N O T I C E This Document Has Been Reproduced From

N O T I C E

THIS DOCUMENT HAS BEEN REPRODUCED FROM MICROFICHE. ALTHOUGH IT IS RECOGNIZED THAT CERTAIN PORTIONS ARE ILLEGIBLE, IT IS BEING RELEASED IN THE INTEREST OF MAKING AVAILABLE AS MUCH INFORMATION AS POSSIBLE ESL INCORPORATEDJA SUBSIDIARY OF TRW INC 495 JAVA DRIVE • P.D. BOX 510 • SUNNYVALE, CA 54086 1 4 1 v O Z w 1 m o .^ m

J-01 a E o ^ t Cc ^10 z , NASA/BLM APPLICATIONS PILOT TEST (APT) PHASE II FINAL REPORT t VOLUME I 1 EXECUTIVE SUMMARY 8 January 1981 a 6. (E82-10254) RASA/BL!! APPLICATIONS PILOT N82-24532 TEST (APT), PHASE 2. VCLURE 1: EXECUTIVE t SUARARY Final Report (Electroaagnetic Systems Labs.) 85 p 8C A05/" A01 CSCL 08B Unclas i 63/43 00254

Copy No. NASA Contract s NAS9-15740 ESL INCORPORATED A Subsidiary of TRW Incorporated Sunnyvale, California

8 January 1981

NASA/BLM APPLICATIONS PILOT TEST (APT) PHASE II FINAL REPORT

VOLUME I EXECUTIVE SUMMARY

This Document Consists of 108 Pages

Copy No. of 40 Copies ORIC-114AL PAGE 5 ^F pGOR QUALITY

CONTENTS

Section Title Page

1.0 INTRODUCTION 1 1.1 Organization of Final Report 1 1.2 Background and Direction of APT 1 1.3 Overview of Phase II Project Objectives 3 1.4 Overview of the Approach 4 2.0 TECHNOLOGY DEMONSTRATION 5 2.1 Landsat Data Processing 5 2.2 Data Collection 11 2.3 Productivity Estimation and Map Verification 22 2.4 Costs Analysis 28 2.5 Landsat Digital Products 28 3.0 TECHNOLOGY TRANSFER ACTIVITIES 31 3.1 Planning Session 31 3.2 Seminars/" Hands-On' Training 31 3.3 Project Status Reviews 33

Appendix I Project Flow Diagram Appendix II Computer Menu Descriptions Appendix III Results of Range Productivity Estimation Appendix IV Results of Woodland Productivity Estimation Appendix V Results of Forest Productivity Estimation

-ii-

NAWBLNI APT FINAL REPORT VOLUME I: EXECUTIVE SUMMARY

1.0 INTRODUCTION.

1.1 Organization of Final Report. I The final report for the Phase II activities on the NASA/BLM Applications Pilot Test (APT) is divided into.) three stand-alone documents. Volume I is the Executive Summary, and as such describes the Phase II project in general terms on objectives, procedures and results. Volume II, Technology Demonstration, describes the methodology and results of the vegetation mapping and inventory activities for the Arizona Test Site. Volume III, Technology Transfer, details the training y courses presented to BLM personnel on the integration and use $.. of remotely sensed data for mapping and inventory requirements as implemented during the Phase II project.

1.2 Background and Direction of APT.

The basis of the current APT program jointly being carried out by NASA and BLM was a remote sensing research project contracted by BLM to the University of California, Berkeley (Colwell 1975). That project, popularly referred to as the Susanville Program, took place in BLM's Susanville District in Northeastern California and-Northwestern Nevada in the 1973-74 time period. Although the program was concerned with remote sensing in its most general terms, a number of specific applications of Landsat and aerial photography at various scales%%ere investigated. An analysis of output products from that study concluded that certain information extracted from satellite data and supported with other stages

r r. 1.2 --Continued.

of data could be used to prepare unit resource analyses and thence for management framework and activity planning. Specifically, the Susanville Program developed vegetation I cover maps and production estimates.

F," ^. As a result of this experience personnel from BLM ! and NASA met in May 1975 to explore the basis for a joint remote sensing project. At BLM's suggestion, an inventory of wildland vegetation was selected since vegetation plays a role

in all BLM resource activities ; : ands, wildlife, minerals, recreation, watershed, range, - ad forestry). By February 1976 a project had been prepared and in November 1976 a memorandum of f' understanding between BLM and NASA was signed. The NASA/BLM program was initiated as an Application System Verification and

Transfer ( ASVT) and was later changes to an Application Project

' Test ( APT). The program undertaken was a phased three year project which encompassed three test sites: in Alaska, in Arizona, and in Idaho respectively. The program was designed to 3 initially expose the user agency ( BLM) to technological l concepts and procedures in remote sensing and thereafter to have BLM assume an ever increasing responsibility for the t- conduct of the program. At its conclusion total responsibility for planning and implementation will reside with BLM. The ^. first (Alaska) phase was carried out by NASA working through a contractor (ESL Incorporated) and was for the purpose of vegetation cover mapping in a test site representative of a —,o.-- northern spruce tundra biome. During this phase BLM assisted A with the field work, wit hga digital processing and analysis and with monitoring the contract. The second (Arizona) phase of the APT program has included both vegetation cover mapping and range, woodland, and forest production estimation. The former task (cover mapping) is being carried out by BLM. The latter task, to which this report is primarily directed, is being

-2- 1.2 --Continued.

carried out by ESL Incorporated under contract with NASA. The Phase II site is located in extreme Northwestern Azizona and is typical of the southwest desert and upland communities. The third (Idaho) phase is to include both vegetation cover mapping and range forage production estimation, both tasks in this case being carried out by BLN. The Phase III site lies in Southwestern Idaho and is typical of a Great Basin sagebrush/ grassland community.

A joint planning session including NASA/JSC, BLM/DSC and Arizona Strip District, and ESL and their subcontractors was conducted on October 3-4, 1978 prior to beginning the contract effort for Phase II in Arizona. The purpose of the session was to finalize project philosophy, objectives and schedules and workshop (training) content, attendees, location and schedule. A similar session at the beginning of Phase I in s - Alaska was very helpful as a beginning point in that the participants were able to exchange views and have an overall j view of the tasks to be accomplished. ESL has prepared a flow chart with a schedule of tasks and tentative dates. Using this as a starting point, a final flow chart of events for the entire Phase II effort was developed. A copy is included as Appendix I. Although changes were required, the flow chart was an invaluable tool as a reference point for identifying the impact of changes on other events and schedules.

1.3 Overview of Phase II Project Objectives.

a The primary objective in Phase II was the demonstration of the integration of remote sensing technology with existing techniquesq producing for a vegetation type map and vegetation productivity estimates. A parallel objective was the transfer of this technology to BLM personnel to promote the implementation and utilization of the procedures and techniques within a -3-

L. 1.3 --Continued.

I the BLM operations framework. The technology being demonstrated emphasized the integration of quantitatively-based remote .I sensing data while maintaining consideration of cost-efficiency in the implementation of the overall design. The technology transfer expanded on this emphasis with particular attention + on further developing the understanding, on the part of BLM, of approaches to inventories that integrate multiple data sources given various resource information objectives. Specif- ically, any considerations based on the candidate approaches were focused on sampling strategies and analyses of the costs of data collection as specified by those strategies.

1.4 Overview of the Approach.

1.4.1 Technology Demonstration.

The technology demonstration aspect of Phase II combined data from Landsat, low-altitude color aerial photography and ground visits to produce the vegetation type map as well as the productivity estimates for range, woodland and forest resources. One Landsat scene covering the Arizona Test Site was selected for computer-aided processing to provide the basis of the type map and to control the productivity sampling efforts (Section 2.1). Photo-interpretation results from large scale (1:750 - 1:1200 scale) color aerial photography was used to describe the vegetation characteristics of the spectral classes derived from the Landsat processing (Section 2.2). Also, estimates of parameters related to productivity were made from the aerial photography to be combined in a double-sampling framework with ground measurements of productivity. Photo data - and ground data were combined through specially designed formulae to produce the necessary productivity estimates and also the vegetation species composition descriptions for each spectral class (Section 2.3). Map-type outputs were created based on

-4- 1.4.1 --Continued. {

the Landsat data and the related class descriptions and, for some maps, productivity information was also included (Section 2.4).

1.4.2 Technology Transfer.

The technology transfer portion of Phase Ii was divided into three parts consisting of a planning session, workshops (two) and project status reviews (four). The planning session was held at the outset of the project to familiarize project participants with the full scope of the work to be performed and to establish a milestone schedule (Section 3.1). The two workshops were held to provide "hands-on" instruction for project participants in specific Phase II technology areas and utilized actual project data (Section 3.2). The four project status reviews were held periodically through the life of the project to maintain continuity in reporting on interim results, to resolve any problems that occurred and to update the project schedule as necessary (Section 3.3).

2.0 TECHNOLOGY DEMONSTRATION.

2.1 Landsat Data Processing. 1 2.1.1 Objectives.

Given the technology objective of a vegetation type map and productivity estimates for range, woodland and forest resources, a. computer-aided digital, classification of Landsat satellite data was selected as a major source of input data. Specifically, tho classification results would provide the basis for the vegetation 4D type map as well as the sampling frame for the productivity estimation tasks. The sample allocation, sample selections and

-5-

2.1.1 --Contir.aed.

tabulations of ostimate summarizations all would use the Landsat classification.

2.1.2 Procedures.

The Landsat data processing for this project was divided into six major categories:

1) Preprocess.,ng 2) 'cr-iini_ng ?) Cl& 3,, , ' ifica :ion 4) Strrtification S) ::Jar— ,escription 6) D"'6 olt ai Terrain Data

Preprocessing. Landsat digital data requires a certain amount ^.)f manipulation before it can be used effectively and

eff.i-iently in a resource assessment /inventory project. To be used more effectively, the data is placed into the proper format, corrected for radiometric and geometric distortions, and related to a preselected ground coordinate system. To be used more efficiently, the project area is extracted from the appropriate Landsat scenes to eliminate unnecessary manipulation of data that is outside project boundaries.

Training. Unsupervised and supervised training procedures were utilized to generate the spectral statistics to be applied in the maximum likelihood classification algorithm used to classify the Landsat data. The unsupervised approach was used first to develop an initial set of training statistics. Preliminary classification based on those statistics performed r on four subsections of the project area was evaluated for complete- ness. Based on that evaluation, specific training sites were r selected and added to the original set of spectral statistics in

I -6-

C ^y 2.1.2 --Continued.

order to improve the results from the preliminary classification. The overall training procedures generated 83 spectral clusters (distinctly separable spectral responses identified in the Landsat data) to be used in classifying the Landsat data. It is important to note that at this point there was only a general vegetation type description associated with each cluster. The subsequent sampling/photo-interpretation procedures (Section 2.2) developed detailed quantitatively based vegetation descriptions.

Classification. The 83 spectral clusters were applied in a maximum likelihood classification algorithm to the raw Landsat data covering the project area. The output of this step was a computer classification with 83 spectral classes generally related to the vegetation of the area. These classes were evaluated by BLM-Arizona personnel for potential of any spectral class representing more than one vegetation type. After all such "confusions" were noted, it was determined that an envircnmental t stratification could be used to eliminate them.

Stratification. Two types of stratification were applied to the Landsat data: administrative and environmental. The administrative stratification partitioned the project area into selected pastures and grazing allotments. This would be used to control the sampling and data sr-&.airization of the productivity estimation tasks. The environmental stratification divided the area into low desert, high desert, mountain, and forest types. This was used to assign new class identifiers to those spectral classes occurring in more than one environmental strata.

Class Description. In order to sample the classification most efficiently and to reduce the total number of samples required, it was necessary to progress from the general vegetation descriptions of the detailed spectral classes to preliminary

-7-

2.1.2 --Continued.

vegetation composition estimates for a set of summary classes. Based on the general vegetation type noted for each detailed spectral class, spectral similarities between classes and proximity of locations between classes within the environmental strata; the detailed spectral classes were grouped into a not of summary classes for subsequent sampling and estimation. These summary classes were then combined with existing photo interpretation data over the project area in a one-way analysis of variance to produce est=mates of the average percent composition of each ' class in terms of trees, shrubs, grass, and non-vegetation. The summary class descriptions were major inputs in determining sample size, allocation and selection.

Digital Terrain Data. The final step in the Landsat processing phase was to input and register digital terrain data with the Landsat data for use in the vegetation mapping task. The detailed descriptions of spectral classes to be generated following the sampling of the above summary classes would be augmented by topographic information. Specifically, for each class the ranges of elevation, slope and aspect would be determined and then included with the species composition descriptions. This terrain information is reported so to further explain any spectral class differences due to topographic influences.

2.1.3 Landsat Processing Results.

t The training steps produced 83 spectral clusters (training statistics) to be used in classifying the raw Landsat data. Elimination of "confusion s classes utilizing the environ- 1 mental stratification resulted in expanding to 117 detailed classes from the original 83 classes. Class description steps grouped the 117 classes into 27 summary categories as shown in e Table 2-1. The sections following describe the sampling,

-I- I Table 2-1. Results of Assignment of All Spectral Clusters to Summary Categories.

ummary Strata Description Spectral Clusters Cate or 1 Low Desert Creosote-Bursage 1,2,7,10,11,87,88, (rocky soil) 89,90,91,92,93,98, 100 2 Low Desert Creosote-Bursage 3,5,9,86,94,95,99 (sandy soil) 3 Low Desert Creosote-Pure 4 4 Low Desert Upland Desert 6,12013,14,96,97 Shrub - Creosote Dominant 5 Low Desert Blackbrush 45 6 Low Desert Mixed Desert 8,74 Shrub - Creosote and Cactus 7 All Riparian 25,26,27 Woodland 8 High Desert Shrub-Grass 28,34,38 9 High Desert Grassland-Shrub 31,39 10 High Desert Snakeweed-Grass 29,30,106,107,108 11 High Desert Sage-Mix Shrub 32,44,102,103,105, 111 12 High Desert Sage 35,49,51,54,55, 110,112 13 High Desert Saltshrub 36,37,104,109 14 High Desert Pinyon-Juniper- 33,50 Sage 15 High Desert Pinyon-Juniper- 53,56,57,58,59,61, Shrub 62,63,84,85 16 High Desert Pinyon-Juniper 40,41,46,52 17 Mountain Mountain Shrub 42 18 Mountain Mixed Chaparral 60,113 19 Loan Desert Agriculture 15,16,17,18,19,20, 21,22,23,24,114 20 Mountain Ponderosa Pine- 66 Oak

-9- Table 2-1. --Continued.

Summary Strata .r Dwscription Cate or Spectral Clusters

21 Mountain Ponderosa Pine- 65 Mix 22 Mountain Ponderosa Pine 64 23 All Shadow 47,83 24 All water 67,68,69,70,71,72, 73,82 25 All Bare 76,77,78,79,80,81 26 Low Desert Upland Desert 43,48,75 Shrub-Blackbrush 27 Hilb Desert Sage-Grass 101,115,116,117

-10- 2.1.3 --Continued.

estimation and vegetation mapping phases and their utilization of the digitized administrative boundaries and digital terrain data.

2.2 Data Collection.

1 2.2.1 Overview of Sample Design.

The general sample design used in this project could be described as two-stage with double sampling. The first stage was a rectangular grid of primary sample units (PSUs) with dimensions of 2000 meters (east-west) by 450 meters (north-south). A set of those PSUs was selected for coverage with large scale (1:750 to 1:1200 nominal scale) aerial photography (LSP) from which to derive the PSU estimates. The second stage was a double sample of photo plots and ground plots from within designated PSUs. Rangeland, woodland and forest types were sampled independently of one another even though the PSUs for each were defined by a common grid. Other features of the design varied with the vegetation type being sampled as described in the following paragraphs.

2.2.1.1 Rangeland.

The quantities to be estimated were pounds per acre and kilograms per hectare of palatable forage for cattle within ten specified allotments. Figure 2-1 illustrates the locations of those allotments within the project area. The BLM-specified precision requirement for the estimates was as follows: the allowable sampling error for the overall forage estimates was ±20% at the 80% confidence level. The results were to be tabulated by vegetation strata (as defined by Landsat processing) and also by pasture and allotment. The primary results were to be total available palatable forage for cattle adjusted for utilization. The secondary results were to be current palatable forage

-11- OF POOR QUALITY

Leld Community ;k ;k/Soapstone Canyon irricane Pank tet iyon Llow-Tarsi Spring int -

Figure 2-1. Range Allotment Boundaries.

-12- 2.2.1.1 --Continued.

whether available or not. The vegetation stratum estimates were F to have 801 confidence intervals. For the total estimates across all strata, the precision requirement would be met if one-half i the length of the 90% confidence interval was no more than 201 i of the actual estimate.

To allocate and select samples, the sample design went through an optimization process. A Survey Planning Model (SPM) developed at the University of California by the Remote Sensing Research Program applied summary class description data (Section 2.1.3 Landsat Processing Results) to create a model representative of the magnitude, variability, and spatial distribution of forage from grasses, shrubs and (orbs for 14 vegetation strata (1, 2, 3, 6, 8, 17/18, 9/12/27, 10/13, 11 0 14, 15, 16, 21, 26 from Table 2-1). The SPM combined this data with cost information for collection and analysis of the photo and ground data as well as the 20!/801 precision requirement by means of a non- linear programming procedure to optimize the sampling method, number of samples and allocation to strata. The result was a stratified two-stage design with Landsat providing the stratification. The first stage was to select 108 primary sample units of size 450 meters by 2000 meters to be used as locations of the second stage sampling units. These PSUs were selected with equal probability from within the 14 vegetation strata. The second stage was in itself a double sample of aerial photo plots and ground plots. Fach of the 108 PSUs was to be flown with large scale aerial photography such that each was composed of 15 photo plots at equal intervals. This resulted in a total of 1620 photo plots which were to then be photo interpreted for assignment into cover type strata (1447 foil within the strata of interest) and 135 sample plots were selected from the 14 vegetation strata of interest for ground measurement. Each ground plot was to be subsampled by five parallel transacts of 40 subplots each for a total of 200 points per ground plot.

-13- 6 i 2.2.1.2 Woodland.

The quantities of interest were cubic foot volume per acre and cubic meter volume per hectare for pinyon pine and juniper. Size (diameter) and age were also of interest. The precision requirement was the same as for rangeland: estimates within t20t at the 801 confidence level. The results were to be tabulated by allotment (see Figure 2-2) and 80• confidence intervals were to be supplied for each. For the total estimates across all three allotments, the precision requirement would be met if one-half the length of the 80% confidence interval was no more than 20• of the actual estimate.

A sample design optimization process was used similar to the rangeland case to specify sampling method, number of samples and allocation to strata. The SPM created a model representative of the magnitude, variability and spatial distribution of pinyon/ juniper volume for the three woodland strata identified and described during the Landsat processing phase. Based on the volume model, the costs associated with data collection and the precision requirements, a two-stage design was again specified. The first stage was to select 45 primary sample units with probability Qf selection proportional to the acreage of pinyon/ juniper within each PSU. In contrast, rangeland PSUs were selected with equal probability. The second stage was again a double sample of photo plots and ground plots. with 15 equally spaced photo plots per flight line and one flight line per PSU, a total of 675 photo plots were to be available from which to select ground plots. Each photo plot was assigned to a cover type strata and samples for ground measurement were selected with equal probabilities from those plots falling in the strata of interest:

i -14- F mountain

Figure 2-2. Woodland Allotment Boundaries.

-is- t 1 2.2.1.2 --Continued.

Photo Stratun► No. of Photo Plots No. of Ground Plots

Pinyon/Juniper/Sage 232 5

Pinyon/Juniper /Shrub 104 6 i Pinyon/Juniper 155 ^S Total 491* 16

*Reflects that the remaining 184 plots of the total 675 photo plots did not fall in the above strata.

There was also to be subsampling within the photo and R ground plots using the line intersect method to select trees to be measured for volume estimation purposes. One of several tem-

plates ( photo scale dependent) each with two parallel lines representing 75 feet each was to be overlayed on the photo plot. Any tree whose crown intersected a line would be photo interpreted for volume. For the 16 plots to be ground visited as well, the same photo interpreted treee were measured for volume on the ground.

2.2.1.3 Forest.

The quanitities of interest were ponderosa pine board foot volume per acre by stand (vegetation strata) for each area illustrated in Figure 2-3. The precision requirement, as in the previous cases, was ±201 at the 801 confidence level. The requirement would be met if for the combined areas one-half the length of the 801 confidence interval was no more than 201 of the actual estimate.

The sampling method, number of samples and their allocation were derived by the sample optimization process described for the woodland case above but used summary data for three ponderosa pine strata. The result was a stratified two-stage design where the first stage was to select 47 PSUs with equal probability from within the three strata. (In contrast, the woodland PSUs were selected with probability proportional to area.)

4 Y . -16-

R i^ r

Figure 2-3. Map Depicting Forest Areas A and H Within the Arizona Test Site.

2.2.1.3 --Continued.

The second stage was a double sample of aerial photo plots and ground plots. Each PSU was covered with 15 photo plots at equal intervals for a total of 690 photo plots ( one PSU photo set was unusable resulting in 15 less plots than anticipated). The photo plots were interpreted to be assigned to cover type strata and samples were selected from the ponderosa pine strata for ground measurements:

# of PSUs # of PSUs # of Photo Plots # of Ground Stratum Selected Available within strata Plots to Visit

Ponderosa Pine/ Oak 4 4 56 4 Ponderosa Pine/ Mixed Shrub 26 25 62 15 Ponderosa Pine 17 17 47 10 Total 47 46 165* 29 *Reflects that 525 of the 690 photo plots were not assigned to ponderosa pine strata.

There was subsampling within the photo and ground plots using the line intersect method as described for the woodland case in Section 2.2.1.2 above.

2.2.2 Summary of Allocation and Selection of Samples.

Results of the procedures described in the sample design overview above are summarized in the following table:

# of Photo Plots # of # of Strata # of PSUs Assigned to Ground Plots Vegetation of Interest to Select * Strata of Interest Selected

Rangeland 14 108 1447 136

Worid land 3 45 491 16

Forest 3 46 165 29

F Total 20 199 2103 18i *Assumes loss of one PSU in forest type.

^F -IA-

2.2.2 --Continued. 1 The number of photo plots above reflects that out of 2985 possible plots, only 2103 were assigned to strata to be sampled for productivity. All photo plots were interpreted for vegetation composition in producing the vegetation map of the area. Of the 181 ground plots selected, two forest plots did not contain any trees and therefore were not visited on the ground= and one range plot was on such steep ground that measurement was impossible. As a result, 178 ground plots were actually visited and the three missing plots were not replaced as this could have introduced unwanted bias into the productivity estimates. 1 2.2.3 Photointerpretation on the Large-Scale Aerial Photography.

` As described above, each photo plot from the PSUs was initially interpreted to assign it to a cover type strata as part of the procedures to select second stage samples. There were, however, two more phases of photo interpretation: vegetation composition and productivity estimation. Each photo plot was first interepreted, independently of cover type assignment, for percent species composition and then used to describe the Landsat spectral classes. Finally, photo plots within the forest and woodland vegetation types were interpreted for tree volume estimates then used to estimate volumes for the Landsat summary classes (cover-type strata). These procedures are described in more detail below.

2.2.3.1 Vegetation Composition.

The large scale aerial photography (LSP) consisted of 35mm format color slides flown to achieve nominal scales ranging from 1:750 to 1:1200 (actual scales ranged from 1:500 to 1:3000 due to rugg_3d terrain in some areas). An LSP photo plot was a stereo pair with the stereo overlap defined as the area of interest. There were 15 equally spaced plots along the center of

-19-

r 2.2.3..1 Continued.

the long dimension of each PSU (2000 maters by 450 meters) resulting in plot centers approximately 125 meters apart and each plot representing 60 feet to 100 feet square on the.ground, depending on photo scale.

s To combine the photo plot interpretation with the Landsat spectral classes, the location of each plot was identified on USGS quads and then digitized and transformed to the same ground coordinate system to which the Landsat data had been registered. This allowed matching of the Landsat classes to the interpreted data on a point-by-point basis.

Within the stereo overlap of each photo plot the follow- s ing vegetation composition information was interpreted and recorded: identification of major plant species occurring on the plot, percent ground cover for each species, and also height class and foliar density by species. The interpretation was performed by one person to eliminate multiple biases that would occur by using more than one interpreter. Thus, any bias in the inter- pretations of the plot characteristics was constant.

2.2.3.2 Woodland and Forest Volume.

in addition to vegetation composition, trees were selected via the line intersect method to interpret for volume estimation on all plots from PSUs covering the forest and woodland vegetation types (676 photo plots out of the total 2985). One of several templates (depending on the plot photo scale) was overlayed in a predetermined manner (to ensure randomness) on to the stereo overlap portion of each photo plot. Each template had two parallel lines with each line representing 75 feet on the ground at the photo scale for which the template was prepared (a total of nine templates were prepared representing photo scales

-20- t

2.2.3.2 --Continued. i from 1:500 to 1:3000 to accommodate the range of actual photo scales noted in the LSP). For a given plot, any tree whose crown touched one of the lines, information was determined and recorded as to its species (pinyon pine, juniper o:c ponderosa pine were the only species of interest), crown diametir (measured) and height (estimated to the nearest five feet).

2.2.4 Ground Data Collection.

The data collection procedures for the ground plots were designed to be as close to BLM procedures as the multistage design would allow. The procedures used on the range plots were modified from! the Soil Vegetation Inventory Method (SVIM) while the forest and woodland procedures were based on the BLM Extensive Forest Inventory Field Handbook.

2.2.4.1 Range Plots.

Plot location was defined as the center point of the stereo overlap of a selected LSP plot. After locating it on the ground, this point was used as the center of a 200 point rectangular grid as formed by 5 lines with 40 points per line spaced evenly over a 42 foot by 48.75 foot area. These 200 points were used in determining species composition of the plot. A systematic set from the 200 points of 20 subplots (.1 square meter) were used to take SV1.4 shrub characterization and ocular weight estimates, recorded on modified SVIM field data fcrms. A subset of the ocular weight estimate plots were clipped, air-dried and weighed to use for adjusting the ocular estimates. This was done for each field crew member so as to have individually developed adjustments for a given plot based on who made the original ocular estimates.

-21- 2.2.6.2 Woodland and Forest Plots.

As described previously, plot location was determined by the parallel transacts in the stereo overlap of each selected LSP plot. These transacts were annotated on the LSP photos to be used in the field with the individual trees to be measured, marked and numbered to ensure correspondence with the photo interpretation data. For each tree, the following information 1 was recorded: diameter (ground and stump height for woodland and breast height for forest), height, crown diameter, number of stems i{ per tree (woodland) and age (on a subset of trees). Measurement of the distance between two photo-identified ground-located points was also recorded for use later in determining actual photo scale.

2.3 Productivity Estimation and Map Verification.

The f ollo•tng sections describe the steps and procedures used to combine the ground measurements, photo estimates and Landsat classification to produce the outputs specified by the project objectives.

2.3.1 Vegetation Type Mapping.

Before combining the Landsat classification results with the photo interpretation data (species composition), the 2985 photo plots were separated into two groups. The first group was composed of 2399 plots which were used to generate the vegetation descriptions of the Landsat classes. The remaining 586 photo plots, the second group, were used to verify those descriptions. This partitioning of photo plots was done in such a manner as to ensure that an adequate number of photo plots would be available . to both describe and then verify the maximum number of Landsat } spectral classes.

-22-

4 2.3.1 --continued. i The photo data from each plot from the first group was then matched with the Landsat classification on a point-by-point class-by-class basis using analysis of variance techniques. This produced the mean and standard deviation of percent cover by species for each Landsat class. The same proc*dures were followed with the verification set of plots. HLM-Arizona personnel, using the vegetation descriptions for the 117 Landsat classes and checking locations of the classes within the project area, assigned each class to a category within the Arizona vegetation framework. This resulted in 14 Level III vegetation framework categories which could be aggregated further into 10 Level II categories. Table 2-2 shows the results of the assignments. The digital terrain information, elevation, slope and aspect was combined with the framework categories resulting from above to augment the species composition descriptions of each. Appendix II contains the "menu" descriptions (vegetation and topography) of each of the 14 Level III categories. The verification data set was compared with the original description not and it was noted that no significant differences existed between them. The descriptions of the vegetation framework classes and their locations (based on the Landsat data) were therefore considered accurate representations of the ground conditions in the project area.

Area estimates by vegetation type with 80 percent confidence intervals at the pasture and allotment level were generated by relating the Landsat class counts to their descriptions from the photo data through linear regression. In the example in Table 2-3, the "adausted estimate" represents the regression coefficients applieu to the Landsat area summaries. The "801 confidence" column represents an interval which has approximately 801 probability of containing the actual value.

-23- Table 2-2. Assignment of the 117 Computer Classes to the Vegetation Framework.

vegetationtation Framework Classi fication Computer Classes - 4 i 1) 1-- Agriculture 15-24, 114

2) 211 Ponderosa Pine Forest 64, 65

3) 311 Pinyon-Juniper Woodland 33, 35, 41-43, 46, 47, 52, 53, 55-59, 61-63, 66, 83-a5

4) 322 Riparian Woodland 2;o, 26

5) 412 Upland Desert Shrub 1-10, 12-14, 74, 75, 81, 86, 87, 89-100

6) 421 Great Basin Sagebrush 11, 28 -30 0 32, 34, 38, 40, 44, 48 -50, 540 101-103, 105 0 106, 108, 110 0 112, 115-117

7) 423 Blackbrush 88

8) 424 Other Tall Shrub 111

9) 425 Half Shrub 45

10) 432 Oakbrush 600 113

11) 433 Other Mountain Shrub 27, 31, 39, 51

12) 511 Perennial Grassland 36, 37, 77, 10%, 107, 109

13) 6-- Barren Land-- 76, 78-20

14) 7-- Water 67-73, 82

-24- ' Table 2-3. Example of Area Estimates in Hectares Allotment and Pasture. fiby ` Vegetation Type: Pinyon/Juniper Woodland

ALLOTMENT LANDSAT-ONLY ADJUSTED PASTURE ESTIMATE ESTIMATE 803 CONFIDENCE LOWER HURRICANE 489 494 285,703 ST. GEORGE 192 194 112,276 COYOTE 109 110 63,156 NORTH GYP 75 76 44,108 WEANING 3 3 2,4 SOUTH GYP 23 23 13,33 GRAVEL 60 60 35,86 PETE 23 23 13,33 HOLDING PEN 0 0 0,0 PARKER 1 1 1,1 FORCE 5 5 3,6 SELLING 0 0 000 TOQUER TANK 44 44 26,63 1 5 5 3,7 2 39 40 23,56 MAINSTREET 6807 6875 3971,9779 CECIL 7 6 4,9 ROUK') POND 1 1 011 SQUARE POND 8 8 5,11 ANTHONY'S HIGLEY 16 16 9,23 CALVING 12 12 7,17 TEMPLE TRAIL 27 27 16,39 SALARATUS 63 64 37,91 MUDHOLE 19 19 11,27 TWIN TANKS 367 371 214,527_ WARDS 23 23 14,33 DUTCHMAN 693 699 404,995 COX-ATKIN 1511 1526 881,2171 ,. ENGLESTEAD 1973 1992 1151,2834 LITTLE JOE 2036 2057 1188,2925 BISHOP i BURR 53 53 31,76` -25- i r*T

2.3.2 Rangeland Productivity Estimation.

The data reduction process started with the air-dry weights of the species clipped on the various subplots of each range ground plot (see Section 2.2.4.1) and related these to the i corresponding ocular estimates through linear regression (R2 -77%). The resulting regression coefficients were then used to adjust the remaining ocular subplot weight estimates to produce current forage weight estimates, "available" or not, the secondary estimates. At t. ,_i point, the availability and utilization factors were applied to produce the primary estimates: forage weights "available" whether "utilized" or not. Both types of estimates were carried through the remainder of the estimation process; however, at this stage all estimates for unpalatable species (for cattle) were deleted from the process. The adjusted estimates were averaged by species, weighted (statistically speaking) by the frequency of species occurrence in all subplots, and then summed to obtain forage weight estimates acroso all palatable species for each of the 135 ground plots.

As described previously (Section 2.2.1.1), the LSP was used as a stratification tool in selecting the plots to be visited on the ground. Attempts were made to correlate the ground-based estimates of forage weight with photo-interpreted values such as crown cover, the cover/density product, and the cover/density/height product, but the relationships were not strong enough to be useful. Consequently, for the range case, the LSP was used only for stratification and not estimation.

The ground-based weight estimates were related to their corresponding Landsat summary classes to produce weight estimates per unit area by class, including the associated 80% confidence inter./al for each. Forage weight estimates for the specified allotments were developed by ratio estimation. (See Appendix III for summary of estimates.)

-26-

2.3.3 Woodland Volume Estimation.

The ground-based measurements of diameter at stump height, total tree height and average crown diameter for trees selected using the line intersect method were converted to cubic foot volumes through BLM pinyon-juniper volume tables. The ground volumes were then related to the Photo values through multiple regression (R2 - 37% fcr pinyon; 341 for juniper) to yield a tree volume prediction equation which could be used for a`1 the photo-interpreted trees. These predictions were then expanded through an estimation procedure that used the tree crown diameters (which were proportional to the probability of selection for a tree) and the transect lengths to develop volume per acre estimates for each photo plot. These were then averaged by PSU and combined with PSU sampling weights to obtain the final cubic foot volume estimates (also converted to metric measure: cubic meters per hectare) for the three allotments of interest. Average diameter and age estimates were obtained from ground measurements alone. Estimates of precision were also produced using methods fitting the sample design and parameter estimation process. (See Appendix IV for summary of estimates.)

2.3.4 Forest Volume Estimation.

The ground-based measurements of diameter breast height and total tree height for trees selected by the line intersect method were measured and used to compu ,_ board foot volume using equations generated by the U.S. Forest Service (BLM-Arizona personnel designated these equations as acceptable ones to use for the area). These ground determined volumes were then re- gressed against the photo-interpreted values (height and crown diameter) for the same trees (R2 - 78% for ponderosa pine) to yield a tree volume prediction equation which could be used for all the photo-interpreted trees. Both the photo and ground tree volume estimates were expanded to board foot per acre estimates for the entire plot by use of the crown diameters (which were proportional

-27- 2.3.4 --Continued.

to the probability of selection for a tree) and transect lengths. The matched photo and ground plot estimates were then related through regression for the Landsat strata sampled (R2 -441 for ponderosa pine/oak and ponderosa pine/mixed shrub; R 2 -761 for i ponderosa pine pure). A statified regression estimator was then- used to obtain the final estimates by Landsat strata, and the associated precision estimates were also produced. (See Appendix V for summary of estimates.)

2.4 Costs Analysis.

The costs associated with this project were separated into recurring and non-recurring components. Only recurring costs were considered in this analysis. In other words, "false" starts, development of procedures, and non-special management costs were omitted to the extent possible. The objective of this analysis was to associate a cost per acre with both the vegetation mapping and the productivity estimation elements of this project. Table 2-4 summarizes this analysis and shows that at the average price per ESL manhour over the life of the project, the mapping effort for the 2.2 million acre site cost 30 per acre. The productivity estimation, performed on 550,000 acres, cost

16.30 per acre iq ven the mapping had been performed. These figures were based on the costs associated with achieving the objectives and accuracies of this project. If similar projects were undertaken but with higher accuracy and precision specifications, it should be anticipated that the costs per acre could be greater.

2.5 Landsat Digital Products.

The Landsat digital hardcopy output products consisted of 4 application maps and 2 sets of color-coded classification images. The application maps were generated from the digital classification quantitative inventory and ancillary data in the

-28-

I Table 2-4. Phase II Cost Summary.

Total ESL = Mapping ♦ Estima Livil nuwj-v-xw&iai Project Element Man-Hours Component Component Costs

Landsat Processing 570 450 120 Multistage Sampling 270 120 150 48001 Data Collection Photo Acquisition 50 _ 50 - 11,3002 Photo Interpretation 700 500 200 Ground Data Collection 100 - 100 31,9703 Productivity Estimation 1250 240 1010 7001 Output Products4 Map Output 180 180 - Tabular Summaries 120 40 80

Totals 3240 1580 1660 $48,770

Vegeta^ion mapping cost [ ( 1580 hrs x $3.17/hr.) + 11,700] + 2.2 million acres - 30/acre

Productivity estimation - ( Given the mapping is completed and available) - [(1660 hrs x $31.70/hr) + 37,470j + 550,000 acres - 16.3t/acre

1. RSRP, U.C. Berkeley, in survey planning model runs and estimation work. 2. Acquisition of 3000 large scale aerial photography plots. 3. Collection of ground data on 179 ground plots. 4. Does not include generation of final report.

-29- 2.5 --Continued. s form of digital terrain data and administrative data. The map products were: a. Rangeland Suitability, defined as: (1) current production of useable forage above 20 lbs. per acre, I (2) located within a four mile radius of water,

(3) slope less than 51 percent.

b. Potential Rangeland Suitability, defined as: (1) current production of useable forage above 20 lbs. per acre,

(2) service area of water is greater than four mile radius, (3) slope less than 51 percent.

c. Sagebrush Treatment Area, defined as: (1) Great Basin sagebrush vegetation community, (2) slope less than 15 percent.

d. Fire Flash Fuel Areas, defined as:

(1) annual grass and forbs cover 30 percent or greater of ground, (2) aspect southwest - south - southeast.

These products were presented with a mapping minimum of 10 acreas (4 hectares) at a scale of 1:126,720 as a halftone grey-coded map on stable base transparency material. Maps 1 and 2 were limited to the 10 allotments sampled for range productivity. Maps 3 and 4 represent data over the full project area.

-30- i 1

2.5 --Continued.

The 5th map output products were for the forest administrative area A and consisted of 8x10 inch Polaroid hardcopies

of the color-coded computer class groupings. These products were presented at 10 acre ( 4 hectare) minimum mapping units for t the Level II framework and for the Level III framework groupings developed by the BLM-Arizona personnel. An IDIMS computer

compatible tape ( CCT) was generated with this data for the full project area as well.

3.0 TECHNOLOGY TRANSFER ACTIVITIES.

3.1 Planning Session - October 3 - 4, 1978.

The objective of this session was to familiarize all project participants with the expected flow of tasks over the life of the Phase II efforts. The project flow chart in Appendix I was presented at this time. Also, the two training courses were discussed and finalized as to scope, content and dates. A preliminary set of milestone dates were established to be used in monitoring progress on the numerous aspects of the Phase II effort.

3.2 Seminars/" Hands-On" Training.

There were two training courses given to BLM-Arizona personnel during Phase II: Multistage Sampling-Field workshop and the Data Analysis workshop.

3.2.1 MultistL sampling-Field workshop, May 21-25, 1979.

This week long workshop/seminar covt ed the sample allocation, data collection and data analysis procedures used on the project. The objectives of the session was:

-31-

j 3.2.1 --Continued.

a. The development of a thorough, practical understanding of sample allocation procedures used on the project and the justification for the procedures. 1 b. To introduce and develop, from a practical stand- point, the statistical procedures and estimators r being used on the project.

c. To further develop understanding of the optimization procedures being used with an emphasis on the factors that must be considered when designing an inventory and mapping project.

d. To further the understanding of the data sources being used on the project, including (1) the information extraction procedures, (2) relative level of accuracy and precision of each data source, (3) relative cost of acquisition and data extraction for each source, and (4) the risk associated with the use of data from each source.

e. To prepare the attendees for the data analysis workshop to be held at ESL in Fall 1979.

Participants in the course stated afterwards that the objectives were, in general, satisfied. The most frequent comment, however, was that of having a much greater appreciation for the complexity of resource inventory programs such as Phase II.

3.2.2 Data Analysis Workshop, November 13-16, 1979.

The objective of this course was to familiarize program participants with the procedures and techniques to be followed in reducing the data collected during Phase II to produce the

M -

-32-

r= ti z ,

3.2.2 --Continued„

vegetation map and the productivity estimates. The course included presentations on analysis of variance, contingency analysis, regression analysis and ratio estimation procedures, all using actual data from Phase II. The most significant result of the course was the grouping of the 117 Landsat spectral classes into 14 Arizona vegetation framework Level III categories performed by the attending BLM-Arizona personnel.

3.3 Project Status Reviews.

The objectives of these periodic reviews were to present progress to date on the Phase II effort and identify and resolve any problems noted or anticipated. The four in-progress reviews were held in January, June, September, and November, 1979. These reviews were considered extremely valuable in maintaining involvement in the project by all BLM, NASA and contractor participants. Coupled with the training course, the reviews provided an excellent means of exchanging information on the project: its objectives, the procedures, the problems encoun- tered and their solutions and the results.

A final project review was held on 29-30 May 1980 to present the accomplishments and results of Phase II to a larger audience than that addressed at the status reviews. In the words of the NASA Project Manager for this APT, "Overall, I considered the Phase II Final Review as a successful culmination of a very rewarding and productive project. The success can be directly attributed to the cooperation and determination of the people involved from both BLM and the contractor."

-33- 3 APPENDIX I PROJECT FLOW DIAGRAM

The following flow diagram illustrates the order of the detailed tasks and their interrelationships with other tasks that were necessary to completely define the approach to the Phase II technology demonstration.

I-1 PAGE IS fl OF POOR QUALM

I_a ORIGINAL F OF POOR

Al l I

s•

t v I

r I

M sS..3v 0

0 dc o^

= = y ^^ I e E E I =

.. i

le° i •

i PAGE IS OR^^INAR QupLiTY OF P00

10 e tiat.

►L

i ^^y

•

i-4 II I j

^ 1 I

I-5 ORIGINAL PAGE IS OF POOR QUALITY

i i

^ - R

Y Y `

I to

tt LS^^ e :Lit e. t ^^^ ^^ w ^

fit

all F It

I.5 ORIGINAL PAGE 19 OF POOR QUALITY

• •yi" i

• l^ / 9

ORIGMAC PAGE iS Of POOR QUALITY

6L61 'IlUdY EZ

t I

6L6T 'IIVdK 6 6L6T ZI2idV 9 — — —

6L61 93.3 9Z

• I-S n 51.6t xznr TT --- -axnr TT OF POaR QUALITY, P FL __ ---- ' Zn^ E-3rinr ^ •

ge^^

t M v

6L61 )Mr I-aKar T

v ^:

•• i ss

•• 7

isr.

6L e, t xKW T_

6L61 'IIHdK £Z I -9 6L6T AON T - . -I- . - - •6L6T S,O 0

ORIGINAL PAGE IS' OF P^OR QUALITY

!i fY

,. i fT

6L6T 100 S

6L6T 3Mnr TT 6L61 3WW 8T I-10 n 9 1 n •

L J

•

6L6T AOM T s

PAGr_ 6S INR^ ^^^^•^•^ .ORIG :t 000a 9.13 Of •

E: t

1 .z i s ;i^^'jt 11.1ts • ^s std

fly rr ^ a ' ;; Z H M

^ .^ 1 . i •c . r•I y w= Y G This appendix contains selected examples of the computer class menu descriptions. See Volume II, Appendix 2-G for the complete set.

II-1

i k F'

ORxmNAL PAGE IS OF POOR QUALITY NASA/BLM APT PHASE II ARIZONA CLASSIFICATION VEGETATION/TERRAIN MENU DESCRIPTION

Summary Class: 3 No. of Photo Samples: 1.222 I of Totalt 58 Name: Pin on-Juni er No. of Acres: 1,056,976 No. of Hectares: woodiana'- Spectral Classes: 33,35,41-43,46-47, Z of Area: -41 52-53,55-59,61-63,66,83-85

I. VEGETATION (x Cover by Species)

A. Trees Mean Std.Err. B. Shrubs--Desert Mean Std.Err.

Ponderosa Pine 1.0 .1 Creosote Pinyon Pine 6.5 .3 Bursage Juniper 17.8 •4 Blackbrush .1 .0 Other Tree 1.5 .1 Big Sagebrush 7.0 .3 Other Shrub 2.1 .1

C. Shrubs--Mountain Mean Std.Err. D. Riparian Woodland Mean Std.Err.

Gambel's Oak 3.2 .3 Cottonwood Turbinella Oak 1.6 .2 Willow Other Shrub 7.0 .4 Other Shrub

E. Grasses Mean Std.Err. F. Cactus Mean Std.Err.

Perennials •3 •1 Yucca .1 0 Annuals 2.7 •3 Other Cactus

G. Jon-Vegetation Mean Std.Err.

Barren (Rocky) 19.7 .7 3:=en (Sandy) 29.1 .8 Water Shadow .4 .1 c -

II-2 - PAGE IS ORIGINAL TV • w OF POOR Q UALI' e N +'1 .6 _I • • d ry ^L i /• ,~` ••••^' ^ Sri •^• r.

O O

1. a c

O rr a v m yam. r ^ ` 1/ './ ')'' '\• 1. ^^/. V1

N W W m, r /

V • O km fD 6m N N

- -

z 0 W w .a w H+ ♦ N X + A er M sued 10 2!)tllm3ud 11-3 S=A 10 ENLOOM

1 ORIGINAL PAGE 13 OF POOR QUALITY i i NASA/BLM APT PHASE II ARIZQNA CLASSIFICATION VEGETATION/TERRAIN MENU DESCRIPTION .

Summary Class: 6 No. of Photo Samples: 535 x of Total: 25 _ Name: Great Basin Sage- No. of Acres: 904,389 No. of Hectares: 361,756 brush Spectral Classes X1.28-30,32,34,38, Z of Area: 40,44,48-50,54,101-103,105-106,108,110,112,115-117

I. VEGETATION (Z Cover by Species)

A. Trees Mean Std.Err. B. Shrubs--Desert Mean Std.Err.

Ponderosa Pine Creosote .1 0 Pinyon Pine .2 0 Bursage .3 .1 Juniper 1.9 .2 Blackbrush .4 .2 Other Tree Big Sagebrush 12.3 .7 Other Shrub 8.7 .4

C. Shrubs--Mountain Mean Std.Err. D. Riparian Woodland Mean Std.Err.

Gambel's Oak Cottonwood Turbinella Oak .2 .1 Willow Other Shrub .8 .2 Other Shrub

E. Grasses Mean Std.Err. F. Cactus Mean Std.Err.

Perennials 3.2 .5 Yucca 1.5 .1 Annuals 14.9 1.0 Other Cactus .3 .1

G. Non-Vegetation Mean Std.Err.

Barren (Rocky) 27.2 1.4 Barren (Sandy) 26.5 1.2 Water Shadow 1.4 .5

II-4 oRitstNAL PA E 103% ~ N OF POOR QUALITY -

CIO

V • O m% = U9 N N

t- i -^--

- - t - - T _ - -1

3 St $ 3 St r=d t0 m" Mad I I- 5 STiffll AO Lwimood PAGE tS ORIGMALPOOR QUALITY 01F APPENDIX III

RESULTS OF RANGE PRODUCTIVITY ESTIMATION

The primary estimates of palatable cattle forage are in Table 1 for Landsat strata and Table 2 for allotment and pasture. These are the amounts available regardless of utilization. The secondary estimates, the current amounts regardless of availablility are in Table 3 for Landsat strata and Table 4 for allotment and pasture. The Landsat strata listed are the summary categories used in sampling and estimation. Results are tabulated in both pounds per acre and kilograms per hectare. The paired values in parentheses are 80 1§ confidence intervals based on the estimated standard errors. To check the precision requirement for the primary results, i.t is seen that one-half the length of the confidence interval across all strata is (16.7 -10.6)/2 : 3.05 lb/acre, which is 2?% of the estimate of 13.66 lb/acre. For the secondary estimates, the result is 238. So the precision requirement was not quite met, although it was fairly close.

The classification results combined into the 27 summary classes and the overlayed allotment and pasture boundaries provide a detailed cover type map illustrating the '_ocations of the various ranqe strata.

III-1 ORIGINAL PAG' 15 N ` _ _ .^. .•. .. OF POOR QUA o ^o In O a • f'f 1p N r O W a M O IA ^ t ^ O ♦ 10 vO vO ^..O vr- vO vO vO vO ft-0a v!^

9 W h- W \• < N co O ^O V N O O O M N 1 r- M 9 .^ W ♦ N N _J H S 1-- J fY J W W ^. .^ ..-. t. LL W W - r' fT ^p CA co co H OfY M •— _ u9 ♦ m fM•1 M tf_ ZW • • O • a 1 .^- h-- ti O W fa M H1 1 CA ♦ O N O W 1 O O O +O H ^i r.• v v v v v v rw v v v V Q

tL Y O m W N fY N ♦ ♦ Q^ .-- M a W W 1-- N O O O M N 1 O ♦ O mJ M N N •^ A J 1-- rr J t^ S WN e W N J HO CD M O N N 1!Y M O ♦ W rf^ M Q M Q .-. r.^Q f^ ^D W Lf_

Y J 3. CO O %O O ^D O N ^P ^O Z W ♦ N O < tD ♦ ^ iJ'f N ^ ♦C44 0\11f ^ • N f^ ♦ P'f W ^ h^ ^ O 1^ N N ^p V W W a N W tr N J y^ U n r W Ln N fT M d O aLf f^ f0 /^ Ln 41% M Of N M - M CL f- Wz: L L v u^ ^" ^ a ^ aLi f Qf ^ OCT >^ 0+ > O 1 0 A 1 %A A M g q N C D N ^ ^ N J N N1 l ^ 1^ N N^ ul H t Z y m 7 O N %A O y it 0 .r E LON ON .8 0%x m N L ^N V =NO tJ Vf NN L N1 Z Li L 7 t/1I v I III-2 M w O ORIGINAL PAGE N Of POOR QUALh D ^O I+f rC" d N IT Z^ ^p 1 1 1 t b O W O 1^f tZ Z t^ r O r• r r it r+ v v v v v v

' Z W ^ O S N Q ^ %o t7 N d ► 1 1 1 i 1 N C Q t+9 W; O 1 1 I 1 \O 1 t0 ar W r N r N O rrJ Z J J W W CL trf Z O^ Li W W t0 n OJ M N n 7 •..i d N ti Li > N • 1 1 1 1 1 o3 ZS . • 1 1 I 1 t La_ (D M N W d Z tD ^ O r r '"' H V vt .^ v v v v d O a 0 cm W ^ tT r N ^O W 1 1 t t 1 J d d r d O 1 1 1 1 M 1 d ^ a r N N CI J cc H Q J NW Q W H M N M ^ ^ Z n t!9 M r r O t0 O Ln O M O O C 1- v+ rN ro r- o o a o 0 0 0 W LA. c7z 0 se

J N n ^D O of n Ô n IH r tp Q^ Z W O M N ^D r ^> O q n p1 N N r N O n < N r r N N N N W Q VW ` V S W f a N W N Ln W W ^^p " Ln ^ N r n Q tY e+l Ô O^ Ln r O Ln CL r Ô d L19 t'! C%; r t!f Ô U., Q 9 i d .rW 01 V LL dc 1i 0^ I L I L I L -C L^ O O O N = C yy C 01 C yy 41 I L A L L L ; dr O G o a d o O. +- m ou u O t W 1 d O L. z JN }^ r OJ 7 r C► O Oi 90 01 b CI Q d 01 C C C C L C C L A L C C Y C C C _C N L• .- 7 A t 7 rL Q1 OHO tq C3, -2 '9 N ^ ^7 N9 C 1 O— da CA 3 cl v III-3 T ^„ a O e•f

c ORI INAL AGE S a UALI Z o 4 OF OOR 40 LAJ M

O2 W ^ M t W ^4^ M M Ln N W N J rr H 2 O i J cc J CLLa n 4. f.1 M 4: I • t0 c r d ^.r d M ^

F- L& O y W V H F- M ^^T yM v p.. C^ Q W V) ao 0 z aY dJ c. O2 m a co yW WJ co a O M F- Q ^ N ^ J m d J yW Q W N J 1-- d Z F•- O^ O O M v. O O W W C52 O

y N J n^D 2 W O dd M N ^J d • N O W 4 V y .r W N W O. y W W J °` aW[ ^ a F- t Q C" N U IL Z W N L V H ^ A i O W

J N e L A ^_ 6.1 ...i III-4 w S 0 RIGI AL AGE Y F P OR UALI da

dW W n N Co n N O M ^ t0 M C!1 1.' O+ O O eT O O O e'f M M Z W — N N r N N N c^+N N N N N O ^L Z OC H W Na CL J_ W F^ 2 O W J J W CC WO W O C9 %n •- Co N N N Co n'! W W O to O Cr H H U .... p- r N N N N V! m 4 pW J CL 3 r CL m O: CT N N e OI L! 1^ to Q^ J N ZO d N O .- t11 r• r• O cn O O • - r- W C H O 1-- W N mJ N J_ W O N %n to ► b d toW O+ Ln O Q coO tO N t0 Co N e' f t- W O^ r- N to O Ot LI! V P9 Oi O d w • • • • • • • • • • • t/1 .- O d V ti Z W W d Z 9 1 J Z to W ON 12% t0 1^ %n t0 1^ 1O CT 1l- O co CD .^ f\ t/f ^ to Ot N N Qt N G^ • • • • w • • • • • w • W c^'f ems ^ Ot ..r Q N ^ O I^ t0 1"+ ^ W ► ► V W CL N W W J O Co O O 4 e•1 to 4 to t0 V qr► Ull I^f er! /h I-- r• r- r• •^ yH C a 0.

v g

]\[ W Vp ~ Z s 7 ^ i

111-J e. ORIG NAL A k--^ OF GR UAL 0 N V a A O

W C A 1O n r r f^ r to %D 14% %0 .re W ,^ N r- O O O O eT e^ 2 N N N N N N N O = N N N N r OCW NQ d n.• J W n.• C9 1- O W J J W 0a O W 00 qw er eh e. co Q co a qr L a I.- d .^ Q^ CT CD C O Cb CD CO CD eT m ty r r r r► r r r r r r' Q J CL > CL CD G c! P^ ^ Q 1A J N O CO M C"! N en N N O W O O N O N ~ WJ r NLai e..• Q N W ^ W 7 N %D p ^ A OT en G7 Q W a n cD N cD ^ co Go O QC^ Nr • • • • • O Z Q u N r' N N ^D to 1A W W W Imp Ci' S W Q c7 Z

J Z O Q kn eT N Cr t N9 e^ C7+ co %D 1A W ^D ^O Q Q CG Q V; e W ^ r- •- r O V W CL N W W J cr J OD fT O .- N en t in .- N r. r N N N N N N kn eA " dc a r a

Nr Z

111-G IGINI 0 I 0 PO L QU kLI A CL

W ID N A IA d d r

N N Z 0 0^ N O N r N O X S r.n cx H W QN 0. W J C7 H O O J I.L O

O IA co d Il9 N N N O W W CC IA O N CC C4 CD CO CD d I+f Il'f C71 C4 I. W d O ^ ^ r r r r r r r r r 1-- W m 4N o7 J CL DC CL ^ CO) CD m 0 Ch c" CID %D d a w 1. N 4 m W zd O N O d O N O O d N O r' q qH Lhi is H Z mJ 1-^ Q N J W r^ B co W zip W O O IO Cl d o 4 ^ 0 co Q C ILn CNI m P1 m 1•• • _' N • • w w • • • • oC H Ill •• .^ I!Y t r9 O W W ^•• N O N Cob LA. CC ^.. W C m I 1 J 2 ul O t!! ^ Il9 I IINf d d CL ^ fd. ^ Ln f. •_ M d d N Vi Vi W I.7 • /. Ill ID N d • CD d r OI p- N In U W Imo.

W J kn M^ d d O Iff V% IA N N N C H ~ CL CL C1► = v a .-. W^ W Z J CD W

J d

111- OR INA PA IS w OF PO4 QU ITY 0 FIT - I r'

w d

W H 4c L!'! N • LA O %n O m 10 OL A !^ OL 1^ ^ ^ r r^ r r 2 O C:^ W C-4 r r r O a[ S rr OC H W Q d N W J O J W J

O W Ul O O O of .- + OL %D a0 (D W W w LH 1A ^ ^D ^ 1-- ^ ^D n ^ LL'1 ► V O F-^ W CD dN o'7 J a DG a d ^ N m O N LD LA O N dt- O O O N r- ^ O O •^ ^ W O O W K ^ J

J N ..r W a W cc f•'! VL r N ON OC Q ^D C Lt! aW ^ ^ Llf n ^ ^ • Ll'L N ^ OC N *1 O O h- < V P9 W W W

}J Z (+f LCL cc a N ^e O 0 W 1+ ^D 0^1 Kf • iD N O N At 'D N N L^1 W K C r u aW N

W W J on ^^pp J M ' LH An O d0 a+

s v

d 1Li-0 AG IS ORI N A ^- UAL TY OF OOR

W CL r d O N N N N CD wl N at N Z ^. u !7 M C4 V; ^ C4 O 0 .^ r r N N N► N N N 9 S CV. F- W CL rJ W O J ^a. J W

W Nf O pC co n ^D P. 1^ O %D N C" ^D W W ^D .- trf O CO lD A O 1A V 1 W; p- N I- P- r r W r N N "ff m d o ..1 D. C6K m O ^D ^ ^ P1 O CH 0^ O O a, ^ Ô N W r' r- O O O O O O O O r W mJ Q r J N r W 4 y O ^D W m 611 m .- ^D ID CD •^ d Ofi f^ r" m e A f a1 ^ ^D ^ ^ C N Nr `^ ^ w Nw w m • WO u C" N N W d I0 J Ô O N O r. CD N ^ ^ r'f0^ ÔDp Ô •- ^ C^ -W p p 1^ N A O O N N w w w w w w w w W w u W a N W J a

d d. S Z u_^ o ^ ^vx W

J III-9 FE I3 LI i .PO 0 ^o

w d

W P1 GC !•f • r• 10• ^ V t^ ^ !rf 1A 04 N N Z N r N O Y S ..r OL QH W N 4 J W O3 J L6 J Li O W W O fy m f %D O ^D O W O N n a O ^ V li t^ ^O N f .^ O • n 1f! N N N do~ W m ^ ^ r r r r 4e-k- O J r a6C ul O O O O J kH 01 O 61 Co N O N 0 r N W H O N O O O ~ W O

W6-1 CJ U'! N ul W ?4 Ln Ul %n N cn O : Ul • . • • • • N F^• • • • • 14 O N N 1ff V C f '' f W W W_ W t ^

Y J 2 CY O W 1lf f^ O N w r,; aO lY O ^ W %a co 2 ^• • • N en N O 1!f ^O N W CO f O N f r. •.r r V W CL I N W W J^^ J fr') — r- 1=- m M en f f f O Ov •_ O a CL w r F ^ Z ..s =O ^

iai-av w 0

RA ORIG ^' ^!. ^v Of F Or• d

Y O_ H N ICL W J_ O H O W J J li OW W W WD cr H H V Nf N W m Q 7 J a Y d co O N A dr O W O ^ W V4 h- mJ 1-- C N J W W W ^D a 4cc N N F^ ^ O t V W Z W W r'

LAJ t

i J Z N r+ O WoiC cft N y W ..r N'! WV aN WJ ^ J N- Mt- <► O ^s CL O.

J O

III-11

$. Uh OF F 00N j;%Llr 0 at fl% N 1A ^ ^ ^ ^• n r f,^ e^! < r A ► ^ ^ • N O w w W OI N uY 1 W b h O O O O n 14 vO vO v v v v v v v v

W W f^ 1 r N N 1^ J W O • co S N O O O N 1 J Y V en N N •^ rr W a F^ S X az W W

O CD C%j CD 4 W J ^ IZ In cn IL > d . C U. W co N ^ 1 ^ f^ 1^ Q1 O . ^ w . N 4 V F^ ► Z O O O ^ O O O ^D 10 ^D i'^ •"• v r„^ ^„^ ^ v v v v v v 4 Q o N aD O N h- aZ Z J Y m Ln 111 Lo 119 q 1 a Cn N Z aWiC 4! W W N O O O e; N 1 f' OC < cn4 N mA cc cc H V J h-N W W O O e•'1 N N 11'f e+f Y 'on -W O J < e" O e•9 O ^ ^O 1i Z 4 A W Q U. N O W K f V W d N Vf ^O n C14 t11 C1 W ^ vi N r e v^+ .- Q W . . J < • m W O r. eV f^ N 10 .-• fn Q < V W Q S Q a

at L1r1 qr D n CD n CD ^ CI W W 1f1 rl^ 10 f"! O^ N eT tr1 OI J

N r fn C®►

W LL ^+ ..r 7 d L W N 1 1 ^ ~ 1 1 ^1 N VV 17 ^ G L W 41 .^ N Cr N L O A C A L N 0 O H 'a 1 0 A 1 ^ 1 I t f^ v1 A se N A N C.0 kA L W 4A .0 NI 4A I J V1 O w ti Q N O 'fl 7 O U al A Q 7 N N 7 W.9 M +^ L al L x L a Q L A iC L A A cm K t^lL 9L L 7 Lfil •^06 L tCJ L A t to t L L C A C L ^C ^O -r mr• U eD V >N V Nf j s, NNN0'n4Az ..r III-12 v.. PAGI 19 0 0 IGINA N QUAI ITY W 0 P00 .-. A N _ D O eh N N ^ z W I rl 1 1 1 1 ; N r- N v v v v v v

W W eh e11 ^t N N J 4i 4 ^; 1 1 1 1 1 ^ V n O to Q1 1 1 / / cnN9 1 J W !' / ^ r rr a z a z ae d VZ `^ OeY d W J N n O z cr. • 1 1 1 / 1 iv O W ^ ^ N N Z ^ H- Q v v v ^ ^/ v Q NO Q Z HN QJ Z

Coo eT qr %O C N H woe ^ ^ 1 1 1 / i W Z f^ rl eao 1 1 / 1 O^ 1 R! H W ` ev N GT eY m •-r Q: J N V W W O J ^^yy cn n IA en .- .-- O ID O IA O /•9 O O O^ N O ^ O O O O O O O W J W O2 W Oz O N K 9 W V W ^ r1 /s ^ o+ Na N r. ^O O to ^C ^O W a N N N A O J N m ^ V N N N a W Q d0 WN arNW e•/ ^D O^ ^! t•f O .^ ID ^ N'1 frf N ^" IA ^D u1

Z .r2W W66 ^ 1 / C A N QQ N C 0 C 41 A 1 L O L L L t^ C V 1 di 1 0) L J N ^aGl ^a7 ^1^ C 7 mr V V i C C C C L C C 7 L L C C ,L C x C C r+ 7 A 7t 7 LL 0• A ^r O - A d7 N -2 N1 ^7 N W dO t^.= d N m v 11I -13 10-ft CR NAL AGE ! •W ^ G ORI ^ OF OOR UALI U. ^ ^ rN

O W OWC ^ J W \ C ^ ]OiL W N r J H = a V d z eae W W eC CL. .•. ^O w O W h t W 1-^ W J ti ? • Cie LV O W N c.; a z •- a 0 4A z

m W c N F- W x H W \ N ^ Z Cr m cic J 1-- 4 N W W

O O W c7 O O Z N W K n. WV a N 4A W n W e N J O Ul V Ul H W ^y a a h W a

NP'1 fAl

Y1 O O~ C D aJi[ ~ 0ac L q 1111• t ^o O S RIGI IAL AGE F POOR UALI a• t d pW^ < W O ^p ^ h !^'f 1!1 h Q^ h C" 1O • ► N N ^ N N N N O N N N N N N N N NCY - C7 W N N N ^^ aL 2 aC dw W J

4s W 0a W h N W p^ ^ .T r, N ^ f+f iO N O N W J yyi cc A O O N N N H ft.. N r N N N N .r m Q 4K J a y

m N a F- ♦ o N N f N r r r N C r ^- r O !• f O O W C H 04 10- ..r 47 F- m Vt W q LA r. O ^ Cp r r O N 1-- ► ^ r. W Of W O+ CC M .- r a N ^f 8 a Le . at O 3 N en r ^ r ^ N^ O r Ir. W Lai

Lai J

W_ h N V +^^pp 0- 1^^ N9 N h O% N N C^ O h C W^ N f O cc en cn W r r N J

H t

O. f 1A ^D co O r

icW

^ W

&^ w nsr

PI aaa .- Io 0 MIN kL P OzI 0 0 PO R Q LI N •

' W Itl f^ ^ 10 ^ f ^ ^" • O Nf ^ f" O O O O O O • N N N N N N N N r N N at W d W J

IL Z}

O W W W J < O 1!1 0 f mo N a at • 1ft N ► aC m ^ J ^ m CD O O • ^• n D O ^► ► H N CD r J

rW V 1/1 Nf A 10 ^ ^ W 1^ Al t^ n f 1 Ip QI 111 IA O N co O • • •► gyp• W I[1 f 10 ^ ID ^ • ^ ^ J O

4 cc 4A N A f 111 N 1"1 N N N N N s• 111 1A N ~O H d

O W

N ..n I A 111-10 %o OR GINR PA` L IS pool t 0 OF Q U " L'T'A

^n d4 L W 0. La H ^ fn O c" $A Qf n n m N fin Ml Ln %C N N9 a% N O^ O e7 r r erf C\9 W N N N N N N Y S OC W W a IW mJ J O 4 W azZIP OC39 W W DC W J ^p en a% co a% N ^ ^O ^ N N cr. ci O C Q fMf to A e+f CD O n O .- O+ O+ h- J r r r N r r r r r N w m N 4 Q J CL ZIP } Q m H N 1-^ --^ ^c a+ e•f W ^D rs a+ v+ e. n e^ CD Lai O N O Q O N O O N O f- N z W W cr en Q W 0. W h L eh ^n en G G^ °C co eh W -- co co n CO co V Vi lc O 4 ^.i 119 r• N O N in •^ .^ 1H eh W W W ^ N Ff CL S W J d 0 z z O 9 NW V W r 1n N Ln O N C" n %n a W ^p P'! {D 1n N to V 1w er N °L n n o n Ln %a eh V) W J F--

0. J N N N Q O 1A 6n vi 0. ~

O G ► V r 0 v r O V J00 N `^ J U.

II1-1/ PAGE IS w GIN IITY, 0 OE P0O Qu •

W CC O Ch a n ses A r M to f N r 0- r- O N m N Qz r !+f ^W N N N N N r" N N O ^ ^ ccW d W J W

O a} s

oG O W W t" co cc W J f 1O }^ f" N 1► r e^f N s ^p r. n O ^ ego ^ r-N V—fT P-O NO p-0; r r r N r► r rLn .r m a s J C6 } s 1 m N e^ N ^O O e\ r r f "o N r C14! rW 5 C O O O O N '- ^ O O f r r F— N W ig N W W co ^h ^p at cm C ^O .^- Oep 9 u9 r N fT e° V 4 ^ en f i9 u1 N O 61 f O ^ ^ r 1c9 t7 erf ^ ^ ^ W W a a LAJ J

W V W u'f ^O f a N ^D O eZ W Mf N cc N of IA O ^O N f *O f V1 . en of at m ^ V m W r f^ -J s a cc p^ O N A f u9 Go to %o %o %o oo %a O a

W v

O 'co ^v

III -10 w0 Y9 Q LI 0 PO R d

O OC ^p r Co co O en 01 + N /o . •...... C4 C4 4 10 N 01 1+ O O O 10 N Z \V r r r r N r N N N N r

W 0. W J W

O W a 0 ui W W N f- N O w /°9 C N r 119 O N a mJ N O O 119 N 4 r e4 119 Ln m r H Q \ r r r r r r N r r r N VJ m a a> J Y a m N O r %D 119 F- r r N9 Co 01 Q1 o O N C r O O to r W Y r r O O O O O O I►- H ^ H N Z W W ^ cc co r m 10 m cc al en D+ Q Q en 61 ^ W O r ^ N LO e•1 4 10 n C1 A N N r r r N r' O 0 W W 1►. Y 0W J

N9 W Or co v r- co 01 r n 119 W N N qw ON N e O O G. LAJ p1 O t^ N f.-. co. . . . . N cc . . . . Co < 1/f 10 r f t!f V ^ W mJ C H 6 4 d r r O Nf yam} a

W ^ "'Z H v p]d O

cr aC J J m TTI -19 .e w0 e RIG It AL P FP R NL!, 0

W •~.• ^ d 1O r r r d 1a7 10 d O N ^" 04 ^[ S O N N r N N N N N W d W W O Q J O 6 W Q 2 OOC W W W J ac CC O %n r 41 O O ^ O d ...... N J m r r r rp1 rM 1!1 O rQ^ r01 rd 4 J r N N d Q Q CID 4A o% O %n co h &n O O r M ^! N Q W ~' O N O O O O N r

H H N Z W W OC N mm W W M ^ 61 m N q k r N Nj C.7 ag 01 M O I t^ r r D\ d f^ r N V Q . • • w w . • • • . • t V r ♦ r r r ^ N N W O 1 r Li ^W WS N W Q O J I WN a.. V M d O O N r P9 !o Co O co n W N t Q. W Co %a d Co C" N oc ^ d O N d C N PZ N r W r M O r cn J r t/'! H Q CL cc 01 O r N M pH HQ N M Q Q n9 M d d d d ^^ r r ^ 1.+ F-F^ CL

.•a r CA r

=^\ ld

cc III-20 APPENDIX IV

RESULTS OF WOODLAND PRODUCTIVITY ESTI14ATION

The juniper estimates by allotment are in Table 1 for standard measures and Table 2 for metric measures. The corresponding pinyon pine estimates are in Tables 3 and 4. In each table the "Total Area" is for the entire area of interest while the "Sampling Frame Area" is for BLM land only since other ownerships were excluded from the sampling. The volumes per unit area are obtained by dividing the total volumes by the Sampling Frame Area. The paired values below the volume estimates are 808 confidence intervals. These are estimated from the standard error of the estimate and the expected deviation for 808 confidence based on a sampled Gaussian distribution. To check the precision requirement for juniper, it is seen on Table 1 that one-half the length of the confidence interval across all allotments is (185.2-142.3)/2 = 21.4, which is only 138 of the estimate of 163.7 cubic feet per acre. Therefore, this precision requirement was met. However, for pinyon pine in Table 3, (37.0-19.0)/2 - 9.0 which is 328 of the 28.0 estimate. This shows that the precision requirement was not met for pinyon pine. The column labeled "Relative Standard Error" contains the standard error divided by the estimate. The other columns in the tables are for average size (diameter) and average age in years, along with associated 808 confidence intervals.

IV-1 qr h ._.. 10^ ^• 2 ;; w+ •• WW ^N r V- ^h r%Dv twot

ORIGINAL PVT r°>^ OF POOR Quv Y..LT ^ ^ ^. N .-. • • N• • N %D

^p O m . • W w+ m v v v

W ^

H N N• N• ^ W H W ^ V1

•^l9 r HN W

N h• . f• &A ♦ P! NN N S W CQ '.CcP'! .• W W ^ ^ ^ N • h .•

•^1 N ^ v W v cc .A^ Ero .rzLa soc 7 N ^+ W w

WWy^ O ^gw N^ W v ^ v

r W v A W ^ A ^ cz • J A F r ._+ 1A ^ P1 86 W r A h, N ws s .. w, r

g V V V^

On* dimb MLg so ^ ORIGINAL PpA*H: sri y/ N 1A 1^ 10 14 11'1 r N OF POOR QLlALITY • N vN v!9 rN ^ ti=w W O N V- M• ^ N N • • t ^ • • atW ►N^ W

w! A N N N Z 2 W in pp r- = W ^ 10 ^ rib f^'llPf Q V J Z N •• (^! O •• ^ •• N = ^^ X1%0 w C) d vm vm v^O vA r-4A NW

H H M•• •.r .sue. N .^pp. 141 °. • 10 iD W N ^ N Pf }}... W .V. ^ 1 1 ^i d rob *0

!^ ^ 10 ^ I^ I^ ^ rN'1 O W N ^ W ~ v v ^ vN

_,.I =Nr W g N N WJW {t~.i N N N N Vol ^M WSW sort

W IV- 3 1^ = N S w^. ^^N ^'1 P %D PN smN W W ^N WA.SVqr M N t V V •••P

N N • N O N !•• in V; %n Y1 N N ! ^O • 1"f • N W ^+W ! • .•N f • .•N ^ .• O • U. P m P %D N N P f^ • _ • . r Nf O W^ P v P P V ' V V

W C3 30Cie {{ oo at 1^ N 0! 4cW 0_ Wa Z N ORIGINAL PAC's IS OF POOR Q1;E+.UTY .- •r O

O Lai W m O A N W tZ.^ P O f^^! W 4w •^+ } W N N 0"42 N qr N^ AL %D P VV v v v v ^ N r^ W N O N Z r. C6 %D y N I^ P W! m t ^ ^ A A W J P P. .0 ^m^ ^O t^ ^P% 00 Z •m ON PP v .r h^ 1^ •.. v v v W

= •rZ t• N N •^•. wi r 6-iJoi ^.^ O^ N hh. r a W N A N ^y 1+! A O W i ~ ^ M ^ N ^+ W

W W iv-

6 A A ^^ N Z N N •+ Q •• .• Ifs ••. 1p^ .• kn

pn • A ! I^ W p.. %D A f 0^ ► W Y1 10 ^f ^ ^ ^ 1N ^O • V V V ri .+

La C3 i ORiGiNAL PACE IS OF POOR QUALITY W !- W GL N

1.9 P_ in Z cc • • • •

H W N N H r Z N W OWN. ^'+ ^ N N wl V f•1 W W F- N O Z _ ^ W Oaf•! N • QJ Af ► N UJ2 NQ W 20 ^? N ♦ ^^ • Z ? W O O fi vO y v v .4 N A W b H N NH i W Z A W Z fa^ r ^ ^ O • N • O • on 9" Z vN • AL Zyj m ^ NM f y^ m O O f'y^ ••• 4n N• ~ s v v ^r 1^'! (i v }

Z ZN ^^W jW 1O !N f0 v^.0 N► N d Z V ^!f bi 1 CW ^ N WN ► J N Z N o a F _ N wrs W 2an

IV-5 APPENDIX V

RESULTS OF FOREST PRODUCTIVITY ESTIMATION

The estimates of Ponderosa pine board foot volume by Landsat strata are given in Table 1 for the Blackrock-Virgin Mountains Area and Table 2 for the Mt. Trumbull-Whitmore Canyon Area. The combined totals are in Table 3. In each table the "Landsat Stratum" column refers to the 27 summa-j classes. "Total Area in Acres" is the entire area of interest and "Sampling Frame Area in Acres" is the set of all PSU's which contain any pixels in any of the three Ponderosa pine strata. The volume estimates are in the next two colum;is. The values in parentheses are 80% confidence intervals. These are estimated from the standard error of the estimate and the expected deviation for an 80% confidence interval based on a sampled Gaussian distribution. To check the precision requirement it is seen that one-half the length of the confidence interval across all strata is (147-113)/2=17, which is only 13% of the estimate of 130 board feet per acre. Therefore, the precision requirement was met easily. The column labeled "Relative Standard Error" is the standard error divided by the estimate.

V-1

.j p ^ O O p• $0 O O O O O O Nf O O O Z O.

JK W N

N i+ N ^ N «. R W Loa: .. --_o %..p vs CP `-' A W cm^ 1 1 0 1 1 1 0 1 o o o •- W =mss rl H t C u.

W N W La< >N Es er W N ` W ^S^ p O O O O O O O ^ O Cp%O O S =F- C a ►- ? V F"' ~ W GZ "W W C. y1 . t G L^ & W O_ 1 N W LAJ 4^ W { .;,. 4c I 1 J CL W< O O N O O O ^ O 4w a • W W ^^ W L n.US N V.. W N 4A W 1 N eh d ^„^ m1•' P", N N r. 01 N N iii O = ^ ..^ nr N W Z

1 1 ^' ^' Lr+ L..r L O v ^^ W W 1 n( O CI q 1 1 0 ? U C 1 GI &AN 1 t C Nf C. ^[ D L ; ^$ N ON d O N d O NI IV q O V 4l O q p^ ^ C L ^+^.^+ y y n, d L U. 7 x L O. O A q ^C O► r. N 1"1 L A L A L O O t .^ L •— O t L C 1► out A A ^ O N N t/'1 m Z N Ct: Z N t9 N N N r

v-2 W N O HJ ^p O i 1!1 10 N N O ^ C%& N O O O O O CL

uh 0C>t ••• O _ _ cc A-% ^ fr/ Irf Nf N J C W 1 Il'I v O / v r t v i v v r ^ C>C N

r N OF 01 A ri-,IALIT O at r., p• W M PY IA A N z cc O v O 1 ^.+ N O v I v v v h^ Cf. N W V >K N W

W N1.- 1•^ W _ _ 8 1-• ^^ N N ^+ N N VII N W O O 1 O O l f- .J C SCr O> OCW W N 1 O N ^ C ^ t OCD ^ ^ r♦ U N La ' Z t N L5 Uj N M M t^^pf at ^O^pp N ' W CK O O O O O O O N O O N c0 I.- d LIMP ^W. Ci 0 } Q in ap H ^ U 3w N La E+ > W d

W ^+ W JI N ^ ~ N N In M1 ^ Q O N N9 I JLa W e•1 M A CT ► A w Z ^ N N ^ Q W W W CL'

uj 22 W ...• I.- V N O A Al 0 f ^1 LAI N IL1 ^ O u L"W ^W 1{'1 • CT qVr ^ ^ • • rO N N 1^« 1L/• f^• r ^ N F=•' O CD N frf ^ ^ Z 66 Z Z = N H wr Z 1 Or •+ N W Z

W La .J J r • tN H^ C G _ L N a L N W •7 L C A N AL N X N . A N O t I L 1. r O ou O•- L Z C N9 C dl q L 7 L O L z L ^ 41C Ca ^QV ac 1 J N C L ?^ •^- C 7 d r D Gi M V f1 A O C at A AA 1•• ^ N M H C ^7 ' t/1 w L O CS O CL O CL W 3 to a N N VVr^ O

V-3 p W O O O O O O O O O O f O m oo'^Z d

•}+ ^ O ^+ 1^ N! ^11 l^+f 1 v 1 v v 0 1 1 1 .j Z I 1 1 r / 1 W N t + N 1 a ppG^ ^ ORS Q'^AL ?ooR Np Of A PI A Nl !^! N Z 1 1 v 1 1 1 1 v ^.. v O 1 O W N V >K W W O ~ W Cie N v v v v W W 1 I O 1 1 1 O / O O O O 1 O No . ^ m a N O W y^ ^_ W N E ^ N^ i ► H m c o o O O O O O O O O O O H La W W N N

t C^1 J J t^^ O ^^'! ` p O ^ O O O W O A r ^ ^

W v^ N ^+ W cc

W«.r W

OZ ^ to N frf N W l~11 N 1^ `^ N9 r• N O •

CD w. N W !

1 1 ^ W W u r L ^` L L ^ ^ > > V L ^ CD ^[ m C ^ H 1 ^ ^ N N p ,r ^ ^ ^W 41 O d A {! L O C C L = ^^] NNNN .+ 1. «+ %n .+ u L q G V ^e L N N N D ^[ D L rA , 1 ii 7 ^O 7 N 7 d J N O d O OI O QJ 4 7 u a N Al Mo a L ^0 L 7 t • t L t G 40 A '^ L ^C a t v1 N v..+v O u ^+ u ^n d v'1 m v'+ ^ V9 ^ N d

V-4 v d1 O1 C J N ♦ 1! pp'$ O O ^- O O h O ZC► ^

W>aca O fw !n a N N

N ORIGI 0R Qa AEA O pF P O y Q^ e7 N O en frl Q• N asz cc 119 N P v 1 N P P ^+ I / O v P d W r H W N Ku W 1n z t W O H ~ W h 4 01 N /^♦ r. a1 ^t . N vh hQ1 S w Cie N m ^^,rie ♦ ^N P P N H '^!^ QNt19 QtV 1p O d N 1nn ti ..- J W .- O 1 cw O le•f ^ 1L1 0 1A I 1 O r ^ C 0 ^ -x ^ N ^ O i W ^ P f°I +1 m CL. P C G O V > t N

1 W Z 1 d tn W ^ m d LA^ h N ^ r ^ 1D P N o ^ O t 1i e^ y9 O ♦ ♦ O AI O S P I^ C)► %D P ^ .-i cc P !^ } r.. O N N CO M C%b • .G fX O 8 N W b N P N N ► > W p O OL W W •.^ I _O 01 O O at N ^ CA P 1s O+ r'► 10 W a N ^ N ♦ • N 1f! I W cn

z W N ^+ c W -, V W/- ^C •.n

ON _N ^D O ^' P O ~ ~ N 1n pp9 ^ e^yy V•.. ^ Wy. W N ♦_ h 1p t. d W n co O ^^f w a ,^ ^ N A ♦ 10 1n O

az 1 s w.s^ N W s

W ^ N W t V ^i ` . N C N L N W 64 N AL %/1 9 M J[ h. , T r 0 .^ 0 P F 17 cc c^ v ate+ L = t. c ma °,c 1c°, c p a^ ab a v ^r Aft ^. C c 7 K L C C c C C C L L O el A vvv O d a 32 < doc>°aa^ 39 as °1rr,

V-5

i i t

O O W

,°^ o 0 0 0 o a 1^1 0 0 0 In o ao

wo 1cc. en Jt W 1 1 v 1 1 1 1 1 v 1 v I w F- at v1

N N t • ORIGINAL FACZ N OF POOR Q A'l O C" N v 1 / 1 O 1 v v Q O I.- ti CL H W U Wx NW 1— N H O W s. WzV •'+ N .^ ..^ N m N W ^ CG ow / / O / / 1 a 1 ^ o 'D O O J W V J y d. dc • CO S ^ O n ^ 2 OC .-1^ W 4 O W W A ^ ^ oc «O., N N H ^ O1c I.- W N N ^ 4un ^ W O W O O O O O O O O Ln O O O O cr NCD O W f.. W 4 D O at a ^O WN =W O o W C W G W J r+ W 2 ~ S N ^ / IL W O 0. W 1 J v O O ^t O O B a0 O o^ m tf a+ N O w 1! cc O H cc U W F- W a QU W 4 Z Li N N N O O+ 1+ w .- G F. &n .- O kn O !l% V- 0-- r M O m N V Z u. W ,q q• O gt 1[ I N N 4l m ••+ Q L6 W w r f an La. '.> > 2 N N cm N M MU. z r

1O^ ►Z+ 1 C6 ^^ J Z • /n w d 1 1 N ++ Q 2 1- Y cc 'C A O L 01 N 1 1 ?C 1 to 't 1 V I C I N q A = 0 N Ct q 41 L O q C L C H wi O •t ^•+ L N U L C^ ^d C S7 R9 y L I z o 0 o a '- q / 3 1 to M^ VI H N C D L 'p 1J L .-- N A d^ N g _.1 %A a n, o a, o a ^v u a, 3 o +n o v+ al ?°, Q $ ^ W N 4` 41 i L q yC L a O < q L ^6 ^C L7^ QI .+ irp i^0i mac ^- -.e •-o t iCCiyr :e A .- ev / 1 f- N N a H CA v1 3 V1 i!1 c7 s/1 n .- ►

} V-6 1 z O O W ^ ^ ^ tff O^ N ^ 1f/ co O= C3% co N O •V- O O A O ^O Z d

.?.. Q p /— O of ^ ^ er'1 a ^ e•f a/ e/ e+/ .—

Z H

J ORIGINAL- r ^:. 4 O OF POCR !^ ^ O O ^ N / '- Z eat .^i v O v 1- N

W $$H 1^ Q n-4 G W 0 CD ^ ^ O r, 12, N e^10N O^MP% r N ~ W p rf Q en ff) ul m v O^ ^D 7 W Z OJT to / O O in ^ N O / n ^ -4 Q Ow ONin 0 r W S C M v V 0 en 1 W go Z Z O to N N Im %a ^p el O $4 ► co e) Ln P12- ^ 1M0 1L N d W I- Go ;1.1 e', an O O O O CD Q w O N o 0.- y r 1^f t! O N N eh N A ^W Ey dO

t W N 1l9 ^^^pp 4 1^ &n h 1^O ..' Q Z eh eh ^D !T A O r O a. ^+' N N O O N ell m W

!\ r O r- u'1 10 v s0 e^f W !Fi tT ► eh Pf t N ^ O ,t ^ N H C Z

e 41

/,. 7 L A .0 N Y N X N Y Al 1 t 1 L L ; O A O •r' O r' L O C N C a rJ L 7 L O L L s Q CO 0 J O O CX ++ M b Li 41 1 40 1 a1 O L 1 ?^•^ 41 p 41 ^) r/C e 4 J H i C 7 4! + 47 d 0 1•• C ar C C 7 L •C L C C C C C C r0 ++ L .^ L tT O .- •- O O t t eT O O ••' Q •r t A G C q 906 .- -^ X cn V 4 d a CL a v: Z r^ O N V-7