In Arizona and the Southwest, Volume 12 (1982)

Hydrology and Water Resources in Arizona and the Southwest, Volume 12 (1982)

Item Type text; Proceedings

Publisher Arizona-Nevada Academy of Science

Journal Hydrology and Water Resources in Arizona and the Southwest

Download date 02/10/2021 21:09:19

Link to Item http://hdl.handle.net/10150/301320 At., War VOLUME 12 HYDROLOGY and WATER RESOURCES in ARIZONA and the SOUTHWEST PROCEEDINGS OF THE 1982 MEETINGS OF THE ARIZONA SECTION - AMERICAN WATER RESOURCES ASSN. AND THE HYDROLOGY SECTION - ARIZONA - NEVADA ACADEMY OF SCIENCE

APRIL 24, 1982, TEMPE, ARIZONA VOLUME 12 HYDROLOGY and WATER RESOURCES in ARIZONA and the SOUTHWEST PROCEEDINGS OF THE 1982 MEETINGS OF THE ARIZONA SECTION - AMERICAN WATER RESOURCES ASSN. AND THE HYDROLOGY SECTION - ARIZONA ACADEMY OF SCIENCE

APRIL 24, 1982, TEMPE, ARIZONA TABLE OF CONTENTS

Page

Publications of the Arizona Section (AWRA) iv Preface v Instructions to Authors vi

Comparison of Methods to Estimate Runoff From Small Rangeland Watersheds 1 H.B. Osborne, C.L. Unkrich and D. J. Busar

ET Measurements over Riparian Saltcedar on the Colorado River 9 L W Gay and R.K. Hartman

A RPN Program for the Generalized Penman Equation 17 L.W. Gay and Robert J. Greenberg

Changes in Streamflow in an Herbicide- Treated Juniper Watershed in Arizona 19 Malchus B. Baker, Jr.

Determining Watershed Conditions and Treatment Priorities 27 Rhey M. Solomon, James R. Maxwell and Larry J. Schmidt

Seasonal Change in Infiltration and Erosion from USLE Plots in Southeastern Arizona 37 J R Simanton and K.G. Renard

Distribution of Loss Rates Implicit in the SCS Runoff Equation 47 Richard H. Hawkins

Quasi Three -Dimensional Finite Element Model of the Madrid Basin in Spain 53 Jesus Carrera and Shlomo P. Neuman

Geostatistical Analysis of Aquifer Test And Water Level Data from the Madrid Basin, Spain 61 Patricia J. Fennessy and Shlomo P. Neuman

Energy and Water Resources Interactions in Arizona, 69 Nathan Buras

Potential Energy Resources of the Gulf of California, Northwestern Mexico 75 Barney P. Popkin

Nonstructural Flood Control Evaluation for Tucson, Arizona (Abstract) 87 Kebba Buckley

Impacts of the Arizona Groundwater Act on Tucson Water 89 Stephen E. Davis

Water Resources - The Primary Factor in Tucson's Future Growth 99 Thomas M. McLean

A Survey and Evaluation of Urban Water Conservation Programs in Arizona (Abstract) 101 Marc Bennett

An Application of the Almon Polynomial Lag to Residential Water Price Analysis 103 Donald E. Agthe

SEDCON: A Model of Nutrient and Heavy Metal Losses in Suspended Sediment 111 William A. Gabbert, Peter F. Ffolliott and William 0. Rasmussen

Techniques for Studying Nonpoint Water Quality 117 Donovan C. Wilkin and Susan J. Hebel

A Revised Phytoplankton Growth Equation for Water Quality Modelling in Lakes and Ponds 123 James Kempf, John Casti and Lucien Duckstein

Mutagenic Activity of Selected Organic Compounds Treated with Ozone 139 Leslie Irwin and Cornelius Steelink

Key Word Index of Arizona Section (AWRA) Proceedings 147 TO ORDER COPIES OF:

HYDROLOGY AND WATER RESOURCES IN ARIZONA AND THE SOUTHWEST

Volume 1, Proceedings of the 1971 Meetings, Tempe, Arizona Edited by Daniel D. Evans, Ph.D. (29 papers) Out of print

Volume 2, Proceedings of the 1972 Meetings, Prescott, Arizona Edited by Martin M. Fogel, Ph.D. (28 papers) Out of print

Volume 3, Proceedings of the 1973 Meetings, Tucson, Arizona Edited by Ron S. Boster, Ph.D. (25 papers) Out of print

Volume 4, Proceedings of the 1974 Meetings, Flagstaff, Arizona Edited by Ron S. Boster, Ph.D. (30 papers) Out of print

Volume 5, Proceedings of the 1975 Meetings, Tempe, Arizona Edited by D. L. Chery, Jr. (27 papers) Out of print

Volume 6, Proceedings of the 1976 Meetings, Tucson, Arizona Edited by D. L. Chery, Jr. (44 papers) 510 per copy

Volume 7, Proceedings of the 1977 Meetings, Las Vegas, Nevada Edited by Linda M. White (32 papers) 510 per copy

Volume 8, Proceedings of the 1978 Meetings, Flagstaff, Arizona Edited by T.R. Verma, Ph.D. (29 papers) $10 per copy

Volume 9, Proceedings of the 1979 Meetings, Tempe, Arizona

Edited by G. Harwood, Ph.D. & K. J. DeCook, Ph.D. (24 papers) . 512 per copy

Volume 10, Proceedings of the 1980 Meetings, Las Vegas, Nevada Edited by Gerald Harwood, Ph.D. (36 papers) 512 per copy

Volume 11, Proceedings of the 1981 Meetings, Tucson, Arizona Edited by Gerald Harwood, Ph.D. (35 papers) 514 per copy

Volume 12, Proceedings of the 1982 Meetings, Tempe, Arizona Edited by Todd C. Rasmussen & Mary L. Tidwell (20 papers) 514 per copy

ARIZONA SECTION SYMPOSIUMS

Water Conservation Alternatives (April 12, 1979) Edited by Daniel D. Evans, Ph.D. & K. James DeCook, Ph.D. 57 per copy

Flood Monitoring and Management (October 26, 1979) Edited by K. James DeCook, Ph.D. & Kennith E. Foster, Ph.D. 57 per copy

Water Quality Monitoring and Management (October 24, 1980) Edited by K. James DeCook, Ph.D., Kennith E. Foster, Ph.D. & Mary L. Tidwell 57 per copy

SEND REMITTANCE TO:

Mary L. Tidwell Office of Arid Lands Studies 845 N. Park Ave. Tucson, AZ85719

Make Checks Payable To:Arizona Section, AWRA

Note: Forcurrent members of the American Water Resources Association, Arizona Section, the curren proceedings may be purchased for a 52.00 discount. Other past proceedings may be purchased for a 51.00 discount. Dues for membership in the AWRA Arizona Section are 53.00 peryear. University students are eligible for joint membership in the AWRA Student Chapter and the Arizona Section for 55.00 per year.

iv PREFACE

Increased desires to meet the future agricultural, industrial and municipal water needs of people living in arid parts of the world has resulted in research which is focusing on how to best obtain water of the highest quality and quantity. As it becomes increasingly apparent that the problems of the southwestern United States are commonly fealt by other portions of the world, research instigated in this country can be used to solve global problems; shown in two articles on the development of water resources in Spain (Carrera and Neuman; Fennessy and Neuman).

Because the development of the Central Arizona Project was conceived at a time of abundant energy resources, the project is confronted with the dilemma of providing the power needed to deliver the water to its final users. The question of where the energy will be obtained is raised by one of the authors in this proceeding (Buras), and is partially answered by a companion article by another author (Popkin). The option of user conservation (Davis), perhaps motivated by water pricing (see the article by Agthe), appears to be an important alternative to importing high -cost water from outside the basin.

The importance of a high quality water source motivates the articles in the remainder of the proceedings. Such questions as sources of heavy metals (Gabbert et al.), nonpoint source water quality impacts (Wilkin and Hebel), and the production of mutagenic activity by treating water with ozone (Irwin and Steelink) demonstrate the concern that is developing in this country for preserving the purity of our waters.

A highlight in this years proceedings is a Key Word Index for all of the articles published under the auspices of the Arizona Section (AWRA).The index can be found at the end of this volume and should provide a bibliographic source for topics of particular interest to researchers concerned with arid land problems.

Todd C. Rasmussen Tucson, AZ. July, 1982 Instructions to Authors

Beginning with Volume 11,instructions have been modified to accommodate the use of word processors. While the use of a word processor is preferred, a typewriter is acceptable. Each submitted manuscript must be camera -ready.Use papers in Volume 11 or later volumes as typing models.

Manuscript Paper

Upon receipt of written or telephoned request, 10 sheets of 9 1/2 x 13 1/2 paper will be sent to authors for use on the final manuscript typing. No other paper will be acceptable.

Page Limit

Manuscripts must be limited to 8 pages, including all figures and tables.Pages in excess of 8 will be accommodated at a charge of $10.00 per page, whether or not the author orders reprints.

Page Margin

Text, tables,and figures must fall within the text area of 8.75 x 12.25 inches. Please attempt to fill the page as completely as possible, keeping the right hand margin as straight as possible. Type from one blue line side margin to the other, and from the top to the bottom blue line of the supplied manuscript paper. Except for the first page, completely fill the space between the blue lines with typing and figures. Take care to use proper hyphenation.The submitted material will be reduced for a printed line length of 5.25 inches.

Right justified text is preferred if text is being typed on a word processor.

Typewriter or Printer Ribbon

Only carbon or film ribbon is acceptable.

Typewriter and Type

It is recommended that an IBM Selectric Typewriter (or equivalent) or a word processor with Letter Gothic type be used for typing the manuscript. If you are preparing the manuscript on a word processor with proportional spacing capabilities, the use of proportional fonts such as Gothic PS is encouraged. Do not use small elite fonts as they will not show clearly when reduced; also, please do not use dot matrix printers.

Spacing

Single space all text material. Double space (or leave a blank line) between paragraphs. Leave 2 blank-5ines etween sections. Indent the beginning of each paragraph 5 characters.

Figures and Tables

Tables and figures must be kept between blue margin lines. If a figure is of such a size that it may be inserted in the text material to conserve space, plan the typing to leave space for thefigure. Paste (use stick glue or rubber cement) the drafted figure in the blank space.

Alternatively, figures may be placed at the end of the manuscript. Please provide complete captions for each figure or table.Figures and tables must be of camera -ready quality and clearly identified. Figures must be originals. Neither photocopf Tigiires and tables nor computer generated tables printed with a dot matrix printer are acceptable.

First Page

When typing the first page of the manuscript, leave 1 1/2" of space between the top blue line and your title.

Place the author affiliation(s) and complete mailing address beneath the author's name(s).Use the papers in this volume as a guide.

vi Abstract

An abstract may be included. If one is included, give it a most important heading and begin it directly below the author's name and affiliation on the first page.

HEADINGS

Titles are centered and capitalized.

Most Important Headings

First letters capitalized, centered, underlined or bold faced, triple- spaced below the last line of the text and double- spaced above the line.

Next Important Headings

First letters capitalized at left margin, underlined or bold faced and triple- spaced below last line of text or previous heading, and double- spaced above the next line.

Least important headings. Lower case (except for first letter of first word), incorporated into the beginning of a paragraph, underlined or bold faced, followed by a period.

Underlining vs. bold face. If text is being typed on a word processor, bold face type is preferred to underlining; be consistent in your choice of methods. However, underlining, ot bold face, is required for italicized words in text and bibliography, unless italic type is available

Endnotes

Endnotes, not footnotes, are to be used.They should be consecutively numbered throughout the paper and appear at the end of the paper, following references cited, within the margin requirements. Avoid the use of end notes entirely, if possible.

Reference Citations

Use the name -and -year system (Smith and Green, 1982). The heading above the list of references at the end of your paper should read References Cited. Use hanging indentation typing format. Do not use underlining except for italicized words such as species names.Capitalize only the first word and proper nouns in article titles. Capitalize all major words of book and journal titles. Also use the following format: Volume(Issue number):page numbers. See the enclosed sample of reference citations for proper citation format.

Smith, Robert and Arthur Green. 1982. Conflicting theories on the reproduction rate of Felis domesticus during the Great Flood. In: Proceedings of the Society of Arcane Studies. R. oT nés (Ed.). San Francisco:Obscure Publications, Inc. 5(14):127 -136.

Page Numbering and Assembling Manuscripts

Do not number pages of the manuscript. Numbering will be done when the volume is assembled. To avoid possible confusion,temporary page numbers should be blue -penciled lightly in the upper right hand corner of the page.

Please do not staple pages.Place pages in order in an envelope; the manuscript should not be bent or folded.

Submission Deadlines

The deadline for manuscript submission is the final day of the annual meetings of the AWRA. At the meeting, manuscripts may be submitted to any of the AWRA Officers. However,it is prefered that manuscripts be mailed directly to the Editor of the Proceedings at the address given on the Preface page of this volume. In past years, deadline extensions have been liberally granted for a variety of reasons, sometimes delaying publication for several months. Frequently,long delays result from the government agency peer review process and improperly prepared manuscripts. Therefore, the editors request that all authors take special care in following the instructions for manuscript preparation and do everything in their power to meet the meeting date deadline.

vii COMPARISON OF METHODS TO ESTIMATE RUNOFF FROM SMALL RANGELAND WATERSHEDS

H. B. Osborn, C. L. Unkrich, and D.J. Busar Research Hydraulic Engineer and Hydrologic Aides USDA, ARS Southwest Rangeland Watershed Research Center, 442 East Seventh Street Tucson, Arizona 85705

Introduction

Many methods have been used to estimate peak discharge and /or storm runoff volumes from small drainages in the United States (Chow 1962; Haan 1982). Most of these methods were developed for urban drainages, or were based on hydrologic data from the eastern and midwestern United States. Several methods have been suggested for possible use on small rangeland watersheds in the southwestern United States. Some, such as the Rational Formula, are easy to use but require subjective estimation of parameters and predict only peak discharge. Others predict both peak and volume, but may require more effort to use. In many cases, model use will determine the needed accuracy and required sophistication. In this paper, several models (methods)are compared and evaluated for use on rangeland watersheds using data from a very small gaged rangeland watershed.

Design of Experiment

The United States Department of Agriculture (USDA), Agricultural Research Service (ARS), Walnut Gulch Experimental Rangeland Watershed is located in southeastern Arizona, near Tombstone (Fig. 1). The 57-mi2 watershed is divided into gaged subdrainages ranging from 0.4 acres to the entire watershed. The study described in this paper was carried out on a 0.45 -acre subwatershed (63.105) located in the Lucky Hills research complex near Tombstone (Fig. 1). The small watershed is fenced and shrub -covered, and erosion pavement dominates the watershed surface. There is a well- defined, but shallow, channel draining the watershed, but channel abstractions were considered insignificant compared to watershed infiltration. The average watershed slope is9 percent, and the channel slope is about 3 percent. There isa 6 -hr weighing -type recording raingage on one edge of the watershed, and runoff is estimated from a FW -1 continuous water -level recorder mounted in a 3 -ft H- flume.

LUCKY HILLS WATERSHEDS

f - 0 I Y S 4 5 SCALE IN MILES CITY OF TOMBSTONE .. WATERSHED BOUNDARIES MAIN CHANNELS

LOCATION OF -_ WALNUT GULCH WATERSHED

Figure 1. Location of the Walnut Gulch Watershed and the Lucky Hills experimental area on Walnut Gulch.

1 Six major runoff -producing events on 63.105 were chosen to compare the five methods. The events were selected to represent a range of temporal rainfall input and antecedent conditions.The models are compared on the basis of their relative accuracy in estimating peak discharge rates and runoff volumes of these storms.

Methods

The five methods discussed in this paper are linear regression, the Rational Method, Illinois Urban Drainage Area Simulator (ILLUDAS), two versions of the Santa Barbara Urban Hydrograph (SBUH), and the Kinematic Cascade Model (KINGEN). The Rational Formula (method) can be used only to predict peak discharge. Linear regression can be used to predict peak discharge and runoff volume separately. The other three methods include a hydrograph output, and can be used to estimate peak discharge, time to peak, flow duration,and storm runoff volume. ILLUDAS and SBUH were both designed for urban drainages, and we expected to make some adjustments in the program subroutines and parameters. KINGEN had been used on both urban and natural watersheds, and considerable information pertaining to rangeland use was already available. In some cases, we estimated the parameters a priori; in other cases, parameters were adjusted by trial and error based on comparisons of observed and computed runoff. In each case, we have adopted the author's notation; so several symbols are not unique in definition.

Linear Regression

Osborn and Lane (1969) developed linear regression equations to predict runoff volume and peak discharge from 63.105 and other Lucky Hills subwatersheds. With additional data, these equations have been modified for 63.105 (Osborn and Simanton, 1981) to the form:

Qp = P15 - 0.68 (r2= 0.80) (1)

Q = 0.7PTOT - 0.24 (r2 = 0.78) (2)

where Qp = peak discharge (in /hr), Q = total storm runoff (in), PTOT = total storm rainfall (in).

The equations were used to estimate peak and volume of discharge from the 6 selected events.

Rational Method

The RationalMethod is stillone of the most popularand simple methods for predicting peak discharge from an ungaged watershed. The equation is:

Qp = CiA (3) where Qp = peak discharge (cfs), C = a constant based on watershed characteristics, A = watershed area (acres), and

i = maximum rainfall(in /hr) for the time of concentration.

A TI -59* hand calculator program for using the Rational Method has been developed by B.C. Yen (1981). In the Yen adaptation of the Rational Method, there are two methods for determining the time of concentration. We determined the maximum average intensity based on the first method, Kerby's (1959) equation, in which: Tck = To + Tf (4) where Tck = time of concentration by Kerby formula, To = time of maximum overland flow from boundary to channel, and Tf = time of flow through the channel.

The factors To and If are based on Kerby's coefficient, slopes, channel velocity, differences in elevation, and the lengths of overland and channel flows. These parameters must be entered into the program.

The second method for determining time of concentration, the Kirpich formula, did not give reasonable values for this watershed, and therefore cannot be recommended for use on such rangeland watersheds. The program TI -59 can be divided into three parts. First, the time of concentration is estimated using the Kerby formula. Second, precipitation intensity is determined by either assuming Td (rainfall duration) = Tc fora statistical event (such as the 30-min, 100 -yr storm), or by entering rainfall depth from an actual storm. Finally, peak discharges are obtained for each event.

*Reference to a specific calculator program does not constitute endorsement of the brand.

2 ILLUDAS

The ILLUDAS model, developed by Terstriep and Stall (1969), is an objective method for the hydrologic design of storm drainage systems in urban areas. The model improves on a method described by Watkins (1962) for the urban drainage, and adds a grassed -area component. The intent of our study was to investigate the "grassed- area" component of the model for possible application to rangeland drainages. In ILLUDAS, an observed time -varying rainfall pattern is uniformly distributed over the basin. The basin can be divided into subbasins which produce hydrographs that are combined and routed downstream from one design point to the next until the outlet is reached. Detention storage can be included as part of the design in any subbasin.

Time versus area curve for each subbasin is represented asa straight line from the origin to a point where the entire subbasin is contributing. The time coordinate of this point is the time of concentration, which can be either entered directly or determined within the program with the following equations by Izzard (1946): ge = 0.0000231IL, (5) where qe = equilibrium overland flow (cfs per ft of width),

I = supply rate (in /hr), and L = length of overland flow (ft), and te = 0.033KLqe- 067, (6) where to = time of equilibrium flow (min), and

K = (0.0007I + C)S-0,33, (7) where S = surface slope (ft), and C = cover coefficient = 0.046 (for bluegrass).

Standard infiltration curves were devised for soils by hydrologic group A, B, C, and D. These curves were calculated from the Horton equation as given by Chow (1964);

f = fc + (fo - fc)e -kt (8) where

f = infiltration at time t, fc = final constant rate, fo = initial infiltration rate (in /hr), k = shape factor (given as 2 in this program), and t = time from start of rainfall.

The program also requires selection of antecedent moisture conditions based on total rainfall during the five days preceeding the storm.

We found that, by changing a few data statements within the program, Izzard and Horton's equations could be easily changed to accommodate conditions other than bluegrass. For the very small watershed, 63.105, the travel times in the channel were so short that the routed hydrograph was the same as the watershed hydrograph.

SBUH Method

The current method for generating a complete hydrograph for retention /detention basin design for storm water management in Santa Barbara County (California)is the Howard Needles version of the Santa Barbara Urban Hydrograph Method (HNV- SBUH), originally developed by J.M. Studchaer of the Santa Barbara County Flood Controland Water Conservation District (Golding 1980). The final design hydrograph is obtained by routing the rainfall excess for each time period through an imaginary linear reservoir with a routing constant equivalent to the time of concentration of the basin.

The model can be described in three parts: (1) calculation of runoff depths,(2) computation of rainfall excess,and (3) routing the rainfall excess through an imaginary linear reservoir. Runoff depths for each time period R(t), are calculated using the following equations:

R(0) - I x P(t) (9)

R(1) = P(t)(1 -I)- f(1 -It) (10)

R(t) = R(0) + R(I) (11)

3 where P(t)= rainfall depth during time increment at (in), f = infiltration during time increment At (in), It = total impervious portion of drainage basin (decimal),

I = directly connected impervious drainage (decimal), and At = incremental time period (hr).

The rainfall excess, I(t)(cfs),is computed in the second step by multiplying the total runoff depth, R(t), for each time period, t, by the drainage basin area, A (acres), and dividing by the time increment At: A I(t) = R(t) x . (12)

In the third part, the design hydrograph is obtained by routing the rainfall excess with a time delay equal to the time of concentration of the drainage basin:

Q(t)= Q(t-1) + K [I(t) + I(t -1) - 2Q(t -1)] (13) where TC - 2Tc°+ At (14)

As was the case with ILLUDAS, the infiltration, f, is computed by the Horton equation (Eq. 8).

The program isalso similar to ILLUDAS in that the standard infiltration curves established by Terstriep and Stall(1969) which adjust foin equation (8)are used to compute infiltration. Again, the infiltration parameters can be changed to accommodate specific soil types.

The original program was written for an HP -67* programmable calculator and used a numerical integration scheme in the infiltration routine. With the help of Dr. Donald Ross Davis, Department of Hydrolo- gy, University of Arizona, the program was rewritten for a TI -59* calculator (DRD- SBUH), and the infiltration curve solved in closed form. Both methods were used in the study.

KIN6EN

Program KINGEN is a modified version of the Kibler -Woolhiser (1970)kinematic cascade model for routing overland flow overa cascade of planes and through trapezoidal channels (Lane and Woolhiser 1977). Input to the program is rainfall excess based on the Phillip (1969) equation:

f = 1/2 St-112 + A (15) where s = sorptivity of the soil, and A = steady state infiltration rate.

The program can be operated in either a simulation or optimization mode. Given slopes and channel characteristics, the program computes flow area, hydraulic radius, velocity, and shear stress for the channel segments. Output is a complete storm hydrograph from which peak time, discharge rate, and volume of runoff are obtained.

Comparison of Methods

Hydrographs were simulated for 6 selected runoff -producing events on subwatershed 63.105, and compared to actual data (Fig. 2 -6). Peaks and volumes of simulated and actual events were also compared (Tables 1 and 2). Both peaks and volumes tended to be overpredicted with ILLUDAS, and peaks underpredicted with either version of SBUH (Fig.2 -4). Both ILLUDAS and SBUH badly underpredicted runoff from the high- intensity, short -duration rainfall of 5July, 1975. KINGEN generally gave a better "fit" of the data, including the 5 July, 1975 event (Fig. 5 -6). We could have improved the "fit" of the actual and simulated peaks from the Rational Method by simply changing the "C" value. Since we had followed the instructions for estimating peak discharge for an ungaged watershed, we felt the strong tendency to over - predict should be noted. The ILLUDAS and SBUH methods were adjusted based on early fittings, and the hydrographs in Fig.2 -4 are based on parameter adjustments. The hydrographs developed with KINGEN are without adjustment. At this point, we clearly have more confidence in KINGEN than in the other methods. However, KINGEN is, by far, the most complex of the methods.

Discussion

Regression equations are easy to develop when rainfall and runoff data are available, and they can be used on similar ungaged watersheds. However, there are several assumptions which limit the value of such equations. First, watershed characteristics such as size, slope, and drainage density must not differ significantly between "similar" gaged and ungaged watersheds. Second, rainfall is assumed uniform over the watershed, bothin time and space. Third, rainfallintensity within the assumed constant

*Reference to a specific calculator program does not constitute endorsement of the brand.

4 1.4 65E274 - 40TU4L 1.4 - - _ IL1.5045

/ - MN 19JUL74 1.2 I %1 ( Ñ1.0

) N+ \

))\ `\1

20 30 40 50 TIME (MINUTES)

20 60 40 60 60 Figure 2. Comparison of three methods for estim- TIME (MINUTES% ating runoff for storms of 24 Sep 67 and 19 July 74. Figure 3. Comparison of three methods for estimating runoff for storms of 24 Sep 74 and 8 Sep 70.

1.2

245E267 1.0

IMUL4TE0 6uN01F 160M /. 6E4% TEINT 641N14LL - lIYYL4TE6 4uN011 ITEY Y4%16 YIM 641MI4LL

19JUL74 .6

5.011.75 242E274

0 O 20 30 40 50 6 TIME (MINUTES) 0 20 30 TIME [MINUTES) Figure 5. Comparison of estimated runoff using breakpoint and maximum 15- min rainfall in a kinematic cascade model (KINGEN) with actual Figure 4. Comparison of three methods for estim- rainfall and runoff data for three storms (from ating runoff for storms of 6 Sep 72 and 5 July 75. Osborn and Simanton, 1981).

5 regression rainfall duration must not significantly affect the estimate of peak and volume of runoff. Because of the extreme variability in rainfall, both in time and space, and the non -homogeneity of rangeland watersheds, linear regression is generally limited to very small rangeland watersheds of no more than a few tens of acres. Regression equations can be used to predict time to peak and flow duration also, but generally with considerably less accuracy than peak discharge and runoff volume.

Table 1. Comparison of estimated peak discharges for 5 methods and 6 selected events on subwatershed 63.105. Rainfall Peak Discharge (cfs) Storm Maximum SBUH Total Actual Regression Rational ILLUDAS KINGEN date intensity DDV HNV (2 min) (in) (in /hr) 24 Sep 67 0.39 4.50 0.64 0.40 0.59 0.70 0.530.40 0.45 8 Sep 70 1.14 4.20 1.04 1.08 2.24 1.45 1.20 1.13 .99 6 Sep 72 .79 3.90 .78 .77 1.38 1.90 1.14 1.12 .89 19 Jul 74 .94 3.00 .55 .63 1.36 .90 .68 .62 .67 24 Sep 74 .47 2.70 .64 .55 .91 .64 .49 .44 .45 5 Jul 75 .59 5.12 .98 .40 .62 .60 .48 .30 1.26

Table 2. Comparison of predicted storm runoff for 5 methods and 6 selected events on subwatershed 63.105. Rainfall Storm runoff (in) Storm Maximum SBUH Total Actual RegressionRational ILLUDAS KINGEN date intensity DDV HNV (2 min) (in) (in/hr) 24 Sep 67 0.39 4.50 0.12 0.04 0.11 0.15 0.11 0.14 8 Sep 70 1.14 4.20 .61 .56 .85 .88 .74 .59 6 Sep 72 .79 3.90 .37 .31 .46 .66 .63 .38 19 Jul 74 .94 3.00 .31 .42 .30 .45 .32 .28 24 Sep 74 .47 2.70 .20 .09 .14 .31 .23 .17 5 Jul 75 .59 5.12 .21 .18 .10 .14 .09 .25 Average .72 -- .30 .27 .33 .43 .36 .30

In this study, the regression coefficients were determined from runoff events for a 15 -yr period and used to predict the 6 selected events, so the fit of actual and estimated volumes was as good as those based on KINGEN, and the fit of actual and predicted peaks as good as all but those based on KINGEN.

The Rational Method is limited to predicting peak discharge independently of peak time and volume. It isa simple method to use, and it is indeed rational in that the units are proper, and the peak discharge depends upon rainfall intensity for the time of concentration. However, like linear regression, rainfallis assumed uniform, both in time and space, and the watershed is considered homogeneous. The "C" value must be representative of the runoff -producing features of the watershed, and is determined, to a large degree, subjectively. Time of concentration must be accurately estimated, and if the rainfall is too variable within the time of concentration, the predicted peaks may be in considerable error. The Rational Method is also limited to very small areas (in terms of acres).

As stated earlier, adjustments were made on the ILLUDAS and SBUH programs, both initially and as the study progressed. It became apparent early in the study that peak discharge and time to peak were extremely sensitive to the infiltration subroutine. Both ILLUDAS and SBUH were developed for urban drainages, with routines for both pervious and impervious areas within the same drainage. We found that the infiltration parameters, which were based on grassed urban areas, did not represent rangeland soils. We had to adjust the parameters considerably to reduce the infiltration for more and faster runoff. In most cases, we predicted no runoff with the original grassed -area infiltration parameters.

The ILLUDAS and SBUH methods are somewhat similar, using the same infiltration equations as well as similar routine techniques. Both are designed for use on larger watersheds, at least up to several square miles. Unlike the Rational Method and linear regression, rainfall can be varied in time, which improves the accuracy of estimation. However, neither program can handle spatial variability, which limits their usefulness for larger watersheds. ILLUDAS does include infiltration from subbasins within the drainage, which allows for non -homogeneous watersheds.We have not, as yet, investigated the sensitivity of shape factor, K, in the Horton equation.

6 Both ILLUDAS and SBUH were fitted to the 6 selected events. Each had infiltration parameters derived from infiltrometer data rather than the standard curves provided. Initial runs of both SBUH and ILLUDAS produced almost no runoff, because Izzard's equation gave a time of concentration four times longer than was reasonable. On subsequent runs, the time of concentration was entered directly.

The output for the 6 events from KINGEN are as shown. We used both breakpoint data and the maximum 15-min rainfall to better illustrate the need for breakpoint input. The program parameters were entered based on previous knowledge from other watershed studies. The parameters were not adjusted to improve the fit of actual and predicted peaks. In this test, KINGEN, with breakpoint data, clearly outperformed ILLUDAS and SBUH, but KINGEN had been used on other rangeland watersheds; whereas, we were starting from scratch with ILLUDAS and SBUH.

Summary

Several suggested methods for estimating runoff from rangeland watersheds were compared using six selected events from a very small gaged watershed on the USDA experimental watershed. These methods included linear regression, the Rational Formula, ILLUDAS (Illinois Urban Drainage Area Simulator), two versions of the SBUH (Santa Barbara Urban Hydrograph) method, and KINGEN (a kinematic cascade model). KINGEN was the most complex of the models, the Rational Formula the simplest. Estimates of both runoff peaks and volumes, based on KINGEN, were significantly more accurate than those from the other methods. Although the linear regression estimates were as accurate as those from ILLUDAS and SBUH, regression equations are not easily transferred to ungaged watersheds. ILLUDAS and SBUH were both designed for urban drainages, but were developed for use on ungaged watersheds. Neither method gave particularly good estimates in this test, but they may be applicable to other watersheds with revisions to accommodate spatial rainfall variability and /or time of concentration. The Rational Method can only be used to make quick peak estimates on very small watersheds. At this point, we have more confidence in KINGEN than in the other methods tested.

6JUtTS

0 I0 20 30 40 50 60 TIME I MINUTES) Figure 6. Comparison of estimated runoff using breakpoint and maximum 15 -min rainfall in a kinematic cascade model (KINGEN) with actual rainfall and runoff for three storms (from Osborn and Simanton, 1981)

7 References

Chow, V. T. 1962. Hydrologic determination of waterway areas for designs of drainage structures in small drainage basins. Univ. of Illinois Engr. Exp. Sta. Bul. 462.

Chow, V. T. 1964. Handbook of Applied Hydrology. McGraw -Hill Book Co., Inc., New York, p. 14 -17.

Golding,B. L. 1980. Hydrograph synthesis by the HNV -SBUH method utilizing a programable calculator. Proc. SWMM Users Group, Toronto, Ontario.

Haan, C. T. 1982. Hydrologic Modeling of small watersheds. ASAE Monograph.

Izzard,C. F. 1946. Hydraulics of runoff from developed surfaces. Proc. 26th annual meeting, Hwy. Research Board 26:129 -146.

Kerby, W. S. 1959. Time of concentration for overland flow.Civil Engineering, Vol. 74.

Kibler, D.F., and Woolhiser, D.A. 1970. The kinematic cascade as a hydrologic model. Colorado State Univ. Hydrology Paper No. 39, 27 pp.

Lane, L.J., and Woolhiser, D. A. 1977. Simplification of watershed geometry affecting simulation of surface runoff. J. Hydrol. 35:173 -190.

Osborn, H. B., and Lane, L.J. 1969. Precipitation- runoff relationships for very small semiarid rangeland watersheds. Water Resources Res., AGU 5(2): 419 -425.

Osborn, H.B., and Simanton, J. R. 1981. Maximum rainfall intensities of southwestern thunderstorms. Proc. Fourth Conference on Hydrometeorology, Am. Meteorological Soc., 166 -173.

Phillip, J.B. 1957. The theory of infiltration. 4. Sorptivity and algebraic infiltration equations. Soil Sci. 84:257 -264.

Terstriep, M. L., and Stall, J. B. 1969. Urban runoff by the road research laboratory method. ASCE J. Hydr. Div. 95(HY6):1809 -1834.

Watkins, L. H. 1962. The design of urban sewer systems. Dept. of Sci. and Industrial Res., London, Her Majesty's Stationery Office, Road Research Tech. Paper No. 55.

Yen,B. C. 1981. TI -59 calculator program for storm sewer design using Rational Method. (Unpublished method used with permission of author. For further information, contact Professor B. C. Yen, Dept. of Civil Engr., Univ. of Illinois, 208 N. Romine St., Urbana, Ill. 61801.)

8 ET MEASUREMENTS OVER RIPARIAN SALTCEDAR O1 TEE COLORADO RIVER

L. W. Gay and R. K. Hartman School of Renewable Natural Resources University of Arizona Tucson, AZ 85721

Abstract

Evapotranspiration (ET) from an extensive stand of saltcedar on the Colorado River floodplain was defined throughout the growing season by a series of Bowen ratio energy budget measurements in 1980 and 1981. The water table depth at the site near Blythe, California, was about 3m during the two summers of measurement. Daily ET totals ranged from 2.9 mm /day in early April up to 11.0 mm /day in late June, and dropped down to 1.8 mm /day in late October. These values are means from two separate measurement systems, averaged over measurement periods of two to four days in length. The highest single day total measured by an individual system was 12.7 mm on June 28,1981. The mid -summer ET rates from the saltcedar at this experimental site are substantial, and rank among the highest rates that have been reported elsewhere for irrigated cropland.The seasonal saltcedar water use of 1727 mm (including 90 mm of annual precipitation) is somewhat lower, however, than earlier, more speculative estimates for saltcedar that ranged up as high as 2100 mm per year.

Introduction The lower Colorado River flows through one of the warmest, driest regions in the United States. For example, the average annual rainfall at Blythe, California airport is only 78.7 mm (3.1 inches), and the maximum temperature of record is 50 °C (122 °F). Maximum temperatures frequently exceed 46 °C in each month from May through September. The Colorado River serves as a large source of irrigation water for this arid region, however, and this supports a prosperous, flourishing agricultural industry. The Colorado River also supports extensive riparian plant communities on its floodplain. The composition of these communities has changed through the years, as native species such as mesquite (Proeopis sp.) have been generally replaced by saltcedar ( Tamarix chinensie). Saltcedar is widely perceived as a heavy user of water that might otherwise be available for more beneficial use by man, and so it has been studied extensively for many years in the southwest.

The studies have evaluated both the consumptive use of saltcedar, and means for eradicating the plant in order to "salvage" water and thus augment existing water supplies. The earlier work on water use was summarized by Horton and Campbell (1974) who concluded (1) that a dense, mature stand of saltcedar would use 1.8 to 2.1 m (6 to 7 ft) annually on the Gila River near Phoenix, (2) the somewhat higher elevation sites at Safford, Arizona, and Carlsbad, New Mexico, would lose about 1.5 to 1.8 m (5 to 6 ft), and (3) use at the still higher elevation Bernardo site near Albuquerque would be about 1.2 to 1.4 m (4 to 4.5 ft) per year. The climatic conditions along the lower Colorado River are even more extreme than on the Gila River, and this should increase riparian water use, especially since the stable flows of the now -controlled Colorado provide for relatively high ground water tables beneath the extensive, heavily vegetated floodplains.

It is surprising to conclude that definitive estimates of water use by riparian species are lacking in the southwest, despite the wide interest in this topic. This is partly due to the complexities of the problem. Water use depends not only upon the climatic regime, but also upon type and density of vegetation and the supply and salinity of water. The lack of water use measurements is also partly due to limitations of the various experimental approaches.For example, Van Hylckama's (1974) lysimeter water budget studies are the basis for the Gila River water use figures

9 cited above. Van Hylckama pointed out that lysimetric measurements are not only time consuming, but the results can be extrapolated only to areas of similar characteristics. The same restriction applies to water budget studies such as the comprehensive measurements described by Hansen, et ál. (1972) along a 15 -mile reach of the Gila River in eastern Arizona. The energy budget approach offers possibilities for generalization, but the energy budget studies in saltcedar have extended over periods of only a few days (Gay and Fritschen, 1979a). It is apparent that improved estimates of water use by riparian communities would contribute to better management of our limited water resources. ET should be estimated from measurements that take into account environmental and vegetative differences between sites. The energy budget is such a method that has proven useful for field estimates of ET in agriculture and in natural communities as well. Consequently, we began several years ago to evaluate ET with the energy budget method in a large stand of saltcedar along the lower Colorado River.

Objectives

The objectives of this study were threefold: (1) to develop an estimate of seasonal ET; (2) to evaluate short term ET rates as an aid for process studies and for later modeling; and (3) to refine and develop measurement techniques for the energy budget method.

Methods

The energy budget method was chosen for this study because its fundamental basis allows generalizations to areas other than where the experiments were conducted. The method requires a moderate level of technical sophistication in order to operate in the field and away from the laboratory. The theory is well -known and has been thoroughly described (see, for example, Monteith, 1973).

The Bowen Ratio Energy Budget Model

The energy balance equation is a statement of the conservation of energy,i.e., the sum of all energy fluxes for a given system must equal zero.The major thermal fluxes in the soil- plant -atmosphere system at the surface of the earth are expressed as: Q* + G + H + LE = O. (1)

The symbols are: Q *, net radiation; G, soil heat flux or change in thermal storage; H, convection or sensible heat; and LE, latent heat of vaporization. Photosyntheti- cally fixed energy is small and is disregarded. The units are either energy flux densities (i.e., W /m2) or energy totals for a specified time period (i.e., J/m2). All fluxes directed to the earth's surface, whether from above or below, are positive and all fluxes away from the surface are negative. Bowen (1926) introduced the ratio of convection to latent energy( a- H /LE) as a means of estimating some of the fluxes in Equation(1). The Bowen ratio reduces to:

a= H /LE = A 30/ 3e (2) whereX is a coefficient equal to 0.66 mb / °C at sea level, and 10/1e is theratio of potential air temperature gradient (30 /3z,°C /m) to vapor pressure gradient (3e /3z, mb /m). The model assumes that the eddy diffusivities for heat and for vapor are equal to unity over the distance across which the temperature and vapor gradients are measured. In practice, 30 /3z and 3e /3z are approximated by measurements of differences 4T and 4e across a vertical distance 4z. The distance 4z is usually of the about 1 m, and the bottom level of measurement is a little above the top canopy. The Bowen ratio method has been widely used for measuring ET, and the basic accuracy of the model is well documented (see,for example, Fritschen, 1965). The measurements do require careful work and high precision, however, in order to achieve the desired accuracy.

10 The Measurement System

The system used for the Bowen ratio measurements is similar to that described earlier by Gay (1979). The key features are: a data acquisition system of excellent precision;a microprocessor computer; precisely calibrated, ceramic wick psychrometers (Gay, 1973; Hartman and Gay, 1981); and an exchange mechanism to interchange the psychrometers between observations. Net radiation and soil heat flux are also measured, and supplementary measurements are made of solar radiation, wind speed and direction. The data acquisition and processing equipment is housed in a small van; power is provided by a portable generator. Two sets of sensors are mounted on separate masts so that two independent ET estimates are obtained.

The psychrometer mechanisms were placed so that the lower psychrometer was just above the tips of the canopy, and the two sampling levels were separated by a vertical distance of 92 cm. The pair of psychrometers was interchanged every 6 minutes so that mean temperature and humidity gradients could be obtained each 12 minutes,free of instrument bias (Sergeant and Tanner, 1967). The exchange mechanism is described in detail by Gay and Fritechen (1979b). The mean 12- minute gradients were based on 40 samples at the tvo levels, after the psychrometers had comme into equilibrium following the interchange.

Data Acquisition and Processing

The data logger houses a 40 channel scanner, an integrating digital voltmeter, a real -time clock, and a strip printer, all in a compact case. The maximum resolu- tion of the voltmeter is better than 0.001% (1 uV in 120 mV full scale), at a scan rate of 2.4 readings per second. The data logger has an averaging option that compresses many sequential samples into a single average for transmission to the microprocessor computer for analysis. The Tektronix 4051 computer is a microprocessor -based system with graphics capability on a CRT screen. Computing and graphics functions are specified in BASIC language. The unit has 32K bytes of internal memory, and has an internal magnetic tape cassette with capacity of 450K bytes.Raw and processed data are stored on the magnetic tape for subsequent retrieval and further analysis when desired.

Site Description

The saltcedar study area was a vast saltcedar thicket of some 10 km2 in area, located on the floodplain of the Colorado River some 50 km south of Blythe, Califor- nia. The elevation was about 90 km. The fetch at the measurement site was about 1 km to the west, and from 2 to 3 km in the other cardinal directions. The height of the vegetation was about 7 m. The sandy soils were spotted with coarse pockets laid down by the meandering river channel in the past. The water table remained constant at about 3.3 m throughout the two summers of measurement.

Sampling

The effects of spatial variability were minimized by selection ofa site ina dense, relatively homogenous portion of the stand. The two masts were separated by about 10 m,and they are considered to sample the same area. Temporal variability was minimized by frequent sampling within a given day and by sampling for several consecutive days within each month. The energy budget (and latent energy) each day was summed from the 12- minute means, each of which was based on 40 measurements.

Results and Discussion

The data collected in this study represent an unusually precise evaluation of ET rates from saltcedar throughout the growing season. We shall first examine some of the daily measurements, and then look at the implications for seasonal water use.

Daily Measurements The energy transfer rates throughout the day are illustrated in Figure 1 for July 28, 1980, for the daylight period when Q *>0. This day was the warmest of all those sampled. The midafternoon air temperature reached 45.1 °C, with a relative humidity of 15.6%. The Figure shows the single value of G, mean values (averaged

11 JULY 28, 1988

ENERGY BUDGET / COLORADO RIVER SALTCEDAR 808

680 /Ne ti 400 LE1 - //// 1 A n It 200 I L.'` ALE2 AI `r 'w 'tr..' %, w__ / p r 0 ---- -1 " r I

-200

-400

-600

-808

-1088 . 1 1 1 1 . I 1 1 1 1 r 1 1 1 1 11 1 . 1 1 1 1 1 1 1 1 600 808 1090 1280 1489 1600 1888 2800

TINE (HRS)

Figure 1. -- Daytime energy budget over saltcedar on the Colorado River floodplain. Net radiation (Q) and convection (H) are means from two sets of sensors; latent energy (LEI and LE2) is shown separately for each set of sensors.Soil heat flux (G) was measured with a single disc.

between masts) of andk,and the separate mast estimates of LEI and LE2. Table 1. -- Energy budget totals (in MJ /m2 and equivalent mm depth of evaporated water) for The smooth trace of Q* confirms the 0624 -1836 July 28, 1980. rather clear sky conditions that day, marred only briefly by clouds in midmorning. The value of G is quite small, con- Flux MJ /m2 mm firming that very little soil heat flux takes place beneath the closed canopy in Q* 18.3 7.32 this stand. The amount of energy stored in G - 0.2 -0.08 the foliage was considered negligible, and 2.16 Ti. 5.4 no correction for this quantity was applied ER -23.5 -9.40 to G. The positive value ofH confirms (LE1/LE2) (-23.8/-23.2) (9.52/9.28) that advection took place all day long as warm air moved from the surrounding desert, across the relatively cool,evaporating canopy. IT is essentially the sum of Q* and H since G a 0. The close correspondence of LEI and LE2 is indicative of the precision with which the two sets of sensors are estimating i!. Itis apparent from the Figure that the two sets of sensors are in exceedingly close agreement, but serial correlation prevents us from establishing confidence limits on the precision. The closeness of the daily totals also serves as an index of precision. Flux totals are summarized for July 28, 1980, in Table 1for the daylight period of positive net radiation. The range in estimates of LE by LEI and LE2 is only ?1.3%. Thisisa very close comparison, especially since the commonly accepted accuracy estimates for Bowen ratio evaluations of LE are ±10% or 15%.

12 COLORADO RIVER SALTCEDAR 14

12 _ ir

DAYTIME ET +

I0 _ NIGHTTIME ET o

+ +* 8 _ t + +

6 _

4 _

2 _ 41.

o m 0 8 0 11I11I111g1I1.11111111111111,tj41I111 0 30 60 90 120 ISO 180 210 240 270 300 330 360

DAY OF YEAR

Figure 2. -- Daytime and nightime evapotranspiration totals from saltcedar on the Colorado River floodplain.

Estimates were obtained for the daylight period of positive net radiation on 21 days throughout the growing season. The number of successive days in each run ranged from 2to 4. Night time data were obtained from 11 nights during this period. The totals obtained from each date (averaged between masts) are plotted in Figure 2,for day and for night. The daylight water consumption ranged from around 2 mm /day in spring and fall up to 12 mm in midsummer. The nighttime water use was small, and ranged from 0.08 mm in spring and fall up to 0.6 mm in midsummer.

Since there are "missing" nights, we shall examine measured daylight totals as an indication of accuracy, and then extrapolate nightime observations as needed to obtain 24 -hour totals for estimating seasonal water use.

The daytime water use measured at site 1 (ET1)is regressed against that measured at site 2 (ET?) in Figure 3. The linear equation ET1 -0.03 + 0.945 E2 has r 0.991. The scatter is very small; 95% confidence limits at ET? 6.45mm are only +0.03%. The offset ( -0.03 mm) and the failure of the slope coefficient to equal unity (0.945) suggest a difference in evaporation rates at the two sites. This was confirmed by exchanging sensors between masts during several successive runs. The small difference remained, thus eliminating instrumentation as the cause of the observed difference. The precision of these measurements derives from improvements described by Hartman and Gay (1981). These include the care used in sensor calibration, the experimental design (exchanging psychrometers, frequent samples, timing of analysis), and the high quality of the data acquisition and processing system.

13 Seasonal Water Use COMPARISON BETWEEN MASTS I AND 2 COLORADO RIVER SALTCEDAR ET The seasonal water use estimate is derived from averages for each of the two to four day runs. The values in Figure 2 are averaged ET1 e -0.03 + 0.94 * ET2 and tabulated in Table 2,thus providing data for evaluating evapotranspiration losses throughout the growing season.

For thispurpose, it is assumed that evapotranspiration is negligible throughout the dormant season. This is likely, since the saltcedar loses its leaves. Fur- ther, winter rainfall is low, and a layer of dry sand at least a meter thick overlies the water table which is 3.3m deep. The spring greenup and fall dormancy date is arbitrarily setat March 23and November 11,based upon inspection of the data in Table 2. 0 1 A simple trapezoidal integra- 12 14 0 2 4 6 8 10 tion ofthe evapotranspiration for MAST $2 DAILY WATER USE,MM. the 233 day growing season yields the following totals: day, -1548 Figure 3. -- Comparison of daytime ET from mast 1 mm; night, -89 mm; total, -1637 mm. versus mast 2. The 1637 mm of evapotranspiration were developed from measurements of water transpired by the saltcedar. This total should be increased by some percentage of the annual rainfall to yield a better estimate of annual evapotranspiration. The annual precipitation at Ehrenberg, Arizona, 4 miles east of Blythe and about 25 miles north of the saltcedar site, averaged 89.7 mm (3.53in) for the 1931 -72 period (Sellers and Rill, 1974). Of this meager amount, 47% or 42.3 mm fell during the April through October growing season,and the remainder during the dormant season. The storms are generally light and infrequent, and it is probable that essentially all of the precipitation falling on the floodplain in this region evaporates,i.e., there is no effective runnoff. Our measurements did not coincide with rain, and there was little evidence of any antecedent soil moisture in the dry blanket of sand at this site. In our judgment, the evapotranspiration estimate should then be increased 42 mm during the growing season and 48 mm during the dormant season to account for direct evaporation of rainfall, and thus yield a "best" estimate of 1727 mm of evapotranspiration over the year. Other estimates from this region are quite generalized. For example, the U.S. Bureau of Recla- Table 2. --Mean daily totals (in mm) based mation (1964) estimated that there upon two separate masts and dates as shown. were 4,593 ha of saltcedar on reach number 4, which extends south from Blythe- Ehrenberg some 30 miles to a day night* 24 -hour point just below the measurement site. The annual wester use was Apr 6,7 - 2.8 -0.1 - 2.9 given as 62,401,704 m', or a depth Apr 28,29 - 6.8 -0.3 - 7.1 equivalent of1359 mm excluding May 28,29,30 - 8.2 -0.4 - 8.6 precipitation. The Bureau esti- Jun 26,27,28 -10.5 -0.5 -11.0 mates were based on the Blaney - Jul 28,29 - 9.0 -0.6 - 9.6 Criddle method, adjusted for den- Aug 15,16,17 - 8.4 -0.6 - 9.0 sity of vegetation throughout the Sep 12,13 - 6.9 -0.5 - 7.4 reach. The agreement is quite Oct 30 -Nov 2 - 1.8 -0.1 - 1.9 good,considering that the saltcedar at our measurement site was quite dense, and should be using *includes interpolated values. water at a near maximum rate. More speculative estimates from other regions range up to 2100

14 mm for the Gila River near Phoenix (Horton and Campbell, 1974). The climatic conditions are warmer and drier on the lower Colorado River than on the Gila, and it seems unlikely that even more favorable vegetation density and water availability combinations could exist to generate the high water use estimates of Horton and Campbell.

References Cited

Bowen,I.S. 1926. The ratio of heat losses by conduction and evaporation from any water surface. Phys. Rev. 27:779 -787.

Fritschen, L. J. 1965. Accuracy of evapotranspiration determinations by the Bowen ratio method. Intl Assoc. Hydrol. Bull. 10:38 -48.

Gay, L. W. 1973. On the construction and use of ceramic wick psychrometers. In: Brown, R.W., and B. P. Van Haveren (Eds.). Psychrometry in Water Relations Research. Pp. 251 -258. Utah Agric. Experiment. Sta., Logan.

Gay, L. W. 1979. A simple system for real -time processing of energy budget data. Proc., Symposium on Forest Meteorology, Ottawa. WMO No. 527, pp. 224 -226. WMO, Geneva.

Gay, L.W. and L.J. Fritschen. 1979b. An exchange system for precise measurements of temperature and humidity gradients in the air near the ground. Proc., Hydrol. Water Resour. Ariz. and SW 9:37 -42.

Hansen, R. L., F. P. Klipple and R. C. Culler. 1972. Changing the consumptive use on the Gila River floodplain, southeastern Arizona. In: Age of Changing Priorities for Land and Water. Proc., ASCE Irr. Drain. Div. Spec. Conf. Spokane. Amer. Soc. Civil Engrs.

Hartman, R. R. and L. W. Gay. 1981. Improvements in the design and calibration of temperature measurement systems. Proc., 15th Conf. Agric. Forest Meteor.,pp. 150 -151. Amer. Meteor. Soc., Boston.

Horton, J. S. and C. J. Campbell. 1974. Management of phreatophyte and riparian vegetation for maximum multiple use values. Res. Pap. RM -117. USDA Forest Service, Rocky Mtn. Forest and Range Exp. Sta., Fort Collins.

Monteith, J. M. 1973. Principles of Environmental Physics. Arnold, London. 241 pp.

Sargeant, D. H. and C.B. Tanner. 1967. A simple psychrometric apparatus for Bowen ratio determinations. J. Appl. Meteor. 6:414 -418.

Sellers, W. D. and R. H. Hill. 1974. Arizona Climate 1931 -1972. Univ. Arizona Press, Tucson. 616 pp.

U. S. Bureau of Reclamation. 1964. Pacific Southwest Water Plan.Supplemental Information Report on Water Salvage Projects - Lower Colorado River. USER, Boulder City, Nevada.

Van Hylckama, T.E.A. 1974. Water use by saltcedar as measured by the water budget method. U.S. Geol. Survey Prof. Pap. 491 -E. GPO, Washington.

Acknowledgements The work upon which this publication is based was supported in part by funds provided by the Office of Water Research and Technology (Project B- 084 -ARIZ), U. S. Department of the Interior, Washington, D.C., as authorized by the Water Research and Development Act of 1978, and in part by the Arizona Agricultural Experiment

Station, Hatch Project 04. Approved for publication as Journal Paper No.417 , Arizona Agricultural Experiment Station.

15 A RPM PROGRAM FOR THE GENUALITXD PENMAN EQUATION

L. W. Gay and R. J. Greenberg School of Renewable Natural Resources University of Arizona Tucson, AZ 85721

Abstract

The Penman potential evapotranspiration model has been expanded and generalized for a vide range ofclimaticconditions byJ.Doorenbos and W. 0. Pruitt (1975, FAO Irrig. á Drain. Pap. 24). Their comprehensive model relies heavily on tabulations and graphs to facilitate its application in remote areas where computer facilities are lacking. Future applications of this model have been substantially enhanced by our development of a program that runs on a HP- 41C /CV calculator. The RPN program occupies 1778 bytes of memory, and will run on a 41CV, or a 41C with a quad memory module. A printer is also needed. The required climatic data can be input in a variety of forms, and program execution was plannedwithparticular attention to the non- technical user.The entire calculator system (41CV, printer, card reader) cost less than $750, and provides for very portable usewitha high degree of computing power.

Introduction

As agricultural water resources become increasingly depleted the need for efficient utilization of irrigation water increases. Potential evapotranspiration predictions have become a vital aid for optimizing irrigation programming. Although extensive research has been achieved on all levels of predicting potential evapotranspiration (PET),a gap exists in relaying this technology between the scientific and agricultural communities.

In recent years large agricultural interests have begun to employ computers to access irrigation requirements. Their use has proven cost- effective in optimizing water input- crop yield relationships. Unfortunately, this computer -based technology has not yet become reasonably accessiblble to the small -scale agriculturist. Small calculators provide an alternative, however.

Hand -held programmable calculators have existed for nearly 10 years. Early models were not capable of supporting software proficient in predicting PET. As with computers, programmable calculator design technology has improved a thousand -fold in the past decade. High -level scientific calculations and impressive data storage capabilities are now available in shirt- pocket size units, and presently only a slim distinction exists between programmable calculators and computers.The one indisputable difference is the portability of hand -held programmable calculators.

With the introduction of the HP -41C calculator (Hewlett- Packard Calculator Div.,Corvallis, OR) in1979, extensive PET prediction capabilities have become accessible to the small -scale agricultural community. The 41C is expandable through a wide range of peripheral devices to suit many calculation and data storage needs.Additionally, the 41C is now capable of interfacing with "real- world" analog signals through a new digital multimeter.

The 41C is the controller of an interactive array of peripheral devices collectively referred to as HP -IL (Hewlett- Packard Interface Loop). These peripheral devices include mass storage tape drives, printer -plotters, digital multimeters and relays.As with computers, the calculator communicates with the user in words as well as numbers. Like all Hewlett -Packard calculators, the 41C utilizes a Reverse Polish Notation (RPN) logic system that is extremely "byte" efficient and user friendly.

A RPN program has been developed for the 41C that utilizes the Penman method for determining PET predictions, based upon Doorenbos and Pruitt's (1975) comprehensive version of the Penman method, which they generalized for a wide range of climatic conditions.

The model occupies 1778 bytes of program memory, with an additional data storage requirement totalling 140 bytes. The program requires a HP -41CV calculator, or a HP -41C calculator with a quad memory module. A peripheral printer is also required, and a magnetic card reader or digital cassette drive is helpful.

The program was designed with particular attention towards the non -technical user.It reduces the complicated calculation procedures that were previously required, down to a simple data input procedure. The needed climatic data can be input in a variety of forms. Furthermore, the program is flexible in the type of units of the input data.

17a The Generalized Penman Hodel

The Penman combination model for evaporation (Penman, 1948) has been widely used for evapotranspiration estimates. The model is a simplification of the energy exchange processes at the evaporating surface, yet its fundamental basis has contributed to a general acceptance of results obtained over a wide range of conditions. Doorenbos and Pruitt (1975) have generalized the Penman equation to extend its application to an even wider range of environments. In doing so they have compiled a useful assortment of graphical and tabular functions that make the application simpler and less laborious. The RPN program in this paper includes algorithms for these functions, and provides a further reduction in the effort required for application.

The basic model, taken from Doorenbos and Pruitt (1975, page 29), is:

ETo* = W Rn + (1- W)f(u)(ea -ed) (not adjusted) (1)

radiation aerodynamic term term where: ETo* is the reference crop evapotranspiration in mm/day (not adjusted);

W is a temperature- related weighting factor;

Rn is net radiation in equivalent evaporation in mm/day;

f(u) is a wind- related function; and

(ea -ed) is the difference between the saturation vapor pressure at mean air temperature and the mean actual vapor pressure of the air, both in mbar.

To find ETo, the reference crop evapotranspiration, ETo* needs to be adjusted for day and night -time weather conditions.

Components of the RPM Program

The computations with this model are often rather tedious, because the wide range of acceptable data may be in many different forms and limits.The programmable calculator reduces this effort significantly. The components of the Penman model will be described separately and the algorithms identified.

Vapor Pressure

The moisture content of the air appears in various forms, and may be expressed in many different sets of units. The model uses millibars for units in calculating the vapor pressure deficit of the air (ea -ed), but relative humidity, dewpoint, or dry- and wetbulb temperatures can also be input into the program. The model calculations are for daylong averages of ea -ed.In practice, this can be obtained from an average of the maximum and minimum values.Even a single measurement of actual vapor content (ed) can give a reasonable estimate of the mean daily ed for use in the program.

The saturation vapor pressure (ea, mb) at mean air temperature ( tmean = (tmax - tmin) /2) is calculated in the program from Murray's (1967) approximation:

ea = A exp [B tmean /(c + [mean)] (2)

for tmean > 0 °C, and where A = 6.1078 mb, B = 17.2691°C and c = 237.3 °C.

Observed vapor pressure (ed) data may be available, in which case subtraction yields the desired (ea -ed).

Relative Humidity Input. The ed is estimated from mean relative humidity (RHmean = (RHmax RHmin) /2) from the relation ed = RH ea /100, where ea is calculated from Murray's formula (Equation at tmean.

Example: given tmax = 35 °C, tmin = 22 °C, RHmax = 80 %, RHmin = 30%. The program calculates ea = 38.91 mb at tmean = 28.5 °C, ed = 21.40 mb at RHmean = 55% and ea -ed = 17.51 mb.

Dewpoint Input.In this case, the ed is estimated directly from mean dewpoint temperature (tdew, °C) with Murray's approximation (Equation 2), and ea is again calculated with Murray's approximation using tmean.

lib Example: given tmax = 35 °C, twin = 22 °C, tdew = 17.5 °C. The program calculates ea = 38.91 mb at tmean = 28.50 °C, ed = 20.00 mb at tdew = 17.5 °C, and ea -ed = 18.91 mb.

Psychrometric Inputs.The program uses the general form of the aspirated (ventilated) psychrometric equation (see, for example, Fritschen and Gay, 1979) to calculate

ed = es -Y (tdry - twet) (3)

where es(mb) is saturation vapor pressure at twet ( °C) from Equation (2),Y is the psychrometric constant (Tr= 0.66 P /Po, mb / °C, and P /Po is the ratio of actual atmospheric pressure to that at sea level where Po = 1013.25 mb), and tdry - twet ( °C) is the wetbulb depression.Actual atmospheric pressure is not normally available,so the program estimates the ratio P /Po from the International Standard Atmosphere (List, 1949):

P /Po = ((288 - 0.00652)/288)5.256 (4)

where Z is elevation above sea level, in meters.

Example: given tmax = 35 °C, twin = 22 °C, and a midmorning observation of tdry 25 °C and twet = 20 °C, at an elevationZ = 0 m (sea level).The program calculates ed = 20.08 mb, ea = 38.91 mb at tmean = 28.5 °C, and ea -ed = 18.83 mb.

Wind Function

The wind function f(u) in Equation 1 expresses the mixing power of the atmosphere, as a function of daily wind travel measured at a height of 2m (U2, km /day). The program uses the formula of Doorenbos and Pruitt (1975):

f(u) = 0.27 (1 + 02 /100). (5)

The program corrects wind measured at other than 2 m height with an empirical approximation:

U2 = U2/(0.1877 Ln(Z) + 0.87025) (6)

where UZ is the wind run (km /day) measured at height Z (m) above the surface.

Example: given that wind run at 3 m height is 250 km /day.While Equation (6) calculates U2 = 232.24 km /day, the program rounds to 2 decimal places to match Doorenbos and Pruitt's (1975) table, and yields U2 = 232.50 km /day and f(u) = 0.90.

Weighting Factor

The weighting factor is Doorenbos and Pruitt's (1975) simplification of variables in Penman's original derivation (i.e.,V = A /( A+ Y), whereA is slope of saturation vapor curve, dea /dt (mb / °C) at tmean, and Y is the psychrometric constant which was defined earlier). The weighting factor W depends upon temperature and pressure. A is calculated in the program by differentiating Murray's approximation for ea and tmean, and incorporating the pressure correction for Y The pressure ratio (P /Po) is estimated in the program by the International Standard Atmosphere given earlier in Equation (4)

The program calculations are:

W = [ea(Bc) /(c + t)2] /[(ea(Bc) /(c + t)2) + 0.66(288 - 0.00652)/288)5'256] (7)

where ea is calculated at t = tmean with Equation 2, the constante B,c are defined in Equation 2, and Z is elevation above sea level in meters.

Example: given tmax = 35 °C, twin= 22 °C, elevation = 95 m. The program calculates W = 0.78, 1 - W = 0.22. Doorenbos and Pruitt (1975) show W = 0.77 for these data; the difference is probably due to rounding.

Net Radiation

The radiation term may lead to some tedious calculations, because the necessary data are often lacking and one must resort to various estimation approaches. If net radiation (Rn) data is available, then the Penman model can be applied quite easily.If not, then estimation procedures usually begin with an evaluation of net solar radiation (Rns), based upon extraterrestrial solar radiation (Ra), estimates of cloud cover from climatological observations, and an estimate of crop albedo. The net longwave exchange (Rol) is then estimated from meteorological characteristics, and the two estimates are combined to yield net radiation:

17c Rn = Rna - Rnl (8)

Note that Rn is shown as the difference between Rns and Rol in Equation (8), since daily Rnl is usually a negative quantity in the crop enviroment.

Solar Declination.If solar radiation (Rs) data is not available, it can be estimated from extraterrestrial solar radiation (Ra), as a function of latitude and time of year. This requires solar declination (DEC).The program calculates DEC( °) as follows (adapted from Ball, 1978, p. 253):

X _ [0.9856(DOY + 0.5)] - [0.2458333((YR Mod 4) -1) - 15.785 + .0095 YR] (9)

Y = (0.02 sin 2X + 1.916 sin X + X) - [(- 47/2750)YR + 111.24] (10)

DEC = arcsin (sin Y sin 23.42) (11)

where DOY is day of year (1 to 366) and YR is calendar year (e.g., 1982). The modulo function returns the remainder of YR divided by 4 (i.e., 1, 2, 3, or 0). When YR Mod 4 = 0, the program uses 4 instead of 0 in Equation (9). The program uses an algorithm from Ball (1978, p. 241) to calculate day of year from calendar date.

Example:given the date of July 20, 1972 (DOY = 202). The program calculates DEC = 20.58° at solar noon.

Extraterrestrial Solar Radiation.Given declination (DEC) and latitude (LAT), extraterrestrial solar radiation (Ra, in mm of evaporation) is calculated from known solar geometry as follows:

S = sin LAT sin DEC (12) C = cos LAT cos DEC (13) N = (arcos (- S /C))/7.5 (14) V = N/24 (15) Ra = (48/)[VS + C sin(57.29578V)] (16)

Example: at 30 °N, on July 20, 1972, Ra = 16.9 mm /day.

Solar Radiation.Solar radiation (Rs) is estimated from the relation

Ra = (0.25 + 0.50 n /N) Ra (17)

where n is observed number of hours of actual bright sunshine and N is maximum possible number of bright sunshine hours. The value of N for a given date and latitude is given in Equation (14).

Net solar radiation (Rns) is then obtained by the simple relation

Rna = (1 - a)Rs (18)

where a is the albedo of the evaporating surface. a is about 0.25 for green crops.

Example:given observed n = 11.5 hrs on July 20, 1972, over a crop with albedo of 0.25. The program calculates Rs = 8.5 mm /day, N = 13.67 hrs, and Rns = 8.5 mm /day.

Longwave Radiation. The next longwave radiation (gal, in mm of evaporation) is obtained in the program as a function of air temperature, vapor pressure and sunshine, using the simple formulation from Doorenbos and Pruitt (1975):

Rnl = oTk4(0.1 + 0.9 n /N) f(ed) (19) -9 where a is 2 X 10 mm /day °R4, Tk is mean daily air temperature in °R (Tk = t( °C) + 273.16), n/N is ratio of observed hours to possible hours of bright sunshine, and f(ed) is 0.56 - 0.0794a in humid climates, and 034 - 0.044/ in dry climates.

Example: given tmean = 28.5 °C, ed = 21.4 mb, n/N = 11.5/13.67, and a dry climate.Rnl = 1.94 mm /day.

Net Radiation. Finally, the desired net radiation is obtained from Equation (8). For this example, Rns = 8.5 mm /day, and Rnl = 1.94 mm /day, so Rn = 6.6 mm /day.

Application of the Model

The program output must be adjusted to account for failure of the aerodynamic portion of the Penman model to fully describe the energy exchange processes.

17d Unadjusted Reference Crop Evapotranspiration

The initial application of the model yields unadjusted reference crop evapotranspiration (ETo *). This is illustrated with the following example from Doorenbos and Pruitt (1975), portions of which have already been used in this paper. Location is Cairo, at latitude 30 °N and elevation 95 m. Data is for July 1972. Solar radiation on the date July 20 is taken to represent the mean for the entire month. Other data for July are: tmean = 28.5 °C, RRmean = 55%, sunshine n = 11.5 hr., U2 = 232 km /day, anda= 0.25.

The program calculates (1 - W)f(u)(ea - ed) = 3.74 mm /day, and W Rn = 5.14 mm /day, so ETo* - 8.88 mm /day.

Adjusting for Climatic Characteristics

Doorenbos and Pruitt (1975) surveyed many data and concluded that the unadjusted values need little correction when daytime wind speeds are about twice those during night -time, and RR maximums are > 60%. They identified six other climatic regimes, however, when ETo* may differ substantially from ETo ('03 ETo* < ETo <-1.2 ETo *). Their classification follows:

Case #1: ETo - 1.2 ETo *. Important in areas with moderate to high radiation (>8 mm /day) during summer months, which consistently have low- humidity winds during much of the day ( >4 m /sec) and calm night -time conditions with high night humidity values approaching 100%.

Case #2: ETo-1.05 to 1.10 ETo *. Applies in areas where daytime winds are <4 m /sec and nights are very calm and humid,i.e. REmax > 75 %. The daily wind distribution should give a day -night ratio of 3 or more.

Case #3: ETo = ETo *. Daytime wind speeds of about twice those during the night -time are generally found and maximum relative humidity > 60 %. No adjustment of Penman ETo* is required.

Case #4: ETo -0.75 to 0.95 ETo *. ETo will be slightly over predicted in areas with moderate tc high radiation and with wind speeds < 4 m /sec but about equal during the day and night.Relative humidity during night -time is > 60 %.

Case #5: ETo-0.65 to 0.80 ETo* 1/.Relatively rare but applies for spring, summer and autum conditions with moderate to high radiation and when wind speeds during the day and night are between and 8 m /sec with maximum relative humidity < 40%.

Case #6: ETo - 0.55 to 0.65 ETo* 1/. Also rare, but applies during spring, summer and autur with moderate to high radiation and when wind speeds during day and night are > 8 m /sec and wi relative humidity day and night < 40 %.

Case #7: ETo - 0.30 to 0.35 ETo* 1/. Very rare, but will apply under very strong winds of > I m /sec during both day and night, while relative humidity, day and night, is low and < 40 %.Case 7 applies only when radiation is low, i.e. during late autumn and winter (Rns < 4 mm /day).

Doorenbos and Pruitt (1975) used graphical adjustments to correct ETo *. The program uses the linear relationship.

ETo = ETo* (1.629 - 0.21046 C)1 /2 (20) where ETo is adjusted reference crop ET (mm /day), ETo* is unadjusted ET (mm /day) and C is case number (1 to 7) as identified above.

Example: Doorenbos and Pruitt (1975) conclude that Case #4 characterizes the Cairo climate. Adjustment of ETo* by Equation (20) yields ETo = 7.55 mm /day.

Conclusions

Doorenbos and Pruitt's (1975) generalized version of the Penman equation should reduce the number of anomalous results that occur when the equation is applied in environmental conditions quite different from those of the original site in the English countryside. The generalizations, bower, introduce some additional considerations which Doorenbos and Pruitt handle neatly with supplementary graphs and tables. The compression of this supplementary material into appropriate algorithms, and organization into the convenient program described in this paper, will further help bring the Penman equation to bear on practical problems of ET estimation.

1/. These drastic corrections occur when relative humidity is low and conditions are very windy. Such conditions are rare and persist only a few days in most climates (Doorenbos and Pruitt, 1975) .

17e The programfits easily into the HP -41CV handheld calculator. This calculator is powerful, portable and relatively easy to operate. While the program works well for our purpose, it could undoubtedly be changed and improved.The identification and inclusion of the basic algorithms in this paper will make it easy for other users to adapt the Penman model to their special needs.

The current version of Robert Greenberg's Penman program with user's guide can be obtained from the authors for the cost of materials ($10 includes 11 magnetic cards in "write all" format; make checks to the University of Arizona). It can also be supplied on a micro -casette for the HP -IL cassette drive. The authors would appreciate comments and suggestions.

References Cited

Ball, John A. 1978. Algorithms for RPM Calculators. John Wiley. 330 pp.

Doorenbos J., and W. D. Pruitt. 1975. Guidelines for Predicting Crop Water Requirements. Irrig. 6 Drainage Pap. 24, FAO, Rome. 179 pp.

Fritachen, L. J., and L. W. Gay. 1979. Environmental Instrumentation. Springer. 216 pp.

List, R. J. 1949. Smithsonian Meteorological Tablee. Smithsonian Misc. Coll. Vol. 114. 6th. Rev. Ed., fifth reprint issued 1971. Smithsonian Institute Press. Washington.

Murray, F. W. 1967. On the computation of saturation vapor pressure. J. Appl. Meteorol. 6:203 -204.

Penman, H. L. 1948. Natural evaporation from open water, soil and grass. Proc. Royal Soc. (London) Ser. A. 193:120 -145.

Acknowledgements

The work in this publication was supported by the Arizona Agricultural Experiment Station. Hatch Project 04. Approved for publication as Journal Paper 420, Arizona Agricultural Experiment Station.

Appendix

The program outputs are listed below for the example problem given in the text. a. Executing programPN generates aheader b. The program prompts for max, tmin, RHmax, and a listing of the inputsrequired to and RHmin, and then calculates tmean, run theprogram, and the output to be RHmean, ed and ea -ed: obtained:

* * * * * * * * * * ** * * * * * * * * * * ** * MODIFIED * 'NOR PRESSURE: * PENMAN * * ET MODEL * T'sax'= 35.88 * * * * * * * * * * ** T'sin'= 22.88 T'sean'= 28.58 ssssssasss+asssssasssss ea= 38.91 RH'sax'= 80.881 t INPUTS * RH'sin'= 38.881 RH'icean'= 55.881 degrees ed= 21.48 C. ea-ed= 17.51 gauge ht. s. wind run ks/day altitude c. The program next prompts for wind travel s. and height of measurement, and calculates radiation wind speed at 2 m and value of f(u): se./day esssts**sss**asssssset** sssssstasstssasttsssss t OUTPUTS t 800 FUNCTION:

E. T. an./day u3.88= 258.88 ************ u2= 232.58 1(u)= 8.98

17f d. Theprogram prompts foraltitude, and calculates weighting factors W and 1 - W using tmean from before:

ssssssssssssssssssssssss

KICKING FACTOR:

altitude= 95.88 N= 0.78 I -N= 8.22

e. The radiationcalculations begin with a prompt for observed n, followed by N. Since N is not known in this example, the program will prompt for year, month and day of month. The program outputs day of year, and prompte for latitude. It then calculates N and n /N. It prompts for ka which if unknown (as in this example) is then calculated and used to estimate solar radiation. The program requests albedo and calculates net shortwaveradiation. The longwave calculations beginwith a prompt for dry or wet climate designation, followedby calculation of the longwave components, net radiation, and evaluation of Penman ETo* (unadjusted).

assasssssssasassssssss

NET RADIATION:

n= 11.58 DATE= 7./ 28.' 72. DAY 282. 26= 28.58 latitude=38.88 N= 13.67 n /N= 8.84 Ra= 16.98 Rs= 11.33 Albedo= 8.25 Rns= 8.58 DRY Climate f(ed)= 8.14 f(n /H)= 8.86 f(t)= 16.31 Rn1= 1.91 Rn= 6.59

ssssssssssssss»sssssses ETos = 8.64 nit./day sees

17g CHANGES IN STREAMFLOW IN AN HERBICIDE- TREATED JUNIPER WATERSHED IN ARIZONA

Malchus B. Baker, Jr. Rocky Mountain Forest and Range Experiment Station Flagstaff, Arizona

Abstract A 147 -ha juniper watershed in north -central Arizona was sprayed with an herbicide mixture to killall overstory vegetation.After the area was sprayed, annual water yield increased significantly when flow was greater than 12 mm.The ratio of mean annual quick flow to event flow prior to treatment was 0.86 and remained essentially the same after treatment.The herbicide treatment reduced evapotranspiration losses and increased water yield by killing the overstory trees and leaving them in place. These dead trees provided some shade and wind resistance and created a microclimate that reduced evaporation and enabled the soil below 30 cm to remain above its soil moisture wilting point. Although mean annual water yield increased by 27% (6 mm, 8 -year mean), this increase may not be practical from a management view point.Therefore,itis unlikely that extensive juniper acreage will be treated.The amount of treated acreage will depend on the demand for water and on the value of water in the market place.The area treated will also be constrained by consideration of other resource values and desires of the public.

Introduction The pinyon -juniper woodland typeis between 1,220 and 1,980 m above sea level and occupies extensive areas in the Southwest [Spencer, 1966; Arnold et al., 1964; U.S. Department of Agricul- ture, Forest Service, 1958].Pinyon -juniper covers about 21 million ha in Arizona, New Mexico, Colo- rado, and Utah [Clary et al.1974].The distribution and number of trees in this woodland type has increased since the early 1900's because of increased livestock grazing and because of the reduced number of forest fires [Arnold and Schroeder, 1955; Barr, 1956]. Arnold and Schroeder [1955] reported that encroachment of pinyon -juniper reduces grazing capacities,increases erosion, increases livestock handling cost, and possibly decreases water yield.Barr [1956] estimated a probable increase in annual water yield of 13 to 25 mm from the removal of pinyon - juniper inArizonaif the land were reseeded with grasses. However, ina study of pinyon -juniper removal,Collings and Myrick [1966] found no significant change in water yield.Skau [1964a and 1964b]reported that any change inwater yield from pinyon -juniper removal would probably result from reduced interception and evapotranspiration losses.Clary et al.[1974], after reviewing all documented hydrologic results from the southwestern pinyon- juniper type, stated that the possibility of increasing water yield through overstory removal was only marginal. The Beaver Creek research watersheds in north -central Arizona were established by the USDA Forest Serviceinthe mid- 1950's to quantify treatment effects on water yield from pinyon -juniper woodland and ponderosa pine forests.The Forest Service has studied two types of vegetation manip- ulationinpinyon -juniper woodlands: clearing of the overstory and killingof the overstory with herbicides [Clary, et al.,1974]. Preliminary results of the herbicide treatment have previously been reported by Clary etal.[1974]. This paper documents results of the 8 -year period following the herbicide application on a 147 -ha watershed.

Study Area The two watersheds usedinthis study are about 80 km south of Flagstaff,Arizona,in the pinyon -juniper woodland type of the Beaver Creek drainage (Figure 1).Watershed 3, which had the herbicide applied toit,is 147 ha in area; watershed 2, an untreated control,is directly adjacent and 51 ha in size (Table 1).

The soils in the area are developed from volcanic basalt parent material. The predominate soilis a Springerville very stony clay that is about 112 cm deep and has a clay texture throughout [Williams and Anderson, 1967]. Most of the clay fraction is montmorillonìte, which produces pronounced swelling

19 TO FLAGSTAF BEAVER CREEK WATERSHED Miles st=asst0 I 2 3 Watershed Boundaries Vegetation Type Lines Improved Roadways --:Unimproved Roadways '..Stoneman

Comp Semi - Desert Utah Juniper AlligatorJuniper I Pond Pine tarde Figure 1. -- Location of watershed 2 and 3 in the Beaver Creek drainage area.

Table 1. Physical Characteristics for Two Experimental Watersheds on Beaver Creek

Herbicide Control Characteristic Watershed Watershed

Size (ha) 147 51 Aspect W NW Mid -Area Elevation (m) 1573 1591

Basal Area (m2 /ha) 13 12 Annual Precipitation Before (mm) 453 466 After (mm) 426 425 Annual Streamflow

Before (mm) 22 25

After (mm) 28 20

and shrinking. After each wet and dry cycle, cracks form, opening as much as 5 cm wide and 1 m deep. Mean elevationof watershed 3 is 1,573 m above sea level,ranging from 1,516 to 1,676 m (Table 1). The general aspect of the watershed is west with 92 percent of its area in slopes of 10 percent or less.

Priorto application of the herbicide, watersheds 2 and 3 were stocked with Utah juniper (Juniperus osteosperma (Torr. )Little) and pinyon pine (Pinus edulis (Engelm. )). Total basal area on watersheds 2 and 3 averaged 12 and 13 m /ha, respectively, with an average of 33 trees per ha (trees 13 cm d.b.h. and greater).

Mean annu$1temperature is13.3° C, and average monthly temperature varies from 3.3° C in January to 25.0C inJuly. Mean annual presítpitation on watershed 3is 442 mm, with a high of

20 691 mm in water year 1973followed by a low of 210 mm in 1974 (Figure 2).Mean precipitation during the seven winter months, October through April, is 270 mm, or 61 percent of the annual total. Two major precipitation seasons characterize this area.The most important period from a water yield standpoint is the winter season from October through April.Most of the remaining precipitation falls during July, August, and September.

Winter precipitation is associated with widespread, protracted, frontal storms.Snowfall commonly begins in November and increases into December.Snowfall usually declines in January and February, followed by an increase to a peak in March.Typically, the snowpack is intermittent throughout the season at this elevation [Baker, 1981]. Nearly all summer precipitation falls during thunderstorms.Although they are frequent, individual storms usually cover relatively small areas.Mean summer rainfall averages 172 mm. Continuous streamflow records on both watersheds are obtained from precalibrated, concrete, trapezoidal flumes [Clary et al., 1974].Mean annual streamflow from the herbicide treated and control watersheds is presented in Table 1.Annual streamflow on watershed 3 has varied from zero in 1959, 1963, and 1974 to 149 mm in 1973 (Figure 3).Mean annual streamflow from watershed 3 prior to the herbicide treatment (1958 -1968) was 22 mm, with 82% occurring from October through April as a result of snowmelt or winter rains. Most of the water yield from the Beaver Creek experimental watershed is produced in the winter. During the period of this study on watershed 3, in 11 of the 19 years, 96% or more of the annual discharge was during the winter season.The soil mantle begins to recharge during the winters, when water from rain or snowmelt is available and evapotranspiration demands are low.Most streamflow isin March and April.The stream channels in these headwater basins become dry before the onset of summer rains. Summer rains are preceded by the two driest months of the year, May and June, during which precipitation amounts average 13 and 10 mm, respectively.Because of the high evapotranspiration demands and the distribution and amount of precipitation during the summer on watershed 3, only 4 summer seasons out of 19 years have had 6 mm or more of streamflow.Although, summmr streamflow isinfrequent, maximum peak discharges from watershed 3 of 7.7 and 5.1 ms km occurred in August, 1964 and Septemlr, _11970respectively. Peak discharges from the same storms on watershed 2 were 9.3 and 4.1 m s km . Much of the streamflow from these upland or headwater basins is the product of quick flow [Hewlett and Hibbert, 1967].The shallow A- horizon (about 8 cm deep, with infiltration rates ranging from 20 to 64 mm /hr) on these watersheds, covers a relatively impermeable soil layer (hydraulic conductivities of less than 1 mm /hr).Rainfall and snowmelt intensities greater than these infiltration and permeability rates are not uncommon on Beaver Creek [U.S. Department of Commerce, 1967; Ffolliott and Hansen, 1968].Water storage capacity of the Springerville soil series is over 46 cm.

1,00

900

800 Summer Winter 700 Winter Average 600 Average

500

400

300

200

100

58 59 60 bl fi2 63 64 65 66 fil 68 69 70 71 73 i74 72 75 76 hi Water Year

Figure 2. -- Average winter (October -April) and summer precipitation on watershed 3.

21 Treatment and Analysis Watershed 3 was treated with a foliage spray application of 2.8 kg acid equivalent of picloram and 5.6 kg acid equivalent of 2, 4 -D [(2, 4- dichlorophenoxy) acetic acid] as triisopropolaminesalts in 94 1 of water per ha.The watershed was treated in August 1968, and the overstory inventory of October 1969 showed that 86 percent of the trees were killed. The paired watershed method [Wilm, 1943; Kovner and Evans, 1954; Hewlett, Lull, andReinhart, 19691 was used to evaluate changes in annual water yield brought about by the herbicide treatment. A covariance analysis was used to test significance of the pretreatment and posttreatmentannual water yield regressions. Confidence limits (95% level) were placed on the pretreatment regression, and posttreatment data were located with reference to these limits to determine significance of treatmentin the individual years [Harr, Frederiksen, and Rothacher, 1979].

Effects of Herbicide Application Streamflow for 19 years on watershed 3 has been divided into two periods (Figure 2).The first 11 years are the pretreatment or calibration years (1958- 1968), and the last 8 years(1969 -1976) are the treatment period following application of the herbicide.

Annual and Seasonal Flow The11 yearsof pretreatment data (1958 -1968) define the regression (correlationcoefficient (r) = 0.99) used in determining posttreatment effects on annual water yield.A covariance analysis ofthepre-and posttreatment regressions (1969 -1976,r = .98),indicatesasignificant increase (0.05 level) in annual water yield following the herbicide application.The analysis shows that the treatment response is multiplicative and is significant only when flow is greater than 12 mm(6 mm on the control watershed).In drier years there is a predicted 8 mm increase in water yield when the control watershed yields 13 mm.In wet years, when the control yields 102 mm, the herbicide treated watershed is predicted to yield an increase of 28 mm or 121 mm. Killing the overstory vegetation caused annual water yield to equal or exceed the 5%prediction limit 3 out of the 8 posttreatment years (years 1969, 1970, and 1973).Of the other 5 years, 3 years (1971,1972, and 1974) received below average winter precipitation, and all but 1976 received below average annual precipitation (Figure 2).The fact that 1975 received normal winter precipitation and

160

140

i 120

100

â 40

72 73 74 75 76 58 59 60 61 62 63 64 65 66 67 68 69 70 71 Water Year

Figure 3. -- Annual water yield from watershed 3.Increase in water yield due to treatment determined from pretreatment regression.

22 1976 received normal winter and annual precipitation without showing a significant increase may indicate that the effect has ended. Prior to the herbicide application, the ratio of mean annual quick flow to event flow was 0.86 on watershed 3 and 0.88 on watershed 2(Table 2).There was little change in this relationship after the application of the herbicide on either the treated or control watershed.

Discussion The hydrologic regimes of the forested areas on volcanic- derived soils on Beaver Creek are different from many other forested areas.Because of the relatively shallow soil depth, the impermeable clay soil, and the predominately short spring streamflow period, manipulation of the forest vegetation has a limited affect upon the streamflow regimes from these upland watersheds.In the hydrograph analyses used, the term quick and delayed flow are used to define the hydrograph components because of the difficulty in defining and quantifying direct runoff and baseflow [Hewlett and Hibbert, 1967] .Most forested areas have a higher portion of their streamflow produced by the delayed flow component. On Beaver Creek the higher percent of streamflow results from quick flow because of their characteristic soil and climatic conditions. The herbicide treatment applied on watershed 3 reduced evapotranspiration losses, as shown by the increase in annual water yield and the decrease in soil moisture loss.The soil mantle (with an available soil water storage of 46 to 56 cm) begins to recharge annually during the fall and winter period on both treated and untreated areas, as observable in the saturated surface soil and overland flow conditions during most spring runoff periods and as apparent in the high ratio of quick flow to event flow. The increase in water yield was achieved by killing the overstory trees and leaving them in place.These dead trees provided shade and wind resistance and created a microclimate which reduced evaporation and enabled the soil below 30 cm to remain above its soil moisture wilting point [Johnsen, 1980].With reduced soil water loss, less precipitation was needed to recharge the soil mantle and consequently more precipitation became available for streamflow. Analysis of the annual water yield response showed a sharp increase after the herbicide application. The increaseisbelieved to be largely the result of reduced evapotranspiration loss. An additional contributing factor would be the increased exposure of soil following the defoliation of the overstory trees.The soil surface would be particularly susceptible to raindrop impact and subsequent soil surface "puddling" and reduced infiltration rates.The initial increase, 65% as reported by Claryetal.(1974), was reduced after 1972 and remained consistent through 1976.By 1973 the watershed would have had the opportunity to stabilize.Herbage production increased following the treatment and the additional ground cover would provide protection for the soil surface and contribute to an increase in soil permeability.

Table 2.Effects of Killing Pinyon -Juniper Trees on Mean Quick Flow and Delayed Flow

Mean Mean Quick flow Delayed flow

Control Herbicide Control Herbicide Period Watershed Watershed Watershed Watershed

Water Year (Oct -Sept) Before Treatment 22 19 3 3 After Treatment 18 23 2 5

Summer Before Treatment 5 4 0 0 After Treatment 2 3 0.4 0

Winter Before Treatment 17 15 3 3 After Treatment 16 20 2 5

23 Although the increase in annual streamflow is statistically significant on the treated watershed for flow events of greater than 12 mm, the 27% (6 mm) increase in mean annual streamflow (8 -year mean) may not be a practical or a significant increase from a management view point.To make this treatment economically feasible,a land manager might need additional increases in other resources such as additional grazing caused by increased herbage production and increased benefits for wildlife. The results also indicate, but were not experimentally verified, that the treatment effect may only last 5 -8 years and that the watershed will then have to be retreated to maintain the increased flow. The reservoir system in Arizona depends on water yields derived primarily from winter precipitation.Although a significant increase in annual streamflow can be realized by killing overstory vegetation on an upland basin,it cannot be assumed that such a treatment will produce an increase in water yield in the valley below. Because of the limited spring streamflow period, most water yield increases resulting from vegetation manipulation will be added to flowing stream channels.Therefore, the water yield increase or a significant portion of the increase is less likely to be lost in transmission from the headwater basins tothe reservoir system.However,as Hibbert [1979] observes, an increase in water yield from treatments on small upland watersheds is no assurance that the increases can be measured at the downstream reservoir even if transmission losses are negligible, because the increased flows may not be detectable after combining with flows from other sources. There is also little likelihood that extensive acreage will ever be treated as in this study. As Hibbert [1979] states, the actual water yield increases from vegetation manipulation will likely be less than the maximum potential because only a portion of each vegetation type can be treated economically for water yield increases.The amount of acreage that can be treated will depend on the demands for water and on the value of water in the market place.The amount of acreage treated will also be constrained by considerations of other resource values and desires of the public. These results show that water yields can be increased on basins with volcanic soil by chemically killing the juniper overstory and leaving the dead trees standing.Therefore a management system could be devised where dispersed upland juniper basins are chemically treated to provide some additional water yield.However, these treated areas may not be esthetically pleasing and the use of chemicals in the environment might be undesirable.

References Cited Arnold, J.F., D. A. Jameson, and E. H. Reid.The pinyon -juniper type in Arizona:Effects of grazing, fire, and tree control.USDA Prod. Res. Rep. 84, 28 p, 1964. Arnold, J. F., and W. L. Schroeder, Juniper control increases forage production on the Fort Apache Reservation. Stn.Pap.18,35 pp., USDA Forest Serv., Rocky Mt.Forest and Range Exp. Sta., Fort Collins, Colo., 1955. Baker, M. B., Jr., Hydrologic regimes of forested areas in the Beaver Creek watershed.Gen. Tech. Rep. RM -88, 8 pp., USDA Forest Serv., Rocky Mt. Forest and Range Exp. Sta., Fort Collins, Colo., 1981. Barr, G. W., Recovering rainfall, part 1, 33 pp., Arizona Watershed Program, Arizona State Land Department, Water Division,Salt River Valley Water Users' Association and the University of Arizona, 1956. Clary, W.P.,et al.,Effects of pinyon -juniper removal on natural resource products and uses in Arizona, Res. Pap. RM -128,28 pp., USDA Forest Serv., Rocky Mt.Forest and Range Exp. Sta., Fort Collins, Colo., 1974. Collings, M. R., and R. M. Myrich, Effects of juniper and pinyon eradication on Pap. 491 -B, 12 pp., U.S. Geol. Surv., 1966. Ffolliott,P.F., and E. A. Hansen, Observations of snowpack accumulation, melt, and runoff on a small Arizona watershed, Res. Note RM -124, 7 pp., USDA Forest Serv., Rocky Mt. Forest and Range Exp. Sta., Fort Collins, Colo., 1968. Harr, R. D., R. L. Fredriksen, and J. Rothacher, Changes in streamflow following timber harvest in southwestern Oregon, Res. Pap. PNW -249, 22 pp., USDA Forest Serv., Pac. Northwest Forest and Range Exp. Sta., Portland, Oreg., 1979. Hewlett, J. D., and A. R. Hibbert, Factors affecting the response of small watersheds to precipita- tionin humid areas,inInternational Symposium on Forest Hydrology, pp.275 -290, Pergamon Press, New York, 1967. Hewlett, J. D., H. W. Lull, and K. G. Reinhart, In defense of experimental watersheds, Water Resour. Res., 5(1), 306 -316, 1969.

24 Hibbert, A. R., Managing vegetation to increase flow in the Colorado River Basin.Gen. Tech. Rep. RM -66, 27 p.USDA Forest Serv., Rocky Mt. For. and Range Exp. Stn., Fort Collins, Colo., 1979. Johnsen, Jr., T. N., Picloram in water and soil from a semiarid pinyon -juniper watershed, J. Environ. Qual. 9(4), 601 -605, 1980. Kovner, J. L., and T. C. Evans, A method for determining the minimum duration of watershedex- periments, Trans. Amer. Geophys. Union, 35(4), 608 -612, 1954. Skau, C. M., Interception, throughfall, and stemflow in Utah and alligator juniper cover types of northern Arizona, For. Sci. 10, 283 -287, 1964a. Skau, C. M.,Soil water storage under natural and cleared stands of alligator and Utah juniper in northern Arizona Res. Note RM -24,3 pp., USDA Forest Serv., Rocky Mt. Forest and Range Exp. Sta., Fort Collins, Colo., 1964b. Spencer, Jr., J.S., Arizona's forests.Resour. Bull. INT -6, 56 pp., USDA Forest Serv., Intermt. Forest and Range Exp. Sta., Ogden, Utah, 1966. U. S. Department of Agriculture.Forest Service, Timber resources for America's future, For. Resour. Rep. 14, 1958. U. S. Department of Commerce, Rainfall frequency atlas of Arizona for durations of 6 -24 hours and return periods from 2 to 100 years, 1967. Williams,J. A., and T. C. Anderson, Jr., Soil Survey of Beaver Creek area, Arizona.USDA Forest Serv., and Soil Conserv. Serv., and Ariz. Agric. Exp. Sta., 75 pp., Washington, D. C., 1967. Wilm,H.G.,Statistical control of hydrological data from experimental watersheds, Trans. Amer. Geophys. Union, 2, 618 -622, 1943.

25 DETERMINING WATERSHED CONDITIONS AND TREATMENT PRIORITIES

by Rhey M. Solomon James R. Maxwell Larry J. Schmidt USDA Forest Service Albuquerque, New Mexico

Abstract

A method is presented for evaluating watershed conditions and alternative watershed treatments. A computer model simulates runoff responses from design storms. The model also simulates runoff changes due to management prescriptions that affect ground cover and structural treatments. Tech- niques are identified for setting watershed tolerance values for acceptable ground cover and establishing treatment priorities based on the inherent potentials of the watershed.

Introduction

Watershed conditions have historically been an important issue in the Southwest and Intermountain West. Early in this century, forest reserves were set aside primarily to sustain favorable conditions of flow for downstream users. During the period 1910 -1960, considerable research was devoted to inves- tigating relationships between vegetation, soil, ground cover, and runoff. Classic studies revealed a strong relationship between ground cover of plants and litter and surface runoff (Croft and Bailey, 1964; Meeuwig, 1960). As ground cover decreases, surface runoff increases especially from intense summer storms (Coleman, 1953). Primed with this knowledge, Federal land management agencies began adjusting livestock numbers to conform with land capability, aggressively suppressing wildfires, and installing runoff controls such as contour trenches. Many of these controls were installed by the CCC in the 1930's to improve or protect vegetation cover.

By the 1960's, many communities that had experienced flood damages associated with poor watershed condition now enjoyed the benefits of managed watersheds with revitalized cover. Many who experienced the consequences of poor watershed condition have passed on. The new generation, and the influx of people unaware of local history, generally take the absence of severe floods in managed areas for granted. In some cases, people have developed lands that were once active floodplains.

In the Southwest, the emphasis of the 1960's was on increasing water yield through vegetation changes. The 1970's spawned a degree of environmental awareness with an emphasis on water quality. Most universities shifted attention to these areas. These two items had scientific and media appeal, and overshadowed the traditional watershed condition concerns. The 1980's have given us stronger emphasis on commodity production, and population shifts to the Southwest. These pressures are re- awakening a concern for watershed conditions among land managers.

Periodic summer floods have renewed interest in the role of watershed condition in regulating flood peaks. This has encouraged a reevaluation of past research to find means of assessing watershed condition. Some of the most relevant research was completed before the age of computer libraries and was difficult to discover. However, a strong body of knowledge exists which relates vegetation cover and soil conditions to runoff.

A technique for linking plant and litter cover to peak flow is fabricated from this knowledge base and builds upon the approach outlined by Lull (1949). We have developed a process model that ties ground cover to peak runoff for a specified design storm typical of the summer season in the South- west. The model can determine the potential for reducing peak flows by increasing ground cover and installing structural treatments. The procedure also proposes a rationale for determining a minimum tolerance cover necessary to protect the site and downstream values.

This paper outlines and documents an approach for quantifying watershed condition.We first discuss the computer model used to simulate runoff. Using this computer tool, we show examples of how changes in watershed conditions alter peak flows, and discuss the ability to improve runoff responses through changes in vegetative cover and structural treatments. To adequately translate

2 7 this modeling tool to an easily understood concept, we introduce the concept of "tolerance" and a "watershed condition index ".

Modeling Approach

A principal indicator of watershed condition is the fluvial system's ability to transport water through a watershed without damaging channel stability, floodplain improvements, or riparian values. Because flow energies, area of inundation, and unstable conditions are greatest during peak flows, we need to assess the effects of land management on the magnitude and frequency of damaging peak flows. Our approach is to model the hydrologic processes affecting runoff and then look at effects management could have on these processes.

The two principal hydrologic processes which management can influence are infiltration and surface detention storage. We can, therefore, propose land treatments that affect these two components. Opti- mum infiltration and storage are best accomplished by assuring an adequate ground cover of plants and litter (Rosa, 1954; Croft and Bailey, 1964). We can also increase them by such activities as soil rip- ping, plowing, or other techniques that break up compacted or sealed surfaces (Croft and Bailey, 1964). These techniques are short lived unless vegetative cover is increased in conjunction with initial treatment.

Horton (1933) concluded that streamflow consists of two components: (1) direct overland runoff and (2) ground water flow. He proposed that storm flows occur when rainfall exceeds infiltration rates and that base flows are fed by ground water aquifers. This exceedance of infiltration is apparent to anyone who has been caught in an intense Southwestern rainstorm. However, this is not the sole source of runoff in forests or where rainfall intensities are moderate (Hibbert, 1975). We focused on overland flow because of our ability to affect this component and its importance in generating peak flows in the Southwest.We also incorporated the variable source area concept (Hewlett and Nutter, 1970) into the modeling approach, because it is important, and perhaps the dominant process, in forests that are in good watershed condition.

Our goal was to model runoff processes and express them in simple terms that have meaning to the manager as well as the scientist. To meet this goal, we searched for a model that best met the following objectives:

1. Sensitive to hydrologic processes that management can affect.

2. Theoretically defensible.

3. Requires minimal data input (i.e., data that is readily available or easily estimated).

4. Applies to the full range of conditions in the Southwest.

5. Applies to small (1 -10 sq. mi.) watersheds where management practices can affect hydrologic responses.

6. Applies to individual rainfall events.

All major hydrologic processes which occur during rainstorms can be simulated in detail with the current state -of- the -art. The problem is to mirror these complex processes using only readily available data. Therefore, our efforts focused on models that consider only the most significant factors in simulating the processes controlling runoff.

Numerous models were reviewed. The "event" oriented models we investigated were: the SCS model (USDA, 1972) and various modifications (Ward et al., 1981; Knisel, 1980); the "Rational" Formula; HYMO (USDA, 1973); the Stanford Model (Crawford and Linsley, 1956); and USDAHL -70 (Holton and Lopez, 1971). Each of these models has merit but none meets all the listed objectives. The principal failing was an inability to model short, intense rainstorms and be sensitive to changes ininfiltration and storage. Accordingly, we were required to piece together parts from various models to best meet the objectives. The result was a simple computer model that is sensitive to land management treatments, requires a moderate amount of data, and appears theoretically sound. The model was developed in the belief that watershed conditions are best reflected by flood flows resulting from intense rainstorms. If satisfactory watershed conditions are a principal objective of management, then we need to demonstrate our ability to affect peak flow and timing of runoff.

The Model

The model SHOWER was developed by incorporating theoretical as well as empirical relationships that describe the processes outlined previously. Where data or theory were vague or absent, assumptions were made to fill these voids. A complete discussion of the model is given by Solomon (1982).

28 Precipitation is the driving input.This input is infiltrated for a designated time increment. Precipitation that does not infiltrate is cumulated in detention storage until the next time increment. Quantities in excess of infiltration and detention storage become surface overland flow. The model places this excess water into the channel system and routes it to the downstream point of interest.

Water that infiltrates the soil is added to the soil water and cumulates until the soil is filled to gravitational free water. At this point water is fed to the channel at a rate dependent on the amount of water in excess of free water holding capacity and the drainage density. This soil excess is added to the surface excess and routed to the downstream point of interest.

Watershed Condition Examples

Two examples portray the concept of watershed condition and our ability to affect storm hydrographs. Agricultural Research Service Watershed 47.003, near Albuquerque, New Mexico (Hickok et al., 1959), and Halfway Creek, near Farmington, Utah (Doty, 1971), were chosen because of available data describing treatment responses. These watersheds contrast sharply as shown in Table 1.

Table 1. Watershed Characteristics for Example Watersheds.

Characteristic Albuquerque 47.003 Halfway Creek

Area: 176 Acres 464 Acres

Average Slope: 10% 50%

Channel Slope: 3% 38%

Soils: Clay loam and silt shallow loam, loam; 2 -3 feet deep sandy loam

Vegetation: sagebrush, grasses, oakbrush, sagebrush, snakeweed, saltbrush aspen

Ground Cover: 25% 60 -80%

Mean Annual Precipitation: 8" 30"

These two watersheds illustrate runoff sensitivity to vegetative cover and structural treatments for actual storm events. We found that it is not the total volume of rainfall that dictates the hydrograph, but rather the distribution and intensity of rainfall. Many models fail to mimic hydrograph peaks because they fail to model rainfall input on a short enough time base.

Ground Cover Changes

Ground cover can have marked effects on surface runoff (Croft and Bailey, 1964; Woodward, 1943; Meeuwig, 1960). Watershed 47.003 shows how the computer model responds to cover changes and how dominant cover can be in determing peak flow responses. Three rainstorms are used: (1) August 3, 1964, (2) September 2, 1965, and (3) June 10, 1966. These storms were of moderate volume and intensity as shown in Table 2.They were simulated for different cover conditions.

Figure 1 shows the considerable influence management can exercise on peak flow. Increasing ground cover from 25 to 35 percent would cause about a 50 percent reduction in peak flow for all the storms modeled. Conversely, allowing ground cover to deteriorate from 25 to 15 percent would increase peak flows about 50 percent. Comparing different storms, the storm of September 2, 1965 would have produced the same peak flow as the June 10, 1966 storm if ground cover were 5 percent rather than 25 percent. This would translate to a 2 year storm producing a peak flow more characteristic of a 10 year storm.

Total storm runoff volumes are plotted in Figure 2 as a function of cover. These changes show the same pattern as Figure 1.Reduced runoff volumes have implications to reservoir management and translate to greater soil water for increased plant growth and base flows.

29 August 3, 1964 September 2, 1965 June 10, 1966

Accumulated Accumulated Accumulated Time Volume Intensity Time Volume Intensity Time Volume Intensity (Minutes) (Inches) (In /hr) (Minutes) (Inches) (In /hr) (Minutes) (Inches) (In /hrl

0 0 0 .48 .36 6 5 .04 5 .03 2 .2 .48 .12 5.1 10 .08 10 .04 4 .37 .48 .12 3.6 15 .12 15 .05 6 .49 .24 .12 5.1 20 .18 20 .06 8 .66 2.28 .12 4.5 25 .37 25 .07 10 .81 2.28 .12 2.7 30 .56 30 .08 12 .90 1.32 .36 2.4 35 .67 35 .11 14 .98 .48 .72 1.5 40 .71 40 .17 16 1.03 .24 1.08 .3 45 .73 45 .26 18 1.04 .12 2.28 .74 50 .45 1.32 55 .56 .06 60 .61 ,24 65 .63

Table 2. Precipitation Data for Three Storms on the Albuquerque Watershed 47.003.

100 80

60 200 % ?, \Q4,o° 100 \.\\ú. \ 4. 80m ":.60 11) o \ \\ 40 \ \'? OJ 30 s\\% LL 4 qu\epó ,\á < 20 . ° '1. W Ó°\\0 Ó °e; 1 o. P 1A ; ¡ ) \'\ N \ 'O °- 2 10 1 8 1

4 10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100 GROUND COVER(percent) GROUND COVER (percent)

Figure 1. Peak Flow Responses to Ground Cover Figure 2. Runoff Responses to Ground Cover Changes for Three Summer Storms. Changes for Three Summer Storms.

30 The next step in model verification and refinement will be to simulate hydrographs on watersheds where significant cover changes have actually taken place and hydrologic data are available. The classic Watersheds A and B of the Great Basin Station in Utah (Meeuwig, 1960) offer this opportunity and hopefully will confirm the preliminary results.

Structural Treatments

When cover decreases on a watershed due to overuse, rehabilitation can be difficult. Reestablish- ment of vegetative cover and favorable soil structure may require decades. In today's economy, managers must be concerned about the resource entropy associated with harvesting resources in excess of the land capability (Rifkin and Howard, 1980). The state of entropy comes from dissipating the soil productivity through erosion. This makes any stabilizing treatment far more costly than proper management would have been. Green (1971) captures this concept in his statement:

The most difficult job is to guard against negating the benefits of spending by accomodating public pressures for certain types of land use that can either act counter to the goals of watershed management or increase the management cost of rehabilitation activity.The public land manager is no less responsible for fiscal integrity than he is for biological, ecological and physical concern for the resource.

In many cases, recoverymay have to be aided with structural treatments. These treatments serve two purposes: (1) they reduce erosion and runoff immediately; and (2) they help reestablish cover by reducing water losses and increasing available water. Such structural treatments were applied to Halfway Creek (Doty, 1971). Before -and -after situations were modeled with SHOWER. Peak flows agree well with actual hydrographs for those storms. Additionally, SHOWER was used to simulate what the hydrograph from the storm in 1945 would have looked like with trenching and what the hydro - graph from the storm in 1965 would have looked like without trenching (Figure 3).

HALFWAY CREEK 20-

18 Storm of Aug. 19,1945 Storm of July 18,1985 16 d14 WITHOUT TREATMENT(edtud) d12 I,`WITHOUTTREATMENT(modeled)

510 I ` J 8 \ LL 6 \ WITH TREATMENT(modeled) ' 4 \ I WITH TREATMENT(actual) 2

1 2 3 4 5 1 2 3 TIME(hrs.) TIME(hrs.)

Figure 3. Pre -and Post -Treatment Hydrographs for Halfway Creek.

This type of comparison visually demonstrates the direct effects of structural treatments for an event rather than by traditional means such as paired watersheds. This method does not lend itself to statistical interpretation as other methods might, but the decision maker can readily identify the tradeoffs being made.

31 Cover Versus Structures

Through modeling, it is possible to compare hydrograph changes from increasing cover versus installing structural treatments. Numerous simulation runs were made on Watershed 47.003 for the storm of August 3, 1965. Figure 4 shows the increase in cover and corresponding amount of contour trenching or furrowing necessary to reduce peak flow by equal amounts. This diagram clearly demonstrates that both structural measures and cover can effectively control surface runoff. However, prudent planning should consider only the improvement in cover for evaluating long term benefits. Structural treatments, while effective, are relatively short lived and provide only temporary site control pending establishment of This vegetation. The structural measures may temporarily be more effective than the long term cover. is because the design of the structures must consider the risk of failure over the period required to attain satisfactory cover. This is a function of downstream consequences of structural failure, a level of risk, and the costs of retreatment if the measures are damaged. It certainly is not prudent or justified to expect more than temporary control of runoff from structural measures.

EQUIVALENT TREATMENT MEASURES

100

wáBÓ Storm of Sept 2,1965 xo, N e watershed 47.003 60 < ó «40 G e Z"I- mi w 20 ¢c) w G. 5 10 15 20 25 30 35 GROUND COVER INCREASE(percent)

Figure 4. Comparison of Treatment Effectiveness.

SHOWER offers a technique for evaluating peak flow alterations due to changes in cover and structural treatments. Additionally, roads can be incorporated into the model by considering the road network as one or more hydrologic units and applying the appropriate infiltration characteristics and runoff routing factors. Roads are also considered an extension of the channel network for soil water flow (Megahan, 1972).

The Concept of Tolerance

The previous section showed how reducing ground cover increases runoff. When do such increases become unacceptable? We must answer this question to distinguish between satisfactory and unsatisfactory watershed conditions.

Watersheds and land units have a maximum potential ground cover that can be achieved naturally. This condition prevailed over most of the Southwest before 1880. Persistent land disturbance progressively reduces ground cover from the potential. Eventually, runoff may increase to a point where it causes economic or environmental impacts which the land manager deems unacceptable. Ultimately, runoff may increase to a point where water flows and land productivity are permanently impaired.

The following discussion presents techniques to derive tolerance levels of ground cover for runoff. Two levels of tolerance are discussed: (1) a variable "management" tolerance responsive to social concerns for selected areas; and (2) an absolute "resource" tolerance responsive to permanent physical damage for all areas.

32 Management Tolerance

SHOWER calculates flood peaks for varying levels of ground cover. For each watershed, the land manager must decide how big an increase in flood peaks (how big a risk of increased flood damages) he is willing to accept (Schmidt, 1978). Watershed priority (Shaw et al., 1981) should be used to determine this marginal level of risk. Many low priority watersheds may not warrant a management tolerance because risks of flood damages are very low. If the flood peak calculated for existing ground cover exceeds the management tolerance set for a watershed, the watershed condition is unsatisfactory.

Resource Tolerance

Channel networks form as a function of the erosive force of runoff and the erodibility of the watershed. A channel forms where runoff becomes sufficient to incise the surface. In any area, there is some minimum drainage area or slope distance required for a channel to form (Schumm, 1977).

As ground cover is progressively reduced from the potential, the erosive force of runoff increases and the resistance of the surface to erosion decreases. Thus, the drainage area or slope distance required for a channel to form is reduced. Eventually, the channel network will expand headward in response to these changes. If unchecked, this process ultimately will severely degrade the watershed and permanently impair the discharge- sediment function of the fluvial system.

One approach for deriving a resource tolerance for runoff computes a critical slope distance, represented by complete expansion of the channel network up all slope depressions to the ridge. The resource tolerance is exceeded when ground cover is reduced enough to permit such a drastic response, which would occur over several decades.

Horton (1945) introduced the concept of the "belt of no erosion." For channel erosion, he defined it as the mean horizontal distance from a channel to the nearest ridge. It can be computed using two equations:

B 5280 (Equation 1) 2d

s1.67 B= 65 (Equation 2) f(sinl'17a)(cos'501a) r

where: B = belt of no erosion (feet), d = drainage density (miles per square mile), f = surface friction factor, a = slope angle (degrees), s = soil shear resistance factor (pounds per square foot), and r = maximum runoff intensity (inches per hour).

A "tolerance" B is computed by extending the channel network to the ridge and using equation 1. Equation 2 can be solved for any level of cover. Eventually, reducing ground cover will increase r and lower s to the point where the existing B (equation 2) declines to the tolerance B (equation 1). At that leveT of cover, watershed condition is unsatisfactory. If ground cover is not increased, radical channel expansion will ultimately occur.

Measure of Watershed Condition

A measure of cover can provide an index of watershed condition that is straightforward, responsive, and meaningful to management. Watershed conditions can be evaluated using ratings that represent the general hydrologic conditions of the land. These condition ratings are determined by comparing measured existing cover against estimated potential and tolerance cover. Lands with different natural capabilities are thus equalized; desert watersheds are not compared with forested watersheds for condition ratings. The watershed condition index is defined by: E - T W 1T Where:

W = watershed condition index, E = existing area -weighted cover, P = the natural maximum weighted cover, and T = the cover conditions necessary to prevent excessive risk of flood damage or channel expansion.

33 If existing cover is below tolerance levels the watershed condition is unsatisfactory (W< 0.0). Satisfactory conditions are defined as having W between 0.0 and 0.5. Optimum conditions are W values greater than 0.5. The distinction between satisfactory and optimum watershed conditions is arbi- trary and serves only to alert management to the conditions relative to potential and tolerance. A watershed in satisfactory condition meets minimum land stewardship requirements but is closer to tolerance than to potential cover. An optimum condition watershed is closer to potential.

A further expansion of the classification system is needed if the system is to have more meaning to management. In addition to knowing how well a watershed performs hydrologically, the manager needs to know where he can invest the fewest dollars for the greatest returns. This is done by putting watersheds into "opportunity classes." These opportunity classes are a measure of the difference between potential and tolerance cover (P -T). The larger this difference is, the greater are the opportunities for affecting favorable conditions of flow. The smaller this difference is, the more limited are the opportunities. We can therefore display watershed conditions as shown in Figure 5. The numerical ratings in Figure 5 give the treatment priority.An UNSATISFACTORY /FULL opportunity watershed would receive first priority for planning treatment measures. Note that all unsatisfactory watersheds regardless of opportunity class, should be planned for restoration before satisfactory watersheds.

WATERSHED CONDITION /OPPORTUNITY CLASS

1.7 Decreasing planning and treatment priority 100 POTENTIAL COVER

C 41 280

W 60 Ift°- 5AT O U 4 40 0 5 I TOLERANCE COVER Z wit-\MTE d UNSAT/FULL UNSATfMODERATE UNSAT/LIMITED I 2 1

3 I I

INCREASING ELEVATION AND PRECIPITATION

Figure 5. Watershed Conditions and Opportunity Classes.

This classification system is purposely made simple by relating watershed performance to a single index, ground cover. It serves to convey the concept of favorable watershed conditions to management. Our feeling is that the concepts for improving hydrologic response have often been needlessly complicated by technical specialists. Over the years, appreciation of our ability to affect runoff and erosion has been given to specialists, and therefore, a commitment to maintaining favorable conditions of flow has been partially lost by management. It is again time that all levels of managers, technical specialists, and decision makers realize their responsibilities for maintaining satisfactory watershed conditions. This can only come about by instilling an understanding of the processes of watershed management in a straightforward way.

Future Research, Equipment, and Data Needs

The process of developing the model and technique identified several needs as follows:

1. Rainfall and runoff data for short time increments (1 to 5 minutes) on small basins (5 to 10 sq. mi.).

34 2. Gaging equipment which is event oriented, solid state (to enhance accuracy), and directly computer -processible to facilitate data handling.

3. An assessment of current ground cover conditions for public lands.

4. Techniques for estimating potential ground cover for various ecosystems in the West.

5. Effects on baseflows and interflow resulting from alterations of infiltration and reduced overland flow.

6. Relationship between cover and activities on infiltration and runoff.

Conclusion

Good watershed conditions benefit everyone. Rain and snow are absorbed into the soil reservoir on slopes and released over a longer period of the year. This increases the effective regulating capacity of downstream reservoirs.Reservoirs and irrigation works require less maintenance because of reduced sediment. Productive soil is maintained on -site providing for sustained production of lumber, fuelwood, livestock, wildlife, recreation and water. There is ample evidence in the literature of the efficacy of improving watershed condition and the relationships between ground cover and runoff. There are examples of successful treatments in New Mexico in pinyon -juniper and juniper -grasslands (USDA, 1960; Columbus, 1980).

To achieve and maintain satisfactory watershed conditions, land uses must be designed to match the inherent capability of the land and be compatible with the needs of downstream communities. Floods, muddy water, and streams that dry up quickly are symptoms of unsatisfactory watershed conditions. Many of these consequences have been inherited by the present generation and are accepted as the natural situation. However, we can in many cases make positive changes in watershed condition and hydrologic responses.

References Cited

Colman, E.A. 1953. Vegetation and watershed management. The Ronald Press Co., New York. 412 p.

Columbus, J.T. 1980. Watershed abuse -the effect on a town. Rangeland 2(4):148 -150.

Crawford, N.H. and R.K. Linsley. 1962. The synthesis of continuous streamflow hydrographs on a digital computer. Dept. of Civil Engineering, Stanford Univ. Technial Report No. 12. 121 p.

Croft, A.R. and R.W. Bailey. 1964. Mountain water. USDA Forest Service, Intermountain Region, Ogden, Utah. 64 p.

Doty, R.D. 1971.Contour trenching effects on streamflow from a Utah watershed. USDA Forest Service Research Paper INT -95. 19 p.

Green, A.W. 1971. Sone economic considerations of watershed stabilization on National Forests. USDA Forest Service Research Paper INT -92. 10 p.

Hewlett, J.D. and W.L. Nutter. 1970. The varying source area of streamflow from upland basins. In: Proc. of the Symp. on Interdisciplinary Aspects of Watershed Management. p. 65 -83. Amer. Soc. of Civil Engr. New York.

Hibbert, A.R. 1975. Percolation and streamflow in range and forest lands. In: Proceedings of Fifth Workshop of the United States /Australia Rangelands Panel, Boise, Idaho. pp. 61 -72.

Hickok, R.B., R.V. Keppel and B.R. Rafferty. 1959. Hydrograph synthesis for small aridland watersheds. Jour. of Agric. Engineering. 40(10):608 -611, 615.

Holton, H.N. and N.C. Lopez. 1971. USDAHL -70 model of watershed hydrology. USDA Agricultural Research Service. Technical Bulletin No. 1435. 84p.

Horton, R.E. 1933. The role of infiltration in the hydrologic cycle. Amer. Geophy. Union. Trans. 14:446 -460.

Horton, R.E. 1945. Erosional development of streams and their drainage basins; hydrophysical approa( to quantitative morphology. Bull. of the Geological Soc. of Amer. 56:275 -370.

35 Knisel, W.G. (ed). 1980. CREAMS -a field scale model for chemical, runoff, and erosion from agricultural management systems. USDA Conservation Research Report No. 26. 643 p.

Meeuwig, R.O. 1960. Watersheds A and B -a study of surface runoff and erosion in the subalpine zone of central Utah. Jour. of Forestry. 58(7):556 -560.

Megahan, W.F. 1972. Subsurface flow interception by a logging road in mountains of central Idaho. In: Proc. Natl. Symp. of Watersheds in Transition. Fort Collins, Co. Am. Water Resour. Assoc. p. 350 -356.

Rifkin, J. and T. Howard. 1980. Entropy. The Viking Press, New York. 305 p.

Rosa, J.M. 1954. Guides for program development flood prevention on small watersheds of the Rocky Mountain area. USDA Forest Service, Intermountain Region, Ogden, Utah. 152 p.

Schmidt, L.J. 1978. The use of risk in specifying job quality. USDA Forest Service, Southwestern Region. Hydrology Note No. 8. 8 p.

Schumm, S.A. 1977. The fluvial system. John Wiley and Sons, New York. 338 p.

Shaw, D., R. Solomon, J. Maxwell, and L. Schmidt. 1981. A system for focusing the watershed management program in the Southwestern Region for the 1980's. USDA Forest Service, Southwestern Region. Hydrology Note No. 11. 17 p.

Solomon, R.M. 1982. SHOWER -storm hydrograph output for watershed evaluation and restoration. USDA Forest Service, Southwestern Region. 30 p.

Ward, A., T. Bridges and B. Wilson. 1981. A simple procedure for developing a design storm hydrograph. Water Res. Bull. 17(2):209 -214.

Woodward, L. 1943. Infiltration -capacity of some plant -soil complexes on Utah range watershed -lands. Trans. Am. Geophy. Union. 24:468:475.

USDA. 1960. Tour guide to Bernalillo watershed protection project. USDA Forest Service, Southwestern Region, Cibola National Forest. 7 p.

USDA. 1972. Hydrology. Section 4, National Engineering Handbook. USDA Soil Cons. Service, Washington, DC.

USDA. 1973. HYMO: problem- oriented computer language for hydrology modeling. USDA Agricultural Research Service. ARS -5 -9. 75 p.

36 SEASONAL CHANGE IN INFILTRATION AND EROSION FROM USLE PLOTS IN SOUTHEASTERN ARIZONA

J. R. Simanton and K. G. Renard

The authors are Hydrologist and Hydraulic Engineer, respectively, U.S. Department of Agriculture, Agri- cultural Research Service, Western Region, Southwest Rangeland Watershed Research Center, 442 East Seven- th Street, Tucson, Arizona85705.

ABSTRACT

A rotating boom rainfall simulator was used on 3 x 10.7 m plots to determine Universal SoilLoss Equation (USLE) parameter values. Simulator runs were made in the spring and fall of 1981 on two replications of four treatments on three soil types in southeastern Arizona. The treatments were:natural, vegetation removed, erosion pavement and vegetation removed,and tilled (moldboard plowed and disked). Runoff, infiltration, and soil loss varied significantlybetween treatments and, most interestingly, between the spring and fall runs. Plot surface characteristics of rock, gravel, soil, litter, and vegetation cover could not explain this seasonal variation in hydrologic response.

INTRODUCTION

Rangelandareas, like many other areas, exhibit extreme variability in the hydrologic processes affecting erosion and sediment yield. As part ofa nation -wide effort to improve application utility of the USLE (Wisch- meier and Smith 1978) to various regions of the United States, the Southwest Rangeland Watershed Research Cen- ter, USDA -Agricultural Research Service,in Tucson, Ari- ARIZONA zona has been using a rainfall simulator and runoff -erosion plots to determine values for the various USLE factors which might be applicable for rangelands. This work is being conducted on the 150 km2 Walnut Gulch experimen- talwatershed near Tombstone, in southeastern Arizona (Fig.1). This watershed is representative of millions of hectares of brush and grass rangeland found throughout the semiarid Southwest. Major vegetation of the watershed includes creosote bush (Larrea divaricata), white - thorn (Acacia constricta),tarbush (Flourensia cernua), black grama Bouteloua eriopoda), blue grama (Bouteloua gracillas) tobosagrass (Hilaria mutica),and bush muhly (Muhlenbergia Porteri). Soils are generally well drain- OCHI SE COUNTY ed, calcareous, gravelly loans with large percentages of WALMUT 6L1LCM ERPERIAIEMTAL WATEREMEM rock and gravel on the soil surface. Average annual pre- PERIPMERAL AREA cipitation on the watershed is about 300 mm, and is bimo- dally distributed, with approximately 70 percent occurring during the summer thunderstorm season of July to mid September.

WATERSHED AREA META.0 Differences between summer and winter precipitation can be tremendous on Walnut Gulch. Summer precipitation is dominated by convective thunderstorms which are limited in areal extent, and characteristically have maximum 5-min rainfall rates greater than 100 mm /hr (Osborn and Simanton 1981). Winter precipitation,though covering larger areas, has maximum 5-min rainfall rates which are Figure 1. Watershed location map. usually less than 10 mm /hr (Osborn etal. 1979). This ten -fold difference between summer and winter precipitation rates hasa significant effect on runoff and erosion. For example, over 99 percent of the annual Walnut Gulch runoff for the past 25 yr has occurred during the summer thunderstorm season.

Rainfall- runoffstudies onWalnutGulch indicate

37 that variables, such as vegetation and soil moisture, are not as significant as precipitation characteristics when used in rainfall- runoff regression models (Schreiber and Kincaid 1967; Osborn and Lane 1969). However, in other areas where precipitation characteristics are notas dominant, vegetation and soil moisture are important factors (SCS 1972).

Dixon (1975) has suggested that easily measured surface characteristics are indirectly related to infiltration and erosion, and has shown that the two surface parameters (microroughness and macroporosi- ty) that do directly influence infiltration are not easily measured.

The possible effects of soil surface compaction, such as would be expected from cattle grazing, have not been determined for soils on Walnut Gulch. However, studies elsewhere, that have related grazing to increased soil bulk density and decreased infiltration and increased erosion, seem mixed in their conclusions (Gifford and Hawkins 1976).

This paper describes and discusses one year's spring and fall variations in runoff and erosion from USLE study plots, and discusses these findings in relation to plot surface, soil, and vegetative characteristics.

EXPERIMENTAL BACKGROUND

Method

The research plan of this study on application of the USLE to rangelands followed procedures used in rainfall simulation studiesto develop values for the USLE soil erodibility factor (Wischmeier and Mannering 1969). This plan includes the use of relatively large plots (3 x 10.7 m long plots are known to show the effects from overland flow erosion), a standard surface treatment (continuous fallow produced by up- and down -slope plowing and disking), standard 9 percent slope, and standard sequences of rainfall inputs. These standard procedures were used so that our results could be compared with results from other USLE research. The study includes seasonal application of simulated rainfall on three treatments and a control that are replicated twice on three soil series, and is expected to continue for at least 3 years.

Soils

The three soil series selected were Bernardino (a thermic Ustollic Haplargid), Cave (thermic, shallow Typic Paleorthid), and Hathaway (thermic Aridic Calciustoll). These are all gravelly loams, and are USDA -Soil Conservation Service bench mark soils for Arizona. They comprise nearly 45 percent of the Wal- nut Gulch watershed area, and are described in detail by Gelderman (1970).

The Bernardino series is a deep, well- drained, fine -textured soil formed in old calcareous alluvium. This soil can have up to 50 percent, by volume, of gravel and cobbles in the surface 10 cm,and usually less than 35 percent gravel in the remainder of the profile.

The Cave series isa shallow, well- drained, medium textured soil with indurated lime hardpans that have developed at less than 45 cm in old, gravelly and cobbly calcareous alluvium. This soil can have up to 60 percent, by volume, of gravel and cobbles in the surface 10 cm,and usually less than 40 percent gravel in the remainder of the profile.

The Hathaway series is a deep, well- drained, gravelly medium and moderately coarse -textured soil over very gravelly, coarse -textured materials of moderate depths. This soil was formed from gravelly or very gravelly calcareous old alluvium, and can have up to 70 percent, by volume, of gravel and occasional cobbles in the surface 10 cm, and usually less than 50 percent in the remainder of the profile.

Plots

Criteria for plot selection were largely based on requirements set forth during the original development of the LISLE, and on constraints of the rainfall simulator. The criteria included:(1) plots had to be in pairs separated by no less than 3 m but no more than 4 m;(2) each pair had to be at least 7 m apart; (3) paired plots had to be parallel; (4) plot slope had to be near 9 percent;(5) plot slope had to be uniform; (6) rills or other obvious drainages must not be present, and (6) plot pairs on each soil series had to be relatively close to one another.

The 24 experimental plots were each 3.1 m x10.7 m in size, with the long axis parallel to the slope. Each plot was delineated on three sides by 15 cm metal borders that were installed so that 3 cm were inserted into the soiland 12 cm extended above the surface. The downslope end of the plot was delineated by a 20 -cm deep metal sheet with a sill plate on one edge. This sheet was inserted into the soil so that the sill plate was flush with the soil surface. The soil -metal interface was sealed with a silicone rubber -paint thinner solution which, upon drying, formed an impervious connecting joint. Runoff from the plots were collected in troughs that divert the water into a runoff -measuring flume equipped with a FW-1 water -level recorder that measures instantaneous discharge. After the plots on each soil

38 series were selected and plot borders installed, plot pairs were randomly assigned their particular treatment.

Treatments

The three treatments imposed included the standard moldboard plowedand disked up and down slope treatment (tilled), a vegetation removed treatment where the vegetation initiallywas clipped at the ground surface and then controlled with a systemic herbicide (clipped), and a vegetation and erosion pavement (rock and gravel > 2 mm) removed treatment where the vegetation was clipped at the ground surface, as previously detailed, and the erosion pavement was hand picked from the plot to minimize soil surface disturbance (fallow). Two natural cover plots were also selected for each of the soilseries. These were used as control plots. The tilled treatment is standard for determining values for the USLE soil erodibility factor (K). The other two treatments and the control were selected to show the effect of both vegetation cover and erosion pavement on soil loss. Prior to treatment, the plots were fenced to exclude cattle grazing.

After treatment, each plot pair was subjected to an initial 60-min rainfall simulation (dry), followed 24 hr later by a 30-min run (wet), which was then followed 30 min later by another 30 -min run (very wet). The simulated rainfall rate for each of these runs was about 60 mm /hr. By combining results from the runs, each series of runs provides runoff and erosion data for 30 mm and 60 mm continuous rains, a 90 mm rain with one 24 -hr interruption, and a 120 -mm rain within approximately a 24 -hr period. This simulation sequence was applied to the 24 plots, or 12 -plot pairs in the late spring (April -May) and early fall (Oct -Nov) within the same year. The test periods, which are periods oflow rainfallprobability, were just before and after the summer rainy season (July- Sept).

Rainfall Simulator

The rainfall simulator used was a trailer -mounted rotating -boom simulator capable of applying either 60 or 120 mm /hr rainfall rates (Fig. 2) (Swanson 1965). There are 10 arms radiating from a central stem. The arms support 30 nozzles that are positioned at radii of 1.5, 3.0, 4.6, 6.1,and 7.6 m, with 2,4, 6, 8, 10 nozzles on each respective radius. The two rainfall intensities (60 mm /hr or 120 mm /hr) are obtained by using either 15 or 30 nozzles. The nozzles spray downward from an average height of 2.4 m, apply about 15 liters /min, and produce drop -size distributions similar to those of natural rainfall (Swanson 1979). Preliminary results of a study of rainfall energies associated with the simulator indicate that the energies are about 80 percent of those of natural rainfall. Because of the simple design and portability of the simulator, many plots can be evaluated in a relatively short time (a complete series of runs can be made on 8 plots in one week).

Field Measurements

Plot surface characteristics and vegetation cover were measured before the initial treatment, and then before the 60-min simulation. Charac- teristics measured were bare soil (particles <2 mm in diameter), gravel (particles 2 to 20 mm in diameter), rock (particles > 20 mm in diameter), litter, vegetative basal cover, andvegetative crown cover. A 3.05 m long pin -point meter with holes spaced at 6 cm intervals was used. The meter is placed perpendicular to the plot slope, and rests on the metal plot border at10 positions along the plot. At each position, 49 pin- point surface and vegetationmeasurements are made by dropping a pin through eachpin hole. Thus, there isa total of 490 point measurements per plot to describe the surface characteristics.

Rainfall amountand application rate were measured witha modified recording raingage that was placed between each plot pair. The raingage was modified to increase its sensitivity to rain - Figure 2. Trailer mounted rotating -boom fall ratebydoublingthe rainfall -collecting rainfall simulator. area and enlarging the recorded time scale. Rainfall distribution over each plot was measured with four small plastic gages that recorded only rainfall amount. One gage was placed near each of the four corners of each plot.

On some early testing of thesimulator, threenarrow slottedtubeswere placedacross

39 each plot to check the appliclation amount. This scheme was abandoned when (1) the amounts were found to agree very closely with those obtained from the four smallplasticgages, and (2) a smallridge was observed to form on the plot surface beneath each slotted tube because of the interception of the simulated rainfall.

The flumes used to measure runoff have a capacity of about 4 liters /sec,and have sloped floors to minimize sediment deposition. The water -level recorders were sensitized so that small changes in runoff are noticeable on the runoff hydrograph.Figure 3 illustrates a complete plot pair field arrangement.

Sediment samples, manually collected in quart sample bottles, were taken at the flume exit periodi- cally during the runoff period. Sampling intervals were dependent on changes in the runoff rate, with frequent sampling when discharge was changing rapidly. The shortest sampling interval was 1 min, usually during the first 10 min of the runoff, and the longest was no more than 15 min, usually toward the end of the run, when runoff rate was relatively constant. The time when the sample was collected was recorded for later correlation to the runoff hydrograph and calculation of sediment discharge rate and amount.

Data Analysis

and surfacecharacteristics ROTATING BOOM RAINFALL SIMULATOR Plot vegetativecover were tabulated and converted to percent cover for subse- sm quent correlation to runoff, erosion, and treatment effect. Rainfall records were digitized and tabulated to give rainfall hyetographs for use in infiltration studies. / a: Analog records from the water -level recorder on each flume // \\\ were digitized, converted to discharge rates, and tabula - '':\ / i N\\\ ted to give a hydrograph and runoff volume. /// /ono* N \ Sediment samples were analyzed for total concentra- // \ tion and particle -size distribution was determined (Haver- \ / /-\ N\I land and Cooper 1981). Sediment discharge rates and total soil losswere calculated using sediment concentration

i values and the runoff hydrograph. Sediment trapped in the I j 1 flume was measured and distributed in proportion to the \ \ \ -- flow rate. \ \ I `, / / \ -/ / RESULTS AND DISCUSSION i i _ s'^roR«,ROINA / Average surface characteristics for the spring and D l/ fall rainfall simulations for the different soil- treatment `WAS.NA / ;;*IN combinations are shown in table 1. Vegetation data for `- the natural control plots are listed in table 2. Spring and fall runoff and sediment treatment averages for the ...WT. dry, wet, and very wet surface simulations are listed in Figure 3. Typical plot layout with schema- tables 3,4,and 5, respectively. Spring and fall hydro - tic of simulator nozzle path. graphs and sedigraphs for a natural plot on two soil types are shown in fig. 4 through 6. Only data from the natural plots are presented because they show only seasonal change and not changes caused by treatment. Since data from the plots on the Cave and Hathaway soils exhibited the same pronounced seasonal shift in runoff,infiltration, and soil loss, only the Hathaway plot data are shown.

Final infiltration rates in the spring decreased 20 percent between the dry and wet surface runs on the Hathaway soil, and a decrease of only 8 percent on the Bernardinosoil. In the fall,the final infiltration rate decreased approximately 20 percent between the dry and wet surface areas on the Hatha- way soil and approximately 15 percent on the Bernardino soil. Seasonal differences in final infiltration rates were measured only during the dry surface simulations. There was about a 60 percent decrease in final infiltration rates from spring to fall on the Hathaway soil, and an approximate increase of 60 percent in the spring to fall simulations on the Bernardino soil(fig. 4 - 6).

Soil loss from the Bernardino natural plots was significantly less in the fall, and there also was a significant decrease in runoff volume. The soilloss from the naturalplots on the other two soils increased, though not significantly, even though there was a significant increase in runoff (table 3).

One of the main objectives of this USLE plot study was to identify and quantify those soil surface characteristics that have a significant influence on runoff and erosion from rangelands. To eliminate treatment effects, only natural plots were used in a multiple linear regression analysis of the data collected. In effect, correlation coefficients were determined between plot runoff and soil loss and surface characteristics for the spring, fall, and combined season data.

Spring runoff volume was positively correlated with erosion pavement (combined percentage of rock

40 and gravel) (r = 0.90) and grass crown cover (r = 0.93), but negatively correlated with litter cover (r = -0.93) and shrub crown cover (r = -0.90). The correlation coefficient needs to be greater than 0.87 to be significant at the 5- percent level. Soil loss was positively correlated to erosion pavement (0.94).

Table 1. Plot surface characteristics( %)

Soil (< 2mm) Gravel 2mm -2cm Rock > 2cm Litter Rock and Gravel Treatment Spring Fall Spring Fall Spring Fall Spring Fall Spring Fall (Avg)

Bernardino Soil Natural 28.06 20.00 37.55 53.00 26.33 21.50 6.32 4.00 63.88 75.00 Tilled 62.45 66.50 21.84 8.70 9.2820.00 6.43 4.80 31.12 28.50 Clipped 32.65 25.00 28.37 43.00 18.68 23.00 20.30 9.00 43.05 66.00 Fallow 63.78 42.00 15.72 46.50 1.63 4.00 18.88 7.50 17.34 50.50 Hathaway Natural 19.80 24.50 33.06 33.00 15.00 24.00 30.00 16.00 48.06 57.00 Tilled 68.68 68.50 12.34 12.00 12.76 18.00 6.22 1.50 25.10 30.00 Clipped 17.34 37.50 36.22 30.00 13.26 26.00 32.86 6.50 49.49 56.00 Fallow 55.51 57.50 6.74 25.00 4.18 6.50 33.57 11.00 10.92 31.50 Cave Natural 29.18 21.50 21.12 39.50 11.22 17.00 35.30 21.00 32.3547.00 Tilled 52.34 31.50 10.52 12.50 28.47 50.50 8.68 5.50 38.98 63.00 Clipped 29.28 34.50 24.28 36.50 9.49 14.50 36.94 14.50 33.78 51.00 Fallow 63.78 59.00 7.24 29.50 3.68 6.50 24.90 6.00 10.92 33.50 All Soils Natural 25.70 22.00 30.60 41.80 17.50 20.80 23.90 13.70 48.10 59.70 Tilled 61.20 55.50 14.90 11.10 16.80 29.50 7.10 3.90 31.70 40.50 Clipped 26.40 32.30 29.60 36.50 13.80 21.20 30.00 10.00 43.40 57.70 Fallow 61.00 52.80 9.90 33.70 3.20 5.70 25.80 8.2 13.10 38.50

Table 2. Vegetative cover( %) natural plots

Grass Forb Shrub Total

Soil Spring Fall Spring Fall Spring Fall Spring Fall

Bernardino Crown 11.2 46.3 0.6 18.4 8.1 3.0 19.9 68.2

Base 1.6 1.0 ------0.1 - -- 1.7 1.0

Hathaway Crown 7.9 36.6 3.1 10.2 10.0 5.8 21.0 52.6

Base 1.68 2.2 0.2 - -- 0.3 - -- 2.1 2.2

Cave Crown 4.1 14.7 0.5 0.8 19.5 15.9 24.1 31.4 Base 1.7 .8 1.3 0.1 3.0 .9

Fall runoff volume and soil loss were poorly correlated to plot surface characteristic. The combined season correlation matrix also did not indicate sufficient correlation between runoff or soil loss and plot surface characteristics.

The erratic nature of the sedigraphs (fig. 4 -6) for all the spring runs may be the result of an observed buildup and breakdown of debris dams on the plots. Under constant rainfall rates, these dams become more a function of surface characteristics than would be expected with varying rainfall intensities of natural storms. The frequency of these debris dam formation and dissipation, and the magnitude of their effects may bedependent onthe plotslope, roughness, and amountof litter.

41 Table 3. Dry surface runoff and soil loss

Runoff Sediment Soil Loss Runoff Soil loss coefficient concentration

Treatment (mm) (Q/p) (gms) (gms /liter) (T /ha) Spring Fall Spring Fall Spring Fall Spring Fall Spring Fall (Avg)

Bernardino Soil Natural 34.1 16.1 0.59 0.28 1483 356 1.4 0.7 0.46 0.11 Tilled 0.2 0.3 0.004 0.005 84 139 6.1 5.5 0.03 0.04 Clipped 27.9 32.1 0.49 0.56 1229 2522 1.4 2.5 0.38 0.78 Fallow* 37.8 34.4 0.67 0.60 6946 8005 5.6 7.1 2.14 2.46 Hathaway Natural 18.0 32.4 0.33 0.56 992 1134 1.7 1.1 0.30 0.35 Tilled 5.5 11.2 0.10 0.20 614 1878 3.4 5.4 0.19 0.58 Clipped 20.4 34.3 0.37 0.60 1310 3524 2.0 2.5 0.40 1.08 Fallow 28.2 39.6 0.52 0.69 6486 13076 7.3 10.1 1.99 4.02 Cave Natural 6.6 19.7 0.12 0.36 462 810 2.6 1.3 0.14 0.25 Tilled 0.8 1.7 0.01 0.03 172 186 3.4 4.8 0.05 0.06 Clipped 11.7 38.0 0.20 0.68 808 3549 1.7 2.8 0.25 1.09 Fallow 17.8 34.0 0.30 0.59 1309 13877 4.8 12.6 0.40 4.27 All Soils Natural 19.6 22.7 0.35 0.40 979 767 1.9 1.0 0.30 0.24 Tilled 2.2 4.4 0.04 0.08 290 734 4.3 5.2 0.09 0.23 Clipped 20.0 34.8 0.35 0.61 1116 3198 1.7 2.6 0.34 0.98 Fallow 27.9 36.0 0.50 0.63 4914 11653 5.9 9.9 1.51 3.58

Table 4. Wet surface runoff and soil loss

Runoff Runoff Sediment Soil loss Soil loss coefficient concentration

Treatment (mm) (Q /p)/ (gms) (gms /liter) (T /ha) Spring Fall Spring Fall Spring Fall Spring Fall Spring Fall (Avg)

Bernardino Soil Natural 15.6 12.8 0.54 0.44 650 302 1.3 0.8 0.20 0.09 Tilled 2.2 10.4 0.09 0.36 194 1526 3.2 4.0 0.06 0.47 Clipped 15.8 17.1 0.59 0.58 534 1192 1.1 3.1 0.16 0.37 Fallow* 15.6 17.6 0.56 0.61 2970 5842 5.9 10.1 0.91 1.80 Hathaway Natural 8.2 13.3 0.29 0.46 430 521 1.6 1.2 0.13 0.16 Tilled 12.8 17.9 0.48 0.61 1882 3153 4.8 5.7 0.58 0.97 Clipped 10.1 19.8 0.37 0.67 452 1750 1.4 2.7 0.14 0.54 Fallow 12.4 18.3 0.43 0.62 3323 6678 8.2 11.4 1.02 2.05 Cave Natural 4.8 13.1 0.16 0.45 314 462 2.2 1.1 0.10 0.14 Tilled 3.6 4.4 0.12 0.16 436 383 3.9 2.9 0.13 0.12 Clipped 9.3 18.5 0.30 0.64 530 1220 1.6 2.0 0.16 0.38 Fallow 14.3 18.2 0.46 0.64 2788 5436 5.9 9.2 0.86 1.67 All Soils Natural 9.5 13.1 0.33 0.45 465 428 1.7 1.0 0.14 0.13 Tilled 6.2 10.8 0.23 0.38 837 1687 4.0 4.2 0.26 0.52 Clipped 11.7 18.5 0.42 0.63 505 1387 1.4 2.6 0.15 0.43 Fallow 14.1 18.0 0.48 0.62 3027 5985 6.7 10.2 0.93 1.84

*Fallow plots had the vegetation clipped,and the erosion pavement was removed from the plot surface. TQ is the surface runoff in mm; P is the applied rainfall in mm.

42 Table 5. Very wet runoff and soil loss

Runoff Sediment Runoff Soil loss Soil loss coefficient concentration

Treatment (mm) (Q /p)/ (gms) (gms /liter) (T /ha) Spring Fall (Avg) Spring Fall Spring Fall Spring Fall Spring Fall

Bernardino Soil Natural 18.4 13.2 0.64 0.46 746 342 1.3 0.8 0.23 0.11 Tilled 5.8 13.4 0.20 0.46 435 1488 2.8 3.2 0.13 0.46 Clipped 17.6 18.9 0.62 0.64 758 1654 1.3 3.3 0.23 0.51 Fallow* 17.6 19.3 0.62 0.67 3871 6328 6.7 10.0 1.19 1.95 Hathaway Natural 10.7 15.8 0.40 0.54 514 507 1.5 1.0 0.16 0.16 Tilled 15.7 18.2 0.57 0.62 2616 3508 5.6 6.0 0.80 1.08 Clipped 15.4 20.0 0.56 0.68 648 1697 1.3 2.7 0.20 0.53 Fallow 15.9 23.9 0.58 0.72 4316 6340 8.2 9.4 1.33 1.95 Cave Natural 7.5 14.3 0.25 0.50 363 442 1.6 1.0 0.11 0.14 Tilled 10.8 9.5 0.36 0.32 952 596 2.5 1.9 0.29 0.18 No veg 13.1 17.3 0.44 0.62 598 1290 1.3 2.2 0.18 0.40 17.5 19.4 0.58 0.66 3384 6146 5.8 9.7 1.04 1.89 All Soils Natural 12.2 14.4 0.43 0.50 541 430 1.5 0.9 0.17 0.13 Tilled 10.8 13.7 0.38 0.47 1334 1864 3.6 3.7 0.41 0.57 Clipped 15.4 18.7 0.54 0.65 668 1547 1.3 2.7 0.21 0.48 Fallow 13.0 20.9 0.59 0.68 3857 6271 6.9 19.7 1.19 1.93 *Fallow plots had the vegetation clipped,and the erosion pavement was removed from the plot surface. TQ is the surface runoff in mm; P is the applied rainfall in mm.

Analysis of variance (ANOVA) (Ouncan's multiple range) among treatments and seasons showed thatrun- off volumes of the dry surface runs were significantly less (1 percent level) from the tilled plots than any other treatment in both seasons. Other significant differences in dry surface runoff volumes were: (1) fall natural treatments < fall clipped and fallow, (2) average spring runoff (all treatments)< average fall runoff (all treatments). Soil loss from the dry surface runs were: (1) significantly greater (1 percent level) from the fallow treatment than any other treatment in both the spring and fall, and (2) significantly less (1 percent level) from the spring fallow treatment than from the fall fallow treatment.

Although the statistical difference between the runoff volume for the wet orvery wet replications was not significant between treatments within a season or between seasons, the SCS Curve Number concept for estimating runoff did indicate a consistent pattern of increasing curve numbers with increasing antecedent moisture (SCS 1972).

The runoff and soil loss differences measured between the spring and fall simulationson the natural plots could not be statistically attributed to any measured surface or soil parameter. However, these seasonal differences might possibly be explained by changes in soil surface compaction (Schumm andLusby 1963; Rauzi and Smith 1973). The spring runs were made on a soil surface loosened throughout the fall and winter by the combined effects of diurnal soil temperature fluctuations and the wetting and dryingof the upper soil layer. This loose soil surface in the spring affected the Hathaway and Cave sites more because of the lower amounts of erosion pavement. This pavement would act as an insulating cover on the Bernardino soil, tend to moderate soil surface temperatures, and reduce soil water evaporation. In the spring, the soils would tend to have a greater initial infiltration rate than the more compacted and crusted soil surface in the fall produced by the high energy associated with the previous summer's thunderstorm rainfall.

To exemplify this, analysis of the natural plot's runoff hydrographs indicates that the time between start of simulated rainfall and runoff is significantly less in the fall than in the spring for the Hath- away and Cave soils, but is not much different for the Bernardino soil (table 6). The loose soil surface also hada dramatic effect on soilloss. Sediment concentration in the runoff in the spring for the natural plots was twice that in the fall (table 3).

Another possibility for these seasonal runoff and soil loss changes could be associated witha recovery from grazing exclusion. The three soil sites chosen had been grazed until the sites were fenced just prior to plot treatments. The recovery from grazing is a combined effect of changes in soil compaction,

43 vegetation density, and litter accumulation. This response may be an important time -related parameter, especially on plot -size areas.

Table 6. Time to beginning of runoff after start of simulated rainfall (min)

Natural Bernardino Hathaway Cave Conditions Spring Fall Spring Fall Spring Fall

Dry 3.50 4.00 5.00 3.75 11.50 4.65 Wet 3.00 2.75 5.25 3.75 5.00 3.00 Very wet 2.25 2.00 2.50 1.90 3.60 2.00

APNICATIQN NATURAL PLOT IS NATURAL PLOT 7 HATHAWAY SOIL BERNARDINO SOIL DRY SURFACE DRY SURFACE 40- RING FA LI_ FALL SPRING 30- SPRING \ r F FALL Q FALL C SPRING 2 2

RUNOFF RUNOFF 10 - - INFILTRATION INFILTRATION

10 20 30 40 50 60 70 0 20 30 40 60 70 TIME (MINUTES) TIME (MINUTES)

4000 NATURAL PLOT IS NATURAL PLOT 7 HATHAWAY SOIL SOIL LOSS BERNARDINO SOIL SOIL LOSS DRY SURFACE SPRING6B99ms DRY SURFACE SPRING 16910ms FALL 7794m, 3000 FALL4014464 1: 3000-

É of

Q , g moo- . Z W -

Ó t0 . 100o- 1000 SPRING FALL

20 30 40 50 60 70 10 20 30 40 50 60 70 TIME (MINUTES) TIME (MINUTES)

Figure 4. Dry surface infiltration, runoff, and sedigraph.

CONCLUDING COMMENTS

A significant feature of this initial interpretation of the runoff, erosion, and surface character - ics relationships is that one season's or year's data are, even under carefully controlled conditions, y difficult to define in terms of easily measured physical parameters when the data are analyzed using mole ANOVA procedures such as were used here.

At this point in the experimental work,it appears that a great deal of additional work will be needed to understand the cause- effect relationships on these plots. Certainly, the differences in runoff and soil loss between spring and fall for the different soils considered would indicate that either the freeze -thaw mechanism or the exclusion of grazing may be changing the bulk density which, in turn, may markedly change the infiltration and erodibility of these rangeland soils.

44 60 APPLICATION IP}ICATION "\ I. \\ NATURAL PLOT T \\ NATURAL PLOT IS 50- \\ BERNARDINO SOIL \ \\ HATHAWAY SOIL \\ WET SURFACE \\ \ \ WET SURFACE 4 \

\ \ \ SPRINS SPRING FALL SPRING 30- ALL FALL SPRING O 20- - RUNOFF -- INFILTRATION - RUNOFF -- INFILTRATION

0 20 30 40 TIME IMINUTES) 10 20 30 40 TIME (MINUTES)

4000 4000 NATURAL PLOT 7 NATURAL PLOT IS BERNARDINO SOIL SOIL LOSS HATHAWAY SOIL SOIL LOSS WET SURFACE SPRING7004ms WET SURFACE SPRING428am4 I3000- FALL3344ms FALL 3199ms 3000

É oÉ W á2000- 2000 z zP 2 iW C Ó W W 1000 - W 1000

00 20 30 40 10 20 30 40 TIME (MINUTES) TIME IMINUTES)

Figure 5. Wet surface infiltration, runoff, and sedigraph.

ACKNOWLEDGMENTS

The authors would like to thank Howard Larsen and Jim Smith for their assistance, dedication, and enthusiasm in the collection of the field data. Also, the efforts of Loel Cooper in the lab analysis of the sediment data and the assistance of Dr. E. D. Shirley in the computer analysis are greatly apprecia- ted.

REFERENCES CITED

Dixon, R. M. 1975. Infiltration control through soil surface management. Proc. Symposium on Watershed Management, ASCE, Logan, Utah. p. 543 -567.

Geldérman, F. W. 1970. Soil Survey, Walnut Gulch Experimental Watershed, Arizona. Special Report USDA - SCS, 55 p.

Gifford, G. F.,and R.H. Hawkins. 1976. Grazing systems and watershed management: A look at the record. J.Soil Water Cons. 31(6):281 -283.

Haverland, R. L., and L. R. Cooper. 1981. Microtrac: A rapid particle -size analyzer of sediments and soils. Proc. Ariz. Sec., AWRA -Hydrology Sec., Ariz. Acad. Sci. 11:207 -211.

Osborn, H. B., and L.J. Lane. 1969. Precipitation- runoff relationships for very small semiarid rangeland watersheds. Water Resources Research 5(2):419 -425.

45 APPLICATION If APPLICATION 50- II NATURAL PLOT 7 t' 50- \) NATURAL PLOTIS II BERNARDINO SOIL \\ HATHAWAY SOIL II VERY WET SURFACE 1\ VERY WET SURFACE 40 - 1 \ ao- \\

\\ FALL \ SPRING O FAIL 30- FALL

SPRING 20 20 -

RUNOFF - - INFILTRATION - RUNOFF - INFILTRATION

O 10 20 30 40 20 30 TIME (MINUTES) TIME (MINUTES)

4000 4000 NATURAL PLOT 7 NATURAL PLOT IS HATHAWAY SOIL BERNARDINO SOIL SOIL LOSS SOIL LOSS VERY WET SURFACE VERY WET SURFACE SPRING 6784ms SPRING 542Im, FALL 3S6Gz. 23000 - FALL 33OOWR 3000-

F É

H H 2000 2000- V F 2 2 W 7 a O w w m m 1000- .00.

10 20 30 40 20 TIME (MINUTES) TIME (MINUTES)

Figure 6. Very wet surface infiltration, runoff, and sedigraph.

Osborn, H. B., and J. R.Simanton. 1981. Maximum rainfall intensity of southwestern thunderstorms. Fourth Conference on Hydrometeorology, Reno, Nevada American Meteorological Society.

Osborn, H. B., R. B. Koehler, and J.R. Simanton. 1979. Winter precipitation on a southeastern Arizona rangeland watershed. Proc. Ariz. Sec., AWRS -Hydrology Sec., Ariz. Acad. Sci. 9:15 -20.

Rauzi, F., and F. M. Smith. 1973. Infiltration rates: Three soils with three grazing levels in northeastern Colorado. J. of Range Management 26(2):126 -129.

Schreiber, H. A., and D.R. Kincaid. 1967. Regression models for predicting on -site runoff from short - duration convective storms. Water Resources Research 3(2):389 -395.

Schumm,S. A., and G.C. Lusby. 1963. Seasonal variation of infiltration capacity and runoff on hill - slopes in western Colorado. J. Geophys. Res. 68(12):3655 -3666.

Soil Conservation Service. 1972. National Engineering Handbook, Sec. 4, Hydrology.

Swanson, H.P. 1965. Rotating -boom rainfall simulator. Trans. Amer. Soc. Agr. Engr. 8:71 -72.

Swanson, N. P. 1979. Field plot rainfall simulation (rotating -boom rainfall simulator), Lincoln, Nebra- ska. Proc. Rainfall Simulator Workshop, Tucson, Arizona March 7 -9, p. 166 -169.

Wischmeier, W.H.,and J.V. Mannering. 1969. Relation of soil properties to its erodibility. Soil Sci. Soc. Amer., Proc. 33(1):131 -137.

Wischmeier, W. H., and D. D.Smith. 1978. Predicting rainfall erosion losses - -A guide to conservation planning. USDA, Agriculture Handbook No. 537.

46 Distribution of Loss Rates Implicit in the SCS Runoff Equation

Richard H. Hawkins

Utah State University Logan, Utah 84322

Introduction

General: It is a practice in applied hydrology, approaching a hallowed tradi- tion, to assume that runoff processes on small watersheds are spatially uniform. The modeling fraternity calls this "lumping ", and it is associated with small watersheds almost by definition. Both practitioners and researchers make the uniformity assumption for reasons of ignorance and simplicity. Not much is known of spatial variability of soil, vegetative, and hydrologic properties of landscapes, and that which is known promises confusing complexity if and when realistically applied. Nonethe- less, it is also acknowledged that even small watersheds are indeed not uniform, though the consequences of the uniformity assumption are not clearly understood. As a simple example, consider a small 100 acre watershed producing 1.00 inch of runoff (or 100 acre -inches) from a rainstorm of 2.00 inches.This rainfall excess could have arisen from a variety of conditions: 1) Half the watershed (50 acres) producing 2 inches of excess; 2) The entire watershed yielding 1.00 inch of excess; or 3) Any number of intermediate combinations, such as 80 acres at 1.25 inches, or mixtures such as 20 acres at 2 inches plus 60 acres at 1 inch. Mass accounting requires that the sum of the area -excess products be equal to the basin runoff, or:

Q =E Qiai/A (Qi S P) [1] Where Q represents the rainfall excess depth from contributing area ai, A is the total watershed area, and Q is the net watershed runoff depth.

This paper will offer an organized though untested approach to dealing with this phenomenon in analysis and synthesis of runoff, with accent on its interpretation in the case of the SCS runoff equation. Background: In a previous paper (3), the writer has developed the notion of distributed effective loss rates, f, on small watersheds, and established the following equivalences. Given that effective loss rates are distributed as g(f), then the rainfall excess rate q for a period of intensity iis shown to be:

i Jr(i-f)g(f)df [2] q = 0

The foundation for this is illustrated in Figure 1. By expanding [2], and making distribution and conservation -of -mass interpretation, it can also be shown that

G(i)r dq /di [31 Where G(i) is the value of the cumulative distribution of g(f) at i =f, or (as a net watershed process), the fraction of the watershed yielding rainfall excess at intensity i. The derivative expression in [3] indicates that this is the slope of the rainfall- runoff rate curve at intensity i. Thus, the slope of the rainfall

47 >- 1-

This portion has i >f . O Generates rainfall excess. >- I- g (f) era Thisportion has i

CCa fn

o 0 f LOSS RATE ( f) AND RAINFALL INTENSITY( i )- 7T

Figure 1. Diagrammatic illustration of probability density for distribution of effective loss rates on a model watershed. While the minimum loss rate is known here as zero, situations in which Emin >0 are also possible, as are polymodal distributions, mixtures, and upper un- bounded distributions.

function is the fractional contributing area. Equation [3] is differentiated to produce the density function g(),so that:

g(i)= d2q /di2 [4] This is the underlying kernel distribution of loss rates.Note that it can be determined through differentiation of the rainfall excess function. This will be explored here for the case of the SCS runoff equation.

When the above reasoning is applied to situations of variable rainfall intensities, net watershed loss rates appear to vary positively with rainfall intensity up to the limit of i =fm Examples of such performances are surprisingly prominent in available data sets from plots and watersheds, thus offering at least indirect support for the general contention of spatially variable loss rates.

SCS Runoff Eauation Background: The SCS equation is the heart of the widely used Curve Number method. The runoff depth function is:

Q = (P- .2S)2 /(P +.8S) P > 0.2S [5]

Q = 0 P < 0.2S Where Q is the direct storm runoff depth (inches), P is the storm rainfall depth (inches), and S is an index of basin retention, equal to the maximum possible difference between effective rainfall (P -0.28) and runoff Q. The land condition index, curve number (CN) is a transformation of S,or:

CN =1000/(10+S) [6]

48 Equation [5] can be standardized on the parameter S, leading to:

QIS = (PIS- .2)21(P /S +.8) [7] which is a much more convenient form and useful for what is to follow. It is shown in Figure 2.

Documentation of the method's development is given in its foundation publication (6) and general subsequent papers (4,7). A critical appraisal of it is given by Chen (1). Despite visible shortcomings, it is the most popular technique of its type in use today, and is applied to a wide variety of situations on an international basis. Its ability to incorporate land use, condition, and site moisture into runoff calculations, its documentation and agency origin, and the lack of suitable alternate techniques have promoted its use.

Assumptions: The distributed loss rate theory requires rainfall and excess rates, not depths (volumes). However,in its original and most robust application, the SCS runoff equation was applied only to daily data, or a time interval ofone day. It was used to transform daily rainfall frequency descriptions to a parallel description of daily runoffs. Thus, it is consistent to assume rainfalls and runoffs as one day intensities, or, to be more general, as average intensities for storm durations used. In what is to follow it is necessary to stretch rates to cover assumed or understood storm durations T, such as 24 hours, so that:

i = P/T [8] and q= Q/T [9] f = i-q [10] As odious as this compromise seems at first glance, it is at least partially justifiable. Equations 8 -10 inflict the presumption of a uniform intensity storm, for which there is some literature support, to achieve the SCS runoff equation (5). Additionally, common use of the SCS runoff equation as an infiltration method achieves the same effect by applying it to individual intervals throughout thepro- gress of design storms.

In addition to the above, all the assumptions required of the foundation equations [2] through [4] are necessary. A major item is that the watershed is composed of an undefined large number of independent loss rate cells or contributing smaller runoff units, each operating with a characteristic time constant loss rate, distributed as g(f). The "runoff -runon" process is ignored within these cells.

Curve Number Loss Rate Distribution

Development: The intellectual strategy in linking the SCS equation and the distributed loss rate concept assumes that the former is empirically correct but that the runoff occurs via the distributed processes imagined in the latter, and notas a lumped uniform phenomenon as historically assumed. The equivalence is drawn by simply equating equations [2] and [5]. Therefore, being careful with limits,

r(P-f)g(f)df=(P-.2S)2/(P+.8S)P Q= J2S

The task is now to solve for g( ). Rather than deal with equation [11] directly, the principles previously discussed and given as equations [2] and [3]may be used. That is, differentiating the right hand of equation [11] (which is equation [5]) with respect to P and simplifying yields:

G(P)= dQ /dP = 1- (P /S +.8) -2 P > 0.2S [12] Of concern is the situation when P = f in the above. Thus substituting f for P gives:

G(f) = 1-(f/S+.8)-2 f > 0.2S [13]

49 1.0

.6 G(f) :1 -'f/5+0.8)2 4

0 f S 0 0.2 1 2 3 4 5

2/S

3 g )= S f/S 4' 0.8

0 >f/S 0 0.2 4 5

Figure 2. Above: Dimensionless expression of the SCS rainfall- runoff equati standardized on the storage index S. Center: Cumulative distribut of effective loss rates implied by the SCS equation. Bottom: Dens function of loss rates. Note the strong positive skew, and the lo limit of 0.2S. The mean is 1.2S, and the median is 0.61425.

50 The density function is obtained by differentiating the above with respect to f, and simplifying, giving:

g(f) _ (2 /S)(f /S +.8) -3 f 2 0.2S [14] Equation [14] above is the distribution of storm loss rates f, which when inserted into equation [2] will give equation [5]. It reconciles the two ideas. Characteristics: Figure 2 shows the original SCS equation, standardized on S, and the simultaneous display of G(f) and g(f), also standardized on the watershed index S. The expected value or mean of the distribution is found by the method of moments as:

E(f) = f* f =1.2S [15] and the median or f50 is found to be:

f50 = (171.8)S = 0.6142S [16] The higher moments, i.e. the variance, skewness, kurtosis, etc. are undefined (they approach infinity). Nonetheless, it exhibits a distinct positive skew. It should be noted that the distribution is valid only for f 2. 0.2S, which asserts that the watershed has no loss rates less than 0.2S.This uncomfortable item springs from the original SCS equation's demand for an initial abstraction of 0.2S. Undoubtedly, a more flexible expression relying on a general value of initial abstraction could also be derived.

Application

Contributing Fraction: Given a storm depth and curve number, the inferred contributing fraction or partial area can be calculated. For example a storm of P = 2.00 inches falling on a watershed of CN =80, yields a calculated runoff of 0.56+ inches. Applying equation [13] for f r 2.00 inches calculates G(f) =1- (1.6)- _ 0.61. This may be interpreted as 61% contributing area, or 39% of the watershed with f > 2 inches, and thus non -contributing. This information would give guidance towards possible land treatment strategies to reduce runoff,erosion, or pollutant pickup.

Loss Rate Equivalences: The relationship in equation [15] can be exploited to provide a translation between mean watershed loss rate and the widely used and handbook -backed Curve Number. Substituting S from equation [6] into equation [15] yields:

AL f' 1.2S r 1.2(1000/CN-10) [17] or CN = 1200/(12+ AO [18] It should be kept in mind that / f is the mean watershed loss rate when distributed as given in equation [13]. An identical result using a 0 -index has been previously demonstrated (2).

Summary

The SCS rainfall -runoff equation may be united with a spatially varied loss rate viathe distribution g(f) _ (2 /S)(f /S + 0.8) -3, which has a mean of 1.2S, a median of 0.6142 S, and undefined higher moments. Use of this notion reconciles the popular concept of partial area contribution with the curve number equation.Interpretations of fractional contributing area may be made for specific conditions.

51 References

1. Chen, C.L. 1981. Assessment of Soil Conservation Service Model with Runoff Curve Numbers.Presented at International Symposium on Rainfall- Runoff Modeling, Mississippi State University, May 18 -21, 1981. (Proceedings in Press)

2. Hawkins,R.H. 1978. Effects of Rainfall Intensity on Runoff Curve Number. In Hydrology and Water Resources of Arizona and the Southwest, Vol. 8, pp. 53 -64. (AWRA and AAAS Arizona Meeting, Flagstaff, April 14 -15, 1978).

3. Hawkins, R. H. 1981. Interpretation of Source Area Variability in Rainfall - Runoff Relationships. Presented at International Symposium on Rainfall Runoff Modeling, Mississippi State University, May 18 -21,1981. (Proceedings in press).

4. Kent, K. M. 1968. A Method for Estimating Volume and Rate of Runoff in Small Watersheds. U.S.D.A., Soil Conservation Service, CSC -TP -149, 58 pp.

5. Morel -Seytoux, H. J. 1981. Extension of the Soil Conservation Service Rainfall- Runoff Methodology for Ungaged Watersheds Report FHWA /RD- 81/060, Federal Highway Administration. 79 pp. (Available from NTIS).

6. U.S. Department of Agriculture, Soil Conservation Service. 1956 (et seq.). National Engineering Handbook, Section 4, Hydrology. USGPO, Washington, D.C.

7. U.S. Department of Agriculture. 1975. Urban Hydrology for Small Watersheds. Tech. Release No. 55.

52 QUASI THREE -DIMENSIONAL FINITE ELEMENT MODEL OF THE MADRID BASIN IN SPAIN

Jesus Carrera and Shlomo P. Neuman Department of Hydrology and Water Resources University of Arizona, Tucson, AZ 85721

Abstract

Groundwater flow in the Tajo River basin surrounding the city of Madrid is studied with theaid of a quasi three -dimensional model. The modelis based on an efficient adaptive explicit -implicit finite element method recently described by Neuman et al. (1982). The top layer is unconfined and interacts with the Tajo River and its tributaries. Reproduction of the existing conditions in the aquifer demonstrates the existence of local and intermediate flow patterns which are superimposed on the regional flow pattern. Such flow patterns could not be identified with a conventional two -dimensional model. The manner in which these patterns are affected by topography and stream configuration is clearly illustrated with the aid of three -dimensional plots constructed from a certain viewing angle. Similar plots are used to illustrate the evolution of drawdown zones due to pumpage at predetermined locationsin the aquifer.

Introduction

The Madrid aquifer is a thick, detritic unit extending over the central part of Spain (Fig. 1). It exhibits a marked vertical anisotropy due to layering. As a result of structural and topographic control, the flow pattern in the aquifer forms a complex three -dimensional picture.Attempts to model various aspects of this flowpattern have been reported byLópez Garcfa (1979) and Sahuquillo et al. (1975). A full -scale three -dimensional model is currently being developed by the Geological Service of Spain, based on a modified version of the Prickett and Lonquist (1971) finite difference computer program.

The purpose of this paper is to investigate the feasibility of modeling three -dimensional flow in the Madrid aquifer with the adaptive explicit -implicit quasi three -dimensional finite element model, FLUMPS, recently developed by Neuman et al. (1982). The results of this model compared favorably with those obtained from the Prickett- Lonquist model by the Spanish Geological Service.

Hydrogeology

The detritus of the Madrid basin has been formed from alluvial fans (López Vera, 1979). It consists of lenses rich in sand whichappear to be randomly distributed within a clay -rich matrix. The granulometric properties of the aquifer material vary according to their source: The three main facies, Madrid, Toledo, and Guadalajara, derive from the west- central range, east -central range, and Toledo Mountains, respectively. As a result of the large spatial variability of the granulometric properties, the hydraulic parameters of the aquifer are also highly variable.

The largest transmissivities (30 -100 m2 /day) are encountered in the Madrid unit of the Madrid facies which extends over the central part of the aquifer. The Tosco unit of th Madrid and Toledo facies, in the western part of the aquifer, has transmissivities between 20 and 50 m4/day. In the east, transmissivities of 10 -30 m /day characterize the Alcala and Guadalajara units of the Guadalajara facies. Least permeable is the evaporite unit extending over the southern and south -eastern part of the aquifer (Fig. 1), with transmissivities around 5 -10 m2 /day.

On the basis of two -dimensional computer modeling studies in the vertical plane conducted by Llamas (1976) and others, vertical hydraulic conductivities appear to be smaller than horizontal ones by a factor of 100. Recharge appears to be around 200 m3 /km2 /day, or 73 mm /year.

The surface drainage system consists of the Tajo river in the south, and its northern tributaries which are perennial except in very dry years.

53 N.

Detrltlefacies

L-= Transition facies Rei Chemical facies

MOCI Bedrock 0.44.. Inverse fault Watershed limit

Figure 1. Location and geology of Madrid basin.

Modeling Approach

Flow in the aquifer is modeled with the adaptive explicit -implicit quasi three -dimensional finite element program, FLUMPS, of Neuman et al. (1982). The quasi three -dimensional aspect of the model requires that the aquifer be subdivided into horizontal layers in which the flow is strictly horizontal. The Madrid aquifer is subdivided into three such layers, as illustrated by the horizontal finite element grids in Fig. 2. Each square in these grids has an area of 25 km2, and is automatically subdivided by the computer program into two triangles. Flow between the layers is strictly vertical. It is handled by vertical strings oflinear finite elements connecting the nodes of the horizontal grids. Ten such strings, each consisting in our case of a single linear element, are shown in Fig. 2.

The entire finite element grid for the Madrid aquifer includes 621 nodes, 285 in layer 1 at the top, and 168 in layers 2 and 3 in the middle and at the bottom, respectively. The grid consists of 981 triangular elements, of which 483 lie in layer 1, and 251in each of layers 2 and 3. In addition, there are 336 line elements connecting the horizontal grids in the vertical direction.

Layers 2 and 3 have impermeable boundaries on all sides. Layer 1 at the top has an impervious boundary on the north side; the remainder of its boundary follows the outline of streams. Streams also run within the interior of layer 1. All grid points along streams are assigned a mixed (Fourier) boundary condition of the following type: Constant rate of aquifer recharge when the water table lies below the stream bottom, and an aquifer discharge rate proportional to the difference between the computed water table and the water levelin the stream when the former lies above the stream bottom. A similar type boundary condition is assigned along ephemeral streams, at spring locations, and over swamp zones or other areas where discharge by evapotranspiration can take place.

The adaptive explicit -implicit aspect of the model consists of solving the finite element equations explicitly at nodes having stability limits in excess of the time step, At, and implicitly by LU- decomposition or iteratively at all other nodes (in this work iterations were used). For details

54 Figure 2. Horizontal finite element grids for layers 1, 2, 3 with a few vertical line elements.

concerning this method of solution, the reader is referred to Neuman et al. (1982). Since explicit solution is much faster than implicit solution, the adaptive scheme leads to considerable savings of computer time. It also leads to improved accuracy because one can use small time steps when hydraulic heads vary most rapidly at relatively little extra computer cost.

In FLUMPS, the size of At is varied automatically by the program. At the beginning when derivatives of heads with respect to time are largest, At is very small, and most nodes are automatically treated by the program as explicit. As time progresses, the large time derivatives imposed by pumpage dissipate, and At increases until at some nodes, At may come very close to (or exceed) the stability limits of these nodes. Such nodes are then automatically reclassified as implicit. The stability limit of each node is a function of the mesh geometry and the hydraulic parameters in the immediate vicinity of the node. Since in our case the mesh is uniform but the parameters vary in space, the stability limit varies from node to node. Fig. 3 is a typical graph showing how the number of implicit nodes in our model varies with the number oftime steps when At is allowed to increase without limit. An un- limited increase in At is allowed in all the steady state simulations reported below. In all transient simulations, the maximum At is restricted to one year, which is less than the smallest stability limit of any node in the grid. Thus, all the nodes during a transient simulation run remain explicit at all times.

Fig. 4 shows the amount of CPU time on the Cyber 175 computer of the University of Arizona versus the total number of time steps required for each of the 9steady state and transient runs performed in our study. Despite the respectable size of our quasi three -dimensional grid, the CPU time is small, being less than one minute in 8 out of the 9 runs. When all nodes remain explicit (steady state run 1, transient runs 6 -9) CPU time increases linearly with the number of time steps. Points departing from this straight line represent steady stateruns in which some, or all of the nodes, become implicit at various stages of the solution process. Departures from the straight line are always in a direction of increased CPU time for a given number of time steps.

Steady State Results

Until the late 1960's, the Madrid aquifer is assumed to be at steady state. The purpose of our steady state simulations was to calibrate the model against recorded data. The finalresult of the five calibration runs performed gave excellent agreement with the data and with the results obtained by the Spanish Geological Service with the Prickett -Lonquist model (1971). A three -dimensional plot of our results is shown in Fig. 5a.

55 600-

100-

10-

100 TIME STEPS

Figure 3. Number of implicit nodes versus time steps for three steady -state runs.Numerals refer to the same runs as those of Figure 4.

The steady state configuration of the watertable in layer 1 at the top isa close reflection of ground surface topography. Since there is no pumpage, convex areas represent recharge, and concave areas represent discharge. All streams are gaining water from the aquifer. Of the total recharge, 60 percent leave the aquifer via river nodes, and 40 percent via discharge nodes.Since many of these discharge nodes are close to the streams, much of the corresponding discharge water must reach these streams via small creeks and superficial alluvium.

The water table configuration in layer 1reflects the existence of local groundwater flow patterns of the kind originally postulated by Toth (1963). The piezometric surface in layer 2 is smoother, ex- hibiting fluctuations of a smaller frequency and amplitude than the water table in layer 1.These fluctuations can be takento represent intermediate flow patterns. The piezometric surface of layer 3 is very smooth, sloping gently from east to west, and indicates the presence of regional flow inthat same direction.

Transient Results

Transient simulations were performed over periodsof 50 years for three different hypothetical pumping regimes. The initial condition in each case is taken to be the steady state. In each case, pumping occurs primarily in layer 1 (to some extent in layer 2) and is heaviest to the northeast, northwest, and southwest of Madrid. A more or less uniform pumping rate similar to the local rate of recharge is imposed on the area east of Madrid. The three pumping regimes differ from each other primarily in the totalrate of extraction from the aquifer.

56 180 5

4 40- 2 W

w 30-

z a-20- c.) 9 X TRANSIENT STEADY-STATE

0 0 2 4 6 8 IO KCYC (x100)

Figure 4. CPU time versus number of time steps (KCYC) for each of nine runs.

Fig. 5b shows the results in all three layers, after fifty years of pumping, at a total rate equal to about 60 percent of total recharge. Maximum drawdown in layer 1 is 130 -140 m (clearly, the drawdown near individual wells may be larger). The local flow regime, as reflected by the water table in layer 1, is now substantially different from what it was at steady state. Whereas at steady state the low points along the water table coincided with streams which effectively divided layer 1 into sub -basins, the lows after 50 years of pumpage are different, and there is some transfer of groundwater from one such sub -basin to another. Layers 2 and 3 again show a gradually smoothed replica of the situation in layer 1. However, the amplitude and frequency of the fluctuations in these layers are larger than at steady state, reflecting the effect of pumpage.

As a consequence of pumping, most of the original discharge zones near centers of pumpageare no longer discharging, and some of the stream reaches lose water to the aquifer. After 50years of pumpage, the total surface discharge is reduced from 100 to 60 percent of total recharge, a rate approximately equal to that of total extraction. This means that about 35 percent of total extraction (equivalent to about 20 percent of total recharge) isderived from storage. If the same rate of pumping continues into the future, the discharge rate will diminish until its sum with the rate of extraction will be equal to the rate of recharge; the system will then be at a new steady state.

Due to the relatively low transmissivity and high storativity of the aquifer, drawdowns are large near pumping centers but the areal extent ofthe drawdown zones is small. To prevent excessive draw - downs in localized areas, wells and wellfields should be spaced at sufficient distances from each other.

Conclusions

The adaptive explicit- implicit quasi three -dimensional finite element modelFLUMPS is well suited for the simulation of steady state and transient three -dimensional flow patterns in theMadrid aquifer. Time steps of one year are treated explicitly with this model, a fact which results in considerable savings of computer time. The model provides a three -dimensional picture of flow in the Madrid basin which clearly illustrates the presence of local and intermediate flow patterns superimposed on the regional pattern. The effect of pumping is shown to be significant but restricted to small areas.

57 Figure 5. Perspective of piezometric surfaces in layers 1, 2, 3 a. at steady -state b. after 50 years of pumpage

Acknowledgments

The authors wish to thank Dr. S. N. Davis for his help and useful comments. This work was supported by Cooperative Research Grant No. T3770171 of the U.S.- Spanish Joint Committee for Scientific and Technological Cooperation. The authors wish to acknowledge M. Busse for drafting.

References

Llamas, M.R., 1976. La Utilización de Aguas Subterráneas en Madrid. De los Mayrat Musulmanes a los Modelos Digitales. Estudios geologicos. Madrid. 32:121 -139.

López García, J., 1979. Historia de la Modelacion del Acuifero de la Cuenca del Tajo. Unpublished progress report of cooperative research between S.G.O.P.U. and University of Arizona.

López Vera, F., 1977. Modelo de Sedimentación de los Medios Detríticos de la Fosa de Madrid. Tecniterrae. Madrid. No. 18.

58 Neuman, S. P., C. Preller and T. N. Narasimhan, 1982. Adaptive Explicit -Implicit Quasi - three- dimensional Finite Element Model of Flow and Subsidence in Multiaquifer Systems.In Press.

Prickett, T. A. and C. G. Lonnquist, 1971. Selected Digital Computer Techniques for Groundwater Resource Evaluation, Illinois State Water Survey Bull. 55, 62 pp.

Sahuquillo, A.,B. Martí, B. López- Camacho, A. Balenilla, L. López García, 1975. Modelo Bidimensional para el Estudio de laComponente Vertical del Flujo en un Acuífero de Gran Espesor. II Congreso Iberoamericano de Geologia Economica. Buenos Aires.

Toth, J., 1963. A Theoretical Analysis of Groundwater Flow in Small DrainageBasins. Journal Geophysics Research. 68(16):4796 -4812.

59 GEOSTATISTICAL ANALYSIS OF AQUIFER TEST AND WATER LEVEL DATA FROM THE MADRID BASIN, SPAIN

Patricia J. Fennessy and Shlomo P. Neuman Department of Hydrology and Water Resources University of Arizona, Tucson, AZ 85721

Abstract

Log -transmissivity, specific capacity, and water -level data from the Tajo River basin surrounding Madrid, Spain, are analyzed by geostatistical methods. The existing log- transmissivity data base is augmented with the aid of regression on log- specific capacities. The augmented set of data is used to obtain estimates of mean log- transmissivities over finite sub -regions of the aquifer by kriging. Krig- ing of water -level data at selected points in the aquifer is accomplished by removing thedrift through an iterative generalized least squares procedure. The covariance matrices of the log- transmissivity and water -level estimation errors are used to investigate the structure of these errors.

Introduction

Ground -water models are becoming increasingly more sophisticated. However, development of a systematic interpretation of the input data has been largely ignored. Kriging may be one way to fill this gap.

In the past, transmissivity and head data were estimated subjectively by hand -contouring, thereby including the modeler's biases. The degree of interpretation varied widely. To minimize this subjec- tivity, automatic contouring and interpolation techniques have been developed. However, most of these methods disregard knowledge of the physical variability of the data (Delhomme, 1978). A measure of the reliability of such contours is lacking. Kriging provides such a measure.

Kriging is an interpolation technique based on the geostatistical theory of Matheron (1963). This technique has the ability to yield unbiased estimates with minimum estimation error variance aswell as a total covariance matrix of the estimation errors. The theory assumes that spatial fluctuation of the data (i.e., log- transmissivity or hydraulic head) about theirmean values can be considered a realization of an intrinsic stochastic process. The process can be characterized by an ensemble mean and a semi -variogram functionwhich are considered uniform over large segments of the aquifer. While log - transmissivity can be assumed to fluctuate about a constant mean value, the fluctuations of head occur about a regional trend, or drift, in the general direction of the gradient. This drift must be removed before the head fluctuations can be treated as an intrinsic process.

Prior to kriging, the semi -variogram is determined based on field observations of data. Using the semi -variogram, kriging computes estimates at selected points, or average values over discrete sub- regions of the aquifer. The method also computes the variance of the estimation error (kriging variance) at each point, or subregion, and the covariance matrix associated with these errors. The kriging error is dependent on the natural fluctuation of the observed data as described by the variogram function, on measurement errors, and on the relative locations of the measurement points. One may therefore expect the kriging errors to be smallest in the vicinity of observation points, and largest far from these points.

This paper describes the application of kriging to log- transmissivity and water -level data from the Madrid basin, Spain. Similar applications to ground -water basins in Arizona were reported earlier by Binsariti (1980) and Clifton and Neuman (1982).

Setting of Study Area

The study area is an elongate northeast -trending alluvial basin surrounding Madrid, Spain (Figure 1). The basin is approximately 160 kilometers long and between 30 and 60 kilometers wide. The study area is bounded to the north and northwest by the Central Range and to the south and east by the Tajo, Jarana, and Henares Rivers.

61 0 25 kilometers AQUIFER PARAMETER MEASUREMENT SITE

Figure 1. Location of Madrid basin, Spain, with aquifer parameter measurement sites.

The climate is continental with a mean annual temperature of 13 °C and a yearlyprecipitation of 400 to 600 millimeters.

The sediments are typical of a continental basin. Fans of coarser detritic deposits spread from the mountains and grade into chemical evaporitic deposits in the southeast. The sediments are hetero- geneous with interfingering of detrital and chemical deposits. Their thickness is believed to reach 2,000 to 3,000 meters (Llamas and Cruces de Abia, 1976).

The aquifer of interest includes the Toledo, Madrid, and Guadalajara groups of thedetrital facies (Llamas and López Vera, 1975). Average horizontal hydraulic conductivity is low, on the order of 0.5 to times smaller (López Vera, Lerman, and Muller, 1 m /day, and vertical conductivities are about 100 1981). in the Recharge occurs by infiltration of precipitation in the interfluvial zones and snowmelt mountains. The aquifer discharges to the rivers.

Kriging Theory

Kriging is a best (minimum variance) linear unbiased estimator (BLUE). As the length of this paper precludes any detailed discussion ofkriging theory, only a brief outline will be given here. Inter- ested readers are directed to David (1977), Delhomme (1978), or Journel andHuijbregts (1978).

has a Let Z(x) be a random function defined in a region, R, which under the intrinsic hypothesis constant mean, E[Z(x)] = mz

and a semi -variogram,

Y(X,x+s) = E{[Z(X*s) - Z(x)]2}

Line where E represents expectation. The semi -variogram depends only on the displacement vector, s. arity implies that the kriging estimator, Zk(x), of the functionZ(x), is of the form

N Zk(x ) = E aiZ *(xi) 1 =1

imposing the where Z *(xi) is thei -th of N observations, and the "kriging weights," ai, are obtained by lead to the following N +1 "kriging conditions of minimum variance and unbiasness. These conditions equations ",

62 N E as((xi -aft i2+u =y(xi ,So) i=1,2 N j=1

N E aj =1 j =1

where Nis Lagrange multiplier, ai2 is variance of measurement error ati -th observation point, y(xi,xa) is semi -variogram between two experimental points xi and x, and y(xi,So) is mean semi -variogram between point xi and subregion So. When kriging at a point, the last term reduces to y(xi,xo), the semi - variogram between experimental point xi and kriging point xo.

The method was applied to log -transmissivity and residual water -level data from the Madrid basin, Spain.

Kriging of Log -Tranmissivity

Regression Analysis

Specific capacity data are available from 299 sites in the Madrid basin (Figure 1). Transmissivi- ties calculated from aquifer tests are available only at 44out of these 299 sites. To increase their number, estimates of log -transmissivities were generated from the more numerous log- specific capacity data by linear regression. The resulting regression equation is

log10T = 0.24685 + 0.70721 log10SC where T is Aransmissivity and SC is specific capacity, both given in m2 /day. The coefficient of determination, R4, for the regression is 0.49. The 95% confidence interval for the predicted log - transmissivity is

1o910T t 2.02 oT(SC)

where aT2 is variance of the error of prediction. These formulae were used to predict log - transmissivities, and to compute the associated error variance at the remaining 255 sites.

Semi -Variogram Calculation and Modelling

The log -transmissivity field is assumed isotropic for the purpose of sample semi -variogram determination. The sample semi -variogram was calculated using thecomputer code SEMIVGM written by Clifton (1981). The resulting sample semi -variogram was fitted with a spherical mode by trial and error (Figure 2). The equation of this spherical modelis

y(s) = .10 + .08[1.5(s/40) - 0.5(s/40)3] s<40 .18 s >40 where s is the distance in kilometers between two points.

The fitted semi -variogram model was calibrated by systematically deleting each data point one at a a time and using the remaining data to estimate the deleted point by kriging. The true errors were unbiased and consistent with the computed kriging errors as outlined by Deihomme (1978).

Kriging of Log -transmissivities

The log -transmissivities were kriged and averaged over 54finite zones of about 100 square kilometers each. The kriging was done using computer code KRIGE developed by Clifton (1981). Each zone was kriged using amaximum of 30 log -transmissivity data points selected from a circular neighborhood of radius 35 kilometers centered on the zone. Data points nearest to the zone centroid were preferentially selected.

Log -transmissivities derived by regression from log- specific capacities have known errors of prediction which are accounted for in the kriging system. Measurement errors associated with log - transmissivities derived from aquifer tests are accounted for indirectly through a jump of the semi - variogram at the origin (nugget effect). Numerical integration was used to calculate mean values of the semi -variogram within the zones.

63 2

ample Semi -variogram o

I G/ Model rn

0 0 10 20 30 40 50 s( Kilometers)

Figure 2. Log -transmissivity sample semi -variogram and fitted model.

After the average log- transmissivities were kriged over each zgne, the corresponding transmissivity estimates, T; xi), and variance of associated estimation errors, aT (xi), were computed fromthe kriged estimates,Zkzi), and the kriging errors, ak(xi). Contours of These transmissivity estimates are plotted on Figure 3. The coefficient of variation, CV, defined as

°T(il) CV * T (xi) is contoured in Figure 4. This figure shows that zones with the most reliable transmissivity estimates (lowest CV values))occur in areas where data points are densely spaced. Where data are scarce, as happens close to the basin margins (particularly on the right and in the lower central areas), the coefficients of variation are relatively large indicating that the transmissivity estimates are highly uncertain.

0 25 kilometers TRANSMISSIVITY IN m2/DAY

Figure 3. Contour map of transmissivity (T *(xi)) derived from kriged log- transmissivity.

64 9 25 kilometers

Figure 4. Contour map of coefficient of variation (CV) of transmissivities derivedfrom kriged log - transmissivities.

It is important to recognize that our kriged transmissivity estimatesare spatial averages over the zones and may thus differ from point transmissivities as determined by pumpingtests or regression from specific capacities.

Without a careful study of kriging errors, or indirectly, coefficients of variation, the kriging results may be misleading. For example, kriged transmissivities in the southeastern portion of the basin are higher than one might expect on the basisof geological considerations; the area consists of evaporites and other chemical deposits which should have relatively low transmissivities. The anoma- lously high kriging estimates in this area are an artifact, caused partly by lack of sufficient data. This lack of data is reflected in the relatively large coefficientsof variation obtained in this area.

Kriging of Heads

Whereas log -transmissivities may be treated as an intrinsic stochasticprocess, the same is usually not true for hydraulic heads. Generally, head data are associated with a non -constant mean, m(x), or drift. In order to use the simple kriging equations given above, the drift must first be removed from the data.

Steady -state water -level measurements are available for 279points in the Madrid basin. Figure 5 shows the location of the measurement wells and water -level contours drawn by the Spanish Geological Service. This contour map was drawn on the basis of more data than those shown inFigure 5. Unfortu- nately, these additional data were not available to us.

An experimental semi -variogram based on the original water -level data demonstrated the need for filtering the drift. We did this by assuming that drift is a polynomial whose coefficients can be determined by generalized least squares. The least squares fitting was done iteratively. We started with a polynomial of first order and assumed, as a first step, that head fluctuations about this polynomial are uncorrelated. After fitting the polynomial to the data, we studied the correlationstructure of the residual head fluctuations by constructing their semi -variogram. This procedurewas repeated with polynomials of second, third, and fourthorder. The third and fourth order polynomials gave similar semi -variograms, and we chose the latter to represent the drift. The least squares fitting, taking the correlation into account, was repeated iteratively untilthere was no significant change from one iteration to the next.

The resulting experimental semi -variogram of the computed residualswas fitted with model (Figure 6). a spherical

y(s) = r1850[1.5(s/18) - 0.5(s/18)3] L1850 s<18 s>l8

The theoretical model was validated, as in the case of log -transmissivities, by systematic deletion of each data point in turn and kriging that value using all remainingpoints.

65 25 kilometers WATER LEVELS IN METERS ABOVE MEAN SEA LEVEL LOCATIONS OF WATER LEVEL MEASUREMENT SITES

Figure 5. Contour map of water -level contours (from Spanish Geological Service) with location of available water -level measurement sites.

Sample Semi -variogram /` 2- . .N- . \ ._. í\Model -

10 20 30 40 50 s( Kilometers)

Figure 6. Hydraulic head residual sample semi -variogram and fitted model.

Head residuals were kriged at 294 nodal points of a finite element grid to be used laterfor aquifer modeling. Kriging was accomplished using a modified version of the computer code KRIGE (Clifton, 1981). Each nodal point was kriged using a maximum of 25 data points in a circular neighborhood of radius 30 kilometers centered on the point. The kriged residuals were then added to the polynomial drift estimates to produce total head values.

66 Figure 7 isa contour map of the kriged steady -state head estimate. The contours are seen to be smoother than those in Figure 5. One reason for the smoothing is the relatively large spacing of kriging points (on the order of 5 kilometers) which precludes the mapping of small scale features. Another reason may be the fact that we used fewer data than those used by the Spanish Geological Service to construct their map.

0 25 WATER LEVELS IN METERS kilometers ABOVE MEAN SEA LEVEL

Figure 7. Contour map of kriged hydraulic heads.

Kriging estimation errors, oH(xi) are plotted on Figure 8. As in the case of log -transmissivities, the errors are seen to be largest in the southeast andnorthern portions of the basin. The reason, as before, is the relatively small amount of data in these areas.

0 25

kilometers I

Figure8. Contour map of hydraulic head kriging errors (aH(xi)).

Conclusions

The following conclusions can be drawn from this study:

1. Kriging is a useful toolfor the interpolation and spatial averaging of log -transmissivities taking into account measurement errors. It provides not only estimates of these values, but also information about the errors of estimation. The method appears to work well for the Madrid basin.

67 2. An iterative method to simultaneously determine the drift and residual variogram of hydraulic heads, using generalized least squares, has been applied with apparent success to the Madrid basin.

Acknowledgments

The authors wish to thank Dr. S.N. Davis for his support, and J. Carrera and E. A. Jacobson for their helpful comments. This work was supported by Cooperative Research Grant No. T3770171 of the U.S. - Spanish Joint Committee for Scientific and Technological Cooperation. The authors also wish to acknowledge S. Adams for typing and M. Busse for drafting.

References

Binsariti, A. A. 1980. Statistical Analysis and Stochastic Modeling of the Cortaro aquifer in Southern Arizona. Ph.D. dissertation, Universtiy of Arizona, Tucson. 242 p.

Clifton, P. M. 1981. Statistical inverse modeling and geostatistical analysis of the Avra Valley aquifer. M.S. thesis, University of Arizona, Tucson. 190 p.

Clifton, P. M. and S.P. Neuman. 1982. Effect of kriging and inverse modeling on conditional simulation of the Avra Valley aquifer in Southern Arizona. Water Resour. Res. In Press.

David, M. 1977. Geostatistical Ore Reserve Estimation: Developments in Geomathematics 2. Elsevier Scientific Publishing Co., New York. 384 p.

Delhomme, J.P. 1978. Kriging in the hydrosciences.Advances in Water Resources. 5:251 -266.

Journel, A.G. and Ch. J. Huijbregts. 1978. Mining Geostatistics. Academic Press, New York. 600 p.

Llamas, M. R. and J. Cruces de Abia. 1976. Conceptual and digital models of the groundwater flow in the Tertiary basin of the Tagus River. Spain. Mem. of the Inter. Assoc. of Hydrogeo. 13. Sym. of Budapest. p. 186 -202.

Llamas, M. R. and C.F. López -Vera. 1975. Estudio de los recursos hidráulicos subterráneos del area metropolitana de Madrid. Avance de las características hidrogeol6gicas del Terciario Detritico de la Cuenca del Jarama -Aqua. Barcelona. 88:36 -55.

López Vera, F., J. C. Lerman and A.B. Muller. 1981. The Madrid basin aquifer: Preliminary isotopic reconnaissance. Jour. of Hydr. 54:151 -166.

Matheron, G. 1963. Principles of geostatistics. Econ. Geol. 58:1246-1266.

68 ENERGY AND WATER RESOURCES INTERACTIONS IN ARIZONA

Nathan Buras, Ph.D. Department of Hydrology and Water Resources University of Arizona, Tucson, AZ 85721

Abstract

Water and energy interact strongly in Arizona. The Arizona State Water Plan mentions that under 1970 normalized conditions 60% of total use in the State was from groundwater aquifers, a proportion which may have increased in the last decade. The utilization of groundwater resources requires substantial amounts of power. In addition, the Central Arizona Project is an energy- intensive pr9ject: its Granite Reef aqueduct will require a pumping lift of 1,084 ft (352 m) using about 1.665 x 10 kwh/ year. The Tucson aqueduct component willhave an additional lift of 997 ft (304 m). The hydropower installations planned within the CAP will have only limited generating capacities: Agua Fria 3 Mw, Granite Reef 3.5 Mw, and Maxwell 11 Mw. The remainder of the load will have to be picked up by thermal power plants and by pumped storage schemes which, by the year 2000, may need over 100,000acre -feet per year to make up evaporative losses. Thus, energy is required to make water available to users, and water isa necessary ingredient in energy -related activities. These and other water -energy interactions in the Lower Colorado Basin are discussed.

Introduction

By and large, the United States has substantial quantities of energy resources and considerable amounts of surface and groundwater. However, in the Southwest and especially in Arizona, water is scarce and most of it is already allocated to various sectors of the economy, the most notable being agriculture and mining. The availability of water resources in reliable quantities and of adequate quality ranks as an important criterion for the siting of thermoelectric power plants.

Water and energy interact in two ways (Buras, 1982). On the one hand, energy is an input to many water resources systems, from pumping of groundwater to treatment of liquid wastes. On the other hand, almost allenergy -related activities use water, either as process water in the production of synthetic fuels or as a transport medium for the removalof waste heat and /or waste matter. Water- energy interactions are shown diagramatically in Figure 1.

Almosobere L

Inter basin Nmpetl Energy transfers w ge

Shea , am outflow Water demand

r Ene

1 MuniclpalenE i

Surface Luoplv Agricultura water treatment

Tra lee Ene. %Warlt tL 1

Ile

Ocean

Figure 1. Water- energy interactions.

69 Water Requirements for Energy -Related Activities

The use of water in energy -related activities is consumptive and it consists of three major components:

1. Process water, such as in the case of synthetic fuels where water contributes to the making of the product.

2. Evaporation, which removes excess (waste) heat from energy -related processes.

3. Waste water, which removes waste matter.

The use of water by energy -related activities is shown in Table 1.

Table 1. Estimated Water Use in Energy -Related Activities (Acre- feet /1015BTU).

Process Waste Activity Water Evaporation Water Total References

Light -Water Reactors 537,200 55,600 592,800a Davis and Wood, 1974

Fossil Fuel Thermal Power 358,100 37,100 395,200b Gold et al, 1977 Stations (No scrubbing)

Coal Gasification, HRTU 32,500 68,100 2,800 103,400 Gold et al, 1977

Oil Shale Conversion 21,700 32,000 8,000 61,700 Gold et al, 1977

Coal Gasification, LBTU 1,000 56,000 700 57,700 Chandra et al, 1978

Coal Liquefaction 2,800 36,700 17,500 57,000 McNamee et al, 1978

Nuclear Fuel Processing 37,400 3,900 41,300 Davis and Wood, 1974

Coal Slurry Pipeline 34,000

Oil Refining 16,000 6,200 22,200 Davis and Wood, 1974

Coal Mining, Underground 7,700 7,700 James and Steele, 1977

Coal Mining, Strip, 3,380 3,380 Gold et al, 1977 Revegetation

Coal Mining, Strip, 1,800 1,800 Gold et al, 1977 No Revegetation

a. 0.66 gal /kwh b. 0.44 gal /kwh

Development of Groundwater

Much of the water currently used in Arizona in energy -related activities is pumpedfrom aquifers, and it is quite possible that this situation will continue in the foreseeable future. It is also con- ceivable that as the demands for water will continue to rise, the different economic sectors in the State will increase their competition for the limited water resources to the point of conflict. It is important, therefore, to develop strategies for the management ofthe groundwater basins so as to meet the competition and avoid the conflict. The groundwatermanagement issue is shown schematically in Figure 2(Domenico, 1972).

The crucial question iswhether toexploit the aquifer within safe -yield limits or to mine the groundwater. An optimal management policy seems to be found between two extreme positions: unregulated mining of groundwater will definitely lead toward the exhaustion of the aquifer within a finite time period, while infinite preservation based on a safe -yield policy may waste groundwater due to the

70 Precipitation Evapotranspiration

1 i

Figure 2. Schematic representation of a groundwater basin.

natural aquifer outflow (Mandel and Shiftan, 1981). Thus the safe -yield concept, based on the natural recharge, should be substituted by the principle of sustained yield. Sustained yield is defined as the maximum rate of groundwater abstractionthat can be maintained over a long period of time (e.g., more than a century) without causing consequences unacceptable from an economic, political or environmental point of view. Examples of undesirable effects are lower water levels and reduced well discharges, deterioration of water quality due to intrusion of saline waters, and land subsidence.The "undesirable effects" may be expressed quantitatively by ranges of values, so thattheir limits can be used as con- straining conditions in mathematical programming models of groundwater systems.

The development of regional groundwater resources should be planned in conjunction with the surface waters. In Arizona in particular, aquifers need to be integrated within the State Water Plan.

The Arizona State Water Plan

The major components of an Arizona Water System could be the following:

surface water: the Salt River Valley Project the Central Arizona Project other surface water projects

groundwater: Tucson Active Management Area Phoenix Active Management Area Prescott Active Management Area Pinal Active Management Area other groundwater basins

The integration of the major (and other) components into a comprehensive system, including its legal framework required for the development, utilization, and management of state's water resources, is a time -consuming process. So far, an inventory of water resources was completed (Arizona Water Commis- sion, 1975), some alternative futures were explored (Arizona Water Commission, 1977), and a number of important water uses were analyzed (Arizona Water Commission, 1978). Currently, the state water planning efforts are concentrated on the groundwater components of the state water resources system (Arizona Water Commission 1980). These efforts should continue so that an integrative plan should emerge in the near future, according to which groundwater aquifers and surface subsystems would be operated conjunc- tively.

The Arizona State Water Plan should include provisions for establishing water quality standards and guidelines for water quality management. Water quantity and water quality are inseparable issues: they influence each other and are integral parts of the natural hydrological processes. Even if it may seem economically attractive and politically acceptable in the immediate or short range to deal with water quality separately from water quantity, the irreversibility of most policy decisions based on this separation may generate considerable social costs in the medium and long time horizon. This becomes particularly obvious when considering that the reclamation of marginal waters (e.g., brackish groundwater, effluent from wastewater treatment plants) is an energy -intensive activity.

71 The Central Arizona Project

The Central Arizona Project will divert from the Coloradoriver about 1.2 Maf /year on the average (U.S. Bureau of Reclamation, 1962). Roughly two -thirds of this amount is allocated to irrigated agriculture and the remainder to the municipal and industrial sectors including electric power generation. The contribution of the CAP to the Arizona state water system will be considerable less than the current groundwater overdraft which isin excess of 2.2 Maf /year.

Energy considerations overshadowed the CAP in the last decade. It was suspected that energy - related development in the Upper Colorado Basin (such as coal -based synthetic fuels, or oil shale exploitation) will deplete allremaining water supplies so that none will be left for CAP. Subsequent studies (Steiner, 1975) refuted these suspicions. Nevertheless, the CAP is an energy -intensive component of the Arizona water resources system. Its Granite Reef aqueduct (capacity, 3,000 cfs; 219 miles long) involves apumping lift of 1,084 ft. The Tucson aqueduct, which will deliver about 100,000 af/ year (capacity, 150 cfs; 56 miles long), will need apumping lift of 97 ft. The annual pumping power requirements for the Granite Reef aqueduct are estimated at 1.665 x 10 kwh. The projected power plants included in the CAP --Agua Fria, 3 Mw; Granite Reef, 3.5 Mw; Maxwell, 11 Mw- -will contribute only mar- ginally to satisfy these requirements. Hence, most of the load needed by the CAP will have to be picked up by power plants outside the project.

Water- Energy Interactions

The Central Arizona Project, one of the most energy -intensive components of the State Water System, is located in a region where electric power requirements are projected to increase more than four -fold during the last two decades of this century: from 6,617 Mw of peak demand in 1980 to 28,532 Mw in year 2000 (Arizona Water Commission, 1971). By the year 2000, it is estimated that the annual consumptive use of water for pumped- storage and thermal power plants in the CAP service area may exceed100,000 af/ year.

A second component where water and energy interact strongly is the desalination plant for the Colorado River International Salinity Control Project. This desalination plant, with a projected output of 101,000 of /year, will require all the power generated by a plant of about 35 Mw capacity (Jacoby, 1975).

Finally, the groundwater basins which supply the bulk of the water currently used in thestate use considerable amounts of power for pumping. The increased demand for water increases pumping rate, which lowers the water table, so that the amount of energy required to lift one unit of water is steadily increasing. It is estimated that each foot of increased pumping head in the southern half of Arizona requires the equivalent of about 6 million kWh of electric energy (U.S. Water Resources Council, 1978).

Summary

The water -energy interaction highlights a very important question: to what extent are water problems in Arizona a barrier to development within the State? Apparently, an answer to this question must take into account four unresolved water issues (Brown and Kneese, 1978): the equity issue, the efficiency issue, the environmental quality issue, and the water development issue. In many cases solutions exist, some of which were already tested, accepted, and arebeing implemented; others are still in the theortical- conceptual stage.

Water is an essential ingredient in almost all energy -related activities, primarily for the removal of excess heat. Energy is a major component of the Arizona State Water Plan, especially when considering that about 60% of the water is pumped from the aquifers and that the major surface water system (the CAP) is energy- intensive. The necessity of studying, planning, designing, and operating the water and energy sectors of the State cannot be overemphasized.

References

Arizona Water Commission. 1971. Lower Colorado Region Comprehensive Framework Study, Appendix XIV: Electric Power. Phoenix, AZ.

Arizona Water Commission. 1975. Arizona State Water Plan, Phase I. Inventory of Resources and Uses. Phoenix, AZ.

Arizona Water Commission. 1977. Arizona State Water Plan, Phase II. Alternative Futures. Phoenix, AZ.

Arizona Water Commission. 1978. Arizona State Water Plan, Phase III- Part 1. Water Conservation. Phoenix, AZ.

72 Arizona Water Commission. 1980. Tenth Annual Report, 1979 -1980. Phoenix, AZ.

Brown, Lee and Allen V. Kneese. 1978. The Southwest: a region under stress. The American Economic Review. 68(2):105 -109.

Buras, Nathan. 1982. Modeling water supply for the energy sector. Electric Power Research Institute report EA -2259. EPRI. Palo Alto, CA.

Chandra, K., B. McElmurry, E.W. Neben and G.E. Pack. 1978. Economic studies of coal Gasification combined cycle systems for electric power generation. Electric Power Research Institute report AF -642. EPRI. Palo Alto, CA.

Davis, G.H. and L.A. Wood. 1974. Water demands for expanding energy development. U.S. Geological Survey circular 703. Reston, VA.

Domencio, Patrick A. 1972. Concepts and Models in Groundwater Hydrology. New York: McGraw -Hill.

Gold, H.,D. J. Goldstein, R.F. Probstein, J. S. Shen and D. Yung. 1977. Water requirements for steam -electric power generation and synethic fuel plants in the Western United States. Environ- mental Protection Agency report no. 600/7 -77 -037. EPA. Washington, D.C.

Jacoby, Jr., G. C. 1975. Overview of water requirements for electric power generation. In: Proceed- ings of the Conference on Water Requirements for Lower Colorado Basin Energy Needs. The University of Arizona. Tucson, AZ.

James, I. C.,II and T. D. Steele. 1977. Application of residuals management for assessing the impacts of coal- development plans on regional water resources. Third Internation Symposium in Hydrology. Colorado State University. Ft. Collins, CO.

Mandel, Samuel and Z.L. Shiftan. 1981. Groundwater Resources. New York: Academic Press.

McNamee, G. P., N.K. Patel, T.R. Roszkowski and G. A. White. 1978. Process engineering evaluations of alternative coal liquefaction concepts. Electric Power Research Institute report AF -741. vol. 1. EPRI. Palo Alto, CA.

Steiner, W. E. 1975. Water for energy as related to water rights in the Colorado RiverBasin. In: Proceedings of the Conference on Water Requirements for Lower ColoradoBasin Energy Needs. The University of Arizona. Tucson, AZ.

U.S. Bureau of Reclamation. 1962. Appraisal Report. Central Arizona Project. Boulder City, NV.

U.S. Water Resources Council. 1978. The Nation's Water Resources 1975 -2000. Washington, D.C.: U.S. Government Printing Office.

73 POTENTIAL ENERGY RESOURCES OF THE GULF OF CALIFORNIA, NORTHWESTERN MEXICO

Barney P. Popkin Water Resources Consultant 5502 Briarwood Forest Drive Houston, Texas 77088

Abstract

The Gulf of California in northwestern Mexico is a tropicalsea of the northeastern Pacific Ocean. Earthquake activity is common in the region, especially towards the northwest where transform faults are associated with volcanic and geothermal activity, and features such asthe San Andreas fault zone. The Gulf contains 132,000 cu km of seawater, witha surface area of 162,000 sq km, and a mean depth of 815 m. Its surface salinity is about 35 ppt. The Gulf is subject to the second highest tides (>10 m in range) in North America. The region is currently undergoing extensive human development and energy exploration. After reviewing the climatic, geologic, soil and vegetation, and oceanographic settings, the potential energy resources of the Gulf are evaluated. These resources include those controlled by climate (solar, wind), geology (hydrocarbon, geothermal), biology (biomass)and oceanography (tidal, wave, hydrothermal). Climatic energy sources (unproven technology) have fair potentialfor modest -scale, onshore and near -use development. Geologic sources are on- line, and have high potentialfor large -scale commercial development with export value. Biological sources(unproven technology)havelow potential for small- scale, near -use development. Oceano- graphic sources have high potential for moderate -scale near -use potentially exportable development.

Introduction

The Gulf of California (Sea of Cortes) in northwestern Mexico is a tropical latitude sea of the northeastern Pacific Ocean adjoined by Baja California Norte and Baja California Sur to the west, and Sonora, Sinaloa and Nayarit to the east (Figure 1). The trough- shaped Gulf contains a very irregular base with severalbasins and hasfree communication with the Pacific at its entrance above depths of 3,000 m. The Gulf was formed from the plate tectonic spreading center on the East Pacific Rise during Tertiary time and exhibits northwesterly transform faults. Earthquake activity is common, especially towards thenorthwest where thefaultsare associated withvolcanicandgeothermal activity and features such as the San Andreas fault zone of California.

The Gulf contains 132,000 cu km of seawater, with a surface area of 162,000 sq km anda mean depth of 815 m. It receives some recharge from rivers to the southeast, although itis excessive winter cooling and summer evaporation which tend to raise the Gulf's surface salinity.The Gulf is subject to the second highest tides in North America, which is partly why novelist John Steinbeck and his marine biologist friend, Ed Ricketts, called the Gulf "a highly dangerous body of water (39)."

Though the region is among the last to be settled in Mexico, it is currently undergoing extensive urban development(based on fishing, tourism, ranching and farming, and mining)and energy exploration. Small scale solar energy research and geothermal energy exploration, and biomass conversion researchbegan in the 1960's and1970's,respectively. Petroleum exploration began in earnest in the early 1980's.

The purpose of this report is to review the potential energy resources of the Gulf of California.

Cultural Setting

In 1900, the population of the Gulf states was 716,100, or 5.3 percent of the national total. By1940 and1980, the population grew from 1,204,100 to 6,292,000 or 6.1 to 8.8 percent of the national total, respectively.

75 This growth represented a 43.2, 51.6, 49.5 and 61.0 percent increase in the Gulf states population in10 -year intervals, while the national population increased by only 31.2, 35.4, 41.0 and 49.1 percent for the same time intervals. The Gulf's population doubled in the past 15 years, while the nationalpopulation has doubledin the past 20 years. Most of this enormous growth occurred in the urban centers. The Gulf states have grown, from fastest to slowest, in this order: Baja California Norte, Baja California Sur, Sinaloa, Sonora and Nayarit. The Baja peninsula is the least densely populated part of Mexico.

Agriculturalproducts from the region include maize, vegetables, cotton and alfalfa (predom- inantly in the Colorado Delta), wheat, beans, sugarcane and rice (Sinaloa and Nayarit) and coffee (Nayarit). Irrigation is commonly practiced. There are also fruit trees, beehives, cattle, horses, mules, donkeys,sheep, goats, hogs and poultry. Fishery production, especially shrimp, is a major activity,with hightonnagefrom the Bajas and highvalue from the Rajas,Sonora andSinaloa. Limited coal and lignite deposits occur southeast of Hermosillo (Sonora). Copper, salt, sand, gravel andlime are important mineral products. Recent tourism and American recreationaland retirement communities have contributed significantly to the Gulf's economy.

The Mexican economy is rapidly expanding and diversifying, while also suffering from the current world -wide recession. Mexico now has an annualfive percent growth (1982)in real gross domestic product, though 10 percent was forecast. There isincreasing demand for professional and skilled workers. While the annual U. S. inflation rate is falling to less than 10 percent, the Mexican inflation is rising to over 30 percent,andthe peso was significantly devalued in early 1982. Inflation in February 1982 reached 53 percent on an annual basis. Mexico, which began exporting energy in the early 1970's, is now the world's fourth -largest producer of oil and gas and the fifth - largest exporter.

The effects of increasing population, agriculture, tourism and supporting industries and activities, coupled with the social expectations of a rapidly evolvingeconomy,hasastronomically increased the demand for energy in the Gulf. Mexico's cultural and economic expectations demand that exportable energy be developed to support growth and security.

Physical Setting

The climate, geology, soiland vegetation, and oceanography of the Gulf of California are the relevant physical components toexamine when evaluating theGulf'spotential energyresources.

Climate

The climate of the Gulf region varies from the arid Rajas and Sonora, the semi -arid Sinaloa, and thesemi -humid Nayarit (20). The Bajas, most of Sonora and north -coastalSinaloa have a hot, desert climate; central -coastalSinaloa and south - coastalSonora have a hot, steppe climate; and Nayarit and south -coastal Sinaloa both have a dry winter, tropical rainy climate (2).

The mean annualair temperature varies from about 15 °C in the mountains of northern Baja to 25 °C in coastal Nayarit. Most of the Gulf hasan average annualair temperature between 20 to 25 °C. The mean annual rainfall varies from less than 250 mm in the northwest, to about 900 mm in the southeast. The area -weighted annualrainfall averages 250 mm, with 25 to 60 percent relative interannualvariability. The average annual evaporation rateis on the order of 100 to 200 mm. Winds are generally less than 10 km /hr, but may exceed 100 km /hr during severe storms.

Geology

The geology of the Gulf is structurally controlled by the San Andreas fracture zone which formed the Gulf (17, 28). The thin, western landmass rises abruptly as a coastal plain to the Baja mountain ranges. The broad, eastern landmass rises gently as a coastal plain of buried mountains. The Gulf drains about 668,800 and 432,000 sq km of land from the U.S. and Mexico, respectively. Nearly all thesurface water runoff from the U.S., and most of that from Mexico, does not contribute directly to the Gulf, but contributes to numerous, small internal basins. Many of these form sahkeh- like playas, especially on the northeast coast of the Gulf.

The Gulf separates the mountainous peninsula of Raja California to the west (a westward -tilted plateau with asteep eastern side) and the broad coastal Sonora plain to the east (a northern late - Cenozoic desert plains and minor ranges, with central and southern modern deltas, lagoons and twisted and elongated plains). Landward from the eastern coastal plain rises basin and range topography to the high range of the Sierra Madre Occidental. The Colorado River has built a large delta with extensive flanking tidalflats at the northwestern end of the Gulf. The terrain sloping toward the Salton Sea from the Colorado Delta is the Valle de Mexicali(Mexicali Valley), and becomes the Imperial Valley in the U. S. The Gulf occupies a structural depression which embraces the Salton Sea in the U.S.

76 The eastern coast is straight, low and sandy, with deep water bays at Topolobampo (Sinaloa) and Guaymas(Sonora), andlagoons enclosed bysand bars or sandy spits. The western coast has the natural bay of LaPaz(Baja California Sur). Guaymas and Mazatlán (Sinaloa) are principal ports.

The southern Gulf has arelatively thin crust, where the Moho is 10 to 11 km below sea level. The Moho is about 23 to 25 km beneath the Baja peninsula and the mainland. The Gulf is structurally part of thePacific Ocean. The Gulf floor's topography suggests that it was formed by lateral movements along northwest- southeast trending fractures associated with the San Andreas fault system. Shallow earthquakes are common. The Gulf was formed from the plate tectonic spreading center on the East Pacific Rise during late Tertiary time. The recent rate of motion between the Pacific and North American plates is about 5.5 cm /yr (28).

Soil and Vegetation

The soilsof the Gulf region are generally poorly developed, well- drained, calcareous sandy alluvial and aeolian soils, with poorly drained flood plain and estuarine silts and lesser, saline and gypsiferous inland -basin silts and clays. Soils tend to be low in organic matter, alkaline and calcium rich. Soil erosion is slight to moderate. There is a yearly soil moisture deficiency, where potential evapotranspirationexceedsprecipitation, except along south -coastal Sinaloaand Nayarit where there is significant soil moisture recharge during summer.

The native vegetationis mostly temperate desert types, except for temperate mesquite grasslands insouth -coastal Sonora. There are also tropical decidous and thorn forests along south - coastal Sonora, Sinaloa andNayarit,andsome tropicalsavannas along south -central Sinaloa and north -coastal Nayarit. Flat coastal marshes support halophytic and estuarine vegetation, especially near the northern Gulf.

Oceanography

The Gulf of California is an arid, tropical latitude sea of the northeastern Pacific Ocean. This elongated, northwest -southeast trough has a very irregular base with ten basins. It has free communication with the Pacific atits entrance above depths of 3,000 m. A central constriction and a group of islands separate the shallow northern section from the deep southern region. The northern section is less than 200 m, while the southern basins are 1,000 to 3,600 m deep. The Gulf contains 132,000 cu km of seawater with a surface area of 162,000 sq km and a mean depth of 815 m. It is 1,500 km long and 100 to 200 km wide.

The Gulf receives some limited flow from the Colorado River and perhaps as much as 18.3 x 109 cu m /yr of recharge and 162 x106 metric tons /yr of sediment from six eastern rivers. Excessive winter cooling and summer evaporation tend to raise the Gulf's surface salinity. Its surface salinity is about 34 to 36 ppt. No permanent streams exist on the western, or northeastern banks, but there are permanent streams in the southeast. Surface water in the northern Gulf ranges from 16 to 19 °C in winter and from 29 to 30 °C in summer.

In winter, the northern Gulf's waters are almost isothermal from surface to bottom because of convective mixing. In summer, evaporation produces marked stratification with a warm, slightly more saline surface layer. In the deeper northern basins, strong tidal mixing produces ahomogeneous water mass. In the southern Gulf, the water is the Equatorial Pacific.

In the central and southern Gulf, water below the thermocline isidentical in salinity and temperature to equatorial Pacific water, witha pronounced oxygen minimum between 400 and 800 m. Above the thermocline, water is modified slightly by evaporation. In the south, water above the thermocline is modified by surface inflow of equatorial Pacific water and water from the California current flowing around Baja.

Surface water circulation is controlled by seasonal wind regimes. Water is driven out of the Gulf by a low- pressure system east of the Gulf and replenished by deeper flow from the Pacific in winter. Water is blown into the Gulf from the Pacific surface by a low- pressure system over the northern peninsula and leaves the Gulf by deeper flow in summer. Upwelling occurs along the eastern bank in winter and along the western bank in summer, resulting in high plankton production.

The mean tidalrange in the Gulf increases from about 1 m in the south to 8 m near the mouth of the Colorado, where it exceeds 10 m at spring tides. The northern Gulf is subject to the second highest tides in North America. The tidal wave isa progressive, northward -traveling wave, where tidal current, velocities are highin the northern Gulf and between the large islands of Tiburón (Sonora) and Angel de la Guarda (Baja California Sur).

Energy Resources

The energy resourcesof the Gulf of California are controlled by climate, geology, biology and oceanography. These willbe either renewable or nonrenewable, and exportable or exclusively domestic. The Appendix presents energy or work, and power unit conversions.

77 Climatologically Controlled Energy

Solar and wind energy are controlled by climate. These energy sources are renewable, currently exclusively domestic, with fair potential for development in the Gulf. Currently, they are merely research topics. These energy sources are available onshore and near ultimate use.

Solar resources. Solar resources are currently used for desalination, heating and cooling, irrigation pumps, corrosion control of pipelines, and to activate photovoltaic cells which directly convert sunlight into electricity. The estimated total annual energy flux from the sun's radiation is about 1041 ergs (26). Photovoltaic power supplies about a million watts world wide for use in remote places such as weather stations, cathodic protection and ocean buoys (6). There are solar villages, solar schools, solar ponds and solar shields (1). Leading research centers and universi- tiesin the Middle East and U.S. Southwest, including the University of Arizona, have extensive on -going solar research projects. The Gulf, because of its abundant sunlight, has afair to good potential for solar energy development.

Wind resources. Wind energy requires a minimum wind speed of 10 to 40 km /hr to be practical (16). The estimated totalannual energy flux from near -surface wind is about 1029 ergs (26). The Gulf, because of its low to moderate winds, has a poor to fair potential for wind energy development.

Geologically Controlled Energy

Hydrocarbon and geothermal energy are controlled by geology. They are nonrenewable forms of energy which are exportable. There is high potential for geologically controlled energy sources in the Gulf on a commercial scale.

Hydrocarbon resources. Hydrocarbon resources are particularly attractive because their conversion to energy isan established technology, refining produces numerous useful by- products, and such resources have an immediate exportable value. A 42- gallon (159 -1) barrelof crude petroleum produces 5 800 x 103 BTU (1016 ergs) and a dry cu ft (0.0283 cu m) of natural gas produces 1.021 x 103 BTU (103 ergs).

Thegeological evidencefavorablysuggests hydrocarbon accumulationand entrapment in the Sebastian Vizcaino Embayment,theEugeniaAbrejos Province and theIray Puríima Regionof the Pacific side of Baja California Sur, peripheral to the Gulf (4, 32).The Sebastian Vizcaino covers about 13, 000 sq km and contains 3.0 km or 29,200 cu km of sediments. The Eugenia Abrejos covers about 5,200 sq km and contains 9.1 km or 25,000 cu km of marine and offshore sediments adjoining the Pacific Ocean. The Iray Purisima covers about 12,000 sq km and contains 3.0 km or 45,900 cu km of sediments adjoining the Pacific opposite La Paz(Baja California Sur). Gas strikes in Baja were announced by Pemex in 1978 (24).

Hydrocarbon shows were reported in the Huichol wildcat drilled by Pemex in the Pacific waters off Nayarit and east of Islas Marias in 1979 (25). Pemex studies found sedimentary basins covering 77, 700 sq km in waters off Sinaloa, Nayarit and Jalisco in 1979. In 1979, Pemex completed Jalisco #1 as an oil discovery on the Pacific Coast (9). The Oanwood Ice drillship was used to drill Totoaba #1 (dryand abandoned) off the Pacific coast. The Zapata Trader drillship (renamed Cora) drilled Huichol#2, which reported gas shows. In 1980,Pemex explored the Gulf and found no shows (9), though researches have found petroleum in offshore geysers (40). Pemex began a seven -well advanced exploration program both on and offshore from the Gulf and past the tip of Baja to Manzanillo. A wildcat 21 km offshore from Nayarit found gas in producible amounts (42).

Guzman (21) reviewed the petroleum possibilities in the Altar Desert of northwestern Sonora, northeast of the Gulf. Gravity, magnetic and seismic surveys indicate that the Altar Basin has good prospects for petroleum production. The basin covers about 15,500 sq km and contains 4.0 km or 46,500 cu km of continental and marine sediments. The Basin, which extends south to Tiburón Island (Sonora), has a very close relationship to the San Andreas fault system to the southwest. Since about 90 percent ofCalifornia's basin petroleum production is from similar Miocene toPliocene sediments, Altar's marine sediments are likely to be productive. Drag folds due to movement of the San Andreas fault produced structural traps. Displacement along the fault system is about 350 km. Should the rocks be returned to their original position, the Altar Basin would be opposite the productive Los Angeles Basin. Drilling near Hermosillo (Sonora) indicated oil shows in Eocene Rocks.

In early May 1981, Pemex announced that it plans astepout from its indicated commercial gas strike in the Gulf off the mouth of the Colorado River (35). It reported that its Extremeño #1 flowed 5.7 MMcfd (0.16 MMcmd) of gas plus some condensate through a 1 /4 -in. (0.64 mm)choke from an intervalof 13,510 to 13,543 ft (4,118 to 4,128 m). The total depth of the well was 15 746 ft (4,799 m). The strike was drilled by the Reforma drillship in 130 feet(40 m) of water, 80 km south of Mexicali (Baja California Norte) and north of Montage Island.

The Extremeño #1 strike was from a Pleistocene sand at 16 km offshore from the mouth of the Colorado River, from a structure thought to cover 60 sq km (42). The strike was the first to establish commercial gas production in Mexico's western offshore area. According to World Oil (42),

78 Pemex has been looking for production in this horizon since 1948 and has drilled about 40 wells in the past seven years. Fourteen of these had shows. In 1975, Pemex found subcom- mercial gas at Catrina I and II onshore near Guerro Negro. Exploration picked up in 1978 when Pemex brought three floating rigs to the Sea of Cortes, in the Pacific, and west and south of the Baja peninsula.

The Extremeño #1strike sparked land leasing for petroleum rights in theImperial Valley of California (36). The Oil& Gas Journal(36) noted that "geothermal hot waters probably have been flushing the hydrocarbons up dip along the flanks" of the Valley, where a number of noncommercial gas shows have been found at 600 to 1 200 m.

Hydrocarbon discoveries are expected in Baja California, Chihuahua, the Gulf of California and offshore Mazatlan in the medium term in northwestern Mexico (42). However, World Oil(43) projects a 13.6 percent decline in the number of wells to be drilled in 1982 over 1981, which may delay new discoveries and encourage cautious production in proven fields in and near the Gulf of Mexico where Mexico's production is secured. Twofactors contribute totheprojected decline: 1) Mexico's "cardinal sin of borrowing heavily against resources in the ground" at the same time that there is a world -wide decline in demand or unexpected- softness of the world crude market; and 2) Mexico's politicalclimate,in which drilling traditionally declines inan election year. Mexico lost at least five billion dollars in anticipated revenues as a result of the 1981- 1982 world oilglut. However, world -wide market or supply changes, such as a new Middle East war or a world -wide economic recovery, would increase the value of Mexican oil.

Mexico became an oil exporting country in the early 1970's. From 1976 to 1981, Mexico annually produced about 1.5,1.6, 2.0,2.3, 3.3 and perhaps 4.5 to 5.0 percent of the world's crude oil. Its anticipated that this percentage will increase in the mid term, with increasing production from existing and new fields, including those yet to be developed in the Gulf.Mexico's proven reserves are estimated to be about twenty percent of those of the U. S.

Mexico'sprospective areasof major hydrocarbon fieldsinclude 60, 000 sqkm offshore from Sinaloa and Nayarit, 68,000 sq km in western Baja California Sur, 84,000 sq km offshore at the mouth of the Gulf, and 245,000 sq km in northwestern Sonora (3).

Geothermal resources. Geothermalenergy is extracted from superheated underground steam or water with temperatures often greater than 300 °C. The estimated total annualenergy flux from thethermalgradient is about1028 ergs(26). At least 22 countries are developing geothermal resources to produce electricity.

The Oil& Gas Journal (34) reported the beginning of a five -year study of the hot -brine geo- thermalresources of the Cerro Prieto area, 32 km southeast of Mexicali and the California- Mexico border in Baja California Norte. At that time, the area was the site of a geothermal plant with an electric generating capacity of 75 MW and an ultimately planned annual production capacity of 440 MW. The Cerro Prieto Geothermal Field in Mexico's Mexicali Valley had its first wells drilled in 1959 and 1961 as a result of hot springs, geysers, mud pots and fumaroles observed near Laguna Volcano about 8 km southeast of the volcanic Cerro Prieto. The Cerro Prieto GeothermalField isa hot -water dominated system. Underground hot brines contain pressured water that "flashes" into steam at the earth's surface. Water and steam from the Cerro Prieto system is reported to be as hot as 330 °C.

The GeologicalSociety of America led a November 1979 field trip to the Salton Trough, and reported that the Cerro Prieto Field completed its 75 -MW power plant in 1973, and increased production to 150 MW in April 1979(12). By November 1979, 65 deep geothermal wells were completed in the field.

Several geothermal anomalies which lack surface expression have been discovered in the Imperial Valley and the Mexicali Valley. By 1979, 63 geothermal production wells were drilled in the Imperial Valley for a 10 -MW plant, and several additional power plants are planned. These include Republic Geothermal's 50 -MW dual -flash power plant at East Mesa, Southern California Edison's 10 -MW plant near Brawley, Southern California Edison /Union Oil's 10 -MW plant and Southern California Edison's 55 -MW plantnear RedHill Volcano at the Salton Sea GeothermalField, and Southern California Edison/ Chevron Resources' 50 -MW unit at the Heber Geothermal Field near El Centro.

Accordingto the U. S. Geological Survey, six identified geothermalfields in the Imperial Valley with reservoir temperatures greater than 150 °C could produce a total of 6,800 MW for 30 years, or 225 MW /yr(5). Additionally, there are numerous smaller experimentalgeothermal plants in the Imperial and MexicaliValleys. Western Services,Inc. estimates that the Imperial Valley could provide up to 10,000 MW of power over the next 30 years (16).

There are350 °C -hot springs on the 2,650 -m deep sea floor ona mid -ocean ridge in a small region along the crest of the East Pacific Rise near the entrance to the Gulf of California (29).

79 Additionally, 315 °C -hot geysers have been found in petroleum -seeping mounds at 2 000 -m depths along transform faults offshore Baja California Sur in the Gulf (40). Numerous indications of hydrothermal activity suggest that productive geothermal fields may be present beneath the Gulf.

Biologically Controlled Energy

Biomass energy is controlled by biology in the Gulf. This is a renewable resource which is currently exclusively domestic. The estimated total annual energy flux from marine biomass is about 1028 ergs (26). Currently, the biomass energy source is merely a research topic, where conversion of fiber to energy is a promise.

Though forest and wood products are scarce, there is an abundance of natural salt -tolerant or halophytic plants which may be irrigated with sea water to produce a locally available, biomass energy source. Cooperative research studies in this regard have been conducted by the Univegity of Arizona's EnvironmentalResearch Laboratory and the Universidad de Sonora near Puerto Penasco (Sonora)since 1978. More recently, the University of Arizona's College of Agriculture and the Office of Arid Lands Studies have entered this research area (33).

Mature Atriplex (salt bush) and Salicornia (pickleweed) contain about 35 and 20 percent of crude fiber, and about 35 and 40 percent ash, respectively (22). Soaking the ground product in 10 volumes of fresh water for 16 hours can reduce the ash content in saltbush and pickleweed to about 15 and 7 percent, respectively,to make a more energy -rich product. These leached plant products could provide a limited, near -site energy supplement.

There are arid -land plants, such as Simmondsia (jojoba) and Euphorbia (gopher plant), which have naturally high hydrocarbon contents (27). Jojoba data are sparse, but gopher plant data from Californiaindicate production of10 to 25 barrels of crudeoil /ha /yr (104 /ergs ha /yr). Addi- tionally, Salsola (common Russian thistle, or tumbleweed) have a high energy content and could yield about 3, 500 cal /gor about 32 billion cal /ha /yr (1015 ergs /ha /yr)from fertilized and irrigated land (19). Unfortunately, jojoba, gopher and tumbleweed plants require fresh water.

Combined agricultural production and biomass energy conversion may be feasible in the Gulf if an agronomic crop, with high energy conversion, could be cultivated. On the big island of Hawaii, for example, over 40 percent of all energy needs are met from biomass energy conversion of sugarcane wastes. Additionally, marine biomass power could be developed in shallow water through eelgrasses and kelps which are efficient producers of organic material and are easily harvested (26).

Oceanographically Controlled Energy

Tidal, wave and hydrothermal energy are controlled by oceanography in the Gulf. These energy sources are renewable and potentially exportable. These energy forms generally require no storage systems. IsaacsandSchmitt (26) discussed the nature and distribution ofnon -petroleum power sources of the sea, including waves, tides, currents, salinity and temperaturegradients, as well as submarine geothermal sources, salt domes, ice and other marine -associatedconcentrations.

Tidal resources. Tidal energy uses the oscillatory flow of water in the filling and emptying of partially enclosed coastalbasins during the semidiurnalrise and fallof the tides (11). This energy maybe partially converted into renewable tidalelectric power by damming such basins to create a differencein water level between the ocean and the basin, and then using the water flow while the basin isfilling or emptying to drive hydraulic turbines to propel electric generators. The hydraulic head for power generation can generally be available for 6 to 12 hrs daily.

The estimated totalannualenergy flux from major ocean currents and feasible tidal power is 1025 ergs each (26). Globally, near -shore areas with a great tidal range and potential for tidal power development are widely distributed; they include the coasts of Alaska andBritish Columbia, the Gulf of California, the Bay of Biscay, the White Sea, the centralIndian Ocean, and the coasts of Maine and eastern Canada. Existing installations are in the French Rance River estuary (producing 240 MW) and at Kislaya Bay in the Soviet Union (producing 440 kW).

The energy, E, produced between high water and low water (10) amounts to:

E = mgh = 8 gpAs2 / T where m is the mass of water, q is the gravitational acceleration, h is the tidal range, p is the water density,A is the cross -sectional area, s is the wave amplitude, and T is the tidal period. The mean tidal range is 2s, and the water volume flowing through a hay during half a tidal period is 2sA. Theoretical power output, TPO (kW), from a hydromechanical device can be calculated from the following equation (30): Q (cfm) x H (ft) O (cfs) x H (ft) _ Q (lps) x H (m) TPO = 708 11.8 102

80 where Qis the flow of water through the perfect power generating system, and H is the hydraulic head. If variable head drives the device, the energy produced is about 90 percent of what is calculated. In practice, head losses from pipe flow and fittings must be subtracted from the hydraulic head, and the overall efficiency must be multiplied by the TPO to derive the operating power output. Actualpower output,APO (kW), can be calculated from APO = OEx TPO, where OE is the overall efficiency of the system (usually 0.50 to 0.80). The electrical output, EPO (kW), from a turbine can be calculated from the following equation (30):

D (ft) x C (cf) x N x R x F D (m) x C (1) x N x R x F EPO 708 102 where Dis the diameter of the turbine, C is the working volumetric capacity of each cell, N is the number of cellsin the turbine, R is the rotational speed in rpm, and F is the turbine efficiency (usually 0.60 to 0.80).

Based ona review of published figures for one existing and six planned power stations (7, 11, 13 -15, 38), it appears that the tidal power plant with 1 -m tidal range and 15 000 -m barrage length produces about 1 000 -MW power output, or power output (MW) equals 15 times tidal range (m) times dam or barrage length (m). The Gulf has good to high potential for tidal energy development.

Wave resources. Wave energy uses the up and down motion of the waves to drive air turbines to propel electricalternators (38). This renewable energy source is still in early experimental stages. The estimated totalannual energy flux from waves is about 1027 ergs(26). Waves could produce about 50 kW per meter of wave, assuming 50 percent efficiency for generation and 50 percent efficiency for transmission losses (37). About 1,000 km of devices could produce 12 kW in sites off the coast of the United Kingdom. The Gulf has fair to good potential for wave energy development.

Hydrothermal resources. Hydrothermal energy uses warm currents from tropical waters to heat a fluid, such as ammonia, and turn it into vapor which drives a turbine to produce electricity. Cooler seawater from depths of about 900 m cools the vapor and condenses it to form water needed to repeat the cycle. Hydrothermal energy could theoretically be harvested by ocean thermal energy conversion (OTEC) plants (6), where a 64 ° -F (36 ° -C) range of water temperature is available. The Gulf has a high potential for hydrothermal energy development.

Technical Feasibility

The technical feasibility of an energy resource may improve with time dueto of technical advances and economic demand.

Climatic energy sources are currently an unproven technology with storage, transmission efficiency,electronic andstructural problems. Solar energy generallyrequiresan energy storage system, and wind energy demands fatigueless metals or ceramics. These sources are currently research areas with fair potential for modest -scale development in the Gulf. Geologic energy sources have a proven technology with technical limits. Hydrocarbon energy has safety and air quality problems, and geothermal energy has turbine efficiency, structural and fatigue, and corrosion problems. Offshore development has unique structuraland corrosion problems. These sources are currently on -line with high potential for large -scale commercial development in the Gulf. Biological energy sources are an unproven technology riddled with biological uncertainties, corrosion and air quality problems. These sources are currently research topics with low potential for small -scale development in the Gulf. Oceanographic energy sources have turbine and transmission efficiency, and corrosion problems. Tidal energy has a proven technology, though most plants are research plants. Wave and hydrothermal energy production are currently research projects. These sources have a high potential for moderate -scale development in the Gulf.

Environmental Impacts

The environmental impacts of energy exploration, development and transmission can be considered in terms of air, water, soil and biota.

Climatic energy has a low impact on air, water, soiland biota. There is no air pollution. Development would require consideration of drainage and erosion impacts. Geologic energy has moderate, high, moderate and fair impact on air, water, soiland biota, respectively. Development of geologic energy would require consideration of brine disposal, spills and water demand. FERTIMEX is currently building a potassium chloride recovery unit for geothermal plants in Cerro Prieto, indicating that waste recovery is feasible. Biological energy has a fair, moderate, fair and small impact on air, water, soiland biota, respectively. Development would require consideration of saltwater agriculture, drainage, soil salinity and water demand. Oceanographic energy has a low impact on air, water, soil and biota. There is no air pollution.Development would benefit by water storage, which would reduce operating costs and prolong operating hours.

81 Any offshore and nearshore energy development to the Gulf must consider the impact of seismic activity. The eastern Pacific Shelf off southern California near Baja has had 24 earthquakes in the past 60 years of a local magnitude 6 (Richter scale) or larger (23). The Pacific and the Gulf coasts are subject to earthquakes, tsunamis seismic sea waves), and slumps and slides.

Mid- 1980's Production Costs

Assuming a 15 percent annual inflation, it will cost about $350 /kW and $2,000 /kW in capital for a natural gas /oil -fired steam plant and a nuclear power plant, respectively (38).Currently designed and nearly completed coal -fired steam plants are costing about $800 /kW and $1,600 /kW, respectively in the Texas gulf coast. The following costs are projected in the Gulf for comparative purposes only.

Climatic energy would cost about $40,000 to $50,000 /kW for solar (6)and wind forms, while geologic energy would cost about $350 to $3,500 /kW for hydrocarbon (38) and geothermal (6, 16) forms in the mid- 1980's. Biological energy would cost more than $50,000 /kW. Oceanographic energy would cost about $7,600,$35,000 and$1,200 /kW for tidal (8, 13 -15, 38), wave (38) and hydrothermal (6, 38).

These cost estimates could be reduced with technical solutions to engineering problems. For example, turbine and metal fatigue problems caused by brine in geothermal power units has recently forced redesign for a second 10 -MW plant operated by Southern California Edison. This has doubled capital costs in the Imperial Valley to $3,000 /kW (16). Technical solutions could reduce this cost in future plants.

Energy Investment Strategies

There are four general methods used for selecting alternative energy investment projects (31, 41). These include accounting (average rate of return, payback) and discounted cash flow (internal rate of return, net present value) methods. Data are unavailable to allow comparison of investment strategies between climatic, geologic, biological and oceanographic energy.

In Mexico, the federal government determines and implements the country's energy policy through Pemex (Petróleos Mexicanos)for hydrocarbons and the Comisi6n Federal del Electricidad (CFE) for electric power at the operating level. On March 18, 1938, Mexican President Lazaro Cárdenas nation- alized the oilindustry, and established Mexican control. More recently, however, Brown & Root, a division of Halliburton, Inc., has become the prime contractor for Pemex, responsible for organizing the engineering, purchasing, building and managing of Mexico's major fields (18).

Mexico's energy program for the 1980's, announced in November 1981, is that export sales will be guided by the need to finance intrastructuralimprovements and to achieve nationalsecurity goals (3). Mexico in the early 1980's plans to conduct exploratory drilling activities in only about 10 percent of the areas which are likely to contain hydrocarbons.

Acknowledgements

I thank the research libraries at Gulf Publishing (World Oil), PennWell Publishing (Oil& Gas Journal), the Fondren Library at Rice University andtheScience Library atthe University of Arizona, the Houston Public Library, the University of Houston Bookstore, and the Brown Book Shop,

Inc. I also thank Georgia A. Henderson for editorialreview, and Ella Cook and Sandy Jackson for manuscript preparation.Dames & Moore supported the preparation of this work.

References Cited

1. Aramco World Magazine. Sept. -Oct.1981. Saudi Arabia andsolar energy -a special section. Aramco World Magazine 32(5): 18 -29.

2. Arbingast, S. A. and others, eds. 1975. Atlas of Mexico,2nd ed. University of Texas at Austin, Bureau of Business Research.

3. Baker,G. Mar.1981. Mexico's energy plan puts exports at1.5 MMbopd in 80s. World Oil193 (3):109 -110, 114, 116, 118.

4. Beal, C. H. 1948. Reconnaissance of the geology and oilpossibilities of Baja California, Mexico. Geological Society of America, Memoir 31.

82 5. Brook, C. A.and others. 1978. Hydrothermal convection systems with reservoir temperatures greater than 90 °C. In: Assessmentof the Geothermal Resourcesof the United States. L. J. P. Muffler, ed. U.S. Geological Survey, Circular 790.

6. Changing Times. Sept. 1979.Eight ideas that promise more energy. CT 33 (9):25 -28.

7. Clark, W. 1987. Energy for Survival. Anchor, Garden City, N. Y.

8. Cousteau, J -Y. 1981. The Cousteau Almanac - An Inventory of Life on Our Water Planet.Double- day /Dolphin, Garden City, N. Y.

9. Deal, C. 1981. South America, Central America, the Caribbean, and Mexico. American Association of Petroleum Geologists, Bulletin 65(10): 1940 -1995.

10. Dietrich, G., K. Kalle, W. Krauss and G. Siedler. 1975. General Oceanography - An Introduction, 3rd ed. Wiley, N. Y.

11. Dorf, R. C. 1981. The Energy Fact Book. McGraw -Hill, N. Y.

12. Elders, W.A., ed. 1979. Guidebook: Geology and Geothermics of the Salton Trough, Field Trip No. 7. Geological Society of America, 92nd Annual Meeting, San Diego.

13. Electric World. May 1981. Interest in small hydro is rising. EW 195 (5): 23 -24.

14. Engineering News Record. Oct. 29, 1981. Tidal power test taps Fundy flow: Nova Scotia energy dreams come true. ENR, p. 39 -40.

15. . Nov. 12, 1981. Alaskan tidal power costly. ENR, p. 40.

16. . Dec. 24, 1981. Geothermal power boom awaits developer push. ENR, p. 11.

17. Encyclopedia Britannica. 1977. Macropadia, 15th ed., Vol. 8.

18. Forbes. Aug. 15, 1977. Pemex to Brown A Root: Yankee, Come In. Forbes 120(4):28.

19. Foster, K. E. Dec. 1979. Arid lands biomass production in Arizona. University of Arizona, Arizona Water Resources Project Information, Project Bulletin 22.

20. Goodrich Euzkadi. 1964. Atlas, Caminos de Mexico.Galas de Mexico, S. A., Mexico.

21. Guzman, A. B.Aug. 24, 1981. Petroleum prospects in Mexico's Altar Desert. Oil & Gas Journal 79(34):110 -126.

22. Hodges, C. N.and others. Dec. 1980. Seawater -based agriculture as afood production defense against climatevariability. University of Arizona, EnvironmentalResearch Laboratory.

23. Howell, D. G., D.S. McCulloch and J. G. Vedder. 1968. General geology, petroleum appraisal, andnature of environmentalhazards, eastern Pacific Shelf,Latitude 28° to38° North. U. S. Geological Survey, Circular 786.

24. International Petroleum Encyclopedia. 1979. Petroleum Publishing, Tulsa, Okla.

25. . 1980. PennWell Publishing, Tulsa, Okla.

26. Isaacs,J.D. and W. R. Schmitt. Jan. 18, 1980. Ocean energy: forms and prospects. Science 207:265 -273.

27. Johnson, J. D. and C. W. Hinman. May 1980. Oils and rubber from arid land plants. Science 208: 460 -464.

28. Kennett, J. P. 1982. Marine Geology. Prentice -Hall, Englewood Cliffs, N. J.

29. Macdonald, K. and B. P. Luyendyk. May 1981. The crest of the East Pacific Rise. Scientific American 244(5):100 -116.

30. McGuigan, D. 1978. Harnessing Water Power from Home Energy. Garden Way, Charlotte, Vt.

31. Megateli, A. 1980. Investment Policies of National Oil Companies. Praeger, N. Y.

32. Mina, F.1951. Mexico - Lower California. In: Possible Future Petroleum Provinces of North America - A Symposium. M. W. Balland others, eds. American Association of Petroleum Geologists, Tulsa, Okla., p. 242 -244.

83 33. Neary, J. June 1981. Pickleweed, Palmer's Grass and Saftwort. Science 81 2(5):38 -43.

34. Oil & Gas Journal. Nov. 21, 1977. Geologists studying hot -brine Cerro Prieto area in Mexico. OGJ 75(48):61.

35. . May 11, 1981. Newsletter. OGJ 79(19).

36. . June 8, 1981. Pemex offshore strike sparks California leasing. OGJ 79(23): 62.

37. Ross, D. June 9, 1978. Will Britain miss out on wave power? New Statesman 95:762 -763.

38. Savage, J. A., E.E. Weynard and W. G. Wyatt. Jan. 1975. Potential of Tidal and Gulf Stream Power Sources. The State of Texas, Governor's Energy Advisory Council, ProjectN /T -9.

39. Steinbeck, J.and E. F. Ricketts. 1941. Sea of Cortez:A Leisurely Journalof Traveland Research. APPEL, Mamoboneck, N.Y.

40. Sullivan, W. Jan. 29, 1982. Divers find natural"oil refineries ". The New York Times, p. Al2.

41. Van Horne,J. C. 1971. Financial Management and Policy, 2nd ed. Prentice -Hall, Englewood Cliffs, N. J.

42. World Oil. Aug.15, 1981. North America: U. S. activity is unprecedented. WO 193(3):57 -.

43. . Feb. 15, 1982. World drilling moves into new countries.WO 194(3):213.

Appendix

Energy or Work

One British thermal unit (BTU1 equals 1.0548 x 1010 ergs or dyne -centimeters (dyne -cm), 777.97 foot -pounds (ft -lb), 3.9292 x 10-4 horse power -hour (hp -hr), 1054.8 joules (J) or newton -meters (N -m), 0.25198 kilogram -calorie (kg -cal) or 2.930 x 10-4 kilowatt -hour (kW -hr).

Power

One horse power (hp) equals 42.418 British thermal units per minute (Btu /min), 7.4560 x 100 ergs per second (erg /sec) or dyne -centimeters per second (dyne -cm /sec), 550 foot -pounds per second (ft -lb/ sec), 10.688 kilogram -calories per minute (kg -cal /min), 0.74570 kilowatt (kW), or 745.70 watts (W) or joules per second (J /sec).

84 CALIFORNIA L SANDIEGO TIAUANA-_ Y U M A AR IZ O N A N E W M E X I C O MEXICALI o-o-o-o- \

ENSENADA e' `` TUCSON v L. úsw \,.___--- -°\\L PASO 2 LJ ` TEXAS o ``` 10 1t A9 \ C 0- ( 0 $ O N O R A ) ÍñA iNCEI !

C H I H U A H U A\N"N. I HERMOSILLO a , (-`---) CHIHUAHUA CDAYMAS l (/ 1 o CUIDÁD MOON . 1 T

DPDLDIAMID S

o D UR A N G O CULIACAN'

DURANGO LA PAZ

TROPIC OF CANCER o

MAZATLAN

o ao ioo iw SCALE IN RRAN[T[AS

JALISCO

GULF OF CALIFORNIA, MEXICO FIGURE 1

85 Nonstructural Flood Control Evaluation for Tucson, Arizona

Kebba Buckley Contractor to U.S. Army Corps of Engineers, Tucson, AZ

Abstract

Nonstructural and nontraditional flood control measures are required to be included in any Corps of Engineers Urban Study. However, Corps policy and practice are not specific as to evaluative techniques. In Tucson,a preliminary study of nonstructural and nontraditional measures was conducted for several dozen sites along eight watercourses.The evaluation showed that such measures appear feasible for about one -third of the sites, and further study of these sites is warranted. When a more complete data base is available for these watercourses, more detailed evaluations may be undertaken.

87 IMPACTS OF THE ARIZONA GROUNDWATER ACT ON TUCSON WATER

Stephen E. Davis Tucson Water P.O. Box 27210 Tucson, Arizona 85726

Abstract

The 1980 Arizona Groundwater Management Act is the product of decades of court decisions and years of work and negotiation by representatives of the State's major water users, members of the State Legislature, and the Governor. The Act intends to conserve and manage the State's groundwater resources through the establishment of the Arizona Department of Water Resources (DWR). The Tucson Active Management Area (TAMA) is one of four geographical areas specifically designated within the legislation. There is established a Tucson office, locally staffed and administered by the DWR.

The legislated goal forthe TAMA is to balance groundwater withdrawal with dependable supplies by the year 2025. A series of time -specific management plans will incorporate conservation, supply augmentation, and farmland retirement to fulfill the 2025 goal.This paper discusses the existing and projected TAMA water supply /demand imbalance assuming no groundwater management and potential impacts of the Act on management and customers of the Tucson Water Utility. Specific positive impacts include increased public awareness, regional metered water use information, and reduction in groundwater overdraft. Specific negative impactsinclude more regulation, higher customer water rates, reduced water quality, and potential growth limitations.

Introduction

An appreciation of the history of Arizona's groundwater law is pertinent to the discussion of the impacts of the 1980 Groundwater Management Act on the Tucson Water Utility. Increasing withdrawals of groundwater by agricultural users in the past resulted in a marked increase in groundwater overdraft in specific areas of the State. During the 1940'sthe Department of the Interior threatened to withhold support for the Central Arizona Project unless something was done locally to limit the unrestricted pumping ofgroundwater. The State Legislature respondedwith the 1948 Groundwater Code prohibiting expansion of agricultural acreage irrigated with groundwater in designated "critical groundwater areas." Three critical groundwater areas were established in the Tucson area, namely the Avra -Marana, Sahuarita- Continental, and Tucson critical groundwater areas. The 1948 Act was only a starting point for controlling groundwater overdraft, and a multitude of water -related court decisions confirmed the Code's inadequacy.

In 1953, the Arizona Supreme Court reversed its earlier decision in Bristor vs. Cheatham. The Court ruled that the doctrine of reasonable beneficial use applied with the constraint that groundwater could not be transported off the Land from which it was pumped if adjacent pumpers were damaged by the transportation. During the 1950's and 1960's the City of Tucson experienced remarkable growth in population and corresponding demand for groundwater. Foreseeing the potential problems associated with water table declines in the Tucson metropolitan area, the City applied for and received a federal grant for $1.4million in 1967 from the Department of Housing and Urban Development to construct a new wellfield and importation pipeline to bring in water from the Avra Valley west of Tucson. In 1969, the Arizona Supreme Court ruled in Jarvis I that the City was illegally transporting groundwater from one critical groundwater area to another, and the wellfield was shut down. The City appealed, and in 1970 the Supreme Court decided in Jarvis II that if the City purchased and retired irrigated farmland in the Avra -Marana Critical Groundwater Area it could transport into the City that amount of water historically used for irrigation.

In 1971 the City purchased the first of over 20 parcels ithas retired from farming in the Avra Valley. The Hill Farm, adjacent to Ryan Field, was purchased for$500 an acre. To date, Tucson has purchased approximately 12,000 acres of Land previously irrigated in Avra Valley at an average price of $800 per acre. In Jarvis III, the Arizona Supreme Court ruled in 1976 that the City couldonly transport the amount of water consumptively used on retired farmland. This amount was determined to be slightly more than half of the water previously pumped for agricultural use. In the FICO suit (1976) the Farmers Investment Company asked that the mining companies and the City of Tucson be restrained from pumping and transporting water out of the Sahuarita- Continental Critical Groundwater Area. The Supreme Court enjoined the transportation of groundwater by the mines to points outside the critical groundwater

89 area, but did not enjoin the City.

The threat of injunction against the City and the economic impacts upon the mines south of Tucson prompted the Legislature in 1977 to enact Senate Bill 1391. This Bill was an interim package which provided the following:

1. A mechanism for transferring groundwater between critical groundwater areas. 2. The right of an injured party to collect damages. 3. The prohibition of injunctions against groundwater pumping by the City and mines. 4. The establishment of a Groundwater Management StudyCommission and a timetable for reformation of the State's groundwater law.

After many drafts of a proposed new code by the Commission, review and approval by the Secretary of the Interior as a condition of Central Arizona Project completion and finalized allocations, and participation by the Governor in writing the final version, the Legislature adopted the 1980 Groundwater Management Act.

Major Elements of the 1980 Groundwater Management Act

The legislative intent of the 1980 Groundwater Management Act is stated in the Declaration of Policy, Section 45 -401. A portion ofthat statement is worth quoting herein. "It is, therefore, declared to bethe public policy of this state that in the interest of protectingand stabilizing the general economy and welfare of this state and its citizens it is necessary to conserve, protect, and allocate the use of groundwater resources of the state and to provide a framework for the comprehensive management and regulation of the withdrawal, transportation, use, conservation, and conveyance of rights to use the groundwater in this state."Primary elements of the Act affecting municipalities in general are discussed in this paper.More comprehensive summaries have been recently written by Pontius (1980) and Johnson (1981).

A key element of the Act was the establishment of a strong state agency calledthe Department of Water Resources with jurisdiction over the management of groundwater, surface water rights, dam safety, flood control, interstate streams, the Central Arizona Project, and, to a limited extent, water quality. The Director is given flexibility and broad powersto organize the Department and administer all laws relating to groundwater in the State. Four Active Management Areas (AMA's) were created by statute including the Tucson AMA, Phoenix AMA, Prescott AMA, and Pinal AMA. These initial AMA's account for 69 percent of the groundwater overdraft in Arizona and contain over 80 percent of the State's population. Local AMA Directors have been selected, and each is advised by a five- member Groundwater Users' Advisory Council appointed by the Governor representing major water -using groups within the AMA.

The goal of the Tucson AMA is to attain safe yield (the long -term balance of groundwater withdrawals and natural and artificial recharge) by January1, 2025. Mandatory conservation, supply augmentation, and purchase and retirementof agricultural Land are the three principal tools through which the Director must attain safe yield. These tools will be applied in varying degrees over the 45 year timeframe for achieving safe yield infive successive management periods. An initial management plan for the Tucson AMA is required under the Act by January 1, 1983, but staffing and data limitations will most likelydelay preparation by one year. During each of the management periods the Act states that "reductions in per capita use and such other conservation measures as may be appropriate for individual users" will be required of all municipal water users. During the second management period supply augmentation including artificial recharge is to be implemented by the Director. If these management tools are not successful by 2010, the end of the third management period, the Director must begin a program of purchase and retirement of irrigated land within the AMA.

Within an Active Management Area all existing legal uses of groundwater are "grandfathered" and may continue subject to conservation requirements imposed by successive management plans. The right to continue pumping groundwater is called an Irrigation Grandfathered Right, a Type 1 Non -Irrigation Grandfathered Right, or a Type 2 Non -Irrigation Grandfathered Right. Certificates of Grandfathered Rights willbe issued by the DWR indicating theannual amount of water to be withdrawn, location of withdrawal, and location of use. Groundwater uses authorizedunder Certificates of Exemption are specifically recognized as legal uses under the Act. A municipality is permittedto increase its pumping and freelytransport water within its service area defined as that area actually being served water for non -irrigation purposes. A city may notdrill new wells nor increase its groundwater pumping outside its service area unless Grandfathered Rights are acquired. A city may extend its service area to serve new development, thereby extending the area from which groundwater may be pumped.

Other groundwater management toolsavailable to the Department of Water Resources include the issuance of Certificates of Assured Water Supply, well construction and spacing regulations, metering of pumpage, groundwater transportation regulation, groundwaterwithdrawal fees, and specific enforcement provisions. Inside an AMA a Certificate of Assured Water Supply is required from the Director prior to sale of subdivided or unsubdivided land. An assured water supply means that "sufficient groundwater or surface water of adequate quality will be continuously availableto satisfy the water needs of the proposed use for at leastone hundred years, the projected water use is consistent with the management plan and achievement of the management goal for the active management area, and the financial capability

90 has been demonstrated to construct the delivery system and anytreatment works necessary to make the supply of water available for the proposed use." The service area of the City of Tucson has been deemed to have an assured water supply on the basis of its filed letter of intent to purchase CAP water.

Tucson Active Management Area Baseline Data

Preliminary to discussion of specific impacts of the Act on Tucson Water, it is salient to understand the regional water supply- demand environment of whichTucson is a part. The Tucson Active Management Area contains a wide diversity of water -using economies, each solely dependent upon groundwater for its water needs. Major water pumpers include the City of Tucson, four mining companies south of the City, Farmers Investment Company (FICO), Cortaro -Marana Irrigation District,Avra Valley Irrigation District, private water companies, andprivate well owners. Staff of the Tucson Active Management Area have recently prepared a baseline estimate and projection of existing and anticipated water uses and supplies from 1980 through 2025. That data is summarized in Table 1 which readily indicates the basic problem of the Tucson AMA: groundwater mining. Much of the tabulateddata was provided by individual water users. The baseline projections for 2000 and 2025 assume no management of groundwater on the part of the Tucson AMA, but do incorporate specific assumptions with regard to populationprojections, irrigated acreage, incidental recharge, Central Arizona Project water availability and usage, andwastewater effluent reuse. These assumptions cannot be realized, however, without management of groundwater inasmuch as the management provides the incentive touse Central Arizona Project water and wastewater effluent.

The population projections used as the basis forpredicting municipal water use are 1979 Arizona Department of Economic Security figures and have recently been updated and increased by that agency for Pima County. Larger population projections result in larger municipal water use and an understatement of the potential overdraft (groundwater mining) problem. Tucson Water is assumed to serve 80 percent of the Tucson Active Management Area,and a per capita water use factor of 160 gallons per day is used to derive municipal water use. Industrial uses include golf courses, power plants, parks, cemeteries, hospitals, schools, sand and gravel industries, dairies, Davis -Monthan AFB, and the University of Arizona. Water use by copper mines in 1980 was established usingdata provided in applications for Certificates of Exemption. Agricultural use is the sum of non -Indian water application at 4.45 acre - feet per acre and Indian application at 5.4 acre -feet per acre.

Municipal incidental recharge is assumed to be 70 percent of the wastewater effluent discharged into the Santa Cruz riverbed. Recharge from mining is assumed to be 20 percent of pumpage except where interceptor wells are operated. Agricultural recharge is assumed to be 20 percent of the total water applied. Central Arizona Project water is the total of DWR staff allocations to municipal, mining, and Indian users based upon a statewide availability of 640,000 acre -feet per year for municipal and industrial users and 160,000 acre -feet per year for Indian users. In 1980, 3,400 acre -feet of wastewater was used by golf courses, and 5,200 acre -feet was used by agriculture. In 1980, approximately 395,000 acre -feet of groundwater and wastewater effluentwas used in the Upper Santa Cruz and Avra Valley subbasins of the Tucson AMA. Analysisof Table 1data indicates that 53.7 percent of the total 1980 water use was agricultural, 23.5 percent was municipal, 15.2 percent was mining, and 7.6 percent was industrial. Groundwater pumpage exceeds natural recharge by a factor of five to one in the Tucson AMA. Consumptive use of water exceeds natural recharge by a factor of four to one. The large imbalance between groundwater supply and demand has resulted in water table declines of over 175 feet in the Avra Valley since 1952 andnearly the same amount along the Santa Cruz riverbed southof Tucson since 1947.

Four major impacts of groundwatertable declines are increased cost for pumping, decreased well capacity due to higher consolidation of aquifer sands and gravels, reduced water quality, and potential Land surface subsidence. For these reasons it is prudent that the Tucson AMA develop a workable plan to reduce overdraft. Under the no groundwater management scenario preparedby the Tucson AMA staff, overdraft is projected to decline from 238,000 acre -feet in 1980 to 87,000 acre -feet in 2025. Key assumptions include full utilization of wastewater effluent by 1990, some gradual conversion of private agricultural land to urban uses, new irrigation of 5,000 acres of San Xavier and 2,000 acres of Schuk Toak Papago Indian Land beginning 1990, and full utilization ofCentral Arizona Project water by requesting entities in the amount allocated by DWR. The no management scenario indicated in Table 1is not the worst case, however, in that projected water uses may be understated, and water supplies utilized from surface sources may be overstated. Tucson AMA staff are preparing additional no management" scenarios which will indicate the full range of overdraft possibilities given other assumptions about wateruses and supplies. Given the severity of the impacts of overdraft the Tucson Groundwater Users Advisory Council is evaluating the possibilityand impacts of moving forward the safe yield date from 2025 to 2005.

Beneficial Impacts of the Groundwater Act

Tucson Water is a municipally owned and operated water utility providing domestic water service both inside and outside the corporate limits of Tucson. Water is provided to a service area of over 300 square miles having an elevation difference of over 1,500 feet. Presently, Tucson Water serves 440,000 people through 126,000 active metered services. This represents 80 percent of the Pima County

91 population and over 85 percent of the metropolitan Tucson population. In 1980 -81, Tucson Water actively pumped 189 wells from four wellfields serving the central metropolitan system and 43 wells serving fifteen isolated systems. Sixty percent of the 78,000 acre -feet pumped by Tucson Water in 1980 -81 was from the central, interior wellfield and seventeen percent was from the Avra Valley. The remainder was pumped from the Southside and Santa Cruz wellfields near the riverbed south of Tucson.

Tucson Water uses population projections officially adopted by the Arizona Department of Economic Security and the Pima Association of Governments for Pima County. The December 1981 figures indicate a County population of 536,100 in 1980 and a 2035 projection of 1,807,400. The associated Tucson Water population and water use projections are shown in Table 2. Water use formunicipal purposes is projected to more than quadruple from an average of 68.6 milliongallons per day in 1980 to 231.3 million gallons per day in 2035 based upon an assumed average daily per capita use of 160 gallons, the current use by Tucson Water customers.

To meet the projected water demands of existing and future customers, Tucson Water will have three primary supplies: groundwater, Central Arizona Project water, and wastewater effluent. Each of these sources is impacted by the passage of the 1980 Groundwater Management Act. Our solepotable source currently, groundwater, will be subjected to increasingly more stringent requirements for conservation, although our dependence upon it will increase until delivery of Central Arizona Project water. Figure 1 is a graph of projectedannual water need assuming 160 gallons per person per day and the expected combination of groundwater and CAP water required to meet that need. The graph indicates that, although a reduction in municipal groundwater pumpage occurs upon acceptance of CAP water, the growth in demand and limitation of Central Arizona Project water allocation to Tucson by the Department of Water Resources (151,000 acre -feet in 2034) will require a gradual increase in groundwater pumping by the City to a level in 2034 equal to the pre -CAPpumping level. Municipal water conservation discussed in the following section will impact the demand for groundwater, but previous successes of theUtility will limit its effectiveness.

The importation of Central Arizona Project water to theTucson AMA and the use of same by Tucson Water has both beneficial and adverse impacts. The immediate positive effect of delivery (assumed to be about 1990) will be a one -for -one reduction in groundwater pumping, thus furthering the goal of groundwater preservation. As Figure 1 indicates, Tucson Water expects to accept and treat 30,000 acre - feet in the first year of delivery. While this amount will provide only thirty percent of our total need in 1990, Tucson Water management will develop expertise in operating the area's first municipal water treatment plant, and customers won't be shocked with a major waterquality change through increased total dissolved solids. By 2035, sixty -two percent of Tucson Water's anticipated need will be met by Central Arizona Project water.

Since the Act's goal is conservation of groundwater pumping, Tucson Water is induced to maximize reuse of its municipal wastewater effluent. In response to a perceived void, the Tucson Water Effluent Reuse Plan (1982) was prepared in draft. This document primarily addresses reuse in the long -term incorporating the Department of Water Resources assumption that Pima County mines will ultimately use a blend of fiftypercent effluent and fifty percent CAP water and that 28,200 acre -feet of effluent per year will be provided to the Papago Indian Tribe in accordance with the Southern Arizona Water Rights Settlement Act of 1982. Additionally, effluent delivery systems were identified and cost- estimated from both the Ina Road and Roger Road WastewaterTreatment Plants to serveexisting and projected golf courses and cemeteries. Subsequent to report preparation, proposed contracts to purchase effluent for agricultural usedownstream of the Ina Road plant were submitted to the City. The Tucson Mayor and Council referred thecontracts and draft effluent reuse report to the Tucson Citizens' Water Advisory Committee for their review and policy recommendation.

The Committee initially evaluated thetotal spectrum of potential effluent users over time and, with Tucson Water staff, developed delivery scenarios indicated in Figure 2. Near -term users include downstream agriculture and metropolitan Tucson golf courses. Mid -term users include the San Xavier Indian Reservation and a Santa Cruz River recharge system. Long -term users require the highest expenditure of funds for delivery systems andpumping costs and include Pima County mines south of Tucson. The Committee has recently restricted its effluent reuse policy considerations to the near -term on the basis of proposals by BIN farms, Incorporated in the Avra Valley and the Cortaro -Marana Irrigation District. Agricultural users can blend the secondary- treated effluent with groundwater to meet quality standards adopted by the Arizona Department of Health Services. The addition of a properly- designed filtration plant downstream of the existing secondary waste treatmentprocess will permit the use of effluent on parks and playgrounds where children play and preclude the requirement to fence golf courses. Tucson Water is presentlyevaluating potential users of advanced waste treatment effluent for inclusion in a second draft of the Tucson Water Effluent Reuse Plan. In conjunction with the Pima County Wastewater Management Department, Tucson Water staff is evaluating theeconomics of constructing subregional waste treatment facilities closer to potential effluent users.

The passage of the Act has provided thenecessary impetus forgroundwater users to foresee the value of converting to wastewater effluent as a reliable water resource. Effluent could prove to be very economical for a varietyof uses in light of the projected costs of Central Arizona Project water and limits imposed ongroundwater pumping. The marketability of effluent by Tucson Water has been hindered in the past by the lack of adequate groundwater regulation and restriction. The new law has prompted increased awareness in the need to maximize use of wastewater effluent and has encouraged the

92 City to adopt generalized policies by which proposed contracts can be evaluated.

Another positive impact ofthe Act on Tucson Water will be improved regional data bases resulting from water rights determinations, well registration, and metering of groundwater pumpage. Heretofore, the Tucson Water Department has been the Leader in basinwide groundwater management and associated data collection activities. The state has now assumed this role through the provisions of the new law. The gradual reduction in overdraft required by forthcoming Tucson AMA groundwater management plans will mitigate costly effects of water table declines,directly affecting the longevity and capability of Tucson's wells. The establishment of the local Tucson AMA office and the employment of competent staff has already provided a renewed public awareness of the groundwater management problem, the importance of water conservation,the need for Central Arizona Project water, and the necessity touse wastewater effluent.

Adverse Impacts of the Groundwater Act

Since the primary tool of the Act is mandatory conservation, Tucson is concerned about the degree to which Tucson Water customers can further reduce water usage beyond that which has already been attained (Davis, 1978). The implementation of an increasing block cost -of- service water rate structure, a winter- summer rate differential, the "Beat- the -Peak "summer demand management program, and public education through the news media have collectively reduced the Tucson Waterper capita use from 205 gallons per day in 1973 -74 to 159 gallons per day in 1980 -81. The lowest use wasobtained in 1978 -79 when Tucson Water's daily per capita use was 147 gallons. In planning for future water facilities staff has assumed 140 gallons per person per day using 50 -year sizing criteria. Since the per capita wastewater flow has stabilized at 90 gallons per day, this lower figure would appear to be the minimum obtainable assuming that all outdoor water uses were eliminated. Potential tools to effectuate additional municipal groundwater conservation include higher water rates, plumbing code regulations, Land use /zoning restrictions on large water uses, and mandated effluent reuse where practical. Forthcoming objectives in groundwatermanagement plans will dictate the timing and extent of municipal conservation.

A potential major impact adversely affecting the Tucson Water Utility is the requirement to accept and treat more Central Arizona Project water than is currently planned for initial years of delivery. Restrictions placed upon municipal groundwater pumping could force the City into this resource usage mode and accelerate those impacts of Tucson's source transition discussed previously by McLean and Davis (1981). The most significant impacts include capital repayment and operation and maintenance costs assessed bythe Central Arizona Water Conservation District (collectively estimated to be $85.00 per acre -foot for municipal water), construction and operation costs associated with a CAP water treatment plant, capital costs associated with rebuilding the Tucson Watertransmission system to distribute a single, large source of water in conjunction with the well supply points, and water quality degradation associated with increased total dissolved solids, hardness, and chlorination. Economic impacts of using Central Arizona Project water in the municipal system are already projected to be significant in terms of increased water rates and assessment of water development fees to newly -developing areas. Increased CAP use will result in even higher rates and considerable water quality degradation.

Additional cost impacts of the Act include the assessment of groundwater withdrawal fees by the DWR and fees for grandfathered rights applications and well registration. Withdrawal fees consist of three components: not more than one dollar per acre -foot pumped for administrative costs, not more than two dollars per acre -foot for water supply augmentation (beginning no sooner than 1984), and not more than two dollars per acre -foot for purchase and permanent retirement of irrigated landbeginning no sooner than 2006. The Tucson Water Department has paid $20.00 each for 248 grandfathered rights applications and $10.00 each for 489 well registration applications. Regulatory impacts include the requirement to file annual pumpage reports, adherence to well construction standards, the requirement to obtain new well drilling permits, and licensing of well drillers.

The ultimate impact of the Groundwater Management Act on Tucson Water will be limitation of growth. For some this result is beneficial rather than adverse. This issue is discussed in more detail by McLean (1982). The total population which can be supported by the projected potable water sources depends upon the quantities ultimately available on a long -term reliable basis and the per capita water use. If the combination of Central Arizona Project water and groundwater availableto the Tucson area is ultimately 200,000 acre -feet per year, the amount of people which canbe supported is 1,785,000 assuming 100 gallons per person per day average water use and 1,116,000 assuming 160 gallons per person per day. Both of these values are lower than the recently- adopted 2035 projection of 1,807,400 for Pima County. The degree to which Tucson area growth will be limited by the lack of water will depend upon the effectiveness ofthe successive Tucson AMA groundwater management plan conservation programs, the amount of reliableCentral Arizona Project water available to the area, and the amount of groundwater which legally can be pumped by municipal users.

A growth- related issue is the determination of Tucson's assured water supply service area via regulations to be promulgated by the DWR. Concepts recentlydistributed by the DWR include the provision that Central Arizona Project allocations for 2034 would determine the extent of the assured water supply service area in which developers would not be required to obtain a Certificate of Assured Water Supply if they received water from Tucson. This projected boundary is shown in Figure 3 along

93 with the existing corporate limits and the area actuallybeing served water by the Utility in 1980. Impacts of this proposal aredifficult to predetermine in that a number of questions still remain with respect to service area extensions and well drilling. The existence of discontiguous, isolated systems both within and without the projected water supply boundary complicate any assessment at this time.

Conclusion

While the overall intent of the 1980 Groundwater Management Act is clear with respect to safe yield within the Tucson Active Management Area, more definitive impacts of the Act on Tucson Water will depend upon specific groundwater management plans to be promulgated in the future. This paper has attempted to assess the major potential impacts that will both benefit and adversely affect the TucsonWater management and rate -payers. The implementation of the Act is in its infancy stage, and only through better definition of time -specific municipal conservation goals and development of municipal service area regulations can amore thorough assessment be accomplished. Tucson Water is highly supportive of groundwater management and will continue to coordinate with the Tucson Active Management Area and the Department of Water Resources to assure customers of the most reliable, cost- effective water service to meet domestic needs.

References Cited

Pontius, Dale E., "Groundwater Management in Arizona: A New Set of Rules," Arizona Bar Journal, October 1980.

Johnson, James W., "Summaryof the 1980 Arizona GroundwaterManagement Act," State Bar of Arizona, Continuing Legal Education paper, 1981.

Tucson Water, "Tucson Water Effluent Reuse Plan (draft)," January 1982.

Davis, Stephen E., "Tucson's Tools for Demand Management," Hydrology and Water Resources in Arizona and the Southwest, Volume 8, pp. 9 -15, 1978.

McLean, ThomasM. and Stephen E. Davis, "The Alternatives and Impacts Associated With a Future Water Source Transition for Tucson Water," Hydrology and Water Resources in Arizona and the Southwest, Volume 11, 1981.

McLean, ThomasM., "Water Resources as a Limiting Factor to Tucson's Growth," Hydrology and Water Resources in Arizona and the Southwest, Volume 12, 1982.

94 TABLE1

TUCSON ACTIVE MANAGEMENT AREA WATER USES AND POTENTIAL SUPPLIES

1980 2000 2025

POPULATION 518,200 828,500 1,225,200 IRRIGATED ACREAGE 47,300 43,000 31,500

TOTAL WATER USE (Acre -Feet) Municipal 93,000 149,000 220,000 Industrial 30,000 53,000 83,000 Mining 60,000 69,000 69,000 Agricultural 212,000 198,000 147,000 TOTAL 395,000 469,000 519,000

CONSUMPTIVE USE Municipal 60,000 78,000 115,000 Industrial 30,000 53,000 83,000 Mining 53,000 61,000 61,000 Agricultural 170,000 158,000 118,000 TOTAL 313,000 350,000 377,000

INCIDENTIAL RECHARGE Municipal 24,000 0 0 Industrial 0 0 0 Mining 7,000 8,000 8,000 Agricultural 42,000 40,000 29,000 TOTAL 73,000 48,000 37,000

POTENTIAL WATER SUPPLIES

Central Arizona Project 0 134,000 215,000 Reused Wastewater 9,000 71,000 105,000 Natural Recharge 75,000 75,000 75,000 Incidental Recharge 73,000 48,000 37,000 TOTAL 157,000 328,000 432,000

OVERDRAFT (MINED GROUNDWATER) 238,000 141,000 87,000

From the Arizona Department of Water Resources Tucson Active Management Area Staff

95 TABLE 2

TUCSON WATER POPULATION AND WATER USE PROTECTIONS

Pima Countyl Tucson Water Year Population Service Population2 Water Use (MGD)3

1980 536,100 428,880 68.6 1985 620,000 496,000 79.4 1990 710,100 568,080 90.9 1995 810,700 648,560 103.8 2000 921,900 737,520 118.0 2005 1,048,400 838,720 134.2 2025 1,554,400 1,243,520 199.0 2035 1,807,400 1,445,920 231.3

1From Arizona Department of Economic Security, December 1981.

2Pina County population times eighty percent.

3Tucson Water service population times 160 gallonsper person per day.

PROJECTED TUCSON WATER NEEDS AND SUPPLIES PER CAPITA USE= 160 GALLONS /DAY 300,000

260,000

200,000 Eß 0-

160,000

GROUNDWATER

100,000

50,000 CENTRAL ARIZONA PROJECT WATER

o 1980 1986 1990 1996 2000 2005 2010 2015 2020 2025 2030 2035 YEAR

FIGURE1

96 EFFLUENT DELIVERY SCENARIOS

- !

N T 10 S CORTARO- MARANA IRRG. DIST. I R 8 E I -. T r,. II .. ,N r`- ..1 i : Upper AVRA VALLEY IRRG. DIST. 11:`' ,, :', . Somrn.rha.en -- ., __._: s= 11.:a its L k. Saverlé !- '2.,... 'i_, ` S _, -1.4-r) . 1

i 1 } V f 1 T i / s '_ 13 rr Oue.ns Well S _ ,...... s .

/ T 14 S LEGEND

T NEAR TERM Sama Rosa 15 MID TERM Ranch S LONG TERM Son Pedro

T i 17 S

eT S

>_R.., Fr.Mwl BASE Canyon OF TR, T-.. RNAL .. _. PIMA 19 enl.nill. \ SANTA S Anvaca . lonclion Sourh Komelik Mad.ra Canyon T Cold held 20 S .C.

m 1 $an M,yu.l T6 21 S

I CORON- 12?r_

.` Oroy BlancRONAOOo i_ , NATIONAL . FOREST R,,br `, Ra!'

FIGURE 2

97 -1144- ö:R - R -11-É r R13-E' 7j--- 1F15-í -FT RB-E _.-( R-17EL- r 10E R-11í Ìr TOWN Of MALIN H ILIIANCO MARAM1 ..+ COMOwA00 jj] SI1J411ELLNFST NCEIIJ

IELVEI NNIiE ins TH""" NO WALLEI ARENAL .4 SALLIOOI G/.. ITIRNIIT Ñ s0Aes1

1. C011M1W0 WOW POMP

lams LA CIFMCA \ It

\\\`'\\ F,.E1E11LEIIIIIIoi1. , A.. ..\\\', I.N ,m®®EI1!!IÉi\\,, NN\ t.ssty_,.\` 1rar'!`:\ ti iiiirkt/\

ROWUroMN Au«cr `,\ \ TALLITVIIIACM \ \ \ \\ CO10NA IETONTON. 'CD \\\' \liF WfirONI,

ASSIED WATER SUPPLY SERVICE AREA BOUNDARY

R-17_E R-9-E R-11-E I R-10-E I R-12E'°'v"1 R13_E ,.... . ' CITY V TUCSON LEGEND: % TUCSON\ WATER ¡3EIfltE IIEA SE11E111 WIIEIW1I11 #I. IFN 1.i9 'ISOEATEI SEIIICE UFA 0 WATERMAINSERVICE AREA

FIGURE 3

98 WATER RESOURCES - THE PRIMARY FACTOR IN TUCSON'S FUTURE GROWTH

Thomas M. McLean, P.E. Tucson Water P.O. Box 27210 Tucson, Arizona 85726

Abstract

The community of Tucsonfaces a tremendous future challenge regarding the management of its local water resources. With the advent of thenew Groundwater Code and a plan to balance thebasin by the year 2025, it is impossible to discuss the growth of the metropolitan area without first questioning the availability of adequate water resources. In Tucson, water will soon become the yardstick by which community expansion will be measured.

The Tucson Water Utility plays a significant role in the management of the local water resource. Although there is currently complete reliance on groundwater, Tucson has received a tentative allocation of Colorado River water by means of the Central Arizona Project to supplement the groundwater supply in the future. In addition, the reuse of wastewatereffluent and further conservation effortsmust be planned in orderto accommodate growth. The key ingredient to regional resource management, however, involves the cooperation that must exist among the major water -using entities of the area: Tucson Water, the mines, farmers, private water companies, and private well owners. This paper addresses the potential favorable and unfavorable impacts of limited water resources on future growth with respect to these concerns.

Introduction

The metropolitan community of Tucson, situated in eastern Pima County (see Figure 1), is experiencing growing pains of a slightly different nature than one would normally expect. These pains come from the realizations that future growth carries with it a high price tag and that the major water - using entities of the region look forward to some difficult water resource -related decisions in the coming years.

A brief overview of the currentwater supply situation in eastern Pima County isappropriate at this point. The Tucson Water Utility is the major purveyor of water for municipal and industrial users in the metropolitan area of Tucson, but it is only one of many local water -using entities all of which subsist entirely on groundwater for their source of supply.There are four copper mining companies in the near vicinity; several agricultural interests suchas the Cortaro -Marana Irrigation District, the Avra Valley Irrigation District and the Farmers Investment Company; a number of private water companies; and various private well owners. Agricultural interests utilize the largest amount of water consumptively followed bythe mines and then the City Water Utility (see Figure 2). Pumpage in this area is estimated to exceed natural recharge by a factor of approximately five to one, and, as a result, there exists a large imbalance betweengroundwater supply and demand which has resulted in long -term declines in local groundwater levels.

Projected Growth of the Tucson Metropolitan Area

The Tucson Water Utility presently provides service to approximately 450,000 people through 130,000 service connections. As the major metropolitan area water utility, Tucson Water's interaction with regional planning efforts can and does provide an important element in the process of formulating projected urban growth patterns. The Water Utility is represented on an intergovernmental planning team along with other governmental agencies and utilities. This group coordinates the necessary technical expertise which contributes to the process of developing regional land use plans. Additionally, local governmental agencies (including Tucson Water) play a significant role in outlining the State of Arizona's regional population projections administered by the Department of Economic Security.

A great deal of development pressure exists in the Tucson vicinity. It is pushing to the peripheral areas of thecommunity in search of large portions of undeveloped land. Urban growth, and, therefore, intensive Land use planning is currentlyfocused on threemajor regions: the northwest, southwest and southeast (see Figure 3). Expansion to these areas must be accompanied by a number of basic support services, and never before has theavailability of water supply been such an integral

99 factor in the decision- making process as it is today. Evidence of this situation is demonstrated by the fact that coordination with land use planning efforts has resulted in the establishment of area -specific water plans (McLean, 1980). These water plans are aimed directly at resource management, but more essentially provide technical and financial solutions to the problem of water delivery to specific areas of the community. The water plans address the need for short and long -term water supply concepts which provide an overview for future residential and associated water system expansion. A financial mechanism is also established in an effort toprovide funds for the capital improvements required to support growth in that particular area as it occurs.

Recently, land use plans for new major urban areas in the southeast and northwest portions of the metropolitan area have been under consideration for adoption by the local governing body. It is apparent that there exists political indecisiveness regarding proposed land use densities dueto the long periods of deliberation required before the plan is adopted. These delays appear to reflect the difficulty with which officials must decide regarding the location and extent of urban expansion and the capability of supporting such growth with the available and projected services.

The propensityfor growth in the Tucson area isnot only displayed by the activity of Land use planning, but is also reflected in the statewide population projections produced by the Arizona Department of Economic Security (DES). In 1979, the DES projected that Pima County would have a population of 818,600 people bythe year 2000 and 1,269,000 people in the year 2030. In 1982, the DES revised these projectionsand now estimates that Pima County will have a population of921,938 in the year 2000 and 1,680,560 by the year 2030 (see Figure 4). Five years beyond that, in 2035, it is predicted the population will increase further to 1,867,400.

Future Water Resources Available to Meet Tucson's Growth Needs

Responsible land use planning together with a reasonable projection of future population distribution obviously provide the key ingredients for planning the community's water system expansion. Major additions to Tucson's watersystem are proposed in these same areas of intense development pressure. However, thequestion that must logically come first can be simply stated --will there be adequate water resources to supply the projected demand? Projected growth translates directly into additional water supply requirements, and the alternativesavailable to this arid region can be enumerated as we see them today.

The continued use ofgroundwater to meet local requirements is envisioned to some extent, but the limitations on the availability of this resource are extremely difficult to quantify due to a number of extenuating circumstances.

1) A New State Groundwater Law - In June of 1980, the State of Arizona passed into law a Groundwater Management Act which sets forth ambitious goals for the regulation of groundwater use within the State. With its eyeon balancing demand with groundwater supply within the Tucson Active Management Area by the year 2025, the State Department of Water Resources proposes to formulate andadminister a series of groundwater management plans designedto impose various conservation regulations for all water using- entities. New subdivisions are required to prove an assured water supply for a period of 100 years.

2) Potential for Land Surface Subsidence - As previously mentioned, a serious overdraft condition exists in eastern Pima County. Historical declines over a 35 -year record in some areas of Tucson's central basin (where the municipalityof Tucson resides) are in excess of 120 feet. Continued long -term pumping from this arearepresents a concern for Utility officials due to the potential for land surface subsidence (particularly differential) and its associated impacts.

3) Legal Limitations - The value of groundwater resources in the Tucson area and the intense competition for rights to its use are exemplified by the current Legal confrontation with the Papago Indian Tribe. Due to the President's veto of the Southern Arizona Water Rights Settlement Act of 1982,negotiations with the Tribe and the Federal Government have been reopened. If the lawsuit persists, injunctions could be issued by the court requiring cessation of pumping in both the City's Avra Valley and Santa Cruz wellfields.

4) Groundwater Contamination - The community of Tucson faces another challenge to its current water supply from Trichloroethylene (TCE) contamination. The historical disposal of industrial waste in thesouthwestern parts of the city has resulted in the contamination of small portions of the aquifer. As a result, a number of public and private wells have had to be shut down until measures can be taken to mitigate the problem.

Facing a relatively complexdisposition of its futuregroundwater supply, Tucson Looks to the importation of a surface water supply as a potential remedy to its water source problems. Tucson has a tentative allocation ofCentral Arizona Project (CAP) water which is proposed to bedelivered by 1990 through the Tucson Aqueduct. The CAP in itself is not a panacea, however. As part of the Central Arizona Project Act of 1968, Arizona is obligated to curtail pumping from the Colorado River during periods of low flow asmeasured by water elevations in reservoirs along the river. The project itself

100a is behind in its original schedule of construction and is subject to constant scrutiny of its projected economic benefits. As can be expected, statewide competition is keen for long -term allocations of this water and will continue as such until the federal government finalizes CAP allocations and sales contracts are signed. The final costs of CAP water are not defined. The Tucson community anticipates that the purchase of CAP water will represent a fairly sizeable impact on the rate payers. Planners and administrators must carefullyassess these economic factors before long -term community commitments can be extended. Lastly, the quality impacts of this new future source are yet another question mark. Tucson, accustomed to relatively high quality, low TDS groundwater must now re- orient its supply system configuration to operate with its first watertreatment facility. While this plant will remove suspended solids and taste and odor problems associated with surface water, hardness and total dissolved solids will increase.

In an effort to maximize the availability of its potable water supplies, Tucson plans to maximize the use of its wastewater effluent available at two secondary treatment plants in the northwest portion of the Tucson metropolitan area. Studies are underway to ascertain the most efficient and practical future use of effluent to replace potable water use wherever possible. Examples of this are already realized throughthe application of effluent on a number of local golf courses. In this manner an extension to the life of local groundwater reserves can be accomplished.

A final option for augmenting Tucson's water resources lies with increased conservation efforts. Since 1977, Tucson Water has successfully advocated a summer demand management program known as "Beat the Peak." A byproduct of this voluntary program has, in fact, been the realization of a significant amount of water conservation. Per capita usage was reduced from 205 gallons per person per day (gpcpd) in 1973 -74 to approximately 150 gpcpd in 1978 -79. This resulted in not only Less pumpage overall, but also a reduction in the expenditure of capital fundsdue to decreased demand. In effect, the life of some water facilities was extended. Since "Beat the Peak" is a summertime campaign, other conservation programs have followed such as "Slow the Flow ". This program is designed to take advantage of the successful summertime effort and broaden the public's awareness of the need for intelligent indoor water conservation to reduce sewage volume and extend the life of wastewater treatment and conveyance facilities. Once again it is noted that these efforts are voluntary, and increasedemphasis in these areas in the future may positively impact the water resourcesituation locally. As previously mentioned, the new groundwater code will bring more attention to water conservation through guidelines that are mandatoryin nature which stem from an overall management plan withtime -specific per capita use goals.

The Impacts of Limited Water Resources on Tucson's Growth

At this point in thediscussion it is pertinent to pause and attempt to put the aforementioned information in perspective through the use of a rudimentary example. Estimates have beenmade that approximately 100,000 acre -feet (a -f) of natural recharge is available annually to the local water table. Assume that the basinis in balance in the year 2030 and all of this recharge is available for use by the municipalityof Tucson (a sizeable assumption in itself). Further assume a reasonable expectation of CAP delivery for that year of 100,000 a -f. If a per capita usage characteristic of 150 gpcpd is applied to this situation, the total population which can feasibly be supported by these resources is 1,200,000 people (see Figure 5). At this point in time (year 2030) it is assumed that Tucson Water is providing retail or wholesale service to essentially all of eastern Pima County and only a very small amount of population exists outside the service area. As you may recall, it was previously indicated in this paper that the most recent planning efforts indicate a population projection for Pima County of 1,680,560 people by the year 2030. We are realizing, therefore, a potential disparity between what is projected as growth for the Tucson area and what can be reasonably accommodated. Since the maximum supply side of the equation is relatively fixed, the demand side must be reducedto match the supply.

The issues surrounding the future of water supply for the southwestern community of Tucson paint a complicated and undefined picture, to say the Least. The governing bodies and the citizens at large must possess a keen awareness of the water resource question at the earliest possible stage in order to adequately address these issues. Prudent and responsibleplanning of the community's growth must carry a high priority with the realization that difficult tradeoffs lie ahead and choices must be made for the benefit of the majority. Competition for local water resources will only increase inthe future, and all areas of the local economy will feel the impact. The agribusiness, being the highest consumptive user, will probably reduce. A limitation on municipal and industrial growth can be reasonably expected.

On the other side ofthe coin the impacts of limited water resources can be viewed favorably. As previously stated, it will encourage intelligent and prudent Land use planning in order to efficiently accommodate the community's needs with the resources available. Since scarcity encourages efficiency, the integrity of local water delivery systems will undoubtedly improve. Similarly, the pricing of the commodity of water will be adjusted upward to a more efficient and proper level. Last, but not least, it is anticipated that the region will cooperatively resolve theoverdraft situation and eventually realize the achievement of a "Safe- Yield" condition.

100b Conclusion The availability ofwater resources is an extremelyimportant factor in the growthof the Tucson area. It isno longerpossibleto discussthe expansionofthe metropolitanarea withoutfirst questioning theavailability ofadequate waterresources.In Tucson,the natureand extentof the community's growthwill be a measureof the ability ofthe local water -using entitiesto effectively manage and conserve the local water resources.

References Cited

McLean, Thomas M. The Northwest Area Water Plan- Tucson, Arizona ", Hydrology andWater Resources in Arizona and the Southwest, Volume 10.

100c PIMA COUNTY

FIGURE 1

1980 WATER USE IN EASTERN PIMA COUNTY

USE CATEGORY POMPAGE CONSUMPTION

QUANTITY CENT QUANTITY PERCENT IACRE FEET I OF TOTAL IACRACRE FEET I OFD

AGRICULTURE 224.953 AF 57.6 151.857 AF 56.7

MUNICIPAL 02.328 21.1 32.841 12.2

INDUSTRIAL 80,144 20.5 80,144 29.9

RECREATIONAL/OTHER 3,282 0.8 3.282 1.2

TOTAL 390.707 AF 100.0 208,124 AF 100.0

FROM WRCC REPORT TO CECIL D. ANDRUS.SECRETARY OF THE INTERIOR, APRIL 14, 1980

FIGURE 2

100d (1)

UMMERHAVEN

100e PIMA COUNTY POPULATION PROJECTIONS (DEPARTMENT OF ECONOMIC SECURITY)

YEAR

2000 2030

DES-1979 818,600 1,269,000

DES-1981 921.938 1,680,560

FIGURE 4

WATER SERVICE CAPABILITY AT VARIOUS PER CAPITA USE RATES I WATER SUPPLY-100,000 ACRE-FEET PER YEAR I

160 150 - 140

130

120

110

100

2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1

SERVICE POPULATION I X 1,000,0001 FIGURE 6

100f 100g A Survey and Evaluation of Urban Water Conservation Programs in Arizona (Abstract)

Marc Bennett Arizona Department of Health Services Phoenix, AZ

Abstract

Numerous cities and towns in Arizona are in the process of developing and implementing water conservation programs aimed at reducing wasteful water use activities of its citizens.Most of these programs have been initiated because of:

1) future water conservation mandates in the State's groundwater law and in contracts for Central Arizona Project (CAP) water; and

2) increasing economic pressures for delivering water because of rising costs and reduced future dependable supplies.

This paper will review these legal requirements and assess the economic pressures on cities and towns in Arizona for water conservation programs. In addition, existing and future urban water conservation programs being developed in the State will be examined. The ability of these programs to reduce water demand will be evaluated based on experiences in Arizona and other states.

101 An Application of the Almon Polynomial Lag to Residential Water Price Analysis

Donald E. Agthe Economic Consultant 1321 East Elm Street Tucson, Arizona 85719

A recent article by Agthe and Billings (1980) concerned itself with the application of Koyck form lagged dynamic economic models to measure the price elasticity of demand for residential water consumption. The Koyck form lagged models have a lagged dependent variable, quantity of water consumed, and assume that the greatest adjustment of quantity to price occurs in the first period after the price change and succeeding adjustments are progressively smaller and less important. This study will consider an alternative dynamic economic model in which the independent variable price is assigned a distributed lag. The distributed lag model chosen is the Almon Poly- nomial lagged model (Almon,1965) as it is expected that the lagged response to a price change is nonlinear in nature. The results of several potential Almon lag forms are compared to the results of the simple Koyck lagged model found in the Agthe and Bill- ings (1980) study. This Koyck form was by far the best result found in the earlier study.

One of the most intriguing advantages of the Almon polynomial lagged model over the Koyck model is the ability to specify the form and the length of the adjustment to the price changes. Given the fairly long periods of stable price structures for water and similarly priced public utility type goods, the consumer has a long time period to adjust and readjust his position to given price structure in the market place. There- fore, initial overadjustment to price changes with later readjustments is a viable hypothesis for consumer behavior. This behavior can be accounted for in the polynomial lag model but not in the Koyck lag model. The Almon process also reduces the possibility of autocorrelation that is frequently associated with the lagged dependent variable that appears in the Koyck model.

The Models and Variables

This study utilizes the general form of the Almon lag model (Gujarati, 1978) as follows:

Q = bo +b1W +b2Y +b3D

+ b4 Pt + b5Pt -1 + + b1Pt and the Koyck model in the form:

Q = bo + b1Qt + b2W + b3Y + b4D + b5Pt -1 where:

Q = Average water consumption (100 cubic feet) per household for active single residential water connections by month (January 1974 thru November 1977) (Tucson, 1978)

D = The difference between what the typical consumer actually pays for water and what would be paid if all of the water were purchased at the marginal rate. (Dollars)

Y = Personal income per household (dollars per month) (Arizona)

W = Evapotranspiration for Bermuda grass minus rainfall (inches) (Boul, 1963; U.S.)

103 P Marginal price facing the average household (cents per 100 cubic feet) (Tucson, 1974, 1975, 1976, 1977)

t Time period in months

n Length of the price lag in months

Water consumption per household (Q) is the dependent variable. Households in this study include all single family residences, apartments, condominiums, mobile homes, duplexes, and triplexes served by individual water connections. Multiple units served by a common water meter were excluded from the study, eliminating those households which receive water at a zero marginal price because its cost is included in their rental payments. The inclusion of a large number of such households would result in an underestimation of the price elasticity of demand for water among households which must pay individual water bills.

A correct specification of the demand model requires the use of two price related variables when there are block rates and /or flat rate service charges in the price schedule (Billings and Agthe, 1980; Nordin, 1976; Taylor, 1975). Since the Tucson water rates include increasing block rates and flat rate charges, two variables must be used in the demand model. The first, marginal price (P) or the per unit price in the marginal block, is the price the average consumer would have to pay for additional unit of water each month, based on average household water use and the City of Tucson water rate schedule. The second price variable (D) is the difference between what the consumer actually pays for water and what would be paid if all water were purchased at the marginal price. This variable measures the flat rate charge imposed on the consumer plus the difference between the amount paid under intra- marginal rates (i.e., block rates below the marginal rate) and what would have been paid at the consumer's marginal price.

The use of the price variable (D) provides a measure of the income effect of any changes in the rate schedule which do not alter the marginal price. It also measures income effects arising from marginal price changes which are in addition to those which would be predicted on the basis of the change in marginal price alone. The difference variable measures a pure income effect which might be expected to have the same impact on water use as any other change in income.

Since this study utilized monthly data, a weather variable (W) to account for variations in sprinkling demand is particularly important. The weather variable in- corporates the effects of rainfall and evapotranspiration which influence the amount of water required for lawn, shrub, and tree irrigation. Since Bermuda grass is the primary residential lawn cover in Tucson, its evapotranspiration rate was selected because it is likely that many residents use the appearance of their lawn to judge their irrigation needs.

Personal income per household (Y) was included in the model to account for variations in the economic situation of water consumers during the study period. This variable was found significant in a previous study of residential water demand in Tucson (Billings and Agthe, 1980) and is normally considered an important factor in most demand analysis.

Because of the substantial inflation which occurred during the study period, the model was estimated using real values of the variables. Real values were derived by dividing the price, difference, and income variables by the consumer price index (Board of Governors). Consumers were found to respond to water prices on the basis of real (Price adjusted) values for water rates and income rather than nominal prices in a previous study (Billings and Agthe, 1980).

If one wishes to model the downward overadjustment with later slight upward consumption adjustments, the Almon Lag Model in cubic form, as plotted in Diagram 1, is probably the best model. Of course, a result with an inflection point is also possible. A potential problem with the Almon model is that a lag of at least 4 periods is required when a cubic equation is tested. Since Hogarty and MacKay (1975) argue that most of the adjustment to marginal price changes occurs in the first three months after the price change, there is some possibility of a cubic equation with a minimum lag of 4 months being too long. To allow for this, a second degree equation with a 3 month lag is considered as an alternative. Results The similarity of the results (Table 1) for the b values of the Almon lagged variables was gratifying while the accompanying low t- values were less satisfying. The low t values may result from the strong multicolinearity inherent in the Almon procedure. In any event, the consistency of results between equations appears to call for

104 acceptance of the b values in spite of the low t ratios.

The majority of the downward adjustment of water consumption in response to marginal price increases appears to be completed by the end of the 3rd month, a result that agrees with Hogarty and MacKay (1975).While there appears to be a slight upward adjustment in consumption for the fourth and fifth months after the price change, the less stable b values and poor t ratios preclude any definitive statement concerning the exact size of this upward adjustment other than it is likely to be small if it exists. It also appears that the adjustment process is complete by the beginning of the fifth month. These results are in agreement with the earlier result using the Koyck model (Agthe and Billings, 1980). This model also demonstrates the greatest decreases occurring in the first period with successively smaller decreases in water and in later months. The long run and short run price elasticity of demand coefficients are presented in Table 2. The short run coefficients for the Almon procedure are calculated for each time period by the formula:

and the long run price elasticity of demand coefficients are calculated by the formula:

4 F E b - i=0

The elasticities were calculated for 5 periods in each case, period to plus 4 lagged periods. The long run results range from .468 to .513. Roughly interpreted, this means that a 1 percent change in price will produce a .5 percent change in water use within 5 periods of the institution of a price change. The Koyck model yielded a long run price elasticity of demand coefficient of .5306. This is slightly larger than the Almon procedure results as the Koyck model does not allow for partial readjustment to a higher consumption level. The Koyck short run price elasticity of demand is - .4164, a coefficient similar to that of the first period Almon procedure values. The price elasticities reported for the Almon process fall well within the range reported in Agthe and Billings (1980).

Conclusion

This study presented the Almon Polynomial Lag as an alternative to the Koyck model to measure consumer response to increased residential water prices because it has the advantage of being able to specify the length and form of the lag. A high amount of multicolinearity inherent in the model precluded obtaining good t ratio results for the most of the lagged prices. Most of the downward adjustment in residential consumption occurs within three months of the price change and any upward readjustment in consumption is likely to be small. Since the upward readjustment produces uncertain statistical results and the price elasticities are about the same for both the polynomial and Koyck results, there is no reason to believe the polynomial results are superior to the Koyck results or that the hypothesized upward adjustment exists.

105 Table 1 - Best results of the application of the Almon Polynomial Lag to Residential Water Demand in Tucson, Arizona

ALMON POLYNOMIAL LAG FORM

Degree and Length of LAG

Variable Koyck or Statistic X2 4 Mo. X3 4 Mo. Lag

Pt -.0964 -.0960 -.0897

t ( -1.91) ( -1.41) (2.32)

Pt-1 -.0347 -.0374 t (-1.404) (.519)

Pt-2 -.0006 -.0001 t (-.017) (.000)

Pt-3 +.0230 +.0161 t (.00) (.012)

Pt-4 +.0064 +.0095 t (.063) (.037)

Pt-5 t

Qt-1 .2154 t (2.15)

W .0053 .0053 -.0042 t (8.09) (7.94) (6.22)

Y .0436 .0436 -.0418 t (2.00) (1.96) (2.27)

D -.0061 -.0061 -.0052 t (-3.46) (-3.40) (-3.30)

Constant -4.589 -4.592 -5.69 t ( -.79) ( -.775) ( -1.15)

Adjusted R2 .793 .787 .817

F 26.52 22.06 39.50

Durbin -Watson 2.04 2.04 h 1.33

106 Table 2- The Short Run and Long Run Price Elasticities of Demand Associated With The Almon Polynomial Lag Results

Model Chosen and LaR

Time Period X2 4 Mo. X3 4 Mo.

t0 -.4475 -.4456

t -1 -.1597 -.1719

t -2 -.0027 -.0005

t -3 +.1037 +.0726

t -41 +.0285 +.0423

Long Run -.4675 -.5131

107 References Cited

Agthe, Donald E. and Bruce R. Billings. 1980. Dynamic models of residential water demand. Water Resources Research. Washington, D.C. 16(3):476 -480.

Almon, Shirley. 1965. The distributed lag between capital appropriations and expend- itures. Econometrica. 33(1):178 -196.

Arizona Department of Economics Security. 1978. Arizona Population and Income Esti- mates. Monthly unpublished data Phoenix, Arizona.

Billings, R. Bruce and Donald E. Agthe. 1980. Price elasticities for water: A case of increasing block rates. Land Economics. 56(3).

Board of Governors. 1977. Federal Reserve System. Federal Reserve Bulletin. Table A -66 (Consumer Price Index). Washington, D.C.

Soul, S. W. 1963. Calculated actual and potential evapotranspiration in Arizona. University of Arizona, College of Agriculture, Tucson, Arizona. Technical Bulle- tin 162.

Gujarati, Danodar. 1978. Basic Econometrics. (McGraw -Hill, New York).

Hogarty, Thomas F. and Robert J. MacKay. 1975. The impact of large temporary rate changes on residential water use. Water Resources Research. 11(6):791 -794.

Houthakker, Hendrick S. and Lester D. Taylor. 1970. Consumer demand in the United States. 2nd Ed. Harvard University Press.

Howe, Charles W. and F. P. Linaweaver, Jr. 1967. The impact of price on residential water demand and its relation to system design and price structure. Water Re- sources Research, 3(1):13 -32.

Nordin, John A. 1976. A proposed modification of Taylor's demand analysis: Comment. The Bell Journal of Economics. 7(2):719 -721.

Taylor, Lester D. 1975. The demand for electricity: A survey. The Bell Journal of Economics. 6(1):74 -110.

Tucson, Arizona. January 1,1974, June 7, 1976 and March 3, 1977. Water Department Rate Schedules. City Ordinance No. 4130, City Ordinance No. 4497 and City Ordi- nance No. 4626.

Tucson. 1978. Arizona Department of Water Sewers. Water use by Month and customer class. Unpublished data on computer tape. Tucson, Arizona.

U.S. Weather Bureau. 1977. Climatological Data Annual Summary, Arizona. Tables 1, 2, 3 and 4. Asheville, N.C. National Oceanic and Atmospheric Administration.

Young, Robert A. 1973. Price elasticity of demand for municipal water: Case study of Tucson, Arizona. Water Resources Research. 9(4):1068 -1072.

108 Diagram 1

The Expected Consumption and Corresponding b- value Patterns for the x3 Almon Polynomial Lag.

Ouantity of Time Lag Water t-0 t-1 t-2 t-3 t-4 t-5

Corresponding 0 Time Leg b- value

109 SEDCON: A MODEL OF NUTRIENT AND HEAVY METAL

LOSSES IN SUSPENDED SEDIMENT

William A. Gabbert, Peter F. Ffolliott, and William O. Rasmussen School of Renewable Natural Resources University of Arizona, Tucson, Arizona 85721

INTRODUCTION

A prototypical computer simulation model has been developed to aid watershed managers in estimating impacts of alternative land management practices on nutrient and heavy metal losses due to transported sediment on forested watersheds of the southwestern United States. The model, called SEDCON, allows users at remote locations with modest computer terminal equipment and commonly available data to obtain reliable estimates of nutrient and heavy metal concentrations in suspended sediment originating on uniformly- stocked,forested watersheds in the Southwest. SEDCON has been structured in an interactive mode to facilitate its use by persons not familiar with computer operations. Written in FORTRAN IV computer language, the model requires approximately 5000 words of core. SEDCON is operative on a DEC -10 computer at the University of Arizona.

FORMULATION OF SIMULATION TECHNIQUE

A study by Gosz, White, and Ffolliótt (1979a, 1979b, 1980) found that combinations of given geologic and vegetative types (other factors of the physical environment were assumed to be reflected by the vegetative type) produce characteristic weathering regimes that influence the physical and chemical characteristics of transported sediment.Using the source data base collected in this study and other appropriate background sources, SEDCON was formulated to estimate nutrient and heavy metal loss by suspended sediment concentrations derived from ponderosa pine and mixed conifer forested watersheds in the southwestern United States (Gabbert, 1982). Geologies evaluated included basalt, sandstone, limestone, and granite. Most of the study areas have been utilized as watershed research areas (Figure 1).

In formulating SEDCON, it was important to select data sets representative of sediment transported during stream flow. Chemical compositions of transported sediments collected in settling tanks on a control watershed (Beaver Creek Watershed Number 19) were compared with sediments taken from the stream channel on one of the study areas (Beaver Creek Watershed Number 13). This comparison was made to determine whether the data used in the computer model were representative of sediment actually transported during stream flow. Geology and vegetation were similar on both watersheds (Gosz, White, and Folliott, 1979a, 1979b, 1980; Gabbert, 1982).

Overall, the results of the above -mentioned comparison support the hypothesis that representative values of transported sediment concentrations can be obtained by sampling channel sediments (Gosz, White, and Ffolliott, 1979a, 1979b, 1980; Gabbert, 1982). As a result, stream channel data were used in the formulation of SEDCON (Gabbert, 1982).

Gosz, White, and Ffolliott (1979a, 1979b, 1980) found that concentrations of chemical constituents varied among soils: under the forest canopy, on stream banks, and in stream channels. Thus, it is important to remember that errors could result when using SEDCON to simulate nutrient and metal loss by transported sediment on disturbed or geologically unstable areas (which could result in bank and surface erosion). Likewise, the model should not be used to simulate losses caused by extreme runnoff events (which also result in bank and surface erosion), as the data could prove to be unrepresentative (Gabbert, 1982).

Chemical constituents transported by suspended sediments in a mixed -conifer forest (Thomas Creek) and a ponderosa pine forest (Beaver Creek) were compared to determine whether the effects of vegetation on nutrient and metal content of sediment could be separated from the effects of geology. Both watersheds were located on similar bedrock. The results of this comparison indicated that it would be difficult to quantify the effects of vegetation. Thus, SEDCON should only be used to estimate nutrient and heavy metal loss on watersheds comprised of: ponderosa pine forests on basalt, granite, sandstone, or limestone; and mixed conifer forests on basalt (Gabbert, 1982).

ln ARIZONA TESUOUE*

O SANTE FE FLAGSTAFF O °ALBUQUERQUE BEAVER CREEK HEBER* THOMAS CREEK* NEW MEXICO O PHOENIX * SACRAMENTO FOREST

O TUCSON

*STUDY AREA 0 60 MILES

Figure 1. Location of study areas.

Predictive equations were developed for each of the above- listed vegetative types and geologies. These equations, which were incorporated into SEDCON, compute 90 percent confidence intervals. General forms of the equations were:

Lower Limit

Y1 =a1Xs + b1Xf where Y1 =lower limit (ppm) of constituent concentration.

al =lower limit coefficient for sand fraction.

bl =lower limit coefficient for fine fraction.

Xs= percent sand in the sediment.

Xf percent fines in the sediment.

Upper Limit

Y2 =a2Xs + b2Xf where Y2 =upper limit (ppm) of constituent concentration.

a2 =upper limit coefficient for sand fraction.

b2 =upper limit coefficient for fine fraction.

In the above equations, it is important to note that there are separate predictive coefficients for the sand (0.061 mm to 2.0 mm) and fine (less than 0.061 mm) fractions of sediment. This separation was essential since Gosz, White, and Ffolliott (1979a, 1979b, 1980) found that constituent concentrations of sediment vary between sands and fines.

APPLICATION OF MODEL

To encourage use of SEDCON, a readily understandable framework was constructed. An illustration of this framework, a simplified flowchart of activities executed in the model, is presented in Figure 2.

112 `VEGETATION ANDGEOLOGY i CHEMICAL CONSTITUENTS /

AREA / i SEDIMENT PRODUCTION / i PERCENT SAND / i

Figure 2. Flowchart of SEDCON.

113 Application of SEDCON can best be demonstrated with a hypothetical example (Figure 3). Operation begins with a question as to which VEGETATIVE TYPE AND GEOLOGY is to be considered. In this example, ponderosa pine forest on basalt was selected.

Next, a list of available TRANSPORTED CHEMICAL CONSTITUENTS is presented. In the example, the terminal operator chose to evaluate all of the constituents. WATERSHED AREA (ACRES), 400 acres, was input next (Figure 3).

The user must then specify whether WINTER OR SUMMER SEDIMENT is to be evaluated. Winter sediments are those that orginate from runoff produced by snowmelt or rain -on -snow events. Summer sediments ori- ginate from runoff produced by thunderstorms. Winter sediments were evaluated in this example. Next, SEASONAL SEDIMENT PRODUCTION (POUNDS PER ACRE) was input at 40 pounds per acre (Figure 3).

SEDCON is not designed to predict sediment production. Instead, this information must be obtained through use of on -site inventory data, appropriate predictive equations, or other simulation models.

An interactive model designed to simulate suspended sediment yields on forested watersheds in central Arizona is available on a DEC -10 computer at the University of Arizona (Rasmussen and Ffolliott, 1979). If desired, this model, called SED, can be linked to SEDCON to provide input data on seasonal sediment production (Gabbert, 1982).

Since concentrations of constituents vary between sands and fines, PERCENTAGE OF SAND in the sediment is requested (Figure 3). Limits have been placed on values that should be selected for PERCENTAGE OF SAND.

Hansen (1966) reported that sands account for 35 to 65 percent of winter suspended sediments. Ad- ditionally, he found a mean concentration of 55 percent of the sediments were sand. As a result, 35 and 65 percent have been incorporated into SEDCON as the range for PERCENTAGE OF SANDS in winter sediment.

A default value of 55 percent sands in winter suspended sediments is available. Acceptance of the default value allows simulation to continue when specific knowledge of the input requested is limited. However, the user has the option of overriding the default value. Since winter sediments were evaluated in the example, the default value of 55 percent sands was utilized (Figure 3).

Hansen (1966) found summer suspended sediment samples varying from 0 to 20 percent sand. Thus, 0 and 20 percent sand in summer suspended sediments were used as limits. A default value of 10 percent sand is available.

At this point, SEDCON will calculate and display estimated values of nutrient and heavy metal transport capacity of suspended sediment. The summary display presents values as acid digestable (primary chemical composition of sediment) and extractable (absorbed to the sediment) nutrients and heavy metals (Figure 3).

At this point, the terminal operator can request another value for percent sand, input a different value for seasonal sediment production, evaluate another season, request to have other constituents evaluated, or select another vegetative type. Anywhere in these steps, the operator can exit the model by answering with a -1. In the example, SEDCON was exited when asked if another value for percent sand was desired (Figure 3).

Since SEDCON calculates seasonal losses of nutrients and heavy metals transported by sediments, it is necessary to execute the model twice, once for winter losses and once for summer losses, to obtain an annual loss. This procedure must be followed for each level management alternative to obtain an estimate of the annual loss of nutrients and heavy metals associated with each alternative.

CONCLUSIONS

SEDCON was developed to aid watershed managers and land use planners estimating the loss of sediment transported nutrients and heavy metals. The model accepts basic watershed data on bedrock geology, vegetative type, and seasonal sediment production. By evaluating management alternatives through the seasonal sediment production function and having an understanding of the significance of a change in site productivity, SEDCON can be used by the land manager to estimate the expected consequences of particular management options.

REFERENCES CITED

Gabbert, William A. 1982. Simulation of nutrient and heavy metal transport capacity of suspended sediment. University of Arizona, Master's Thesis, 49 p.

Gosz, J. R., C.S. White, and P.F. Ffolliott. 1979a. Nutrient and heavy metal transport capabilities of sediment in the Southwest. Final report for Eisenhower Consortium, Project Number 255, 58 p.

114 Gosz, J. R., C.S. White, and P.F. Ffolliott. 1979b. Nutrient and heavy metal transport capabilities of sediment in the Southwest. Final report for the Eisenhower Consortium, Project Number 340, 51 p.

Gosz, J.R., C.S. White, and P.F. Ffolliott. 1980. Nutrient and heavy metal transport capabilities of sediment in the southwestern United States. Water Resources Bulletin, 16:927 -933.

Hansen, E. A. 1966. Field test of an automatic suspended- sediment pumping sampler. American Society of Agricultural Engineers, Transactions 9:738 -743.

Rasmussen, W. 0., and P.F. Ffolliott. 1979. An interactive model of suspended sediment yield on forested watersheds in central Arizona. Hydrology and Water Resources in Arizona and the Southwest, 9:43 -47.

Figure 3. Hypothetical example of SEDCON.

EXECUTE SEDCON

LINK: Loading [LNKXCT SEDCON execution]

VEGETATIVE TYPES AND GEOLOGIES ARE:

1. PONDEROSA PINE FOREST ON BASALT 2. PONDEROSA PINE FOREST ON SANDSTONE 3. PONDEROSA PINE FOREST ON LIMESTONE 4. PONDEROSA PINE FOREST ON GRANITE 5. MIXED CONIFER FOREST ON BASALT

ENTER NUMBER OF VEGETATIVE TYPE AND GEOLOGY TO BE EVALUATED. 1

TRANSPORTED CHEMICAL CONSTITUENTS ARE:

ACID DIGESTABLE

1. CALCIUM 7. COPPER 2. MAGNESIUM 8. MAGANESE 3. SODIUM 9. LEAD 4. POTASSIUM 10. CADIUM 5. ZINC 11. TOTAL NITROGEN 6. IRON 12. TOTAL PHOSPHORUS

EXTRACTABLE

13. CALCIUM 16. CATION EXCHANGE CAPACITY 14. MAGNESIUM 17. ORGANIC MATTER 15. POTASSIUM

18. ALL ACID DIGESTABLE AND EXTRACTABLE CONSTITUENTS

ENTER NUMBER OF CONSTITUENT(S) YOU WISH TO EVALUATE. 18

WHAT IS WATERSHED AREA IN ACRES? 400

DO YOU WANT TO EVALUATE WINTER(1) OR SUMMER(2) SEDIMENT?

WHAT IS WINTER SEDIMENT PRODUCTION IN POUNDS PER ACRE? 40

WHAT IS PERCENT SAND IN SEDIMENT (RANGE:35- 65, GIVES:55)?

115 CONSTITUENT CONCENTRATION (PPM) VOLUME (POUNDS)

ACID DIGESTABLE CALCIUM 3885.-5910. 62.16- 94.56 MAGNESIUM 18090.-29440. 289.44-471.04 SODIUM 381.- 658. 6.10- 10.52 POTASSIUM 895.- 1342. 14.32- 21.47 ZINC 71.- 90. 1.14- 1.43 IRON 57300.-87050. 916.80-1392.80 COPPER 58.- 81. 0.94- 1.29 MANGANESE 1018.-4610. 16.30- 73.76 LEAD 6.- 15. 0.10- 0.24 CADIUM 0.26- 0.26 0.00- 0.00 TOTAL NITROGEN 335.- 766. 5.36- 12.25 TOTAL PHOSPHORUS 542.-1138. 8.66- 18.21 EXTRACTABLE CALCIUM 1413.-2260. 22.60- 36.16 MAGNESIUM 378.-1016. 6.05- 16.26 POTASSIUM 126.- 202. 2.02- 3.23 ORGANIC MATTER 46250.-80200. 740.00-1283.20 CEC(MEQ /100G)= 8.78-46.45

DO YOU WANT TO EVALUATE OTHER CONSTITUENTS? (1 =YES, 2 =NO, -1 =EXIT) 2

DO YOU WANT TO EVALUATE ANOTHER PERCENT SAND VALUE? (1 =YES, 2 =NO, -1 =EXIT) -1

STOP

END OF EXECUTION CPU TIME: 0.27 ELAPSED TIME: 2:8.28 EXIT

116 TECHNIQUES FOR STUDYING NOMPOINT WATER QUALITY

Donovan C. Wilkin Susan J. Hebel

School of Renewable Natural Resources College of Agriculture University of Arizona Tucson, Arizona

Introduction

While the results and techniques described in this paper apply to Midwestern watersheds, there isno reason they can not equally apply to western and eastern watersheds as well. This paper might more properly be titled "Techniques for Studying Undefined Contributions to Water Quality ".The more common distinction between sources for water quality is, of course, between point- as opposed to nonpoint sources. The authors, in their work to help develop a more efficient water quality control program in the State of Illinois, have found the more useful distinction to be between defined sources of water quality and undefined sources. A defined source is one for which the amount of a water quality constituent delivered per unit time to the active stream channel and its location of delivery are known. The active stream channel is defined as that part of the stream bed and floodplain with greater than 50% probability of being flooded in any given year, i.e. the 2 -year floodplain.

The distinction between defined and undefined sources is more important than between point- and nonpoint sources. Effective and efficient water quality management can only be applied to defined sources. Water quality models to deal with management options can only simulate defined sources. Defined sources can be either point sources as, for example, waste treatment facilities, or nonpoint sources as is floodplain farming. Undefined sources, as well, can be either point sources, as are some illegal gravel operation discharges, or nonpoint sources such as upland farming.

In our work in Illinois, we studied several hundred watersheds to distinguish defined and undefined sources of water quality. Prospects for effective and efficient water quality control were seen to be poor given the very large proportion of water quality deriving from undefined sources (Wilkin and Flemal, 1980). Our analyses suggested that, except in a few urban -dominated watersheds, undefined sources were accounting for nearly 100% of the total constituent load in the stream and that total control of defined, and mostly nonpoint, sources would effect little in the way of water quality improvement.

The work suggested several crucial kinds of information needed to convert undefined sources of water quality to defined sources. These are:

1. A fixed network of monitoring locations, including both instream and all effluent locations; 2. periodic sampling for chemical water quality analysis using both filtered and unfiltered analysis techniques; 3. a record of effluent discharge or stream flow at the time of water quality sampling; 4. land use information for critical locations in the watershed; 5. watershed loading information on the amounts and locations and timing of chemical constituents introduced to the watershed; and 6. further theoretical and empirical studies on the movement of chemical constituents in the watershed.

No state has the ideal program for gathering this information. Nonetheless, Illinois has had an unusually good monitoring program for water quality that provides a significant amount of this required information. This paper describes our use of these data and what we have learned therefrom about defining presently undefined sources of water quality.

The Illinois Water Quality Monitoring Network

At times, the State of Illinois has monitored over 400 stations in a fixed network around the State. Each station was sampled monthly, and each sample was analyzed for a suite of over 50 constituents or characteristics. These fixed stations allowed us to divide the state into as many water quality sub - basins, with drainage divides determined by the station locations. Within each subbasin, the location of any defined sources was noted including the flow distance from that source down to the water quality monitoring station. In addition, we attempted to identify the theoretical "center" of the subbasin for ni estimating the average flow distance for nonpoint contributions within the subbasins. Figure 1is a schematic representation of the upper Sangamon River in east -central Illinois showing the subbasinor subsegment breakdown, the location of point sources, monitoring stations, and subsegment centers.

Subsegment or Subbasin Drainage Divides

Point Source

IEPA Ambient Water Quality Monitoring Station

Hypothetical Center of Undefined Subbasin Inputs

Figure 1. Schematic representation of the upper Sangamon River in east-central Illinois. Station E18 is the downstream delimiter.

Mass Balance Pccounting

Mass balance accounting (Wilkin and Flemal, 1979) is the primary technique used for identifying the relative importance of defined and undefined sources in the water quality subbasins. In effect, this procedure, which was ultimately computerized, uses the monitored constituent concentration at each monitoring station, multiplied by flow or discharge and corrected to an annualized rate of metric tons per year of chemical constituent flowing past that point in the stream. In the Illinois network, flow data were not routinely taken, so estimates of flow had to be made based on U.S. Geological Survey stream flow records and drainage area at the sampling station.

Each defined water quality contributor to the subbasin, including the next upstream water quality monitoring station or stations, was accounted for assuming first order assimilation. The rate of assimilation had to be obtained by solving a series of simultaneous equations, one for each subbasin in the analysis of which there had to be at least two. The computer does this by successive approximation.

When the assimilation rate is determined, it can be applied to the constituent load at the upstream station or stations and to all intervening defined sources. These calculations result in the total defined portion of the subbasin load arriving at the downstream monitoring station. Anything left over by simple difference between the total constituent load measured at the downstream station and the various defined contributions from upstream is presumed to be an undefined load originating in the subbasin. This is the starting point for the search for undefined sources.

Wilkin and Flemal have studied the fraction of load deriving from undefined sources across Illinois. Findings from the upper Sangamon River were typical of the state as a whole, outside urban areas. The upper Sangamon consists of 494 square miles of drainage area with mean total discharge of 1.03 cubic feet per second per square mile. The basin consists of five subbasins and has eight point sources contributing, on the average, 0.5 percent of the toal stream flow. Land use in the basin is 2% urban and industrial, 96% agricultural, and 2% other. Table 1shows the proportion of constituent loads from undefined sources in the upper Sangamon watershed.

118 Table 1. Fraction of load from undefined sources, upper Sangamon River, east -central Illinois.

Constituent Fraction I!ndefined

Ammonium Nitrogen 0.84 Barium 0.99 Boron 0.99 Chloride 0.99 Copper 0.99 Fecal coliform 0.62 Fluoride 0.99 Iron, total 0.99 Lead 0.99 Manganese 0.99 MBPS 0.99 Mercury 0.94 Nitrate nitrogen 0.99 Phenol 0.99 Phosphorus, total 0.93 Sulfate 0.99 Total dissolved solids 0.99 Zinc 0.99

An analysis of individual water quality constituents is instructive. The mean basin concentration for iron is 1.09 mg /1 with 32% of the samples violating the Illinois general use standard of 1.0 mg /l. Only 1% of the total iron load derives from defined sources. Even if there were total removal of defined sources, there would be no significant reduction in the violation rate for iron.

Mean basin concentration for fecal coliform organisms is 3,200 MPN /100 ml.Sixty -nine percent violate the Illinois general use standard of 200 MPN /100 ml. Total load from defined sources amounts to 38 %, with 62% undefined. Even with total elimination of defined sources, the minimum feasible violation rate would be in the range of 59 %.

Phosphorus, again, has a mean basin concentration of 0.29 mg /1 with 90% of the samples violating the Illinois general use standard of 0.05 mg /1. Only seven percent of the total load is from defined sources, so the minimum feasible violation rate, with total removal of defined sources, is still 87 %.

Other constituents show the same general story. Clearly, until we can define a larger proportion of the total constituent load, we have no chance of controlling these sources to accomplish our water quality goals. The following describes the various means by which we have attempted to define previously undefined water quality contributions.

Concentration /Discharge Relationships

The importance of monitoring concentration and discharge together cannot he overestimated. For point sources, of course, it is essential to know both concentration and discharge in order to determine the amount of the constituent being introduced to the stream. For streamflow measurements, it is the only way to estimate the total load of the constituent arriving at that point in the stream. Both are essential in knowing what fraction of the load is deriving from defined and undefined sources. Further benefits, however, can be gained in the attempt to identify and define water quality sources. Especially where the monitoring is frequent and where individual flood events can he recorded, we can begin to gain clues to the delivery mechanisms of the constituent to the stream.

For example, where the constituent is being delivered to the stream at a constant rate, as might be the case for constituents entering with ground water seepage, there should be a clear inverse relationship between concentration and discharge. This is the result of dilution. Where a constituent shows such a pattern, the presumption is that of relatively constant delivery. Nitrate nitrogen seems to follow this general pattern.

Where the constituent is delivered to the stream more rapidly at higher stream discharges, while the concentration may do anything from raising to reducing on the rising limb of the hydrograph, the product of discharge and concentration will increase significantly. Phosphorus is a constituent whose concentration tends to increase with discharge. This could be a combination of bed load entrainment and delivery of phosphorus- bearing eroded sediments to the stream channel. Where high concentrations tend to persist after return of the hydrograph to baseline levels, the probable cause is bed load entrainment.

A different pattern is observed for iron. Iron concentrations in our study streams increase sharply with increased flow. Then, at the peak of the hydrograph, they decrease sharply to levels equal to or below baseline levels. This is thought to indicate a "first flush" phenomenon. Temporarily higher flows entrain material that was not otherwise involved in streamflow. This can happen either for

119 materials being deposited on banks or for materials washed in as urban storm runoff. In the case of iron, ground water contains high concentrations of soluble ferrous iron. Seepage of ground water along stream channels allows oxidation to insoluble ferric iron form which precipitates onto the dry banks. Not until the early stages of the next major flood event is this material delivered downstream.

The point of relating discharge and concentration is that we can begin to infer certain characteristics of delivery of the constituent to the stream. This can be of significant help in pinpointing the source.

Sediment- Related Pollution

Sediment has been identified as our most serious water quality problem. The list of problems and costs associated with sediment is lengthy. It includes reservoir sedimentation, disruption of navigation, disturbance of fisheries and wildlife habitat, and added treatment costs for water supplies. Sediment is also implicated in the delivery of a significant chemical fraction to water quality. The Illinois Environmental Protection Agency Task Force on Nonpoint Sources of Pollution (1978) estimated that over 90% of the organic nitrogen and phosphorus from upland agriculture delivered to streams is adsorbed to eroded soil sediment. Clearly, other pollutants can either be delivered to surface waters adsorbed to sediment or can move as unadsorbed particulate by the same pathways as soil sediments. Our groups attempted to study sediment movement in order to gain some ideas about the sources of this kind of water quality constituent.

This work will be reported at length elsewhere, but it can be summarized here. It was necessary to find a tracer that would allow us to study sediment movement across the watershed, to find its areas of erosion and its areas of redeposition.Sediment delivery was defined, for our work, as the process of sediment movement to the active flood plain. Once sediment is at the active floodplain, it can be considered delivered to surface waters since it is only a matter of months, normally, before a flood event suspends the material and moves it downstream. Work had been done with fallout Cesium -137, a byproduct of the atmospheric testing of atomic weapons. Fallout Cesium -137 adsorbs strongly to surface soils. Little of significance is removed with crops. The inference is that, in areas of strong erosion, the surface Cesium -137 concentrations are lower than the average for the watershed. In areas of strong redeposition, they should be higher than average.

The surprising results were that, for a large part of the watershed similar to the upper Sangamon, eroded sediments appear to be trapped by such things as upland depressions, fence rows, hedge rows, roadside ditches, and other obstructions. The fraction of upland sediment delivered to the active floodplain seems very low indeed. No attempt was made to estimate this accurately, but figures in the range of 10% delivery seem appropriate based on our field observations. Sediments eroded from lands very near the floodplain, particularly on the steeply sloping land immediately next to it, are redeposited by obstructions much less frequently. Thus, these eroded soils are delivered at a much higher rate. Floodplain soils, in areas of floodplain farming, are not only among the most severely eroded soils in the watershed, but, since this sediment is by definition already delivered when it is eroded, it has maximum effect on instream sediment levels.

Similar logic, of course, can be applied to those chemical constituents delivered to the stream adsorbed to soil particles, or to those whose physical form causes them to move in the same manner as eroded soil particles. Phosphorus, of course, is one such chemical constituent. It remains now to identify not only what chemicals use such a delivery pathway, but to find out where, when and how they are loaded into the watershed in the first place, since this has a strong influence on the probability of delivery to surface waters.

Land Use /Water Quality Relationships

Early statistical studies relating land use over the entire watershed and the undefined portion of the constituent loading gave largely negative results. The only constituent for which positive results were obtained was nitrate nitrogen. This model explained just.over 30% of the variability for nitrate nitrogen. This can be explained logically. Nitrate nitrogen is highly soluble and is primarily loaded into the watershed as nitrogen fertilizer, an activity occurring relatively uniformly over the watershed. Its high solubility and its tendency to persist in ground water suggests that a relationship should be found between land use over the entire watershed and stream water quality.

As noted previously, a constituent like phosphorus, however, being relatively insoluble and strongly adsorbed to soil particles, should show a very different pattern. Based on the foregoing sediment work and our understanding of the chemical properties of phosphorus, land use close to the stream should be more important in determining instream phosphorus levels.

LANDSAT imagery was used to obtain land use within the 100 -year floodplain of the upper Sangamon. To this was added any urban land use found in each subbasin. These, then, were used in a statistical analysis of the effects of land use on two water quality constituents. Because nitrate nitrogen and total phosphorus seem to be so different in delivery pathways, the former moving with groundwater and the latter with eroded sediment, these were chosen. The early expectation was that phosphorus levels

120 would be better predicted by land use close to the stream than would nitrate nitrogen levels. Results, however, suggested that both instream constituents are strongly predicted by knowing land use close to the streams.

For phosphorus, of course, these findings are consistent with those from the sediment movement work mentioned previously. For nitrate nitrogen, however, based on highly significant regression coefficients, the inference is that enough can happen to ground water constituents, between the point of watershed loading and the stream, that land use close to the stream still tends strongly to control constituent concentrations instream. These effects can be both additive and subtractive, depending on the land use.

Conclusions

These results only scratch the surface. We can not claim to have truly defined any previously undefined water quality sources. This work suggests some fruitful directions to take this kind of research. It seems obvious, both from a theoretical standpoint, and from our empirical results, that land uses close to the streams seem to be unusually important in determining instream constituent concentrations. This would suggest that more work should be done closer to the streams than before.

Much more can be done, however. Constituent loadings into the watershed, their locations and timing could be easily monitored. More effort could be spent in monitoring water quality during the course of storm runoff events to give better definition to the relationship between discharge and concentration. More effort could also be given to distinguishing the water quality constituents whose delivery pathway primarily involves movement as, or adsorbed to, eroded sediment, and to those moving primarily as groundwater. Of course, more could be done in long -term, fixed station monitoring of instream water quality.

References

Illinois Environmental Protection Agency. 1978. Task Force on Agricultural Nonpoint Sources of Pollution. Summary of the Agricultural Task Force Water Quality Plan Recommendation. 31 pp.

Wilkin, D.C. and R.C. Flemal. 1979. A mass balance accounting procedure for estimating contributions to water quality. J. Environmental Systems. 9(l):1 -16.

Wilkin, D.C. and R.C. Flemal. 1980. Feasibility of water quality improvement in three Illinois rivers. J. Water Poll. Control Fed. 52(2): 293 -298.

121 A REVISED PHYTOPLANKTON GROWTH EQUATION FOR WATER QUALITY MODELLING IN LAKES AND PONDS

James Kempf, John Casti and Lucien Duckstein Department of Systems and Industrial Engineering University of Arizona Tucson, Arizona 85721

Abstract

A physiological model of nutrient uptake, based on membrane transport is combined with a phytoplankton biomass growth equation, based on internal nutrient limitation, to form a system of equations modeling phytoplankton growth which are capable of considerably richer dynamics than the Michaelis- Menton -Monod model(1l3) or the Droop model. In particular, since the characteristic time scale of nutrient uptake is considerably faster than that of biomass increase, a singular perturbation problem results, leading to a relaxation oscillation similar to the van der Pol oscillator. In contrast with both the Michaelis -Mentor. Monod model and the Droop model, which were developed using steady state chemostat data, the present model would seem to be appropriate for batch cultures and lakes with long turnover times, where the assumptions of the chemostat steady state are not fulfilled. The qualitative behavior of the model compares favorably with data on batch growth of phytoplankton from the literature.

Introduction

Algal cells raised under the physical -chemical conditions which, in the phytoplankton nutrition and aquatic ecosystem modeling literature, are called the chemostat "steady state ", are usually quite far fron steady state in a physiological sense. The chemostat steady state is maintained by supply- ing enough nutrient externally so that the internal nutrient concentration of the phytoplankton does not change and harvesting off the resultirs growth of phytoplankton, so that the population size remains constant.

If phytoplankton are removed from a chemostat at steady state and allowed to increase in numbers until their integral nutrieri supply is exhausted, a quite different type of phytoplankton population results. The final population size is often far higher than that of the chemostat steady state, and the cellular composition of the population is quite different (Bienfang, 1975). A population at physiological equilibrium is more similar to a batch culture Ïrstationary phase, in which the cells may undergo morphological changes when the external and internal nutrient supply is exhausted (Fogg, 1975). This distinction between the chemostat steady state and the ihysiological steady state is often not clearly drawn in the algal nutrition and aquatic ecosystem modeling literature.

The distinction becomes crucial when modeling lakes with very high turnover times (Hutchinson, 1957) or for explaining algal species succession. In very large lakes and marine environments with low turnover times, the input of nutrients and removal of water containing phytoplankton occurs swiftly enough so that the chemostat is probably a good approximation. In smaller lakes, however, out- flows and inflows ten,' to be restricted, so that turnover times are higher, and the environmental conditions thus seer, better approximated by a batch culture.

Similarly, the transition in an algal sperien succession from one species to another usually occurs when the first species has exhausted the external nutrient supply. Thus the physiological condition of the species being replaced is probably better approx.inzted by that of a stationary phase batch culture than that of a culture at the chemostat steady state, in which the cells are maintained in log phase through nutrient input.

The two most widely used r.oCcls for nutrient limited phytoplankton growth, the Michaelis- Menton- Monod model (M3) (DiToro et al., 1971, 1977) are: the Droop internal nutrient model (Rhee, 1980) were both developed using experimental data from chemostat experiments. In both models, the specific growth rate of the phytoplankton population, u, is a rectangular hyperbolic function of a nutrient concentration, q; being approximately linear in q for q small and appro>.inately constant in q for q large. The difference is that, for the M model, q is the external nutrient concentration, while for the Droop model, q is the internal concentration or cell quota. In addition, the specific growth rate becomes negative in the internal nutrient model, if q falls below q , which is taken to be the level of q necessary for maintaining the cell's metabolic machinery inOgood order.

123 The internal nutrient nodel also requires an equation for the rate of nutrient uptake by the cell, and, in most cases, this is also of a rectangular hyperbolic form in the external nutrient concentration. When both models are exatir:ed analytically near the chemostat stea4y state, the resulting dynamical behavior is quite similar (DiToro, 1980), suggesting that the M model would be the better choice, since it is the simpler.

In this paper, a model of phytoplankton nutrient uptake and growth, based on a simplified model of nutrient uptake, is developed and examined numerically and analytically. The potential for relaxation oscillations similar to those in the van der Pol equation (Howard, 1979) or those discovered in a model for a grazing ecosystem (May, 1977; Kempf, 1981) is demonstrated numerically and a lower bound on the total nutrient in the culture or ecosystem is derived, above which relaxation oscillations can occur. Numerical integration of the model results in a trajectory which is qualitatively similar to the results of a batch experiment (DeMarche et al., 1979) that measured internal nitrate concentration, and also to a case from the literature in which no measure;ert of internal nutrient concentration was made (Fogg, 1975) but which displayed an oscillating population size during the stationary phase of a batch culture. The implications for aquatic ecosysi.er mcceling, in particular, and theoretical ecology, in general, are discussed.

The Model

The transport of nutrient across the algal cell membrane is a crucial step in nutrient uptake. Membrane transport can be separated into two types, depending on whether or not the cell expends energy to facilitate the flow df transported substances. If the process occurs strictly as a result of the electrochemical gradient for the transported substance, then transport is passive, while if energy is expended in the transport of the substance itself or to maintain an ionic gradient for a co- transported solute, the term active transport is used. Membrane transport is perhaps best understood for the sodium -potassium ion pump in nerve cells (Hodgkin, 1964); but for phytoplankton nutrient transport, there is very little specific data.

Various models have been developed to explain both passive and active transport. The most widely accepted is that enzymatic type proteins within the cell membrane latch onto substances outside the cell and either migrate across the cell membrane or change shape so that the portion of the molecule carrying the bound substance protrudes into the cytoplasm, where the substance is released. In active transport, the affinity of the membrane protein for the transported substance is assumed to be enhanced either through the hydrolysis of ATP or through the binding of a co- transported solute (Heinz, 1978). Mathematical models for both passive transport (Heinz, 1978) and active transport (Verhoff and Sudaresan, 1972; Heinz, 1978) have been developed.

In contrast with the above models, an experimental system with glucose pumping activity has been demonstrated in which no movement of a membrane protein is involved. The system consists of a membrane in which two enzymes (a hexokinase and a phosphatase) are immobilized between two layers of a material impermeable to the enzymes' common intermediate (Thomas, 1976). When the membrane is charged with ATP, glucose is pumped from the hexokinase side to the phosphatase side without any translocation of membrane proteins. A mathematical model of this system, based on reaction -diffusion equations, has also been developed and studied both numerically and analytically (Kernevez, 1980).

Rather than embrace any of the above formulations, general dynamical features of phytoplankton nutrient uptake and mass balance considerations will be used to develop the nutrient uptake equation. Evidence exists that phosphate is transported actively in Nitella translucens (Kuhl, 1974) and other species and that assimilation of nitrate and ammonia also requires ATP (Morris, 1974); however, since the exact nature of the mechanism behind active transport is not too well known in general (and may, in fact, be quite different for different types of transported solutes), it seems safer to base the model on the general dynamics of the process rather than on any particular mechanism.

Considering, for the moment, that no conversion of internal nutrient into cell material is taking place, the net flow of nutrient across the cell membrane will be the difference between the gross flux and the backward leakage rate (Heinz, 1978):

q = gross uptake rate - backward leakage rate where

q = internal cellular nutrient concentration (wtvl- 1cell -1)

Data from short term uptake e:,.Fneiments indicates that the gross uptake rate is approximately Michaelis -Menton in fore in the external nutrient concentration, n (Rhee, 1980). Theoretical considerations also suggest that the gross uptake rate probably obeys Michaelis- Menton kinetics, since enzymatic

124 type proteins are probably involved in the forward transport process (Heinz, 1978).

Since phytopiarkton have been observed, both in culture and in nature, to concentrate nutrient very strongly against the electrochemical gradient (Rhee, 1973; Bienfang, 1975), a Michaelis -Menton form for the backward leakage reaction seems inappropriate. A more likely possibility is that the backward leakage reaction is substrate inhibited, so that at very high internal nutrient concentrations, the leakage rate decreases rather than saturating to some maximum rate. The leakage rate would thus be approximately linear in q, for q near zero, achieve a maximum, for some intermediate q, and decrease toward zero, as q approaches infinity (see Fig. 1). Such substrate inhibited kinetics are displayed by the uricase reaction, in which uric acid is converted by the enzyme uricase into allantoin in the presence of oxygen (Kernevez, 1980). If substrate inhibited kinetics are assumed for the leakage rate, then the equation for uptake takes the form:

vnn vqq q (Kn + n) qll + K ) (1) where vn, vg = respective maximum uptake rates (wtv1- 1time1)

K1 = inhibition constant (wtvl- 1ce11 -1)

Kn, Kg = respective half saturation constants (wtvl -1 -cell-1)

Removing the assumption that no conversion of internal nutrient into cell materiallis taking place, the growth of particulate or fixed nutrient, p, will take place at a rate of up wtvl- time , where u is the specific growth ratelof theytop]ankton population. Thus the cell pool will be "thinned out" at a rate equal to uq wtvl cell '.time . Including this term in Eq. 1 gives the final equation for the change in the cell pool:

vn'n vgq uq (2) q (Kn + n) - 2 (Kg+q+K

Before proceeding further it might be helpful to summarize the two assumptions made in developing Eq. 2:

(1) The uptake rate is the difference between a forward transport rate and a backward leakage rate. This is in contrast to most developments in the literature (Rhee, 1980) which assume that the uptake rate is Michaelis -Menton in external nutrient, with perhaps an inhibition term due to increasing internal nutrient concentration, but in line with more general modelins of passive and active transport processes (Heinz, 1979).

(2) The gross uptake rate is Michaelis- Menton and the backward leakage rate is substrate inhibited. Both assumptions seem reasonable from a physiological standpoint, and are supported by general dynamic considerations (i.e., the ability of phytoplankton to concentrate nutrient strongly), although further experiments would be the only way to confirm this.

The functional form of the specific growth rate, u, was left unspecifie(; in the above discussion. A reasonable form, developed by Droop (Rhee, 1980), expresses the specific growth rate as a saturating function of q, which becomes negative if q falls below a cellular maintenance concentration. The equation for the increase in particulate or fixed nutrient would then be

c10 p =um(1 -q) P =uP (3) where pm = maximum growth rate (time-1)

q0 = maintenance level of cellular pool (wtv1-1cel1 -1)

p = particulate or fixed nutrient(wt-v1-1)

In batch culture or in lakes with long turnover times, the total amount of nutrient in the system will be constant. Thus, assuming the lake or culture volume is constant, the sum of the external concentration, the total internal pool concentration, and the particulate concentration must be constant, in order to assure mass balance:

n + a'g'P + p = nT (4)

125 where nT = total system nutrient concentration (wt-v1-1)

a = volume density of the cell (v1cellwt -1); (e.g., reciprical of the fraction of cell volume which is protein)

To facilitate the analysis, it is helpful if all the variables and parameters are rendered dimensionless, so comparisons between quantities are about at the same order of magnitude (Fife, 1979). Let -

g = , P = n = 17,-1-, and T = tum, Eqs. 2 and 3 become: q 'n n

P = (1 - g)P = pP (5a)

ßn Yq = (1 +n) - PR (5b) k2)

1 1 where a = q0, ß = y ( ), y = y ( ), and k = . A now means derivative with respect to T. K q nKqpm qKqum KI

To convert Eq. 4 to dimensionless form, note that if the volume density of the cell, a, is assumed constant, then Eq. 4 becomes:

NT = n+ wQ + P (5c)

where w = aKand NT = KT; and Q = Pg is the nondimensional total cellular pool. n The nondimensional external nutrient concentration, n, can be eliminated from Eq. 5b by solving Eq. 5c for n and substituting:

3(NT - wQ - P) Yq g = ( 1 úg (6) + NT - w 0 1- P) (1 + +42)

Eqs. 5a and 6 can be rendered entirely in terms of Q by substituting g = P and using the chain rule to calculate Q. The resulting equation system will be

P = (1- I) -P (7a)

0(8T - wQ - P) = (7b) (1 + NT - wQ - P) (p2 + PQ + kQ2)) P

Bifurcation Geometry

Experimental evidence (Droop, 1973; DiToro, 1980; Rhee, 1980) indicates that nutrient uptake is often an order of magnitude or more faster than growth. Since the characteristic time of Eqs. 7 will be that of the growth reaction, ß and y will be very large. Dividing both sides of Eq. 7b by ß, the system becomes (Fife, 1979):

R= (1- Q) p (8a)

NT - wQ - P aFQ (8b) EQ = P) (p2 02) 7 + NT - wQ - PQ+ where o =ßand e = ß « 1.

Eqs. 8a and 8b constitute a singular perturbation problem, since the time derivative of Q is multiplied by a small parameter. In the next section the dynamics of the system are described, using a nonstandard approach advocated in Diener and Poston (1980) and Lutz and Goze (1982), and the existence 12fi of a relaxation type oscillation is demonstrated. In order for a relaxation type oscillation to be possible, however, the zero set of Eq. 8b must have the folded structure shown in Fig. 2.

The zero set of Eq. 8b will be given by:

NT - wQ - P aPQ r={(P,Q):(1 + NT - w01 - P) 0} (9) (p2+ PQ + kQ2)

r = {(P,Q): h(P,Q) - g(P,Q) = f(P,4) = 01

Since the characteristic time scale of Q will be considerably faster than that of P, P can be considered a parameter for purposes of examining the geometry of r. Because of the difference in time scales, r is often called the slow manifold (Zeeman, 1977).

Kubicek's method of continuation (Kubicek, 1976; Kernevez, 1980) was used to numerically calculate r for appropriate parameter values (see Fig. 2). The method of continuation forms a set of differential equations for P and Q using the arc length along the curve as the independent variable. The equations are numerically' integrated from a known point using the Adams -Bashforth method with maximal order four and a variable step size. A correction step using Newton's method follows the integration. Details can be found in Kernevez (1980).

The points (P *,Q *) and (P * *, Q * *) in Fig. 2 will be such that:

(P*,Q*),(p**,Q**) E r (10)

(P*,(1*),(P**,(1**) e {(P,(1): aQ = 0}

Following Stewart (1976) and Poston and Stewart (1978), these points will be called fold points. An analytical solution of the two conditions in Eq. 10 for the fold points is impossible; however, the bifurcation geometry of r can be studied using an approach similar to that in Kernevez (1980).

For a particular species of algae, the parameters a, 8, y, w, and k will only change slightly with temperature. The parameter which is most likely to determine whether or not r has the folded structure necessary for relaxation oscillations is therefore NT. Changes in Nwill correspond to changes in the trophic state of the lake or other experimental systém under investigation, and therefore the bifurcation geometry of r with changes in NT will give information on how the dynamics of a particular algal species can change with trophic state.

From Eq.9, if (P,Q) e r then

(P,Q) e {(P,Q): h(P,Q) = g(P,Q)}

This means that points on r will occur when the uptake function, h, and the leakage function, g, are, equal. Considering g, Fig. 1 shows that, for a fixed value of P, g will achieve a maximum at some Qm and will approach zero as Q approaches infinity. On the other hand, h will be monotonically decreasing in both P and Q, as can be seen by examining its derivatives:

2h _ -1 2P (1 + NT - wQ - P)2

8Q (1 + NT - wQ - P)2

Thus, if P is considered to be a slowly varying parameter, h will have a maximum when P = 0 and Q = 0 and will decrease thereafter.

To motivate the following theorem, consider Figs. 3e and 3b, which show how r could not have two fold points. Let the points on the graph of g and h where g and h achieve their maximum be called gm

127 and h , respectively. Then, as shown in Fig. 3a, if g is above hfor all values of P and Q, the only intersection point between the two will be on the tailmof g. Simi9ariy, if P7 is the value of P where

the graph of g, at its inflection point, intersects the graph of h and if Pr is the value of P where

the graph of g, at its maximum, intersects the graph of h, then if Pig < Pm9, only one intersection will occur, as shown in Fig. 3b.

From the above comments, the following two criteria must be satisfied in order for three intersections between h and g to occur:

(1) gay < hm

(2) Pmg < phg

The sequence in Fig. 4 shows how, for increasing P, three such intersections are generated.

These conditions on g h Phg, and 879 are translated into conditions on the bifurcation parameter NT by the following theoYem:m

Theorem 1: In order for r to exhibit two fold points, NT must satisfy:

NT > max[N1, N2] (12) where a N1 2K + (1- a)

a (K+2nw 2n(K+w)

N2 ( 1 - 2n) w [2K + ( 1 - o )] [K + 2n(1 - a) + 4K12]

= cos {3 cos-1 (2K)} k > 4 K= N

Remark: For the values of the parameters used to generate Fig. 4, namely a = 4.9, k =9, w = 0.5, NT = 5:

N1 = 2.33 N2 = -4.594 (13)

and since NT = 5, Inequality (12) is satisfied.

Proof: With respect tR the conditions in Points (1) and (2) above, the equations for g and h can be solved for gm, hm, Pm9, and IT in the following way.

Considering P as being fixed and differentiating g once with respect to Q gives:

02) aP(P2 - (14) k42)2 4 (P2 + PQ

Setting Eq.IL to zero and solving for Q as a function of P gives:

P9 g= (15) Qm K

At which g will have the value:

(16) 9m - 2K+1G

128 Note that Qm will be monotonically increasing with P, so that Qm will move to the right as P increases. On the other hand, gm will be fixed with respect to changes in P.

Differentiating g again with respect to Q gives:

_gLg. 2 k2 Q3 - 6kOP2 - 2P3) = QP (17) aQ2 (P2 + PQ + kQ2)3

2 Inflection points will occur where is zero. Setting Eq. 17 to zero gives the following zero set: aQ

Pg 0 3- kQ9P9 2-k2 = 0 (18)

This cubic will have three real roots if the cubic discriminait is less than zero:

4 k3< 0 (19)

Solving for k gives:

k > (20)

which is the condition on k in the statement of the thorem.

The middle root of the cubic in (18) will be giver by:

2Pg -1 Q9 = cos {3 COS (21) K (2K)}

Q9221 n

Again, Q9 will be monotone increasing with P.

The va'.er of g at the inflection point will be

2an 9i (22) (K + 2n + 4Kn2) which is, again, constant with respect to changes in P.

At P = 0, Q = 0,h achieves its maximum:

NT (23) hm 1 + NT

Substituting (16) and (23) into the condition in Point (1) gives:

o NT 2K +1 <1 +NT (24)

Solving for NT gives N1 in (12).

Phg can be found by equating gm and hm at Q = Qm:

129 1,11g(1 a NT - + K) 2K + 1 [1 + (25) - Pm-(1 + K)]

(26) P9m K + w(NT [2K +1 - a)]

Pig can be found in a similar way;

NT - P4(1 + 2K) 2an (27) (K + 2n + 4Kn2) [1 + NT - P1P(1 + 2K)]

K 2an (28) ( N Pir K+ 2nw T [K + 2n(1 - a) + 4Kn2]

The inequality in Point (2) would then be:

K a \ K // 2an ( (29) K +w`NT+w [2K +(1 -o)] /< K +2nw \NT [K +2 (1 -a) +4Kn2]

Solving for NT gives N2 in (12).

An additional condition for relaxation oscillations is that the P = 0 isocline intersect the slow manifold on the unstable section between the fold points, as shown in Fig. 5. If the intersection point between r and the set:

A = {(P,Q): P = á } _ {(P,Q): P = 0, P # 0} (30)

is denoted by (PI,Q1), then the condition for relaxation oscillation is:

** p**** Q* < Q < Q < PI (31)

The following theorem gives Q1 as a function of the parameters:

a[NT {kw(ak -1) + (1- a)} - a] (32) Q1 [Kw(a2 + a- 1) + a(w - a) - (a - 1)]

Proof: Since Q1E{An r }, substituting the value of P in (30) into the equation for r gives:

1 1 a -1 a 2 QI QI KwQl3 + QI (33) 13-[°15(1-6)+k+0 NT[k á] á }QI + {NT a 2

Factoring out a QI2 and solving for Q1 gives Eq. 33.

Dynamics

As mentioned in the previous section, the singular perturbation problem posed by Eqs. 7 will evolve on two time scales: a fast scale for Q and a slow scale for P. The initial change in Q will be rapid; so that, before P changes appreciably, the trajectory will have already approached to within an infinitesimal distance of r.

There are a number of analytical approaches which have been used to study singular perturbation problems of this type (Fife, 1979; Diener and Poston, 1981; Lutz and Goze, 1982; Nayfeh, 1981). One of the most attractive is nonstandard analysis (Diener and Poston, 1981; Lutz and Goze, 1982). The small parameter E in Eq. 7b is formally set to an infinitesimal number (Robinson, 1974) instead of considering limitirc, behavior for small finite E. While, in the numerical example in Fig. 6, a is anything but infinitesimal, the use of infinitesimals makes it easier to talk about what happens when a finite e becomes small. Lutz and Goze (1982) formally prove the existence of relaxation oscillations in systems such as (7), so comments here will be limited to a discussion of the general behavior of the flow in the

130 state space.

Define the halo of a set, SeIR n as the set of points infinitely near S. Also, with reference to the system in Eqs. 7, define:

A+ _ {(P,Q): f(P,Q) > 01 (34a)

A" _ {(P,Q): f(P,Q) < 0} (34b) and:

r+ = {(P,Q): P > -10;011 (35a)

r = {(P,Q): P <-1,;01} (35b)

+, Then for (P,Q)eA Q > 0; while for (P,Q)cfl ,Q < O. Since the Q equation will equilibrate much faster than the P equation, the flow in the Q direction will be almost perpendicular to the slow manifold, until it is within the halo of the slow manifold. There it sill level out and flow along the slow manifold, it the direction of increasing P, if the flow is in r or in the direction of decreasing P, if the flow is in r".

When the flow on the slow manifold reaches the upper fold point gt (P**,Q * *), the fast foliation in the Q direction will push the flow down onto the lower sheet. Here, P < 0, so the flow will be to the left, in the direction of decreasing P. At the lower fold point, (P *,Q *), the opposite process takes place and the result is a relaxation oscillation, as shown in Fig. 5. Such behavior is called "slaving" by Haken (1978).

lFigs. 6 area numerical integration of Eqs. 7, with NT = 5.0, o = 4.9, w = 0.5, k = 9.0, a = 0.8, and -= 29.5. Initial conditions were P(0) = 0.1 and Q(0) = 0.1. Gear's stiffly stable method was used in the integration, since the difference in time scales between the two equations makes the system stiff. The oscillating nature of the solution is obvious both from the flow in the state space, in Fig. 6a, and from the time trajectories, in Figs. 6b and 6c.

Comparison with Experimental Data

DeMarche et al.(1979) report on an experiment in which the diatom Skeletoma costatem was raised in batch culture arc both the population size, p, and the internalnutrient poor -o Vtrate,q, were monitored. Samples of the culture were taken at regular time intervals, and the population size was determined by measuring particulate nitrogen concentration in the samples. The internal nitrate pool concentration was measured by sonicating a portion of the sample and measuring the resulting increase in the medium nitrate concentration.

In Fig. 7, the data of DeMarche et al.(1979) are replotted as a state space plot, with q against p. Initially, the internal nutrient concentration rose very rapidly, over a period of less than 0.2 day, without any change in the population size. The population size then began to increase more slowly; and, as it did so, the internal nitrate concentration also decreased slowly, until about 2.8 day into the experiment. At this point, q decreased quite suddenly to near zero, again over a period of about 0.2 day, while no large change in p occured. The experiment was terminated at this point.

The qualitative similarity between the experimental data, plotted in Fig. 7, and the numerical integration, in Fig. 6, is striking.The fast initial rise in the internal concentration and the drop at the start of the oscillation match in both, as does the movement along the slow manifold. Unfor- tunately, the experiment was terminated before the oscillation could really develop.

Fig. 8 displays the results of an experiment in which log phase Monodus subtemaneus cells were transferred into a medium containing no nitrogen (Fogg, 1975). After the internal nitrogen pool was exhausted, the population size began to oscillate. Since the cells initially had no external nitrogen source, the oscillation developed sooner than would have been the case if nitrogen had been available.

Discussion

In aquatic ecosystem models which use the M3 model or the Droop model, it is not possible to obtain an unforced limit cycle or relaxation oscillation without another "external" state variable. Addition of another trophic level, for example, zooplankton predators, can lead to population cycles of a predator -prey type (Adachi and Ikeda, 1978; Arnold, 1978) which have been widely discussed in the theoretical ecological literature (May, 1978). Such a system is ecologically quite distinct from the system presented above, in which a rapid change in the physiological state of a single species is

131 causing a population cycle.

Another ecosystem in which a physiological variable might be causing oscillations is the cycles of small mammal populations in the Arctic (Finerty, 1980). The lynx -hare pelt data from the Hudson Bay Company records are one of the most widely cited pieces of evidence for limit cycle oscillations in predator -prey systems (but see May (1980) for questions on the data's validity).Statistical examination of the data for a number of small mammal populations in the Arctic (Finerty, 1980) suggests that the predator population cycles are actually being forced by a population cycle in their chief food supply, the Arctic hare, rather than causing the cycle directly. Finerty speculates on a number of possible causes for the hare population cycle, one of which is the kind of physiologically caused periodicity discussed above.

In general, the internal physiological state of an organism would be expected to equilibrate on a more rapid time scale than the reproductive state of the entire population. To the extent that the reproductive state is influenced by the physiological state, a separation of the dynamics might occur, in which the internal physiological state forms a slow manifold upon which the population size evolves.

Wherever the physiological state manifold has a folded structure, like that of the internal nutrient manifold in the above model, the physiological state variable will change rapidly to a new quasi -equilibrium, and the population size will adjust more slowly. Such a phenomenon could also occur between two different species with radically different characteristic growth rates, for example, whales and krill; with the equation for the species having the higher growth rate forming the slow manifold upon which the system evolves.

Thus, coupling between an organism's physiology and the environment might be responsible for extremely complex population dynamics, which have previously been ascribed to general mass action type interactions between organisms at two trophic levels. A three variable system, with one physiological variable forming a slow manifold, might exhibit the type of chaotic dynamics reported by Rbssler (1979). Two possible examples might be the bloom -dieoff cycles in eutrophic ponds, reported by Barica (1974), and the annual species succession of phytoplankton in lakes (Hutchinson, 1957). For such systems, the population trajectories would look as if they were being generated by a stochastic process, even though the underlying dynamics were deterministic.

Conclusions

A physiologically derived phytoplankton growth model has been presented in which the phytoplankton population size can exhibit relaxation oscillations for certain values of the parareters. The model approximates conditions in batch culture or in a self -contained lake, for which the total nutrient concentration in the ecosystem is relatively constant. The internal nutrient equation isocline forms a slow manifold, to which the flow of the dynamical system is strongly and quickly attracted, and upon which the system evolves. If the isocline for the phytoplankton population size intersects the slow manifold on the unstable section, oscillations in the population size can occur. Evidence supporting the model was presented from a batch experiment in which both the population size and the internal nutrient concentration were measured, and from the large body of data in the literature on batch growth of phytoplankton, where no internal nutrient concentration measurements were made.

The implications of the model for ecosystem modeling in general have been discussed. A comparison with another system, the Arctic small mammal populations, in which population oscillations occurred was made. The possible role of physiological variables in ecosystem models has also been mentioned.

Acknowledgements

This research was supported by NSF Grant CEE 8110778, "Modern Stability and Dynamical Concepts in Water Resources Management." We also wish to thank Dr. H. Haken of the Institut fur Theoretische Physik, Universität Stuttgart, Stuttgart, West Germany, for the opportunity to visit in the summer of 1981, during which many of the seminal ideas in this paper were developed. Dr. Tim Poston also played a large role in clarifying the authors' thoughts on the dynamics of aquatic ecosystems.

References

Adachi, N. and S. Ikeda. 1978. Stability Analysis of Eutrophication Models. IIASA Research Memoran- dum RM- 78 -53, International Institute for Applied Systems Analysis, Laxenburg, Austria.

Arnold, E.M. 1978. Aspects of a zooplankton, phytoplankton, phosphorous system. Ecological Modelling, 5:293 -300.

Barica, J. 1974. Some observations on the internal recycling, regeneration, and oscillation of dissolved nitrogen and phosphorous in shallow, self- contained lakes. Archives of Hydrobiology 73(3):334 -360.

132 Bienfang, P. 1975. Steady state analysis of nitrate -ammonium assimilation by phytoplankton. Limnology and Oceanography 20:402 -411.

DeMarche, J. M., H. C. Curl, Jr., P. W. Lundy and P. L. Donaghay. 1979. The rapid response of the marine diatom Skeletonema costatum to changes in external and internal nutrient concentration. Marine Biology 53:323 -333.

Diener, M. and T. Poston. 1982. On the perfect delay convention or the revolt of the slaved variables. To appear in Chaos and Order in Nature, H. Haken, ed., Springer, Heidelberg.

DiToro, D., D. O'Connor and R. Thomann. 1971. A dynamic model of the phytoplankton population in the Sacramento -San Joaquin Delta. Nonequilibrium Systems in Water Chemistry, Advances in'Chemistry Series, 106, American Chemical Society, Washington, 131 -180.

DiToro, D., R. Thomann, D. O'Connor and J. Mancini. 1977. Esturine phytoplankton biomass models- - verication, analysis, and preliminary applications. The Sea 6:969 -1020, E. Goldberg, ed., John Wiley, Inc., New York.

DiToro, D. 1980. Applicability of cellular equilibrium and Monod theory to phytoplankton growth kinetics. Ecological Modelling 8:201 -218.

Droop, M.R. 1973. Some thoughts on nutrient limitation in algae. Journal of Phycology 9:264 -272.

Fife, P. 1979. Mathematical Aspects of Reacting and Diffusing Systems. Lecture Notes in Biomathema- tics 28, Springer, Heidelberg.

Finerty, J. P. 1980. The Population Ecology of Cycles in Small Mammals. Yale University Press, New Haven, CT.

Fogg, G.E. 1975. Algal Cultures and Phytoplarkton Ecology. University of Wisconsin Press, Madison, WI.

Haken, H. 1978. Synergetics: An Introduction. Springer, Heidelberg.

Heinz, E. 1978. Mechanics and Energetics of Biological Transport. Molecular Biology, Biochemistry, and Biophysics 29, A. Kleinzeller, G.F. Springer, and H. G. Wittman, eds., Springer, Heidelberg.

Hodgkin, A. L. 1964. The Conduction of the Nerve Impulse. Liverpool University Press, Liverpool.

Howard, L.N. 1979. Nonlinear oscillations. Nonlinear Oscillations in Biology, F. C. Hoppenstadt, ed., Lectures in Applied Mathematics 17:1 -67, American Mathematical Society, Providence, RI.

Hutchinson, G.E. 1957. A Treatise on Limnology. Wiley, New York.

Kempf, J. 1980. Multiple steady states and catastrophes in ecological models. ISEM Journal 2(3 -4): 55 -79.

Kernevez, J. P. 1980. Enzyme Mathematics. Studies in Mathematics and its Applications, 10, J. L. Lions, G. Papanicolaov, and R. T. Rockefeller, eds., North Holland, Amsterdam.

Kubicek, M. 1976. Dependence of solution of nonlinear systems on a parameter. ACM Transactions on Mathematical Software 2:98 -107.

Kuhl, A. 1974. Phosphorous. Algal Physiology and Biochemistry, Botanical Monographs, 10:636 -654, W. Stewart. ed., University of California Press, Berkeley, CA.

Lutz, R. and M. Goze. 1982. Nonstandard Analysis: A Practical Guide with Applications. Lecture Notes in Mathematics, 881, Sprinyer, Heidelberg.

May, R. 1977. Thresholds and breakpoints in ecosystems with a multiplicity of stable states. Nature 269:471 -477.

May, R. 1978. Mathematical aspects of the dynamics of animal populations. Studies in Mathematical Biology Part II: Populations and Communities, M.A.A. Studies in Mathematics, 16:317 -366, S. A. Levin, ed.

May, R. 1980. Cree -Ojibwa hunting and the hare -lynx cycle. Nature 286:108 -109.

Morris, I. 1974. Nitrogen assimilation and protein synthesis. Algal Physiology and Biochemistry, Botanical Monographs 10:583 -609, W. Stewart, ed., University of California Press, Berkeley, CA.

133 Nayfeh, A. H. 1981. Introduction to Perturbation Techniques.Wiley -Interscience, New York.

Poston, T. and I. N. Stewart. 1978. Catastrophe Theory and Its Applications. Pitman, London.

Rhee, G -Y. 1980. Continuous culture in phytoplankton ecology. Advances in Aquatic Microbiology 2:151 -204, M.R. Droop and H. W. Jannasch, eds., Academic Press, London.

Robinson, A. 1974. Nonstandard Analysis. Second Edition. Elsevier, New York.

Rössler, 0. 1979. Continuous chaos -four prototype equations. Annals of the New York Academy of Sciences 316:376 -392.

Stewart, I. N. 1976. Catastrophe theory and equations of state: the butterfly in the stirred tank. Mathematical Institute, University of Warwick, Coventry, England.

Thomas, D. 1976. Artificial enzyme membranes: transport, memory and oscillating phenomena. Analysis and Control of Immobilized Enzyme Systems, D. Thomas and J. P. Kernevez, eds., North Holland, Amsterdam.

Verhoff, F.H. and K.R. Sundaresan. 1972. A theory of coupled transport in cells. Biochimica et Biophysica Acta, 255:425 -441.

Zeeman, C. 1977. Catastrophe Theory: Selected Papers (1972- 1977). Addison -Wesley, Reading, MA.

134 0.828

0.724

10.000

0.621

8.333

0.517

6.687 F(0) 0.414

05.000 0.310

3.333 0.207

1.867 0.103

0.000 0.000 1.600 3.200 4.800 6400 8.000 9.600 o

Figure 1. Shape of the substrate inhibited Figure 2. Slow manifold, f(P,Q) = 0, determined by leakage function Kubicek's method of continuation for NT = 5.0, a = 4.9, w = 0.5, and k = 9.0

0.828 0.697

0.724 0.610

0.621 0.523

0.517 0.435

F(0) 0.414 F(0)0.348

0.310 0.281

0.207 0.174

0.103 0087

0.000 0.000 0.000 1.600 3.200 4.800 6.400 8.000 9800 0.000 0.333 0.667 1.000 1333 .1.667 2.000 0 0

Figure 3a. Relation between g(P,Q) and h(P,Q) Figure 3b. Relation between g(P,Q) and h(P,Q) for for gm < hm Pmg > Phg 135 1.033 3.011 e.011 1411 0411 e.000 0.013 o.311 1.311 3403 e.300 an 1411

Figure 4a. Relation between g(P,Q) and h(P,Q) for Figure 4b. Relation between g(P,Q) and h(P,Q) for P < P* P* < P < P **

0400 an 1400 I.e00 3.400 0.110 3.011

Figure 4c. Relation between g(P,Q)andh(P,Q) for Figure 5. Flow in the (P,Q) state plane, showing P ** < P fast flow perpendicular to slow manifold and relaxation oscillation

136 6410 1.0

1.0 I.a

1.0 1410

100

000

0.00

-1410 -1.00 -140-040 0410 04101410 14101410 1410 1.00 140 Ca001 000 0.0 0.40 040 040 1.00 1.0 1.40 1410 1410141005' net

Figure 6a. State plane plot of P vs. Q for Figure 6b. Time trajectory plot of P for numerical numerical integration of Eqs. 8 showing relaxation integration of Eqs. 8 oscillation with all parameters the same as in Fig. 2 and e = 2915. Gear's stiffly stable method was used for the integration and P(0) = 0.1, Q(0) = 0.1.

6.00

6410

1.60

1410

1 1 1 I 1 1 1 1 1 1 i__I 1.00 20 40 '660 l' 80 Particulate nitrogen (p) (yg atoms) 1m i i

0410 Figure 7. Replotting of data from DeMarche et al.

-1410 (1979), Fig. 3, as a state space plot of particu- 0-00 0.0 0.400.0 0.00140 140 1.0 14101410 0.000* late nitrogen, p, vs. internal nitrate pool, q, TOE showing fast -slow transitions prior to extensive growth in p and prior to aborted oscillation. Fast transitions are shown as double arrows, while slow flow is shown as single arrows. Figure 6c. Time trajectory plot of Q fornumerical integration of Eqs. 8

137 400 o o á o o E 300 o

743 200 >ar 100 0-

0 I 2 3 4 5 6 7 40 Days of nitrogen starvation

Figure 8.Replotting of data from Fogg (1975), Fig. 19, p. 56, showing oscillation of population size in a nitrogen limited batch cul ture during "steady state"

138 MUTAGENIC ACTIVITY OF SELECTED ORGANIC COMPOUNDS TREATED WITH OZONE

Leslie Irwin and Cornelius Steelink

Department of Chemistry, University of Arizona Tucson, Arizona 85721

Introduction

In the early 1970's it was discovered by several groups of investigators that trihalomethanes (THM) were present in some American drinking water supplies, and that their presence was due primarily to the process of chlorination. Bellar and co- workers (1974) pinpointed the appearance of trihalomethanes to the addition of chlorine in the water treatment process. Symons and co- workers, in the National Organ- ics Reconnaissance Survey (1975), reached the conclusion that treated water contained higher levels of trihalomethanes than untreated water supplies, as surveyed in cities across the United States. In 1978 Glatz and co- workers reported finding increased mutagenic activity in chlorine -treated water in several U.S. cities.

In 1976 and 1977 Rook published evidence that certain natural organic substances, such as humic and fulvic acids, might be serving as precursors for trihalomethane formation. He pointed out that aro- matic compounds with meta dihydroxy groups produce a high level of chloroform subsequent to chlorination. Other investigators have succeeded in demonstrating formation of chloroform and other chlorinated organics in treated waters (Cheh et al., 1980; Maruoka and Yamanaka, 1980).

Because of these findings, alternatives to chlorination are being considered and studied. One of these alternatives is ozone treatment. Ozone is a proven water disinfectant, but little is known about the products of its reaction with complex organic compounds. This project examined one aspect of the effect of ozone on natural water; that is, the effect on the mutagenic activity of organic compounds that may be present in ozone -treated waters.

Model Compounds

To study this effect, we chose six organic compounds to serve as models for organic material in natural or industrial process water.These compounds include (figure 1):

1. Phenol. 2. Resorcinol. 3. Rutin, chosen to represent fulvic acid because of its meta -dihydroxylated benzene ring. Quercetin, a known mutagenic flavanol, is obtained by cleavage of the glycosidic bond of rutin. 4. Lignin, represented here by coniferyl alcohol which is one of its monomers. Lignin is present in pulp mill effluent and in leachate from woody plants. 5. Tannic acid, a hydrolyzable tannin, which is an ester of a sugar (usually glucose). Tannic acid is also present in leachate. 6. Humic acid, the last model compound, is not represented in the figure, since it is a large complex molecule containing many phenolic, quinoid and benzenecarboxylic acid groups. For this project a commercially prepared salt of humic acid was used (Aldrich sodium salt).

Experimental

These compounds were tested for mutagenic activity in the Ames Salmonella /mammalian microsome assay. The Ames test is a reverse mutation assay utilizing specially developed strains of Salmonella typhimum/- um. These strains require histidine for growth, but can be reverted to histidine prototrophy by certain mutagens. The strains used in these experiments were TA98 which detects frameshift mutation, and TA100 which detects base -pair substitution types of mutation.

A standard agar incorporation assay was done (Ames et al., 1975), with bioactivation supplied by rat liver S9 fraction (induced with phenobarbital). Each compound was tested over a range of concentrations with no ozone treatment; this was a dose -response experiment. Next an ozonation experiment was

139 done for each compound: a single concentration (100 ppm) of the compound was tested for mutagenic activity against varying times of ozonation. Positive controls were N- methyl -N'- nitro -N- nitrosoguanidine, an alkylating agent that is positive in strain TA100 withand without bioactivation, and benzidine, an aro- matic amine which is positive in strain TA98 with bioactivation. The results of each experiment were expressed as the mutation index: this is the ratio of revertant colonies on the test plate to revertant colonies on the control plate. This ratio must be greater than 2.0 in order to demonstrate an increase in mutagenicity over control. An analysis of variance was performed on all data from dose -response experiments and from the ozonation experiments.

For the ozonation experiments ozone was generated from oxygen by means of electrical discharge. The ozone was passed into a 200 -ml aqueous solution of the organic compound for a period of 2, 5, 10 or 15 minutes. Ozone output was measured by a standard iodometric titration (Standard Methods for the Examina- tion of Water and Wastewater, 14th edition). The ozone available to react with the model compounds was determined by measuring the amount of ozone that was not utilized by a water blank in a given time period. The amount of ozone that reacted with the model compound was determined by the difference between the excess ozone given off by the organic solution and by the water blank.

Results And Discussion

Ozone Utilization

The results of ozone utilization are shown in figure 2; ozone consumption was recorded as moles o- zone utilized per mole of organic substrate. The amount of ozone consumed by each model compound appears to fall into a natural grouping based on the structural complexities of the compunds.Phenol and resorcinol utilized the least amount of ozone and their rate of ozone consumption had stabilized within 15 minutes. At the other end of the spectrum humic acid utilized more ozone than any other model compound and its rate was still increasing after 15 minutes of ozone treatment. This is probably a conservative estimate of ozone utilization by humic acid, because a conservative value was used for the molecular weight when calculating the molar ratio of ozone consumption.

The decrease of model compound concentration during the course of ozonation was measured bytheFolin- Ciocalteu test for phenolics (Standard Methods for the Examination of Water and Wastewater, 14th edition). This was done for four model compounds and the results are shown in figure 3. The decrease in concentration of phenol was nearly linear over 15 minutes and ozone had destroyed virtually all measur- able phenol. The other compounds exhibited a rapid decrease to a concentration between 10 to 30 ppm, followed by a much slower rate of decrease.

Mutagenicity Studies

In the dose -response experiments (in which compounds were tested for mutagenicity without ozone treatment) the analysis of variance detected no correlation between increasing concentration and increase in mutagenic activity. All compounds (except rutin) exhibited levels of mutagenic activity that were similar to control levels (The analysis of variance did indicate a significant interaction between strain, bioactivating system and compound; this interaction is due to slightly increased mutagenic activity in strain TA100 (with $9) on the part of rutin (figure 4). This mutagenic activity was not reproducible between experiments, and the increase in activity was attributed to an unidentified impurity). The lack of a dose -response relationship indicates that, under the conditions of this study, these model compounds are not mutagenic prior to water treatment.

In the ozonation experiments the model compounds demonstrated no increase in mutagenic activity after ozone treatment, with one exception. This exception was lignin, which did exhibit an increase in mutagenic activity after two minutes of treatment with ozone; this activity returned to control levels with longer ozone treatment. The analysis of variance clearly indicates the increased response of lignin at two minutes (figure 5).

One interesting point about this observed increase in mutagenic activity is that the responses of the bacterial tester strains differed. Lignin treated with ozone for two minutes exhibited increased mutagenicity in strain TA98 with and without bioactivation, but the response without the bioactivating system was somewhat greater (figure 6). Tested against strain TA100, the treated lignin demonstrated a weak response only in the absence of the bioactivating system; there was no increase in mutagenic activity in the presence of the bioactivating system (figure 7). It appears from this that a transient direct -acting mutagen is being formed and then rapidly destroyed by ozonation.While these increases in mutagenic acitivity are statistically significant, they represent only a weak mutagenic response and may not have biological significance; but it will require further testing to determine this.

This project has demonstrated several things: first, the usefulness of the Ames test in assessing the potential mutagenic effects of a water treatment process, and second, the production of a transient mutagenic species upon ozonation of lignin. The model system presented here could be used to study more completely the effect of ozone treatment on lignin in water supplies.

140 Acknowledgment

This work was supported in part by the Office of Water Research and Technology, A -094, U.S. Depart- ment of Interior, Washington, D.C., as authorized by the Water Research and Development Act of 1978, and in part by the U.S. Environmental Protection Agency, Office of Research and Development,EPAN0.CR- 807363.

References Cited

Ames, B.N., J. McCann and E. Yamasaki.1975 . Methods for detecting carcinogens and mutagens with the Salmonella /mammalian microsome mutagenicity test. Mutation Research, 31: 347 -364.

Bellar, T.A., J.J. Lichtenberg and R.C. Kroner. 1974 . The occurrence of organohalides in chlorinated drinking waters. J. American Water Works Association, 66: 703 -706.

Cheh, A.M., J. Skochdopole, P. Koski and L. Cole. 1980. Nonvolatile mutagens in drinking water: production by chlorination and destruction by sulfite. Science, 207: 90 -92.

Glatz, B.A., C.D. Chriswell, M.D. Arguello, H.J. Svec, J.S. Fritz, S.M. Grimm and M.A. Thomson. 1978. Examination of drinking water for mutagenic activity. J. American Water Works Association, 70: 465 -468.

Maruoka, S. and S. Yamanaka. 1980. Production of mutagenic substances by chlorination of waters. Mu- tation Research, 79: 381 -386.

Rook, J.J. 1976. Haloforms in drinking water. J. American Water Works Association, 68: 168 -172.

Rook, J.J. 1977. Chlorination reactions of fulvic acids in natural waters. Environmental Science and Technology, 11: 478 -482.

"Standard Methods for the Examination of Water and Wastewater," 14th edition, American Public Health Association., Washington, D.C., 1974, pp. 456 -457, 607 -608.

Symons, J.M., T.A. Bellar, J.K. Carswell, J. DeMarco, K.L. Kropp, G.G. Robeck, D.R. Seeger, C.J. Slocum, B.L. Smith and A.A. Stevens. 1975. National Organics Reconnaissance Survey for Halogenated Organ- ics. J. American Water Works Association, 67: 634 -647.

141 H OH o

Phenol Resorcinol

OH HC= CHCH2OH

Rutin Coniferyl Alcohol (monomer of Lignin) COON GGGG

H Tannic Acid G= Gallic Acid

Figure 1. Model compounds. Humic acid (Aldrich sodium salt) is the only model compound not shown.

142 Figure 2. Utilization of ozone by model organic compounds. Ozone utilization is calculated as a molar ratio. All conditions of ozonation were held constant except for time. Ozone dose = 7.1 mg /L.

i0 15 Time of Ozonotion (mini

Figure 3. Decrease in model compound concentration during ozonation. Standard curves of model compound concentration were constructed using the Folin -Ciocalteu (F -C) test for hydroxylated aromatics; the same compounds were then measured by the F -C test subsequent to ozonation.

Time of Ozonotion (min)

145 .5 59 i.pnImaN 5.9 7.59. .59 Figure 4. Analysis of variance for dose - O. ram.un 59 response experiment. The significant interaction (S9 x strain x compound) does not 03 involve concentration, but indicated a weak mutagenic response in strain TA100 with bioactivation when tested against rutin. Mutation index must exceed 2.0 to indicate a significant increase in mutagenic activity (log 2 = 0.301).

iz151. ixln. II xja iz5i. II Iz151.

I RNm I h9n.adel usmn qummx5 0 ieil

07-

a6 -

as Figure 5. Analysis of variance for ozonation experiment. Only lignin at two minutes of oe ozone treatment exceeded

level of significance. as Further ozonation of lignin caused mutagenic activity to return to control levels. = oz

s a 9 7 r T, 0 r14-

-a2

-03

:5 = xx 2'4 -oa )¢ 2mirt 15nnJ ImóI= I£i5min I Compound x Time Interaction (Ozonation Experiment)

144 Figure 6. Mutagenic response of strain TA98 to ozone- treated lignin. The mutagenic activity is greater in the absence of the bioactivating system than in its presence.

Figure 7. Mutagenic response of strain TA100 to ozone- treated lignin. The mutagenic activity is increased over control only in the absence of the bioactivating system.

145 Key Word Index of Arizona Section (AWRA) Proceedings

The following index is a compilation of all articles published through the auspices of the Arizona Section of the American Water Resources Association (AWRA).The index contains articles from the proceedings of meetings held jointly with the Hydrology Section of the Arizona Academy of Science. In addition, proceedings of three symposia sponsored by the Arizona Section (AWRA) are also included.

The listing is arranged in alphabetical order with each key word in a title appearing as the leading entry in the index. Each title may have as many as three keywords within the index. Because an entry may begin midway through a title, the end of a title is identified by the symbol S.

Following the title, the author(s) is identified. The volume or symposia number is listed after the author. Volumes (identified by the label Y #) relate to publications of the series Hydrology and Water Resources in Arizona and the Southwest (Volumes 1;1971 through Volume 11; 1981). Symposia can be identified from the following list:

P1 - Symposium on Water Conservation Alternatives (April 12, 1979; Phoenix, Arizona)

P2 - Symposium on Flood Monitoring and Management (October 26, 1979; Tucson, Arizona)

P3 - Symposium on Water Quality Monitoring and Management (October 24, 1980; Tucson, Arizona)

EXAMPLE:

Collective Utility Viewpoint $ Comparison of Water Pricing Structures from a Bill Metter and Lucien Duckstein - V1

The actual title of the article is "Comparison of Water Pricing Structures from a Collective Utility Viewpoint ";the $ symbol indicating the title beginning (or ending). The authors appear on the following line along with the publication source, V1 (indicating Volume 1;1971 in Hydrology and Water Resources in Arizona and the Southwest). Additionally, the title could be found using the keyword Water Pricing.

147 The following list contains the keywords which were used to compile the index. Associated with some of the keywords are related keywords which may be used to locate desired material.

Acid Price (also; Water Rates) Agriculture (also; Irrigation) Probability Aquifer (also; Ground -Water) Public Arid (also; Desert, Semiarid) Pumping Attitudes (also; Water Awareness) Quality Bacteria Rain (also; Precipitation) Biologic Range Burning (also; Fire) Recharge (also; Transmission Losses) CAP Reclamation (also; Rehabilitation, Revegetation) Central Arizona Project Recreation Chaparral Regression (also; Error Analysis) Chemical Rehabilitation (also; Reclamation, Revegetation) Coal Reservoir Collective Utility Residential (also; Municipal, Urban) Colorado River Revegetation (also; Reclamation, Rehabilitation) Conservation Runoff Contamination (also; Pollution) Safe Drinking Water Act Copper Salinity Desalination (also; Osmosis) Salt Desalt Saltcedar Desert (also; Arid, Semiarid) Sediment Development Seepage Earth Fissure (also; Subsidence) Semiarid (also; Arid, Desert) Economic Simulation (also; Models) Effluent (also; Waste -Water) Snow (also; Precipitation) Environment Soil Ephemeral Solar Erosion Stochastic Error Analysis (also; Regression) Stock Evaporation (also; Transpiration) Storm (also; Flood) Fire (also; Burning) Streamflow Flood (also; Storm) Strip -mined Forage Subsidence (also; Earth -fissures) Forest Temperature Geomorphic Thunderstorms Goals Time Ground -Water (also; Aquifer) Tracer Hydrologic Transmission Losses (also; Recharge) Hydrology Transmissivity Infiltration Transpiration (also; Evaporation) Information Tree -Ring Irrigation (also; Agriculture) Uncertainty Lysimeter Unit Hydrograph Management Universal Soil Loss Equation (also; USLE) Mexico Urban (also; Municipal, Residential) Mine USLE (also; Universal Soil Loss Equation) Mining Vegetation (also; Plant) Models (also; Simulation) Waste -Water (also; Effluent) Multi - Water Awareness (also; Attitudes) Multiple -Use Water Consumption Municipal (also; Residential, Urban) Water Harvesting (also; Water Impoundment, Wax) Nitrogen Water Impoundment (also; Water Harvesting) Nutrient Water Quality (see Quality) Osmosis (also; Desalination) Water Rates (also; Price) Pine Water Resource Planning Water Rights Plant (also; Vegetation) Water Supply Policy Wax (also; Water Harvesting) Politics Weather Modification Pollutants Weathering Pollution (also; Contamination) Wetlands Precipitation (also; Rain, Snow) Well

148 Key Word Index of Arizona Section (AWRA) Proceedings

Acid Drainage From Abandoned Mines in the Patagonia Mountains, Arizona $ Sheila A. Dean - P3 Acid Solutions in Selected Arizona Soils $ Penetrability and Hydraulic Conductivity of Dilute Sulfuric S. Miyamoto, J. Ryan and H. L. Bohn - V3 Acid:Its Potential for Improving Irrigation Water Quality $ Sulfuric H. L. Bohn and R. L. Westerman - V1 Agricultural and Municipal Groundwater Requirements in the Tucson Area, Arizona$ Jojoba: An Alternative to the Conflict Between ... Kennith E. Foster and N. Gene Wright - P1 Agricultural Water Budget$ A Potential for Water -Efficient, C44 Halophytes in Arizona's Edward P. Glenn, James W. O'Leary and Barney P. Popkin - V11 Agriculture, and Implications for Urban Water Supply: The Tucson Case $ Rising Energy Prices, Water Demand by Peri -Urban ... H. W. Ayer and D. W. Gapp - V8 Agriculture s The Price of Water in Western David L. Wilson and Harry W. Ayer - V11 Aquifer Analysis$ Application of Finite Element and Computer Graphics Techniques in D. F. O'Donnell, L. G. Wilson and W. O. Rasmussen - V5 Aquifer Characteristics Using Drillers' Logs $ Estimations of Kandy G.Kisser and JillS. Haimson - VII Aquifer Test Error Analysis $ Aspects of Ahmed M. Benbarka and Donald R. Davis - V11 Aquifer Tests $ Correcting Tidal Responses in Observed Water Well Levels During Coastal Barney P. Popkin - V11 Aquifer $ Geostatistical Analysis and Inverse Modeling of the Avra Valley Peter M. Clifton and Shlomo P. Neuman - V11 Aquifer $ Transmissivity Distribution in the Tucson Basin D. J. Supkow - V2 Aquifers s The Application of Step -Drawdown Pumping Tests for Determining Well Losses in Consolidated Rock ... V. W. Uhl, Jr., V. G. Joshi, A. Alpheus and G. Sharma -V5 Arid Environment. Tucson, Arizona $ Water Quality Problems of the Urban Area in an G. Hansen - V8 Arid Environments $ Estimating Potential Evapotranspiration in Z. Osmolski and L. W. Gay - V11 Arid Lands Vegetable Production $ Mulching Techniques for R. W. Peebles and Norman F. Oebker - V1 Attitudes of Arizona Water Resource Leaders $ Man -Nature Roger A. Kanerva and David A. King - V2 Bacteria Removal and Infiltration in Soil Columns$ The Effect of Increasing the Organic Carbon Content of Sewage on Nitrogen, Carbon, and ... J. L. Lance and F. D. Whisler - V5 Bacterial Indicators in Assessment of Water Quality of the East Verde River$ Use of Patrick V. Athery, Marilyn J. Urbina and Milton R. Sommerfeld - 1111 Bacterial Water Quality Investigation of Canyon Lake $ A W. F. Horak and G. S. Lehman - V4 Biological Problems in the Grand Canyon $ Chemical and G. C. Slawson, Jr. and L. G. Everett - V3 Burning on Surface Water Quality $ Some Effects of Controlled Bruce D. Sims, Gordon S. Lehman and Peter F. Ffolliott - V11 CAP Intake Area $ Water Quality Study of Lake Havasu, Arizona near the Simon Ince, David L. Kreamer, Don W. Young and Charles L. Constant - V6 CAP on Lake Havasu's Thermal Regime $ Future Effects of the David Kenneth Kreamer - V6 Central Arizona Project Concept of Operation$ Frank C. Springer, Jr. and Albert L. Graves - V9 Central Arizona Project Water Allocation Model System$ The Arizona Water Commission's Philip C. Briggs - V7 Central Arizona Project $ An Economic Analysis of the James L. Barr and David E. Pingry - V7 Central Arizona Project $ Economic Adjustment to a New Irrigation Water Source: Pinai County, Arizona and the ... Mark A. Boster and William E. Martin - V5 Chaparral to Grass to Increase Streamflow$ Converting Paul A. Ingebo - V2 Chaparral Watershed in Central Arizona $ Sediment Production from a Thomas E.Hook and Alden R. Hibbert - V9 Chaparral $ Root System of Shrub Live Oak in Relation to Water Yield by Edwin A. Davis - V7 Chemical and Biological Problems in the Grand Canyon $ G. C. Slawson, Jr. and L. G. Everett - Y3 Chemical Character of a Perched Water Zone, Harquahala Valley, Arizona$ Origin, Development and Charles G. Graf - V10

149 Chemical Hydrographs in Groundwater Quality StudiesS The Use of Kenneth D. Schmidt - V1 Chemical Oxygen Demand of the Antitranspirant, Folicote $ R. H. Garrett and B. E. Kynard - V7 Chemical Quality of Streamflow by an Interactive Computer Model S Prediction of the William O. Rasmussen and Peter F. Ffolliott - V10 Coal Mine Lands $ Hydrologic Evaluation of Topsoiling for Rehabilitating Black Mesa Frank G. Postillion - V10 Coal Mining Effects on Water Quality of the Tongue River, WyomingS Preliminary Results from a Study of Richard D. Olsen and Edward H. Dettmann - V6 Collective Utility of Exchanging Treated Sewage Effluent for Irrigation and Mining WaterS Stephen C. Ko and Lucien Duckstein - V2 Collective Utility Viewpoint $ Comparison of Water Pricing Structures from a Bill Metler and Lucien Duckstein - Vi Collective Utility: A Systems Approach for the Utilization of Water Resources $ Edwin Dupnick and Lucien Duckstein - Vi Colorado River Corridor of Grand Canyon $ Water Quality Analyses of the Brock Tunnicliff and Stanley K. Brickler - V10 Colorado River Driftwood in the Grand CanyonS Tree -Ring Dating of C. W. Ferguson - V1 Colorado River System $ Salinity Control Planning in the John T. Maletic - V4 Colorado River Trips: A User Study $ An Investigation of Mark A. Boster and RussellL. Gum - V2 Colorado River$ Politics and the Wesley E. Steiner - Vi Conserve Water in ArizonaS Irrigation Management and Water Policy: Opportunities to Harry W. Ayer and PaulG. Hoyt - V10 Conservation in Arizona's Future S Role of Laurence Linser - P1 Conservation in ArizonaS Future Outlooks for Water Sol Resnick - Pi Conservation in the President's Water Policy S Water Gary D. Cobb - P1 Conservation in Water Supply PlanningS The Role of Augustine J. Fredrich - P1 Conservation S Land Treatment of Primary Sewage Effluent: Water and Energy R. C. Rice and R. G. Gilbert - V8 Conserving Water and Energy in Irrigation $ J. A. Replogle and D. D. Fangmeier - P1 Conserving Water with Drought -Tolerant Crops $ D. L. Johnson and W. L. Ehrler - P1 Contamination in the Cortaro Area, Pima County, Arizona $ Groundwater Kenneth D. Schmidt - V2 Desalination S Fresh Water for Arizona by Salt Replacement Anthony B. Muller - V4 Desalting Plant $ The Yuma Dana B. Hill and K. M. Trompeter - P3 Desert Cities $ A Rational Water Policy for W.G. Matlock - V4 Desert FarmerS Use and Abuse of Southwestern Rivers: The J. E. Ayres - V1 Desert Stream Channel $ Nitrogen Species Transformations of Sewage Effluent Releases in A P. G. Sebenik, C. B. Cluff and K. J. DeCook - V2 Desert -Household Gardening $ Salvaging Wasted Waters for D. H. Fink and W. L. Ehrler - V8 Develop Snowmelt- Runoff Forecasts in Arizona$ Use of Satellite Data to Peter F. Ffolliott and William O. Rasmussen - V6 Developed in Conjunction with the Water Use Inventory on BLM Administe $ Quantification Methods Marvin Goss - V10 Developing Forest Management Guidelines for Increasing Snowpack Water YieldsS Progress in David B. Thorud and Peter F.Ffolliott - V1 Development (Abstract) $ Impact on the Environment by Water Resources Sol Resnick - V3 Development and Chemical Character of a Perched Water Zone, Harquahala Valley, ArizonaS Origin, Charles G. Graf - V10 Development and Testing of a Laser Rain Gage $ Arnold D. Ozment - V5 Development by the Appropriation Doctrine S Constraints on Water William L. Lorah - V4 Development of a Groundwater Quality Protection Strategy for Arizona $ Toward Marc M. Bennett and Larry K. Stephenson - V11

150 Development on Groundwater in the Parker StripS The Effect of L. G. Everett and T. R. Schultz - V4 Development on Stream Discharge in Navajo and Apache Counties, ArizonaS The Effects of Second -Home and

Resort -Town . . . T. D. Hogan and M. E. Bond - V9 Development on Stream Flows$ Impact of PaulD. Trotta, James J. Rodgers and William B. Vandivere - V9 Development S Legal Aspects of Urban Runoff D. A. Chudnoff - V8 DevelopmentS Storm Flows Management in Relation to Industrial Robert E. Smith - P2 Development S Watershed Indicators of Landform Burchard H. Heede - V5 Development$ Well -Field Design Criteria for Coastal Seawater Barney P. Popkin - V10 Earth Fissure Movement in South -Central Arizona S Character of M. C. Carpenter and J. K. Boling, Jr. - V10 Economic Adjustment to a New Irrigation Water Source: Pinal County, Arizona and the Central Arizona Project $ . Mark A. Boster and William E. Martin - V5 Economic Alternatives in Solving the U.S.- Mexico Colorado River Water Salinity Problem S William E. Martin - V4 Economic and Energy Opportunities for Municipal Waste -Water Utilization in Arizona $ E. J. Weber -P3 Economic Analysis of the Central Arizona Project S An James L.Barr and David E.Pingry - V7 Economic Impacts of the Safe Drinking Water Act on Arizona's Water Systems$ Socio- Richard S. Williamson - V10 Effluent - An Element of Total Water Resource Planning $ Wastewater J. D. Goff - V8 Effluent - Irrigation Water Exchange S Metropolitan Operated District for Sewage C. Brent Cluff and K. James DeCook - V4 Effluent Applied to a Soil -Turf FilterS Nitrogen Removal from Secondary E. L. Anderson, I. L. Pepper and G. V. Johnson - V8 Effluent by Ground -Water Recharge $ Renovating Sewage Herman Bouwer, J. C. Lance and R. C. Rice - V1 Effluent in the Santa Cruz River $ Transformations in Quality of Recharging L. G. Wilson, R. A. Herbert and C. R. Ramsey - V5 Effluent Lakes, Puerto Penasco, Sonora, Mexico $ A Study of Salinity in Alison L. Dunn - V11 Effluent Recharge S Effect of Illuviated Deposits on Infiltration Rates and Denitrification During Sewage ... Richard Alan Herbert - 47 Effluent Releases in A Desert Stream Channel $ Nitrogen Species Transformations of Sewage P. G. Sebenik, C. B. Cl uff and K. J. DeCook - V2 Effluent $ Effect of Algal Growth and Dissolved Oxygen on Redox Potentials in Soil Flooded with

Secondary Sewage . . . R. G. Gilbert and R. C. Rice - V8 Effluent $ Rehabilitation of Copper Mine Tailing Slopes Using Municipal Sewage Tika R. Verma, Kenneth L. Ludeke and A. D. Day - V7 Effluent $ The Use of a Computer Model to Predict Water Quality Transformations During Subsurface

Movement of Oxidation Pond . G.G. Small and L.G. Wilson - V3 Effluent: Water and Energy Conservation $ Land Treatment of Primary Sewage R. C. Rice and R. G. Gilbert - V8 Environment by Water Resources Development (Abstract) $ Impact on the Sol Resnick - V3 Environmental Impact Statement (USDA - SCS, 1978): A Look at State -of$ An Examination of the Buckhorn- Mesa Watershed ... Dale A. Altshul - V9 Environmental Impact Statement S Systematic Assessment of Uncertainties in an Soronadi Nnaji, Donald R.Davis and Lucien Duckstein - V6 Environmental Perspective $ Water for Food, Energy and Municipal Use in the Colorado Basin: A Consumer - Barbara Tellman - V6 Environments $ Estimating Potential Evapotranspiration in Arid Z. Osmolski and L. W. Gay - V11 Ephemeral Channels $ A Proposed Model for Flood Routing in Abstracting Leonard J. Lane - V2 Ephemeral Flow and Water Quality Problems: A Case Study of the San Pedro River in Southeastern $ S. J. Keith - 88 Ephemeral Flow in the Tucson Basin: Implications for Ground Water RechargeS Seasonal and Spatial Trends of ... Susan J. Keith - V10 Ephemeral Mountain Stream $ Equilibrium Condition and Sediment Transport in an Burchard H. Heede - V6 Ephemeral Stream Channels$ Estimating Transmission Losses in Leonard J.Lane, Virginia A. Ferreira and Edward D. Shirley - V10 Ephemeral Stream Channels$ Water Disposition in T. W. Sammis - V2

151 Ephemeral StreamS Bed Material Characteristics and Transmission Losses in an J. B. Murphey, L. J. Lane and M. H. Diskin - V2 Ephemeral Stream $ Objective and Subjective Analysis of Transition Probabilities of Monthly Flow on an William Dvoranchik, Lucien Duckstein and Chester C. Kisiel - V2 Erosion and Sediment Control on the Reclaimed Coal Mine Lands of Semiarid SouthwestS Soil Tika R.Verma, John L. Thames and John E. Mills - V7 Erosion and Sedimentation in the Upper Gila Drainage, A Case Study $ R. L.. Kingston and R. M. Solomon - V6 Erosion Index of the Universal Soil Loss Equation S Thunderstorm Precipitation Effects on the Rainfall - Kenneth G. Renard and J. Roger Simanton - V5 Erosion of a Semiarid Southwestern Rangeland Watershed $ Effects of Brush to Grass Conversion on the Hydrology and ... J. R. Simanton, H. B. Osborn and K. G. Renard - V7 Erosion Simulation Model S A Sediment Yield Equation From an E. D. Shirley and L. J. Lane - V8 Error Analysis of Evapotranspiration Measurements S Robert K. Hartman - V10 Error Analysis $ Aspects of Aquifer Test Ahmed M. Benbarka and Donald R. Davis - V11 Evaporation Via Surface Temperature Measurements S Assessing Bare Soil Sherwood B.Idso, Robert J. Reginato and Ray D. Jackson - V5 Evaporation S Color It M. J. Dvoracek - V2 Evapotranspiration from SaltcedarS An Energy Budget Analysis of L. W. Gay, T. W. Sammis and J. Ben -Asher - V6 Evapotranspiration in Arid Environments S Estimating Potential Z. Osmolski and L. W. Gay - V11 Evapotranspiration MeasurementsS Error Analysis of Robert K. Hartman - ViO Evapotranspiration Models (Abstract)S How to Select T. E. A. van Hylckama, R. M. Turner and O. M. Grasz - V9 Fire on Water Infiltration Rates in a Ponderosa Pine StandS Effects of Malcom J. Zwolinski - V1 Flood Control Problem Area, State of Arizona S William D. Mathews - P2 Flood Estimates S Tests on Arizona's New Brian M. Reich, Herbert B. Osborn and Malchus C. Baker, Jr.- V9 Flood Estimation $ A Solution to Small Sample Bias in William Metler - V2 Flood Management Problems: Kassandra and the Sirens $ Two Flash - Susan Zevin and Jose Marrero - P2 Flood Routing in Abstracting Ephemeral ChannelsS A Proposed Model for Leonard J. Lane - V2 Flood Warning System for Tucson Basin $ Early Dan Chudnoff and Robert Reynolds - P2 Flooded with Secondary Sewage Effluent S Effect of Algal Growth and Dissolved Oxygen on Redox Potentials

in Soil . R. G. Gilbert and R. C. Rice - V8 Flooded with Sewage Water S Addition of a Carbon Pulse to Stimulate Denitrification in Soil Columns J. C. Lance and R. G. Gilbert - V6 Floodplain Management in Sonora, MexicoS Traditional Technology for Thomas E. Sheridan and Gary P.Nabhan - P2 Floodplain Management Within Pima County, Arizona S Present Practices and Future Goals for Brian M. Reich and MichaelE. Zeller - P2 Floodway Delineation $ Application of Remote Sensing in Robin B. Clark - V4 Forage Production on a Semiarid Rangeland Watershed$ Increasing J. M. Tromble - V4 Forest and Rangelands in Arizona $ Water Resources Research on Alden R. Hibbert - V4 Forest Density on Bedload Movement in a Small Mountain Stream$ Influence of Burchard H. Heede - V7 Forest Lands in Arizona S Water Yield Opportunities on National Rhey M. Solomon and Larry J. Schmidt - V11 Forest Management Guidelines for Increasing Snowpack Water Yields$ Progress in Developing David B. Thorud and Peter F.Ffolliott - V1 Forest Openings$ A Technique to Evaluate Snowpack Profiles in and Adjacent to Peter F. Ffolliott and David B.Thorud - V4 Forest Watershed$ Snowpack Density on an Arizona Mixed Conifer Peter F. Ffolliott and J.R. Thompson - V7 Forested Watersheds in Central ArizonaS An Interactive Model of Suspended Sediment Yield on William O. Rasmussen and Peter F. Ffolliott - V9 Forested Watersheds on Sedimentary Soils S Water Quality of Streamflow from PaulW. Gregory and Peter F. Ffolliott - V6

152 Forests f Evaluation of the Use of Soil Conservation Service Snow Course Data in Describing Local Snow

Conditions in Arizona . Gerald J. Gottfried and Peter F. Ffolliott - V11 Forests $ Lysimeter Snowmelt in Arizona Ponderosa Pine MikealE. Jones, Peter F. Ffolliott and David B. Thorud - V6 Forests $ Snowpack Dynamics in Arizona's Aspen Michael J. Timmer, Peter F. Ffolliott and William O. Rasmussen - V10 Forests $ Windbreaks May Increase Water Yield from the Grassland Islands in Arizona J. R. Thompson, O. D. Knipe and Phil M. Johnson - V6 Geomorphic and Hydraulic Response of Rivers $ The D. B. Simons - V5 Geomorphic Features Affecting Transmission Loss Potential $ D. E. Wallace and L. J. Lane - V8 Geomorphic Thresholds and Their Influence on Surface Runoff from Small Semiarid Watersheds$ D. E. Wallace and L. J. Lane - V6 Goals for Floodplain Management Within Pima County, Arizona$ Present Practices and Future Brian M. Reich and MichaelE. Zeller - P2 Goals in Arizona and Oregon $ A Public Weighting of Four Societal D. B. Kimball, R. L. Gum and T. G. Roefs - V3 Ground Water in the Santa Cruz Valley $ Marshall Flug - V9 Ground Water Recharge$ Seasonal and Spatial Trends of Ephemeral Flow in the Tucson Basin: Implications

for . . . Susan J. Keith - V10 Ground Water Systems with Analog Models (Abstract) $ Simulation of E. P. Patten - V3 Ground $ An Exchange System for Precise Measurements of Temperature and Humidity Gradients in the Air

Near the . . . L. W. Gay and L. J. Fritschen - V9 Ground- Recharged Water $ Organic Pollutants in MichaelA. Mikita, Kevin Thorn, James Hobson, Suzanne Lo and Cornelius Steelink - V11 Ground -Water Basins $ Uncertainties in Digital- Computer Modeling of Joseph S.Gates and Chester C.Kisiel - V1 Ground -Water Recharge $ Renovating Sewage Effluent by Herman Bouwer, J. C. Lance and R. C. Rice - V1 Ground -Water Reform: Forces and Consequences of Change in State Water Policy$ Arizona Floyd L.Marsh and Scott A.Hansen - V11 Groundwater Characteristics from Well Driller Logs$ Computerized Depth Interval Determination of Mike Long and Stephen Erb - V10 Groundwater Contamination in the Cortaro Area, Pima County, Arizona $ Kenneth D. Schmidt - V2 Groundwater Exploration in Northeastern Arizona Using LANDSAT Imagery $ Kennith E. Foster and K. James DeCook - V10 Groundwater from Municipal Wells, Flagstaff, Arizona $ Chemistry of Effervescing John C. Germ and ErrolL. Montgomery - V5 Groundwater Geology of Fort Valley, Coconino County, Arizona$ Ronald H. DeWitt - V3 Groundwater in the Basin and Range Province of Arizona$ The Occurrence of Thermal Jerome J. Wright - V1 Groundwater in the Parker Strip $ The Effect of Development on L. G. Everett and T. R. Schultz - V4 Groundwater Law Reform - An Urban Perspective$ Arizona H. Holub - V8 Groundwater Management for the City of Tucson, Arizona$ Hydrologic Factors Affecting R. B. Johnson - V8 Groundwater of the Tulare Lake Basin, California $ Salt Balance in Kenneth D. Schmidt - V5 Groundwater Quality Control $ Management of Artificial Recharge Wells for L. G. Wilson -V1 Groundwater Quality Protection Strategy for Arizona $ Toward Development of a Marc M. Bennett and Larry K. Stephenson - V11 Groundwater Quality Specialists $ Academic Training for Kenneth D. Schmidt - V6 Groundwater Quality Studies $ The Use of Chemical Hydrographs in Kenneth D. Schmidt - V1 Groundwater Recharge From a Portion of the Santa Catalina Mountains$ R. A. Bel an and W. G. Matlock - V3 Groundwater Requirements in the Tucson Area, Arizona $ Jojoba: An Alternative to the Conflict Between

Agricultural and Municipal . Kennith E. Foster and N.Gene Wright -P1 Groundwater Resource $ A Method for Maximizing the Present Value of a Reuben N. Weisz and Charles L. Towle, Jr.- V6 Groundwater Supply of Little Chino Valley $ The W. G. Matlock and P. R. Davis - V2 Groundwater Usage from the Navajo Sandstone $ Competitive F. H. Dove and T. G. Roefs - V3

153 Groundwaters in Arizona $ Factors to Consider in Drafting Standards to Protect Marc M. Bennett - V10 Groundwater f Management Alternatives for Santa Cruz Basin K. E. Foster - V8 Groundwater S Three -Dimensional Velocity Log in Richard P. Chagnon - VIO Hydrology (Abstract) S The Significance of Logistics to J. R. McCarthy - V3 Hydrology as a Science? S M. J. Dvoracek and D. D. Evans - V2 Hydrology and Erosion of a Semiarid Southwestern Rangeland WatershedS Effects of Brush to Grass

Conversion on the . . . J. R. Simanton, H. B. Osborn and K. G. Renard - V7 Hydrologic Aspects of Land -Use Planning at Tumamoc Hill, Tucson, ArizonaS Barney Paul Popkin - V4 Hydrologic Data in Arizona S Role of Modern Methods of Data Analysis for Interpretation of Chester C. Kisiel, Lucien Duckstein and Martin M. Fogel - V2 Hydrologic Effects of Soil Surface Micro -Flora $ William F. Faust - V1 Hydrologic Evaluation of Topsoiling for Rehabilitating Black Mesa Coal Mine Lands S Frank G. Postillion - V10 Hydrologic Factors Affecting Groundwater Management for the City of Tucson, Arizona$ R. B. Johnson - V8 Hydrologic Investigations of the Dry Lake Region in East Central Arizo S James J. Lemmon, Thomas R. Schultz and Don W. Young - V9 Hydrologic Properties of Diatremes in the Hopi Buttes, ArizonaS Preliminary Investigation of the Kenneth C. Scott, R. J. Edmonds and E. L. Montgomery - V4 Hydrologic Regimes of Three Vegetation Types Across the Mogollon RimS Malchus B. Baker, Jr.- Vii Hydrologic Research S Lake Powell Research Project: Gordon C. Jacoby - V3 Hydrologic Research on Southwest RangelandsS Use of Stock Ponds for J. R. Simanton and H. B. Osborn - V3 Hydrologic Time Series S On the Statistics of Sidney Yakowitz and Jack Denny - V3 Hydrologic Tracers, A New Technology $ Chlorofluorocarbons as J. H. Randall and T. R. Schultz -V6 Hydrological Design $ Using Linear Regression in G. D. Peterson, D. R. Davis and J. Weber - V4 Hydrology: State -of- the -Art $ Urban Warren Viessman, Jr. - V3 Infiltration and Runoff $ Simple Time -Power Functions for Rainwater R. M. Dixon, J. R. Simanton and L. J. Lane - V8 Infiltration Characteristics and Water Yield of a Semiarid Catchment$ Variability of Soronadi Nnaji, Ted W. Sammis and DanielD. Evans - V5 Infiltration Characteristics of a Ponderosa Pine Soil S Effects of a Wetting Agent on the Marc G. Kaplan and Malcolm J. Zwolinski - V3 Infiltration Control $ A Land Imprinter for Revegetation of Barren Land Areas Through R. M. Dixon and J. R. Si manton - V7 Infiltration in Soil Columns $ The Effect of Increasing the Organic Carbon Content of Sewage on Nitrogen, Carbon, and Bacteria Removal and ... J. L. Lance and F. D. Whisler - V5 Infiltration Rates and Denitrification During Sewage Effluent RechargeS Effect of Illuviated Deposits

on . Richard Alan Herbert - V7 Infiltration Rates in a Ponderosa Pine Stand á Effects of Fire on Water Malcom J. Zwolinski - V1 Infiltration Rates of Contributing Watersheds to the Lower Gila Below S The Use of a Realistic Rainfall

Simulator to Determine Relative . C. B. Cluff and D. G. Boyer - V1 Infiltration Response to Surface Plant Cover and Soil Invertebrate Population $ IsobelR. McGowan - V10 Infiltration $ A Microroughness Meter for Evaluating Rainwater J. R. Simanton, R. M. Dixon and I. McGowan - V8 Information Sources: Highlights $ Water -Related Linda M. White - V6 Information System - 1975 $ The Arizona Resources Carl C. Winikka - V5 Information System for Water Yield Improvement Practices$ A Bibliographic Linda M. White - V4 Information System$ Design and Pilot Study of an Arizona Water K. E. Foster and J. D. Johnson - V2 Irrigated Alfalfa S Energy Budget Measurements Over L. W. Gay and R. K. Hartman - V11 Irrigated Farm with Limited Water S Planning Models of an Herbert G. Blank - V6

154 Irrigation (Poster Session) $ Augmenting Water Supply for Home Barney P. Popkin - V9 Irrigation and Mining Water $ Collective Utility of Exchanging Treated Sewage Effluent for Stephen C.Ko and Lucien Duckstein - V2 Irrigation Management and Water Policy: Opportunities to Conserve Water in Arizona$ Harry W. Ayer and PaulG. Hoyt - V10 Irrigation Pumping Equipment $ Solar Powered Dennis L. Larson and C. D. Sands, Jr. - V9 Irrigation Pumping $ Feasibility of Using Solar Energy for Dennis Larson, D.D. Fangmeier, W.G. Matlock, John Day and C. D. Sands II- V6 Irrigation Water Exchange $ Metropolitan Operated District for Sewage Effluent - C. Brent Cluff and K. James DeCook - V4 Irrigation Water Quality $ Sulfuric Acid: Its Potential for Improving H. L. Bohn and R. L. Westerman - V1 Irrigation Water Source: Pinal County, Arizona and the Central Arizona Project $ Economic Adjustment to

a New . Mark A. Boster and William E.Martin - V5 Irrigation $ Conserving Water and Energy in J. A. Replogle and D. D. Fangmeier - P1 Irrigation $ Optimal Utilization of Playa Lake Water in M. J. Dvoracek and T. G. Roefs -V1 Irrigation $ Seasonal Effects on Soil Drying After B. A. Kimball and R. D. Jackson -V1 Lysimeter$ Construction, Calibration, and Operation of a Monolith Weighing Theodore W. Sammis, Don W. Young and Charles L. Constant - V6 Lysimeter Snowmelt in Arizona Ponderosa Pine Forests $ MikealE. Jones, Peter F. Ffolliott and David B. Thorud - V6 Lysimeters in an Arizona Mixed Conifer Stand $ An Evaluation of Snowmelt Gerald J. Gottfried and Peter F.Ffolliott - V10 Management Alternatives for Santa Cruz Basin Groundwater $ K. E. Foster - V8 Management Analysis$ Canyon Creek L. E. Siverts, R. D. Gale and J. W. Russell - V4 Management for the City of Tucson, Arizona$ Hydrologic Factors Affecting Groundwater R. B. Johnson - V8 Management for the San Carlos Apache Indian Reservation: Gila River$ Application of Multidisciplinary Water Resources Planning and ... M. E. Norvelle, D. J. Pervious and N. G. Wright -V7 Management in Relation to Industrial Development $ Storm Flows Robert. E. Smith - P2 Management in Sonora, Mexico$ Traditional Technology for Floodplain Thomas E.Sheridan and Gary P.Nabhan - P2 Management Plan for Tucson, Arizona $ Impacts of a New Water Resources R. Bruce Johnson - V10 Management Problems: Kassandra and the Sirens $ Two Flash- Flood Susan Zevin and Jose Marrero - P2 Management Systems for the Sonoita Creek Watershed$ Evaluation of Water Hugh B. Robotham - V9 Management Versus the Reasonable Man Test$ The Prejudices, Polemics, and Politics of Water Barbara A. Stribling - V6 Management Within Pima County, Arizona $ Present Practices and Future Goals for Floodplain Brian M. Reich and MichaelE. Zeller - P2 Management $ Evaluating and Displaying Watershed Tradeoffs for Rhey M. Solomon and Larry J. Schmidt - V10 Management $ Land Treatment for Urban Waste Water William L. Lorah and Kenneth R. Wright - V3 Management $ Tucson's Tools for Demand S. T. Davis -V8 Mexico Colorado River Water Salinity Problem $ Economic Alternatives in Solving the U.S. - William E. Martin - V4 Mexico Water Agreements and Related Water Use in Mexicali Valley: A S$ United States - K. J. DeCook - V4 Mexico $ A Study of Salinity in Effluent Lakes, Puerto Penasco, Sonora, Alison L. Dunn - V11 Mexico $ Exploration for Saltwater Supply for Shrimp Aquaculture, Puerto Penasco, Sonora, K. J. DeCook, S. Ince, B.P. Popkin, J. F. Shreiber, Jr. and J. S. Sumner - V10 Mexico $ Traditional Technology for Floodplain Management in Sonora, Thomas E.Sheridan and Gary P.Nabhan - P2 Mine Lands of Semiarid Southwest $ Soil Erosion and Sediment Control on the Reclaimed Coal Tika R. Verma, John L. Thames and John E. Mills - V7 Mine Lands$ Optimal Livestock Production of Rehabilitated Fritz H. Brinck, Martin M. Fogel and Lucien Duckstein - V6 Mine Reclamation Alternative $ The Mound and Valley Water Harvesting System: A Potential Charles L. Constant and John Thames - V10

155 Mine Sites and Their Effect on the Water Quality of the Lynx Creek WatershedS Reclamation of Orphaned Tika R. Verma and Ernesto N.Felix - V7 Mine Spoils of the Arid Southwest$ Stochastic Prediction of Sediment Yields from Strip Mark E. Auernhamer, Martin M. Fogel, Louis H. Hekman, Jr. and John L. Thames - V7 Mine Tailing Slopes Using Municipal Sewage Effluent$ Rehabilitation of Copper Tika R. Verma, Kenneth L. Ludeke and A. D. Day - V7 Mines in the Patagonia Mountains, Arizona $ Acid Drainage From Abandoned Sheila A. Dean - P3 Mined Lands in Northern Arizona S Water Quality of Runoff From Surface J. Kempf, L. Leonhart, M. Fogel and L. Duckstein - V8 Mining Activity in the Miami, Arizona Region$ The Effects on Water Quality by D. W. Young and R. B. Clark - VB Mining Activities and Water Quality in the Lynx Creek Watershed $ Past E. N. Felix, T. R. Verma, E. E. McCrary and J. L. Thames - V6 Mining District of Arizona - An Institutional Approach$ Addressing Water Quality Problems in the Globe - Miami . Robert J. Cinq -Mars, Daniel R. Mayercek andEdwin K.Swanson - P3 Mining Effects on Water Quality of the Tongue River, Wyoming$ Preliminary Results from a Study of Coal Richard D. Olsen and Edward H. Dettmann - V6 Mining Operations on the Black Mesa$ Visual Impacts: Perception and Modification of Surface Jon Rodiek - V9 Mining Water $ Collective Utility of Exchanging Treated Sewage Effluent for Irrigation and Stephen C.Ko and Lucien Duckstein - V2 Model for Flood Routing in Abstracting Ephemeral Channels$ A Proposed Leonard J. Lane - V2 Model for Semi -Arid Catchments $ A Deterministic S. Nnaji, D. R. Davis and M. M. Fogel - V4 Model of Suspended Sediment Yield on Forested Watersheds in Central Arizona$ An Interactive William O. Rasmussen and Peter F. Ffolliott - V9 Model System $ The Arizona Water Commission's Central Arizona Project Water Allocation Philip C. Briggs - V7 Model to Predict Water Quality Transformations During Subsurface Movement of Oxidation Pond Effluent

s The Use of a Computer . . . G. G. Small and L. G. Wilson - V3 Model $ A Sediment Yield Equation From an Erosion Simulation E. D. Shirley and L. J. Lane - V8 Models (Abstract) $ How to Select Evapotranspiration T. E. A. van Hylckama, R. M. Turner and O. M. Grasz - V9 Models (Abstract) $ Simulation of Ground Water Systems with Analog E. P. Patten -V3 Models and Methods for Rivers in the Southwest $ Statistical Sidney Yakowitz - V7 Model S Input Specifications to a Stochastic Decision D. M. Cl ai nos, L. Duckstein and T. G. Roefs - V2 Models of an Irrigated Farm with Limited Water$ Planning Herbert G. Blank - V6 Model $ Prediction of the Chemical Quality of Streamflow by an Interactive Computer William O. Rasmussen and Peter F. Ffolliott - V10 Modeling and Computational Aspects of Dynamic Programming with ApplicaReservoir Control$ On the Moshe Sniedovich and Sidney J. Yakowitz - V6 Modeling for Capital Improvements Planning$ Hydraulic Stephen E. Davis - V10 Modeling of a Forward Osmosis Extractor $ Mathematical C. D. Moody and J. O. Kessler - V6 Modeling of Ground -Water Basins $ Uncertainties in Digital- Computer Joseph S.Gates and Chester C.Kisiel- V1 Modeling of the Avra Valley Aquifer $ Geostatistical Analysis and Inverse Peter M. Clifton and Shlomo P. Neuman - Vii Modeling $ Parameter Influence on Runoff Samual E. Kao, Theodore G. Roefs and Simon Ince - V5 Modeling $ Solar Radiation as Indexed by Clouds for Snowmelt D. P. McAda and P. F. Ffolliott - V8 Multiattribute Approach to the Reclamation of Stripmined Lands$ A Fritz H. Brinck, Lucien Duckstein and John L. Thames - V9 Multidisciplinary Water Resources Planning and Management for the San Carlos Apache Indian Reservation: G i l a River $ An Application of ... M. E. Norvelle, D. J. Percious and N. G. Wright -V7 Multiobjective Approach to Managing a Southern Arizona Watershed$ A Ambrose Goicoechea, Lucien Duckstein and Martin M. Fogel - V6 Multi- Objective Approach to River Basin Planning $ A Mark Gershon, Richard McAniff and Lucien Duckstein - V10 Multiple -use Approach to the Reclamation of Strip- mined Lands$ Decision Making in a Ambrose Goicoechea, Lucien Duckstein and Martin Fogel - V7 Municipal Groundwater Requirements in the Tucson Area, Arizona $ Jojoba: An Alternative to the Conflict Between Agricultural and ... Kennith E. Foster and N.Gene Wright - P1

156 Municipal Use in the Colorado Basin: A Consumer- Environmental Perspective$ Water for Food, Energy and Barbara Tellman - V6 Municipal Waste Water $ Evaluation of a Turfgrass - Soil System to Utilize and Purify R. C. Sidle and G. V. Johnson - V2 Municipal Waste -Water Utilization in Arizona $ Economic and Energy Opportunities for E. J. Weber -P3 Municipal Wells $ Barometric Response of Water Levels in Flagstaff ErrolL. Montgomery, Emily Dordosz, Russell O. Dalton, Jr. and Ronald H. DeWitt - V7 Municipal Wells, Flagstaff, Arizona$ Chemistry of Effervescing Groundwater from John C. Germ and ErrolL. Montgomery - V5 Nitrogen Balance for a 23- Square Mile Minnesota Watershed $ Jack D. Johnson - V1 Nitrogen Removal from Secondary Effluent Applied to a Soil -Turf Filter $ E. L. Anderson, I. L. Pepper and G. V. Johnson - V8 Nitrogen Species Transformations of Sewage Effluent Releases in A Desert Stream Channel $ P. G. Sebenik, C. B. Cl uff and K. J. DeCook - V2 Nitrogen, Carbon, and Bacteria Removal and Infiltration in Soil Columns$ The Effect of Increasing the

Organic Carbon Content of Sewage on . . . J. L. Lance and F. D. Whisler -V5 Nutrient Levels on the Verde River Watershed with Recommended Standard $ Timothy D. Love - V11 Osmosis Extractor $ Mathematical Modeling of a Forward C. D. Moody and J. O. Kessler - V6 Osmosis: Design Characteristics for Hydration and Dehydration$ Application of Direct J. O. Kessler and C. D. Moody -V5 Osmosis: Possibilities for Reclaiming Wellton- Mohawk Drainage Water $ Application of Direct C. D. Moody and J. O. Kessler -V5 Pine Forest $ A Preliminary Assessment of Snowfall Interception in Arizona Ponderosa Larry C. Tennyson, Peter F. Ffolliott and David B. Thorud - V3 Pine Soil $ Effects of a Wetting Agent on the Infiltration Characteristics of a Ponderosa Marc G. Kaplan and Malcolm J. Zwolinski - V3 Pine Stand $ Effects of Fire on Water Infiltration Rates in a Ponderosa Malcom J. Zwolinski - V1 Plan - Tucson, Arizona$ The Northwest Area Water Thomas M. McLean - V10 Plan for Tucson, Arizona$ Impacts of a New Water Resources Management R. Bruce Johnson - V10 Plant Cover and Soil Invertebrate Population $ Infiltration Response to Surface IsobelR. McGowan - V10 Plans for the Santa Cruz River Basin by Q- Analysis $ Ranking Alternative Ronald T. Pfaff and Lucien Duckstein - V11 Planning $ A Multi- Objective Approach to River Basin Mark Gershon, Richard McAniff and Lucien Duckstein - V10 Planning at Tumamoc Hill, Tucson, Arizona $ Hydrologic Aspects of Land -Use Barney Paul Popkin - V4 Planning and Management for the San Carlos Apache Indian Reservation: Gila River$ An Application of Multidisciplinary Water Resources . .. M. E. Norvelle, D. J. Percious and N. G. Wright -V7 Planning for the San Tiburcio Watershed $ Land Use Roberto Armijo and Robert Bulfin - V9 Planning in the Colorado River System $ Salinity Control John T. Maletic - V4 Planning Methodology: Public Input $ The Cognitive Strawman Weston W. Wilson, RussellL. Gum and T.G. Roefs - V3 Planning Models of an Irrigated Farm with Limited Water $ Herbert G. Blank - V6 Planning Tool $ Public Perception of Water Quality as a R. M. Judge and R.L. Gum -V3 Planning $ Establishing a Process Framework for Land Use Lloyd J. Lundeen - V4 Planning $ Hydraulic Modeling for Capital Improvements Stephen E. Davis - V10 Planning$ State Water Wesley E. Steiner - V5 Planning $ The Role of Conservation in Water Supply Augustine J. Fredrich - P1 Planning$ Wastewater Effluent - An Element of Total Water Resource J. D. Goff -V8 Plants: A Possible Paleohygrometer $ Stable Isotopes of Oxygen in A. M. Ferhi, A. Long and J. C. Lerman - V7 Policy for Desert Cities $ A Rational Water W.G. Matlock - V4 Policy $ Arizona Ground -Water Reform: Forces and Consequences of Change in State Water Floyd L.Marsh and Scott A.Hansen - V11

157 Policy: Changing Decision Agendas and Political Styles S Arizona Water Hanna J. Cortner and Mary P. Berry - V7 Policy: Opportunities to Conserve Water in Arizona $ Irrigation Management and Water Harry W. Ayer and PaulG. Hoyt - V10 Policy $ Water Conservation in the President's Water Gary D. Cobb - P1 Politics and the Colorado River S Wesley E. Steiner - V1 Politics of Water in Arizona S Senator Morris Farr - V6 Politics of Water Management Versus the Reasonable Man TestS The Prejudices, Polemics, and Barbara A. Stribling - V6 Political Constraints to ImplementationS Action Programs for Water Yield Improvement on Arizona's

Watersheds: . . H. J. Cortner and M. P. Berry - V8 Political Styles S Arizona Water Policy: Changing Decision Agendas and Hanna J. Cortner and Mary P.Berry - V7 Pollutants Determine Runoff Source Areas S Nonpoint- Source L. J. Lane, H. L. Morton, D. E. Wallace, R. E. Wilson and R. D. Martin - V7 Pollutants in Ground- Recharged Water S Organic Michael A. Mikita, Kevin Thorn, James Hobson, Suzanne Lo and Cornelius Steelink - V11 Pollution Control Program S Highlights of the State's Water Marc M. Bennett and Larry K. Stephenson - P3 Pollution S Saline and Organic Water Hinrich L.Bohn and Gordon V.Johnson - V2 Precipitation in Lake Powell (Abstract) S Calcite Robert C. Reynolds, Jr.- V3 Precipitation in the Basin and Range Province of Southern Arizona - So S Determining Areal J. Ben- Asher, J. Randall andS. Resnick - V6 Precipitation on a Southeastern Arizona Rangeland Watershed S Winter H. B. Osborn, R. B. Koehler and J. R. Simanton - V9 Precipitation on Rugged Terrain in Central Arizona S Distribution of Alden R. Hibbert - V7 Price of Water in Western Agriculture S The David L. Wilson and Harry W. Ayer - V11 Prices, Water Demand by Peri -Urban Agriculture, and Implications for Urban Water Supply: The Tucson

Case$ Rising Energy . . . H. W. Ayer and D. W. Gapp - V8 Pricing Structures from a Collective Utility Viewpoint $ Comparison of Water Bill Metter and Lucien Duckstein - V1 Probability Distribution for Rainfall on a Watershed by Simulation$ The Construction of a Gary Williamson and Donald Ross Davis - V2 Probability Distributions of Snow Course Data for Central ArizonaS Lawrence E.Cary and Robert L.Beschta - V3 Probabilities of Monthly Flow on an Ephemeral Stream S Objective and Subjective Analysis of Transition William Dvoranchik, Lucien Duckstein and Chester C. Kisiel - V2 Probabilities $ Conditional Streamflow T. G. Roefs and D. M. Clainos -V1 Public Input S The Cognitive Strawman Planning Methodology: Weston W. Wilson, RussellL. Gum and T.G. Roefs - V3 Public Involvement in Federal Water Resource Projects S Early Freda Johnson and Michael Thuss - V9 Public Perception of Water Quality as a Planning Tool $ R.M. Judge and R.L. Gum - V3 Public Weighting of Four Societal Goals in Arizona and Oregon S A D.B. Kimball, R. L. Gum and T. G. Roefs - V3 Pumping Equipment$ Solar Powered Irrigation Dennis L. Larson and C. D. Sands, Jr. - V9 Pumping Project: Operating Experiences $ Arizona Solar Powered Dennis L. Larson - Vii Pumping Tests for Determining Well Losses in Consolidated Rock Aquifers S The Application of Step -

Drawdown . V. W. Uhl, Jr., V. G. Joshi, A. Alpheus and G. Sharma - V5 Pumping S Feasibility of Using Solar Energy for Irrigation Dennis Larson, D.D. Fangmeier, W.G. Matlock, John Day and C. D. Sands II- V6 Quality as a Planning Tool $ Public Perception of Water R. M. Judge and R.L. Gum -V3 Quality Analyses of the Colorado River Corridor of Grand Canyon S Water Brock Tunnicliff and Stanley K.Brickler - V10 Quality by Mining Activity in the Miami, Arizona RegionS The Effects on Water D. W. Young and R.B. Clark - V8 Quality Considerations for Landfill Siting in Arizona S Water L. G. Wilson - P3 Quality Control S Management of Artificial Recharge Wells for Groundwater L. G. Wilson - V1

158 Quality in Air Bubbling to Eliminate Thermal Stratification in Upper L S Improvement of Water E. H. McGavock and J. W. H. Blee - V3 Quality in Southeastern Arizona Rangeland $ Rainwater Herbert B.Osborn, Loel R. Cooper and Jeff Billings - V11 Quality in the Lynx Creek Watershed $ Past Mining Activities and Water E. N. Felix, T. R. Verma, E. E. McCrary and J. L. Thames - V6 Quality Investigation of Canyon LakeS A Bacterial Water W. F. Horak and G. S. Lehman - V4 Quality of Recharging Effluent in the Santa Cruz River S Transformations in L. G. Wilson, R. A. Herbert and C. R. Ramsey - V5 Quality of Runoff From Surface Mined Lands in Northern Arizona S Water J. Kempf, L. Leonhart, M. Fogel and L. Duckstein - V8 Quality of Stock Tanks of Southeastern Arizona $ Time -Related Changes in Water D. E. Wallace and H. A. Schreiber - V4 Quality of Streamflow by an Interactive Computer Model $ Prediction of the Chemical William O. Rasmussen and Peter F. Ffolliott - V10 Quality of Streamflow from Forested Watersheds on Sedimentary Soils S Water Paul W. Gregory and Peter F. Ffolliott - V6 Quality of the East Verde River $ Impact of Recreation on the Water Milton R. Sommerfeld, Patrick V. Athey and Bradley C. Mueller - P3 Quality of the East Verde River S Use of Bacterial Indicators in Assessment of Water Patrick V. Athery, Marilyn J. Urbina and Milton R. Sommerfeld - V11 Quality of the Lynx Creek Watershed$ Reclamation of Orphaned Mine Sites and Their Effect on the Water Tika R.Verma and Ernesto N. Felix - V7 Quality of the Tongue River, WyomingS Preliminary Results from a Study of Coal Mining Effects on Water Richard D. Olsen and Edward H. Dettmann - V6 Quality Problems in the Globe -Miami Mining District of Arizona - An Institutional Approach $ Addressing Robert J. Cinq -Mars, Daniel R. Mayercek and Edwin K. Swanson - P3 Water . . . Quality Problems in the Puerco River, Arizona $ Water Edwin K. Swanson - P3 Quality Problems of the Urban Area in an Arid Environment. Tucson, Arizona S Water G. Hansen - V8 Quality Problems: A Case Study of the San Pedro River in Southeastern S Ephemeral Flow and Water S. J. Keith -V8 Quality Protection Strategy for Arizona S Toward Development of a Groundwater Marc M. Bennett and Larry K. Stephenson - V11 Quality Sampling Schedules Using Fecal Coliform Concentrations in Sabi S Evaluating Water Robert M. Motschall, Stanley D. Brickler and Robert A. Phillips - V6 Quality Specialists $ Academic Training for Groundwater Kenneth D. Schmidt - V6 Quality Study of Lake Havasu, Arizona near the CAP Intake Area $ Water Simon Ince, David L. Kreamer, Don W. Young and Charles L. Constant - V6 Quality Studies Today and Tomorrow $ Water - John D. Hem - V3 Quality Studies $ The Use of Chemical Hydrographs in Groundwater Kenneth D. Schmidt - V1 Quality Transformations During Subsurface Movement of Oxidation Pond Effluent$ The Use of a Computer

Model to Predict Water . . . G. G. Small and L. G. Wilson - V3 Quality $ Some Effects of Controlled Burning on Surface Water Bruce D. Sims, Gordon S. Lehman and Peter F. Ffolliott - V11 Quality $ Sulfuric Acid: Its Potential for Improving Irrigation Water H. L. Bohn and R. L. Westerman -V1 Rain Gage $ Development and Testing of a Laser Arnold D. Ozment - V5 Rainfall in Southeastern Arizona $ Point- Area -Frequency Conversions for Summer Herbert B.Osborn and Leonard J.Lane - V11 Rainfall in the Southwest$ Regional Differences in Runoff -Producing Thunderstorms H. B. Osborn -V1 Rainfall in the Southwest$ Stationarity in Thunderstorm William C. Mills and Herbert B. Osborn - V3 Rainfall Intensity on Runoff Curve NumbersS Effects of R. H. Hawkins - V8 Rainfall on a Watershed by Simulation $ The Construction of a Probability Distribution for Gary Williamson and Donald Ross Davis - V2 Rainfall Occurrence in Arizona and New Mexico S Simulation of Summer Herbert B.Osborn and Donald Ross Davis - V7 Rainfall Records Relative to Chart Scales $ Resolutions of Analog Donald L. Chery, Jr. and Dave G. Beaver - V6 Rainfall Simulating Infiltrometer$ A Jeep- Mounted William R. Henkle - V3 Rainfall Simulator to Determine Relative Infiltration Rates of Contributing Watersheds to the Lower Gila Below $ The Use of a Realistic ... C. B. Cluff and D. G. Boyer - V1

159 Rainfall- Erosion Index of the Universal Soil Loss Equation$ Thunderstorm Precipitation Effects on the Kenneth G. Renard and J. Roger Simanton - V5 Rainfall- Runoff Relation S Significance of Antecedent Soil Moisture to a Semiarid Watershed D. L. Chery, Jr. - V2 Rainfall- Runoff Relationships for a Mountain Watershed in Southern Arizona$ M. Myhrman, C. B. Cl uff and F. Putman - V8 Raingage Networks $ Use of Radar as a Supplement to Herbert B.Osborn and J.Roger Simanton - V10 Rainwater Infiltration and Runoff$ Simple Time -Power Functions for

. M. Dixon, J. R. Simanton and L. J. Lane - V8 Rainwater Infiltration $ A Microroughness Meter for Evaluating J. R. Simanton, R. M. Dixon and I. McGowan - V8 Rainwater Quality in Southeastern Arizona Rangeland S Herbert B. Osborn, LoelR. Cooper and Jeff Billings - V11 Rangeland Conditions in the SouthwestS Applicability of the Universal Soil Loss Equation to Semiarid K. G. Renard, J. R. Si manton and H. B. Osborn - V4 Rangeland WatershedS Increasing Forage Production on a Semiarid J. M. Tromble - V4 Rangeland WatershedS Winter Precipitation on a Southeastern Arizona H. B. Osborn, R. B. Koehler and J. R. Simanton - V9 Rangelands in Arizona $ Water Resources Research on Forest and Alden R. Hibbert - V4 Rangeland S Rainwater Quality in Southeastern Arizona Herbert B.Osborn, LoelR. Cooper and Jeff Billings - V11 Rangelands $ Application of the USLE to Southwestern J. Roger Simanton, Herbert B. Osborn and Kenneth G. Renard - V10 Rangelands S Use of Stock Ponds for Hydrologic Research on Southwest J. R. Simanton and H. B. Osborn - V3 Recharge S Effect of Illuviated Deposits on Infiltration Rates and Denitrification During Sewage

Effluent . Richard Alan Herbert - V7 Recharge From a Portion of the Santa Catalina Mountains $ Groundwater R. A. Belan and W. G. Matlock - V3 Recharge Through Soils in a Mountain Region: A Case Study on the Empire and the Sonoit$ Evaluation of U. Kafri and J. Ben -Asher - V6 Recharge Wells for Groundwater Quality Control S Management of Artificial L. G. Wilson - V1 RechargeS Renovating Sewage Effluent by Ground -Water Herman Bouwer, J. C. Lance and R. C. Rice - Y1 RechargeS Seasonal and Spatial Trends of Ephemeral Flow in the Tucson Basin: Implications for Ground Water . Susan J.Keith - V10 Recharged Water $ Organic Pollutants in Ground - Michael A. Mikita, Kevin Thorn, James Hobson, Suzanne Lo and Cornelius Steelink - V11 Recharging Effluent in the Santa Cruz River $ Transformations in Quality of L. G. Wilson, R. A. Herbert and C. R. Ramsey - Y5 Recharging the Ogallala Formation Using Shallow Holes $ M. J. Dvoracek and S. H. Peterson - V1 Reclaiming Wellton- Mohawk Drainage Water $ Application of Direct Osmosis: Possibilities for C. D. Moody and J. O. Kessler - V5 Reclamation Alternative $ The Mound and Valley Water Harvesting System: A Potential Mine Charles L. Constant and John Thames - V10 Reclamation of Orphaned Mine Sites and Their Effect on the Water Quality of the Lynx Creek Watershed$ Tika R. Verma and Ernesto N. Felix - V7 Reclamation of Strip- mined Lands$ Decision Making in a Multiple -use Approach to the Ambrose Goicoechea, Lucien Duckstein and Martin Fogel - V7 Reclamation of Stripmined Lands$ A Multiattribute Approach to the Fritz H. Brinck, Lucien Duckstein and John L. Thames - V9 Recreation on the Water Quality of the East Verde River $ Impact of Milton R. Sommerfeld, Patrick V. Athey and Bradley C. Mueller - P3 Recreational Waters of Upper Sabino Creek $ Bottom Sediment Analysis of the Patrick L.McKee and Stanley K. Brickler - V7 Regression in Hydrological Design $ Using Linear G. D. Peterson, D. R. Davis and J. Weber - V4 Rehabilitated Mine Lands $ Optimal Livestock Production of Fritz H. Brinck, Martin M. Fogel and Lucien Duckstein - V6 Rehabilitating Black Mesa Coal Mine Lands $ Hydrologic Evaluation of Topsoiling for Frank G. Postillion - V10 Rehabilitation of Copper Mine Tailing Slopes Using Municipal Sewage Effluent$ Tika R. Verma, Kenneth L. Ludeke and A. D. Day - V7 Reservoir Control S On the Modeling and Computational Aspects of Dynamic Programming with Applica Moshe Sniedovich and Sidney J. Yakowitz - V6 Reservoir Design under Random Sediment Yield$ L. Duckstein, F. Szidarovszky and S. Yakowitz - V6

160 Reservoir Operation $ A Utility Criterion for Real -time Lucien Duckstein and Roman Krysztofowicz - V7 Reservoir: Efficient Water Storage in Flat Terrain Areas of Arizona $ The Compartmented C. B. Cluff -V8 Residential Water Use: A Cross -Section Time -Series Analysis of Tucson, Arizona $ The Impact of Socioeconomic Status on ... R.Bruce Billings and Donald E.Agthe - V9 Revegetation of Barren Land Areas Through Infiltration Control $ A Land Imprinter for R. M. Dixon and J. R. Si manton -V7 Runoff Curve Numbers $ Effects of Rainfall Intensity on R. H. Hawkins - V8 Runoff Development$ Legal Aspects of Urban D. A. Chudnoff - V8 Runoff Forecasts in Arizona $ Use of Satellite Data to Develop Snowmelt- Peter F. Ffolliott and William O. Rasmussen - V6 Runoff from Small Semiarid Watersheds $ Geomorphic Thresholds and Their Influence on Surface D. E. Wallace and L. J. Lane -V6 Runoff from Small Watersheds$ Effect of Urbanization on Samuel E. Kao, Martin M. Fogel and SolD. Resnick - V3 Runoff From Surface Mined Lands in Northern Arizona $ Water Quality of J. Kempf, L. Leonhart, M. Fogel and L. Duckstei n - V8 Runoff Modeling s Parameter Influence on Samual E.Kao, Theodore G. Roefs and Simon Ince - V5 Runoff Records Using Tree -Ring Data $ Augmenting Annual Charles W.Stockton and Harold C. Fritts - V1 Runoff Relation $ Significance of Antecedent Soil Moisture to a Semiarid Watershed Rainfall - D. L. Chery, Jr. - V2 Runoff Relationships for a Mountain Watershed in Southern ArizonaS Rainfall - M. Myhrman, C. B. Cl uff and F. Putman - V8 Runoff Source Areas $ Nonpoint- Source Pollutants Determine L. J. Lane, H. L. Morton, D. E. Wallace, R. E. Wilson and R. D. Martin - V7 Runoff: A Preliminary Evaluation $ Effect of a Grass and Soil Filter on Tucson Urban Barney Paul Popkin - V2 Runoff? $ Some Legal Problems of Urban Hugh Holub - V2 Runoff $ Simple Time -Power Functions for Rainwater Infiltration and R. M. Dixon, J. R. Simanton and L. J. Lane - V8 Runoff -Producing Thunderstorms Rainfall in the Southwest$ Regional Differences in H. B. Osborn -V1 Safe Drinking Water Act on Arizona's Water Systems $ Socio- Economic Impacts of the Richard S. Williamson - V10 Salt Balance in Groundwater of the Tulare Lake Basin, California $ Kenneth D. Schmidt - V5 Salt Replacement Desalination$ Fresh Water for Arizona by Anthony B. Muller - V4 Salt Treatment $ Water Harvesting: Soil /Water Impacts of Albert Todd - V10 Saline and Organic Water Pollution $ Hinrich L.Bohn and Gordon V. Johnson - V2 Salinity Control Planning in the Colorado River System $ John T. Maletic - V4 Salinity in Effluent Lakes, Puerto Penasco, Sonora, Mexico $ A Study of Alison L. Dunn - V11 Salinity Loading Relationships in the Lower Colorado River Basin $ Intermittent Flow Events - William Woessner - V10 Salinity Problem $ Economic Alternatives in Solving the U.S.- Mexico Colorado River Water William E. Martin - V4 Salinity Problems of the Safford Valley: An Interdisciplinary Analysis$ Anthony B. Muller - V3 Saltcedar Thickets $ Antitranspirants as a Possible Alternative to the Eradication of Robert S. Cunningham, Kenneth N. Brooks and David B. Thorud - V5 Saltcedar $ An Energy Budget Analysis of Evapotranspiration from L. W. Gay, T. W. Sammis and J. Ben -Asher - V6 Saltcedar $ Diurnal Trends in Water Status, Transpiration, and Photosyntheses of Mary Ellen Williams and Jay E. Anderson - V7 Saltcedar $ Transpiration and Photosynthesis in Jay E. Anderson - V7 Saltwater Supply for Shrimp Aquaculture, Puerto Penasco, Sonora, Mexico$ Exploration for K. J. DeCook, S. Ince, B. P. Popkin, J. F. Shreiber, Jr. and J. S. Sumner - V10 Sediment Analysis of the Recreational Waters of Upper Sabino Creek $ Bottom Patrick L.McKee and Stanley K. Brickler - V7 Sediment Control on the Reclaimed Coal Mine Lands of Semiarid Southwest$ Soil Erosion and Tika R. Verna, John L. Thames and John E. Mills - V7

161 Sediment Production from a Chaparral Watershed in Central Arizona$ Thomas E.Hook and Alden R. Hibbert - V9 Sediment Sources of Midwestern Surface Waters S Donovan C. Wilkin and Susan J.Hebel - V11 Sediment Transport in an Ephemeral Mountain StreamS Equilibrium Condition and Burchard H. Heede - V6 Sediment Yield Equation From an Erosion Simulation Model S A E. D. Shirley and L. J. Lane - V8 Sediment Yield from a Semi -Arid Watershed S Uncertainty in J. H. Smith, M. Fogel and L. Duckstein - V4 Sediment Yield on Forested Watersheds in Central ArizonaS An Interactive Model of Suspended William O. Rasmussen and Peter F. Ffolliott - V9 Sediment Yields from Strip Mine Spoils of the Arid Southwest$ Stochastic Prediction of Mark E. Auernhamer, Martin M. Fogel, Louis H. Hekman, Jr. and John L. Thames - V7 Sediment Yield S Reservoir Design under Random L. Duckstein, F. Szidarovszky and S. Yakowitz - V6 Sediments and Soils S Microtrac: A Rapid Particle -Size Analyzer of R. L. Haverl and and L. R. Cooper - V11 Sedimentation in the Upper Gila Drainage, A Case Study S Erosion and R. L.. Kingston and R. M. Solomon - V6 Seepage Control S An Evaluation of Current Practices in D. G. Boyer and C. B. Cluff - V2 Semiarid Catchment S Variability of Infiltration Characteristics and Water Yield of a Soronadi Nnaji, Ted W. Sammis and Daniel D. Evans - V5 Semi -Arid CatchmentsS A Deterministic Model for S. Nnaji, D. R. Davis and M. M. Fogel - V4 Semi -Arid Watershed $ Uncertainty in Sediment Yield from a J. H. Smith, M. Fogel and L. Duckstein - V4 Semiarid Rangeland Conditions in the SouthwestS Applicability of the Universal Soil Loss Equation to K. G. Renard, J. R. Si manton and H. B. Osborn - V4 Semiarid Rangeland WatershedS Increasing Forage Production on a J. M. Tromble - V4 Semiarid Southwestern Rangeland Watershed S Effects of Brush to Grass Conversion on the Hydrology and Erosion of a ... J.R. Simanton, H. B. Osborn and K. G. Renard - V7 Semiarid Urbanized Watershed S The Effect of an Intensive Summer Thunderstorm on a D. G. Boyer and K. J. DeCook - V6 Semiarid WatershedS A Water Budget for a Severo R. Saplaco, Peter F. Ffolliott and William O. Rasmussen - V9 Semiarid Watershed S Application of a Double Triangle Unit Hydrograph to a Small M. H. Di ski n and L. J. Lane - V6 Semiarid WatershedS Simulation of Partial Area Response from a Small Leonard J. Lane and Delmer E. Wallace - V6 Semiarid Watersheds$ Geomorphic Thresholds and Their Influence on Surface Runoff from Small D. E. Wallace and L. J. Lane - V6 Simulating Infiltrometer$ A Jeep- Mounted Rainfall William R. Henkle - V3 Simulation of Ground Water Systems with Analog Models (Abstract) $ E. P. Patten -V3 Simulation of Partial Area Response from a Small Semiarid WatershedS Leonard J. Lane and Delmer E. Wallace - V6 Simulation of Summer Rainfall Occurrence in Arizona and New Mexico$ Herbert B.Osborn and Donald Ross Davis - V7 Simulation$ The Construction of a Probability Distribution for Rainfall on a Watershed by Gary Williamson and Donald Ross Davis - V2 Simulator to Determine Relative Infiltration Rates of Contributing Watersheds to the Lower Gila Below

$ The Use of a Realistic Rainfall . . . C. B. Cluff and D. G. Boyer - V1 Snow Course Data for Central ArizonaS Probability Distributions of Lawrence E. Cary and Robert L. Beschta - V3 Snow Course Data in Describing Local Snow Conditions in Arizona ForestsS Evaluation of the Use of Soil Conservation Service ... Gerald J. Gottfried and Peter F.Ffolliott - VII Snow Cover from ERTS Imagery on the Black River Basin S Measuring Jerry S.Aul and Peter F.Ffolliott - V5 Snowfall Interception in Arizona Ponderosa Pine Forest S A Preliminary Assessment of Larry C. Tennyson, Peter F. Ffolliott and David B. Thorud - V3 Snowmelt in Arizona Ponderosa Pine Forests $ Lysimeter MikealE. Jones, Peter F. Ffolliott and David B. Thorud - V6 Snowmelt Lysimeters in an Arizona Mixed Conifer StandS An Evaluation of Gerald J. Gottfried and Peter F.Ffolliott - V10 Snowmelt Modeling $ Solar Radiation as Indexed by Clouds for D. P. McAda and P. F. Ffolliott - V8 Snowmelt- Runoff Forecasts in Arizona S Use of Satellite Data to Develop Peter F. Ffolliott and William O. Rasmussen - V6

162 Snowpack Density on an Arizona Mixed Conifer Forest Watershed $ Peter F. Ffolliott and J.R. Thompson - V7 Snowpack Dynamics in Arizona's Aspen Forests S Michael J. Timmer, Peter F. Ffolliott and William O. Rasmussen - V10 Snowpack Inventory- Prediction Relationships S An Analysis of Yearly Differences in Peter F.Ffolliott, David B. Thorud and Richard W. Enz - V2 Snowpack Mapping $ Aerial William L. Warskow - V5 Snowpack Profiles in and Adjacent to Forest Openings$ A Technique to Evaluate Peter F. Ffolliott and David B.Thorud - V4 Snowpack Water YieldsS Progress in Developing Forest Management Guidelines for Increasing David B. Thorud and Peter F.Ffolliott - V1 Soil Columns Flooded with Sewage Water S Addition of a Carbon Pulse to Stimulate Denitrification in J. C. Lance and R. G. Gilbert - V6 Soil Drying After Irrigation $ Seasonal Effects on B. A. Kimball and R. D. Jackson - V1 Soil Evaporation Via Surface Temperature MeasurementsS Assessing Bare Sherwood B.Idso, Robert J. Reginato and Ray D. Jackson - V5 Soil Filter on Tucson Urban Runoff: A Preliminary Evaluation$ Effect of a Grass and Barney Paul Popkin - V2 Soil Flooded with Secondary Sewage Effluent S Effect of Algal Growth and Dissolved Oxygen on Redox Potentials in ... R. G. Gilbert and R. C. Rice - V8 Soil Invertebrate Population S Infiltration Response to Surface Plant Cover and IsobelR. McGowan - V10 Soil Moisture RemotelyS Assessing Robert J.Reginato, Sherwood B.Idso and Ray D. Jackson - V5 Soil Moisture to a Semiarid Watershed Rainfall- Runoff Relation S Significance of Antecedent D.L. Chery, Jr. - V2 Soil Moisture Under Natural Vegetation S Variations in T. W. Sammis and D. L. Weeks - V7 Soil Porosity Characteristics of Arizona Soils S Relationships of Soil Texture with Soil Water Content

and . . . Donald F. Post -V11 Soil Stabilizer $ Wax Water Harvesting Treatment Improved with Antistripping Agent and Dwayne H. Fink - V10 Soil Surface Micro -Flora S Hydrologic Effects of William F. Faust - V1 Soil System to Utilize and Purify Municipal Waste Water $ Evaluation of a Turfgrass - R. C. Sidle and G. V. Johnson - V2 Soil Texture with Soil Water Content and Soil Porosity Characteristics of Arizona SoilsS Relationships of ... Donald F. Post - V11 Soil $ Effects of a Wetting Agent on the Infiltration Characteristics of a Ponderosa Pine Marc G. Kaplan and Malcolm J. Zwolinski - V3 Soils for Water Harvesting S Candelilla /Petroleum Wax Mixtures for Treating Dwayne H. Fink - Vii Soils for Water Harvesting $ Laboratory Evaluation of Water -Repellent Dwayne H. Fink - V4 Soils in a Mountain Region: A Case Study on the Empire and the Sonoit$ Evaluation of Recharge Through U. Kafri and J. Ben -Asher - V6 Soil $ Laboratory Weathering of Water -Repellent Wax- Treated Dwayne H. Fink - V6 Soils Treated for Water Repellency $ Freeze -Thaw Effects on Dwayne H. Fink and Stanley T. Mitchell - V5 Soil -Turf Filter S Nitrogen Removal from Secondary Effluent Applied to a E. L. Anderson, I. L. Pepper and G. V. Johnson - V8 Soil -Water Content and Soil -Water Pressure $ Field Measurements of R. J. Reginato and R. D. Jackson - V1 Soil /Water Impacts of Salt Treatment S Water Harvesting: Albert Todd - V10 Soil -Water Status Via Albedo Measurement $ Assessing Sherwood B.Idso and Robert J. Reginato - V4 Soils S Microtrac: A Rapid Particle -Size Analyzer of Sediments and R. L. Haverland and L. R. Cooper - V11 Soils $ Penetrability and Hydraulic Conductivity of Dilute Sulfuric Acid Solutions in Selected Arizona S. Miyamoto, J. Ryan and H. L. Bohn - V3 Soils $ Water Quality of Streamflow from Forested Watersheds on Sedimentary Paul W. Gregory and Peter F. Ffolliott - V6 Solar Energy for Irrigation PumpingS Feasibility of Using Dennis Larson, D.D. Fangmeier, W.G. Matlock, John Daffy and C. D.Sands II- V6 Solar Powered Irrigation Pumping Equipment S Dennis L. Larson and C. D. Sands, Jr. - V9 Solar Powered Pumping Project: Operating Experiences$ Arizona Dennis L.Larson - V11

163 Solar Radiation as Indexed by Clouds for Snowmelt Modeling $ D. P. McAda and P. F. Ffollíott -V8 Stock Ponds for Hydrologic Research on Southwest Rangelands $ Use of J. R. Si manton and H. B. Osborn -V3 Stock Ponds with Soda Ash $ Effectiveness of Sealing Southeastern Arizona H. B. Osborn, J. R. Si manton and R. B. Koehi er - V8 Stock Tanks of Southeastern Arizona $ Time -Related Changes in Water Quality of D. E. Wallace and H. A. Schreiber - V4 Stochastic Analysis of Flows of Rillito Creek $ A N. E. Baran, C. C. Kisiel and L. Duckstein - Vi Stochastic Decision Model $ Input Specifications to a D. M. Clainos, L. Duckstein and T. G. Roefs -V2 Stochastic Prediction of Sediment Yields from Strip Mine Spoils of the Arid Southwest $ Mark E. Auernhamer, Martin M. Fogel, Louis H.Hekman, Jr. and John L. Thames - V7 Stock- Water Harvesting with Wax on the Arizona Strip $ Keith R. Cooley, Loren N. Brazell, Gary W. Frasier and Dwayne H.Fink - V6 Storm Flows Management in Relation to Industrial Development $ Robert E. Smith - P2 Storm Operations System for Salt River and Verde River Watersheds $ Salt River Project Emergency Dick Juetten and Don Wessner - P2 Stormflow as a Function of Watershed Impervious Area $ Jan M. Pankey and Richard H. Hawkins - V11 Stream Discharge in Navajo and Apache Counties, Arizona $ The Effects of Second -Home and Resort -Town

Development on . . . T. D. Hogan and M. E. Bond - V9 Stream Flows $ Impact of Development on Paul D. Trotta, James J. Rodgers and William B. Vandivere - V9 Streamflow Probabilities $ Conditional T. G. Roefs and D. M. Clainos -V1 Streamflow $ Converting Chaparral to Grass to Increase Paul A. Ingebo - V2 Stripmined Lands $ A Multiattribute Approach to the Reclamation of Fritz H. Brink, Lucien Duckstein and John L. Thames - V9 Subsidence Damage in Southern Arizona $ Charles A. McCauley and Russell L. Gum - 42 Temperature and Humidity Gradients in the Air Near the Ground$ An Exchange System for Precise Measurements of ... L. W. Gay and L. J. Fritschen - V9 Temperature Measurements$ Assessing Bare Soil Evaporation Via Surface Sherwood B.Idso, Robert J. Reginato and Ray D. Jackson - V5 Thunderstorm on a Semiarid Urbanized Watershed$ The Effect of an Intensive Summer D. G. Boyer and K. J. DeCook - V6 Thunderstorm Rainfall in the Southwest$ Stationarity in William C. Mills and Herbert B. Osborn - V3 Thunderstorms Rainfall in the Southwest $ Regional Differences in Runoff -Producing H. B. Osborn - YI Time Series $ On the Statistics of Hydrologic Sidney Yakowitz and Jack Denny - V3 Time -Related Changes in Water Quality of Stock Tanks of Southeastern Arizona $ D. E. Wallace and H. A. Schreiber - V4 Time- Series Analysis of Tucson, Arizona $ The Impact of Socioeconomic Status on Residential Water Use: A Cross- Section ... R. Bruce Billings and Donald E. Agthe - Y9 Tracers for Waste Monitoring $ New Organic Stephen L. Jensen and Glenn Thompson - P3 Tracers, A New Technology $ Chlorofluorocarbons as Hydrologic J. H. Randall and T. R. Schultz - V6 Transmission Loss Potential $ Geomorphic Features Affecting D. E. Wallace and L. J. Lane - V8 Transmission Losses in an Ephemeral Stream$ Bed Material Characteristics and J. B. Murphey, L. J. Lane and M. H. Diskin - V2 Transmission Losses in Ephemeral Stream Channels $ Estimating Leonard J.Lane, Virginia A. Ferreira and Edward D. Shirley - V10 Transmissivity Distribution in the Tucson Basin Aquifer $ D. J. Supkow - V2 Transmissivity Values in the Salt River Valley Using Recovery Tests $ Determination of Mary Ann Niccoli and MichaelR. Long - V11 Transpiration and Photosynthesis in Saltcedar $ Jay E. Anderson - V7 Transpiration, and Photosyntheses of Saltcedar $ Diurnal Trends in Water Status, Mary Ellen Williams and Jay E. Anderson - V7 Transpiration $ Estimating Phreatophyte Lloyd W. Gay and Theodore W. Sammis - V7 Transpiration $ Reducing Phreatophyte David C. Davenport and Robert M. Hagen - V7

164 Tree -Ring Data $ Augmenting Annual Runoff Records Using Charles W. Stockton and Harold C. Fritts - V1 Tree -Ring Dating of Colorado River Driftwood in the Grand Canyon $ C. W. Ferguson -V1 Uncertainties in an Environmental Impact Statement$ Systematic Assessment of Soronadi Nnaji, Donald R.Davis and Lucien Duckstein - V6 Uncertainties in Digital- Computer Modeling of Ground -Water Basins $ Joseph S. Gates and Chester C.Kisiel- V1 Unit Hydrograph to a Small Semiarid Watershed $ Application of a Double Triangle M. H. Di skin and L. J. Lane - V6 Universal Soil Loss Equation in the Tropics $ Use of the Todd C.Rasmussen and Fred C. Tracy - V11 Universal Soil Loss Equation to Semiarid Rangeland Conditions in the Southwest$ Applicability of the K. G. Renard, J. R. Simanton and H. B. Osborn - V4 Universal Soil Loss Equation $ Thunderstorm Precipitation Effects on the Rainfall- Erosion Index of the Kenneth G. Renard and J.Roger Simanton - V5 Urban Area in an Arid Environment. Tucson, Arizona $ Water Quality Problems of the G. Hansen - V8 Urban Area $ Analysis of Wastewater Land Treatment Systems in the Phoenix R. L. Ewing - V8 Urban Hydrology: State -of- the -Art $ Warren Viessman, Jr.- V3 Urban Perspective $ Arizona Groundwater Law Reform - An H. Holub - V8 Urban Runoff Development$ Legal Aspects of D. A. Chudnoff - V8 Urban Runoff: A Preliminary Evaluation $ Effect of a Grass and Soil Filter on Tucson Barney Paul Popkin - V2 Urban Runoff? $ Some Legal Problems of Hugh Holub - V2 Urban Waste Water Management$ Land Treatment for William L. Lorah and Kenneth R. Wright - V3 Urban Water Supply: The Tucson Case $ Rising Energy Prices, Water Demand by Peri -Urban Agriculture, and

Implications for . . . H. W. Ayer and D. W. Gapp - V8 Urbanization on Runoff from Small Watersheds$ Effect of Samuel E. Kao, Martin M. Fogel and Sol D. Resnick - V3 Urbanized Watershed $ The Effect of an Intensive Summer Thunderstorm on a Semiarid D. G. Boyer and K. J. DeCook - V6 USLE to Southwestern Rangelands$ Application of the J. Roger Simanton, Herbert B. Osborn and Kenneth G. Renard - V10 Vegetable Production $ Mulching Techniques for Arid Lands R. W. Peebles and Norman F. Oebker - V1 Vegetation Changes Along the Santa Cruz River Channel Near Tumacacori,$ Hydraulic Effects of Lee H. Applegate - V11 Vegetation Types Across the Mogollon Rim$ Hydrologic Regimes of Three Malchus B. Baker, Jr.- V11 Vegetation Types $ An Analysis of Recession Flows From Different Wan Norazmin bin Sulaiman and Peter F. Ffolliott - V11 Vegetation $ Variations in Soil Moisture Under Natural T. W. Sammis and D. L. Weeks - V7 Wax Mixtures for Treating Soils for Water Harvesting $ Candelilla /Petroleum Dwayne H. Fink - V11 Wax on the Arizona Strip $ Stock- Water Harvesting with Keith R. Cooley, Loren N. Brazell, Gary W. Frasier and Dwayne H. Fink - V6 Wax Water Harvesting Treatment Improved with Antistripping Agent and Soil Stabilizer$ Dwayne H. Fink - V10 Waste Monitoring$ New Organic Tracers for Stephen L. Jensen and Glenn Thompson - P3 Waste Water Management $ Land Treatment for Urban William L. Lorah and Kenneth R. Wright - V3 Waste Water $ Evaluation of a Turfgrass - Soil System to Utilize and Purify Municipal R. C. Sidle and G. V. Johnson - V2 Wasted Waters for Desert -Household Gardening $ Salvaging D. H. Fink and W. L. Ehrler - V8 Waste -Water Utilization in Arizona$ Economic and Energy Opportunities for Municipal E. J. Weber - P3 Wastewater Effluent - An Element of Total Water Resource Planning $ J. D. Goff - V8 Wastewater Land Treatment Systems in the Phoenix Urban Area$ Analysis of R. L. Ewing - V8 Wastewater Reuse Alternatives $ Herman Bouwer - P1

165 Wastewater Reuse$ Heavy Metals 8 T. E. Higgins - V8 Wastewater Reuse- How Viable Is It? Another Look $ W. L. Chase and J. Fulton - V8 Wastewater to LandS Health Effects of Application of James D. Goff - V9 Water Consumption in Tucson, 1974 to 1978$ Changes in Water Rates and Adrian H. Griffin, James C. Wade and William E. Martin - V10 Water Consumption Patterns of Arizona Second -Home Owners $ Current and Forecasted M. E. Bond and R. H. Dunikoski - V8 Water Harvesting System: A Potential Mine Reclamation Alternative $ The Mound and Valley Charles L. Constant and John Thames - V10 Water Harvesting Treatment Improved with Antistripping Agent and Soil Stabilizer $ Wax Dwayne H. Fink - V10 Water Harvesting with Wax on the Arizona Strip $ Stock - Keith R. Cooley, Loren N. Brazell, Gary W. Frasier and Dwayne H. Fink - V6 Water Harvesting$ Candelilla /Petroleum Wax Mixtures for Treating Soils for Dwayne H. Fink - V11 Water Harvesting $ Laboratory Evaluation of Water -Repellent Soils for Dwayne H. Fink - V4 Water Harvesting $ Residual Waxes for Dwayne H. Fink - V7 Water Harvesting: Soil /Water Impacts of Salt Treatment $ Albert Todd - V10 Water Impoundment Applications for SBR /Asphalt Membrane Systems$ Carlon C. Chambers - V10 Water Rates and Water Consumption in Tucson, 1974 to 1978S Changes in Adrian H. Griffin, James C. Wade and William E. Martin - V10 Water Resource Alternatives for Power Generation in Arizona $ Stephen E. Smith, K. James DeCook and Rocco A. Fazzolare - V4 Water Resource Leaders $ Man -Nature Attitudes of Arizona Roger A. Kanerva and David A.King - V2 Water Resource Projects $ Early Public Involvement in Federal Freda Johnson and Michael Thuss - V9 Water Resources Development (Abstract) S Impact on the Environment by Sol Resnick - V3 Water Resources Management Plan for Tucson, Arizona $ Impacts of a New R. Bruce Johnson - V10 Water Resources of the Inner Basin of San Francisco Volcano, Coconino S E. L. Montgomery and R. H. DeWitt -V4 Water Resources of the Woody Mountain Well Field Area, Coconino County, Arizona S Errol L. Montgomery and Robert H. DeWitt - V5 Water Resources Research on Forest and Rangelands in Arizona $ Alden R. Hibbert - V4 Water Resources $ Collective Utility: A Systems Approach for the Utilization of Edwin Dupnick and Lucien Duckstein - V1 Water Rights of the Bureau of Land Management in Colorado$ Federal Reserved Richard A. Herbert and Anthony L. Martinez - V11 Water Rights: The Bureaucratic Response $ Indian Daniel C. McCool - V11 Water Supply Data Base $ A J.F. Nunamaker, David E.Pingry and Rex Riley - V7 Water Supply for Home Irrigation (Poster Session) $ Augmenting Barney P. Popkin - V9 Water Supply for the Carefree -Cave Creek Area S Study of the Adequacy of the Edward A. Nemecek and Philip C. Briggs - V6 Water Supply Planning $ The Role of Conservation in Augustine J. Fredrich - P1 Water -Awareness in Tucson, Arizona: A Case Study $ Kebba Buckley - V10 Wax- Treated Soil $ Laboratory Weathering of Water -Repellent Dwayne H. Fink - V6 Waxes for Water Harvesting $ Residual Dwayne H. Fink - V7 Weather Modification in Arizona, 1971 $ Herbert B. Osborn - V2 Weather Modification S The Role of the States in Control of Ray Jay Davis - V6 Weathering of Water -Repellent Wax- Treated Soil$ Laboratory Dwayne H. Fink - V6 Well Driller Logs S Computerized Depth Interval Determination of Groundwater Characteristics from Mike Long and Stephen Erb - V10

166 Well Field Area, Coconino County, Arizona S Water Resources of the Woody Mountain Errol L. Montgomery and Robert H. DeWitt - V5 Well Field Design S Application of Bayesian Decision Theory in Charles A. Bostock and Donald R. Davis - V5 Well Field, Coconino County, Arizona$ Structural Relations Determined from Interpretation of

Geophysical Surveys: Woody Mountain . . . Phyllis K. Scott and E.L. Montgomery - V4 Well Levels During Coastal Aquifer Tests $ Correcting Tidal Responses in Observed Water Barney P. Popkin - V11 Well Losses in Consolidated Rock Aquifers $ The Application of Step -Drawdown Pumping Tests for Determining ... V. W. Uhl, Jr., V. G. Joshi, A. Alpheus and G. Sharma - V5 Well -Field Design Criteria for Coastal Seawater Development$ Barney P. Popkin - V10 WellsS Barometric Response of Water Levels in Flagstaff Municipal ErrolL. Montgomery, Emily Dordosz, Russell O. Dalton, Jr. and Ronald H. DeWitt - V7 Wells, Flagstaff, Arizona $ Chemistry of Effervescing Groundwater from Municipal John C. Germ and Errol L. Montgomery - V5 Wetlands S The Importance of Arizona's Jon Rodiek - V1O

167