 m, the velocities are assumed to increase to a maximum value of 6096 m/s at 120G m.

191 Ue7i Ue7h 0

200 -

LoJ 400 =• o 600 V _J mLJ

CL 800 -

1000 -

1200 250 500 750 1000 1250 1500 DISTANCE ALONG PROFILE W\

Fig. B. Velocity structure used for ray trace modeling between holes Ue7h and Ue7i - vertical velocity gradient. Ue/i Ue7h

200 •

UJ u < 400 • a. / ^^ in S 600 D I a 800 • jj

1000 •

1 r?nn 250 500 750 1000 1250 1500 DlSlANCE ALONG PROMlC Fig. 9a. Best fit structural model for assur.ied constant velocity in Pz.

LINE 1 EAST FROM HOLE Vl~> I .600

Ue7n tie .700

.600

500

D calculated .400 M e inusurad

.300 '0 250 500 750 1000 1250 15Or 1750 2000 DISTANCE ALONG PROruE (u) Fig. 9a. Calculated and measured travel times along line Ue7i-1E.

LIMC WNW TROW HOLC UC7H a» 800

700

.500

250 500 750 1000 1250 1500 1750 2000 DISTANCE ALONG PROFILf (M)

Fig. 9c. Calculated and measured travel times along line Ue?h-WNW.

193 Ue7h

12000 600 7b0 1000 12LC 1500 DISTANCE ALONr PROr IL T Fig. 103. Best fit structural model using all available velocity measurements. L t*t[ I f AST rROM HOLi Ut7' ^' .800

Ue7h tie .700

600

500

400 I«I O calculated o measured

300 o 2SO SOO 7bO 1000 1250 1 1750 2000 DISTANCE ALONG PROr IL I (M) Fig. 10b. Calculated an.1 measured travel times along line Ue7i-1E.

LINE WM* fROM HOLE UE7H ^ .800

Ue7i tiiee — •• ^f*

.600

.500

.400 a calculated

O measured

.300 0 250 500 750 1000 1250 1500 1750 2000 DISTANCE ALONG PROFILE (M)

Fig. lDc. Calculated and measured travel times along line Ue7h-WNW.

194 Figures 10b and 10c are the fits to the data. We now have a fit to all data available to us at this time. The interpreted structure is almost identical to the one obtained under the constant velocity assumption.

The profile reversal gives us added confidence that strong horizontal variations in velocity are not present. For example, if there were large horizontal changes through the Pz hill, and we were not accounting for them in the modeling, the hill structure interpreted from line Ue7h-WNW would have a shape different than that interpreted from line Ue7i-1E. We get good, but not perfect, matches with the data from both lines using the same model, leading us to conclude that there are no significant horizontal velocity changes present.

STRUCTURAL INTERPRETATION

Using knowledge gained on velocities from the reversed lines between Ue7h and Ue7i, we have guidelines for modeling along lines that do not have Pz tags or other velocity control. Figure 2a, shown earlier as an example of ray tracing, represents the best fit to all of the data from line Ue7i-1E. Prior discussion of this line dealt only with the region between Ue7h and Ue7i, while in fact the subsurface coverage extends some distance east of Ue7h. The velocity model of Figure 8 (vertical gradient in Pz) was used in the interpretation. Note the large Pz scarp just east of Ue7h. Figures' 11 through 14 are structural interpretations along four of the remaining six profile lines radiating out from hole Ue7h. The lines were interpreted using the same velocity model that was used for Figure 2. There are Pz tags along lines Ue7h-NW (Figure 13) and Ue7h-SW (Figure 14). Line Ue7h-NE (Figure 11) and Ue7h-N (Figure 12) have no Pz control. Ths tie at Ue7ax along line Ue7h-NW is very good, indicating that our velocity function is reasonable in that area. However, there is a 15 ms mis-tie at Ue7j along line Ue7h-SW, suggesting that some modifications to the velocity functions are required along that line. Bear in mind that all interpretations assume two- dimensionality. We are fairly confident that line. Ue7i-1E is roughly

195 200

«00

600 IE 1

so:

100C

'20C 25C SOC 750 DISTANCE ALONG PROFILE

pig. lla. Best fit structural model along line Ue7n-NE.

LINE UE7H - NE .500

.450 o UJ to .400

° o 6 .350

D calculated

e measured .300 250 500 750 01 STANCE (M)

Fig. lib. Calculated and measured travel times along line Ue?h-NE.

196 ?t>0 bOO DISTANCl ALONG PROf =ig. 12a. .Best fit structural model along line Ue7h-M. LINE UE7H - N . JUU

.450

O in .400

.350 - D calculated « measured

•*nn 250 500 750 DISTANCE (M)

Fig. 12b. Calculated and measured travel times along line Ue7h-N.

197 Ue7ax U7at

200

400

600

i n

BOO

1000

1200 0 2S0 500 750 1000 DISTANCE ALONG PROFILE ig. 13a. Best fit structural modsl along line Ue7tn-NW.

LINE UE7H - NW

.450 Ue7ax tie—i

oo °o o o o UJ in o • 8 .400 a o

.350 o • o ooo B O calculated

• 250 500 750 1000 DISTANCE (M) Fig. 13b. Calculated and measured travel times along line Ue7h-rjW.

198 Ue7h LJe7j LINE UE7H - SW

1200, 250 500 750 1000 1250 1500 DISTANCE ALONG PROFILE

Fig. 14a. Best fit structural model along line Ue7h-SW. LINE UE7H - SW .700 Ue7j mis-tie .650

.600

.550

.500

.450

.400 B calculated e raasurBd .350

.300 250 500 750 1000 1250 1500 DISTANCE ALONG PROFILE (M)

Fig. 14b. Calculated and measured travel times along line Ue7h-SW.

199 perpendicular to strike. Line Ue7h-N is probably sub-parallel to strike and thus is subject to "sideswipe"; i.e., the first arrivals result not from rays traveling directly beneath the profile line but rather from the side, from the scarps and ridges paralleling the profile line. Currently we have no way for making accurate interpretations from such data. As a general statement, it is important when we go into a new area that we establish the direction of strike of the major features. Foreknowledge of structural trends (strikes) will vary from area to area. A very evident characteristic of the area studied is that there is major relief on the Pz surface. The large ridge between Ue7h and Ue7i was unknown to us prior to these surveys. Also, we had no way of knowing the magnitude of throw on the scarp to the east of Ue7h, or its location. In our current Interpretation, the scarp lies directly beneath, and very near, a previously conducted underground test (ALIGGTE in U7bg). Figure 15 is an isopach of the Pz surface showing the Pz structural trends in the area as they are now interpreted. Currently the only line interpreted from Ue7i is the one discussed earlier, line Ue7i-1E. The isopach map will probably be modified as additional interpretations are made from the Ue7i lines, and-perhaps further modified when data become available from Ue7j.

SUMMARY

At this point, the half-refraction results are most encouraging. We are successfully mapping a geologically complex area. With proper attention to Pz control in the area, we are able to minimize structural uncertainties due to incomplete knowledge of Pz velocities to an acceptable level. We have found a major structure very near the working point of a previous underground test. When all surveys in the area are complete, we feel confident that we will be able to make very efficient use of this part of Area 7 for nuclear testing. From what we have seen so far, the area can acconrmodate both deep and shallow tests, if they are carefully sited.

200 Depths to Paleozoic surface, metres

424

.652

S3 o 783

•652

Fig. 15. Isopach of the surface from hole tags and half-refract ion surveys. As we conduct more half-refraction surveys, we probably will find additional structural settings that are quite different from previously mapped. Based on what we have seen so far, we believe we shall find that some of the older underground tests were conducted in unusual, and by today's standards alarming, structural environments. From such knowledge, we may be able to form better-founded opinions on the containment threat posed by such structures.

202 REFERENCES 1. Exploration Services Division, Geosource, Inc., 1978, Seismic Wave- Experiment - Area No. 10/Nevada Test Site: Final Report under Subcontract No. 848-CUC-78.

2. App, Frederick N., 1981, Progress in Seismic Exploration at Los Alamos: Proceedings of the Monterey Containment Symposium, Los Alamos Report LA-9211-C, v. 1., p. 343-353.

3. Burkhard, Norman L., 1982, Private Connmunication.

203 LCRL- 89411 PREPRINT

Determination of Subsurface Geological Structure with Borehole Gravimetry

S. R. Clark J. R Hearst

This paper was prepared for submittal to the Second Containment Symposium, Albuquerque, New Mexico, August 2-4, 1983.

July, 1983

This is a preprint of a paper intended for publication in a journal or proceedings. Since changes tuny be made before publication, this preprint is made available with the understanding that it will not be ciled or reproduced without the permission of the author.

205 DETERMINATION OF SUBSURFACE GEOLOGICAL STRUCTURE WITH BOREHOLE GRAVIMETRY* CLARK, S. R., EG&G, 2801 Old Crow Canyon Road, San Ramon, CA 94583 and HEARST, J. R, Lawrence Livermore National Laboratory, P. 0. Box 808, Livermore, CA 94550 Abstract: Conventional gamma-gamma and gravimetric density measurements are routinely gathered for most holes used for underground nuclear tests. The logs serve to determine the subsurface structural geology near the borehole. The gamma-gamma density log measures density of the rock within about 15 cm of the borehole wall. The difference in gravity measured at two depths in a borehole can be interpreted in terms of the density of an infinite, homogeneous, horizontal bed between those depths. When the gravimetric density matches the gamma-gamma density over a given interval it is assumed that the bed actually exists, and that rocks far from the hole must be the same as those encountered adjacent to the borehole. Conversely, when the gra/imetric density differs from the gamma-gamma density it is apparent that the gravimeter is being influenced by a rock mass of different density than that at the hole wall. This mismatch can be a powerful tool to deduce the local structural geology. The geology deduced from gravity measurements in emplacement hole, U4al, and the associated exploratory hole, UE4al, is an excellent example of the power of the method.

Underground nuclear tests conducted at the Nevada Test Site require a complete understanding of the subsurface geology in order to assure containment of radioactive gases. A large diameter emplacement hole for the nuclear device is drilled and studied extensively to assure that no unacceptable features that could compromise containment are present. For example, it is possible that high-density, high-acoustic-velocity rock near the event could reflect the shock wave towards the surface, or that a fault could provide a gas travel path to the surface. Conventional gamma-gamma and gravimetric density measurements are routinely gathered for most emplacement holes. The logs serve to determine the subsurface structural geology near the borehole. The gamma-gamma density log measures density of the rock within about 15 cm of the borehole wall. The difference in gravity measured at two depths in a borehole can be interpreted in terms of the density of an infinite, homogeneous, horizontal bed between those depths. When the gravimetric density matches the gamma-gamma density over a given interval it is assumed that the bed actually exists, and that rocks far from the hole must be the same as those encountered adjacent to the borehole. Conversely, when the gravimetric density differs from the gamma-gamma density it is apparent that the gravimeter is being influenced by a rock mass of different density than that at the hole wall. This mismatch can be a powerful tool to deduce the local structural geology. The geology deduced from gravity measurements in emplacement hole, U4al, and the associated exploratory hole, UE4al, is an excellent example of the power of the method.

* Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract number W-7405-ENG-48.

206 The mismatch between measured gravimetric density and gamma-gamma density in U4al (Fig. la) increasing with depth, is evidently the result of a high density mass in the vicinity of the borehole. The local geology would suggest this to be a buried topographic high in the high-density carbonate Paleozoic basement complex. This feature can conservatively be considered the footwall of a scarp (Fig. 2). Therefore, it was of paramount importance to adequately define its position so that the possible gas travel path that could be caused by a projection of the fault to the surface could be evaluated. Applying a correction to the measured gravimetric density to compensate for the influence of the correct structure should result in a good match between the two curves. A poor match would imply that the structure is still not well understood and should be reconsidered. Before a structure correction can be made, the density contrast between the various rock layers must be determined. The overburden density is available from the gamma-gamma log, but the basement density must be 1) obtained from Paleozoic cores from the area, 2) assigned a reasonable value or 3) measured in a nearby exploratory hole drilled to obtain the density. In this case, surface gravity studies had indicated that the density contrast was 0.70 g/ cc. This contrast produced an adequate match between the two curves when used with a structure inferred from the surface gravity. A different density contrast would, however, have provided an equally good match when used with a different structure. Since it was imperative to accurately define the density of the Paleozoic basement which would allow interpretation of the structure, the exploratory hole UE4al was drilled. The gamma-gamma density log in the Paleozoic section of that hole was not usable due to gaps between the sonde and the borehole wall. The density was, therefore, determined from core samples. The depth to the Paleozoic surface in UE4al was 500 meters. This was approximately 130 meters greater than that predicted by surface gravimetry. The Paleozoic density of 2.69 produced a density contrast of 0.82 g/cc. When this density contrast, along with-the Paleozoic tag at 500 meters shown in Figure 2 was used to correct the gravimetric density, the agreement was good (Fig. 1b). The electric log from the exploratory hole, however, indicated that a fault intersected the hole in the basement complex. This required the inclusion of another fault scarp to the west of the hole (Fig. 3). This cross section provided a correction equally as good as the first cross section (Fig. lc). In fact, it was subjectively felt that the revised cross section provided a better fit at the bottom of the hole than the original. Some overcorrection of the gravimetric density between 275 meters and 350 meters tended to make the root mean square error of the fit supplied by the model virtually the same as that of the previous model. When a borehole gravimeter was run in the exploratory hole, the disagreement between the gravimetric density and the gamma-gamma density was quite large (Fig. 4a). The disagreement is of the sense such that above the Paleozoic rocks (^500 meters) the gravimetric density is greater than trie gamma-gamma density. Below that depth, the opposite is true. Gaps in the gamma-gamma density curve occur when data are absent because of gaps between the tool and the wall of the borehole. When corrections derived from the same cross section used for the emplacement hole (Fig. 3) were applied, the agreement was

207 quite good (Fig. 4b). Below 500 meters the correction is positive, which is consistent with the position of the scarp to the east. Above 500 meters the correction is negative which also verifies the location of the smaller scarp to the west. If the second scarp was not included (Fig. 2) the agreement was far worse (Fig. 4c), tending to confirm the existence of the second scarp. Note that below 500 meters the correction for the low density mass to the east is retained. Above 500 meters the correction is lost when the small scarp is removed from the cross section. This case history is a good illustration of the ability to use the comparison of gamma-gamma and gravimetric densities to choose among possible subsurface structures. Knowledge of the density contrast allowed the verification of the structural model. The small fault scarp adjacent to hole UE4al was predicted and verified by borehole gravimetry. It is often possible to use a method like this to test the validity of a structure determined from geologic or geophysical data.

DISCLAIMER - This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the University of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government thereof, and shall not be used for advertising or product endorsement purposes.

208 DENSITY...GM/CC DENSITY...GM/CC DENSITY...GM/CC M -• -MM "MM W Zl O » - 0 O 1 o 1 m 1 TO O 50 ( V \ 100 i

ISO o200 ( m •o i -I 1 A : 250 I m n [ W300

350 ] V L 1t r 400 ( 450 \ /

500 FIGURE 2 U4AL -100, WEST EAST —100 o 100 aoo soo «oo soo OVERBUPDEM •00

7OO PALEOZOIC PASfUtNT •00

•00

1000

1100 ? f ? ° S 8 § DISTANCE FROM HOLE LUAL (METRES)

FIGURE 3 U4AL WEST

? B S DISTANCE FROM HOLE U4AL (METRES)

210 FIGURE 4A 3.O

2.5 L-

1 /—-fr 1—11 "

FIGURE 4B 3.O

2.5 r i 2.O

•

1 .5

FIGURE 4C 3.O

2.5

2.O

1 .O

DEPTH (METRES) LA-UR-S3-2229

Los Alamos National Laboratory is operated by the University of California for the United States Department of Energy under contract W-7405-ENG-36.

TITLE: GEOCHEMICAL SIMILARITIES BETWEEN VOLCANIC UNITS AT YUCCA MOUNTAIN AND PAHUTE MESA: EVIDENCE FOR A COMMON MAGMATIC ORIGIN FOR VOLCANIC SEQUENCES THAT FLANK THE TIMBER MOUNTAIN CALDERA

AUTHOR(S): R. G. Warren, ESS-2

SUBMITTED TO: 2nd Containment Symposium 2-4 August 1983 Kirtland AFB Albuquerque, MM

The Los Alamos National Laboratory requests that the publisher identify this article as work performed under the auspices of the US Department ol Energy

Los Alamos National Laboratory Los Alamos, New Mexico 87545

FORM NO »36 R4 ST NO 2628 5/81 213 GEOCHEMICAL SIMILARITIES BETWEEN VOLCANIC UNITS AT YUCCA MOUNTAIN AND RAHUTE MESA: EVIDENCE FOR A COMMON MAGMATIC ORIGIN FOR VOLCANIC SEQUENCES THAT FLANK THE TIMBER MOUNTAIN CALDERA

R. G. Warren Los Alamos National Laboratory Los Alamos, NM 87545

214 ABSTRACT Chemical compositions have been determined for sanidine, plagioclase, biotite, and hornblende phenocrysts by electron microprobe for a comprehensive set of samples of Crater Flat Tuff and tuffs of Calico Hills. Most of these samples were obtained from drill holes at Yucca Mountain. Samples of tuffs and lavas of Area 20, obtained from locations at Pahute Mesa, have similarly been subjected to microprobe analysis. Complete modal petrography has been determined for all samples. Biotite and hornblende in the samples from both Yucca Mountain and Pahute Mesa have Fe-rich compositions that cortrast strikingly with Fe-poor compositions in the overlying Paintbrush Tuff and the underlying Lithic Ridge Tuff at Yucca Mountain. Each unit from Yucca Mountain has distinctive compositions for both sanidine and plagioclase that very closely match compositions for a corresponding unit identified within the lower, middle and upper portions of the Area 20 tuffs and lavas from Pahute Mesa. Each of these paired units probably originated from a common parental magma and was erupted contempor- aneously or nearly so. Each pair of units with matching phenocryst chemistries has a similar, but not identical set of petrographic characteristics. The petrographic differences, as well as small differences in phenocryst chemistry, result from a zonal distribution of phenocrysts within the parent magma chamber and eruption through earlier units that differ markedly between Yucca Mountain and Pahute Mesa.

INTRODUCTION During the past two years, intensive research into the geochemistry of the volcanic rocks of Pahute Mesa has been pursued at Los Alamos. The purpose of this research has been to establish the usefulness of major, minor, and trace element geochemistry (determined mostly by neutron activation analysis, or NAA) and mineral chemistry (determined by electron microprobe) in establish- ing stratigraphic correlations among units in Pahute Mesa drill holes. This geochemical approach has been highly successful and allows the definition of units based on chemistry and petrography rather than lithology; units so defined are termed "Petrologic Units." A single Petrologic Unit may include several lithologic units, as seen in Table 1. For example, the Pyroxene- bearing Rhyolite of the Scrugham Peak Quadrangle (Petrologic Unit symbol TRPP)

215 Table 1. Nomenclature and symbols for Petrologic Units discussed in this report. {Units within each area listed in stratigraphic succession, youngest unit at the top. Complete stratigraphic sequence not given.)

Petrologic Lithologic Unit Unit Volcanic Volcanic Symbol Symbol Subunit Unit Group

TMRu Tmr quartz latite Rainier Mesa Member Timber Mountain Tuff TMR1 Tmr rhyolite

TPCu Tpc quartz latite Tiva Canyon Member Paintbrush Tuff

TRPP Trpp, Tp(Tb) * Pyroxene-bearing rhyolite Tpl of Scrugham Peak Quadrangle

'TRA(ul) Trau, Trab mafic-poor upper i

TRA(u2) Tral,b Trab mafic-rich upper

c r>ahute TRA(p) Tra1, Trab plagioclase-rich Tuffs Mac a rlcba ' and and TRA(m) Trab middle lavas vicinity of TRA(lr) Trat lithic-rich Area 20

TRA(sw) Tsw Stockade Wash Tuff

TCT Tp(Tb) Tram Member Crater Flat Tuff

r THl Tht, Thl mafic-poor tuffs and lavas of Calico Hills TH2 Tht, Thl mafic-rich

TCP Tcp Prow Pass Member

Yucca TCB Tcb Bullfrog Member . Crater Flat Tuff fountain and TCT Tct Tram Member vicinity TLR Tlr Lithic Ridge Tuff

TTA Tta Unit A, USM-Gl

TTB Ttb Unit B, USW-Gl

TTC Ttc Unit C, USW-Gl h Present common usage for NTS volcanic units. Upper part of lower lavas (Figure 5 in Byers et al., 1976a). c Lower part of lower lavas (Figure 5 in Byers et al., 1976a).

216 includes the Pyroxene-bearing Lava of the Scrugham Peak Quadrangle (Lithologic Unit symbol Trpp), bedded tuff that underlies the Trpp unit, and the local ash- flow tuff of the Paintbrush Tuff (Tpl Lithologic Unit). All three of these lithologic units have identical mineral compositions and very similar (although not identical) phenocryst contents that contrast strikingly with those of bordering units. Symbols for Petrologic Units (Table 1) are utilized extensively throughout this paper. These symbols are fully capitalized and therefore distinct from symbols for lithologic units, which have only the first letter capitalized (compare symbols in Table 1). The Petrologic Unit has been particularly successful in the recognition and identification of correlatable horizons within bedded tuffs and lavas beneath Pahute Mesa. Such lithologies generally show poor correlations between drill holes (if correlation is attempted at all), due to intertonguing of lavas with bedded tuffs and to misidentifications which often result due to a lack of distinctive hand-sample characteristics among bedded tuffs and lavas. However, early results indicate that Petrologic Units do not show significant thickness variations where such lithologies interfinger. For example, the TRA(ul) Petrologic Unit in U19aS consists of 229 m of lava and 21 m of bedded tuff but in U19q, only 444 m distant, it consists of 34 m of lava and 213 m of bedded tuff (Warren, in prep.). Although the thicknesses of the individual lithologic components differ strikingly, the TRA(ul) unit is 250 m thick in U19aS and 247 m thick in U19q and the correlation of this Petrologic Unit between these adjacent drill holes is excellent. It is likely that lavas simply represent near-source eruption of magma, and tuffs tend to "fill in" the paleotopography for some distance away from this source. The purpose of this report is to demonstrate that the Petrologic Unit provides a powerful means to correlate volcanic units by relating units that occur in widely separated areas of the NTS. This correlation could not be achieved using lithologic units, due to striking dissimilarities between lithologies within the volcanic sequences correlated between the two areas. These volcanic sequences are the tuffs and lavas of Area 20 of the Pahute Mesa area and the tuffs and lavas of Calico Hills and the Crater Flat Tuff of Yucca Mountain and vicinity (Byers et a!., 1976a; Carr et al., in press; see Figure 1 for locations). In order to establish the correlation of volcanic sequences between these areas, it is necessary to subdivide the tuffs and lavas of Area 20 into members, and to demonstrate the petrographic and chemical equivalence

217 ,.' KlvwchRinge'.

Ue19p \ • Ue19i U19ab \U»20f «# Ue19x,

117°W

mm n\ ? N?;."^>v-.-;:-^3S "

; ••••••.•'U'/"s-.•••• T^:--. •".'.•:• •/^^^^S^lJ^l»opw»iii7^§•••••/•-. r*^^ ;

Fig. 1 Location of sample sites. Outcron locations are open symbols, drill hole locations are closed symbols. Volcanic units of NTS are present throughout region except in stippled areas. Exact locations are given for samples south of Timber Mountain in Warren et al. (in prep.), and for"samples of the TRA(ul) unit in Table 2. Locations for samples of other units north of Timber Mountain are Tpb: RW19f; TRA(u2): Ue20f, RW19f; TRA(p): Uel9p, Uel9x; TRA(p)?: Ue20f; TRA(m): Uel9p, U19ab, Uel9fS; TRA(lr): RW19r, Uel9p, U19ab, Uel9i; TRA(sw): RW19r, RW18a; TCT: RW18a. Mineral chemical data are available for all samples from each of these locations (USW-G2 and Ue25b-lh from Broxton et al., 1982), and modal petrographic data are available for all locations (except USW-G2 and Ue25b-lh.)

218" of tuffs and lavas that comprise a single Petrologic Unit. The following section is devoted to these topics.

DESCRIPTION OF PETROLOGIC UNITS OF THE TUFFS AND LAVAS OF AREA 20 The tuffs and lavas of Area 20 (Byers et al., 1976a) are here divided into four major Petrologic Units, termed the upper, plagioclase-rich, middle, and lower members (Table I). The upper member is further subdivided into an upper, mafic-poor portion [symbol TRA(ul)] and a lower, mafic-rich portion [symbol TRA(u2)]. The lower member includes highly lithic-rich tuff [symbol TRA(lr)] and the Stockade Wash Tuff [TRA(sw)]. An additional Petrologic Unit related to the tuffs and lavas of Area 20 is present beneath the Stockade Wash Tuff at Rattlesnake Ridge (see Fig. 1 for location), southeast of Pahute Mesa. On the basis of a single sample, this unit (symbol TCT) is correlated with the Tram Member of the Crater Flat Tuff (Tct). Each Petrologic Unit is defined on the basis of well-defined range in certain petrographic and mineral chemical parameters; the most useful and reliable of these is the sanidine composition. Other silicic- to intermediate- composition lavas also occur within the TRA sequence; preliminary study indicates that they are associated with the uppermost portions of the TRA(p) and/or TRA(lr) units. These lavas include the olivine latite of Kawich Valley (Sargent et al., 1966), andesites of U19e and U19g, and a lava misidentified as the pre-Pah lava in U19aj. These lavas are not treated in this report. Below, summary tables of petrographic and mineral compositional data are given for individual samples of the TRA(ul) unit to illustrate their ranges within a Petrologic Unit (Tables 2-5), but only combined data for all samples of other Petrologic Units are presented. Tables 6-11 include data for each Petrologic Unit of the tuffs and lavas of Area 20, as well as its correlative unit from Yucca Mountain. In each of these tables, data for Petrologic Units appear in groups; the Petrologic Unit at the bottom of each group is found at Yucca Mountain and all those above it are correlative units found at Pahute Mesa and vicinity. Petrologic Units increase in age towards the bottom of each table. For comparison, Table 11 al^j contains data for units both younger and older than the tuffs and lavas of Area 20.

219 TRA(ul) Petrologic Unit ) The upper, mafic-poor member of the tuffs and lavas of Area 20 consists of ash-flow tuff, bedded tuff, and lava. The TRA(ul) unit is present in most drill holes in Pahute Mesa (where penetrated), and attains a thickness exceeding 1000 m in Ue20h, Petrographic results are given for twelve samples of this unit in Table 2, including (in sequence) three samples of lava, eight samples of bedded tuff, and a single sample of ash-flow tuff. For comparison, results for a single sample of the tuff of Blacktop Buttes (Tpb unit) is also presented at the bottom of Table 2. This unit is positioned strati graphically within the Paintbrush Tuff by Byers et al. (1976a), but has petrographic and chemical characteristics identical to those of the TRA(ul) unit, and strikingly different from those of other Paintbrush Tuffs. Field work is in progess by Byers and Warren to investigate the possibility that the presently accepted strati graphic position of the Tpb unit is incorrect. All samples of the TRA(ul) unit show a consistently low phenocryst content, ranging from 1.4 to 4.5%, and consistently high proportion of quartz phenocrysts. They also have a strikingly low content of mafic minerals, iron- titanium oxides, and accessory minerals. Biotite is the only mafic mineral present in most samples. Ilmenite, sphene, and perrierite (all abundant in most units of the Paintbrush Tuff and many other volcanic units of the NTS) were not found in any of these samples, but allanite is present in many samples. Except for the presence of appreciable hornblende, pyroxene, and apatite in sample RW19r-7, there are no significant differences among the primary mineral contents of the samples of TRA(ul) unit (Table 2). Sanidine compositions, represented by their orthoclase (Or, KAlSi'Og) + celsian (Cn, BaAl2Si20g) molecular percents (mol%) and BaO contents (Table 3) are also very similar for all samples of the TRA(ul) unit. Compared to other units of the NTS, sanidine of the TRA(ul) unit is potassium-rich and barium- poor. There is a slight, but consistent, decrease in the Or+Cn content of sanidine stratigraphically upward at each drill core location, from Or+Cn = 69 in the strati graphically lowest sample at each location to a value as potassium-poor as Or+Cn = 65 in Ue20e-l. The sanidine compositions near the base of the TRA(ul) unit match those of the underlying TRA(u2) unit, and those near the top match those at the base of the overlying Topopah Spring Member of the Paintbrush Tuff (Warren et al., in prep.; Caporuscio et al., 1982).

220 Table 2. Summary of petrography for individual samples of TRA(ul) unit and Tpb unit. (ppmV = parts per million by volume. See below for explanation of symbols.)

RESULTS OF POINT COUNT DETERMINED BY SCAN; all values in ppmV Major conponents. » Thin titho- section Felsic logic Sample Rock Alter- area? Points Pheno- Relative I felsics Mafic Phenocrysts Fe-Ti oxides Accessory minerals Sample no. unit type tyi ation (mm ) counted Voids Pumice Lithics crysts Q K Biot hbH Cpx Opx ' 61 Ac Arf Ht/Hir Urn Sph All Per Ap Zr Other

RH20g-3 Trau 0 1 Gl.Sp 724 595 3.4 0.0 1.9 18* 34* 48" 86 0 0 0.6 0 0 0 330 0 0 10 0 9 14 Ue20e-l-1715 Trau c 1 Gl 617 457 22 0.0 1.5 44* 39* 17* 1100 0 0 0 0 0 0 360 0 0 0 0 6 16 -3525 Trau c 1 Sp.Zc 533 528 4.0 0.0 4.5 46 42 12 1400 0 0 0 0 2 0 440 0 0 23 0 0 26 a

RH19r-7 Trab 0 b Gl,mZc 348 435 17 3.4 4.1 42* 42* 16* 44U 37 210 29 0 0 0 660 0 0 0 0 33 3

RU2Og-5 Trab 0 b Gl 724 595 12 12 3.9 1.4 63* 23" 14* 280 0 0 0 0 0 0 100 0 0 2 0 0.6 0.3 Uel9p-16OO Trab c b Gl 753 559 5.2 28 11 2.3 52* 37* 11* 82 0 0 0 0 0 0 120 0 0 5 0 2 3 -1652 Trab c b Gl 723 535 11 40 7.7 2.1 46* 44* 10* 170 0 0 0 0 0 0 120 0 0 7 0 2 5 Uel9x-2301 Trab c b Zc 640 475 2.5 0.4 2.5 51* 43* 6* 77 0 0 0 0 0 0 93 0 0 0 0 0 3 b -2474 Trab c b Zc 626 515 2.3 14 2.2 33* 47* 20* 360 0 0 0 0 0 0 140 12 0 0 0 0 5 U19ab-165O-6O Trab da b Gl 153 378 2.6 22 13 2.1 39* 33* 28* 02 0 0 0 0 0 0 32 0 0 0 0 0 0 -1740-50 Trab da b Zc.Gl 234 580 2.1 23 6.4 3.1 39* 26* 35* 110 0 0 0 0 0 0 180 0 0 8 0 2 5

Ue20e-l-3155 Trab c nw Zc 504 499 2.8 1.2 2.6 28* 29* 43* 140 0 0 0 0 0 0 84 0 0 27 0 0 0.2

RU19f-12 Tpb 0 m Gl 617 508 8.1 25 3.0 2.7 35* 28* 37* 410 71 0 7 0 0 0 280 0 0 92 0 9 7 ? Monazite 2 ppmV Pseudobrookite Explanation of symbols:

Lithologic unit: Trau = upper rhyolite lava of Area 20; Trab * bedded and ash-flow tuffs of Area 20; Tpb = Tuff of Blacktop Buttes

Sample type: c = drill core; da - cuttings that are representative of Hthology; o = outcrop (Nevada State coordinates, in meters: RW20g-3 285195N, 175750E; RU20g-5 286161N, 176010E; RW19r-7 286067N, 189982E; RU19M2 268818N, 179649E). Numbers for core and cuttings Indicate sample depth, in feet. Rock type: 1 = lava, b - bedded tuff, nwt - non-welded ash-flow tuff

Alteration: Gl * glassy (vitric), Sp * spheruHtic, Zc * zeoUttc (clinoptiloUte). Modifier V indicates "minor."

Minerals: Q = quartz; K - sanidine; P • plagioclase; Biot = biotite; Hbld = hornblende; Cpx = clinopyroxene; Opx = orthopyroxene; 01 * olivine; Ac = acmite; Arf = arfvedsonite; Ht = titanomagnetite; Hm = hematite; llm = ilmentte; Sph = sphene; All • allanite; Per » perrierite; Ap = apatite; Zr = zircon.

Asterisks: indicate that relative proportions of felsic phenocrysts were determined from largest phenocrysts only. These phenocrysts comprise 70-1001 of the felsic phenocrysts in these samples. Table 3. Frequency distribution of Or+Cn and BaO contents for sanidfne ptienocrysts in individual samples of TRA(ul) and single sample of Tpb unit (RW19f-12). (Each value represents the number of analyses that have orthoclase+celsian (Or+Cn) end-member contents, and barium oxide weight contents, within the interval indicated. Symbols for rock type defined in Table 2.)

Rock Or+Cn, MolS BaO, wt% Sample no. type 72 70 68 66 64 62 60 58 0.0 0.3 0,6 0.9 1.2 1 1 1 1 1 I RW2Og-3 1 7 1 3 2 12 -5 2 10 1 12

RW19r-7 4 5 4 1 1 11

Uel9p-1600 b 7 7 13 -1652 b 14 1 13

Uel9x-2301 b 5 9 12 -2474 b 5 9 11

U19ab-165.-60 b 1 13 12 -1740-50 b 13 1 12

Ue20e-l-1715 , 2 12 13 -3155 nwt 9 5 12 1 -3525 1 7 2 3 1 1 3 10

RW19N12 nwt 3 13 , 1 , 8 1 1 1 1

Plagioclase compositions, represented by their anorthite (An, ^gg molecular percents (mol%) (Table 4) also show a well-defined ran^e for all samples of the TRA(ul) unit. Compared to other units of the NTS, plagioclase of the TRA(ul) unit is moderately calcium-rich (particularly for a mafic-poor unit) and has a relatively narrow compositional range. There is a slight, but consistent, decrease in the An content of plagioclase stratigraphically upward at each location that parallels the change in sanidine compositions (in both cases the chemical change can alternatively be regarded as an increase in sodium content stratigraphically upward). Similarly, plagioclase compositions near the base of -the TRA(ul) unit match those of the underlying TRA(u2) unit, and those near the top match those within the rhyolitic portion of the overlying Topopah Spring Tuff (Warren et al., in prep.; Broxton et al., 1982). Biotite compositions, represented by their molecular Mg/(Mg+Fe) contents ("Mg number," Mg*), show a well-defined range for all samples of the TRA(ul) unit (Table 5). Biotite compositions are very similar among all Petrologic

222 Table 4. Frequency distribution of An contents for plagioclase phenocrysts in individual samples of TRA(ul) and single sample of Tpb unit (RW19f-12). (Each value represents the number of analyses that have anorthite (An) end-member contents within the interval indicated. Symbols for rock type defined in Table 2.)

Rock An, Mol% Sample no. type 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 • 40 45 50 55 60 1 1 1 1 1 1 1 1 1 1 1 i— l RW2Og-3 1 1 4 7 1 -5 b 1 1 2 9 1

RW19r-7 1 2 2 7 1 1

Uel9p-1600 b 3 10 -1652 b 1 5 7 1

Uel9x-2301 b 1 7 4 -2474 b 1 1 5 2 2 1 1 1

U19ab-1650-60 b 1 6 3 3 1 -1740-50 b 1 4 8 1

Ue20e-l-1715 1 1 8 4 1 -3155 nwt 1 4 6 2 1 -3525 1 4 8 2

RW19f-12 nwt 5 3| 3| 1 I ! III! |

Units of the tuffs and lavas of Area 20, and are strikingly magnesium-poor (MgM).4) compared to most NTS units. Petrographic results for the TRA(ul) unit are summarized as median values in Table 6. Mineral compositional data discussed above are combined for all samples of the TRA{ul) unit in Table 7 for sanidine, in Table 8 for plagioclase, in Table 9 for biotite, and in Table 10 for hornblende. Dominant values for the compositional parameters represented in Tables 7-10 are given in Table 11; the domfnant value is equivalent to the statistical mode.

TRA(u2) Petrologic Unit The upper, mafic-rich member of the tuffs and lavas of Area 20 consists of ash-flow tuff, bedded tuff, and lava. The TRA(u2) unit is about 230 m thick in Ue20f but is absent in Uel9p. Samples of the TRA(u2) unit contain the same primary minerals as those of the TRA(ul) unit, except that the

223 Table 5. Frequency distribution of molecular Mg/(Mg+Fe) contents for biotite phenocrysts in individual samples of TRA(ul) and single sample of Tpb unit (RW19f-12). (Each value represents the number of analyses that have molecular Mg/(Mg+Fe) end-member contents within the interval indicated. Symbols for rock type defined in Table 2.)

Rock Mg/(Mg+Fe) Sample no. type .30 .32 .34 .36 .38 .40 .42 .44 .46 .48 .50 .52 .54 .56 .58 .60 .62 I r RW2Og-3 1 2 1 -5 b

RW19r-7

Uel9p-16OO 2 4 -1652 2 I

Uel9x-2301 1 4 -2474 4

U19ab-1650-60 b 1 3 -1740-50 b 1 4

Ue20e-l-1715 1 -3155 nwt 1 1 -3525 1

RW19f-12 nwt

concentrations of these minerals are much higher (Table 6). The complete absence of all other mafic minerals despite a high concentration of biotite (up to 3.6% in one sample) is particularly distinctive'. Sanidine (Table 7) and plagioclase (Table 8) compositions of the TRA(u2) unit are very similar to those at the base of the TRA(ul) unit, as previously noted. Biotite of the TRA(u2) unit has a slightly, but significantly higher Mg content than biotite of the TRA(ul) unit. Compared to other NTS units (Table 11), sanidine is highly potassium-rich and moderately barium-rich, and plagioclase is calcium-rich.

TRA(p) Petrologic Unit The plagioclase-rich member of the tuffs and lavas of- Area 20 consists of ash-flow tuff, bedded tuff, and lava. The stratigraphic position of the TRA(p) unit is well established in Uel9p between the TRA(ul) and TRA(m) units,

224 Table 6. Median values for volume contents of lithic fragments and phenocrysts for Petrologic Units of Pahute Mesa (PM), Yucca Mountain (YM), and Rattlesnake Ridge (RR). (ppmV = parts per million by volume. Symbols for minerals are defined in Table 2. Samples of TRA(sw) unit are from both Ammonia Tanks quadrangle and Quartet Dome quadrangle locations discussed in text.)

Major components, •

Number Felsic Relative I felsics Mafic phenocrysts (ppmV) (ppmV) Accessory Minerals (ppmV) Petrologic of pheno- Unit Location Samples Lithics crysts Q K P Biot Hbld Cpx Opx 01 AC Arf Mt/Hm llm Sph All Per Ap Zr

TRA(ul) PM 12 5.2 2.3 44 39 17 155 0 0 0 0 0 0 130 0 0 4 0 1 4 Tpb PM 1 3.0 2.7 35 28 37 410 71 0 7 0 0 0 280 0 0 92 0 9 7 THl YM 5 2.8 2.4 52 22 27 270 0 0 0 0 0 0 125 0 0 0 0.2 4

TRA(u2) PM 3 0.7 12 35 29 26 6500 0 0 0 0 0 0 1300 0 0 260 0 130 33 TH2 YM 1 3.0 25 32 14 54 3000 0 0 0 0 0 0 1100 0 0 0 17 53

TRA(p) PM 3 3.7 9.8 16 20 64 3800 2400 0 0 0 0 0 820 0 0 510 0 20 49

TRA(p)? PM '•} 0.0 13 15 25 60 6100 4800 0 0 0 0 0 1500 0 0 280 0 170 100

TRA(m) PM 5 2.0 4.3 5 67 28 190 170 0 0 0 0 0 480 2 0 0 0 0 12 TCP YM 22 0.9 9.6 13 44 43 250 0 0 a 0 0 0 880 0 0 0 ' 8 25

TRA(ir) PM 10 21 6.9 28 47 25 560 0 0 0 0 6 0 370 0 0 5 0 2 19 TRA(sw) RR 2 5.8 5.0 22 50 28 1400 900 15 0 0 9 0.5 410 0 0 40 0 16 11 TCB YM 35 0.5 12 19 36 45 2900 a 0 0 0 0 0 1200 0 0 0 28 34

TCT RR 1 1.4 1 . 30 38 32 4000 9 0 0 0 0 0 770 0 0 99 4 77 23 TCT YM 31 6.3 11 33 34 3500 0 (1 n n n 0 1000 n 0 n 25

Present as pseudomorphic forms in most samples. Table 7. Frequency distribution of Or+Cn and BaO contents for sanidine phenocrysts in samples of Petrologic Units from Pahute Mesa, Yucca Mountain, and Rattlesnake Ridge. (Each value represents the number of analyses that have orthoclase+celsian (Or+Cn) end-member contents, and barium oxide weight contents, within the interval indicated. Results combined for all samples of each unit. Data for Yucca Mountain from Warren et al., in prep.)

Number Petrologic of Or+Cn, MoU BaO wtt Unit Location Samples 70 60 50 40 30 0 0 6 1 2 1.8 2.4 n i i 1 1 1 1 1 1 ' 1 1 1 1 1 1 1 11 1 r I i 1 I1 1 ! 1 TRA(ul) PM 12 11 81 48 20 5 4 136 27 4 2 Tpb PM 1 8 3 3 13 1 TH1 YM 10 6 38 25 3 1 6 67 11

TRA(u2) PM 3 I 13 22 2 1 1 1 1 1 6 17 11 6 2 1 TH2 YM 6 1 3 18 17 22 36 4 2 9 6 4

TRA(p) PM 3 9 9 t 3 7 3 2 1 1 2 20 7 10 5 1 TRA(p)? PM 3 2 11 10 10 3 2 3 1 1 26 9 8

TRA(m) PM 5 1 1 2 3 14 34 16 7 1 1 70 8 1 1 TCP YM 20 1 1 21 74 23 3 3 1 1 I 118 10 1

TRA(lr) PM 10 1 5 35 22 13 5 5 1 3 7 14 11 7 1 3 2 98 26 7 1 1 1 1 TRA(sw) RR 2 13 16 1 15 13 2 TCB YM 26 2 59 185 37 9 3 2 1 1 1 1 1 45 129 98 26 3 1 " TCT RR 1 12 4 7 3 5 1 c TCT YM 25 1 1 3 38 157 8 1 1 69 66 33 14 II ! , "l 1 1 1 1 1 I 1 1 1 1 1 1 i i I I t but its position with respect to the TRA(u2) unit is uncertain. A 500-m-thick unit in Ue20f has been identified as the TRA(p)? unit (see Tables 7-11); this unit directly underlies the TRA(u2) unit. Samples of the TRA(p) unit characteristically have consistently high plagioclase contents, and most contain abundant hornblende in addition to biotite (see Table 6). Sanidine (Table 7) and plagioclase (Table 8) compositions of the TRA(p) unit are significantly more sodium-rich than those of the TRA(u2) unit, and sanidine is considerably more potassium-rich than sanidine of the underlying TRA(m) unit. However, a sample of the TRA(p) unit cannot be confidently distinguished from one of the TRA(lr) unit on the basis of mineral chemistry (compare units in Tables 7-10). Furthermore, some individual samples of the TRA(lr) unit that have relatively high plagioclase contents and relatively low contents of lithic fragments are petrograpMcally similar to samples of the TRA(p) unit, and so a distinction between these units based on petrography is also uncertain. The TRA(lr) and TRA(p) units can be confidently distinguished only if their stratigraphic relation to the distinctive TRA(m) unit is known.

TRA(m) Petrologic Unit The middle member of the tuffs and lavas of Area 20 consists of ash-flow tuff and bedded tuff. Lavas have not presently been recognized for the TRA(m) unit, which consists mostly of very fine (ash fall) tuff where characterized. The unit is 65 m thick in Uel9p. The TRA(m) unit is distinct due Lo a consistently and distinctly low content of quartz and high content of sanidine (Table 6). In addition, samples of the TRA(m) unit exhibit extremely low biotite contents, but contain hornblende contents that approximately equal that of biotite. Sanidine (Table 7) and plagioclase (Table 8) compositions of the TRA(m) unit are relatively sodium-rich. Sanidine compositions are particularly distinct from those of any other TRA unit (Table 7). In addition, one sample of the TRA(m) unit contains an appreciable content of Fe-rich orthopyroxene (Mg* ^.S). Such orthopyroxene is rarely found in NTS rocks and has been presently identified only in single samples of the TRA(p), Tpb, and TCP units.

TRA(lr) Petrologic Unit The lithic-rich member of the tuffs and lavas of Area 20 consists of non- to partially welded ash-flow tuff. The TRA(lr) unit is widely distributed

227 Table 8. Frequency distribution of An contents for plagiociase phenocrysts in samples of Petrologic Units of Pahute Mesa, Yucca Mountain, and Rattlesnake Ridge. (Each value represents the number of analyses that have anorthite (An) end-member contents within the interval indicated. Results combined for all samples of each unit. Data for Yucca Mountain from Warren et al., in prep.). Number of An, moU Unit Samples Location 10 20 30 40 50 60 70 •j— 1 1— I— T ~T 1 • 1 I 1 1 1 'I 1 !! 1 — TRA(ul) 12 PM 1 1 2 21 53 57 18 5 1 2 1 1 1 Tpb 1 PM - 1 1 5 3 3 1 TH1 8 YM 1 1 6 11 25 19 3 1 1

TRA{u2) 3 PM 3 B 15 5 2 3 2 1 1 1 TH2 3 YM 5 11 7 13 9 6 4 1 1 2 5 3 3 2 1 3 3 1

TRA(p) 3 PM 6 1 5 7 7 7 1 2 1 2 1 2 1 1 1 2 1 1 TRA(p)? 3 PM 1 3 13 19 2 1 1 1 3 1 1 1 1

TRA(m) 5 PM 1 19 23 14 4 1 2 1 1 2 1 1 1 1 1 TCP 16 YM 19 59 17 1 1 1 4 2 3 3 3 2 4 1 1 1 1 3 1 1 2 1

TRA(lr) 10 PM 2 26 45 13 8 3 3 2 2 2 1 5 4 3 1 8 3 1 1 1 1 TRA(sw) 2 RR 1 4 13 5 1 1 1 1 1 TCB 29 YM 3 35 127 79 16 15 3 4 6 6 6 6 6 4 6 5 1 5 5 1 1

TCT 1 RR 2 4 4 1 C 1 1 I 1

TCT 26 YM 2 5, 32 58 28 17 15, 14 9 4 | 18 0 1 13 13 4 1 5 3 3( 3| 1 1 2 2 J J 1 1 , , l i l

Table 9. Frequency distribution of molecular Mg/(Mg+Fe) contents for biotite phenocrysts in samples of Petrologic units of Pahute Mesa, Yucca Mountain, and Rattlesnake Ridge. (Each value represents the number of analyses that have molecular Mg/(Mg+Fe) end-member contents within the interval indicated. Results combined for all samples of each unit. Data for Yucca Mountain from Warren et al., in prep.).

Number of c molecular Mg/(Mg+Fe) Unit Samples Location .30 .40 .50 .60 .70 I I I I 1r i r i 1 i r IT r TRA(ul) 12 PM 3 2 13 24 6 10 2 Tpb 1 PM 1 5 1 TH1 7 YM 6 13 13 2 1 1

TRA(u2) 6 PM 7 8 12 4 TH2 6 YM 1 11 4 7 9 9 1

TRA(p) 3 PM 2 3 2 3 17 8 TRA(p)? 3 PM 3 11 7 1

TRA(m) 4 PM 2 13 3 1 2 1 TCP 6 YM 1 2 6 3 5 2 2

TRA(lr) 10 PM 1 1 7 27 12 4*134 1 3 5 TRA(sw) 2 RR 1 4 5 3 TCB 20 YM 1 3 5 24 34 4 1 I1 5 2 1 1

TCT 1 RR 2 1 I TCT 18 YM 1, 4. 5, 3! V 2, 3

226 throughout Pahute Mesa. It is >400 m thick in Uel9p and about 500 m thick in Ue20f. Except near the uppermost .portion of the unit, samples of the TRA(lr) unit consistently contain exceptionally high contents of lithic fragments (Table 6) derived almost exclusively from the underlying peralkaline units of the Silent Canyon area. These lithic fragments are unusually poorly sorted; their sizes range down to microscopic, as illustrated in Figure 2, and up to meters. Their presence provides a highly distinct petrographic characteristic of the TRA(lr) unit. Samples of the TRA(lr) unit characteristically contain a low to moderate content of felsic phenocrysts with plagioclase subordinate to sanidine and quartz (Table 6). The biotite content is low, and hornblende is absent. However, near both the top and bottom of the unit, samples have phenocryst contents as high as 26%, plagioclase is dominant, and both biotite and hornblende are abundant. Two very distinct compositions occur for sanidine in samples of the TRA(lr) unit; one with a dominant Or+Cn value of 62 mol% and the other with an Or+Cn value of 40 mol% (see Table 7). The former composition is similar to dominant compositions for other TRA units (see Tables 7,11), and represents the true phenocryst chemistry. The latter composition matches those for sanidine in samples of the underlying peralkaline rocks and for sanidine phenocrysts within the lithic fragments of the TRA(lr) unit; sanidine grains with these compositions are clearly xenocrysts. No other unit of the NTS is known to contain sanidine xenocrysts in such abundance. Plagioclase compositions (Table 8) are not unusual compared to other NTS units (Table 11), and both biotite (Table 9) and hornblende (Table 10) compositions are magnesium-poor (Table 11), typical of the tuffs and rhyolites of Area 20.

TRA(sw) Petrologic Unit The TRA(sw) Petrologic Unit is equivalent to the Stockade Wash Tuff(?) of Byers et al. (1976a) and may additionally include bedded tuff stratigraphically bounding the ash-flow cooling unit. Two samples of the Stockade Wash Tuff were examined; these were collected from locations (RW19r and RW18a, Fig. 1) that Byers et al. (1976a) consider probably not stratigraphically equivalent. The strati graphic position of the TRA(sw) unit in the Quartet Dome quadrangle (Sargent et al., 1966) has been confidently established at location RW19r, where it directly underlies the TRA(lr) Petrologic Unit ("conglomerate" of Sargent et al., 1966). The petrography and mineral

229 a.

Fig. 2 Photomicrographs of sample Uel9p--2204, TRA(lr) unit, Note the abundant lithic fragments (mottled appearance), particularly those with very small sizes, Field of view 1.9 x 2.8 mm. (a) reflected light, (b) transmitted light.

230 Table 10. Frequency distribution of molecular Mg/(Mg+Fe) contents for hornblende phenocrysts in samples of Petrologic units of Pahute Mesa, Yucca Mountain, and Rattlesnake Ridge. (Each value represents the number of analyses that have molecular Mg/(Mg+Fe) end-member contents within the interval indicated. Results combined for all samples of each unit. Data for Yucca Mountain from Warren et al., in prep.). Number of molecular Mg/(Mg+Fe) Unit Samples Location .30 .40 .50 .60 .70 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 TRA(ul) 1 PM 1 1 Tpb 1 PM 1 2 2

TRA(p) 3 PM 1 3 4 3 5 5 1 2 1 TRA(p)? 2 PM 2 10 7 1 2 1 1 1 1

TRA(m) 4 PM 1 1 3 8 4 1 1 TCP 2 YM 1 1 1 1

TRA(lr) 4 PM 2 1 2 1 4 2 2 1 TRA(sw) 2 RR 1 1 3 5 1 TCB 4 YM 1 7 5

fCT RR 1 1 1 1 1 1 1 1 ! | 1 1 t I chemistry of the sample at location RWICa from the Ammonia Tanks quadrangle (Hinrichs et al., 1967) identically matches the sample of this unit from the Quartet Dome quadrangle. Nonetheless, the correlation of the unit between these areas is considered uncertain due to a possible confusion with the TRA(p) unit.

TCT Petrologic Unit (Rattlesnake Ridge Location) A single sample of bedded tuff from Rattlesnake Ridge (location in Figure 1) is tentatively correlated with the TCT unit of Yucca Mountain. This thin (18 m thick) bedded tuff is bounded stratigraphically below by the Grouse Canyon Member of the Belted Range Tuff and above by the TRA(sw) unit. It is conceivable that the TRA(sw) unit at this location correlates with the TRA(p) rather than the TRA(lr) unit [see discussion of TRA(sw) unit]; in this case the sample of bedded tuff might represent the base of the TRA(p) unit. Such is considered unlikely because the petrography and mineral chemistry of this sample of TCT unit differs distinctly from all known or tentatively assigned samples of TRA(p) unit (see Tables 6-11).

231 Table 11. Dominant values for compositional parameter? of phenocrysts for Petrologic Units of Pahute Mesa, Yucca Mountain, and Rattlesnake Ridge. (The dominant value is equivalent to the statistical mode for the frequency distribution of analyses represented in Tables 7-10. Symbols for all Petrologic Units defined in Table 1.)

Mafic Minerals Sanidine Plagioclase Number Molecular Mg/{Mg+Fe) Petrologic of MolS wtt MoU Unit Location Samples Or+Cn BaO An Biot Hbld

TMRu a 6 60 0.00, 1.13 19, 24" 0.62 0.64

TMR1 upper PM 1 53 0.00 13 lower b 7 62 0.00 13 0.60 0.72

TPCu upper PM 1 49 0.05 0.68 0.65 lower PM 3 36 0.05 0.66 0.70 TRPP PM 9 49 0.75 13 0.65

lKAtul) PM 12 69 0.12 20 0.37 0.46 Tpb PM 1 71 0.09 20 0.35 0.39 THl YM 10 68 0.10 20 0.38

TRA(u2) PM 3 69 0.58 25 0.44 TH2 YM 7 71 0.21 24 0.43 TRA(p) PM 3 63 0.36 17 0.43 0.47 TRA(p)? PM 3 61 0.25 13 0.42 0.44

TRA(m) PM 5 53 0.05 12 0.37 0.37 TCP YM 22 53 0.14 11 0.42 TRA(lr) PM 10 62, 4Cb 0.13, 0.00b 15 0.37 0.43 TRA(sw) RR 2 62 0.29 15 0.40 0.44 TCB YM 35 61 0.56 15 0.40 0.44

TCT RR 1 69 0.27 22 0.42 TCT YM 31 67 0.55 21 0.42

TLR YM 16 65 0.67 18 0.59 TTA YM 8 64 0.55 18 0.55 TTB YM 2 66 0.96 26 0.59 TTC YM 6 72 3.4 31 0.62

Locations throughout NTS Bimodal population K-rich anorthoclase. Ca-rich plagioclase (mantled by K-rich feldspar) also present.

232 Other Petroiogic Units . Most other Petrologic Units discussed in this report (Table 1) consist of well-described ash-flow cooling units, mostly from Yucca Mountain (location shown in Figure 1). Each such Petrologic Unit also includes stratigraphically bounding bedded tuff that can be petrographically and chemically related to the ash flow, but this added tuff is so volumetrically insignificant that the distinction between Petrologic and Lithologic Units is unimportant. Descriptions of units younger than the TRA units are found in Byers et al. (1976a), and descriptions for Petrologic Units of Yucca Mountain are contained in Warren et al. (in prep.).

COMPARISON OF TRA PETROLOGIC UNITS OF PAHUTE MESA WITH UNITS OF YUCCA MOUNTAIN The occurrence of substantial petrographic variations within single lithologic units of the NTS are well documented. Many NTS ash-flow cooling units have mafic-rich caprocks (e.g., Lipman et al. 1966; Byers et al., 1976a). Substantial petrographic variations are now well documented for ash- flow cooling units that do not have mafic-rich caprocks; for example, tht upper portion of the TCB unit of USW-G1 is quartz-rich, but the lower portion is quartz-poor (Carr et al., in press; Warren et al., in prep.). The TPCu unit is a single cooling unit only 22 m thick in Uel9p, but the phenocryst content differs by a factor of 3.5 between top and bottom of the unit and mafic mineral and trace mineral contents differ by an order of magnitude or more (Warren, in prep.). It is clear that primary minerals (phenocrysts) are not homogeneously distributed within most magmas. Furthermore, it might be expected that the upper portion of a magma might concentrate volatiles. Explosive release of these volatiles would result in eruption of the upper portion of the magma as tuff, whereas the lower, more volatile-depleted portion of the magma would tend to erupt more quietly, as a lava flow. Such an eruption would result in both tuff and lava from the same magma. Historical eruptions in such a manner are known (Hildreth and Drake, 1983). Additionally, large-volume eruptions of tuff will occur near-source as massive and poorly sorted ash flows, but as well sorted (bedded) tuffs farther from the source (see, for example, Figure 8 in Sheridan, 1979). Overprinted on these primary features due to emplacement mechanics are the zonal cooling features of compaction and crystallization described by Smith (1960). Thus, although eruption of a given magma might provide a time-stratigraphic marker,

233 the recognition of this "pulse" might be very difficult due to vertical ami lateral variations both in the lithology and in the primary mineral content in the erupted material. Fortunately, the mineral chemistry of most eruptive sequences of the NTS has been found to be unaffected by petrographic and lithologic variations. For .example, substantial petrographic variations occur both vertically and horizontally within the TCB unit (Warren et al., in prep.). Although samples from this unit were obtained from numerous locations, separated by as much as 55 km, compositions for sanidine are identical at all locations within the small analytical uncertainty of about 1 mol% Or+Cn. Similar results have been obtained for other minerals., particularly biotite and plagioclase. In general, each unit has a unique set of mineral compositions that differs from those of all other NTS units end serves as an invaluable aid in its recognition. The similarity in mineral chemistries between the TRA(p) and TRA(sw) units noted in their descriptions is an unusual, and highly significant exception, it should be noted, however, that not all units show as remarkable a vertical and lateral consistency in mineral compositions as does the TCB unit. There are significant systematic variations in sanidine composition within the Topopah Spring (Broxton et al., 1982) and Rainier Mesa rhyolites that could be employed to define stratigraphic levels within these units.

Comparison of Mineral Chemistry Although there are small (but significant) differences among individual units, all TRA units have distinctively Mg-poor mafic minerals. In contrast, mafic minerals in units of the overlying Paintbrush Tuff and allied lavas of Pahute Mesa have characteristically Mg-rich compositions (see Table 11). The peralkaline rocks that underlie the TRA sequence are extremely biotite-poor (most samples lack biotite) and instead contain olivine and clinopyroxene phenocrysts that in many cases are nearly pure Fe end members. At Yucca Mountain, mafic minerals in the Crater Flat Tuff and overlying tuffs and lavas of Calico Hills have Mg-poor compositions that are identical to those of the TRA units of Pahute Mesa (see Tables 9-11). The overlying Paintbrush Tuffs, except for the Topopah Spring rhyolite, have characteristically Mg-rich mafic minerals identical to those in the Paintbrush Tuffs and allied lavas of Pahute Mesa. The Crater Flat Tuff of Yucca Mountain, however, overlies the Lithic Ridge Tuff and associated lavas rather than the

234 peraikaline rocks that underlie the TRA units of Pahute Mesa. The Lithic Ridge and older units are generally plagioclase-rich units (Warren et al., in prep.) that bear Mg-rich mafic minerals (Table 11). It is clear that the TRA units of Pahute Mesa and the Crater Flat and Calico Hills tuffs of Yucca Mountain have compositions for mafic minerals that are indistinguishable from each other but are highly different from those of bounding volcanic groups. In contrast to mafic mineral compositions, sanidine and plagioclase compositions differ distinctly among units, and can be compared unit-by-unit throughout the stratigraphic columns of Pahute Mesa and Yucca Mountain (Figure 3). Sanidine compositions (Table 7) are identical within the small analytical uncertainties for each pair of units in the stratigraphic sequence at both locations. All compositions of individual phenocrysts that differ substantially from those of the dominant composition (Table 11) are due to xenocrysts

PAHUTE MESA

Paintbrush Tufts & Allied Lavas |200m .

YUCCA MOUNTAIN TRA(u1)

Paintbrush Tuffs TH1 TH2 TRA

TCP TRA (pi* TCB TRA (m)

TCT TRA llr) /TCT Lavas TLR TSCP

* Relative stratigraphic positions uncertain

Fig. 3 Correlation diagram between Petrologic Units of Yucca Mountain and Pahute Mesa. (Symbols for units defined in Table i. TSCP = peralkaline rock of the Silent Canyon area. Greatest known or inferred thickness for each unit illustrated.)

235 or to the occurrence of narrow sodium-rich rims (see Figure 4) produced from late-stage deuteric• alteration (Warren et al., in prep.). Matching units, stratigraphically downward (unit present at 'Pahute Mesa listed first), are TRA(ul) and TH1; TRA(u2) and TH2; TRA(m) and TCP; TRA(lr) and TCB; and TCT and TCT. A unit equivalent to the TRA(p) unit of Pahute Mesa is not known for Yucca Mountain. Plagioclase compositions (Table 8) match in similar fashion. Although the TRA(lr) unit of Pahute Mesa contains sanidine whose composition (dominant Or+Cn value = 62, Tables 7 and 11) matches the TCB unit of Yucca Mountain, it also contains a population of sodium-rich sanidine xenocrysts that do not occur in the TCB unit. Additionally, there are significant differences in the barium contents of sanidine for many paired units, most notably between the TRA(lr) and TCB units (see Table 11). Discussion of these differences is deferred to a following section.

Comparison of Lithology and Petrography The primary mineral contents of the TRA units of Pahute Mesa are quite similar to their paired unit of Yucca Mountain (Table 6). At both locations,

Fig. 4 Transmitted light photomicrograph (crossed nicols), of sodium-enriched rim of sanidine in sample CFLSM-5, TCB unit. Field of view 0.24 x 0.35 mm.

236 these units contain abundant quartz, and variable amounts of biotite, hornblende, and allanite. Pyroxene, sphene, and perrierite rarely occur in samples of these units. These petrographic characteristics contrast strikingly with those of stratigraphically bounding units of both areas (Byers et a!., 1976a; Warren et al., in prep.; Warren, in prep.). The match in primary mineral contents is excellent for several of the paired units of Figure 3, particularly for the TRA(ul) and TH1 units. The most substantial lithologic and petrographic differences between paired units occur between the TRA(lr) and TCB units. Samples of these units are grossly dissimilar in hand sample; the TRA(lr) unit is exceptionally lithic-rich and generally non-welded whereas the TCB unit is quite lithic-poor and always occurs as a well-defined cooling unit. Portions of the TCB cooling unit may be densely welded or even vitrophyric. Differences in the primary mineral contents of these units are not as great, but nonetheless they are appreciable. Most conspicuous is the much higher median biotite content of the TCB unit (0.20%) compared to that of the TRA(lr) unit (0.06%). However, highly phenocryst- and mafic-rich zones occur within the TRA(lr) but not within the TCB unit. Thus, the average phenocryst and mafic contents of the TRA(lr) unit are considerably higher than median values, and averages probably match much better for the two units than the medians. The TRA(m) unit of Pahute Mesa also differs appreciably in lithology and primary mineral content from the TCP unit of Yucca Mountain. The TRA(m) unit generally has a low to moierate content of phenocrysts, mostly sanidine (Table 6), and occurs as a well-sorted (bedded) tuff where presently characterized. It is distinctly quartz-poor. The TCP unit contains a considerably higher phenocryst content, and occurs as a well-defined ash-flow cooling unit. Both units, however, contain a highly distinctive, magnesium-poor orthopyroxene that provides a striking correlation of primary mineral contents.

DISCUSSION OF RESULTS: EVIDENCE FOR A COMMON MAGMATIC ORIGIN OF VOLCANIC SEQUENCES THAT FLANK THE TIMBER MOUNTAIN CALDERA Each TRA unit of Pahute Mesa and correlative unit of Yucca Mountain (Figure 3) contains Mg-poor mafic minerals, consisting primarily of biotite and/or hornblende. The chemistry of these minerals differs distinctly from those of bounding volcanic groups, and suggests that the sequences at each location were derived from closely similar magmas or from the same magma. By

237 contrast, the chemistry of sanidine and plagioclase differs markedly between successive Petrologic Units. These large chemical differences within each stratigraphic succession are parallel at each location. It is considered highly unlikely that separate magmas could evolve in such a manner to produce the observed correspondence in feldspar chemistry. The mineral chemical data thus strongly indicates that each of the paired units of Yucca Mountain and Pahute Mesa was derived from the same magma. It is unlikely that Petrologic Units that include lavas have been emplaced very far from their source; such units include the TRA(ul), TRA(u2), and TRA(p) units of Pahute Mesa and the TH1 and TH2 units of Yucca Mountain. Thick exposures of the TH units occur north of Yucca Mountain and at Calico Hills (Byers et al., 1976b). If these lavas ascended directly upward from a magma chamber, a very large magma chamber must have been present during eruption of the TH and TRA(u) units if indeed these units represent magma from the same chamber. The distribution of TRA and TH lavas (shown in Figure 5) may then represent the approximate limits of this magma chamber. The Timber Mountain Caldera, a much younger feature, is located near its center. Lavas intermediate in age between the tuffs and lavas of Area 20 and the Timber Mountain Tuff have also been correlated across thre Timber Mountain Caldera (see Figure 19 in Byers et al., 1976a). This suggests that the Timber Mountain area has been the center of a sequence of chemically different magma systems throughout an extensive period of time. It is probable that these magma systems, which produced a thick sequence of NTS volcanic rocks including the tuffs and lavas of Calico Hills, tuffs and lavas of Area 20, Paintbrush Tuff and Timber Mountain Tuff are related as a single, large evolving magma system centered beneath Timber Mountain. A model that might describe such a magma system is described by Hildreth (1981). Each pair of units found on opposite sides of Timber Mountain might have been erupted from a central source, the present location of Timber Mountain, or from separate sources. If the former is true, Pahute Mesa and Yucca Mountain are marginal to a major structure (e.g., caldera) associated with eruption of these units. The latter possibility, eruption from separate sources, is consistent with presently accepted views. In this case, the Crater Flat Tuffs and tuffs and lavas of Area 20 were erupted from structures that flank Timber Mountain (Carr, 1982; Orkild et al., 1968). It is likely that the magma chamber was not fully emptied during the final stages of eruption and the remaining liquid has

238 %:

Fig. 5 Location of lavas of the TRA(u) unit (diagonal pattern north of Timber Mountain) and of the TH unit (same pattern south of Timber Mountain). Non- welded tuff shows a wider distribution for both units. The single magma chamber inferred to be the source of both lavas is centered beneath Timber Mountain; the distribution of the lavas may approximate the location of the outer portion of a magma chamber.

239 solidified to form a large piuton beneath Timber Mountain. If the size and shape of the magma chamber did not change substantially through time, then the distribution of lavas shown in Figure 5 may also approximate the subsurface location of such a pluton, which presumably would be a granitic body. The nearly identical mineral chemistry for units erupted from opposite sides of this hypothetical magma chamber indicates that chemical equilibrium between phenocrysts and liquid is closely maintained throughout an extraordinarily large volume of magma. This is true even where substantial variations of phenocryst concentrations occur within a unit. This indicates that the mayma has an inhomogeneous distribution of phenocrysts, and the concentration observed within a unit depends upon the precise position that the sample occupied as a liquid in the magma chamber. Probable causes for variations in phenocryst content throughout a magma chamber are temperature and water pressure variations. Such factors might also affect the mineral chemistry, but the data at hand indicate that for many units they do not. There is good evidence (Warren, in preparation) that mixing of magmas occurs at the margins of the magma system. This is not an equilibrium process and non-equilibrium conditions are evident in units a^ected by magma mixing. The crystallization sequence is preserved, however, by variations in the barium contents of sanidine. Sanidioe has a marked tendency to concentrate barium (Leeman and Pheips, 1981; Hanson, 1978), which readily substitutes for potassium because these elements are very large ions of similar size. As crystallization proceeds, changing physiochemical conditions require the equilibrium composition of sanidine to change. The primary change usually requires progressive replacement of orthoclase component (KAlSi-CL) by albite component (NaAlSi^Og) as temperature decreases. This is rapidly accomplished under magmatic conditions simply by exchange of Na with K. Early crystallized sanidine contains relatively high Ba concentrations compared to sanidine formed later. However, equilibrium for Ba between early and late sanidines requires an exchange of celsian component (BaAlpSipOg) with albite. This equilibrium cannot be attained simply by exchange of Ba with Na, because it also requires a coupled substitution of Al with Si. The latter elements are present in a much more tightly bound (tetrahedral) structural site within sanidine, and require considerably greater energies for exchange. Consequently, although the Na and K exchange is rapid and complete up to the time of eruption and reflects equilibrium, Ba exchange is limited or does not

240 occur. This process is here termed "limited equilibrium." Thus, portions of the magma that began to crystallize sanidine earliest will have the highest Ba contents, and those that formed sanidine latest will have the lowest. Thus, the differing Ba contents for sanidine between paired Petrologic Units of Yucca Mountain and Pahute Mesa simply relate to the crystallization history of the magma, and not to inherent compositional differences between the units. The distribution and concentration of volatiles (primarily water) within the magma chamber is very important, because they certainly provide the driving force of the eruption. The higher their concentration, the more violent an eruption might be expected. The element cesium concentrates strongly in the upper portion of the magma that produced the Bishop Tuff (Hildreth, 1979); this element shows an extremely strong association with volatiles. Cesium concentrations in the TRA(lr) unit are extraordinarily high, generally more than an order of magnitude higher than those of other NTS units, including the TCB unit (Warren, in prep.). The cesium contents suggest that volatiles concentrated strongly in the portion of the magma chamber that erupted the TRA(lr) unit, but not in the portion of this same magma chamber that erupted the TCB unit. These volatiles drove an unusually violent eruption of the TRA(lr) unit that resulted ii the incorporation of remarkably large and poorly sorted lithic fragments within the TRA(lr) unit.

CONCLUSIONS AND FUTURE RESEARCH A very striking geochemical similarity, based on chemistry of primary mineral phases, has been demonstrated between volcanic groups that flank the Timber Mountain Caldera. It is concluded that paired units within the two locations were derived from the same magma. Lithologic and petrographic differences between units flanking Timber Mountain that have similar mineral chemistries are attributable to an inhomogeneous distribution of phenocrysts and volatiles within the magma. The distribution of paired units that contain thick lavas indicates that the magma body was very large and centered on Timber Mountain. The similarities in mafic mineral compositions among an entire group of units suggest that each successive unit represents the evolution of a single, large magma system. It is the author's belief that the hypothesized magma system centered at Timber Mountain was also associated with the eruption of the oldest known NTS volcanic rocks. The mineral chemistry is presently being investigated for

241 such units of Pahute Mesa for comparison with well-characterized units of Yucca Mountain (Warren et al., in prep.). The Petrologic Unit introduced here serves as a valuable correlative tool for bedded tuffs, and is particularly valuable for drill holes that do not penetrate the ash-flow cooling units used as marker beds. Volcanic units in other areas of the NTS (e.g., Yucca Flat) can probably be related to those of Pahute Mesa and Yucca Mountain by utilizing the Petrologic Unit. Such use is required to correlate older, poorly exposed units effectively over a large region such as the NTS due to marked lithologic differences that may occur amonp such units at different locations.

ACKNOWLEDGMENTS This work has built on 20 years of geologic mapping and study, largely by geologists of the U.S. Geological Survey but also geologists of several other organizations. Many of their names are found among the references cited. The correlations described herein could not have been accomplished without the basic stratigraphic framework developed by previous NTS workers. David E. Broxton, Frank M. Byers, Jr., and B. W. Smith provided very timely and thoughtful reviews. The advice and counsel of Frank M. Byers, Jr., has been very helpful. Roland C. Hagan and Lois F. Gritzo have greatly aided in the microprobe analyses. The ability of Marcia A. Jones to produce complex tables without error and with remarkable speed is simply astounding. Similarly, the ability of David A. Mann and Tino Lucero to create wonderful probe sections from rocks that fall apart in one's hands is remarkable. This research has been vigorously supported by Thomas A. Weaver and Wayne A. Morris, past and present group leaders of the ESS-2 (Geochemistry) group, and by Jack W. House, program leader for containment. I am very grateful to all for their generous contribution.

REFERENCES Broxton, D. E., D. Vaniman, F. Caporuscio, B. Arney, and G. Heiken, "Detailed Petrographic Descriptions and Microprobe Data for Drill Holes USW-G2 and Ue25b-1H, Yucca Mountain, Nevada," Los Alamos National Laboratory report LA-9324-MS (1982). Byers, F. M., Jr., W. J. Carr, P. P. Orkild, W. D. Quinlivan and K. A. Sargent, "Volcanic Suites and Related Cauldrons of Timber Mountain - Oasis Valley Caldera Complex, Southern Nevada," U.S. Geol. Suirv. Prof. Papr- 919, 69 pp. (1976a).

242 Byers, F. M., Jr., W. J. Carr, R. L. Christiansen, P. VI. Lipman, P. P. Orkild, and W. D. Quinlivan, "Geologic Map of the Timber Mountain Caldera Area, Nye County, Nevada," U.S. Geol. Surv., Denver, CO, Map 1-891, scale 1:48,000 (1976b).

Caporuscio, F., D. Vaniman, D. Bish, D. Broxton, B. Arney, G. Heiken, F. Byers, R. Gooley, and E. Semarge, "Petrologic Studies of Drill Cores USW-G2 and Ue25b-1H, Yucca Mountain, Nevada," Los Alamos National Laboratory report LA-9255-MS (July 1982).

Carr, W. J., "Volcano-Tectonic History of Crater Flat, Southwestern Nevada, as Suggested by New Evidence From Drill Hole USW-VH-1 and Vicinity," U.S. Geol. Surv. Open-file report 82-457, 23 p. (1982).

Carr, W. J., F. M. Byers, Jr., and P. P. Orkild, "Stratigraphic and Volcano- Tectonic Relations of Crater Flat Tuff, Nevada Test Site Region, Nye County, Nevada," U.S. Geol. Surv. Prof. Paper (in press).

Hanson, G. N., "The Application of Trace Elements to the Petrogenesis of Igneous Rocks of Granitic Composition," Earth Planet. Sci. Letters, 38, 26-43 (1978). ~~

Hildreth, W., "Gradients in Silicic Magma Chambers: Implications for Lithospheric Magmatism," J. Geophys. Res. 86(B11)9 10153-10192, Nov. 10 (1981).

Hildreth, W., "The Bishop Tuff: Evidence for the Origin of Compositional Zonation in Silicic Magma Chambers," in Ash Flow Tuffs, C. E. Chapin and W. E. Elston, eds., Geol. Soc. Amer. Special paper 180 (Geol. Soc. Amer., Inc., Boulder, CO), pp. 43-76 (1979).

Hildreth, W. and R. E. Drake, "1932 Eruption of Quizapu, Central Chilean Andes," Geol. Soc. Amer. Rocky Mountain and Cordilleran Sections, Annual Meeting, Salt Lake City, UT (1983).

Hindrichs, E. N., R. D. Krushensky, and S. J. Luft, "Geologic Map of the Ammonia Tanks Quadrangle, Nye County, Nevada," U.S. Geol. Surv., Denver, CO, Map GQ-638, scale 1:24,000 (1967).

Leeman, W. P. and D. W. Phelps, "Partitioning of Rare Earths and Other Trace Elements Between Sanidine and Coexisting Volcanic Glass," J. Geophys. Res. 86(311), 10193-10199, Nov. 10 (1981).

Lipman, P. W., R. L. Christiansen, and J. T. O'Connor, "A Compositionally Zoned Ash-Flow Sheet in Southern Nevada," U.S. Geol. Surv. Prof. Paper 524-F, 47 pp. (1966).

Orkild, P. P., F. M. Byers, Jr., D. L. Hoover and K. A. Sargent, "Subsurface Geology of Silent Canyon Caldera, Nevada Test Site, Nevada,11 in E. B. Eckel (Editor), "Nevada Test Site," Geol. Soc. Am. Memoir 110, 77-86 (1968).

Sargent, K. A., S. J. Luft, A. B. Gibbons, and D. L. Hoover, "Geologic Map of the Quartet Dome Quadrangle, Nye County, Nevada," U.S. Geol. Surv., Denver, CO, Map GQ-496, scale 1:24,000 (1966).

243 Sheridan, M. F., "Emplacement of Pyroclastic Flows: A Review," in Ash Flow Tuffs, C. E. Chapin and W. E. Elston, eds., Geol. Soc. Amer. Special paper WTGeol. Soc. Amer., Inc., Boulder, CO), pp. 125-136 (1979). Smith, R. L., "Zones and Zonal Variations in Welded Ash Flows," U.S. Geol. Surv. Prof. Paper 354-F, p. 149-159 (1960).

Warren, R. G., "Use of Petrology and Geochemistry to Define Petrographic Units in Drill Holes Uel9p, Uel9p-1, and Uel9ab, Southeastern Pahute Mesa, Nevada Test Site," Los Alamos National Laboratory report (in preparation).

Warren, R. G., F. M. Byers, Jr., and F. A. Caporuscio, "Petrography and Mineral Chemistry of Units of the Topopah Spring, Calico Hills, and Crater Flat Tuffs, and Older Volcanic Units, with Emphasis on Samples from Drill Hole USW-G1, Yucca Mountain, Nevada Test Site," Los Alamos National Laboratory report (in preparation).

244 SOIL PROPERTIES AS AGE INDICATORS FOR QCIATERNARY SURHCIAL DEPOSITS AND FACJLTS IN THE NEVADA TEST SITE AREA, SOUTHERN NEVADA Shroba, R. R., U. S. Geological Survey, Denver, CO 80225 Faults are common geologic features in the NTS area. Trios*' in the vicinity of subsurface nuclear tests mCj pose a potential threat to cor.t^inment. The tectonic history of local faults, especially the age of the last movement, is therefore an important consideration for nuclear testing in Yucca Flat or in adjacent areas.

Soils are useful criteria for estimating the ages of Quaternary surficial deposits and faults in the southwestern United States. Many soil properties change systematically with time, and therefore reflect the age of the parent material(s). Ages of surficial deposits estin kited from soils can be used as constraints on ages of fault movement.

Aye-related properties of .soils in the NTS area include secondary clay content; clay- film morphology: structure, color, and thickness of B horizons; and the morphology, secondary carbonate content, and thickness of calcic (Cca and K) horizons. The presence and thickness of vesicular A (Av) horizons, and the degree of development of desert varnish on surface siones. are also useful age indicators.

In the NTS area soils with Al. weak cambic B (Bs), and Stage-I Cca horizons form in about 10 thousand years (ky) or less. These soils lack the thin, silty colian mantles in which the Av horizons of elder soils are formed. Stones on the surface of soils of this age are unstained to slightly stained by desert varnish. Soils with Av, Bs, and Stage-ll Cca horizons form in less than 100 ky. These and older soils have well-varnished stones on their surfaces. Soils about 100 to 250 ky old have Av, Bs, or argillic 3 (Bt), and strong Stage-ll Cca or weak Stage-Ill K horizons. Soils containing K horizons with strong Stage-Ill or Stage-IV carbonate are greater than 250 ky old. Rates of soil development in the NTS area are strongly influenced by aridity and the addition of secondary clay and calcium carbonate from atmospheric sources. These rates appear to be slower than those reported for other areas of the southwestern United States.

245 UCRL-89415 PREPRINT

INFLUENCE OF GEOLOGIC STRUCTURE ON ALLUVIAL SEDIMENTATION IN NORTHWESTERN YUCCA FLAT, NYE COUNTY, NEVADA

J. L. Wagoner

This paper was prepared for submittal to Second Symposium on Containment of Underground Nuclear Explosions, Kirtland AFB, Albuquerque, New Mexico, August 2-4, 1983.

August 11, 1983

This is a preprint of a paper intended for publication in a journal or proceedings. Since changes may be made before publication, this preprint is made available with the understanding that »t will not be cited or reproduced without the permission of the author.

247 INFLUENCE OF GEOLOGIC STRUCTURE ON ALLUVIAL SEDIMENTATION IN NORTHWESTERN YUCCA FLAT, NYE COUNTY, NEVADA*

ABSTRACT With the aid of the LLNL downhole fisheye movie camera, alluvial sediments are described from five boreholes in Yucca Flat. Stratigraphic units are delineated based on overall grain size and maximum clast size. The five boreholes are located on or near the present-day Grouse Canyon alluvial fan in northwestern Yucca Flat. In the subsurface, the fan is transected by a north-south-trending buried ridge of Paleozoic sedimentary rocks. During and after deposition of the Tertiary tuffs, this horst was gradually uplifted, becoming a topographic feature, and dividing Yucca Flat into a western and eastern basin. The relatively thin mantle of soft volcanics was quickly eroded off the horst., forming a fine-grained basal tuffaceous alluvium in both basins. The Paleozoic sedimentary rocks were then exposed, providing a source of coarse detritus for the proximal alluvial deposits along each side of the ridge. Uplift of the horst eventually abated, and in conjunction with rapid sedimentation in the basins, the ridge was buried by alluvium. The main source of detritus then shifted to the western basin margin, with proximal sedimentation of coarse alluvium continuing along the range front and becoming finer-grained distal sedimentation toward the center of the present Yucca Flat basin. These changes in the alluvial sedimentation are documented by the downhole photography, a tool which provides a unique opportunity to study subsurface alluvial deposits and their response to tectonic development of a basin. INTRODUCTION In support of the LLNL Containment Program, many types of geological and geophysical data are collected in order to understand the geologic setting and insure containment of underground nuclear tests. A relatively new and unique exploration technique is the LLNL downhole fisheye motion picture camera (Brugman, 1979). This device produces a continuous color movie of the entire emplacement hole and provides a clear picture of the rocks and sediments which occur at the site. The photography allows the geologist to identify faults, occurrence and character of carbonate constituents, and other geologic parameters which are important to containment. Indeed, many faults would not be detected without downhole photography. The fisheye camera also provides a unique opportunity to study the subsurface alluvial deposits of a nonmarine basin such as Yucca Flat. The geologist can measure bed thickness, color, estimate clast size, describe sedimentary textures, and classify alluvial sediments. All of these parameters are important to understand the source of and depositional environment of the sediments. Three-dimensional analysis of the deposits defines facies changes and variation of physical properties of the alluvium. This type of analysis is especially important in northern Yucca Flat, where alluvium is typically very coarse-grained and locally contains large amounts of detrital and secondary carbonate. * Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract number W-7405-ENG-48.

248 The purpose of this study is to document the alluvial stratigraphy in a part of Yucca Flat.By using the downhole photography, we can define the significant changes in alluvium clast size and changes in source areas. Through detailed documentation of the alluvium stratigraphy, we can learn more about the structural development and evolution of Yucca Flat basin.

METHODS Downhole fisheye motion pictures are routinely run in all LLNL emplacement holes at the Nevada Test Site. The LLNL drilling method of dual string-reverse-air and water circulation generally results in clean borehole walls, which a^e well-adapted to photographic analysis. Each movie frame consists of a circular downhole-view photo, an =izimuthal heading, and a depth reading (Fig. 1). Using a special projector, each frame can be stopped for detailed viewing and analysis.

Fig. 1 Dov;nhole photograph from the LLNL fisheye camera showing azimuth (213°) and depth (655.81). The five emplacement holes selected for this study are located in southwestern Area 2 and northern Area 4 (Fig. 2). Except for one hole, all sites are located on the alluvial fan which emanates from Grouse Canyon in the Eleana Range. The other hole is located off the far eastern edge of the fan near the center of the basin. The alluvial stratigraphy is described in detail for each site. The sediments were classified into four groups based on grain size. Each sediment classification is based on individual percentages of size classes (Table 1), estimated visually from the photography. Changes in color are also noted,

249 Table 1

Sediment Classification Description

Gravel >75% clasts >2cm diameter, generally clast-supported, commonly contains very large boulders. Sandy Gravel >50% clasts >2cm diameter, usually clast-supported, contains more sand and finer grain sizes than above. Gravelly Sand <50% clasts >2cm diameter, usually matrix-supported, with variable %'s of gravel-size clasts. Sand <5-10% clasts >2cm diameter, matrix-supported, contains variable amounts of sand, silt, and clay. Because of the limited resolution of the fisheye camera photography, clast sizes less than a few centimeters diameter cannot be accurately measured. This sediment classification scheme has been devised to meet the needs of this study, and thus is different from classification systems based on grain size distribution analyses.

250 •'//.-.£ .

•'; »rs

*• »

AIIU¥l.,

-'..-.- ' :•• *, - y

!> .-*

V, -- "j ; A\^ •" -.'• V \" ' ) %\ M Fig. 2 Location of Grouse Canyon fan and holes in northwestern Yucca Flat. Also shown are individual fan boundaries, Carpetbag and faults, and location of alluvial fan facies. although separate units are not defined based on this criterion. Contacts between stratigraphic units are characterized as either conformable or erosional. In addition to defining the general stratigraphy at each site, we also measured the maximum clast size observed in individual units throughout the section. The maximum clast size is the measurement of the maximum cross-sectional projection of the largest clast observed in a stratigraphic horizon. The m~vie frame is projected onto a calibrated circular grid, from which the clast measurement is made (Fig. 3). Maximum clast size is plotted versus depth on the stratigraphic column for each hole.

Fig. 3 Calibrated grid used in measuring clast size from the fisheye movie. Each division (A) equals 21 cm diameter. A transparent template is used to overlay the movie frame, which is projected on a screen.

252 GEOLOGY OF YUCCA FLAT Yucca Flat is a north-south-trending arid basin in the southwestern Basin and Range geologic province. The basin is surrounded by ranges composed of Paleozoic and Precambrian clastic and carbonate sedimentary rocks, Mesozoic granitic rocks, and Tertiary volcanic tuff (Fig. 4). Within the basin, there is a complex system of normal faults, with many faults penetrating up through the alluvium. Two large faults control the geometry of the basin, and both have surficial expressions on the valley surface (Fig. 2). The Yucca fault trends north-south and is located in the east-central part of the basin. The fault dips to the east and forms a large westward-tilting structural block. The Carpetbag fault strikes north-south and occurs in the western part of the basin. This fault also dips to the east and represents the eastern edge of a large north-south trending buried horst of Paleozoic carbonate rocks.

The general geometry of the horst is defined by both surface gravity and seismic reflection data. At the shallowest measured point, the Paleozoic rocks are buried by only 40 m of alluvium. The alluvium thickness is quite variable along the horst, due to faulting and the irregular erosional surface of the Paleozoic rocks. There are remnant patches of Tertiary tuff along the sides of the ridge, but in general, most of the volcanics probably were eroded off the high.

TYPES OF ALLUVIAL DEPOSITS The Tertiary-Quaternary alluvial deposits in the Yucca Flat basin are derived from detritus eroded and washed from the ranges surrounding the basin. Deposition of sediment onto alluvial fans takes several forms. Channel and braided stream deposition are probably the most important, occurring during summer and winter storms. The resulting deposits are generally moderately to well bedded and consist of fine- to coarse-grained sand, gravelly sand, and sandy gravel, Another important mode of deposition is debris flow. A debris flow is a viscous mass of coarse sediment in a muddy matrix which flows downslope. Debris flows occur where sediment sources provide abundant silt and clay-size material, slopes are steep, and during periods of intermittent heavy rainfall. The resulting deposit consists of cobbles and boulders imbedded in a matrix of fine-grained sediment, are poorly sorted, and usually internally unstratified (Fig. 5). Debris flows can be confined to channels or spread out laterally as sheets or lobes over the fan.

253 Fig. 4 Location of Yucca Flat at the Nevada Test Site. 254 Fig. 5 Debris Flow deposits a) 1.5 m thick boulder bed in L)2co. The deposit is clast-supported and may represent a reworked debris flow.

b) Thin debris flow (arrows identify upper and lower contacts) interbedded with fine-grained sediments at 326m in U2ff. Note the coarse clasts (b) at the top of the unit, suggestive of inverse grading.

c) Thin debris flow at 303m in U2ff. Arrows show top and basal contacts. 255 U2co

MAXIMUM CLASTSIZE 0 100 200cm i I i

20-

40-

80- Fig. 6 Stratigraphic column of alluvium in U2co, The width of each stratigraphic unit 100- designates the sediment type, as shown at the base of the column. The mixed alluvium/basal 120- tuffaceous alluvium contact is at 157 m. O • O O C • o • p-o"

160-

I6ir

200-

Sand — Gravelly Sand-' Sandy Grovel- Grovel

256 ALLUVIAL STRATIGRAPHY Several studies have been done on the variation of grain size on alluvial fans (Bluck, 1964; Denny, 1965; Lustig, 1965; van Wie, 1976). In general, alluvial fans can be divided into 3 segments, called proximal, mid-fan, and distal facies (Fig. 2). The proximal facies contains the coarsest sediments and is deposited in the upper segment of the fan, where the main stream emerges from the range front. The mid-fan facies occurs in the central area of the fan, and consists of interbedded coarse gravels and gravelly sand. The distal facies occurs in the lower part of the fan and contains the finest-grained sediments. This facies typically coalesces with the fine-grained deposits of the central alluvial plain (Fig. 2). Two basic alluvial units have been identified in the Yucca Flat basin. These have been termed mixed alluvium and basal tuffaceous alluvium (Wagoner et a!., 1982). Each unit is defined by the source rocks from which it is derived. Mixed alluvium is derived from both Paleozoic sedimentary rocks and Tertiary volcanic tuff. The basal tuffacecas alluvium was derived almost exclusively from-the volcanic sources, and is generally a much finer-grained deposit than tlie younger mixed alluvium. The contact between the mixed alluvium and basal unit is generally a sharp erosional contact and represents the time of unroofing of the Paleozoic rocks by erosion of the Tertiary tuff, which blanketed most of pre-Yucca Flat topography.

Fig. 7 Contact (arrows) of the mixed alluvium (m) and basal tuffaceous alluvium (t) in U2co.

The stratigraphic sections for the five holes are shown in Figs. 6,8,10,11,12. The hole, L)2co, is the most proximally located on the Grouse Canyon fan (Fig. 2). As expected, this section contains the coarsest mixed alluvium, with 83% gravel and sandy gravel and 17% gravelly sand and sand. The measured maximum clast size ranges up to 1 m diameter (Fig. 6). There is a sharp erosional contact between the mixed and basal tuffaceous alluvium. The basal tuffaceous unit has a characteristic light coloration and is mainly fine- to coarse-grained sand and gravelly sand (Fig. 7). 257 U2cp

MAXIMUM CLAST SIZE 0 100 200 300cm I I ' '

20"

40-

60-

Fig. 8 Stratigraphic column of alluvium in U2cp. The mixed/basal tuffaceous alluvium contact is 163 m.

120-

140-

180"

200"

Sand ' Gravelly Sand-J Sandy Grovel Gravel 258 Hole U2cp is located 440 m southeast of U2co and contains only 4.?% gravel and sandy gravel in the mixed alluvium. The maximum clast size ranges up to 1.75 m diameter (Fig. 8). The basal tuffaceous unit is basically the same as at U2co, except that it contains some very large ( 2 m) landslide(?) boulders (Fig. 9).

Fig. 9 Large landslide(?) boulder in the basal tuffaceous alluvium in U2cp.

Hole U2cq is located 228 m southeast of U2cp and contains 48% gravel and sandy gravel and 52% gravelly sand and sand in the mixed alluvium. The maximum clast size ranges up to 1 m diameter (Fig. 10). The stratigraphic section at hole U2ff is located near the distal edge of the Grouse Canyon fan. The mixed alluvium consists of 40% gravel ana sandy gravel and 60% gravelly sand and sand. The sediments from 0-238 m are mainly sand and gravelly sand; the section below 238 m is dominantly gravel and sandy gravel. The maximum clast size ranges up to 70 cm diameter (Fig. 11). The hole terminates into 9 m of basal tuffaceous alluvium. Hole U4aj is located 120 m off the distal edge of the fan, on the central alluv'al plain of Yucca Flat, 2.2 km southeast of U2ff (Fig. 2). The mixed alluvium contains 21% gravel and sandy gravel and 79% gravelly sand and sand. This is the finest-grained section in the study area. The maximum clast size ranges up to 50 cm diameter. The mixed alluvium becomes coarser grained toward the base of the section (Fig. 12).

259 U2c£ MAXIMUM CLAST SIZE 0 50 100c00cm i

40-

Fig. 10 Strati graphic section oif alluvium at U2cq. The mixed/basal tuffaceous alluvium contact is 146 m.

120 —

140

180-

Sand !J Gravelly Sand- Sandy Grovel Gravel 260 Fig. 11 Stratigraphic section of alluvium at U2ff. The mixed/basal U2ff tuffaceous alluvium contact is 350 m. MAXIMUM CLAST SIZE (m) 0 100cm 0-

200- 20-

220-

240 60

260 80 —

280 100

300 120

320" 140-

340- 160-

360 180- Sond Grovelly Sond Sandy Gravel ' Grovel —

261 U4aj

MAXiMUM CLAST SIZE 0 50 100cm I i I

200

20—

220-

40 —----.

240-

60—=-

260-

280-

100— •--=

300-

120 Sond ' Gravelly Sand Sandy Gravel- Gravel

140-

160- Fig. 12 Strati graphic section of alluvium in U4aj. The mixed/basal tuffaceous alluvium 180- contact is 266 m.

262 STRUCTURAL DEVELOPMENT OF THE BASIN AND ALLUVIAL STRATIGRAPHY Alluvial fan sedimentation is directly influenced by tectonic movements in the basin and surrounding ranges. As a result of these movements, alluvial fans undergo periods of growth and erosion, and since fan sedimentation occurs adjacent to the source area, changes in base level and subsequent flow capacity should be reflected in the sediments. The area! relationship of the five stratigraphic sections and correlations of generalized stratigraphic units are shown in Figure 13. From west to east, the cross-section shows a gently sloping basin margin, a deep faulted western basin, a large structural block (horst), and a deep faulted eastern basin. Alluvial deposits in the western basin are 122-183 m thick, whereas east of the horst, the alluvium is greater than 366 m thick. The Tertiary volcanic deposits are greater than 300 m thick west of the ridge; to the east, the tuff thickens to greater than 500 m. There is probably \jery little tuff present on top of the horst. Stratigraphic thicknesses of the Tertiary volcanics in the Yucca Flat basin suggest that significant topography (probably due to faulting) was present on the pre-Tertiary surface. It is possible that the structural ridge was a topographic feature during deposition of the Tertiary tuffs and that significant thinning of the volcanics occurred over this high. As structural uplift of the western basin margin and horst accelerated, the relatively soft tuff deposits were rapidly eroded and transported to both basins. The resulting deposit is represented by the basal tuffaceous alluvium. The Paleozoic clastic rocks of the western basin margin and the Paleozoic carbonate rocks of the horst were eventually unroofed by erosion, creating new sources of Paleozoic detritus for the alluvium. At this point, the character of the alluvium changed from the highly tuffaceous basal unit, to the mixed alluvium. Alluvial sedimentation along the western basin margin east of U2co was certainly dominated by coarse proximal and mid-fan sediments (Fig. 13). The U2co stratigraphic section is 80% gravel and sandy gravel. There is a general decrease in grain size downfan from U2co to U2cp and U2cq, indicating that most of the sediment was derived f^om the Grouse Canyon area and not the uplifted horst. Assuming that the basal tuffaceous alluvium in both basins is approximately time-strati graphic, then the oldest mixed alluvium deposits should also be approximately the same age. At U2ff and U4aj, the oldest mixed alluvium should be represented by fine-grained distal Grouse Canyon fan sediments. In fact, they are coarse gravels derived from a nearby source (Fig. 11,12,14). This suggests the horst was actively uplifting, shedding coarse Paleozoic carbonate detritus into the eastern basin. The horst was a topographic high effectively dividing Yucca Flat into two subordinate basins.

263 N 262000

tN261,000 [23 Tuff

I jPalwtolc S«dlm«nlary Roclo

N260.OO0 f

Fig. 13 Geologic cross section along the Grc oe Canyon fan. The alluvial units are generalized from the detailed stratigraphic columns for each hole. Vertical scale = horizontal scale. Fig. 14 Major alluvial facies at U2ff and U4aj. a) Fine-grained distal sands at 104 m in U2ff b) Coarse-grained proximal sandy gravels at 320 m in U2ff c) Fine-grained distal sands at 77 m in U4aj d) Coarse-grained proximal sandy gravels at 206 m in U4aj.

265 At U2ff and U4aj, the major change in sedimentation occurred at 237 m and 197 m, respectively (Fig. 11,12). The sediments changed from coarse gravel to gravelly sand and sand. This represents a change from proximal and mid-fan deposition to distal deposition (Fig. 14). Paleozoic carbonate rock fragments derived from the horst became less common. Uplift of the horst slowed down, until the rate of alluvial sedimentation from the western basin margin eventually exceeded the rate of uplift. The Grouse Canyon fan and other fans eventually buried the horst (Fig. 15).

SUMMARY Using downhole photography, alluvial sediments are described in 5 emplacement holes in northwestern Yucca Flat. The holes are located on or near the Grouse Canyon fan. The 3 most proximally located holes: contain the coarsest sediments and display a general decrease in grain size in the downfan direction. The 2 most distally located holes contain fine-grained distal facies sediment in the upper parts of the holes and coarse-grained proximal facies gravels lower in the holes. The proximal gravels in the lower half of the sections were derived from the gravity high, a north-south-trending horst which was exposed early during the history of Yucca Flat basin. Alluvial sedimentation eventually exceeded uplift of the horst, which was buried by distal facies sediments, derived from the western basin margin.

REFERENCES

1. Bluck, 3. J., 1964, Sedimentation of an alluvial fan in southern Nevada; Journ. Sed. Petrology, V. 34:395-400. 2. Brugman, V. P., 1979, Downhole fisheye motion picture camera: Lawrence Livermore National Laboratory, Livermore, California, UCRL-82534. 3. Denny, C. S., 1965, Alluvial fans in the Death Valley Region, California and Nevada: U.S. Geological Survey, Professional Paper 466: pp. 1-62. 4. Lustig, L. K., 1965, Clastic sedimentation in Deep Springs Valley, California: U.S. Geological Survey, Professional Paper 352-F: pp. 131-192. 5. Van Wie, W. A., 1976, Sedimentology of aluvial fans in southern Nevada (Ph.D. thesis): University of Cincinnati, Cincinnati, Ohio, 249 p. 6. Wagoner, J. L., McKague, H. L., Howard, N. W., 1982, Variation of Physical properties of alluvium in an arid basin: reflection of source and transport: Geol. Soc. Amer. Abs. with Programs, V. 14, #7, p. 640.

266 b

d Fig. 15 Structural development of northwestern Yucca Flat basin. a) During the Tertiary, volcanic tuff was deposited on an irregular erosional surface of Paleozoic sedimentary rocks. The gravity high may have been a gentle topographic high, over which the tuff deposits were somewhat thinner. b) In the Late Tertiary, structural development of the Yucca Flat basin began, with uplift of the horst and Eleana Range and corresponding subsidence of the western basin and eastern basin. Volcanic tuff was eroded off the uplifted areas and highly tuffaceous detritus was deposited into both basins. The resulting deposit was the basal tuffaceous alluvium. c) The tuff deposits were eventually eroded off the uplifted areas, unroofing the Pa'eozic rocks, which were then eroded. This marked the change from basal tuffaceous alluvium to mixed alluvium. Coarse proximal gravels were deposited along the flanks of the gravity high and along the western basin margin. d) The rate of alluvial sedimentation from the western basin margin eventually exceeded uplift of the horst. Alluvial deposits of the Grouse Canyon fan effectively buried the Paleozoic rocks of the horst, and the proximal sedimentation along the flanks of the horst changed to finer-grained distal alluvial deposition.

268 DISCLAIMER - This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the University of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government thereof, and shall not be used for advertising or product endorsement purposes.

269 1X10 LA-UR -83-2281

lot •limoi Neiiongi UboraiO'r ft Dpt'iltd 6y trie Uni»«i»ity ol Cti'fomn tot mo United tlltll Popinmem pi tnttfj untfr contract W-74QSENG

TITLE GEDLDGIC INVESTIGATIONS DF DRILL HOLE SLOUGHING PROBLEMS, NEVADA TEST SITE

AUTMOR(S) 5. L. Drellack, Jr., F&S hi. J. Davies, F&S 3. L. Gonzales, F&S k). L. Hawkins, LDS Alamos

SUBMITTED TO 2nd Containment Symposium Kirtland AFB Albuquerque, MM August 2-k, 1983

ty *cc«punc« of trin (tidi th» publiihc racognltti ttt«i lh» U S 6ovtmm«ni riltini • nonaiclunt*. reytHr-(ro* hevnM to pubi>«A or f«p'oduc» tfi* pubinxtd form pf tin contribution, or (o silo* ethri to do to. (or U.S Oovtmrrwnt purpoits TIM Lot tlimoi K»non«i utomlory raqmtti toil lh« pubN»n»t toanufy thil irticH u work performed undc tf»t tvtplCM of the U.t Oaptimtnt o< Encg,

Los Alamos National Laboratory Los Alamos,New Mexico 87545

271 HO GEOLOGIC INVESTIGATIONS OF DRILL HOLE SLOUGHING PROBLEMS, NEVADA TEST SITE

S. L. Drellack, Jr. , W. J. Davies , J, L. Gonzales

and

W. L. Hawkins2

Severe sloughing zones encountered while drilling large diameter emplacement holes in Yucca Flat, Nevada Test Site, have been identified, correlated and predicted through detailed geologic investigations. In central and southeastern Area 7 and in northern Area 3, the unstable zones are a very fine-grained, well-sorted, unconsolidated sand deposit, probably eolian in orgin, which will readily flow into large diameter drill holes. Other areas exhibit hole erosion related to poor induration or extensive zeolitization of the Tertiary tuff units which are very friable and porous. By examining drill hole samples, geophysical logs, caliper logs and drilling histories, these problem zones can be characterized, correlated and then projected into nearby sites. Maps have been generated to show the depth, thickness and areal extent of these strata. In some cases, they are local and have a lenticular geometry, while in others they are quite extensive. The ability to predict such features can enhance the quality of the hole construction and completion operations to avoid costly delays and the loss of valuable testing real estate. The control of hole enlargements will also eliminate related containment concerns, such as stemming uncertainties.

Venix and Scisson, Inc, Los Alamos National Laboratory

272 INTRODUCTION

This report addresses the hole stability problems encountered while drilling large diameter emplacement holes in southeast Area 7 (Figure 1); however, problems have also been experienced in other parts of Yucca Flat which are pertinent to this specific area (In Pocket). The available data was studied and synthesized into a model which may be used to predict expensive and potentially dangerous hole stability problems. Use of this model may allow preventive measures to be taken to avoid costly disruptions in drilling schedules. Unfortunately, the ability to accurately predict problem areas is limited by the quantity and quality of available data. Also included in this report are several recommendations which may be applicable depending on geologic conditions and hole specifications.

This report is a follow-up to the DOE/NTSSO conference held on January 14, 1982 shortly after U7bp developed severe sloughing problems.

Hole enlargements within recent memory which may be confused with this particular problem occurred within a strongly zeolitized section of the Paintbrush tuff. This problem will not be addressed in this report except to say that it appears to be controllable with proper treatment of the drilling fluid.

HISTORY

Los Alamos experienced several hole stability problems in recent years which resulted in the abandonment of emplacement holes along with various amounts of hardware. The most recent were the U7bc - U7bp experiences. LLNL has also experienced severe sloughing problems in their use areas. This problem, along with suggested solutions, was addressed in a LLNL interdepartmental memo from A. C. Douglas dated March 20, 1979.

GEOLOGY

Yucca Flat is a typical north-south trending basin and range grah\">. Following the deposition of Tertiary volcanic rocks the valley is being filled by coalescing alluvial fans formed from sediments derived frorr the surrounding up- faulted mountains. These alluvial fans include deposits from debris flows, sheetwash, and braided streams. Other oediments which contribute to the alluvium section are eolian sands (sedimentary deposits which are transported and laid down

273 LEGEND AREA •OUNDARY "0*0 OlRT ROAD-

Figure 1. -- Index map of the Nevada Test Site showing the area of investigation.

274 to wind), and playas or lacustrine deposits. These various deposits are interfingered and interbedded to such an extent that their locations are extremely difficult to predict.

In southeast Area 7, the problem sloughing zone is a very fine grained, well sorted (poorly graded), unconsolidated sand deposit (Figure 2). Grain size, degree of sorting, and bed geometry suggest that it is eolian. It seems to be associated with, or just above, a thin, fine grained, sandy alluvium rich in mafic minerals, especially magnetite (Figure 8). This mafic-rich stratum can be identified by magnetometer and natural gamma ray well logs. It must be emphasized, however, that this association of the sand zone with a detectable mafic-rich zone has only been recognized in a few holes around U7bp - U7bc. It therefore cannot be relied on to identify or predict troublesome sand zone outside this local neighborhood. These nonindurated eolian sands also have a high primary permeability. This combination of properties frequently enables this type of deposit to behave almost like a fluid.

METHODOLOGY

The data search involved the examination of samples, geophysical logs, and hole histories from roughly 130 drill holes in and adjacent to area? which historically have had sloughing problems (Table 1). For emphasis on drill holes in southeast Area 7 see Table 2. Data through the alluvium is incomplete because, in the past, geophysical logs were used sparingly and the alluvium was effectively ignored by sampling programs.

Flowline cuttings samples are usually available. These, however, have limited use because of grain siz3 biasing introduced during the drilling process and the sample preparation (washing). In cutting samples, the fine and very fine sand size fraction is often lost in the sampling and sample handling operations. Sidewall sampling of these intervals is virtually impossible because of hole enlargement that occurs in these zones.

The best source of data available for identifying a sloughing zone is the caliper log. These logs were examined and the large washouts noted. Washout "signatures" were then evaluated for correlatability with the caliper signatures of known problem areas. Data points were then plotted using various presentations in order to synthesize a model which might be used as a predictive tool.

275 r^tx-i "S*

Surface sample from the flank of Busted Flow line sample from sloughing zone in 8utte west of Jackass Flat, NTS. abandoned emplacement hole U7bt (^-240 m)

* "

A f 4 '/ % S" Jf hi V H

Sample from "flowing" sand zone encoun- Example of localized eoiian deposit having tered during mining operations in shaft a nearby tuffaceous source. Note partially U3ev #3 (^43 m) rounded glass shards.

Fig. 2 -- Photographs of eolian sands typical of problem sloughing zones magnification). 276 The resultant isopach and structural contour maps depict the subsurface geometry and geographic location of sloughing sand units (Figures 3, 4, 5, and 6). These maps show the approximate positions of the troublesome subsurface units. Their usefulness for prediction varies based on drill hole control and data quality.

CONCLUSIONS

Based on spatial relationships, the model which best fits the alluvium data is one of discrete sand lenses of eolian orgin deposited on the eastern side of Yucca Valley by prevailing westerly winds. The exact position of the lenses appears to be at least partially controlled by the morphology of the tuff surface (Figure 7) as well as the deposition of pre-eolian alluvium (Figure 6).

After examining many caliper logs and hole histories, we observed that the emplacement holes which have had severe sloughing problems in alluvium were those that penetrated sand zones thicker than 25 feet (Figure 5). Holes which penetrated zones thinner than this developed large washouts but generally stayed open. Other factors which may contribute to hole instability are depth to the sand zone, drilling techniques, hole diameter, and natural or induced seismic activity.

RECOMMENDATIONS

The procedures for preventing sloughing problems should be dictated by the predicted severity. Each location will have to be assessed and recommendations made on an individual basis. Several of the procedures discussed at the DOE/NTSSO conference, such as (I) the use of a sand stabilizing additive in the drilling fluid, (2) maintaining a fluid level above the problem zone, and (3) not dewatering the hole until intermediate casing is run to below the problem zone, may be viable cures Lo the hole sloughing problem. In the extreme case, sites which are highly suspect may require exploratory holes in order to localize problem zones so that exotic and expensive drilling fluids and casing may be conserved and the risk to large diameter bottom hole assemblies minimized.

277 Table 1.--Abridged Drill Hole Data Cave Zone Interval Tuff/Alluvium Drill Hole Present (Feet) Contact Remarks U3cn Yes 820 - 910 981 U3cp Yes 215 - 230, 800 262 - 278 U3cq Yes - Hole Abandoned U3cw Yes - Hole Abandoned U3cy Yes 240 - 380 879 U3da 928 Insufficient Data U3dh No 669 U3dl 929 Insufficient Data U3do Yes 276 •- 292, 690 308 - 328 U3dq Yes 258 •- 294 770 U3ei Yes 270 •- 278, 610 296 •- 304, 318 •- 334 U3er Yes 300 •- 305, 630 322 •- 330, 332 •- 336 U3ev Yes 146 •- 186 220 U3ev #3 Shaft Yes 126 •- 186 230 U3fc No 600 U3ff No 161 U3fg No 710 U3fh No 500 U3fi Yes 358 -- 442 - Hole Abandoned U3fj Yes 3 between 410 270 - 330 U3fk 249 Insufficient Data U3fl Yes 302 -• 332 - U3fp - Insufficient Data U3fq Yes 584 -• 592 710 U3fx No 4,30 U3gv No 960 U3gw No - U3gy - Insufficient Data U3gz Yes 232 -• 238 570 U3jq Yes 276 -• 286 fabb U3ke Yes 394 -• 400, - Hole Abandoned 407 - 417 U3kl Yes 758 - 770 - Hole Abandoned U3kt No - U3kx Yes Numerous wash- 900 outs between 320 - 900 U31a No 600 U31c No 675 U31d No 1295 U31k Yes 496 - 514, 1250 1174 - 1200

278 Table l.--Cont.

Cave Zone Interval Tuff/Alluvium Drill Hole Present (Feet) Contact Remarks U31m Yes Numerous washouts 560 - 750

Ue3a 223 Insufficient Data U4j No 762 U4p Yes 214 - 230 630 (small wachout) U7b Yes 278 - 29L 570 U7c 771 Insufficient Data U7d No 794 U7e Yes 302 - 328 886 U7f 942 Insufficient Data U7g 719 Insufficient Data U7h ho 758 U7i Yes 475 - 525 850 U7j Yes 90 - 170, 486 360 - 435 U7m No 351 U7n No 259 U7o Yes 128 - 210, 1060 662 - 742 U7p No 800 U7q 1060 Unable to locate caliper log U7r No 700 U7s No 280 U7t Yes 336 - 364, 620 380 - 402 U7u No 860 U7v No 690 U7w No 230 U7x No 220 U7y Yes 258 - 264, 650 271 - 280, 283 - 292 U7z Yes 150 - 230 - U7aa 151 Insufficient Data U7ab Yes 154 - 156, 310 194 - 208 U7ac Yes 310 .- 350 565 U7ad No 5/b U7ae Yes 366 - 378 400 U7af Yes 120 - 230, 640 420 - 470 U7ah Yes 640 - 656, 865 066 - 680, 712 - 730 U7al Yes 538 - 542, 554 - 564

279 Table 1.—Cont.

Cave Zone Interval Tuff/Alluvium Drill Hole Present (Feet) Contact Remarks

U7am Yes 625 -• 645 630 U7ao No bUi; U7ap No 236 U7aq No 9b5 U7at 259 Insufficient Data U7av No 990 U7aw Yes 283 -• 290; 806 rougfi hole 540 -• 790 U7ax Yes 140 -• 160 2Ub U7ay No 420 U7az - TD 114' U7bc Yes 230 -• 258 415 U7bd Yes 483 -• 495, 695 598 -• 616 U7bf Yes 295 -• 450, 670 575 -• 589 U7bh No 180 U7bm No 303 U7bp Yes 372 - 400 455 U7bx - Insufficient Data Ue 7g r?la No Ue7h No 170 Ue7i No 450 Ue7j No 386 Ue7n 177 Insufficient Data Ue7aa 264 Insufficient Data Ue7ax Insufficient Data Ue7az 740 Insufficient Data Ue7bc Yes Lost Circulation E-Log shows low at 330 resistivity 300 - 330' Ue7bh No (6) 60 Ue7bi 165 No logs, poor cuttings Ue7ns No 187 U7ajs Yes 300 - 31C U9ae Yes 422 - 436, 456 - 486, 872 - 904 U9bd No 230 U9be 479 Insufficient Data U9bf No 348 Insufficient Data U9bg Yes 120 - 135 669 U9bj No 308 U9bk No U9bp No 692 U9bt No 308 U9cl No 220 U9cm No 141 U9cn No U9cq No

280 Table l.--Cont.

Cave Zone Interval Tuff/Alluvium Dril 1 Hole Present (Feet) Contact Remarks U9cs No U9x 718 Insufficient Data lest Hole #1 No 581 Test Hole #2 Yes 95 114, 587 225 232, 243 248, 312 318, 373 382 Test Hole #3 Yes 8 - 50, 367 96 - 104, 106 - 114, 192 - 201 Test Hole #4 Yes 160 - 172 203 Test Hole #5 No 446 Test Hole #6 No 515 lest Hole #7 No 427 lest Hole #8 No 453 Test Hole #9 Yes 166 - 188 354 80-215 Lost circulation Possibly at 330 _at.._33g_' USGS HTH E Insufficient Data U3ev - 9 Yes 130 - 151 Caliper #1 150 - 350 Caliper #3

281 Table 2.—Southeast Area 7 Drill Hole Data

HOLE COMPLETION DRILLING HOLE SIZE DEPTH GEOLOGY GEOPHYSICAL (STATUS) DATE FLUID AND DEPTH CAVED ZONE(S) TO TUFF NOTES NOTES U7ab 6-29-72 Air and Mud 86" to 328' 154' - 156' 310' Normal alluvium to Neutron log shows (Expended) 329'. Air 52" to 1700' 194' - 208' 290". 290' - 310' no water at foam' to TD is tuffaceous, washout. sandy, finer grained. U7ae 7-29-76 Air and 86" to 671' 119' - 155' 400' 340' cuttings are Caliper only. (Expended) Bentonite 64" to 2260' 366" - 378' sandy, fine grained, mud. 384' - 390' well sorted. 380 -400' more magnetite and horn- blend, fine grained, well sorted. 400' very fine grained, v 1 sorted. ^ U7al 3-27-77 Air and 52" to 1510' 538' - 542' 59T Caliper only. £ (Expended) Bentonite 554" - 564' mud. U7ao 4-14-77 Air and 52" to 1510' none 600' Good log (Expended) Bentonite coverage. mud. Ue7az 7-22-80 Air foam, 6V to 1021' 740" No caliper log (Plugged) polymer, am 3.9" to 2363' in alluvium. bentonite 2.98" to 2441' mud. U7b 5-22-64- ? 64" to 1545' 278' - 298' 570' Caliper log (Expended) only. Ue7bc 9-05-80 Bentonite 6V to 740' none 320' (Plugged) mud to 695'. 3.9" to 2707' Polymer to TD. Table 2.—Cont.

HOLE COMPLETION DRILLING HOLE SIZE | DEPTH GEOLOGY GEOPHYSICAL (STATUS) DATE FLUID AND DEPTH CAVED ZONE(S) TO TUFF NOTES r-'OTES U7bc 11-10-81 Polymer 85" to 2250' 230' - 258' 415' 220 - 240' samples Caved zone (Caved to seems slightly associated with surface, finer grained, gamma ray log stemmed better sorted. response. with gravel) Neutron log shows H2O increase in caved zone. U7bd 6-29-78 Air and 72" to 2000' 488' - 495' 695' Caved zone Caved zones (Expended) light 598' - 616' associated with associated with bentonite mafic rich zone. magnetometer and mud. Poisible caliche gamma ray log cemented cobble responses. bed at base of oo mafic rich zone. Ue7bi 11-03-80 Air foam 6k" to 2144' 1 165' Hole caved (Plugged) before logs .•/ere run. U7bm 12-16-79 Air and 72" to 155O1 Rough hole 330' 200' s/w sample Inconslusi ve. (Expended) mud. 117" - 196" may contain more magnetite. U7bp 3-24-82 Polymer 120" to 520' 372' - 400" 455' Caved zone is very Caliper only. (Expended) 86" to 2250' fine sa.id, well sorted, unconsolidated; mafic rich zone at base. U7m 6-2-66 Air and 86" to 475' Slightly 351' Caliper only. (Expended) Davis Mix. 64" to 1700' enlarged (?) 118' - 210' U7p 1-14-69 Air and 64" to 1900" Rough hole 800' Caliper only. (Expended) Davis Mix. 117' - 450" U7s 10-26-66 Air foam 86" to 721' Rough hole 280' Caliper only. (Expended) 64" to 1900' 120" - 245' Table 2.--Cont.

HOLE COMPLETION DRILLING HOLE SIZE DEPTH GEOLOGY GEOPHYSICAL (STATUSJ DATE FLUID AND DEPTH CAVED ZONE(S) TO TUFF NOTES NOTES U7t 9-04-66 Air foam 86" to 873' 336' - 364' 620' Very fine, Caliper only. (Expended) 64" to 1900' 380' - 402' unconsolidated sand from cave zones. U7w 10-27-67 Air and 86" to 480' Small 240' Caliper cnly. (Expended) Davis Mix 64" to 1750' enlargement 120' - 154' U7y 1-03-70 Air foam 86" to 670' 258' - 264' 650' Poor samples. Caliper only. (Expended) 64" to 1900' 271' - 280' 283' - 292" U3do 6-30-64 Air and 64" to 1215" 276' - 292' 690" Caliper only. £ (Expended) Davis Mix 308' - 328' U3ei 4-10-70 Air and 64" to 1630" 270" - 278' 610' Caliper only. (Expended) Davis Mix 296" - 304' 318' - 334' U3jq 4-13-71 Air and mud. 64" to 1900' 276' - 286' 656' Caliper only. (Expended) i LEGEND • Drill hole, data point / Approximate edge of / sloughing sand zone Thickness in feet Contour interval 50 feet

U7r

U7m

AREA 7 AREA 3 ngure3.--ISOPACH MAP OF ALLUVIUM ABOVE THE SLOUGHING SAND ZONE 1000 0 I0CO 2000

GRAPHIC SCALE IN FEET

285 LEGEND • Drill hole, data point / / Approximate edge of > sloughing sand zone Elevations in feet above mean sea level Contour interval 50 feet

U7m

U7w

AREA 7 AREA 3 Rgure4.--STRUCTURE CONTOUR MAP ON TOP OF THE SLOUGHING SAND ZONE 1000 0 1000 2000 GRAPHIC SCALE IN FEET

286 -LEGEND- • Drill hole, data point / Approximate edge of / sloughing sand zone Thickness in feet Contour interval 5 feet

U7r

U7s

U7m

U7w U7bm

\ AREA 7 Figures.-- AREA 3 ISOPACH MAP OF THE SLOUGHING SAND ZONE IOOO 0 1000 2000

GRAPHIC SCALE IN FEET

287 •LEGEND- • Drill hole, data point / Approximate edge of ' sloughing sand zone Thickness in feet Contour interval 50 feet

U7r

U7s

U7m

U7w U7bm

\\\ AREA 7 AREA 3 Figure 6.--IS0PACH MAP OF ALLUVIUM BELOW THE SLOUGHING SAND ZONE 1000 0 1000 2000

GRAPHIC SCALE IN FEET 288 AREA 3 Figure 7.--ALLUVIUM ISOPACH MAP IOOO O 1000 2000

GRAPHIC SCALE IN FEET Contour Interval 50 feet 289 Ground Leve!

"Normol" Tuffaceous Alluvium

372'

Sonct~ ^ery fine $

400* Sandy Alluvium -5 to 10% magnetile, 410' unconsolidoted, fair sorting

Sandy Alluvium-magnetite decreasing

v t A -, 4 W •J > A *» i- V " i. A A -* A 7 r v . t(V '7 J '* «J 5 J>-1 t •» •^ A 4 'A -1 A A V J 1 * ^ *• •»

* *• J > k c *- * k-* k *- ^ < V A V A >L" >v >r V'-\/ >/ 1 V > < A V ^ v » 4. A A A l> J 1 -Rainier Mesa Member, Timber ^*V < 1 > W -I <" 1 < 7 1 •* £. 1 * < ^ I. 7 r v J V Mountain Tuffs-Ash-flow tuff V/ 1 <^ ^ V LA*

V •»< » A * *- > A ' ^ "7 ? 7 v r i. *- r V -1 ' v A 1 i_ ' > < > . •* > A A A A U A ^ f > A f > VT r' y «- A < <- * „! ^ ' 7 > ^ *1 ^ A \, -j > U t ; -> " A . -1 < •• „ ' A **'".! A k 7 1* M 1 J A v „ r ' w < J* C v A i. V ^ > A < ' r r A A 4 A ^ r V ~> >•>

Figure 8.-- Schematic stratigraphy through the alluvium of emplacement hole U7bp.

290 i imp v \ U9cl "J? \ tfibi U9Bd « • EXPLANATION

* Drill holes with duti jement on cfllipec tog

on coiipei log X Dull holes lost due to cav;ng probems

\ | AcriQi e*tent o( delmetJ sand \en

["_^AenQl e»tent of sontl tenses needing

Roods

N850.000-

TH7 AUSGS TH9 yU3ev-3 shotl XU3C* )

N830.000 -

Map of Sloughing Sand Zones.Within the Alluvium in LANL Use Areas, Yucca Flat, NTS

291 UCRL-89416 PREPRINT

LLNL Site Selection Procedures

C. W. 01 sen

This paper was prepared for submittal to the Second Containment Symposium, Albuquerque, New Mexico, August 2-4, 1983.

July, 1983

This is a preprint of a paper intended for publication in a journal or proceedings. Since changes may be made before publication, this preprint is made available with the understanding that it will not be cited or reproduced without the permission of the author.

293 LLNL SITE SELECTION PROCEDURES OLSEN, C.W., University of California, Lawrence Livermore National Laboratory Livermore., CA 94550 The LLNL has established a cooperative effort between the Containment and Field Operations Programs for improved site selection. Each proposed site is carefully reviewed for projected geologic structure and medium properties, nearby holes, programmatic need for holes of a given depth, possible drilling problems, and event schedules. By using our data bank and new information (such as seismic surveys), we avoid drilling useless, marginal, or expensive to use holes,, and have managed to establish a stockpile, allowing better use of each site by having an available hole to fit a new event. The containment evaluation of the sites can, therefore, be done in a much more efficient manner. We will discuss our criteria for site selection, our use of NTS real estate, and our plans for development of underutilized areas. The differences between selecting a new drill site and selecting a W. P. location in an existing drill hole are also discussed.

Work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract W-7405-Eng-48.

294 There are three subsets of site selection, namely: selection of an existing stockpile hole for a specific event, selection of a new drill site for a stockpile hole, and selection of a new drill site for a specific event. The same information is, in general, used for all three, but the sequence and, sometimes, the quality of data at a given time are different. Factors considered are (1) scheduling of drill rigs, (2) scheduling of site preparation (dirt work, auger hole, surface casing, cementing), (3) schedule of event (when are drill hole data needed?), (4) depth range of proposed W.P., (5) geologic structure (faults, Pz contact, etc.), (6) stratigraphy (alluvium, location of Grouse Canyon Tuff, etc.), (7) material properties (particulary montmorillonite and CO? content), (8) water table depth, (9) potential drilling problems (caving), (10) adjacent collapse craters and chimneys, (11) adjacent expended but uncollapsed sites, (12) adjacent post-shot or other small diameter holes, (13) adjacent stockpile emplacement holes, (14) adjacent planned events (including LANL), (15) projected needs of Test Program for various DOB's and operational separations, and (16) optimal use of NTS real estate. The most straightforward of the three classes of site selection, although not necessarily easy, is choosing a stockpile hole for an event. There are available firm data (although possibly not yet complete) from the hole or holes under consideration. The depth of the hole is known so that a suitable W.P. can be chosen to avoid wasting drill hole, which costs several hundred dollars per foot. The stratigraphic contacts are known, significant faulting in the hole has been photographed, there is at least some indication of hole stabilty, correlation with nearby drill holes can be investigated, geophysical logs can be reviewed for quality of data, and problem areas such as horizons with high CO;? contents are known. The problem is basically one of fitting the event, with a given yield, to the site, taking into account operational as well as containment issues. It is LLNL policy to have firm data from the hole under consideration for at least a maximum credible yield melt radius below the proposed W.P. It is also policy to have good control, either from the hole or adjacent holes, to at least one cavity radius below the W.P. for both structure and material properties. For containment, the preferred W.P. would be in a "normal", low C0£ content medium with horizontal stratigraphic bedding across which only minor material properties changes occur. The material above the W.P., the overburden, has essentially the same wish list although the CO2 may be higher and the individual properties may be different from the W.P. region. All of these factors are basically subjective and, although there is general agreement on what is bad and what is acceptable, the large transition region is ill-defined and often not well understood. The main areas, material properties, CO2 content, and structure will be discussed separately, but there is clearly overlap within these areas and also with operational considerations.

295 The question of W.P. material properties is difficult for several reasons. First of all, there is no feasible way to determine within the entire region of interest (say within two to three cavity radii of the W.P. j_n all directions) even the properties we know how to measure. The only routine technique which can penetrate more than a few; metres beyond the hole wall is the gravimetric density measurement. Second is the fact that no measurable "containment parameter" has been discovered (it probably does not exist); strength parameters, which may be the closest thing, are extremely difficult to determine in situ. Third, the interaction with the overburden and structure beyond the immediate cavity region can be complex and important; in other words, the region of interest is larger than the 2-3 cavity radii just mentioned, but the features of prime interest may be different. There have been two basic approaches to determining what W.P. materials are acceptable. The main one is simply to see if they are similar to prior successful W.P. regions. This is done by comparing measurable (or calculable) properties such as bulk density, grain density, water content, saturation, porosity, and sonic velocity with the same parameters from other sites. The aim is not so much to show that the site is good as it is to show that the site is not anomalous. The other approach is to develop material models and codes, and, using properties for a given site, calculate the behavior. The latter approach is very time consuming and not generally warranted for most events. It is not clear that the capability even exists to truly calculate the phenomena resulting from an underground detonation. One property of considerable significance is saturation or, its companion, gas-filled porosity. It has been found both experimentally and calculationally that shock coupling-efficiency is very dependent on gas-filled porosity. Zero gas-filled porosity results in coupling roughly five times that of a material with > ^ 5% gas-filled porosity. Rainier Mesa tunnel events, virtually saturated, use the enhanced ground shock for LOS pipe closure. On Pahute Mesa or Yucca Flat, events with W.P.'s below the static water level (hence expected to be saturated) have all behaved well from a containment standpoint, probably because of the great depth of the water table (> 520 m). The spectacular counter-example is BANEBERRY, which had an unexpectedly saturated W.P. at 278 m depth. Ground motion, both surface motion at the site and seismic signals, were equivalent to several times the actual 10 kt yield. Saturation is also of operational significance, since wet holes require special hardware and cable treatment. The LLNL has quit doing events below the SWL in Yucca Flat, by doing larger yield events on Pahute Mesa, thereby avoiding hole caving problems common below the SWL in Yucca Flat, and reducing operational problems resulting from large ground motion at shot-time. The grain density is used in calculating porosity and saturation and is also used for lithologic determinations. High grain density can, for example, indicate the presence, of clay. Bulk density is also used as input for the calculated parameters. Bulk density determines the vapor and melt radii, since it determines the mass near the W.P., and the bulk-density integrated over the DOB determines the overburden pressure. Bulk-density along with sonic velocity determines the acoustic impedance of the material, important in determining shock response.

296 Although there is no routine measurement of strength, strength related parameters, particularly shear strength, are clearly of considerable importance. A method of inferring shear strength based on measured cavity radii has been developed^'>, but it obviously is not applicable for site selection unless one assumes some areal strength distribution. The most suspected contributor to weak material is montmorillonite clay because of its association with BANEBERRY. Montmorillonite not only holds considerable amounts of water (increasing saturation) it is weak when wet. It is not known how much montmorillonite is too much in any given case. The large zone around the BANEBERRY W.P. of 50-90% montmorillonite was too much. Generally a sample of < 20% montmorillonite in an otherwise competent matrix is considered to be of no concern. The 20% level was that chosen by the CEP to be the threshold for reporting the presence of montmorillonite. Large amounts of montmorillonite can also lead to drilling problems. For these reasons, areas with known montmorillonite contents of more than ^ 30% in an appreciable depth range are avoided.

The question of CO? content is another where acceptable and unacceptable are separated by a transition region. Experience^) indicates CO? contents greater than 7% pose a distinct containment risk, presumably by generating large amounts of non-condensible gas which causes flow away from the cavity or chimney region. Experience also indicates that CO2 contents up to ^ 5% are safe, although concern generally sets in at levels of about 3.5 - 4.0%. The typical release ascribed to CO? is relatively late starting and lasts for an extended period. The classic CO2 release was the NASH event (365 m DOB in U2ce on 19 January 1967), which was fired in dolomite (45.8 wt% CO?). Collapse occurred at 23.9 minutes after zero-time. At just less than 10 hours post-shot seepage was detected in the crater area. The release continued for over 40 hours, with a total release of 6.9 x 10^ Ci . The HANDCAR event (403 m DOB UlOb on 5 November 1964) was also fired in dolomite (47.3 wt% CO2); in this case there was no collapse. There was a small early release attributed to cables and another release starting at ^ 16 hours post-shot which lasted ^ 8 hours, with a relase of 57 Ci. Strangely, there were two other events KANKAKEE .(455 m DOB in UlOp on 15 June 1966) and BOURBON (560 m DOB in U7n on 20 January 1967) fired in dolomite which did not have detectable releases. The reason for the latter two events being contained is not really understood., though they were deeper and of somewhat higher yield than the two which leaked. On the other hand; the amounts of CO2 generated were considerably higher than on HANDCAR, based on yield.

The effect of structure on containment, and hence on site selection, is also complex. Several features are considered to be bad, but there are always mitigating circumstances. A general rule in siting a oil! hole is that no W.P. shall be closer than two cavity radii to the Paleozoic rocks, but the actual distance can be decreased almost to one radius • \'~ the Pz contact is well determined and the surface is essentially flat; LAM used the same argument while presenting ATRISCO to the CEP. One rec-io-i to avoid some Paleozoic rocks is the CO? problem already discussed, but, for that reason alone, one could probably approach to within one cavity -adius or even somewhat closer. The more common reason is to avoid enhanced ground motion caused by large shock reflections from the high impedance surface. There are

297 two aspects of the reflection problem also; first is surface motion, second is motion and stress near the cavity. Surface motion can be reduced by an • increased SDOB regardless of proximity to a reflector, so that aspect is yield dependent. The problem of stress near the cavity is more complex. A large contrast in material properties or in acoustic impedance anywhere near the cavity can alter normal phenomena. A marked drop in impedance will cause a rarefaction to return to the cavity region and can have a significant effect on the residual stress. By influencing the details of event phenomena, simple layering may affect W.P. selection in a given hole, but rarely does it make a site unacceptable. The presence of Paleozoic rocks, especially if the configuration is more complex than a simple layer-cake, can have strong influence on the resulting behavior, generally requiring 2-D calculations to assess.

The other major structure problem is faulting. Small, tight faults are agreed to be of no concern, a feeling confirmed by experience. Major faults with zones of fault gouge, marked contrasts in material properties across the fault plane, and uncertain stress changes near and across the fault are generally believed to be bad. A fault could be close enough to a W.P. to be a potential weakness either mechanically or for gas flow, it could conceivably alter the stress field such that anomalous phenomena occur, or it could alter collapse phenomena, by providing a plane of weakness in the chimney region, for example.

The real influence of faults is not clear. While there have been releases attributed to faults, e.g., BANEBERRY and PINSTRIPE, there are no releases blamed solely on the presence of a fault; there have always been additional factors involved. Since there is no agreement on when a fault becomes "major", the process of site selection tends toward the conservative side. If one has an existing hole, any fault cutting the hole can be examined photographically and by logging. When siting a new drill hole we stay a hole depth away, both at depth and on the surface, from major, named faults. It is possible that this requirement can be relaxed somewhat, especially in the down-thrown block. The subject of nearby holes is treated by rules-of-thumb which, while generally reasonable, have some demonstrable flaws. The rules used are: (1) no closer than one-half DOB to a collapsed site (hole-to-hole separation), and, (2) no closer than one DOB to an expended, uncollapsed site (hole-to-ho'le separation; zero yield events not counted). Limited, although real, experience indicates these separations are conservative; there has never been radioactivity detected on the surface at an adjacent site, and only rarely at depth. The rule-of-thumb that if the path exceeds one DOB it is of no concern, because straight up would be preferred, is also applied to small diameter, open holes (post-shot and exploratory holes), but it is used as a criterion for careful review, not as a cut-off for what is acceptable. Since an open hole to the surface at a distance of one-half DOB is undoubtedly of less concern than an open hole going from the W.P. to one-half DOB, even though

29& both have the same impeded path, the guidelines must, obviously, be applied on a case-by-case basis. The LANL uses 7.6 cavity radii as the minimum safe impeded path, but this, too, must be considered in light of the nature of the open part of the path. The LLNL feels that if most or all of the impeded path is between the W.P. and the open hole that about one-half DOB (roughly five cavity radii) is ample; there is successful experience with such distances. In light of the "containment cage" concept, one could presumably approach with an open hole to just the outside of the residual stress field, say three cavity radii, and be safe from a major failure.

The process of selecting a new drill site is essentially as follows. The programmatic need for hole of a given depth, based on anticipated /ields of future events, is compared with existing holes in stockpile. Taking account of operational considerations, such as separation from other sites, power lines, craters, etc., as well as containment concerns, a site is proposed by the Field Operations representative in the LLNL-N drilling group. The site is chosen to meet the following guidelines: (1) no closer than a hole depth to a major fault, (2) no closer than one-half a DOB to a collapsed event and no closer than one DOB to an uncollapsed event (these guidelines are often shaded), (3) Paleozoic rocks at least two cavity radii below a proposed W.P., (4) no nearby small diameter holes (which cannot be plugged) providing a minimum path from than W.P. to the surface of less the 6 cavity radii, (5) projected medium properties acceptable (CO2, clay, etc.), (6) and no major drilling problems anticipated.

When a site is chosen, a check list is filled out (Figure 1) and various information is collected. Data related to structure is assembled from existing drill holes, seismic survey lines, and surface gravity data, along with any other data available. A search of the Data Bank is made to determine where existing drill holes are with respect to the new site. The information is then forwarded to the Containment Program in Livermore. Containment personnel (geologists, geophysicists, and containment scientists) review the data and assemble other relevant information such as projected medium properties at various horizons and containment experience in the area. Every effort is made to ensure that no surprises will be found when the hole is drilled, thereby avoiding extra studies or possibly even an unusable hole. When review has shown that satisfactory containment is expected at the site, it is formally accepted by the Containment Program, and the new drill hole is put on the drilling schedule. In some cases special requests are made when it is deemed necessary or cost effective. Examples are drilling to the Paleozoic contact or obtaining rapid CO2 analyses of cuttings, while the rig is still at the hole, to assure that a usable W.P. region is reached without unnecessary overdrilling.

In summary, the LLNL Site Selection procedures, by carefully considering the needs of containment and the test program, as well as operational concerns, have permitted us to attain a good stockpile of usable holes, and to enjoy the accompanying flexibility.

299 References

1. R. W. Terhune and H. D. Glenn, Estimate of Earth Media Shear Strength at the Nevada Test Site, UCRL-52358, November 3, 1977. 2. N. W. Howard, A. New Look at Experience with CO2 in the Work Point Region, UCRL-89413, 1983, to be presented at the 2nd Containment Symposium, Albuquerque, NM, August 2-4, 1983.

DISCLAIMER

This document was prepared i>s an account of work sponsored hy an agency of the I'nifed Slates Government. Neither the I nited Slates Government nor the I niversily of California nor any of Iheir employees, makes any warrants, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the I niU'd Slates Government or the Inivcrsity of California. Tht> views and opinions of authors expressed herein do not necessarily state or reflect those of the ( nited States Government thereof, and shall not he used for advertising or product endorsement purposes.

300 SITE SELECTION FLU SHEET HOLE NO. EVENT NAME EVENT SCIENTIST Feet Metres Date 1. Coordinates N_ t: E 2. Elevation 3. Proposed DOB 4. Check Nearby Cross Sections from Previous Events 5. Alluvium/Tuff Contact 6. Chec!; Alluvium/Tuff Characteristic 7. Tuff/Paleozoic Contact 8. Check Depth & Thickness of Grouse i.nyon 9. Static Water Level 10. Distance from W.P. to Paleozoic 11. Distance from W.P. to SWL 12. Distance to Nearest Seismic Line 13. Distance to Nearest Pz Control H:t>e 14. Distance to Major Fault 15. Distance to Other Faults, Inferred, etc. 16. Distance to Nearest Collapsed Event 17. Distance to Uncollapsed Events Within .1 DOB 18. Distance to Nearest Stock Hole 19. Distance to All Open Holes Within 1 DOB 20. Distance to'Nearest Drilling Activity 21. Date Submitted to NTED for Area Motion Effects 22. Date Submitted to USGS for Approval

301 SITE SELECTION FLOW SHEET HOLE NO. EVENT NAME Page 2

23. Date Archeological ana Endangered Species Search Completed 24. Date Approval is Received from USGS 25. Date Submitted to Containment for Approval 26. Date Approval is Received from Containment 27. Date Data Submitted to LLNL Data Bank for K Search 28. Date Data Received from LLNL Data Bank 29. Date Paleotopo Received from USGS - Healey 30. Date Location Submitted to H&N 31. Data Location is Field Staked by H&N 32. Date Field Checked 33. Date Post-Shot Plugbacks are Checked 34. Date Location Submitted to C.E. for Approval 35. Date Location Submitted to Drilling Dept. for Approval

REMARKS

BY DATE

302 The normal separation from a collapsed event is 0.5 DOB

0.5 DOB

I I I I I I I

I I I I I I I I I I I I DOB I I I I I I I I I I I I I I I I I I I I I I I I I

The normal separation from an uncollapsed event of non-zero yield is about 1 DOB

DOB

303 The "impeded path" criterion is not so simple as it seems The paths "L" are the same length, but they are certainly not equivalent for containment

6 LA-UR-B3-2262

Lot Ai»i unma tmci i of fn*'g

TITLE THERMODYNAMICS OF HYDROGEN GENERATION

AUTHORui E. S. GAFFNEY

SUBMITTED TO 2ND CONTAINMENT SYI-IPOSIUM 2-4 AUGUST 1983 KIRTLAND AFB ALBUQUERQUE, NM

•r teevpttne* e* thu «iiei» fM pubiiihar roeognlni ttui tfi» US Oovtrnmtni r«t*mi« nontteludvt. reyiHy.tro* term lo publish or reproduce tf>« Publnhtd <0-m of this coninfution, or to *"a* Vff't to 00 »0. fo' US Oovtrnmtnt pwrpotti Th« lot Altmoi Mtiontl Ubordory roqutltl ">»' •*• publ>|f«r iMniify ffi« tnicl* U arort ptrformod under tfx iutSICtt e) tft* U.t Dapinmm

Los Alamos National Laboratory Los Alamos,New Mexico 87545

HO 305 THERMODYNAMICS OF HYDROGEN GENERATION

E. S. Gaffney

Earth Space and Sciences Division Los Alamos National Laboratory

Abstract

Hydrogen generation in underground tests is usually estimated using the reaction of water and iron to produce ferrous oxide (wustite) plus hydrogen. This reaction yields large amounts of hydrogen at temperatures below about 1000 K. However, underground nuclear tests can never be approximated as closed systems with only iron, water and their reaction products. Inevitably silicates and commonly a few percent of ferric iron occur in the media surrounding the working point, and large amounts of magnetite are frequently used as stemming and shielding. Studies of the thermodynamics of iron, oxygen and hydrogen in a wide variety of systems have shown that the iron-wustite system is intermediate in its potential for hydrogen generation because wustite is a relatively unstable compound with a very restricted stability field. Incorporation of silicates makes possible the formation of fayalite (Fe2SiO^) which has a much larger stability field. Reactions involving fayalite serve as hydrogen buffers maintaining the hydrogen pressure at particular levels until either all the iron or all the magnslite is consumed. Even in the absence of silicate and ferric iron (i.e., with the chemistry considered previously), the buffered hydrogen pressure would be about 80 per cent of total pressure. Incorporation of excess magnetite, but not silicate, reduces hydrogen pressure very little at about 600°C but to only 15 percent of the total pressure at about 1000°C. If the system has excess iron and fayalite, buffered hydrogen pressures are even higher than in the absence of silicate; the gas will be virtually pure hydrogen. When magnetite is in excess instead of metallic iron,

306 hydrogen will be buffered to only about 2 percent of the ambient total. In systems where all of the iron and all of the magnetite have been comsumed in the formation of fayalite, no buffer exists and the hydrogen pressure may vary between 2 and 100 percent of ambient and will be controlled by mass balance considerations.

I. INTRODUCTION

Generation of non-condensable gases, such as CO2 and H£, by an underground nuclear test can present a serious containment threat. As a result, zones with high carbonate contents are avoided by both Laboratories when selecting sites for tests. It has also been customary to consider that Ho is generated solely by chemical reaction of iron in the test rack with water from the surrounding formation, especially at low yields where the rack mass is comparable to the mass of rock melted. This paper presents a more complete analysis of the thermodynamics controlling hydrogen generation in underground tests.

The standard approach to hydrogen generation is described by Dreiling and Hudson (1), who consider the reaction of iron and water to yield wustite (FeO) and hydrogen. With this as the controlling reaction, one mole of Ho (.002 kg) will be produced by each moie of Fe (.056 kg) or about 18000 moles of Ho per tonne of Fe. (They assumed wustite was stoichiometric FeO whereas the compound is actually slightly iron deficient; thus H2 generation will be slightly underestimated in their model). Because each mole of H2 presents the same containment threat as a mole of CO2 (mass .044 kg), each tonne of Fe will produce H2 equivalent to about 800 kg of CO2. By considering that mass of "CO2 equivalent" to be dispersed in a region of rock heated to 1100 K (a temperature high enough to release the CO2 from calcite) they derived the relation:

H2 (wt% CO2 equivalent) = MFe/(40W)

where Mpe is the mass of iron in tonnes (Mg) and W is the device yield in kilotons (kt). This relation has been used by both Laboratories since it was published for predictions of gas generation on low yield events.

307 In 1982, Gaffney (2) considered the effects of Fe+' present in the formation and in the stemming on H2 generation and concluded that less gas would ususally be present than predicted by the model of Dreiling and Hudson. However, the actual chemical systems controlling hydrogen generation in underground tests are much more complicated than either of these previous treatments have assumed. This paper describes, these systems in more detail and derives more realistic relations for prediction of hydrogen generation.

II. CHEMICAL ENVIRONMENT OF UNDERGROUND TESTS

The chemical environment of an underground nuclear test includes the materials in the test apparatus plus some part of the surrounding medium. A typical underground nuclear test canister contains, in addition to negligable quantities of chemical and nuclear explosives, large amounts of steel, lead and aluminum and lesser but frequently substantial, amounts of copper, polyethylene and various other alloys and compounds. In addition, large amount of magnetite (Fe-jO^,) and borates are frequently used as stemming around the canister. Although stemming materials are usually oxidized, the materials in the canister are usually in unoxidized metallic or covalent states. Upon detonation, all or most of this material is vaporized or melted and is mixed with material from the surrounding medium.

Table 1 shows some of the elements found in canisters and indicates their oxidization states. Two of the metals (Cu, Fe) can form oxides of various valence states. For such metals oxidation to the lower state (+1 for copper and +2 for iron) is most important for containment considerations. For each mole of n-valent metal oxidized by reaction with water (HoO), n/2 moles of hydrogen gas will be produced. However, as will be shown below, many of these metals cannot be oxidized by water and, hence, present no threat of noncondensable gas production.

The geologic media surrounding the working point are also part of the chemical environment of an underground test. Detonation will vaporize hundreds of tonnes of rock and will melt even more rock. Dreiling and Hudson (1) estimate that in a typical test 1350 Mg of rock are melted per kilDton of yield. Because of the turbulence associated with the high velocity flow, this melted and vaporized rock will be intimately mixed with the vaporized and melted canister materials- Thus this material is also part of the chemical system we must consider.

308 Table 1. Elements common in underground tests

Valance reaction ' examples

+1 1 2M + .502*5=* M20 M - K,Na,Cu ,H,Li

+2 +2 2 M + .5O2=?5=-MO M = Fe ,Ni,Be,Cu ,Pb

+3 3 2M + 1.5O2 ==^-M2O3 M « Al,Fe ,B

A K + 02 ^^M02 M - Si.C

309 Test media at NTS are almost always tuff or alluvium. Because the alluvia are derived largely from tuff sources, the two rocks are chemically similar. The tuffs are rhyolite with typical compositions as in Table 2 (after Warren [3]). Very few chemical analyses of NTS alluvium are available. Bass et al. (4) reported Area 3 alluvium with the composition shown at the bottom of Table 2. All of these rocks have more than 66% silica '(SiC^) and substantial amounts of I^O, A^O-j, and

H^O. As will be shown below, this material has an appreciable effect on H2 generation.

III. BUFFERING IN GEOCHEMICAL SYSTEMS

A. Fe-O-H

When a metal such as-iron appears on both sides of a chemical reaction, but in different oxidation states, the possibility exists for the reaction to maintain the fugacity* of oxygen at a fixed level as long as both products and reactants are present. For example, if we have a system consisting of iron metal, ferrous oxide (wustite), water and hydrogen, we may write the chemical reaction of those constituents as follows

Fe + H2O 3=^ FeO + H2

While both metallic and ferrous iron are present, the composition of the fluid at equilibrium will be fixed for a given temperature and pressure. If the initial mixture of H2 and H2O differs from the equilibrium value, either Fe or FeO will react with the fluid to change the ratio H2/H2O toward the equilibrium value. If there is an excess of the needed reactants, the reaction will cease when equilibrium is attained. Such a system is said to be buffered; as long as both Fe and FeO are present the fluid composition depends only on temperature and pressure and not on the proportions of Fe and FeO.

In a system composed of Fe, O and H there are four chemical reactions that can buffer the fluid composition. Each deals with the oxidation/reduction of a pair of the four compounds iron (Fe), wustite (Fe O) magnetite (Fe Fe2^+O^) and hematite (Fe2 O-j). The buffering reactions are:

*Fugacity is essentially partial pressure corrected for non-ideal gas effects.

310 Table 2. Chemical analyses of typical NTS rocks.

Rock Unit SiO2 T102 A12O3 Fe203 FeO MgO CaO K20 Na20 H20 Rainier Mesa (u) 70.6 0.3 14.2 1.6 * 0.5 1.3 5.2 3.6 1.4 (1) 76.0 0.1 12.3 0.8 * 0.2 0.6 4.8 3.4 1.4 Ammonia Tanks (u) 66.3 0.4 15.3 2.0 * 0.6 1.5 5.0 4.2 0.9 (1) 72.8 0.2 12.7 1.1 * 0.4 1.0 5.0 3.9 0.7 Tiva Canyon 72.8 0.2 13.0 1.1 * 0.2 0.6 5.5 3.9 0.9 Scrugham Peak Lava — 0.3 12.7 0.9 * 1.0 1.0 4.8 2.6 — Topopah Spring (u) 67.4 0.3 14.7 1.6 * 0.3 0.9 5.8 3.9 2.0 (1) 74.3 0.1 12.7 0.9 * 0.1 0.6 4.8 3.2 2.7 Area 20 67.4 0.2 13.2 1.9 * 0.8 1.5 3.6 2.2 7.2 Grouse Canyon 72.3 0.2 11.2 3.3 * 0.1 0.4 4.3 2.7 9.8 74.0 0.2 11.3 2.5 1.0 0.1 0.3 4.6 4.8 0.4 Tunnel Beds 4 69.6 0.2 12.3 1.6 * 0.3 0.9 4.4 2.5 7.3 3 70.0 0.2 12.5 1.3 * 0.4 1.2 4.3 1.8 8.2 2 70.0 0.2 13.5 2.4 * 0.3 0.9 3.4 2.3 6.5 1 69.1 0.4 15.0 3.8 * 0.4 1.5 4.2 2.4 3.7 Tub Spring 73.0 0.1 li.O 2.0 * 0.6 .0.9 3.8 2.4 3.7 Yucca Flat — 0.3 14.7 2.1 * — 1.1 5.6 3.2 __ Red Rock Valley 71.7 0.2 14.4 1.8 * 0.4 0.8 5.2 3.6 1.9 Area 3 alluvium 71.6 0.2 12.1 1.7 0.2 1.0 2.4 3.5 2.0 2.5

relative amounts on FeO and Fe2O3 not determined, all Fe calculated as Fe203

311 H2O ^=* FeO + H2 (I-W)

3Fe + 4H2O *F^ Fe3O4 + 4H2 (I-M)

3FeO + H2O ^ Fe3O4 + H2 (W-M)

2Fe3O4 + H2O ^^fe 3Fe2O3 + H2 (M-H)

Any system initially containing both Fe° and Fe will eventually be buffered by one of these reactions unless the ratio Fe°/Fe corresponds exactly to that of a single iron phase. Which buffer will control depends on the ratio of Fe°/Fe .

Earlier studies of hydrogen generation have all restricted attention to the system Fe-O-H. Dreiling and Hudson (1) considered only the (I-W) reaction and assumed that all the Fe would be consumed, thus leading to an unbuffered rystem. Gaffney (2) pointed out the existence of the other reactions, made possible by the incorporation of Fe in the stemming and surrounding media, but again did not acknowledge the buffered nature of the system. In this paper, the buffer concept is used to predict the fugacity of hydrogen for underground tests.

The buffer concept has been used for decades in experimental petrology to control oxidation states of iron and other multi-valent elements in reactions at high temperatures and high pressures. Originally described by Eugster (5), the method depends on the dissociation of water

and the very high mobility of protons to control the amount of oxygen available for reactions in solids.

From thermodynamic data (6), the hydrogen fugacity for each of the buffers can be calculated. Figure 1 shows these values. Note that the hydrogen fugacity can vary over about 5 orders of magnitude in the present s of magnetite. More practically, we never expect excess hematite so the hydrog ;n fugacity can vary by a factor of about 20 depending on whether iron or wus ite is excess. At low temperatures (below about 850 K, 577°C) only small va iat. ons are possible at equilibrium. In the system Fe-O-H, the field of stability of feu-rci iron is restricted. Wustite occurs only above about 850 K and over a lin>iti" range of hydrogen fugacity. This is because ideal FeO is not stable. All ivir.tites have actual

312 I I T r

0 -M-I

s -2

a. -3

o r—t —4

H-M -5

-6 I I I I 300 400 500 600 700 800 90010001100 Temperature (°C)

Figure i. Hydrogen mole fractions for buffers in the system Fe-O-H compositions of Fej_xO where 0.02< x <0.15. The deficiency of iron is due to the presence of some ferric iron. Thus wustite is a mixed ferrous-ferric oxide like magnetite. B. Fe-O-H-Si

Addition of silica to the system makes possible the formation of stable ferrous silicates and increases both the temperature range and the hydrogen fugacity range for ferrous iron.The stable ferrous silicate is the iron olivine, fayalite, Fe2SiO^. Fayalite can be substituted for wustite to obtain two new hydrogen buffering systems with the following reactions (5,6):

2Fe + SiO2 + 2 H2O 5==="Fe2 SiO4 + 2 H2 (QFI)

We2 SiO4 + 2 H2O =P=^2 Fe3O4 + 3 SiO2 + 2 H2 (QFM)

Figure 2 shows the temperature-hydrogen fugacity loci of these two buffers plus HM which is unaffected by silica. Note that systems wibh fayalite can have hydrogen pressures that differ by factors as large as 50, depending on whether there is excess iron or magnetite.

C. Fe-O-H-Si-K-Al

The components of all systems considered to this point are anhydrous. A further degree of complexity is required when potassium and aluminum are added to 2+ the system. A hydrous ferrous iron compound, K Fe 3 Al Si-j O-^Q (OH)2, or the mica annite, is now possible. Annite, like most hydroxylated minerals, is not stable at very high temperatures. Hydrogen fugacities in this system have been studied by Eugster and Wones (6) and, as anticipated, the appearance of annite below about

700°C ("975 K) changes the buffered f^_j2 of the system. The new reactions involve annite and the feldspar sanidine, KA1 Si-jOg:

2+ KFe 3 Al Si3 O1Q (OH)2 5=^KA1 Si3O8 + Fe3O4 + H2 (ASM)

2+ 2K Fe 3 Al Si3 O1Q (OH)2 + 3 SiO2^=^ 2KA£Si3Og + 3Fe2 SiO4 + 2H2

Even though the ASF reaction involves no change in valence state it still controls hydrogen indirectly because the dissociation of H2O is temperature dependent as is

314 0 -

-1 - o a Q _

a -3 -

00 J2o -4 -

-6 300 400 500 600 700 800 90010001100 Temperature (°C)

Figure 2. Hydrogen mole fractions for buffers in the system Fe-0-H-Si the dehydration of annite. An analogous reaction with iron in place of fayalite occurs at very high hydrogen concentrations and low temperatures

2+ 3Fe° + KAlSi3O8 + 2H2O 5=^ KFe 3 AlSijOg (OH)2 + H2 (AS1)

Figure 3 shows the location of the hydrogen buffers in the system Fe-O-H-Si-K-Al when excess quartz is present. If quartz is not in excess then the sanidine must be replaced by kalsilite (KA£SiO^) and the annite field is somewhat larger; because quartz is the most abundant constituent of most of our containment media and also a common stemming material we will not examine kalsilite systems further.

IV. HYDROGEN BUFFERING IN UNDERGROUND TESTS

As we have seen in the preceeding section, hydrogen generation in underground tests is a complicated function of the temperature, pressure and composition of the chemical system involved. In this section consider some typical situations and calculate expected hydrogen generation for hypothetical test environments.

A. Case with no added magnetite.

When no magnetite is added to the stemming the only oxidizers for the iron in the rack are water and the small amount of Fe^+ in the surrounding medium. If the Fe in the melted or vaporized rock is available as an oxidizing agent, we can estimate its mass given the percent Fe2O-j in the rock and the yield. If we have a relatively low content of iron in the geologic medium surrounding the working point, say one per cent Fe2O3, then we expect about 13.5 tonnes of Fe^3^ per kiloton of yield. For a relatively iron-rich medium, say four percent Fe2O3, about

54 tonnes of Fe2.O3 per kiioton is available. From the reaction

Fe + Fe2 O3 ^=^ 3 FeO we see that 2.86 tonnes of Fe^D-^ will oxidize one tonne of Fe to FeO. (The presence of SiO2 will result in formation of Fe2SiO^ instead of FeO, but the mass balance of Fe to Fe° will not be affected.) Thus ferric iron in the geologic medium can oxidize 4.7 Mg of metallic iron per kiloton for each percent of Fe9O, - l\.l Mq per kiloton for our hypothetical low-iron site and 18.9 Mg per kiloton for our high-iron site.

316 1 i r r r

QFS. 0 AS I

-1 h 3 QFM.

10 h-* -3

2 -4

H-M -5

-8 I I I I I I 300400 500 600 700 800 90010001100 Temperature (°C)

Figure 3. Hydrogen mole fractions for buffers in the system Fe-O-H-Si-Al-K If more metallic iron is present than can be oxidized by ferric iron in the geologic environment, free iron will be present, and the hydrogen concentration will be controlled by the QFI and ASI buffers. Because these buffers lead to high fugacities of hydrogen, hydrogen production may be a serious containment concern on such an event. If there is adequate ferric iron to oxidize all of the metallic iron, then the controlling buffers will be QFM and ASM which keep hydrogen at fairly low levels. In such cases, hydrogen production will not be a threat to containment. In the intermediate case where there is just barely enough ferric iron to oxidize the rack, same water may also be involved in the process and hydrogen fugacities will likely be controlled by ASF. For this buffer, the hydrogen concentration increases rapidly with decreasing temperature (see Figure 3). This situation may present the most serious threat to containment.

It is clear that any action that can increase the amount of ferric iron available for oxidation of the rack will be beneficial to containment by decreasing the concentration of hydrogen in the cavity gas.

B. Case With Added Magnetite

Magnetite is frequently used as a stemming medium around the diagnostics canister because its high density provides radiation shielding. Two-thirds of the iron in magnetite is ferric iron, so it also serves as an oxidizer. For the reaction

o Fe + Fe3O4^ 2 FeO

we need 5.26 Mg of magnetite to oxidize 1 Mg of metallic iron. Magnetite (plus formation ferric iron) in excess of this amount will restrict hydrogen concentrations to the QFM or ASM buffers. Note that the effect of magnetite in the stemming is (nearly) independent of yield, so that low yield events present no more of a threat to containment than high yield events, if there is excess magnetite. If there is excess metallic iron, then the controlling buffers are QFI and ASI but the total mass of metallic iron available to produce H2 is still reduced by 0.19 times the mass of magnetite.

C. Effect of Yield

The effect of yield on pressure of hydrogen generated by and underground nuclear test is well known. In the standard analysis [e.g., Dreiling and Hudson (1)]

318 the mass of H2 is related to the mass of metallic iron. That amount of hydrogen, which is independent of yield, must be contained in a cavity volume that is proportional to yield so that pressure will be inversely proportional to yield. The present work points to another effect of yield - that of changing the composition of the chemical system to be considered. As the yield varies so does the radius of the cavity.

At very low yields (a few 100 tons or less) the melted material may include only a portion of the diagnostics and stemming, but virtually none of the surrounding formation. If the stemming included is magnetite, the chemical system may closely approximate the Fe-O-H system described earlier. In this case the proper buffers will be WI, WM and MI rather than QFI or QFM. Furthermore, the masses of iron and magnetite will be complicated functions of the yield because large amounts of each will be some distance from the melt region. A detailed analysis of such systems would require 2- or 3-D radiation/hydrodynamic calculations and is beyond the scope of this paper.

The main effect of increasing yields above several 100 tons is to increase the amount of the surrounding formation included in the chemical system. This has two important effects. First, it adds silica so that fayalite replaces wustite as the ferrous phase. If there is excess magnetite, this result? in substantial lowering of the hydrogen threat. If there is excess iron then hydrogen will be somewhat more of a threat. In addition to silica, feldspar is also added, which allows formation of annite. The second effect is to add ferric iron to the system. This will oxidize metallic iron and reduce to total mass of hydrogen produced. Thus at high yields, not only is any hydrogen produced contained in a larger volume, but less of it will be produced because the overall chemical system will include more of the highly oxidizing geologic medium. A 10 kiloton event in an iron-poor formation with only one percent Fe2O3 will produce no H2 if the rack is smaller than about 135 tonnes. In an iron-rich formation (4 percent FeoOj) the rack could exceed 500 tonnes with no threat of hydrogen production. This phenomenon may be responsible, at least in part, for the marked improvement in containment experience at yields greater than 10 kt.

V. FURTHER COMPLICATING FACTORS

Although a considerably more complicated set of chemical systems has been addressed here than by previous authors, even the system Fe-O-H-Si-Al-K is far

319 less complicated than the real underground systems. It has been assumed that all of the rack mass is metallic iron whereas up to several tons of other metals and compounds may be present. Metals more readily oxidized than iron such as aluminum or magnesium should be added in a mole equivalent to the mass of metallic iron to be oxidized. One mole of aluminum will be equivalent to about 3 moles of iron, whereas one mole of Mg will be equivalent to only about 2 moles of iron. This is because aluminum oxidizes to a trivalent state but iron and magnesium, only to divalent ones. Unoxidized compounds such as B^C and polyethylene should be treated similarly. Metals such as copper and lead which oxidize less readily than iron can be omitted from the rack mass for purposes of hydrogen production.

The silicate portions of the real chemical systems have also been idealized, the most complicated system dealt with here is Fe-O-H-Si-Al-K whereas the rocks at NTS contain appreciable amounts of other components, most notably sodium and carbonate. Inclusion of sodium permits solid solution in the feldspar, resulting in a slight movement of the mica-feldspar buffers into the mica field of Figure 3(7). Inclusion of carbonates will introduce COo into the gas phase which, in turn, makes possible the reaction

CO, + H, s?=* CO + H,O which has been ignored above. There is an exceeding large literature on the thermodynamics and kinetics of this "water-gas shift" reaction which must be addressed in a complete study of non-condensable gas generation by underground tests. \

The discussion above deals with an idealization of the mineral phases present in the chemical system. In reality, large amounts of glass are produced by the test and the effect of glass on hydrogen buffers is essentially unknown. Furthermore, the surrounding formations frequently contain large amounts of clay minerals, some of which can have iron in either ferric or ferrous states. Thermodynamic methods have been proposed (8) which could provide estimates of the hydrogen fugacity for systems buffered by iron-bearing clay minerals such as chlorite, vermiculite or nontronite but experimental data are lacking so that application to this problem is not justified. However, ultimately such low temperature phases should be considered because they may that control the hydrogen concentrations at late times when cavity temperatures fall below 400°C.

320 Reaction kinetics present a final complication. Kinetics of silicate reactions at temperatures below 700-800°C are notoriously sluqgish; reaction times of days or weeks are often required. On the other hand, reactions involving only proton transport can occur much more rapidly. When considered together we should, perhaps expect that on a time scale of minutes to hours that buffering will be by high temperature assemblages (such as QFM instead of ASM) because the reaction of sanidine and magnetite to form annite will be kinetically hindered. But, which ever buffer is in operation, it should respond rapidly (seconds to minutes) to temperature or pressure changes because only protons need diffuse.

VI. CONCLUSIONS The above discussion of the thermodynamics of hydrogen generation by underground nuclnar tests has shown that the ratio of hydrogen produced to metallic iron present in t. e rack is a function of several variables. The single most important variable is the composition of the chemical system. The composition, in turn, depends on the rack mass, the stemming/shielding materials, the yieid and the composition of the surrounding geologic medium. The containment scientist rarely has any input into either the composition and mass of the rack or the device yield. However, by judicious choice of the working point location and the stemming medium around the canister, the threat of leakage of non-condensable hydrogen gas can be substantially reduced. Working point media with high iron contents should be used if possible. Due to their generally high Fe contents (Table 2) thr Grouse Canyon and Tunnel Beds 1 and 2 should be preferred and the lower Rainier Mesa, lower Ammonia Tanks, Tiva Canyon, and lower Topopah Spring formations should be avoided if hydrogen generation presents a containment concern. Use of magnetite stemming in the vicinity of the device should be encouraged whenever large racks are used and low yields are likely.

This study has described hydrogen buffering systems likely to function within the environment of an underground test. Over the next year or so I anticipate that predictions calculated from chemical equations presented herein will be compared with gas sampling observations.

321 VII. REFERENCES

(1) L. A. Dreiling and B. C. Hudson (1980) "Non-condensible Gas Generation." Lawrence Livermore National Laboratory Report UOPKL-80-17 (CFRD).

(2) E. S. Gaffney (1982) "Hydrogen Gas Generation on Underground Nuclear Tests." Los Alamos National Laboratory Report ES/CPO-82-28 (CFRD).

(3) R. G. Warren (1982) "Chemical Compositions of Volcanic Units in the Vicinity of Yucca Flat." Los Alamos National Laboratory Memo ESS-2-82-39.

(4) R. G. Bass, H. L. Hawk and A. J. Chabai (1963) "Hugoniot Data for Some Geologic Materials." Sandia Laboratories Research Report SC-49G3.

(5) H. P. Eugster (1957) Heterogeneous reactions involving oxidation and reduction at high pressures and temperatures. J. Chem. Phys. 26, 1760.

(6) H. P. Eugster and D. R. Wones (1962) Stability relations of the ferruginous biotite, annite. J. Petrology 3, 82-125 and references therein.

(7) M. J. Rutherford (1969) An experimental determination of iron biotite - alkali feldspar equilibrium. J. Petrology 10, 381-408.

(8) J. O. Nriagu (1975) Thermochemical approximations for clay minerals. Amer. Mineral 60, 834-839.

322 A CHEMICAL INVESTIGATION OF GLASSES PRODUCED BY THE RAINIER UNDERGROUND NUCLEAR EXPLOSION

A. 3. Piwinskii F. 2. Ryerson and W. F. Beiriger

323 ABSTRACT In order to better understand the chemical and physical processes taking place both during, and subsequent to, underground nuclear explosions, we have commenced analysis of glasses and residual crystalline phases obtained from post-shot drill back samples. The samples were chosen in order to assess the effects of 1) shot medium chemistry, 2) initial H2O content of the medium, 3) yield of the event, and U) depth of burial. Glasses preserved in samples collected from the Rainier underground nuclear explosion are dehydrated relative to both the host media and hydrated glasses found within the host media. The post-shot glasses are extremely heterogeneous with respect to major element chemistry on a 20 urn scale. One ubiquitous feature characteristic of the glasses from this low-yield event is the presence of dark "marble-cake" regions within the glass. Results indicate that these regions are locally enriched in iron due to the presence of a discrete iron-rich phase which may be related to device debris. In a number of cases, optically amorphous regions with chemistries of crystalline phases are present as evidence of shock melting. In general, the glass compositions do not indicate any migration of the major elements from the cavity, and only limited redistribution and homogenization within the cavity. Although temperatures were far in excess of that of the dry liquidus for these compositions, the short duration of the thermal pulse and high viscosity of the melts produced precluded rehomogenization by diffusion or convection, respectively.

INTRODUCTION One of the major goals of our research effort in support of the LLNL Test Program is an understanding of the phenomenology of underground nuclear explosions and its relationship to the various lithologies in which the nuclear events occur. Included here are: 1) melt production, 2) gas evolution, and 3) radionuclide migration in post-shot iithologies. Initial stages of our investigation have focused on the following tasks: • characterization of pre-shot country rock lithologies in whic^ well-documented nuclear events have occurred * characterization of post-shot core samples as a function of distance from the working point.

The chemical study of the glasses formed as a result of underground nuclear explosions is a major thrust of our current research activity. A comparison of the major element chemistry of glasses found in post-shot samples with pre-shot lithologies should enable one to distinguish which rocks and minerals are melting, and perhaps more importantly, how they are melting. Some key questions include the following. Are glass compositions indicative of total melting or of equilibrium partial fusion? Are there any indications of volatile losses in the glasses? In this paper, we report the results of our chemical study of glas^s * produced by the Rainier underground nuclear explosion. Our investigat; , the glasses formed during the Duryea and Cannikin events is currently being coi pieced and results will be reported shortly (Piwinskii et alia, 1983).

324 THE RAINIER EVENT The Rainier event was the first completely contained underground nuclear explosion (Warner and Violet, 1959). It occurred at the U.S. Atomic Energy Commission's Nevada Test Site (NTS), Nye Country, Nevada on 19 September. 1957 (See Figure 1). The nuclear device was detonated 899 feet beneath the surface of Rainier Mesa and approximately 1950 feet horizontally in from the slope (See Figure 2) (Flangas and Shaffer, 1960; Wadman and Richards, 1961). The explosion occurred in the bottom of a shallow east-northeast trending syncline (See Figure 3), 90 feet above the base of a bedded, indurated, porphyritic rhyolitic to quartz dacitic tuff of the Oak Springs formation (TOS7) of Tertiary age (Wilmarth et alia, 1960; Thompson and Misz, 1959). The Rainier explosion had an energy release equivalent to 1,700 tons of high explosive, and according to calculations of Johnson et alia (1959), it produced temperatures and pressures of 1 x 10° K and 7 Megabars, respectively, a few microseconds after detonation. Rock Chemistry and Mineralogy The petrochemistry of the Oak Springs Tuff formation in which the Rainier event occurred has been extensively documented by workers from the U.S. Geological Survey, University of California at Berkeley, and the Lawrence Livermore National Laboratory. The tuffaceous rocks located near the Rainier event tunnel (See Figures 2 and 3) have been divided into eight lithologic units (Figure 3). Chemical analyses of tuff units 3, 4, 6, and 7 are given in Table 1. The tuffs occurring near the Rainier explosion chamber and the exploratory tunnel were also analyzed. Results are provided in tables 2 and 3, respectively. Mineralogical modal analyses of tuffs occurring close to the working point (WP) were determined by Wilmarth and his colleagues (1960). The mean of 52 analyzed tuff samples is recorded in Table 4. A perusal of the petrochemical data contained in Tables 1, 2, and 3 leads to the following conclusions: • the Oak Springs tuff units are very heterogeneous in their major element chemistry, especially with respect to silicon dioxide • the tenor of water is very high in all tuff units. Post-Shot Investigations The effects of the Rainier underground nuclear explosion on the surrounding country rocks was very extensively studied by tunneling back towards the working point and by sinking 13 re-entry drill holes (Thompson and Misz, 1959). Disposition of these drifts and drill holes is indicated in Figure 4. Two geological cross sections, A-A1 and B-B', which were constructed orthogonal to each other, are illustrated in Figures 5 and 6, respectively. In perusing the latter figures, the reader is asked to note the locations of drill holes H (Figures 5 and 6), T-l 1 (Figure 5) and K-l (Figure 5). Glasses Formed by the Explosion The intense pressures and temperatures generated by the Rainier nuclear explosion resulted in the fusion of some of the surrounding tuffs which, on cooling, formed glasses. There are at least three varieties of glass formed in contained underground nuclear explosions (Short, 1964; Rawson, 1966). The first is a "dense, dark glass which is high in radioactivity and is believed to represent puddle accumulation in which mixing of melts flowing and dropping from the cavity aids in concentrating the radionuclides "(Short, 1964). The second type is "light in color, variably vesicular, and low in radioactivity.... It develops by superheat fusion of blocks incorporated into the melt puddle after the cavity expansion phase" (Short, 1964). The third variety of glass is a "highly radioactive, dense form that appears as globules, blebs, smears, and coatings along fracture surfaces in fissures and along emplacement drift walls.... Most of this glass is believed to result from rapid condensation of gases composed primarily of vaporized silicates" (Short, 1964).

325 RESULTS The identification and classification of the various types of glasses produced by the Rainier nuclear explosion were based on observations made by previous workers in the exploratory tunnel and various drifts (See Figures 4, 5, and 6). Unfortunately, we were unable to gain access to the tunnel or to any of these drifts; thus, we did not have the opportunity to collect the various types of glasses detailed by previous workers _in situ. Ail of our studies were carried out on specimens chipped from cores containing visible glass from drill holes H, T-ll, and K-l (See Figures 5 and 6 for drill hole locations). We are confident that we collected glass specimens of the first type, which occur in the puddle zone region of the cavity. According to Figures 5 and 6, drill Holes H, K-l and T-l 1 penetrate the cavity puddle zone. We were unable to obtain glass samples of the third variety, not having access to the exploratory tunnel walls and various drill back drifts. We may have sampled the second type of glass, but it cannot be shown unequivocally at the present time. Chemical analyses reported in this paper were obtained using a JEOL-733 electron microprobe at 15 Kv and 15 nanoamps sample current measured on a Faraday cup. The resulting glass analyses were corrected using a modified Bence-Albee scheme. Excessive alkali losses were avoided by employing a wide beam size. Hundreds of chemical analyses were obtained on glasses occurring in drill holes T-ll, K-l and H. Because of ipace limitations, however, we shall present and discuss only those new analytical results which we feel are of general interest and provide the most insight on the phenomenology of underground nuclear explosions. Rainier T-l 1 Drill Hole • In the T-l 1 drill hole, glass was found at the following depth intervals as measured from the working point (WP): 6.51 - 9.3' interval 10.0' - 13.0' interval 23.0'- 25.0'interval 25.0' - 26..01 interval 26.0' - 27.0' interval 27.0" - 28.5' interval 28.5' - 30.0' interval. Chemical analysis of glasses occurring in the 10.0' - 13.0' and 27.0' - 28.51 depth intervals are given in Table 5. Figures 7, 8, and 9 are photographs of regions in the glass from which we obtained our analyses. The glasses from the T-ll hole contain both colorless and brown regions (See Figures 7 and 9). Distinct chemical signatures characterize each of these regions (See Table 5 and compare analyses 3-7 (BR) with others). The brown regions in general, have a greater tenor of FeO and TiO2 than colorless glass and much lower SiO2 contents. Flow structures occur in some of the glasses as manifested by the bubble train lineations observed in Figure 8 and the elongated brown regions illustrated in Figure 9. Evidently these tufiaceous melts were capable of very limited flow, even though anhydrous, silica-rich melts are extremely viscous at, and above, their liquidus temperatures. Piwinskii and Weed (1980) experimentally determined the dynamic viscosity of anhydrous tuff melts obtained by fusing tuffaceous rocks in which the Schooner event occurred. In the temperature interval 1300° to 1700°C, measured viscosities ranged from 10 to 10 Pa.s.

326 Rainier K-I Drill HoJe In the K-l drill hole, glass was observed at the following depth intervals as measured from the WP: 19.5' interval 32.8' - 33.01 interval 33.0' - 33.21 interval. Chemical analysis of glasses occurring at these depths are recorded in Table 6. Figures 10 through 14 are photographs of regions in these glasses on which we have obtained chemical analyses. Glasses from the K-l hole also prominently display both colorless and brown regions (See Figures 12 and 14). Each area is characterized by a unique major element chemistry (See Table 6 and compare analysis of samples 8-6(BR) and 2-2(BR) with the others). Brown glass regions generally have higher FeO and Tin2 contents than the colorless glass and are lower in SiO2- However, there are wide variations in major oxides within any one region testifying to large chomical heterogeneities within these nuclear glasses. From the data reported in Table 6, it is clear that glasses from the K-l drill hole are essentially anhydrous. Well developed banded flow features are present in some of the glasses. Figures 10 and 11 illustrate two varieties observed at the 19,5' depth level in the K-l drill hole. The glasses from the T-ll and K-l drill holes which we have analyzed are generally devoid of any crystals. However, one occasionally encounters an isolated "crystal island" floating in the "glassy sea", such as that shown in Figure 13. Almost always these refractory "crystal islands" are quartz, almost pure Sin2« We are of the opinion that these relict quartz crystals are unmelted phenocrysts which are present in the original Rainier tuff. (According to the modal analysis given in Table 4, 2.6 volume per cent quartz occurs as phenocrysts in the Rainier tuff.) Rainier H-Drill Hole In the H drill hole, glass occurs at the following depth intervals as measured from WP: 32.6'- 34.11 interval 39.3' - 40.3' interval 40.3' - 43.2' interval 51.4'-52.5" interval Chemical analysis of glasses found at these depths are listed in Table 7. Some specimens collected from the 32.6' - 34.1' interval are especially interesting because they illustrate the transition zone between the unmelted Oak Springs tuff unit and the fused rock. Figures 15 and 16 depict the nature of this transition which is relatively sharp and clean. We did not observe any extensive interfingering of silicate melt into the surrounding unmelted tuff. Glasses from the H hole are characterized by colorless and brown regions (See Figures 17, 18, and 19). As previously observed in drill Holes K-l and T-ll, a distinctive chemical signature is characteristic of each area (See Table 7). The brown glass invariably possesses a higher concentration of FeO than the colorless glass and lower S1O2 contents. However, there is a large variation in these oxides, as well as in titania, as manifested by analyses II-l(BR), II-2(BR), D-ll(BR), and 3-2(BR) recorded in Table 7. Colorless glass areas are characterized by high silica and alumina contents and also display wide variations in these oxides (Note analyses VI-1, P-2, and 3-3 in Table 7). The new analytical data recorded in Table 7 clearly testify to widespread chemical heterogeneities present in these nuclear glasses over distances of approximately 20 to 40 irn. Glasses analyzed from the H drill hole are also essentially anhydrous. Features diagnostic of flow are visible in some of the glasses recovered from the H drill hole. Elongated and diffuse areas of brown glass are shown in Figures 17, 18, 19, and 20. Linear extended bubble trains in a glass sample from the 32.6' - 34.1' interval are illustrated in Figure 21.

327 Another rare instance of an isolated crystal sitting in a "glass sea" was observed in the transition region interval (32.6* -34') of the "H" drill hole. An subhedral quartz crystal (SiO2) is situated in the bottom of Figure 22. Glass Chemistry As A Function of Distance From the Working Point It is of interest to investigate the chemistry of the glasses as a function of distance from the WP. The H and K-l drill holes commence approximately 30 feet below the WP (See Figures 5 and 6). The T-l 1 drill hole, on the other hand, is situated to the left of the WP, but also starts approximately 30 feet from it (Figures 5 and 6). Variations in MgO, Na20 + K20, and FeO of glasses from the T-l 1, K-l, and H drill holes as a function of distance from the WP are illustrated in Figure 23. From the data portrayed, we can conclude the following: • glasses from the Rainier event are extremely heterogeneous • the wide variation in Na?0, K20, are FeO of glasses recovered from the 33.0' - 33.2' interval of the K-[ drill hole (See Figure 24) masks any systematic change in these oxides as a function of distance from the WP • the amount of magnesia in ail Rainier glasses is apprcv

328 CONCLUSIONS Our chemical investigation of glasses produced by the Rainier underground nuclear explosion has revealed that: • glasses produced are dehydrated relative to the water-rich tuffaceous host rocks • glasses formed as a result of the explosion are extremely heterogeneous in bulk element composition • electron microprobe studies are necessary to document the highly variable glass chemistry due to the scale (^ 20i_m) of iruhomogeneity dark, iron-rich, silica-poor regions characterize some Rainier glasses there was no appreciable migration of major elements from the cavity region • shock melting was one important process occurring during the nuclear event. Some unanswered questions which arose as a result of our studies are the following: does the chemical disequilibrium and heterogeneous glass chemistry persist in high yield events? • what is the effect of lower silica and high refractory oxide bulk compositions on melt production, glass heterogeneity, and chemical disequilibrium? • are the dark, iron-rich, silica-poor regions found in Rainier glasses residual remnants of device debris? • are the chemical signatures of the low Rainier explosion characteristic of other low yield events in tuff media? ACKNOWLEDGMENTS

We thank K. Oswald, E. Lesses, R; MacArthur, D. Richards, W. Richards, 3. Schweiger, A. Denning, R. Wilcox, R. Carroll for assistance in sample collection, information, and logistics support at NTS and the U.S. Geological Survey, Denver office, J. Olson and C. Violet for discussions of the Rainier event, and Ms. D. Olson for the efficiently typing the manuscript.

329 REFERENCES

Diment, W., Wilcox, R., Keller, G., Dobrovolny, E. Kracek, F., Roller, 3., Peselnick, L., Robertson, E., Lachenbruch, A. and Bunker, C. (1958) United States Geol. Survey TEI Report 672.

Diment, W.. Wilmarth, V., Wilcox, R., Clebsch, A., Manger, G., Hawley, C, Keller, G., Robertson, E., Peselnick, L., Bunker, Healey, D,, Kane, ML, Roth, E., Stewart, S., Roller, 3., Jackson, W., Oliver, H., Byerly, P. and Mabey, D. (1959) United States Geology Survey TEI Report 355.

Flangas, W. and Shaffer, L. (1960) University of California LLNL Report, UCRL-5949.

Johnson, G., Higgins, G., and Violet, C. (1959) Jour. Geophys. Res., Vol. 64, No. 10, p. 1457-1*70. Piwinskii, A. 3., Ryerson, F. 3., and Beiriger, W. (1983) In Preparation. Piwinskii, A. J. and Weed, H. C. (1980) Thermochemica Acta, Vol. 37, No. 2, p. 189-195.

Rawson, D. (1966) University of California LLNL Abs't. 14730-T.

Short, N. (1964) University of California LLNL Report, UCRL 7787.

Thompson, T. and Misz, 3. (1959) University of California LLNL Report, UCRL-5757.

Wadman, R. and Richards, W. (1961) University of California LLNL Report, UCRL-6586.

Warner, S. and Violet, C. (1959) University of California LLNL Report, UCRL-5542.

Wilmarth, V., Botinelly, T., and Wilcox R. (1960) United States Geol. Survey Prof. Paper 400-B, p. B149-B151.

330 •a- —

o o o o d

CM q a c5 c5

•3- — oo c5 a d O

— a a — — o od d d

d d o

o o o CM o —• —• d d d n O _ u-\ -t 00 rvi —* —• o -J — d d

o — — d o I- 1-5

331 Table 2. Chemical Analyses of Tuffs Near the Rainier Explosion Chamber, Weight Per Cent (after Diment et alia, 1959)

S T U V W X Y 2 Above Z

SiO2 66.* 67.2 64.9 66.7 58.5 62 .4 67.5 69.5 69.5

A12°3 12.0 10.2 10.5 11.9 16.4 14.2 12.7 11.9 12.8

Fe203 2.7 2.0 2.4 2.5 4.0 3.4 1.9 1.5 1.5

FeO 0.07 0.06 0.05 0.23 0.65 0.34 0.27 0.15 0.13

MgO 0.8 0.85 0.80 1.0 1.4 1.4 1.0 1.0 0.94

CaO 2.0 2.1 2.0 2.4 3.8 2.9 2.5 2.3 1.8 Na2° 1.0 0.42 0.58 1.2 2.3 1.4 1.6 1.0 1.8 K2° 2.5 2.3 2.2 2.1 1.9 2.3 2.6 2.3 2.6

TiO2 0.28 0.11 0.2 0.33 0.61 0.48 0.27 0.19 0.18 co2 0.03 0.00 0.00 0.07 0.21 0.10 0.10 0.05 0.01

P2°5 0.10 0.08 0.06 0.10 0.15 0.12 0.08 0.09 0.08 MnO 0.12 0.18 0.11 0.17 0.28 0.16 0.08 0.12 0.06

+ H20 6.0 7.4 5.6 4.6 4.8 4.9 5.3 4.7 } 16.2 H20" 6.0 7.1 5.7 5.2 6.0 4.5 4.6 3.9

TOTAL 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

SiO2 = 65.84 ± 3.52

FeO + Fe203 = 2.65 ± 0.99

K20 + Na20= 3.56 ±0.60

CaO + MgO = 3.44 ± 0.82

332 Table 3. Chemical Analyses of Tuffs From the Rainier Exploratory Tunnel, Weight Per Cent (after Diment al alia, 1959)

Sample Sample Sample Sample 15 + 00 J5 + 50 J5 + 60 16 + 05

SiO2 67.85 66.00 68.08 69.21

A12°3 11.93 13.01 13.09 11.28

Fe203 2.67 2.56 2.85 2.23

FeO 0.07 0.00 0.10 0.07

MgO 0.95 0.96 0.96 0.89

CaO 2.07 2.24 2.16 2.13

Na20 1.76 1.59 2.18 1.36

K20 3.09 3.21 3.37 2.83

TiO2 0.27 0.18 0.31 0.17

P2°5 0.07 0.03 0.08 0.08 MnO 0.08 0.08 0.10 0.07 co2 0.0 J 0.01 0.01 0.02

H20+ 4.10 5.48 3.35 4.86

H20" 4.69 4.36 3.62 4.49 *

TOTAL 99.61 99.71 100.19 99.69

SiO2 = 67.77 ±1.33

FeO + Fe203= 2.64 ±0.28

Na20 + K20 = 4.85 ±0.56

CaO + MgO = 3.09 ± 0.09 333 Table 4. Modal Analysis of Tuff Collected Close to the Rainier Explosion Chamber, in Volume Per Cent (after Wilmarth et alia, 1960). t

Phenocrysts 17 A Quartz 2.6 K-feldspar . 6.1 Piagioclase 6.8

Biotite 1.3 Pyroxene and Amphibole 0.2 Magnetite 0A Xenoliths 6.9 6.9 Shards and Lapil.'i 68.0 Zeolite: Clinoptilite (?) 23.0 Clay: Montmorillonite 12.0 3-cristobalite 11.0 Amorphous Material 22.0 Vesicles 7.7 7.7

TOTAL 100.00 100.00

Mean of 52 analysed samples

334 Table 5. Chemical Analyses (Weight Per Cent) of Glasses from the Rainier T-l 1 Hole.

10.0' to 13.0' Depth 27.01 to 28.51 Depth

Sample Sample Sample Sample Sample Sample Sample Sample 3-2 3-3 3-5 3-8 3-7(BR) 1-2(A) 11-6(B) 11-8(B)

SiO2 78.22 72.60 68.16 65.77 58.85 71.58 69.25 67.79

A12O3 10.48 14.58 19.71 21.29 12.49 14.82 18.08 17.30

MgO 0.72 0.90 0.42 0.58 1.04 1.16 1.60 0.92

FeO 0.92 1.49 0.56 0.59 14.72 2.47 3.16 2.97

CaO 1.61 3.19 3.17 5.25 3.65 3.89 3.68 3.98

Na2O 2.02 1.41 1.74 2.18 1.43 1.29 1.36 1.48

P 0.00 2°5 0.00 0.04 0.00 0.43 0.00 0.37 0.19 S 0.05 0.00 0.02 0.00 0.00 0.00 0.00 0.00

K20 3.08 3.04 3.66 2.78 2.27 2.41 2.94 3.16

TiO2 0.17 0.09 0.00 0.00 1.52 0.51 0.25 0.64

Cr 0.00 0.00 0.00 0.09 0.17 2°3 o.-oo 0.05 0.00 MnO 0.00 0.00 0.02 0.00 0.83 0.00 0.00 0.00

BaO 0.00 0.13 0.00 0.00 0.19 0.45 0.44 0.00

NiO 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

ZrO2 0.19 0.00 0.33 0.00 0.18 0.00 0.00 0.00

TOTAL 97.46 97.43 97.84 98.44 97.68 98.74 101.18 98.45

335 O — —' d

a Tr-, o o o

00 >r\ o o —. a. d d d a C r-i

E 7 —i CS —*

C cr>

a. m o o — t*\ —. ». — d d d d Er d d o o o ^ O o — do

-^ — u-\ "5. — o d

— o d d

t/7 S

iios \ o M5 —. O n) 7 — do

\0 1A O o o o o »• - d d d d t/1 >

60 ir\ o 00 o "D. 1A -. O a- o o ? d

_4J o v "5. o o E1(M) —< -^ f^ *-w d d

o o — VC —t —4 d d d

0 Z Z IJ

336 O o OO — tN O d

if. a. fciN

o — » - o — a. —. o d d V. -.

-J. K O o o o — ^ O O o o N. — r-i — d d d

o •a -j IA o q q q <43 _ |N1 d d d d

O o IN —. q q q m —. o d d

—• —'VD CM—• —

a • «00 q0> q d d d

< o ..T1 n O

337 FIGURE CAPTIONS Figure 1. Index Map of Nevada Test Site illustrating the location of the Rainier U12B event (after Wadman and Richards, 1967) Figure 2. General Geological Map of the Rainier Area, 12 Event (after Thompson and Misz, 1959) Figure 3. Geoiogical Section of the Rainier Event (after Flengas and Shaffer, 1960) Figure 4. Plan View of the Rainier Event (after Thompson and Misz, 1959) Figure 5. Geological Cross Section A-A1 Through the Rainier Event Chamber (after Thompson and Misz, 1959). See Figure 4 for location of Section A-A1. Figure 6. Geological Cross Section B-B1 through the Rainier Event Chamber (after Thompson and Misz, 1959). See Figure. 4 for location of Section B-B1. Figure 7. Glass Sample 3 from The 10.11 - 13.0* Interval of the Rainier T-ll Drill Hole. See Figure 5 for the T-l 1 drill hole position. Locations of chemical analyses are shown by black arrowheads; arabic numbers indicate the analyses as recorded in Table 5. Note the irregular shaped, dark brown glass region in the centre of the Figure. Figure 8. Glass Sample 1 from The 27.01 - 28.5' Interval of the Rainier T-l 1 Drill Hole. See Figure *> for the T-l4 drill hole position. The location of the chemical analysis is shown by the black arrowhead. The arabic number identifies the analysis as recorded in Table 5. Note the dark elongated bubble ellipsoids in the central portion of the photograph. Figure 9. G:ass Sample 11 from The 27.01 - 28.5' Interval of the Rainier T-l 1 Drill Hole. See Figure 5 for the T-l 1 drill hole position. The locations of the two chemical analyses are shown by the black arrowheads. The arabic numbers indicate the analyses as recorded in Table 5. Note the elongated dark brown glass region in the left central portion of the photograph and the fine-scale flow structure occurring above it. Figure 10. Glass Sample 12 from The 19.5' Interval of the Rainier K-l Drill Hole. See Figure 5 for the K-l drill hole position. The location of the chemical analysis is marked by the black arrowhead. The arabic number identifies this analyses in Table 6. Note the pronounced flow structure exhibited by the dark brown glass regions. Figure J J. Glass Sample 13 from The 19.51 Interval of the Rainier K-l Drill Hole. See Figure 5 for the K-l drill hole position. The location of the chemical analysis is illustrated by the black arrowhead. The arabic number keys this analysis as recorded in Table 6. The photograph contains another example of flow structure observed in Rainier glasses.

338 Figure 12. Glass Sample 2 from The 32.8' - 33.0' Interval of the Rainier K-l Drill Hole. See Figure 5 for the K-l drill hole position. The locations of the three chemical analyses are shown by the black arrowheads. The arabic numbers indicate these analyses as given in Table 6. Note the irregular shaped dark brown glass region in the center of the photograph. Figure 13. Glass Sample 27 from The 33.0" - 33.2' Interval of the Rainier K-l Drill Hole. See Figure 5 for the K-l drill hole position. The locations of the two chemical analyses are marked by the black arrowheads. The arabic numbers identify these analyses in Table 6. Note the relict anhedral quartz grain in the central portion of the photograph. Figure 1*. Glass Sample 8 from The 33.0' - 33.2' Interval of the Rainier K-l Drill Hole. See Figure 5 for the K-l drill hole position. The locations of the three chemical analyses are illustrated by the black arrowheads. The arabic numbers key these analyses in Tabie 6. Note the large, irregular dark brown glass region in the center of the photograph and the flow structure present in the specimen. Figi -e 15. Specimen V from The 32.6' - 34.9' Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. Note the presence of the sharp transition zone, as indicated by the thin white line, between the Oak Springs tuff and the glass produced by the explosion. Figure 16. Specimen VI from The 32.6' - 34.1' Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. Thin white lines trace outlines of plagioclase grains in the tuffaceous transition zone. The location of the chemical analysis in the glass region is shown by the black arrowhead. The arabic number indicates the analyses in Table 7. Figure 17. Glass sample II from The 32.6' -34.1' Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the three chemical analyses are marked by the black arrowheads. The arabic numbers identify these analyses in Table 7. Note the presence of flow structures as manifested by the elongated, dark brown glass bands in left-central portion of the photograph. Figure 18. Glass Sample D from The 39.3' - 40.3' Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the three chemical analyses are illustrated by the black arrowheads. The arabic numbers key these analyses in Table 7. Note the presence of dark brown glass areas in top central portion of Figure and the presence of flow banding throughout the glass. Figure 19. Glass Sample 3 from The 51.4' - 52.5' Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the two chemical analyses are shown by dark black arrowheads. The arabic numbers indicate these analyses as recorded in Table 7. Note the presence of the large, irregular dark brown glass region in the central portion of the photograph.

339 Figure 20. Glass Sample 13 from The 40.3" - 43.21 Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the two chemical analyses are shown by dark black arrowheads. The arabic numbers identify these analyses in Table 7. Note the thin flow bands in the glass. Figure 21. Glass Sample P from The 32.61 - 34.1* Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the two chemical analyses are illustrated by the dark black arrowheads. The arabic numbers key these analyses in Table 7. Note the presence of multisized dark bubbles aligned in linear patterns, especially in the central part of the photograph. Figure 22. Glass Sample III from The 32.6' - 3*. I1 Interval of the Rainier H Drill Hole. See Figure 5 for the H drill hole position. The locations of the three chemical analyses are shown by the dark black arrowheads. The arabic numbers indicate these analyses in Table 7. Note the relict quartz grain, outlined in white, which occurs in the glass. Figure 23. AFM Diagram of Rainier Glasses Obtained From the H, T-l 1 and K-i Drill Holes. See Figures 5 and 6 for drill ho.le positions. A = Na2n + K2O in weight percent; F = FeO in weight percent; M = MgO in weight percent. Figure 24. Qz-Ab-Or Diagram of Rainier Glasses Obtained From the H, T-l 1, and K-l Drill Holes. See Figures 5 and 6 for drill hole positions. Qz = normative quartz (SiC^); Ab = normative albite (NaAlSi3<)g); Or = normative orthoclase (KAlSi3Og).

Figure 25. SiO2 Content (Weight Percent) of Rainier Glasses Obtained From The H, T-l 1 and K-I Drill Holes As a Function of Distance from the Working Point. See Figures 5 and 6 for drill hole positions. Figure 26. NaoO + K2O Content (Weight Percent) of Rainier Glasses Obtained from* The H, T-ll, and K-l Drill Holes As a Function of Distance From the Working Point. See Figures 5 and 6 for drill hole positions. Figure 27. FeO Content (Weight Percent) of Rainier Glasses Obtained from the H, T-ll, and K-l Drill Holes As a Function of Distance From the Working Point. Open symbols represent data obtained on colorless glass regions. Solid symbols indicate results of analyses of brown glass regions. See Figures 5 and 6 for drill hole positions.

340 I I- cr zo

FIGURE 1 INDEX MAP OF THE NEVADA TEST SITE SHOWING THE LOCATION OF THE UI2B RAINIER EVENT

341 NEVADA TEST SITE AREA 12 UI2B.UI2C TUNNEL ARE*

INDEX I GENERAL GEOLOGICAL MAP 100 0 200 «00

j GEOLOGY AND ENGINEERING 0AT« FAULT iL= ROAO .....•- TUNNEL '" BOUNDARY _/• TOPOGRAPHIC CONTOUR

6400 6300 u\

FIGURE 2

342 DESCRIPTION OF 3SDPIHGS

Tofg - Welded Cuff; rhyollce Co quartt latlte 7500- Bedded cuff; aoatly looaely cemented and ''tandy"; light gray Co graylth brown 7400- To»6 Welded Cuff; light gray to brovnlih gray Bedded tuff; well cemented; light yellow green To»4 - Bedded Cuff; veil cemented; light gray to buff, aoae pink 7300- Bedded tuff; veil cemented; red aC Cop and bate, pink Co buff lnterbedj Bedded Cuff; noatly light gray to buff 7200- Io$2 To* | Bedded tuff; purpllih Co plnklih red (Approximate top of weakly 7100- camenied gronular tuff , LtaeaCone; hard, deme, crystalline; medtua to dark gray

7000- Tos 6900'-

6800- Approximole base of weakly cemented granular tuff w 6700 ^ 6600'

6500''

6400' 6300'

6200'

6100

6000

5900- FIGURE 3 UI2B RAINER PLAN VlCW S i TWB1AS L TMQUP9M. (Hwyr-gftj MHN I WI3Z. Gto-«*# -1KL-H

LEGEND

16.00 SURVEY STATION NUMBERS (OST FWiJW POWTALJ

I'' STRIKE .« OP Of BEDS

FAULT SHOWING NATURE OF DISPLACEMENT

JOINTS

" >_-.. ..•...:• L

FIGURE 4

344 'I MC-SKIT BHU. MOLE U01.- UI2B RAINIER CROSS • SECTION A-A*

MATERIALS CORE MOLE ANALTSIS

FIGURE 5

34b ,L'TMOLOGY 1E ENGIhEER:NG Mt7 UI2B RAINIER CROSS SECTION B-B

. =~ * *•»»».£ '

/'•• * .....1

TO5?7 • V ,

FIGURE 6

346 1 •A Ji '

•*«

347 w Bi ED

348 ON w o

349 "'%\>%

»>VA •|:.4 < I ^V

w OS

350 PS op I

351 •fc •**

«-..;»

352 o

353 c o* w ea O

354 r—i

355 o M

356 OS o n

357 M

358 359 m 6

E o o3. o hi 3 o

360 361 362 o 10.0-13.0 * 23.0-25.0 25.0-26.0 26.0-27.0 * 27.0-28.5 * 28.5-30.0

RAINIER "T-11" HOLE 39.3-40.3 o 40.3-43.2 51.4-52.5 V \/ V M RAINIER "H" HOLE 19.5 o 32.8-33.0 • 33.0-33.2

RAINIER "K-1" HOLE

FIGURE 23 Qz

39.3-40.3 o 40.3-43.2

-P- RAINIER "H"HOLE V V V \/ \/ Or

° 6.5-9.3 10.0-13.0 23.0-25.0 o 25.0-26.0 19.5 « 26.0-27.0 o 32.8-33.0 • 27.0-28.5 • 33.0-33.2 RAINIER "T-11" HOLE RAINIER "K-1" HOLE FIGURE 24 100 ^ i ' r RAINIER GLASSES Hole 90 OT-11 Q H 5 80 0 K-1

O 70 55 I 60

50 i i 0 20 40 60 80 Distance from WP (Ft)

FIGURE 25 12 r Hole RAINIER GLASSES 10 O T-11 n H 0 K-1

1 o

I I I 20 40 60 80 Distance from WP (Ft)

FIGURE 26 RAINIER GLASSES

20 40 Distance from WP (Ft) FIGURE 2/

367 I'CRL-89406 PREPRINT

CALCULATION OF EARLY TIME SUBSIDENCE PHENOMENA (A CALCULATIONAL STUDY OF THE CARPETBAG EVENT)

J. T. Rambo D. E. Burton F. A. Morrison, Jr. R. W. Terhune

This paper was prepared for submittal to the Proceedings, Containment Symposium, Aug. 2-4, 1983, Albuquerque, NM.

August 1983

This is a preprint of a paper intended Tor publication in a journal or proceedings. Since changes may be made before publication, this preprint is made available with the understanding that it will not he cited or reproduced without the permission of the author.

369 CALCULATION OF EARLY TIME SUBSIDENCE PHENOMENA (A CALCULATIONAL STUDY OF THE CARPETBAG EVENT) J. T. Rambo, D. E. Burton, F. A. Morrison, Jr., and R. W. Terhune Lawrence Livermore National Laboratory Livermore, California 94550

Abstract We initially hypothesized that a nuclear detonation below the water table in a high porosity medium will preferentially compact the void space above and along the water table, which may result in a broad early-time subsidence at the surface. Ground motion measurements show rapid compaction above such events.^ Hydrodynamic calculations are presented, which provide some insight into these subsidence phenomena. The work is in progress and conclusions are still tentative. Introduction Our objective is to gain a better understanding of prompt dynamic subsidence by calculating a simulation of the Carpetbag event. Available data and a 2-D calculational tool (TENSOR)^ are used to help isolate parameters which affect the observed subsidence. The subsidence on Carpetbag appears to be composed of two types. One, a long term subsidence which has been measured over a period of years. At Carpetbag the pre-event value was approximately 1 m and may have been stimulated by nearby events. Two, a prompt subsidence which occurs in less than a few seconds of the event. For Carpetbag this portion of the subsidence reached a value of 3.7 m at SGZ. A previous calculational study^ used a compaction damage model as an approach to the subsidence problem. In this study a Pz scarp was modeled in the horizontal direction over which a reduction in the horizontal stress field was calculated. We have not yet treated this possibility in our calculations. We are initially using the calculations to sort out the possible causes of the dynamic nature of the prompt subsidence. In the initial subsidence work,^ it was suggested that modeling calculations of Carpetbag would be useful to better understand the subsidence phenomena. Our efforts are currently focused toward this objective. Subsidence Data Figure 1 shows a surface map of the Carpetbag area in which the data shown in brackets are from survey measurements between the years 1956 to the pre— Carpetbag date of November 1970. A maximum value of 1.3 m is observed near Carpetbag. This long term subsidence indicates that there are some very weak subsurface layers or that overall material in the region was very weak.

370 SUBSIDENCE DUE TO CARPETBAG M 3000 N • T O 0.043 [0.003]

00.052 2000 O 0.052

00.076

1000 O 0.146

-1000 © 0.061

0 0.067

[] Pre-shot subsidence -2000 - 1956 - 1970 @ 0.076 O Survey measurements El Estimates from gauges 0.052-^D Q-058 Cc 0.055^6 o0.052 -3000 J , 0.049 |[0.533]^->O -1000 1000 2000 3000 Distance - m

Figure 1

371 The circles are subsidence from survey measurements spanning the pre- to post-Carpetbag event. At the closest approach the value is a maximum of 0.57 m. The gage data at the trailer park and SGZ show 1.06 m and 3.66 m of subsidence, respectively. This data is replotted as subsidence versus radial range in Figure 2. The subsidence beyond 1200 m is minor (less than 0.1 m). Between 1200 m and SGZ the subsidence begins a noticeable increase. The survey data tends to merge into the gage data and describes the largest subsidence at SGZ.

The integration of velocity gages shewn in Fig. 3 describes the displacements vs. time at three locations. The stemming velocity gage was located approximately half way from the WP to surface and showed 0.72 m of subsidence. This gage was located in the porous stemming of the cased emplacement hole. While the stemming material is more compactible than the surrounding geology, this effect on prompt subsidence at the gage cannot presently be estimated. There have been other free field data from satellite holes showing subsidence deep in the hole. The SGZ, trailer park, and the stemming velocity gages all show that subsidence occurred within the time of slap-down. Calculational Model Carpetbag was executed in November 1970 just prior to Baneberry. Its yield was 220 kt with a shallow SDOB of 110 m/kt'/3. Because of the pre- Baneberry execution there were no modern standard geological data on material properties at this site. A modern nearby geological site was required to model U2dg. A recent cross section is shown in Fig. 4. The line of sections show that the geologies to the west are perturbed by Pz scarps. For this reason we chose to ignore events on tne west side and chose events in a general north- south direction as shown by arrows in the plan view. Two events, U2dr and L)2dz, which are shown on a recent crack map of the area (Fig. 5) have about the correct direction and were located north and south of U2dg, respectively. Similar layers were picked for both U2dr and U2dz based on similar breaks in the density and significant porosity changes such as the water table interface. Average properties were calculated for both sites with good agreement being noted for equivalent zonal averages between the two sites. The U2dr site was used for the zonal averages because it penetrated the water table. The selected average geophysical properties are shown in Fig. 6 for the U2dr site. The significant material properties are the gas porosity below and above the water-table which would focus the incident wave" and the high surface gas porosity from which a weakness in strength was implied.

The strength properties were reviewed from several sources. Recalling that this site exhibited subsidence pre-Carpetbag, we looked for any evidence that might be related in some way to weak strength properties. Four types of .displays vs. depth are shown side by side in Fig. 7 which relate to strength. The far right display indicates what we used in each layer for strength in our first 2-0 TENSOR calculation. Tne caliper log and drilling rate display tend to show a qualitative relation to strength where hole enlargements and high drilling rates may indicate weak regions in the hole. The general trend of

372 DYNAMIC GAGES INDICATED THE LARGEST

SUBSIDENCE TO BE NEAR SGZ.

u _ . . - -

• - • : • •' • • ' -O- SOUTH SURVEY DATA u

\ O Q NORTH SURVEY DATA 2.0 —- SURFACE VELOCITY GAGES

3.0 -

CO 3 o

5.0 ! I. I I 0 1000 2000 3000 4000 S000 RANGE FROM SGZ - M

Figure 2 DISPLACEMENT - M DISPLACEMENT - M

ro co .£» Oi .,._ co CO C73 en TO r~j 1 > ! 3s •< , m m en "" i— —C/1l O CO 3 o 1— — t-H O3 —1 CO 73 -< »^ o o 2 ~ CD zz o CO £73 m G> — m o i^i o o cr> cPOr PO a DISPLACEMENT - M

m3:

CO o T| CO oa o o CT.L

co sze

m o 73 O V5

CO m o

m a

\\ \\ o m ? c;

o -a po

CO m o

33 O

i O

a: o HOLES U2dr AND U2dz WERE THE MOST APPROPRIATE GEOLOGIES TO USE AS A CARPETBAG MODEL.

U2dr

N 265,000 •

U2dz

N 264,000..

E 203,000

E205OO0

1OOO WEtHES

Figure 5

376 ZONAL AVERAGE PROPERTIES USED IN CARPETBAG MODEL.

MATERIAL PROPERTIES HOLE_JU_2dlL p p o H20 g ^ z* ^ 3 3 Mg/m Wt% Mg/m VQ1% VQ1% yon

1 .93 8.7 2.63 33 16 16

2112 9.3 2.63 27 73

1.94 13.8 2.96 35 77 8

2.01 12.7 2.61 33 77 7

1..74 19.2 2.47 43 77 10

1 .90 21.7 2.53 41 100 0

4> is total porosity ui \txxl% I is saturation in vol% 4> is gas porosity in vo1% Figure 6 STRENGTH PROPERTIES WERE DERIVED FROM VELOCITY GAGES, CAVITY RADIUS, AND JUDGMENT

STRENGTHS FROM CAVITY RADIUS DRILLING RATE .... M/HR DIAMETER .... H ELASTIC STRENGTH MEASUREMENTS FOR STRENGTHS FROM STEMMING EVENTS WITHIN (Tau) USED o 1 Z 3 4- 5 N> O GAGES (PT) 3000 m of U2dg IN TENSOR

...i.. - ':-• I I I I I __ |"j • 0.5 MPa

1 i

4 '

—c 5.5 MPa 1.1 MPa •U tl I)::. • - •

—'•• 385 m O ...X m 346 m f >1-5MPj w F 4.7 MPa 3.5 MPa : J... 300 m > 5.8 MPc 00 •ti •y- r- . 1 • i > 7 MPa '.I • •- : i'r 215 m 5 MPa 3.5 MPa 1 • j ._. - . ;r~ > 15 MPa ..I. 7.8 MPa 11.2 MPa SWL SWL 86 m SWL_ ft" — i -- ri. I 10 MPa 60.0 MPa i WP IWWP cREGION 60 MPa ; I I I I U2DG CALIPER o 4 8 RUN 1 7-2-69 x MPa Fiaure 7 these two logs show some signs of weaker material at the surface and near the bottom in the WP region. The drilling rate log also shows a small region that is weak between 465 and 475 m depth. The very high drilling rate near the bottom is of uncertain interpretation because the drill bit cutters and rollers were changed at the location (595 m) where tne drilling rate increased significantly. The lower cavity radius measurement further confuses the strength of the working point region with ai anomalously low value of 56 m. This is equivalent in calculational strength to 60 MPa7 in which a calculated cavity of the same size would be produced in the lower direction. A more normal cavity radius measurement in the lower direction would require a 30 m error in the survey. The drilling rates indicate that stronger material (lower drilling rates) was present on either side of the weak region. One could speculate that this combination of high and low strengths could produce an anomalously shaped cavity.

Other strength data were estimated from the elastic peaks of particle velocity gages at 215, 300, 346, and 385 m range. Only two of these gages (215 and 346 in) survived past the elastic portion of the signal. Our estimates of the strengths were based on particle velocity and elastic sound speed with some verification from 1-D calculations. The region from WP to 215 m requires elastic strengths 2: 15 MPa. From 215 to 300 m it is >_1 MPa and from 300 to 346 m the value was >. 5.8 MPa. The value from ?46 to 385 m was a low value of 1.5 MPa and may be in error because of being isolated by the cased hole in the low stress region. Strengths based on cavity radius measurements' of nearby holes (3000 m square) of the U2dg were averaged over the selected layers. These strengths show a g^n-jral increase toward greater depths and generally correlate with increasing strength observed from the velocity gages. Our initial calculational strengths are shown in the last column of Figure 7. The calculation which we call the base case emphasized a strong WP region (60 MPa) which is probably unrealistically strong. However, it was used to match the lower cavity radius measurement as a basis for comparison to other sensitivity calculations. The strengths in the upper portion of the hole were set to low strength values to emphasize the crush up in the upper regions. The base case calculation was set up to maximize the 'crush up of the surface material as a possible important subsidence phenomenon. Tne surface material (16% gas porosity) was arbitrarily set to a possible low strength of 0.5 MPa. Calculational Results The comparison between data and the initial calculdtion is shown in Fig. 8. The available velocity signals are at the left and displacement from tne integrated velocities are to the right. The lowest gage showed a slightly reduced peak arriving ahead of the data. The lower material may have been slightly more porous than modeled. The 215 m gage ceased to operate past 0.45 sec. and thus a displacement plot was omitted.

379 RESULTS OF THE INITIAL TENSOR CALCULATION SHOWED SOME SUBSIDENCE, BUT NOT IN A SIMILAR TIME FRAME.

PARTICLE VELOCITY - M/SEC DISPLACEMENT - M

rk7*1 fDATA c -2 -4 -6

TIME - go Till TIME - S 40 215 M H v, 30 £20 '•%• 10

0 . _^^ATA -)0 -20 J I I 0 .1 .2 .3 .4 .5 TIME - S

Figure 8

380 The calculated velocity peak was slightly higher at the 346 m gage with the displacement peak being close to the data peak. The final displacement does not capture the prompt subsidence observed at this depth. Tne calculated SGZ velocity shows reasonable agreement until the free fall terminates at ^2 sec. Just after the peaK there is a portion of the signal at about 1 second that is flat. This portion of the velocity wave was not present in our 1-D calculations and is attributed to focusing of the plastic portion of the wave from the water table. The fit to the initial portion of the velocity wave is reasonably good considering the uncertainties in the data. However, the extent and depth of subsidence as evidenced from the displacement data was not calculationally captured. The calculated displacement at SGZ did, however, show some subsidence had occurred and that it was continuing in a downward direction with time. In order to ascertain where the downward drift in subsidence was coming from the calculation was continued to 10 seconds after thoroughly dezoning the problem to reduce calculational cost. In Fig. 9 we show a comparison in displacement vs. range between the calculations at 2.5 sec. and 10 sec. The displacement for both times reduces to zero by 570 m range and into subsidence above and to the surface. For the 10 second time the displacement decreased by an additional 1/2 metre at about 200 m range and further decreased by a metre at the near surface locations. The two subsidence data points indicate a slope not too different from the calculation. However, they tend to lead one to the conclusion that the bulk of the subsidence occurred from a phenomenon deep in the formation.

Sensitivity Calculations Our intent on running two variations to the base case calculation was to investigate the effects of weak zones near the working point as suggested by the high drilling rates. One variation was to include a weak porous layer at 190-200 m range. A single physical property sample at this elevation from hole U2dv showed 36% gas porosity. As this weak zone may not have existed in reality, our primary justification for this calculation was to better understand the possible effect on subsidence from a very porous layer. Tensile strength in the working point region was also included in this calculation to observe any sensitivity. Figure 10 shows the geological layout of the problem to the left (CAR07) and th^ displacement results to the right compared to the base case. The dotted line on the displacement vs. range relationship indicates that tensile strength may have reduced the displacement substantially, even though the material was failed by the outgoing wave (no tensile strength after failure). At 190 m the incident wave entered the porous region sending back a rarefaction wave toward the cavity. As the shock wave moved into this region the material from below spalled into the layer. This is apparent from the sudden decrease in displacement across the layer. How- ever, not much of the reduced displacement occurs at the surface. The outgoing stress wave is severely attenuated in the porous layer and is unable to crush up the material toward the surface. This is shown by the different slopes between the base case (CAR04) and the porous layer case (CAR07). Generally, porous weak layers near the cavity region allow greater cavity growth into them

381 THE LOWER GAGE INDICATES SUBSIDENCE OCCURRED BELOW 346 m.

CAR04 CYCLE 7773. TIME 2.50+00 m ~i—i—i—r—r—T—i—i—i—i—i—i—i—r" m T

CAVITY

O--

m --

1 o I

N c>l

10 SEC,

2 « X I X10'

Figure 9

382 CAR07 CALCULATION INCORPORATED A WEAK LAYER AND TENSILE STRENGTH IN THE WORKING POINT REGION.

CARD* CAR07 CYCLE 7773. TIME 2.50+00 CYCLE 14*23. TIME 2.50+00 MATERIAL I ' ' •'

TWEAK 10 M LAYER (_6AS POROSITY =

x I -L, 0123*5678 ,2 xio R AXIS

XIO

Figure 10 resulting in greater positive displacement between cavity and layer. The crusn up is not effectively passed on to the surface as a subsidence. We have not yet sorted out the tensile strength effects of this calculation. In our second sensitivity calculation a portion of the working point region was weakened to conform to a possible weakness observed in the drilling. The calculation was intended to flatten the top of the cavity so that the cavity rebound would be large because the radial convergence is reduced. Figure 11 shows tne geology, representation after full cavity growtn. Some of the desired result was seen in the displacement below the 140 m range. How- ever, variations in materials near the working point resulted in significant changes to the shape and size of the particle velocity wave near the surface. The great sensitivity of the surface velocity and displacements to working point properties is caused by the focusing across fie water table.

The deep subsidence phenomenon may indeed have come from the cavity region. Our single calculation merely indicates the complex effects that worKing point material properties have on the surface displacement. The cavity region remains the source of unresolved phenomena. The early cavity collapse to the surface (15 minutes), the small cavity radius in the lower direction, the high drilling rates in the cavity region and data showing deep and early subsidence all indicate an abnormal cavity. Future calculations are required to better understand near cavity phenomena. Surface Phenomena That portion of the total prompt subsidence which we calculated was mostly due to a surface compaction. Reviewing the base case calculation as shown in Fig. 10 we note that the incident stress at the base of the weak surface alluvium (0.5 MPa strength) was a relatively high value of 3 MPa. This caused a portion of the surface layer to cr'usnup and was followed by second higher stress due to slapdown crushup. The higher second stress pulse due to slapdown is shown by the dashed line in the peak mean stress plot of Fig. 12. The actual SGZ data accelerations which are related to stress indicate slapdown accelerations which are twice the value of the incident acceleration. The calculated compression versus range relationship is shown below the peak stress curve. Most of the compression in the upper one half of the site occurs in the surface layer. Had the free fall of the calculation been extended to the time observed in the data, then greater slapdown stress would have furtner compressed the surface alluvium. The calculated surface subsidence versus range from SGZ is compared to tne data in Fig. 13. The deeper subsidence is calculated between SGZ and 200 in range. There is some similarity of a general nature between the calculation and the data. However, the calculation shows subsidence in a more irregular fashion along the surface. THE CAR08 CALCULATION CONTAINED A WEAK LAYER IN THE CAVITY REGION.

CARCH CYCLE 7773. TIWE 2.50+00 CARO8 ' 1I '—• <—• r• • I CYCLE 8789. TIME 2.50+00 MATERIAL 51.

CAVITY

FR:E SURFACE

x T \, , , , i , , , , i ,,,,[,,, M GAGE X10 _i l • ' • • -i—I i | i i i i ( i—i i—i ]—1_ G i 3 4

X10 RA/^GE - M

Figure 11 A PORTION OF THE SUBSIDENCE MAY HAVE COME FROM THE SURFACE LAYER.

PEAK MEAN STRESS 100, i r T HIGH INCIDENT STRESS COMPRESSED THE SURFACE LAYER, ACCELERATION DATA

CPSlC-S-3- I 5iV SLAPDOWN .STRESS 10 a.

3 MPA O o

-4 1 2 SURFACE time - sec Acceleration data indicates SURFACE STRENGTH = 0.5 siapdown force (stress) could be a factor of two 0.1 I I I I higher than the incident 100 200 300 400 500 600 700 stress.

1 Range - m r r i ' 1 •• i • " T~

FREE SURFACE

SURFACE LAYER

IS) to 2 o. O I "

. . . | t I 1 L_J 1 1 1 1 1 L. 5 S X10 RANGE - M 386 Figure 12 SURFACE SUBSIDENCE COMPARISONS

CAR04 CYCLE 22145. TIME 1.00+01 o . T CALCULATION AT 10 SEC, SGZ

CALCULATED MAXIMUM POSSIBLE SUBSIDENCE

14 16 X1O

Figure 13

387 The lower curve represents a maximum possible subsidence which would occur if all the stresses were returned to over.burden and the resulting volumes were accumulated for each zone down to a depth of the Pz interface 340 m below the working point. The summation process accounts for the specific maximum unloading compression relationship of each zone. The upward bend in the curve between 200 m and SGZ is related to the cavity. This maximum subsidence would not be realized except over geologic time. Surface cracking was not an intent of our calculations, however it is interesting to note similarities between observed surface radial cracking and calculated surface cracking. Figure 14 shows displacement vectors and tensile stress. Figure 15 shows the surface crack map. The displacement vectors at the surface indicate outward radial displacements exceeding two metres which extend out to ^00 in range. This radial displacement would put the surface material into tension. The stress plot of T3, that is, the stress out of the plain of the paper shows tension stress was initiated along the surface out to about 600 m range. The cracking is of a shallow nature occurring down to about the depth of one zone (10 m). By 40 m deptn the radial displacement vectors are greatly reduced. The 600 m circle as shown on tne crack map tends to be the outward boundary of the surface radial cracks and is in reasonable agreement with tne calculation. Some concentric cracking was also noted in the calculations between 200 and 500 m surface range. However, the extensive concentric cracks outside the 600 m circle were not calculated and may be due to other long term subsidence phenomena. Tentative Conclusions Our analysis of the data and the calculations to date suggest that prompt subsidence for this site was from two regions. Surface compaction from a high incident stress and a nigh recompaction stress was calculated. Some contributing factors to this compaction are: a somewhat smaller cavity in the upward direction (small displacement) because of high strength; focusing from the water table which in turn increases the surface stresses, peak surface displacements and the recompaction stress on slapdown; and finally, the modeling of a weak compressible surface material on which the higher stresses act. The deep subsidence from near the cavity region was not successfully calculated, however, tnis region is of primary interest because of anomalous cavity related data and our suspicion that the deep subsidence may have started nere. Some fu'ure work may involve further calculational studies in which we will attempt to sort out the effects of various strength combinations in the cavity region, tensile strength, and possibly some surface strength variations. Some consideration will be given to a rate dependent pore collapse model if it is essential for prompt subsidence.

3&S THE CALCULATIONS SHOWED SURFACE RADIAL CRACKING EXTENDED OUT TO 600 M

VCCtOR RH ZH O 2 RZ iO u RADIAL SURFACE DISPLACEMENT > 2 M CARO4 CYCLE 22145. 7IME VLC1ORS RH- 2H -,-r-y- ^ ' I ' ' ' '. 2.00D+O0 1.750+00

1 .500+00

.i i i t « « i i t f t • 1.25D+00 i i t i i • t ! ! I r > . _iit»»»i 1 .000+00 1 1 1 I ! I > n-TTT-T 7*7 7.50O-01 u t I t I t I ! I >*> .Ill MM 5.000-01

2.500-01

. i »

CM O

0 1 2 3^ 5 6 7 a 9' 10 X1O H AXIS M CON 13 O X SHADE M 600 M CARO4 CYCLE 221«. TIUE 1 00+01 T3

RADIAL, TENSION CRACKS

o Figure 14 O 1 xio2 R AX IS POST SHOT SURFACE CRACKS*

600m circle

500

N 870 000

UEdc-2

1000 2000 Feel 1 1 1 300 600 Meters

*USGS Report USGS-474-160 Surface Fracturing of the Carpetbag Event, Yucca Flat, NTS, Nevada, E.Jenkins.

Figure 15

390 ACKNOWLEDGMENTS We wish to acknowledge the work of J. Wagoner for geological discussions and for the cross section; to EG&G personnel at LLNL for processing of average properties in the area near U2dg; to V. Wheeler, LLNL, for providing Carpetbag velocity gage data; and to Miriam Lohmann for providing production support. References 1. Private communication, V. Wheeler, February 1983. 2. Burton, D.E., L. A. Lettis, Jr., J. B. Bryan, and N. R. Frary, "Physics and Numerics of the TENSOR Code [Incomplete Preliminary Documentation]," Lawrence Livermore National Laboratory Report UCID-19428, July 15, 1982. 3. Allen, R.T., "Differential Compaction," Proceedings of the Monterey Containment Symposium, Monterey, California, Aug. 26-28, 1981, LA-9211-C, Vol. II. 4. Holmes & Narver, Inc., Internal Memorandum, Area 2 Bench Mark Elevations: Pre & Post Carpetbag, July 8, 1971. 5. Private communication with J. Wagoner, May 1983. 6. Rambo, J.T., and J. B. Bryan, "Calculation of High Surface Velocity Due to Focusing on the Tyoo Event", Lawrence Livermore National Laboratory Report UCRL-89405, Preprint, Aug. 1983. Proceedings of the 2nd Containment Symposium, Albuquerque, NM, Aug. 2-4, 1983. 7. Terhune, R.W., and H. D. Glenn, "Estimate of Earth Media Shear Strength at the Nevada Test Site", Lawrence Livermore National Laboratory Report UCRL-52358, November 3, 1977.

391 DISC I.AI\II:R

This document »as prepared as an account of work sponv>rcd bv an agency at the I niled Stales (.oUTiinicnl. Neither the I niled Stales (iovernmenl nor the I niversity of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, pnxluct. or process disclosed, or represents that its use Mould not infringe privatelv owned rights. Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does nol necessarily constitute or imply its endorsement, recommendation, or favoring by the I niict] States fi'otcrtiment or the I niversity of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the I niled States (iovernmenl thereof, and shall not be used for advertising or product endorsement purposes.

392 THE PG-2 PHOTOGRAMMETRIC PLOTTER: A RAPID AND ACCURATE MEANS OF NAPPING SURFACE EFFECTS PRODUCED BY SUBSURFACE NUCLEAR TESTING AT THE NEVADA TEST SITE, NEVADA Van de Werken, Martha Garcia, U.S. Geological Survey, Box 25046, Federal Center, Denver, Co 80225

ABSTRACT

Accurate documentation of surface effects is important for site selection and structural studies. Past and present practice has been to map surface effects in the field shortly after detonation using pre-test aerial photographs. Because this method is time consuming and prone to error, an alternative method to field mapping was developed. Since October 1981, the U.S. Geological Survey has been using the Kern PG-2 photogrammetric plotter to map surface effects using post-test aerial photographs. The main goal of this pilot program was to compare the two mapping methods and to determine if field observations are necessary. Preliminary results indicate that only questionable sma>l-scale features need to be field checked. Mapping on the plotter is highly reliable if aerial photographs obtained immediately after detonation are used. If photography is delayed, surface effects may be obliterated by natural processes and construction activities. Disadvantages to the plotter method relate to the quality and coverage of aerial photographs. The main problem concerns the scale of aerial photographs. Because of the large scale, the photographs lack adequate control points to properly orient the photographs to a map base. In addition, the paper print photographs used were often distorted. Once the problems were

393 recognized and corrected, the method was greatly improved. Generally, the PG- 2 offers a precise method for determining the distribution of surface effects.

INTRODUCTION Surface effects in themselves make interesting studies, but more important to understanding the mechanics of containment, surface effects may reveal bu *ied active geologic structures and contribute to the understanding of aberrant surface collapse behavior. Accurate documentation of surface effects is recognized as an important and necessary activity for site selection and structural studies. However, field mapping of surface effects is often tedious and time consuming and has been done with differing degrees of quality over the past years. Because of the quality inconsistencies, a faster, more reliable method was sought. The U.S. Geological Survey has developed a method for mapping surface effects utilizing Kern PG-2 photogrammetric plotters. The PG-2 plotter was chosen because it provides a means of plotting, data on a base map accurately and offers a high quality viewing system for stereoscopic inspection. Although the PG-2 plotter was designed for topographic mapping, it has been successfully adapted for surface effects studies. The PG-2 allows mapping of effects from post-test photographs more thoroughly and faster than does the method of walking over the ground.

FORMATION OF SURFACE EFFECTS Since underground nuclear testing began in 1957 at the Nevada Test Site (NTS), more than 400 announced tests have been conducted. Surface effects produced from the tests are sinks, fractures, and fault movement. Surface effect formation is dependent on the yield of the explosion, the depth of

394 burial, and the test medium. Yields at the test site range from less than 1 kiloton to 150 kilotons, and tests are conducted in alluvium, tuff, basalt, or granite. Nuclear devices are commonly emplaced in vertical drill holes at depths of 180 to 760 m (Figure 1). The drill holes are filled and tamped following emplacement and the device detonated (Houser, 1968). The energy of the explosion is released in about a tenth of a micro-second, the temperature is increased to several million degrees kelvin, and the pressure to many kilobars. Everything around the device is immediately vaporized and melted. The initial cavity expands and the neighboring alluvium begins to act hydro- dynamically. Shock energy begins to propagate in all directions and creates seismic waves. A slight bulge occurs at the ground surface and fractures begin to form (Figure 2a). Immediately after the cavity ceases to expand the molten rock drains downward to form a puddle and ground surface returns to a position near or above its pre-test level (Houser, 1970). At a later time, ranging from several seconds to hours later, the explosion cavity collapses to form a chimney (Figure 2b). If the test is in very deep alluvium or moderately deep in competent rock the chimney will not reach ground surface. But if the test is at a moderate depth in alluvium or at a moderate to shallow depth in competent rock, the collapse will continue to the surface and form a sink (Figure 2c). The sinks of Yucca Flat, appear as an agglomerate of craters and depressions. Aerial photographs of areas of intensive testing slightly resemble a cratered lunar surface (Figure 3). Few sinks are perfect circles in plan view; most are slightly oval or elliptical in shape and have scalloped rims. The sinks in Yucca Flat vary from 3 to 244 m in radius, and from 0.6 to 61 m in depth. The larger sinks represent volumes approaching 6.5 x 10^ m^ (Houser, 1968).

395 INSTRUMENT HOLES

ALLUVIUM

WELDED ASH-FLOW TUFF

BEDDED ASH-FALL TUFF

ZEOLITIZED TUFF

CANISTER CONTAINING DEVICE AND INSTRUMENTS

Static Water Level

ELEV. 730 m

PALEOZOIC ROCKS

Figure 1.—Example of drill-hole emplacement of a nuclear test in Yucca Flat, Nevada'fest Site.

396 SURFACE EXPANSION s. POST-TEST SURFACE MUCH EXAGGERATED DEPRESSION C. TOTAL AREA OF SUBSIDENCE ZONE

VISIBLE SINK J

SURFACE CRACKS FORMED BY EXPANSION PRE-TEST SURFACE COLLAPSE A-/CHIMNEY —I

ZERO POINT —v\

SHOCK FRACTURES/^l l\ ^.FUSED ROCK PUDDLE OF MELTED ROCK •- —J ""*- CAVITY X> ^y pVhDLF OF /^^\7\3«EiT£D BOCK 1 V\ ffl>*D(OXCT;VE / ^ | \ Gl/lSSJ

Figure 2.--Formation of a sink, cross-section view (modified from Houser, 1970). A.—Generalized early reaction of rock medium to a nuclear explosion. B.—Generalized collapse of explosion cavity and formation of chimney. C—Generalized collapse of surface chimney. No scale. Figure 3.—Aerial photograph of Yucca Flat, Nevada, showing craters produced by underground nuclear explosions.

3S-8 Fractures occur in the alluvium overlying the site of an explosion regardless of the formation of a sink (Figure 4). Fractures not only form in radial and concentric patterns, but are also commonly a lined along certain preferential directions. These preferentially alined fracture trends form two groups: those that show a spatial relationship to the Yucca Fault and other related faults, and those that do not. The fracture trends unrelated to the Yucca Fault parallel joint trends observed in the bedrock at the adjacent edge of the valley; the most conspicuous preferential fracture trends in the alluvium appear to parallel closely spaced fractures in the bedrock. Fractures generated by explosions detonated in tuff beneath Yucca Flat show a greater degree of preferred orientation than those caused by explosions detonated in alluvium (Barosh, 1968).

METHODS OF STUDY Field Methods The U.S. Geological Survey (USGS) began mapping the surface effects of underground nuclear testing in the late 1960's. At the present time (1983), Los Alamos National Laboratory and Lawrence Livermore National Laboratory are responsible for mapping surface effects from their respective events. Effects of tests are mapped in the field on pre-test photos shortly after detonation, Pre-test photos rather than post-test photos are used at this stage because the processing time for post-test photos is commonly 2 weeks. If mapping is delayed, surface effects may be obliterated by weather or construction activities. Once the fractures have been mapped in the field, they are compiled at a reduced scale onto a composite surface effects map using a Kail plam'metric plotter.

399 £ 208,000 I

- N 266,000

500 1000 I METERS

Figure 4.—Fractures produced above an underground nuclear explosion, Yucca Flat, Nevada. Mapped by T. L. Prather, USGS.

400 This method is effective but involves time, scale, and accessibility problems. First„ the time problem is related to a ground control problem. Field mapping on photos and compiling the data accurately are difficult at the NTS, because the area is constantly undergoing change: within a week roads, vegetation, drilling pads, or trailers may appear or disappear, making it difficult to locate field stations. Second, the scale of pre- and post-test photography is 'arge, usually 1:2400 or 1:4800, providing inadequate control points to properly annotate photographs. Without adequate control the accurate annotation of surface effects is questionable. Third, not all of the affected area is accessible for mapping. Due to safety reasons the area within the perimeter fence around the collapsed area is usually left unmapped and dimensions of craters that form must be mechanically surveyed. Secause the method described above is time consuming and in many ways unsatisfactory, an alternative method has been developed. Since October 1981, che USGS has been using the Kern PG-? photogrammetric plotter (Figure 5) to map surface effects using post-test photos. The main goal of the pilot program was to compare the two methods and to determine if continues field mapping is necessary.

Photogrammetric Method The Kern PG-2 photogrammetric plotter is capable of reconstructing a nearly distortion free stereoscopic model of the terrain from which data can be transferred to maps. Aerial photographs are mounted on two plates that can be moved and rotated sc that the photograghs effectively recapture the orientation of the camera in the airplane at the time of exposure. The chotography is observed orthogonally, such that the projecting lines are perpendicular to the plane of projection.

401 The PG-2 plotter is designed to accommodate glass or film transparencies as well as paper prints. White light is used for illumination, allowing color photographs to be used as well as black and white. A high-qua'iity viewing system incorporating a variable magnification that can be readily changed is available, as well as an illuminated floating mark that can easily be seen in dark portions of the image. Not only does the PG-2 plotter provide a means of accurately plotting data on a base map, but it also offers an excellent viewing system for stereoscopic inspection. A plotting table (Figure 5), connected to the viewing instrument, is situated within easy reach of the operator. When enlargement or reduction is required, the horizontal scanning motions are transferred to the drawing table by means of a stable polar pantograph, and by means of an arm when plotting at map scale only. These features together with the compactness of the instrument insure an ease and speed of operation that permits geologists who are relatively untrained in photogrammetric operations to use the machine for geologic observation after the photographs have been oriented (Pillmore, 1979). Though designed for topographic mapping, the PG-2 plotter has been successfully used for surface-effects mapping. The method is successful primarily because the photos that were used in the field can be set into the plotter, and additional field data can be observed directly in the sterao- scopic model and then plotted onto a map. Several additional advantages of the PG-2 method are: 1) fracture mapping can be done comfortably in a well- lighted r om; 2) a larger area can be checked for effects within a limited amount of time; and 3) inaccessible sinks and fractures enclosed by the perimeter fence can be accurately mapped and measured.

402 Figure 5.—The Kern PG-2 photogrammetric plotter.

403 In the pilot study, quality and coverage of aerial photograghs presented disadvantages to the photogrammetric method. As previously stated, photographic coverage was at a large scale; thus, stereoscopic viewing of the total disturbed area was difficult. Also, a sufficient number of control points to properly orient the photographs to a map base were not provided. In addition, distortions along the borders of paper prints cause problems especially in areas of high relief. Figure 6 illustrates the problem caused by insufficient photo coverage. The fractures shown in gray were mapped on the P6-2 and the other fractures shown in black were mapped in the field. Because of the relatively large sink and very large scale of the photos (1:2400), the coverage was limited to the area of the sink and post-test photos were virtually useless for mapping the outlying area surrounding the sink. After the control, scale, and distortion problems were recognized, steps were taken to correct them. First, plans were initiated to place control points east and north of the event site to provide at least two control points in addition to surface ground zero. Second, scale-stable positive film prints were used, which greatly improved the quality image compared to paper prints and eliminated distortion problems. The scale problem could not be easily solved. Large-scale photographs are needed because they offer high resolution of the surface effects; whereas, resolution decreases when smaller scale photographs are used. Perhaps the best solution would be to provide extended aerial coverage for events where extensive surface effects are expected. Figure 7 demonstrates the accuracy of the PG-2 plotter. The field mapped results are shown in black and plotter mapped results are in gray. Viewing the results together illustrates the effectiveness of the PG-2 to map nearly all of the field-detected effects plus additional features. The plotter also detected the presence of a northwest-southeast-trending scarp. Although part

404 E 206 000 m E 208 000 m I

N 260000 mH

N259000mH

Figure 6.—Effects from an underground nuclear test mapped in the field (in black) and on the PG-2 plotter (in gray). The area enclosed indicates the extent of stereoscopic photo coverage. E209.500 E2I0.000 1 I

- N260.000

T 1 N V \ •< A

ik I!

\ — N239.000

Figure 7.—Effects from &n underground nuclear test mapped in the field (in black) and on the PG-2 plotter (in gray).

406 of the scarp trace was undetected on the plotter, a simple field check would be able to verify the additional scarp trace. Another example is shown on Figure 8. Collapse to the surface occured after post-test photography was flown. Thus, the area prior to collapse could be mapped on the plotter (Figure 8a). After collapse occurred, the resulting sink was mapped in the field and again on the plotter (Figure 8b). The fractures shown on Figure 8b appear to be radials because they are roughly perpendicular to the sink; however, comparison with the PG-2 mapped results on Figure 8a shows the fractures to be northeast-southwest linears that reflect the stress field in the area, placing greater significance on the pre-collapse surface features.

DISCUSSION The purpose of the pilot program was to determine the feasibility of using PG-2 photogrammetric plotters for surface-effects studies. Results indicate that only questionable small-scale features identified by the photogrammetric method need to be field checked. The plotter method is more thorough and faster than the former method of field mapping when quality photographic coverage is available and control points are pre-determined. The greatest advantage to the photogrammetric method is that the PG-2 plotter provides a more accurate method of transferring surface effects data from photographs than the previous field method. On the basis of the pilot study, the USGS, in conjunction with Los Alamos and Lawrence Livermore National Laboratories, has developed the following procedures for mapping surface effects following underground nuclear testing. Pre-test aerial photographs are stereoscopically reviewed to determine areas of past and possible future fracturing. After photocopies of annotated,

407 E206,000

- N258.5OO

- N258,OOO

Figure 8a.--Effects from an underground nuclear test mapped on the PG-2 plotter.

408 E206.000

- N 258,500

- N258.OOO

Figure 8b.--Effects from an underground nuclear test mapped in the field (in black) and on the PG-2 plotter (in gray).

409 field-mapped photographs are' prepared by *the laboratories and post-test aerial photographs are received by the USGS, surface effects are mapped on the PG-2 plotter. Any questionable features identified on the PG-2 plotter are field checked by USGS personnel. Plotter results are modified as necessary after the field check and final results are compiled for transmittal to the laboratories. These procedures would be more efficient if plotter mapping was done first, in order to provide a gr.ide for the field check. This would reduce field time and the attendant cost for surface-effects mapping.

REFERENCES

Barosh, P. J., 1968, Relationship of explosion-produced fracture patterns to

geologic structure in Yucca Flat, Nevada Test Site, _vn_ Nevada Test Site:

Geologic Society of America Memoir 110, p. 199-217. Houser, F. N., 1968, Application of geology to underground nuclear explosions,

Nevada Test Site, vn_ Nevada Test Site: Geologic Society of America Memoir

110, p. 21-33.

Houser, F. N., 1970, A summary of information and ideas regarding sinks and

collapse, Nevada Test Site: U.S. Geologic Survey Report USGS-474-41, 30p.

Pillmore, C. L., 1979, The history and function of the U.S. Geological Survey

Photogrammetric Laboratory for Geologic Studies: American Society of

Photogrammetry Annual Meeting, 45th, March 1979, Proceedings, p. 465-468.

410 CONTAINMENT ANALYSIS FOR THE QUESO NUCLEAR EVENT GLENN, H.D., STUBBS, T. F.*, KALINOWSKI, J.A.*, and WOODWARD, E.C. Lawrence Livermore National Laboratory, Livermore, CA 94550 QUESO was a low yield event detonated at a depth of 216 m in hole UlObf at the Nevada Test Site on August 11, 1982. Radiation and pressure measurements in the emplacement hole indicated radioactivity was contained in the cavity region until a subsurface subsidence occurred at 25.55 minutes after the nuclear detonation. Then radioactive gases slowly migrated up the emplacement hole but were contained below the second lowest plug. Peak shock and residual stress measurements were obtained in the emplacement and satellite holes. Peak shock stresses in the emplacement hole were a factor of three lower than in the satellite hole. The difference is attributed to attenuation of the shock wave by the low impedance stemming material used to backfill the emplacement hole. The tangential stress measurements indicated a strong residual stress field was produced above the cavity by 200 ms following rebound of that region. Acceleration and velocity profiles for three stations, at comparable ranges from the nuclear explosion, in the emplacement and satellite holes were in excellent agreement. A motion canister bolted to the top of the surface casing exhibited damped motion in comparison to a free-field measurements for the surface station 15 m from surface-ground- zero. The strong residual stress field above the cavity in conjunction with the cable fanout and cable gas blocks in the lowest fines layer are considered responsible for the excellent containment results that were achieved. * Employees of EG&G, Inc., San Ramon, CA

411 1. introduction QUESO was a low yield event conducted in hole UiObf at 0800 hours on August 11, 1982. Jhe 216.4 m depth of burial (DOB) was considered conservative from a containment standpoint. The measured and calculated lithologic zonal average properties for QUESO are shown^ in Table 1. The device canister was located in a layer of tuffaceous alluvium. The region above the working point (WP) from 207 m depth to the surface consisted of mixed alluvium. Within this region all of the experimental measurements contained in this report were taken. This report will describe experimental results from acceleration, velocity, and stress sensors located in the emplacement hole (UlObf) and the satellite hole (UElObf) 15.24 m to the west of UlObf. The design of the stemming plan and performance diagnostics fielded for the QUESO event is reviewed in Section 2. The containment analysis based on experimental results is presented in Section 3. Section 4 presents a brief summary and conclusions.

2. QUESO Stemming Plan and Performance Diagnostics Figure 1 shows the QUESO (as-built) stemming plan. The stemming plan is in basic agreement with that from the QUESO ProspectusJ Coarse and fines layers and four coal-tar-epoxy (CTE) plugs were located to prevent the flow of cavity gases up the emplacement hole. The three lower CTE plugs were keyed to the formation via hole enlargements. A semi liquid plug was emplaced above but in contact with the bottom and top CTE plugs to seal any holes in or at their perimeters as a further inhibitor to gas flow. Impedance to the flow of radioactive gases up the cables or cable bundles was provided by three sets of cable gas blocks and six cable fanouts. In particular, a set of cable gas blocks and a cable bundle fanout were located in the lowest fines layer where experience2 indicated the residual stress field should be the strongest. The objective was to restrain the flow of cavity gas as deep in the hole as possible to avoid challenges to containment systems closer to the surface. Five radiation ana pressure canisters (Stations 31-35) were emplaced and supported by a cable pendant (Fig. 1) to monitor the possible flow of cavity gases up the emplacement hole. The lowest station (31) was at a depth of 136 m in the LLL coarse fill below the lowest CTE plug but above the lowest two fines layers. The second station (32) was at a depth of 121,6 m in an approximately 3 in thick layer of LLL coarse fill that was sandwiched between a 15 m layer of LLL stemming mix and a 2 m thick semi liquid plug. The semi liquid plug was above but in contact with the lowest CTE plug. The object of .Station 32 was to provide an evaluation of the CTE and semiliquid plug combination in the eventuality tnis system was challenged. The remaining three stations (33-35) were emplaced to provide information on the effectiveness of cable fanouts, fines layers, and the middle two CTE plugs (Fig. 1). Piezoresistant stress gages were located in the emplacement hole (Fig. 1) and the satellite hole (Fig. 2). Table 2 summarizes the types of gages, their location and orientation, and the material used in fabrication of the piezoresistance grids. Because of orientation requirements, the stress gages in the emplacement hole were supported by brackets that were bolted to the emplacement pipe before being lowered into place. The two stress gages in the satellite hole were supported by brackets that were bolted to the fiberglas pipe ui^ed in the grout backfill operation. The stress gages were'^nstalTed to measure the radial and tangential stress profiles associated with trie'ground shock and the magnitude of the residual stress field above the cavity following rebound of the region around the cavity. The primary difficulty in obtaining these measurements is survival of j;he stress gage through the ground shock and subsequent large displacements induced by the ground shock and cavity expansion. The ?bove dynamic measurements were conducted to obtain aata for comparison with future predictions for the numerical simulation of the ground motion following the nuclear detonation. In addition to supporting the five radiation and pressure canisters, the cable pendant also supported a motion canister located in the middle of each of the four CTE plugs. Each motion canister contained a vertical velocity gage, a vertical acceleration gage, and a thermistor. The thermistor provided a temperature reading for correction of the velocity gage record. A similar motion canister was emplaced at three different depths in the satellite hole (Fig. 2). The depths were chosen so that each motion canister would be at a radial distance from the WP corresponding to the location of the motion canisters in each of the top three CTE plugs of the emplacement hole. This provided a direct comparison of ground motion measurements in the satellite hole to those in the emplacement hole. Experience^ has shown that the large majority of releases havo occurred for events with yields less than 12 kt. For such events it has been customary to emplace acceleration and velocity gages in each plug to monitor the motion associated with each plug. Since the emplacement hole is generally not included in numerical simulations, there arises the question of how representative are these measurements of plug motion to those predicted for the free-field at the comparable radial distance from the WP. To explore this question further, let us examine certain criteria generally imposed on a good free-field measurement. The following discussion enumerates a few prominent criteria that one commonly strives to satisfy. Cl. To reduce any perturbation to the shock wave propagation the sensor and its housing should be constructed of material to match the impedance of the surrounding media. The sensor should not'affect the parameter that is being measured. C2. The hole diameter should be comparable to the width of the sensor emplaced or as small as possible. Any hole represents a man-made alteration of the geologic media that cannot be totally compensated for by remedial effort. C3. The hole is generally backfilled with a grout that has been carefully selected in an attempt to match the impedance of the surrounding medium following emplacement of the sensors. This effort is made to address the problem cited in the second criterion. C4. Generally, the sensor is emplaced so that it interacts with the shock propagating in a direct line from the explosive center. This criterion is normally implicit but is stated explicitly here because of certain variances that exist in the comparisons to be made.

413 A surface motion canister (Station 62) was boltec' to the wall of the surface casing at a depth of 0.6 m. This velocity measurement will be compared with that from a surface motion canister (Station 61) buried at a depth of 1.22 m and a.horizontal distance of 15.24 m from surface ground zero (SGZ). Until shortly after the first spall occurred, the particle velocity profile is determined by the superposition of the incident compressive wave and reflective tensile wave from the free surface.4 An unusually large ground shock or su,rface motion could significantly increase the spall depth^ and induce extensive radial and tangential fracturing of the medium around the upper portion of the emplacement hole.

3. Containment Analysis Initially, experimental findings for radiation and gas pressure measurements will be presented. This is followed by stress measurements in the satellite (free-field) hole and emplacement hole which provided not only peak stresses associated with the shock front but also values for the residual stress field following rebound of the region around the cavity. Then experimental results from acceleration and velocity gages in the emplacement hole will be correlated with corresponding gages in the satellite hole and at the surface. 3.1. Radiation and Pressure Measurements Station 31 (80.4 m above the WP) was located in |_LL coarse fill below the lowest plug but above the lowest two fines layers. After a few seconds of EMP generated noise, the record for the radiation gage returned to its previous baseline. Neither the radiation or pressure gages registered an increase above normal until their outputs were terminated by subsurface subsidence at 25.55 minutes after the nuclear detonation. The above results indicate that cavity gases did not reach this station prior to the occurrence of subsurface subsidence. Pressure gages in Station 32 (94.8 m above the WP) and Station 33 (121.9 m above the WP) registered large reductions in their ambient pressure readings following the first subsurface subsidence, Fig. 3. As material below the lowest plug falls into the cavity it leaves a void region in its wake. The rapid depressurization of the emplacement hole between the bottom two plugs indicates good communication between this region and the void volume below the lowest plug. These results imply that the subsurface subsidence compromised the lowest plug sufficiently that it would unlikely provide a gas seal if challenged b-> cavity gases. The fact that Station 34 (Fig. 3) above the second plug aid not record any depressurization would suggest the second plug retained its gas seal capabilities following the first subsurface subsidence. The radiation gages in Stations 32, 33, and 34 provide a check on these suppositions. Figure 4 illustrates the radiation records for all three stations. Currently there exists no definitive explanation why the radiation levels at Station 33 are nearly an order of magnitude greater than the radiation levels at Station 32. Figure 4 represents the radiation history for approximately a full day following the detonation. Monitoring was continued for an additional four days. During this time Stations 32 and 33 decayed to 0.01 R/hr (background) and 0.05 R/hr (five times background), respectively. Throughout this five day recording period neither Stations 34 or 35 registered increased radiation readings above their background levels. This demonstrates that the second plug was an effective gas seal.

414 3.2 Stress Profiles in the Free-Field and Emplacement Hole Configurations Figure 5 presents the radial stress profile for the flat pack$ (gage 52) located in the satellite hole (UElObf). First shock arrival occurs at 13 ms and then peaks at 65 MPa by 18 ms. By 35 ms the pulse has decayed to half maximum. As the shock wave continues to propagate, radial stress at the gage location continues to decay to 2.5 MPa at 125 ms. Then the medium appears to be rebounding to a peak radial stress of 6.0 MPa at 192 ms. Afterwards the medium seems to relax to a fairly steady state in the 2-4 MPa range until the record is terminated at 653 ms. Figure 6 gives the tangential stress profile for the flat pack (gage 53) located in the satellite hole. First shock arrival occurs at 14 ms and then peaks at 64 MPa by 18 ms. By 30 ms the pulse has decayed to half maximum. As the shock wave continues to propagate, tangential stress at the gage location decays to 4.5 MPa at 140 ms. Then the medium begins to rebound to a peak of 9.0 MPa at 192 ms. Shortly after 200 ms the-reading becomes erratic due to what we believe are problems with the junction between the cables and gage. Orientation of the flat pack and cables are important in reducing differential motions at the junction between the two systems when they are impacted by the shock wave.5 This problem can be corrected in the future with a 90 degree twist incorporated into the steel strips that form the body of the flat packs. Then the cables would exit the flat pack in a plane parallel with that for the radial stress gage. Comparing the radial and tangential stress measurements in the satellite hole reveals some important features. Although the peak radial stress slightly exceeds the peak tangential stress they are roughly comparable in magnitude and occur at the same time. Calculational results normally predict^j7 an even higher peak value for the radial stress with respect to the peak value for the tangential stress in the shock front. The relatively high peak reading for the tangential stress may be a gage interaction effect and/or an inclusion effect.8 The pulse width for the initial shock appears to be broader for the radial stress than for the tangential stress. Rebound appears nearly complete by 192 ms.with peak tangential and radial stresses of 9.0 MPa and 6.0 MPa, respectively. The fact that the residual tangential stress exceeds the residual radial stress by 50% for this distance from the WP is consistent with.the, model proposed9»10 to characterize the residual stress field following a nuclear explosion. Measurements have also shown the presence of a residual stress field around the cavity for rock matching grout in laboratory" and tuff in field^ experiments. Both studies employed high explosives as the driving source. Of the three gages (numbers 11, 12, and 13) located 22 m from the WP in the emplacement hole only the composite tangential gage^3 (number 12) survived the initial ground shock. Figure 7 illustrates the tangential composite records (number 12) for the ytterbium and the ytterbium reading after correction for longitudinal strain. First shock arrival occurs at 20 ms and the peak of 19 MPa at 25 ms. The pulse decays relatively slowly and reaches half maximum at about 110 ms. The pulse reaches a minimum of 8.5 MPa at atcjt 140 ms during a transient noise disturbance. The transient noise pulse is attributed to electrical problems and not grid related. Following rebound, the tangential stress peaks at 9.5 MPa at about 193 ms. Afterward the compacted stemming relaxes to a steady state in the 8.3-8.7 MPa range until the record is terminated at 653 ms. A composite gage on VIDc, located 30.1 m above the 415 WP, survived to measure about 16 MPa for the residual stress field.13 The geologic medium in the vicinity of the gage and VIDE WP was tuff, which normally possesses a greater shear strength than alluviumJ4- This fact may explain the higher residual stress field that was measured on VIDE compared to QUESO. The tangential stress measurement in the emplacement hole indicated a later TOA for shock arrival and a factor of three lower peak stress value than was recorded in the satellite hole. These differences are attributed to the failure to satisfy the last three criteria cited earlier for obtaining a good free-field measurement. The shock wave arriving at the gages does not propagate in a straight line through the stemming from the WP but is coupled in from the medium (criterion 4). The dimensions of the gage are small with respect to the emplacement hole (criterion 2) and located at 0.41 m, about half the radius from the center of the emplacement pipe to the wall of the emplacement hole. The LLL stemming mix surrounding the gage has much lower impedance than the mixed alluvium medium (criterion 3). Consequently as the shock propagates from the medium into the hole it takes a few milliseconds and is attenuated by compaction of the stemming before reaching the gage. The delay and attenuation effects are felt to be the explanation for the later TOA and lower magnitude of the peak stress readings for the stress record of the tangential gage (number 12) with respect to that observed in the satellite hole. The three stress gages (numbers 14, 15, and 16) .in the emplacement hole at 50 m above the VJP had signal levels too low for reliable data reduction. At this distance the shock wave has attenuated1^ to less than 10 MPa and is therefore near the clastic limit^,17 fOr the mixed alluvium layer above the WP. Consequently not much compaction of the stemming material would be expected. Such conditions would explain the observed low stress levels and low signal to noise ratios. 3.3 Vertical Acceleration Records for the Emplacement Hole and Satellite Hole The vertical acceleration (AV) records for the four plug stations (21 through 24) and surface casing station (62) are shown in Figs. 8 through 12, respectively. The first few rapid oscillations for the above stations were induced by the motion of the emplacement pipe. Such rapid oscillations are quickly attenuated with distance and thus are not seen by the accelerometers in the satellite hole stations (41,-42, and 43) or at the surface station (61). This fact is illustrated in the vertical acceleration (AV) records for these stations shown in Figs. 13 through 16. The reader is reminded that the three satellite stations (41, 42, and 43) correspond in range from the WP to the motion canisters (22, 23, and 24) located in the top three CTE plugs. The motion canister (21) in the lowest CTE plug has no counterpart in the satellite hole. Another aspect of the motion canister records in the satellite hole is the rapid oscillations that persist for about 40 ms and are centered at approximately 200 ms. This transient signal has been observed in comparable satellite hole measurements on previous events and is attributed to stress relaxation of the fiberglas pipe used to emplace the sensors and for grouting the satellite hole.'8 m comparing the free-field and emplacement hole accelerometer records we will limit the discussion to certain prominent signals and not try to cover small differences such as the transient oscillations cited above.

416 The first large acceleration attributed to ground motion was produced by the elastic precursor portion of the shock wave. Following this positive acceleration phase the gravitational forces become evident with a period of negative acceleration. During this negative acceleration period the upward motion of the ground slows, comes to rest and then descends until recompaction or slerpdown occurs. Approxi- mately 200 ms elapses between initiation of the first positive phase and the initiation time for slapdown. During this period the negative acceleration phase lasts about 125-150 ms. Following slapdown and rebound, the ground experiences a series of rapidly damped oscillations. The free-field surface station has oscillations following slapdown that do not dampen as rapidly due to the absence of overburden which acts as a restraining force. The accelerometer records for the comparable emplacement and free-field stations are in excellent agreement. This statement applies not only to the timing but the general pattern of the various signals. Although the amplitude of the more prominent accelerations may generally differ to a degree, even these results are in good agreement at the 164.9 and 181.5 m stations. The above comments must be qualified in certain cases. For example, the first positive acceleration induced by ground motion at Station 22 (Fig. 9) has been modified by the superposition of the emplacement pipe signal. The stress relaxation effect for the fiberglas emplacement pipe at about 200 ms on the accelerometers in the satellite hole has already been mentioned. The restraining force and damping effect of the surface casing on Station 62 (Fig. 12) explains many of the differences with Station 61 (Fig. 16). In light of the criteria cited earlier for good free-field measurements it might seem remarkable that such close agreement was recorded. We will defer arguments for the good agreement obtained until the comparisons between the corresponding free-field and emplacement hole results are given for the velocity profiles. 3.4 Comparison of Vertical Velocity Profiles for the Emplacement Hole Gages and Free-Field Gages Figure 17 presents the particle velocity profiles from the velocity gages in the four CTE plugs. For purpose of comparison, the particle velocities are given for the three velocity gages in the satellite hole and at the surface station. In spite of the failure to satisfy the free-field criteria cited earlier the particle velocity profiles of the top three plugs in the emplacement hole track closely the corresponding three free-field measurements in the satellite hole. This agreement can be better understood if we examine the time frame of the profiles. The 15-20 ms rise time portion at Station 22 for the first two velocity peaks correspond to the most rapidly changing segment of any of the profiles. Even for these short times the shock has propagated 30-40 m which is more than an order of magnitude greater than the radius (0.82 m) of the emplacement hole. Consequently there is more than adequate time for the shock wave to transit this distance several times to bring the gage into equilibrium with the surrounding medium despite the impedance differences associated with the stemming materials in the emplacement hole. The rise time of the shock front becomes progressively shorter as we approach the WP. Closer to the WP the dimensions cf the emplacement hole and differenres in physical properties of thu stemming compared to the surrounding medium wHl impact segments of the measurement profile

417 differently than in the free-field setting. This effect arose earlier in the discussion of the stress gage results. The comparison of the particle velocity profile for the motion canister attached to the surface casing (Station 62) with the surface motion canister (Station 61) is an example of how the emplacement procedure can affect the measurement. The 34.9 m long surface casing tends to move somewhat like a rigid body and thus provides a restraining or damping force, this damping effect is particularly evident on the timing and magnitudes of the initial fall back and subsequent oscillatory behavior of the free surface.

4. Summary and Conclusions QUESO was a low yield event conducted in hole UlOb.f on 11 August 1982. The main emplacement hole was stemmed with coarse and fines layers and four coal-tar-epoxy (CTE) plugs to inhibit the flow of cavity gases up the hole. To detect the possible flow of radioactive gases the emplacement hole contained five pressure and radiation stations. Radial and tangential stress gages were located in the emplacement and satellite holes to record dynamic stresses associated with the shock wave and residual stresses following rebound of the region around the cavity. A motion anister was installed in each of the four CTE plugs and another bolted t » the top of the surface casing to record accelerations and particle velocities induced by the shock wave from the nuclear explosion. To determine how accurately the above measurements represent free-field motions there were three motion canisters located in the satellite hole at comparable ranges from the WP as the motion canisters in the top three CTE plugs. In addition, a motion canister was emplaced at a surface station 15 m from surface-ground-zero for comparison with results from the motion canister on the surface casing. Data returned from the above diagnostic systems were excellent and provided a dynamic picture of the event that is coherent and consistent. Radiation and pressure records indicate that radioactive gases were contained in the cavity region until after the first subsurface subsidence at 25.55 minutes. Containment during this time is believed to have been achieved by a combination of: shock wave compaction of the lowest fines layer, high residual stress field above the cavity following rebound of that region after passage of the shock wave, cable fanout and cable gas blocks in the lowest fines layer. Records from the two pressure gages located between the lowest two plugs indicated rapid depressurization occurred immediately following the subsurface subsidence. This suggests the integrity of the lowest plug was compromised at the time of the subsurface subsidence. Radioactive gases slowly migrated up the hole past the lowest plug but were stopped at the second plug. A radiation gage just above the second plug was monitored for five days with no increase in reading above background indicating the integrity of the second plug was not compromised. During this period the radiation level just below this plug decayed from a peak of 186 R/hr to 0.05 R/hr (five times background). Radial and tangential measurements in the satellite hole provided peak stress readings. Following rebound of the region above the cavity the residual tangential and radial stress levels were 9.0 and 6.0 MPa, respectively. These values are in good agreement with the ratio generally cited in theoretical models of the post-shot residual stress field. At the same radial distance from the WP for the tangential stress

418 measurement in the emplacement hole a peak stress for the shock wave was a factor of three lower than measured in the satellite hole. The fines layer in which this gage is located has approximately 40% gas filled porosity and is very compressible. These characteristics should result in significant attenuation of the shock wave coupled in from the media and may possibly explain the observed differences in peak stress values for the tangential gages in the satellite and emplacement hole. The residual stress portion of the tangential stress profile are in close agreement for these two gages. Comparison of velocity profiles from motion canisters, at comparable ranges from the WP, in the emplacement hole and free-field stations were in excellent agreement. The few major differences that did exist could generally be explained. The net result is the motion canisters in the emplacement hole represent a fairly accurate picture of free-field response despite the significant differences in the emplacement conditions. The explanation lies in the time frames for which the velocity changes occur. The velocity changes relatively slowly giving time for the velocity gages in the emplacement hole to reach equilibrium with the motion of the surrounding medium.

ACKNOWLEDGMENTS The authors are indebted to F. A. Morrison, Jr., A. G. Duba, and C. W. Olsen for their critical review and constructive criticisms; W. Webb for field electronic support; R. S. Salazar and C. A. Cordill for Mechanical Engineering support.

Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract No. W-7405-ENG-48.

419 References 1. B. C. Hudson, "Containment Prospectus for the QUESO Event", Lawrence Livermore National Laboratory, Internal Report UOPKL 82-42, (1981) (Title U, Report CFRD). 2. H. D. Glenn, T. F. Stubbs, 0. A. Kalinowski, and J. F. Rambo, "Containment Analysis for the CABOC Nuclear Event", Lawrence Livermore National Laboratory, UCRL-53274 (Title U, Report CFRD) 3. B. C. Hudson and L. A. Dreiling, "An Historical Look at the Containment of Low Yield Events", Lawrence Livermore National Laboratory, Internal Report UOPKL 80-12 (1980) (Title U, Report CFRD). 4. H. D. Glenn, "Spall Study in One Dimension", Lawrence Livermore National Laboratory, UCID-17144 (1976). 5. H. D. Glenn, "Development of Diagnostic Techniques for Application to Underground Nuclear Tests", Defense Nuclear Agency Report DNA 3257F, Systems, Science and Software, La Jolla, CA, (1973). 6. T. R. Butkovich, "Project SCROLL - Final Report", Lawrence Livermore National Laboratory, UCRL-50588 (February, 1969), (T'tle U, Report CFRD). 7. R. w, Terhune, Lawrence Livermore National Laboratory, Private Communication, (1982). 8. A. L. Florence, D. D. Keough, and R. Mak, "Calculational Evaluation of the Inclusion Effects on Stress Gage Measurements in Rock and Soil", Proc. Interaction of Non-Nuclear Munitions in Structures Symp., May 10-13, 1983, USAF Academy, Colorado Springs, CO. 9. R. W. Terhune, H. D. Glenn, p. E. Burton, H. L. McKague, and J. T. Rambo, "Numerical Simulation of the BANEBERRY Event", Nuclear Technology, A6_, (November, 1979). 10. R. W. Terhune, H. D. Glenn, D. E. Burton, and Ji T. Rambo, "Containment Analysis for the Simultaneous Detonation of two Nuclear Explosives", Lawrence Livermore National Laboratory, UCRL-52268, (April 1977). 11. A. L. Florence and R. W. Gates, "Explosive Cavity Residual Stress Measurements", Stanford Research Institute, Report TR-76-3702-3 (1976). 12. C. W. Smith, "PUFF TOO: A Residual Stress Experiment," Sandia National Laboratories, Albuquerque, NM 87185, Report SAND 79-1674 (1980).

420 References (Continued)

13. C, W. Olsen, "Stress Measurement Development at LLNL", Proc. Monterey Containment Symp., Vol. II, August 26-28, 1981, Naval Postgraduate School, Monterey, CA-, (T.itle U, Report CFRD). 14. R. W. Terhune and K. D. Glenn, "Estimate of Residual Shear Strength Around Explosive Cavities in Earth Media", Lawrence Livermore National Laboratory, UCRL-52358 (1977). Also presented at the 19th U.S. Smyposium on Rock Mechanics, Stateline, NV, 1978. 15. T. R. Butkovich and A. E. Lewis, "Aids for Estimating Effects of Underground Nuclear Explosives", Lawrence Livermore National Laboratory, UCRL-5O929, Rev. 1 (February 1973). 16. B. P. Bonner, A. E. Abey, H. C. Heard, and R. N. Schock, "High- Pressure Mechanical Properties of Merlin Alluvium", Lawrence Livermore National Laboratory, UCRL-51252 (July 1972). 17. B. P. Bonner, Lawrence Livermore National Laboratory, Private communication (1983). 18. T. F. Stubbs, "PERA, Final Report on Free-Field Motion and Collapse", Lawrence Livermore National Laboratory, Report to be published.

421 Table 1. Measured and calculated average properties for the lithologic zones identified in QUESO. The values shown are based on sidewall samples and logs run in hole UElObf.

_^ Measured ^__ Computed Sample Water BuTir Gram Gas Lithologic Depth Content Density Density Porosity Saturation Porosity Zone (m) (Wt. %) (Mg/m ) (Mg/rn ) (Vol. %) (Vol. %) (Vol. %) Identification

0-207 6 .1 2.13 2.67 25 .1 51 .8 12.1 Mixed Alluvium 207-245 14.3 1.62 2.50 44 .5 51 .8 21.5 Tuffaceous Alluvium 245-388 13 .0' 1.57 2.48 44 .9 45 .3 24.6 Paintbrush Tuff 388-408 13.8 1.21 .2.37 56 .2 29 .7 39.5 Grouse Canyon Airfall | 408-678 15 .8 1.85 2.49 37 .6 77 .9 8.3 Tunnel Beds Tuff 678-691 - 2.66 - - - Paleozoic Dolomite

422 Table 2. Stress gages installed in the emplacement hole (UlObf) and the satellite hole (UElObf).

Radial Distance Hole Gage Gage from WP Orientation Grid I.D. Number Design (m) of Measurement Material UlObf 11 Composite* 22.0 Radial Ytterbium and Carbon 12 Composite 22.0 Tangential Ytterbium and Carbon 13 Flat-Pack* 22.0 Tangential Ytterbium 14 Composite 50.0 Radial Ytterbium and Carbon 15 Composite 50.0 Tangential Ytterbium and Carbon 16 Flat-Pack 50.0 Tangential Ytterbium UElObf 52 Flat-Pack 22.0 Radial Ytterbium 53 Flat Pack 22.0 Tangential Ytterbium

* The Composite gage was constructed by Dynasen, Inc., Goleta, CA, and the Flat-Pack gage was constructed by SRI International, Menlo Park, CA.

423 Surface -LLNL backfill mail -2.24 M dia surface OM — conductor Semi-liquid plug Rigid plastic plug 29 M with cable 31 M gas blocks 36M-J? 37MJ13M —Station 35 (RA, PX) 49 M -Rigid plastic plug 53 M-I Station 34 (RA, PX) 177.7 MM dia X 11.50 MM wall pipe filled with 86 M overton sand from DM 90 M to 157 M -filled with plastic from 157 M to 3 207 M O 15M 120M- -Rigid plastic plug CO . 123 M 125 M-{ -Station 33 (RA, PX) as 129 LLL stemming mix O typ 8 pics -Station 32 (RA, PX) -Semi-liquid plug Rigid plastic plug with cable gas blocks Station 31 (RA, PX) Drag ring 1.06 M dia v^~ Three stress gages 216.4 MWP 1.63 M dia uncased dole Ll.L coarse fill typ 4 pics \ -Three stress gages Radiation and pressure stations -Magnetite Three stress gage locations -Diagnostics canister Device canister Overton sand -1.22 M dia liner

Fig. 1. The QUESO (as built) stemming plan. Shown also are the five radiation and pressure stations (31-35) and two locations where stress gages were emplaced.

424 Depth UEIOBF Depth STA Meas Feat (meter) Feet (meter)

\W ///III Casing dia 13 3/8"- III Bottom of surface casing 80.0 (24.38) 76.5 (23.32)

116.9(35.63) 43 UV, AV

171.3(52.21) 42 UV, AV

Hole dia 12' 293.4 (89.43) 41 UV, AV

Medium matching grout

00—657.9 (200.53) 53 52

D- 710.0(216.41) 51 CL

740.0 (255.55) Compass

750.S (228.47) Bottom weight

Fig. 2. Stemming plan (as-buiit) for hole UEIOBF located 15.24 m west of hole UIOBF. Shown also are the vertical acceieration (AV) and velocity (UV) stations (41-43) and stress gage (aR, or) stations (52, 53). 0.10 1—r II 1 Till I I | I 1000 rr

Station 33

0.08 100

0.06

o 4> 0.04 l*o to 1.0 cc V \ Station 34 0.03 0.1 £_

0.02 I l I I i i i I i i i 0.01 I 24 30 40 50 SO 200 400 600 800 1000 1200 1400 Time (min) Time (min) Fig. 3. Pressure profiles for Stations 32, above the lowest Fig. 4. Radiation levels for Station 32, above the lowest plug. Stations 33 and 34, below and above the second lowest plug, Stations 33 and 34, below and above the second lowest plug, respectively plug, respectively. Fig. 5. Radial stress profile in the satellite hole at a range of 22 m from the WP.

-20 -0.1

Fig. 6. Tangential stress profile in the satellite hole at a range of 22 m from the WP.

427 -5 -0.1 0.1 0.2 0.3 0.4 0.5 0.6 Time (s) Fig. 7. Tangential strees profile (lower curve) at a range of 22 m above the WP. The upper curve shows the signal following correction for longitudinal strain.

0.60

Fig. 8. Vertical acceleraition (AV) profile in the lowest CTE plug and at a distance of 89.0 m above the WP.

428 4 1 r • r 1 ' i ' i 1 i • 3 A — f I 1 II - 2 a 0 1 \ U \V-s-^ 1 - — 1 I — -2 ,I.I -3 1 1 1 i i • i "0 0.10 0.20 0.30 0.40 0.50 0.60 0 0.10 0.20 0.30 0.40 0.50 0.60 Time (s) Time (s) Fig. 9. AV profile 128.0 m above the WP. Fig. 13. Satellite AV profile at 127.9 m range.

10 I ' I ' I ' 8 I o 6 c o

2 2 J — Si g 0 < -2 -4 -6 I , I I I I 0 0.10 0.20 0.30 0.40 0.50 0.60 0 0.10 0.20 0.30 0.40 0.45 0.50 Time (s) Time (s)', Fig. 10. AV profile 164.9 m above the WP. Fig. 14. Satellite AV profile at 164.9 m range.

12 1 n | i | r 10

111, 0 0.10 0.20 0.30 0.40 0.50 0.60 0 0.10 0.20 0.30 0.40 0.50 0.60 Time (s) Tirre (s) Fig. 11. AV profile 181.5 m above the WP. Fig. 15. Satellite AV profile at 181.4 m range

5 1 1 1 1 1 1 1 | ' 1 ' 4 3 3 — 1 - c •B 2 2 S 1 - 01

< ^ _ -1 - i , 1 -2 , 1 ,1,1, | | 0 0.10 0.20 0.30 0.40 0.50 0.60 0 0.10 0.20 0.30 0.40 0.50 0.60 Time (s) Time (s) Fig. 16. Surface AV profile at 215.7 m range. Fig. 12. AV profile 215.8 m above the WP.

429 61 UVS-F

62 UVS-F 215.8

CM 181.5

164.9

E - Pipe motion arrivals o Ground motion arrivals a uat 22 c UVS-F < 41 128.0 5

21 UVS-F 89.0

200 400 600 800 Time (ms)

Fig. 17. Particle velocity profiles in the emplacement hole (Stations 21-24 and 62), satellite hole (Stations 41-43) and at the surface (Station 61). There exists no velocity measurement in the satellite hole corresponding to Station 21 in the emplacement hole.

430 AUTHOR INDEX

Author Volume/Page Allen, R. T. 3/139 Ander, H. D. 1/155 App, F. N. 1/177 Barthel, J. R. 2/307 Bass, R. C. , . . 2/413 3/53 Batra, R. 1/103 Beiriger, W. F 1/323 Brown, W. T. 2/177 Brownlee, R. R 3/59 3/197 Bruesch, G. T. 1/17 Bryan, J. B. 3/35 Burton, D. E. . 1/369 Carothers, J. E 1/1 Chavez, P. F. 2/177 Cheney, J. A. 2/365 Cherry, J. T. 2/41 Cizek, 0. C. 2/1 2/179 Clark, S. R. 1/205 Cogbill, A. J. 1/177 Cook, C. W. 3/53 Davies, W. J. 1/271 Dockery, H. A. 1/155 Drellack.Jr., S. L 1/271 Duff, R. E. 2/413 Florence, A. L 2/1 2/179 Fogel, M. B. 2/71 Funston, R. J. 2/335 Gaffney, E. S. 1/305 2/365 Geil, R. G. 3/207 Glenn, H. D. 1/411 Glenn, L. A. , 4/1 Gonzales, J. L. 1/271 Griffiths, S. K. 2/281 Gulick, C. W. 2/379 , Hawkins, W. L. 1/271 3/115 Hearst, J. R. 1/113 1/205 Heusinkveld, M 2/123 Higgins, G. H. 3/63 Hill, L. R. 3/187 Howard, N. W. 3/83 Huang, L. 1/113 Hudson, B. C. 3/73

431 AUTHOR INDEX (Continued)

Author Volume/Page Jones, E. M. . . . ,\ 3/197 Kalinowski, J. A 1/411 Kunkle, T. D. 3/197 Lagus, P. L. r~^~. . . . 2/197 Lie, K. 2/41 Lowry, W. E. 1/49 Lundber-g, A. L 1/49 LaComb, J. 2/109 2/413 LaDelfe, CM. 1/121 Margolin, L. G 2/165 Mathews, .M. A. 1/121 Meadows, W. R. 1/177 Miller, L. R. E 2/335 Montgomery, S. T 2/401 Morrison.Jr., F. A 1/369 McKague, H. L. 1/175 Nilson, R. H. 2/253 2/281 Oliver, R. D. 1/97 Olsen, c. W. 1/293 Patch, D. F. 3/1 3/139 Pawloski, G. A. 1/75 Peterson, P.. W 2/197 Piwinskii, A. J ... 1/323 Rambo, J. T. 1/369 3/35 Rimer, N. 2/41 Ryerson, F. J. 1/323 Schatz, 0. F. 2/109 Shroba, R. R. 1/245 Smith, B. W. 2/165 Smith, C. W. , . 2/87 2/281 Smith, R. H. 2/109 Starrh, L. I. 1/49 Stubbs, T. F. 1/411 .Summa, W. J. 3/187 Terhune, R. W. 1/369 2/123 Thomsen, J. M. 2/335 Travis, B. J. 2/231 Van de Werken,M. G 1/393 Wagoner, J. L. 1/247 Warren, R. G. 1/213 Welch, J. E. 3/1 3/139 Wheeler, V. E. 3/103 Woodward, E. C ' 1/411 Wu, H. E. 2/197

432 Dv-i-ibution: UNCLASSIFIED Air Force Weapons Laboratory Field Command/DNA Albuquerque, New Mexico Albuquerque, NM Rinehart, Eric Ashbaugh, Laurence J. Aut, Warren E., RADM Desert Research Instutute Ballantine, George A. Reno, Nevada Gabriel, Lawrence J. Case, Clinton M. Keller, Carl E. (15) Fenske, Paul Sunir.a, Williair. J. EG&G, Inc. Field Command/DNA Las Vegas, Nevada Mercury, Nevada Bellow, Bernard W., Jr. LaComb, Joseph W. Fannin, David C. Farthing, Larry E. Headquarters/DNA Hill, George C. Washington, D.D. Hunsaker, Travis D. Gillespie, Milton, ADDST Noyes, Roger P. Reed, Melvin E. Linger, Don, SPSS Tanner, Paul A. Los Alamos National Laboratory Wallace, Delmont Los Alamos, New Mexico Williams, Troy L. Adams, Thomas F. App, Frederick N. FG&G, Inc. Batra, Ravi Mercury, Nevada Brownlee, Robert R. Coleman, Ronald R. Cogbill, Allen H. Jorgenson, Eric D. Diven, Benjamin C. Dockery, Holly A. EG&G, Inc. Edwards, C. L. San Ramon, CA Gaffney, Edward S. Broderick, Nicholas E. Hawkins, Ward L. Brown, Terence F. Homuth, Emil F. Clark, Steven R. House, Jack W. (15) Hatch, Melton, A. Jones, Eric M. Kalinowski, John A. Keller, Charles F. Longinotti, David B. Krier, Donathan J. Rienter, Jerome Kunkle, Thomas D. Stubbs, Theodore F. LaDelfe, Carol M. Library Services (3) Fenix & Scisson, Inc. Malik, John S. Mercury, Nevada Margolin, Leonard G. Baldwin, Margaret J. Marusak, Nancy L. Bruesch, Grant T. Mathews, Mark A. Davies, William J. Meadows, Wayne R. Douglass, Carl W. Oliver, Ronald D. Drellack, Sigmund L. Olwin, Richard B. Gonzales, Jose L. Planner, Harry H. Morrison, Jerry N. Saunders, B. Roy O'Brien, Margaret B. Scolman, Theodore T. Smith, Barham W. Townsend, Dean R. Travis, Bryan, J. Warren, Richard G. Weaver, Thomas A. Wohletz, Kenneth H.

UNCLASSIFIED UNCLASSIFIED Distribution (Continued): 1 Lawrence Livermore National Laboratory U.S. DOE, Nevada Operations Office Livermore, CA Las Vegas, Nevada Behne, Joseph C. Hague, Richard S. (3) Berryraan, James 6. Mudra, Paul J. Burkhard, Norman R. Burton, Donald E. Navarro, Richard Carlson, Richard C. Physics International Carothers, James E. San Leandro, CA Coyle, Philip E., Ill Miller, Lori R. Dong, Richard G. Thomsen, Jeffrey M. Geil, Robert G. Glenn, H. David Pacifica Technology Glenn, Lewis A. Del Mar, CA Grinnell, Stephen E. Allen, Robert T. Hearst, Joseph R. Fogel, Martin B. Heusinkveld, Myron E. Patch, Dan F. Higgins, Gary H. Holzer, Alfred S-Cubed Howard, Nancy W. La Jolla, CA Hudson, Billy C. Ide, Roger H. Barthel, James R. Larson, Donald B. Dismukes, Charles R. Lowry, William E. Duff, Russell E. Lundberg, Anders W. Higginbotham, Michael L. McKague, H. Lawrence Lagus, P. C. McNairy, Ray M., Jr. Lie, Kyoon-Haeng Maranville, Billy L. Nilson, Robert H. Moran, Bill Peterson, Edward W. Morrison, Frank A. Proffer, William J. Olsen, Clifford W. (15) Reed, Leonard Pawloski, Gayle A. Rimer, Norton •-• Piwinskii, Alf J. Wiehe, John W. Rambo, John T. Wu, H. E. Richardson, William U. Saunders, Kenneth A. Sandia National Laboratory Schock, Robert N. Albuquerque, New Mexico Spataro, Samuel J. Banister, John R. Springer, Donald L. Bass, Robert C. Starrh, Laurence I. Brown, William I". Stout, Ray B. Broyles, Carter D. Terhune, Robert.W. Chabai, Albert J, Thigpen, Lewis Chavez, Patrick f. Toman, John Cook, C. Wayne TID Library (3) Dolce, Sanders R. Wagoner, Jeffrey L. Griffiths, Stewart K. Wheeler, Vernon E. Gullick, Charles W. Wilder, Dale G. Hill, Leslie R. (15) Woodward, Ervin C, Jr. Kempka, Steven N. Mehl, Clarance R. Montgomery, Stephen T. Lawrence Livermore National Laboratory Plimpton, James D. Mercury, Nevada Smith, Carl W. Donithan, James D. Twenhofel, William S. Ferderber, Lawrence J. Yarrington, Paul McKinnis, William B. Schmidt, Casey S.

UNCLASSIFIED UNCLASSIFIED Distribution (Continued): Science & Engineers Assoc, Inc. McLean, VA Cramer, James D. Stanford Research Institute Menlo Park, CA Cizek, John C. Florence, Alexander L. Terra Tek, Inc. Salt Lake City, Utah Cooley, Craig H. Schatz, John F. Smith, Redd H. University of California Davis, California Cheney, James A. U.S. Geological Survey Denver, Colorado Carroll, Roderick D. Covington, Harry R. Dixon, Gary Healey, Don L. Jenkins, Evan C. Orkild, Paul P. Sharps, Joseph A. Shroba, Ralph R. Van de Werken, Martha G. Willaims, Van S. Tecnnical Information Center P. 0. Box 62 Oak Ridge, TN 37830

UNCLASSIFIED 3