EPA/600/R-93/135
                                         July 1993
         GLOBAL CHANGE RESEARCH PROGRAM


  NORTH AMERICAN  LANDSCAPE CHARACTERIZATION

            (NALC)  -  PATHFINDER PROJECT




                  RESEARCH PLAN
     US Environmental Protection Agency

       Office of  Research Development




              Prepared by:


     Remote and Air Monitoring Branch

Environmental Monitoring Systems Laboratory

           Las  Vegas, Nevada




  Wayne N. Marchant, Laboratory Director




        Bill Forte', Branch Chief




   Ross Lunetta,  NALC Technical Director

              702-798-2175

                                 re&> Printed on Recycled Paper

-------
                       LIST OF AUTHORS
   Ross  S.  Lunetta,  US  Environmental Protection  Agency  (EPA),
Environmental Monitoring Systems Laboratory  (EMSL), Post Office Box
93478, Las Vegas, NV  89193-3478.

   John  G.  Lyon,   Ohio  State  University,   Department  of  Civil
Engineering, 2070 Neil  Ave., Columbus, OH 43210-1275.

   James A.  Sturdevant, US Geological Survey,  EROS Data Center,
Sioux Falls, SD, 57198.

   John L. Dwyer, Hughes  STX Corporation, EROS Data Center, Sioux
Falls, SD, 57198.

   Christopher D. Elvidge, Desert Research Institute and University
of Nevada-Reno, Post  Office Box  60220, Reno, NV  89506.

   Lynn K.  Fenstermaker, Desert Research Institute,  Post Office Box
19040, Las Vegas, NV, 89132-0040.

   Ding Yuan, Desert Research Institute, Post Office Box 19040, Las
Vegas, NV, 89132-0040.

   Steve R. Hoffer,  Lockheed Environmental Systems and Technology,
980 Kelly Johnson Drive,  Las Vegas, NV, 89119.

   Ridgeway Weerackoon, Desert Research Institute, Post Office Box
19040, Las Vegas, NV, 89132-0040.

-------
                              NOTICE
     The information in this document has been funded wholly by the
U.S. Environmental Protection Agency under a contract.   It has been
subjected to Agency review and approved for publication.
                         ACKNOWLEDGMENT


     The assistance of Dorsey Worthy,  U.S.  EPA/EMSL Las Vegas, Len
Gaydos, U.S.  Geological Survey, Ames  Research Center,  CA, David
Peter.son, NASA Ames Research Center,  CA,  and Russell Congalton,
University  of  New Hampshire with the  review of this document is
greatly appreciated.

     The  NALC  Peer Review  Panel is also  acknowledged  for their
assistance  and thoughtful comments.   The  Panel includes:  James
Lawless,  Chair,  Jack Estes,  Ed Sheffner,  David  Peterson,  Len
Gaydos, Douglas Muchoney, John  Townshend,  and  Charles Dull.

-------

-------
                              CONTENTS
NOTICE/ACKNOWLEDGMENT








PROJECT OVERVIEW






  Introduction




  Areas of Coverage



  Satellite Data Acquisition




  Standardized Data and Products



  Methods Development



  Project Schedule




  Global Change Science Issues




  References
Page 1
CHARACTERISTICS OF SENSORS AND NALC-PATHFINDER DATA SETS
                                                      Page 11
  Landsat Program History



  Multispectral Scanner  (MSS) Data Characteristics




  NALC Landsat MSS Triplicate Image Data Sets



  Image Composites



  Categorized Images of  Land Cover



  Change Detection Images

-------
METHODS AND RESPONSIBILITIES FOR PRODUCTION OF DATA PRODUCTS






                                                      Page 14






  Acquisition of Current Data Products



  Identification of Archival Data Products



  Image Processing of Triplicate Scene Data Products



  Landsat MSS Data Characteristics



  Image Processing for NALC MSS Triplicate Production



  Background



  Data Acquisition



  Triplicate Processing



  Image Rectification Analysis Forms










IMAGE CATEGORIZATION AND DEVELOPMENT OF LAND COVER DATA PRODUCTS






                                                      Page 28



  Training Set Development



  Evaluation of Training Sets



  Categorization



  Identification of Land Cover and Classification System



  Data Production and Distribution



  Quality and Accuracy Assessment



  NALC Products

-------
METHODS FOR MANAGING AND DISTRIBUTING NALC DATA AND PRODUCTS
                                                      Page 34
  NALC Data Archive, Management, and Distribution




  NALC Data Distribution



  Query Procedures



  Product Media Available to the User



  Standard Product Ordering Procedures



  Alternative Products and Ordering Procedures
IDENTIFICATION OF CHANGE DETECTION METHODS
                                                      Page 48
  Introduction



  Change Detection Literature Review



  Optimizing for Digital Change Detection



  NALC Scene Selection and Preprocessing



  NALC Change Detection Strategy



  NALC Change Detection Procedure



  NALC Change Detection Evaluation Criteria



  NALC Change Detection Research: Technical Work Plan




  Study Sites



  Methods  to be Tested and Evaluation Criteria



  Post-Categorization Change Detection



  Enhancement Techniques for Change Detection



  Change Detection Through Vegetation Index Differencing



  Change Detection Through Principal Components Analysis



  Change Detection Through Regression Analysis

-------
                                                  Page 63
  Change Detection Study Time Frame

  Change Detection Correspondences

  References


NALC QUALITY ASSURANCE AND QUALITY CONTROL



  Standard Operating Procedures

  Quality Control

  Data Analysis Tracking

  Spatial Accuracy Verification

  Thematic Accuracy Assessment

  Field Verification

  Sampling Design

  Sampling Scheme

  Sample Size

  Verification Procedures

  Thematic Accuracy Assessment Reporting

  References


APPENDICES



  Appendix    I:

   NALC Technical Review

    NALC Technical Review Session Agenda

    List of Invited Participants

    NALC Review Panel
Report of the Technical Review Panel, North American
 Landscape Characterization, February, 1993

NALC Project, Response to the NALC Panel Review Report
                                                  Page 76

-------
Appendix   II:
                                                    Page 111

 Agreement Between INEGI and USEPA
Appendix  III:
                                                   Page 128

 Path/Row Data Acquisition List, Triplicate Assembly Priorities
Appendix   IV:


 Triplicate Image Selection for All Path/Rows
Page 140
  Caribbean

  Central America

  Chesapeake Bay Watershed

  Mexico

  Hawaii

  Alaska

  Western US

  Eastern and Southern US

  Midwest and Great Plains US
Appendix    V:

 Interagency Agreement (IAG) Documentation
Page 344

-------
Appendix   VI:
                                                   Page 356
 Mexico Worksheet for Acquisition and Determining Temporal
  Windows for Acquisition of Remotely Sensed Imagery
Appendix  VII:
                                                   Page 363
 NALC / GCRP Coordination / Collaboration Documents
Appendix VIII:

 NALC Product Refinement Workshop

  Agenda

  Participants

  Letter Reports

   Houghton

   Lawrence

   Salas
Page 378
Appendix    IX:
                                                   Page 396
 Landsat  Pathfinder  Initiative and the Canada Centre for
  Remote  Sensing  (CCRS), and Documents from the December,
  1992 Meeting with  CCRS.
Appendix  X:
                                                   Page 406
  Land Cover Categorization System for Use in the NALC-Pathfinder
   Proj ect

-------
                       FIGURES
Number
5.
 Concept
                                   *
Schematic Scatter Diagram of Date 1 versus
 Date 2 for a Given Band
                                                      Paoe
1.

2.
3.
4.
Processing Flow for NALC - Pathfinder, Landsat
MSS Triplicate Production
NALC Pathfinder Data Flow
NALC Pathfinder Categorization
Landsat Pathfinder (LP) Program,


System
Data Management
21
22
33

                                                             38
                                                             62
                          TABLES
Number
2.  Interim NALC IMS Metadata Database Schema

3.  Definitions for NALC Interim IMS Database Schema

4.  8mm Tape Mapper for NALC Triplicate

5.  Example of Data Descriptor Record  (DDR)

6.  Explanation of NALC Tape Mapper

7.  NALC MSS Interpretation Verification Form

8.  NALC Photographic Interpretation Verification Form

9.  An Example Error Matrix Table
                                                      Page
1.  Schedule for Triplicate Assembly and Application
     Project Starts
                                                         5

                                                        39

                                                        40

                                                        43

                                                        44

                                                        47

                                                        72

                                                        73

                                                        74

-------

-------
        NORTH AMERICAN LANDSCAPE CHARACTERIZATION  (NALC)

                        LANDSAT-PATHFINDER



                         PROJECT OVERVIEW
  Introduction
     The North American Landscape Characterization  (NALC) project
is a component of the National Aeronautics and Space Administration
(NASA) Landsat Pathfinder  program of experiments to study global
change issues.  The NALC program is  funded principally by the U.S.
Environmental Protection Agency's  (EPA's)  Office of Research and
Development  (ORD)  Global Warming Research Program (GWRP),  and by
the U.S. Geological Survey EROS Data Center.

     The purpose of the project is to produce land cover and land
cover change data products  at a 3.2-to-5.8 hectare (8.0-14.2 acre)
spatial resolution across  a major portion  of  the North American
continent (Central America, Mexico, Caribbean and Hawaiian Islands,
and  the  United States) .   The NALC  -  Pathfinder is  designed to
contribute  to an  important  objective  of the U.S.  Global  Change
Research Program by "documenting global change."  In  support of the
EPA's GWRP, NALC will assist in providing primary land cover type
and  extent  data  for use  in national inventories  of  terrestrial
carbon  stocks  and trace  greenhouse   gas  (methane,   nitrogen
compounds) emissions, and  support  the carbon cycle  and trace gas
dynamics assessment objectives of  the  Intergovernmental Panel on
Climate Change (Anonymous 1993) .  Principal U.S. NALC collaborators
include the  EPA's  Environmental  Monitoring Systems  Laboratory at
Las Vegas,  Nevada (EMSL-LV),  the U.S. Geological  Survey's EROS Data
Center  (EDC)  in  South Dakota,  and  NASA Ames Research  Center at
Moffett Field, California.

     EPA  research funds  will  be  made  available  to  support
application project research beginning  in Fiscal Year (FY)  1993.
Application project research will focus  on the development of land
cover and land cover change digital data base products and on the
assessment of product accuracies.  Two remote sensing consortia are
being developed to provide application  project  support in Mexico
and Central America.  Funding  for application project research in
the  United  States will be determined  on  a  competitive  basis
beginning in Fiscal Year 1995.  A total of seven funding assistance
agreements are expected to be  initiated  in support  of the  NALC -
Pathfinder project by 1996.

-------
     The objectives of the NALC project are to produce standardized
remote  sensing  data sets, develop standardized analysis methods,
and derive standardized land cover change products for the majority
of the North American continent.   All data, methods and derivative
products  will be  made available  to the global  change research
community.   Data will  be vended at  the  cost  of duplication and
distribution.  All data and products will be archived by EDC and
listed  in  their  Information Management  System  (IMS)  so  as  to
provide user  access and data distribution support.

     Information about  NALC  products  and ordering may be obtained
from:  Customer Services, EROS Data Center, USGS, Sioux Falls, SD,
57198,  (605)  594-6151.
  Areas of Coverage
     The NALC project includes all of the North American continent,
defined as the area north of the Panama-Colombia border including
the Caribbean and the Hawaiian Islands.   Due to funding restraints
and  limited  Landsat data  holdings  in U.S.  archives,  NALC  is
currently not funding the assembly of standard NALC data sets for
Canada.   Canadian  archives contain  an extensive  collection  of
Landsat Multispectral Scanner  (MSS) data from the 1990's, 1980's,
and 1970's that could be used in support of a NALC/Canada effort.
Currently, work is under way on a Great Lakes Watershed Pilot Study
under  the leadership  of the  Canada  Centre  for Remote Sensing
(CCRS).  The future expansion of NALC to other priority watershed
locations in Canada  will be dependent on the results of the CCRS
Great Lakes Watershed Pilot Study.
  Satellite Data Acquisition
     Satellite data used  in the NALC project include Landsat MSS
data from the years of 1991,  1986, and 1973, plus or minus one year
(whenever  possible) .    MSS data  are  the  principal  source  of
satellite measurements for  the NALC program.  This is for several
reasons, including the 20 year MSS digital  data archive and the
relatively low data costs as compared to Landsat Thematic Mapper
(TM) or SPOT Multi-Spectral  (MS) data.  A total  of  803 Landsat MSS
scenes were acquired  during  1992 to help complete  the twenty year
MSS data archive.  Data collections for the 1990's  epoch included:
Mexico  (102  scenes);  Central America  (34  scenes); the Caribbean
islands  (48   scenes);   Alaska  (182  scenes);  Western U.S.  (138
scenes); Chesapeake Bay Watershed (18 scenes); Eastern and Southern
U.S.  (142  scenes); Midwest and Great Plains  (131  scenes);  and
Hawaiian islands  (8  scenes) .  In addition to  the 803  MSS scenes
acquired for 1990's by NALC, a comprehensive  review of the Landsat

-------
MSS archive has been conducted, and historical MSS data have been
selected from the 1970's and 1980's to complement the 1990"s data
acquisitions.
  Standardized Data and Products
     Standardized  data sets  used  in the  NALC project  feature
Laridsat MSS  triplicate images  that have been  georeferenced and
presented  in earth  coordinates.    In addition,  image to  image
coregistration  between  the  triplicate   image  scenes  has  been
stressed to provide high  quality  and standard data sets for land
cover change analysis.  The NALC standard triplicate data sets will
include the following:   1)  Coregistered/georeferenced Landsat MSS
triplicates;  2) Digital Elevation Model  (DEM) data;   3) spectral
cluster or categorized data coverages corresponding to individual
NALC triplicate MSS scenes; and 4)  change detection  images.  All
data sets will be clipped to the Landsat World Reference System 2
(WRS2) image  sampling  frame.   In the case  of the  1970'a  images,
which  conform to  the  World Reference  System 1 sampling  frame,
multiple scenes are being assembled and clipped to  the WRS2 sample
frame boundaries to provide complete coverage.

     An important aspect of the NALC project is the generation of
standardized data products for the North American continent.  These
derivative products include:  A  nearly border-to-border land cover
(vegetation)  type  categorization  for  priority  North  American
locations using the 1990's MSS imagery; and  land cover change data
from the early 1970's to the early 1990's.   Land cover change data
sets will be processed only for those areas where vegetation change
has  occurred between  successive  dates  of  the triplicate. .  For
example, scene by  scene change  detection analysis  will yield two
products:  1) A change  detection image indicating the locations of
change within a  scene between 1990'a  to  1980's  and  1990's  to
1970's;  and  2)  a raster  data coverage identifying the land cover
type(s) that existed at specific locations and that had undergone
change.

     It should be noted that the CCRS Great Lakes Watershed Pilot
Study differs from non-Canadian NALC project activities  in that the
Great Lakes Pilot will  result  in the development of a 50-meter MSS
mosaic data  product.   This  study is  actually a parallel  project
being conducted in association with NALC and using NALC products.
Subsequent to the completion of the pilot, the CCRS will evaluate
the utility of a NALC type effort  for Canada.  EPA is hopeful that
the  Great  Lakes  pilot  will  eventually  develop  into a  larger
NALC/Canada effort.

-------
  Methods Development
     Specific issues related to development of NALC remote sensing
standard  methods have  been addressed.   Methods  issues  include
determining:   1)  The time periods when  satellite  data should be
acquired for most of North America (by path/row  image scenes) ;  2)
the characteristics of automated Landsat  MSS scene georegistration
and image-to-image coregistration;  3)  the  procedures necessary to
develop  cloud  reduced MSS  image  composite  products;   4)  the
algorithm for spectral clustering;  5) standard methods development
for land  cover class labeling;   6)  image  difference algorithms,
post-categorization procedures, and standard methods to be used for
land  cover change  analyses;   7)   data  indexing/archiving  and
metadata  Information Management Systems (IMS);    8)  NALC Quality
Assurance/Quality  Control   (QA/QC)   procedures;     9)  accuracy
assessment methods;  9)  land cover classification system;  and 10)
Landsat  Thematic Mapper  (TM)  swath  data  acquisition and data
processing methods.

     Whenever possible  NALC will utilize existing remote sensing
data  analysis methods  that have wide acceptance in  the  remote
sensing  community.    A specific  goal  is  to draw upon  the most
conservative  methods that  will accomplish NALC  data  processing
objectives.   This  emphasis towards  conservative  methods  is  to
provide  standard  methods   that  can  be   employed  by  as  many
researchers  as  possible.    However,  there are numerous  remote
sensing methods development issues that will require new approaches
or significant improvement to existing methods.  Methods currently
under development  include:   1)  The  automated  georeferencing and
coregistration  of  Landsat  MSS imagery;    2)  the  development  of
accuracy assessment procedures for large  area accuracy assessment;
3) the development of a NALC land cover classification system; and
4) the development of Landsat TM swath data acquisition and data
processing methods.
  Project Schedule
     The  schedule  for  the  assembly  of  NALC  triplicates  and
application project starts is presented in Table 1.

-------
                              TABLE 1.
   SCHEDULE FOR TRIPLICATE ASSEMBLY AND APPLICATION PROJECT STARTS
  Project Schedule
     The following  schedule  is based on  continued-level  funding
through FY 98.
TRIPLICATE ASSEMBLY

1993  Mexico

      Chesapeake Bay Watershed

      Caribbean

      Central America
1994  Alaska

      Hawaii

      Western U.S.
SCENES

  102

   18

   48

   34


  202




  182

    8

  138


  328
APPLICATION START

       1994
       1995
1995  Midwest and
      Great Plains

      Eastern-Southern U.S
  131
  142
       1996
                                   273

-------
  Global Change Science Issues


     The principal clients of the NALC-Pathfinder project are the
U.S. Global Change Research  Program (USGCRP) ,  and the U.S. EPA's
Global Wanning Research Program  (GWRP).  NALC fulfills an important
goal of the USGCRP by "Documenting Global Change".   An important
objective of  the USGCRP  is  to document global  change through the
establishment of an integrated, comprehensive, long term program of
observing and analyzing earth  systems change  on a  global scale
(Committee on  Earth  and  Environmental Sciences 1993).    The NALC
project is part of  a  larger global effort known as  the Landsat
Pathfinder  Science  Working  Group.   Other   companion  projects
included in the Landsat Pathfinder  series are being conducted in:
Amazonia  (NASA),  Central  Africa  (NASA),  S.E. Asia   (NASA/EPA),
Coastal Brazil  (EPA)  and Eastern Europe (NASA).

     The Environmental Protection Agency's GWRP research projects
will benefit  from NALC  data products focused on  science issues
related to  carbon and trace greenhouse  gas fluxes.   The carbon
measurements will be  used in  global climate change models to refine
predicted changes in  land cover distributions.   These contributions
will be used to estimate changes in carbon stocks and carbon fluxes
from the terrestrial system  to the  atmosphere,  and in attempts to
identify carbon sources and  sinks.

    NALC products will provide  inventories of  the current status
and changes over the past 20 years  for specific land cover types.
These data will be  used by  global  change research  scientists in
combination with literature  values  or new field measurement data,
to calculate standing carbon stocks and estimate carbon and trace
greenhouse gas fluxes over the past  20 years. NALC land cover data
base products could also be  applied in the identification of North
American  locations   suitable  for   growing   land   cover  types
(vegetation)  to capture or sequester carbon.

     The general change in carbon abundance or flux estimates for
carbon in North America over  the past 20 years will be based on the
changes in land cover from one type to another  (i.e., forest cover
or "woody" to agriculture  or "developed  land") .  NALC will also
attempt to categorize the different age classes during  the process
of forest regrowth.   These data  will be particularly important for
providing more accurate estimates of carbon content for individual
forest stands.

     Digital data products to be provided in support of carbon and
greenhouse gas studies include the NALC "triplicate data sets", and
nearly border-to-border  land cover (LC)  and  LC-change  data for
priority locations in North America.  NALC standard land cover and
land  cover  change   digital  data  base products  will have  both
immediate  and  future  applications in  support of  global change
science issues.  The  1991 +/- one year border-to-border, land cover

-------
product for North America will have  the following applications to:
1)  Provide data  input  for  the calculation  of standing  carbon
stocks;  2)  provide data  to determine areas of anthropogenic change
in  land  cover;  and  3)  provide a baseline data set  to evaluate
future shifts  in land cover  attributable  to  future  changes  in
global climate.

     The importance of land cover type data for calculating carbon
stocks is illustrated  in the following  equation (Sheffner et al.
1993) .

     Total North American Carbon (LCi)  = A (LCi) x C (LCi)
       for land cover  type "i"
        where:
             A  (LCi) = Area of land cover type i, and

             C  (LCi) = Carbon per unit area or carbon density
                         for land cover type i
     The level of  confidence  of total NALC North American carbon
estimates will be no greater than the uncertainty associated with
either or both of  the above terms, A or  C.   A general consensus
among researchers is that the  uncertainty  associated with the area
term, is at  least  as great as  that of  the carbon stock per unit
area or carbon density measure  for a given land cover type (Skole
1993).  Also, the amount of carbon contained in forest land cover
is much greater than other land cover types. Hence, measurement of
forest cover types  has the highest priority (Houghton et al. 1992).

     Current estimates of deforestation in the United States are
largely  based on  statistics  data  compiled by  the  U.S.  Forest
Service and state government forest management agencies.  Most of
these  statistics  provide  limited  information   on   the  spatial
distribution of forest resources.  Although important data (e.g.,
timber type and diameter breast height, DBH) for the calculation of
carbon stocks are typically compiled, data from state inventories
are  not easily  accessed and  inventory  data  are not  typically
collected for private land holdings.   South  of  the  U.S. border,
forest  statistical  data  in  Mexico  and  the  Central  American
countries are, for  the most part, unavailable.

     To provide a consistent source of data to support  the study of
global deforestation rates and to provide data inputs to greenhouse
warming  and  global vegetation models requires a  project such as
NALC.   Scientists  have  been utilizing NOAA Advanced  Very  High
Resolution Radiometer (AVHRR)  data for related purposes.   AVHRR
data have and will  continue to provide an  excellent source of data
for  global  change  research.    The  temporal  frequency of  data

-------
collections  (daily) , the low data volume,  and inexpensive costs of
acquiring these data contribute to its utility.  EPA global change
modelers  addressing  terrestrial  systems  response  and vegetation
redistribution  find  the 1.1  km by 1.1 km  spatial  resolution of
AVHRR data applicable to their models which require data input in
one-degree longitude by one-degree latitude  cells.  Although AVHRR
data meet many  of the  remote sensing data input requirements for
current global change modeling efforts,  serious problems can occur
when limited to only one remote sensing data  source.

     An  important factor in  evaluating the  utility  of  any data
source being considered as data input for monitoring, inventory, or
modeling studies is the accuracy and precision associated with the
data.   In remote sensing studies  data  are  validated  as  to their
accuracy and precision in one of three ways.   The first two methods
involve  comparing  remote  sensing  results  to  in  situ  field
measurements or comparing  the results  to another  remote sensing
data source  of known accuracy  and precision.  The  third method
involves  comparing two  remote sensing data results of  unknown
accuracy and precision.  By evaluating the  agreement  between two
data sets, a quantitative evaluation of data  quality can frequently
be made.  It is best  to employ one of these  three approaches, as
there is a significant potential for error, when data are assumed to
be accurate when in fact they have not been validated.

     An  analysis  of  various  remote  sensing  data  types  and
approaches  for  monitoring tropical  deforestation were  evaluated
under a NASA-sponsored  research project titled  "Landsat Tropical
Deforestation Project".   Results  revealed  that  Landsat  Thematic
Mapper  (TM)  and  SPOT  Multi-Spectral  data  provide  comparable
estimates of standing  forests (Skole and  Tucker  1993).   However,
when compared to AVHRR results, the AVHRR overestimated the area of
tropical forest  stands  by 40  to 90 percent.  A  comparison of the
rates of  tropical deforestation  between  Landsat and AVHRR data
indicated  that  AVHRR  results often  overestimated  the  rates of
tropical deforestation by approximately 200 to 400 percent  (Skole
and Justice 1992).

     The NALC data products will provide data of known  accuracy and
precision  to facilitate  the evaluation  of  other commonly used
remote sensing data for global change research, e.g., AVHRR.  NALC
MSS data products will also facilitate the study of sub-kilometer
scale  science  issues,  i.e.,  "patch"  sizes  and  the change  or
migration  of land cover types.    In particular  the NALC data
products will be useful in the study of  "landscape pattern change"
or "early indicators" of global change, such  as induced shifts in
biogeographic regions or spatial distribution changes of land cover
within a given  region.   These changes that  would not  be detected
with coarse resolution  sensors,  and they  could  have  an impact on
ecological function prior to  migration events.

-------
      The Landsat  MSS  "triplicate  data  sets"  or  "triplicates"
 represent the NALC standard data set.   These  data  sets  alone will
 represent a substantial contribution to the study of global  change
 and will do so well into the next century.   In addition  to their
 application in global change research, NALC  triplicate data sets
 will provide  an unparalleled  data  record and  will potentially
 contribute   to  the  study of  numerous science  issues  in North
 America.  Examples of future science applications include:  1)  The
 documentation of changes in biodiversity;  2)  contributions to  the
 future mapping of  North American biotic community zones; and   3)
 the analysis of future geographic shifts in vegetative communities
 related  to  changes  in global climate.

      The  standard  NALC  LC-change products will provide important
 data for  global  change  scientists  as they provide  a  retrospective
 "picture"  or view  in time useful  in  evaluation  of a  number  of
 important global change science issues. Issues applicable to NALC
 LC-change products  include:  1)  The measurement of change in land
 cover or change in carbon abundance (carbon flux) over the past 20
 years;  2) estimating the potential for storing carbon in  the form
 of  land cover  (carbon  sequestration)   in  North America;  and  3)
 estimating  the flux of trace greenhouse gases  using change in land
 cover as an  indicator.  Carbon and trace greenhouse gas  fluxes  are
 defined  as   the  net exchange  between  terrestrial sinks  and  the
 atmosphere  over  time.

      The development of  a  border-to-border baseline land  cover data
 set  .for  much  of   North  America  will potentially be  a  major
 accomplishment.  The current Landsat Pathfinder series is the first
 attempt at processing high resolution land cover type data  sets  for
 significant  portions of the global  land mass.  Border-to-border
 baseline triplicate data sets are not currently being  developed  for
 Canada.  However, subsequent to  the completion of  the Great Lakes
 Watershed Pilot  study by the CCRS, it  is  hoped  that Canada will
 embark on a NALC/Canada  effort  to complete the border-to-border
 baseline data assembly for North America.
                          REFERENCES


Anonymous,  1993,   U.S.   EPA  Global  Change  Research  Program:
Strategic Plan, April, 1993.

Committee on Earth and Environmental Sciences, 1993, Our Changing
Planet:  The FY 1993 U.S. Global Change Research Program, 79 pp.

Houghton,  J.,  Callander,  B.,  and  Varney,   S.,  Editors,   1992,
Climate  Change  1992:    The  Supplementary  Report  to   the  IPCC
Scientific Assessment, Intergovernmental Panel on Climate Change,
WMO/UNEP, Cambridge University Press, UK, 198 pp.

-------
Sheffner,  E.,  Dull,  D.,  Estes,   J.,  Gaydos,  L.,   Lawless,  J.,
Muchoney, D., Peterson,  D.,  Skole,  D.,  and Townshend,  J.,  1993,
Report of the Technical Peer Review Panel:  North American Landscape
Characterization Project, Unpublished Report, 19 pp.

Skole, D., 1993, Tropical Deforestation Project, Briefing Document,
U.S. EPA Headquarters, January, 1993.

Skole,  D.  and  C.  Justice,    1992,    NASA  Landsat Pathfinder:
Tropical   Deforestation   Project,   Briefing   Document,   NASA
Headquarters, February, 1992.

Skole, D. and C. Tucker, 1993,  Tropical Deforestation and Habitat
Fragmentation in  the  Amazon:   Satellite  Data from  1978  to 1988,
Science 260:1905-1909.
                                10

-------
                   CHARACTERISTICS OF SENSORS AND

                    NALC - PATHFINDER DATA SETS
  Landsat Program History

     The   Landsat  program   of   satellites  gathered   digital
Multispectral  Scanner  (MSS)  remote  sensor  data from  July, '1972
through September, 1992.   The result  is a twenty year time span of
data and the opportunity to evaluate earth resources from space.

     Because  the  archive is  composed  of   digital  data,  these
products present a useful source for image processing and analysis.
These data will support evaluations of change in landscapes or land
cover over time.  No other digital archive of similar time length,
similar resolution, or of  similar potential information content is
available for North America.

     Landsat data  have been  indexed and archived in a number of
facilities  throughout  the world including  North America.   Data
collection  and indexing  has  been  discontinuous for most  areas,
however, and has .required the NALC  Project  to select  particular
years in the past  twenty  for analysis.   These  certain  years hold
many scenes of interest in the archive.

     For the United States the historical Landsat MSS data resides
at  the  National  Satellite   Land  Remote  Sensing  Data  Archive
(NSLRSDA) at the EROS Data Center.   In particular,  the years 1973
and 1986 have very complete coverage of North America.   They have
been selected as focal years  to  include in the production of NALC-
Pathfinder images.  The archive from these years, plus or minus a
one year period,  will provide a unique opportunity to develop a
NALC data set for North America.

     To assure  that high  quality MSS data would be  available for
the  current period  of  time,   a  major  acquisition  program  was
executed.  The program involved purchase of MSS coverage for 1992
based on  the  existing purchase agreement between EDC  and EOSAT.
From February through September  1992  a number of acquisitions were
made.   This action successfully populated  the archive  and will
support production of images for the current time period.
                                11

-------
  Multispectral Scanner  (MSS) Data Characteristics

     Multispectral Scanner (MSS) instruments have flown on Landsats
using the same or  similar spectral  and spatial resolution.   This
consistent  period  of instrument  deployment  provides  a  unique
opportunity to conduct comparisons over time.

     The MSS has four spectral bands or spectral bandpass windows
that are measured by detectors.  The bands include a green band or
band 1 (0.5 - 0.6 urn) , a red band  or band 2  (0.6  - 0.7 urn) , a near
infrared band 3  (0.7 -  0.8 urn)  and  a  second near infrared band 4
(0.8 - 1.1 urn).   Bands 1  through 3 are digitized on a scale of 0 to
127, and band 4 is digitized from 0 to 63.   On later Landsats the
band labelling was changed so that old system bands 4 through 7 are
now new system bands 1 through 4.  The new system is used here.

     The  MSS sensor  measures  spectral radiance over  a  nominal
instantaneous field  of view  (IFOV) of approximately 80 m by 80 m.
The pixel resolution was resampled  to  57 m  by 80 m during ground
processing as that  was  the true geometric  or spatial resolution.
The difference resulted from a systematic over-sampling of radiance
along the scan line.   This action served  to sharpen or convolute
the  image.    The  difference between  the  spectral   and  spatial
resolution has been addressed and corrected in the  standard NALC
processing procedures by pixel resampling.    The different pixel
dimensions are rectified in the course of generating the geocoded
triplicates, and the resultant output pixels are  resampled to 60 m
by 60 m.  This procedure is described later.


  NALC Landsat MSS Triplicate Image Data Sets

     A major goal  of NALC-Pathfinder was  the development  of a
"triplicate" product of  historical  and current data  from the MSS
instruments.  The triplicate would include a scene from 1973 plus
or minus one year,  1986  plus or minus  one year,  and 1991 plus or
minus one year.  The triplicate  is a "stack"  of three image scenes
for each individual path/row scene of North  America.  The path/row
scenes form a world wide coverage that can be  indexed  via the
Landsat World Reference System  (WRS).  For each individual path/row
scene of the WRS for MSS sensor, the areal coverage on the ground
is approximately 185  km  by 185 km.

     The NALC program is dedicated  to  building coverage of North
America in triplicate MSS data sets.  From this triplicate product
it will be possible to evaluate change between three dates over the
last twenty years.
                                12

-------
  Image Composites

     The standard NALC products  are  to be thirty percent or less
cloud  covered.   In  some  areas of North  America it will  not be
possible to obtain a given scene that meets the standard.  Hence,
reduced cloud cover composites  (RCCC) will be made from two scenes.
Composite  scenes  will be  necessary  to provide  appropriate land
coverage by path/row image scene  and  to supply similar information
for  evaluations of  change between  1986  and  1991  scenes in  a
triplicate.  This product  will be  made by EROS Data Center (EDC)
and is described later.
  Categorized Images of Land Cover

     To address variables of interest to Global Change Researchers,
it  is  necessary  to  process  NALC  triplicates  into  derivative
products.  Of particular utility  is  the Land Cover (LC)  thematic
map that displays the  land cover types found in a given scene.  The
land cover classes are keyed to a  land cover categorization system
of standard  design (see section on  image  categorization for the
system).

     Evaluations of single-date NALC  scenes provide inventory data
on land cover or landscape characteristics. Comparisons of two or
more dates of NALC scenes allows  determination of  change in land
cover classes.


  Change Detection Images

     The determination  of change  in land  cover over time  is one
goal of  the  GCRP.   Comparisons of land cover change between the
1970's, 1980's and 1990's image scenes will be undertaken by NALC.
The products are described in the section on change detection.
                                13

-------
                 METHODS AND RESPONSIBILITIES FOR

                    PRODUCTION OF DATA PRODUCTS
  Acquisition of Current Data Products
     Landsat MSS  data products for use in  the  NALC - Pathfinder
project have already been acquired between February  and the end of
September, 1992.  This effort was initiated early to populate the
recent archive with suitable MSS images before the cessation of MSS
data acquisition.   Past activities by other  MSS clients had not
resulted in acquisitions on a  continental  basis  for North America.
Thus, the NALC project acted quickly to avoid  a  similar deficiency
in the 1992 timeframe.

     The acquisition program was developed between USEPA, USGS and
EOSAT.  EOSAT formerly had copyright on the raw MSS data for a two-
year period after acquisition, and they formerly operated much of
the system for data collection and initial data  processing.  Their
role  diminished  with the  transfer of  the  existing  processing
capability and MSS archive to EDC in March,  1993.

     The  coordination  and  implementation  of  data  collection
activities involved an Interagency Agreement (IAG) between USGS and
USEPA  (Appendix V)  and subsequent use  of  the;standard USGS-EOSAT
agreement   covering  Landsat   data  purchases  and  processing
activities.

     The MSS data were scheduled for acquisition based on a number
of land and cloud-cover criteria.  Considerable study was made of
historical periods of low cloud-cover acquisitions throughout the
NALC area of interest. Details on the seasonality of forest cover
and crop cover, the  seasonality of snow cover,  and  the timing of
growing or  rainy seasons  were used to  select periods  for  data
acquisition.  This effort is further described  in Appendix VI.
  Identification of Archival Data Products
     The MSS  triplicate images make  use  of a 1991  +/-  one year
period data set and two archival image scenes or "path/rows" from
the archive.  The archival images were obtained from the National
Satellite Land Remote Sensing Data Archive  (NSLRSDA) at EROS Data
Center in Sioux Falls, South Dakota.
                                14

-------
     Selection of  two historical scenes for  each  triplicate was
made from the 1970's  and  1980's  archive.   The selection has been
made by efforts of EMSL-LV,  cooperators  and EROS Data Center  (EDC)
personnel.   The  criteria for  selection included:   1)  The three
scenes  should  be from  similar seasonal periods; 2)  they should
contain thirty per  cent or less cloud  cover;  3) one  image scene
should be from 1973 plus or minus one year; 4) one scene should be
from 1986 plus or minus one year;  and  5)  the images should be of
high quality.

     The years 1973 and 1986 were selected based on the population
of the MSS data archive.  A complete or  majority coverage of North
America was not commonly achieved.  These years represent the best
examples in terms of data availability.

     The main objective is to obtain the  highest  quality images
from the  three  periods and to have  the images match  closely in
seasonal or phenological time.  Naturally,  the varying conditions
of weather,  image collection priorities, population of the archive
and other factors detract from this objective.  A best effort was
made to meet the criteria and produce high quality products.


  Image Processing of Triplicate Scene Data Products


     A  variety  of activities  are  involved in production  of the
triplicate products from  three MSS data scenes.   Data processing
activities include geometric corrections, radiometric corrections.,
data    formatting,    scene-to-ground    image   registration   or
"georegistration",  image-scene-to-image-scene  registration,  data
indexing and archiving, and data distribution.

     An overall objective  is to develop an optimal data product for
the user.   Efforts have  been focused  to  provide a uniform and
standard product that  resulted from both geometric and radiometric
corrections.   In general,  the corrections were conservative in
design and necessary to reduce sources of noise and to present the
image as  a  map-like product.   The procedures  were intentionally
conservative  to  ensure that  they represented  approaches  with
demonstrated  capabilities  based   on  widespread  current  and
historical use.
  Landsat MSS Data Characteristics


     Landsat data have been archived and delivered in a variety of
formats over the years.   The initial format for purchase was the
"X"  format.    This  gave way  to  the  "A"  format   that  included
additional processing to yield a partially corrected product.  The
most recent and current standard is the "P" format.

                                15

-------
     Landsat  MSS  data  in  X  format  have  been  corrected  for
variations  in detector gains  and  offsets.   They  have  also been
corrected  for line  length variations  by  the addition  of extra
pixels at regular intervals at the ends of lines.

     Additional X  format  characteristics  include a pixel storage
format which was band-interleaved-by-pixel.  The  individual pixels
("n") were  stored as pairs  Cn, n + 1) for the four bands, and the
next two pixels were stored (n + 2, n + 3)  for  each  band.  The data
were also stored in four strips or tracks of approximately 810 to
830 columns.  These strips need to be patched together to develop
a whole image product.

     Landsat MSS data were subsequently stored in A format.  These
data were  available in either band-interleaved-by-line  or band-
sequential  formats.   Band-interleaved-by-line format had all the
pixels in one scan line (n,  . . ., n + 3200) stored one  band at a
time.   Band-sequential  images  stored all  the pixel   rows  and
columns, by individual band, as individual files.

     Landsat MSS data  of  current origin are stored as a P format
product (band-interleaved-by-line or band-sequential formats).  The
data have   either  been rectified  with  ground-control points  or
ground-control chips (precision corrected)' or without  use  of ground
control  reference points  in  a  systematic correction  based  on
satellite orbital  characteristic  data.   The P  format  data  are
provided in either Universal  Transverse Mercator,  Hotine Oblique
Mercator, or the Space Oblique Mercator projections.

     Due  to  the variety  of  formats  over   the  years  and  the
historical  nature of the NALC effort, it was necessary to address
all these differences and organize  processing procedures  for them.

     For NALC, the systematic corrections applied  to the triplicate
images by EDC included:

     Band-to-band geometric offsets;


     Line-length adjustments;


     Detector-to-detector radiometric offsets;


     Correction for earth rotation during image acquisition;


     Non-linear acceleration / deceleration of  scanning mirror; and


     Satellite ephemeris or orbital characteristics.

                               16

-------
       Image Processing for NALC MSS Triplicate Production

  Background

     Restoration of geometric fidelity and accurate registration of
entire images, pairs or triplicates of images, or component parts
of Reduced Cloud Cover Composites  (RCCC) raises important issues.
The use of ground control points and image rectification procedures
are used to transform the image data to  a map projection.  The use
of ground control point library (GCPLIB) image chips can expedite
the process of georeferencing and scene-to-scene registration.

     A number  of  processing procedures  have  been established to
facilitate  the production of  standard  NALC  triplicates.   These
include the systematic  radiometric and  geometric corrections and
precision image-to-map registration.  The procedures presented here
have been implemented by the EROS Data  Center  for  production of
standard  triplicates.    These  standard  procedures  have , been
developed from more than twenty-years of experience in the digital
processing of Landsat data.  A  number of these procedures are well
known and have been documented in the literature.

     The 1970's data selected for use in triplicate production are
pre-processed  to  correct for  line  length  adjustments,  variable
detector  response,   band  registration,  nonlinear  mirror-scan
velocity, earth-rotational skew, and detector-to-detector offsets.
Both .the line length adjustments and detector-to-detector offsets
require a resampling in the along-scan direction  (one-dimensional
cubic  convolution).    The  parameters  for  these    systematic
radiometric  and  geometric  corrections  are  derived  from  the
satellite ephemeris and payload correction  data.   The individual
bands are also "destriped" to remove noise due  to variable detector
response.  The resultant  product  is analogous to the "P" product
currently generated by  the EROS Digital Image  Processing System
(EDIPS).    In accordance  with  the Landsat  Pathfinder  Program
concept, the Pathfinder "basic  data sets" will be composed of data
which have been radiometrically and geometrically processed to the
"P" level.  NALC triplicates, however, will be  processed beyond the
"P" level as described below.
  Data Acquisition

     The  generation of  NALC  triplicates  is  initiated once  the
appropriate scenes have been identified by EPA-EMSL and purchased
from EDC.  EPA staff are responsible  for determining all scenes to
be  purchased,  both  current  and  historical  acquisitions,   in
accordance  with  the scene  selection  criteria described  in  a
previous  section.   This process  involves  searching databases  and
reviewing browse  or  microfiche imagery  to assess  cloud cover  and
image quality.  EPA submits the scene-identification numbers to EDC

                               17

-------
 for the subsequent  submission  of  CCT  orders.  The original scenes
 are then archived into  the  Landsat  Pathfinder Archive  (LPA) under
 a  project  code  which  uniquely  identifies the  scenes  as  NALC
 Pathfinder  data.

      In parallel with the ordering of scenes, the geographic corner
 points  for  the selected scenes are used with EDC database  software
 to  identify  the  topographic  map  sheets  corresponding  to  the
 respective  scene  path/rows.  A subset of these maps is retrieved
 for use in  ground  control  point  selection  for  image-to-map
 registration.     EDC  has   developed   a  relational  database  to
 facilitate  archiving  and  retrieving  all  U.S.  1:24,000  scale
 quadrangle maps as well  as the  1:250,000 and 1:50,000 maps  received
 from Mexico's National Institute  for Statistics,  Geography,  and
 Information (INEGI) .  This  database will be augmented as maps are
 acquired for  Central America and  the  Caribbean.
  Triplicate Processing

     The process  of generating the triplicates involves a multi-
stream  approach  (Figures  1  and 2).   Ground  control  points are
selected from  the 1986  images and maps for use in developing the
model for  geometric transformation (geocoding).   The map control
points  contain X  and Y  positional values and elevation values to
correct for relief displacement.  A cubic convolution registration
is used to rectify and resample the image to a UTM projected output
image.  The image is comprised of 60 m x 60  m pixels, with an Root
Mean Square Error (RMSE) of less than  1.0 pixel.

     Concurrent   with  this   activity,  the  1970's  scenes  are
radiometrically and geometrically pre-processed to convert the CCT-
X  data  to  a  P-level product.   The  images are  "destriped"  to
compensate  for variations in  the  radiometric  response of  the
individual detectors prior to geometric registration because the
noise is scan-line dependent.   An interim systematic correction is
applied, using the satellite ephemeris data and platform navigation
model,  to generate  a UTM projected output image that is oriented
north-up.

     Automated cross-correlation procedures are then implemented to
extract control points from the 1970's and 1980's images to compute
coefficients for  image-to-image registration.  This involves the
use of a single band from each of the 1970's input  images.  Once an
accurate transformation  is developed,  the grid  for  the interim
systematic  correction   is convolved  with  the  image-to-image
transformation  grid  to  produce  coefficients that  facilitate  a
single  step  registration  of  the  1970's  P-product with  the map
registered 1980's image.   The  net result is an image registration
procedure that  only involves  one step of resampling.   The target
RMSE for the  image-to-image registration is  0.5  pixels  or less.
Previous   studies  have   shown  that  the   use   of  polynomial

                                18

-------
transformations alone on  CCT-X format data yields only a 1.5-2.0
pixel internal image accuracy,  once  map projected.   The use of a
satellite model overcomes this problem and assists in meeting the
registration criteria.

     The last step involves creation of a pixel identity image to
accompany each of the 1970's images.  Each pixel value identifies
the specific 1970's scene  used to fill-out the WRS-2 path/row scene
or "tile".  This step  is  necessary due to the east-west shift in
scene position from WRS-1 and WRS-2.

     The procedures for the 1990's image registration are similar
to those  for the  1970's  data, except  that the  1990's  data are
acquired  as P-level products.  An interim systematic correction is
applied, based on  the  satellite orbital data platform navigation
model,  to  generate a  north-up UTM  projected image.   Automated
cross-correlation procedures are used to select control points, and
a  transformation   grid  is  computed   for  the  image-to-image
registration.  The  two grids are convolved into a  single grid which
is then applied to the 1990's data to create an equivalent of the
1980's image.   Similar to  the 1970's processing effort, this action
involves only  one  resampling step.  The  RMSE  objective for this
registration is 0.5 pixels or less.

     Reduced Cloud Cover Composites (RCCC)  are  made after all data
have been coregistered.  This step is only performed in cases where
19907s scenes with 30% or less cloud cover are not available.  To
minimize  the  amount  of  cloud cover  in  the 1990's  triplicate
component,  EDC  has adapted AVHRR  cloud-reduction  compositing
procedures for use in NALC triplicate generation.  The compositing
process operates on  image pairs.   The procedure  is based on the
Normalized Difference Vegetation Index (NDVI):

            (band 4 - band 2)/(band 4 + band 2)

     This   index   is   sensitive   to   variations   in   surface
characteristics, such as the presence or  absence of green biomass,
and other  scene characteristics  such  as clouds.   The  NDVI  is
computed for each of the  images to be used for compositing.   The
maximum NDVI value determines  which input image  pixel  brightness
values  (BV) will  be used to  constitute the output image.   This
maximum NDVI decision rule  is  efficient for computing  and yields
consistent  results  as  judged by comparisons of  image  and ground
surface features.   A 1990's triplicate  component which has been
composited will have two additional bands, a pixel identity image
and the resulting RCCC or maximum NDVI composite image.

     The  final  task  in  triplicate  production  involves  making
mosaics and a projection  transformation  of the Digital Elevation
Model (DEM) data.  The  DEM data being used is derived from the DMA
DTED data  which  were digitized from  the  standard NTMS 1:250,000
scale topographic maps, for which  there  is complete  coverage for

                                19

-------
the United States and Mexico.  These data, often referred to as the
3  arc-second  DTED   data,  are  made  available  as  grid  files
corresponding  to  blocks in dimensions of 1° of latitude by 1° of
longitude.  These blocks are formed into  a mosaic for the path/row
of interest,  and then re-projected and resampled to 60  m x 60 m
pixels in UTM projection.
  Image Rectification Analysis Forms
     As explained previously, the image registration process will
use  ground control  points  or  the  ground control  point library
(GCPLIB) or "chips". The georegistration process will quantify the
errors associated with the modelled transformation.  An independent
set of ground control points will be used  to validate the accuracy
of georegistration.

     The following  forms document the results of the registration
of a NALC image.  The forms  will provide part of the documentation
of quality and heritage of  the  image.  These forms will be stored
at  the origin  of  production  and  may be  accessed for quality
assurance and quality control (QA/QC) requirements.

     A  sample  of the Image Rectification Analysis  Forms will  be
evaluated as part of the QA/QC  efforts.  It  is anticipated that
every  tenth  rectification  will be  evaluated by QA/QC  personnel.
Images  or  groups of  images that fail  the  criteria will be re-
processed.   It is  anticipated  that the  acceptance  criteria for
rectification will  be + or -  1.0 pixel  in comparison  to ground
position and + or - 0.5 pixel in image - to - image registration.
                                20

-------
CD
    o
 OS "D
0.  O

OQ.
    co
a)



I
PL,
   'co
U- (O
 c +-

"co 5
 CO
 CD
 O
 O
                                  21

-------
«
•a
*-» *-
o o
•a «j
1 ii
—I "- ra
/' 'jJ U j-


CCT-X |
Reprocessing j
GO
(0
T3
13 o
=> 9
•% £
{ '• LLCP 3480
:v.v.^
• ; Browse Pn
f.V:.:JData Base


0.
o
<
(h
D.
Q
ai
O

Reprocessing
O
(A
TJ
1 8
•So?
{ } LLCP 3480
'.'.'.I'.v:
• I Browse Pr
r';;JData Base


EDIPS-A to -P
O

Reprocessing
D



                                          o
                                          o


                                         "5"
    D)
            OT
                           OT
-
o il
HI   HI
Q:3i=
    0)
    DC
                22

-------
 o
 0,
                                IMAGE RECTIFICATION  ANALYSIS  FORMS
  o
 •MO
 O I)
 O.CS
     The  following  forms record  the  characteristics of  the  image
 registration and  rectification activities.   These documents will
 be held at  the point  of  origin.
  CO
  o>
CO
via
o
  §
  o
CM
JC
u
o
o
irt
in
f-
m at,
»-4J a
>o •-«.
       a  -a  u
       C  t)  L.
       a.t- o  v
         V U  4J
     « c « x  u
          I V  E
                         E E XC  t.
                         •»-  .00  o
                           O  t-  O.

                         u  «o  c
                                         u
                                         u
                                         c
M  O.
O L. U
U 41 U
O w O
i. 3 e
a  i
  xti
«i e
E a •'
act)
          U a
          -> C
                 w—<—> t.

                 tfo «*-
                 U O 3 tl
                 O E C.Q
                 £.  a
                 at> e t>
                  j=  -a
                 O*->T3 O
  „              c  «o

  cu  t)*eEwt-  IJCQ.CI
  •c-«r- t._J*r* O      t)«f t) a
  O4-> «J.Q E t-Xl-  "D t- O^J
  O.  O «  **-J3 O   O 3 OH-
   •a c t-i   c_   B-O a
  c t)  a ow L.      i (i
  •r-4-i o a—> ct)t>ut>c.v>
  W O O E O «>i- I Of U U-r-
                         • c u    a •
                           V V_J  _j
                         *-t-D.ejl.C
                       w O*J • CO t.
                       C  « t- M L.
                       •*-»  BBSO
                       o a c 3 o  u
                       O.T3T- act)
                                -
                      •O t.  U  C'T-'B 3
o>
o
u ••
veo
a
       00
       00
       oo
       oo

C OU.  OO
Oi-     • .
•»-4JT3  00

a u c.
 C  M4J
                                    X  C «J
                    O U V « a C_i O.OLC     -
                    o a  au o.an B4-> x*j a.c
                    _«-x_a  •  o  —i    c
•—--•	"•  -•	^~ K ••— tr^'l ' ^% «iJT—
_^ X—> O XOT3 34->.a BJQ «l 01

"-   04J«JC_.^  X-r-XoH
  X-r- OT-4J a X4J—«4-._. V O
t)-. W-i-a QIC « 0-> B-.-rn
> -j a t)T- coecaeaOM
•»• «j E t. « e a    3 o a t. «
*-• 3 O t-J3 C. a t) C C *• C  4J
o c<-> o 3*»f-.c.o e 3 a xa
< a 3 U «• M-D4JT- 6 a E—< U
  1 «J I I I I   will B-r-
     i I i I   a i i i
                                                               23

-------
                         IMAGE  RECTIFICATION ANALYSIS  FORMS  (continued)
o
0,
   I
   J
  o
   o.

 ur-

 ea

 o

 • M
    O
   •r*
   Ui
   U ••
   oca
   3 B
                                            4-r    f
                                            C    f                    O*—tSIMi^* irvOf-^co O.O*wc\lKK~*trt*O f~cOO>p
                                            ^~    f vpj fi^^ifi-O'^. co o-^~ *"~*^*~^^^~^~*~«~ ^t\j ivj csx<\nvj-* r a,a.0.0.0.0.0.a.da;0.0,0.0.0.0.0.0.a0.0.0.0.0.0.0.0.0.0.0.



                                            o    i
                                            o

                                            —i-a i                        KI                           •oco                           -OOOT)'nTIt3T3TIT31Ot3-DT3-aT3OKlWXjTI
                                              t» I 33333333333033 333333333 3 3««««««•««««« M                        CI-OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
a,                  »> c                        o i oooooooooooooooooooooooooooooo
C                    o                       t-l'--»  ...........................
^-                    OX>-                   4J I OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                                               C I
o                                             > <
4J                                              tl I
                                               «^ I
AJ                                             U] *

Q.                    OCVJ«-                      t OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
4j   ,_                OOO                      I OOOOOOOOOOOOOOOOOOOOOOOOOO OOO O
3                    +  i  i                       I OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
O                    «««                      |..»*.».........................
     „                O^T-^-                      I OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
a,   c                ^O^                   X f

g   ^*              ^xCO«4'^O                    O I
     *J              COKIKI                    C r
•k   o              ^^O^**'^*                    O (

 C   i-                O.PJO                      I
 U   X                MO-CO              "       >
•v-   O                CO*OOt             —'       '

.r_   o                O*O*Ot   CM          X       I OOOOOOOOOOOOOOOOOOOOOOOOOO OOOO
«.   a.                ...             t-       i oooooooooooooooooooooooooooooo
,»_   a               -o<-o>              0-      i  - - * ' -	
 o          >.   >_       1      ..                >- I OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
 O   t_     *~-  "•             4JOOOS     C     "S. I
O   O     (V   CM             •>-»-«-    ••-     ** I
     »-     ^^  ^x    oOKt   t-                a i
  r          <   m    OOO              c     _i t
  t   u               + + i   e •• -     «>       i
     O     -I-   +    VOOT-X>*     >      I
 c   t.               o*t*.co             "**       i
 O   Or     XX    ^-»-   M'r"T~     „       i ooooppppppppppopoooocDooooooooo
                               3EB     t-     «JI

 C.   V             OV     M   UEI..	.-
 ^   ^               »OOr*-   C €> to     t.  <3 a I       v-«-«—CMCMCMKIKkK>****-*«rt      ^-1  «  (M CM CJ K> K> K> *
 «   ,-   	^    ^-0*<-   -r-C-t-     O   CLUI
 c    a   «»o  o    MOM   O*-H-     c-  co  i
  O    C.   4^ ^.^  ^.^    ^^ O *£*   O-         C-      '
  t.    «   c<  m      ...     s-»-     c  w
 Ha,o          O.T-V   »•— o o         3
          ^-«i     i       o          —•  a.

 „    5   ^_ ^^                 L.tof>     3  3CI****'**'^''^""*^***'**'**''^'^"**'"*^*'^'COC^O3COCOOO0OOOOOCOOOCOOOCO
 CC   C   H-X  >-             «€)*»     TOO-^I'

 >   x   uu.  u.             ecu a     H
 z   a   o                   3 u u     v
 •••  =c   o                   zaa     <£
                                                                             24

-------
                          IMAGE RECTIFICATION ANALYSIS FORMS  (continued)
o
o.
  o
  o
O V
O.O
V a
  eo
  o>
 CO
 vin
 o

 a M
 4J _»
 «-J
 a a
  u.



 ui o

  »
 CJ
 o^
 Ox
  o ••
  II CO
  o

   • V
  3 •
                                                                                               .^^.CO 00 CO CO CO 00

                          Q.Q.Q.o.Q.o.o.o.o.cxa.Q.o.a.o.Q.a.a.Q.cxo.Q.cxa.Q.a.o.a.a.o.Q.Q.a.cxc





                     Om  CMKl^O                     *o*o    mo.                    o  4n*--*
                     CMm  OOCOO                     ^.**-    ^in                    r-  m*o*o
                     K>CM  K»»— «O                     KlfO    CM «-                    «-  CMK»«-

                     oo  ooo                     o«-    oo                    o  ooo
««««««««*«     «        ««««««««««     ««     ««««««««««   «       ««««««««
««*«««««««     «        ««««««««««     « «•    «<««««««<«   «       <««<««««
•****««*««*     «        ««««««««««     ««     ««««««««««   «       ««•««*««

"OT3T) T3"OT3"O TJTJT3    "D       "OT3"D^*DT3T3^"O"n    "D"O    ~O~UT3T3T3"a^"OT3"a  T3      "D"X3T3*O "OT3"OT3




ooooooooooooooooooooooooooooooooooooooooooooooc3"o^j^jT)^jTj7)7j
CCCCCCCCCC     C     ICCCCCCCCCCIICC     CCCCCCCCCCIC     ICCCCCC'CC


4J 4J 4J 4-» 4J 4J 4J 4J 4J 4J    4J       4_* 4J 4J 4^ 4J ) • | • 4^ | • ^ .    4J4J    4-I4J4J4J4J4J 4J 4J 4J 4J  4_|      xj 4.1 4.1 4^ 4^ 4_i jj 4^
OOOOOOOOOO     O       OOOOOOOOOO    OO   .OOOOOOOOOO   O       OOOOOOOO

                     CM**  ^TKIN       •             h-IA    O>K»        .            tw  inKlin
**********CMf\J* Ki«~ l/\ 4t4I4I4I4I«I«4I4I«K1O*4:«—O*******«4I4IO*CMCM«— «««**«««

««*«««**««OO«OOO*'«««4i«««««O«^««OO«««««««4t«4iO«OOO««««<«««
                      III                        III                     lit




O OOOOOOOOOO*-OOtin«—OOO OOOOOOOO «— OOO«OcoOOOOOOOOOO^O*4-COm OOOOOOOO
	—_	   	._  Jr^l^oaaoooao

                                                                                            30

                                                                              -*-«-CMCMCMIOK>»O-*-*-»-in      «-^-




OOOOOOOOOOSDCJOCOCOT-OOOOOOOOOO-OI—OOOOT-OO OOOOOOOO OCOCOmOOOO OOO O
OOOOOOOOOOKt(MOK)O>OOOOOOOOOOOO>N-OO4-OOOOOOOOOOOinOr>.OOOOOOOO

                                                                                                        t -f -» -4- ~*
                                                                                          J r- »- T- CM CM CM 00 CO CO CO CO
                                                                                          •kKifOfOFO




                                                                                          DOOOO
                                                                                          DOOOO

                                                                                          3OOOO






                                                                                          DOOOC
                                                                                          3000C
                                                                                          DOOOC

                                                                                          noooc









                                                                                          3000C
                                                                                          3OOOC
                                                                                          ooooc

                                                                                          ooooc









                                                                                          ooooc
                                                                                          ooooc

                                                                                          o n ** ** ** in      T- «- V-CM CM CM fO K> Kl ** -* ** in       «-»-«—CM CM CM KIM IO-* >
                                                                                                                   Co cocococo
                                                                          25

-------
                                          IMAGE  RECTIFICATION  ANALYSIS  FORMS  (continued)
t-
_j
o
 ,o
  o
 o «
 D.O
co eo 00 eoo-o-o«cs o-oo-o-o-<>.»

 O. O. Q. Q. O. Q. O. O. Q. O. O. O. Q. a O. D. CL O. O. Q. Q. Q. O. D. Q. [J. 0.
           CU**«*'r-KI«-«-«r-*J-O'O>O




                  KI  «-»»O

                  o  ooo
           T3 T3 "O •* "D O M »- 13 TJ T3 "D "O "O T3 T3 TJ TJ 13 "D T3 T3 T3 TI T3 TJ T3
            uuoouooouuuuuuuuuuuouuuuuuu
            CCC   Cl 1   CCCCCCCCCCCCCCCCCCC
            ooo   o
                             ooooooooooooooooooo
                 : O« OOO<
                   I       I
   CO
   o>
           oooeoo«-r~«-ooooooooooooooooooo
           OOOOOK1K10OOOOOOOOOOOOOOOOOOO
 CO
 out
 o
            OOOI^-OK>-cocococD(X>vr^-vr-j--j--T-i--T-i- M KI M Kl M M-»-*-a--*-*-»•-#-»•-»•-
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo

            ooooooooooooooooooooooooooo
                                                                      D.D.
                                                                      Old)
                                                                      o o
                                                                      a. a.
                                                                                 ,
                                                                                 co^-
                                                                                 «-K»
                                                                                 o-o
                                                                                 I0~»
                                                                                 •OKI
                                                                                  c- t.
                                                                                  o o
                                                                                  L. i.
                                                                                  1. 1.
                                                                                 U1U1
                                                              r-UMM   V V
                                                              irtQ^^  Kcc
                                                              Oif\~f
                                                              COf*-*^  CM O

 ooooooooooooooooooooooooooo   in*—o  co-a-
                                                               • • •  KICJ
                                                              OOO  KIKI
                                                                      •OCO
                                                                                 o
                                                                                 o
            OOOOOOOOOOOOOOOOOOOOOOOOOOO
            ooooooooooooooooooooooooooo
            OOOOOOOOOOOOOOOOOOOOOOOOOOO
                                                                                         ooo
                                                                                         ooo
                                                                                         ooo
                                                                                         ooo
                                                                                         ooo   oo
                                                                                         ooo   oo
                                                                                          • •  •   oo
                                                                                         000   00
                                                                                                  00
                                                                                                  00

                                                                                           _.     OO
                                                                                          " •>
          3      "O *•      -• a
                  f V     -J 3
          •a       vtc.      cxj
          Cl       O        3 *r*
          4J       CC V     T3 M
          a         —i  ••  -r- ii
          v«4-o   o cxui   Mce
          U.CJ-J   C EM   t)
                                                                                                     _i  a.
                                                                                                              C E
            ooooooooooooooooooooooooooo
            ooooooooooooooooooooooooooo
   KI
  u ••
  ueo
  o

   . u
  3 a
  JC—
                   CO CJ in 00  KI KI KI KI K> KI KI-* ^"-J-•*-»-*•* ^^-*•*-»••*•*•*-»•-*
                                                              BT3
                                                              3-»-
                                                              T3 «•
                                                              •»- o
                                                              UK
                                                              

v a.
C E
•*- a
          •a .<•— •   a
          •^ «> a.  — •
           U C E   O
          o*»- «   a.
             — ' M   L.
           a      v
           C *-» 4-*   4->
          t- 3 a   c
           Q.Q.Q.  -c
                                                                                    oo'
                                                                                                              o
                                                                                                              a.
        tl
        4J
        c
                                                                                             *-• t_ t_
V
a
                                                                                             EJQ.TI   i.
                                                                                             o E E   v
                                                                                             «j 3 3   >
                                                                                             19ZZ   •<
                                                                            26

-------
                            IMAGE  RECTIFICATION ANALYSIS  FORMS  (continued)
         V O
         E-1
o
a
 (W CM «*• *» O> m CO -i»- m in «- »- O KICMCMK»KllO»OCM«MtOCMCMtOCMK>CMMK>CVlCMCMCMCMCMM   CWI-O
        Via t     •                                                                    H «l. «J "O
        t.   i                                                                           awo
        v!                                                                         t~-   u  •

        MVI	•	"Oco
         a   i r-ooocoo-orj<^a-ooootoo»-Kiooo.»--*eoK>o.c>.oocymooo>Kieoo

         1—• I r«I»OeOeOinC>t>CO»-lftCM-*T-l>.>OCMOCO«>>O«*-O«O-i a. i
    c
    •*-

II  U
—*  N
T-  •»-   E a
4-  M


c  —•
o  o
i-  X
4-1  «^
a  a
u
f-  
•r-   «3
 L.   E
  3   a
  a.  "
  c   3
  —   o
     IOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                                                                                       CM ••*• V
                                                                                        I -O«- X
                                                                                            • T-
                                                                                        •  • •* a.
                                                                                           •o
                                                                                        no   c
>-i e I •OOm«OCOCMin«OO-*-*OOM'*i>-'rtacs01 aicsaaoacnadoucao»CDOocacaoiaicso
                                                                            27

-------
                 IMAGE CATEGORIZATION AND DEVELOPMENT

                     OF LAND COVER DATA PRODUCTS
  Training Set Development

     The development of  training sets for categorization requires
a method  that will yield high  quality  and  unbiased training set
characteristics in an automated  fashion.  The unsupervised training
set  development  approach has been proven successful in providing
good, homogeneous  spectral  clusters in applications such as land
cover or  landscape characterization.   The  unsupervised approach
negates many systematic  problems  associated with atmospheric and
terrain  radiometric distortions.   This  approach has  been used
successfully for over twenty years, and it can also be implemented
with a minimum of  effort.  Consistent results can be expected for
most applications.

     Unsupervised  training set  development will be conducted with
an automated spectral  clustering  algorithm.   A common clustering
method has  been selected from available choices in the literature
of  remote  sensing  applications and has  been developed  intb  an
algorithm by EMSL-LV personnel.

     Training set  development will   be  conducted  with  some
assumptions and  input  parameters.    Individual  scenes  of  the
triplicate  will  be clustered  separately.   Each scene will  be
clustered on an iterative basis  for as many as one-hundred or more
clusters or classes.  Criteria for  linking or breaking of clusters
will be established empirically.

     Several approaches  to  unsupervised training set development
will be  tested.    These  include  clustering using an  EMSL-LV
procedure   and  using  procedures   employed  by  the  commercially
available image processing software currently in place at the EMSL.
These procedures will  be tested as part  of  the  standard methods
development activities that  form the Chesapeake Bay Watershed and
State of Chiapas, Mexico Pilot  Studies.
  Evaluation of Training Sets

     The results of unsupervised training set development will be
evaluated in several manners.  A preliminary categorization image
will be produced from the unsupervised training set development or
clustering activity.  This  image will help indicate the location of
many clusters or classes.

                                28

-------
     The preliminary  categorized image will  be used  along  with
other procedures  to  evaluate the quality of  individual  training
sets or clusters using aerial photo or other data.  Procedures will
include  divergence  calculations,  bi-variate   scatter  plots  of
cluster  means  and variance-covariance,  and  comparison  of  the
preliminary categorized  image  with ground information or aerial
photographs.

     Following  analyses  of  the quality  of  training sets,  the
contents may be edited to remove examples of poor clusters,  e.g.,
clusters that  represent  very few  pixels  in the  original image,
overlapping clusters, and clusters that are composed of more than
one distinct,  spectrally homogeneous land cover class.


  Categorization

     The edited training sets will be used  for categorization of
all  three  original  scenes  from  the  triplicate.   The  maximum
likelihood / Bayes decision rule approach to categorization will be
evaluated.   There are a  number of reasons to  use this approach,
among which  is the  fact  that it  has been successfully  used in
remote sensing applications for over twenty years.  This approach
makes use of both Euclidean distance measurement and probability-
based criteria  in the determination  of the class identity  of a
given pixel.   The use of  these criteria makes the  categorizing
algorithm more  sensitive to  class  characteristics as compared to
single-criterion approaches.

     After evaluation of training set quality, the scene from the
triplicate will be categorized using all four bands as input to the
maximum  likelihood classification algorithm.   The work  will be
completed at EMSL-LV and will employ the scene triplicates produced
by EDC.

     The resulting product of one  hundred  or  more  classes  will
represent a Preliminary Land Cover (PLC)  product.   It remains to
identify the  land cover  type of  each cluster or class.   This
activity is known as  "labelling" and will be perform by using local
knowledge,  ground information, aerial photographs and maps.   The
PLC  product will  require  that  certain  clusters  or  classes be
aggregated  or  "lumped"  to  render  the  best  product.     This
aggregation will  allow inventory of  several  clusters or  classes
that  may  represent   the  same  functional  land  cover  class  as
represented in the NALC land cover categorization system.
                                29

-------
  Identification of Land Cover and Classification System

     The land cover classes in the PLC image will be  labelled as to
cover  type  using a land  cover  classification system.   The NALC
classification  system was  developed to specifically support NALC
project objectives and to be  compatible with the other major land
cover  classification systems.  The  NALC classification system is
compatible with the Anderson et  al.  (1976),  Cowardin et al. (1979)
and Brown et al. (1979)  systems.  The system has been optimized to
support carbon inventory applications.

     The standard NALC characterization data base products will be
based  on Level 2  classes   (Figure  3) .    For  a   more  complete
description of the NALC system and definitions of land cover types,
refer  to Appendix X.

     The assignment of class types to PLC classes will be performed
by NALC cooperators.  The scheme is presented below, with Level 2
as  the final  land cover  type  detail  to  be  identified.   This
approach will supply sufficient detail while using the capabilities
of  MSS  data to  their  best measure.    It  is recognized  that
differentiation between forest and shrub/scrub classes at Level 2
will be difficult and in some  cases not possible without the aid of
ancillary data.   However,  a  best effort  will  be made throughout
North America to accomplish this differentiation.

     The resulting  product will be  the Land  Cover  thematic data
coverage products (LC)  for each path  / row.  It will  be supplied to
EMSL-LV for validation.  Subsequent to data validation, LC products
will be archived at EMSL-LV.  After validation, the products will
be archived and distributed by EDC.   The image will be accompanied
by details on the training sets including cluster or  class means in
each  band,   variance-covariance matrix,  class  identities,  and
details related to quality and accuracy.
  Data Production and Distribution

     The data are to be processed by three groups,  USGS-EDC, EPA
EMSL-LV, and the cooperators.  The triplicates produced by EDC will
be distributed  initially to  EMSL-LV.   EMSL-LV will  conduct image
pre-processing,  training set selection,  training set evaluation,
and maximum likelihood categorization.   EMSL-LV will distribute
NALC triplicate  scenes,  the categorized or clustered image scenes
(Preliminary Land Cover, PLC),  as well as  documents on standard
processing methods to the cooperators.  Cooperators will identify
the clusters or classes of the PLC maps, aggregate and assign land
cover types  from the categorization system,  and create the Land
Cover (LC) products.  The cooperators will supply the LC products
to EMSL-LV  for  archive.  The LC  products will be tested against
standards and for quality  (data validation), and then sent to EDC
for archive and  distribution  to users.

                                30

-------
  Quality and Accuracy Assessment

     The  evaluation  of data  quality has  been addressed  in two
manners: the systematic sources of  errors  in production;  and the
accuracy of derived products.  Once the systematic or within-system
evaluations have  been performed  and the characteristics  of the
triplicates and derived products determined to meet the standards,
the data characteristics will be recorded and stored at the point
of origin  (EDC).

     The  accuracy of  derived products,  such  as  the  Land Coyer
products,  will be  tested.    The results  will be  stored  as  an
attribute file and maintained at  the point  of origin.  Accuracy of
each land cover class, the error matrix,  and other' details will be
included.

     The accuracy assessment  information of the LC data sets must
pass validation.  EMSL-LV personnel  will  take  the data and develop
the accuracy assessment calculations and results. The criteria are
described in the section on QA/QC.

     The LC data sets will be re-evaluated if they fail the above
criteria.   The evaluation will focus on determining  whether the
initial  classes  were incorrectly  aggregated, were  incorrectly
labelled,  or  were subject  to other  sources  of errors.   If the
source of error can be identified and a correction can be made, the
image will  enter  the stream for  further  evaluation.   If it fails
again,  a  new  attempt will be made at categorization,  and results
will be evaluated  for  accuracy until a solution of the problem is
identified  and the criteria are met.
  NALC Products

     Products  provided by  the NALC  project will  include post-
processed  remote  sensor  data, post-processed  digital elevation
model data  (DEM),  intermediate categorization products, and final
database  categorization  products.    All NALC  products  will be
"clipped" to the World Reference System 2 (WRS 2) path/row scene
configuration.

     NALC  standard  data products  will  include  the  following:
triplicate  data sets  georeferenced  and coregistered Landsat MSS
imagery  (1991,  1986,  1973  plus  or minus  one  year);  spectral
clustered  categorized images  for  each  Landsat  MSS  image;  and
coregistered DEM  data.   Categorization database  products  will
include  .complete  (100%)  image  categorizations  for  Landsat MSS
scenes  from 1991  plus  or  minus  one  year,  and  partial image
categorization products for  Landsat  MSS scenes  from 1986  and  1973
plus or minus  one  year.   For the Landsat MSS  1986  and 1973 scenes,
only the areas that have undergone  significant  levels of  spectral
change will  be available in  categorized form.

                                31

-------
Anderson, J., Hardy, E., Roach, J., and Witmer, R., 1976,  A Land
Use Classification  System  for Use With Remote-Sensor Data,  U.S.
Department  of the Interior,  U.S.  Geological Survey Professional
Paper 964, Washington, DC, 28 pp.

Brown,  D.,   Lowe,   C.,   and  Pase,   C.,   1979,     A  Digitized
Classification System for the Biotic Communities of North America,
With Community (Series) and Association Examples, for the Southwest,
Journal of the Arizona-Nevada Academy of Science 14:1-16.

Cowardin,  L.,  Carter,   V.,   Golet,  F.,   and LaRoe,  E.,  1979,
Classification of Wetlands and Deepwater  Habitats of the United
States,   U.S.  Department of Interior,  U.S.  Fish  and  Wildlife
Service, Rep. No. FWS/OBS-79/31, Washington, DC, 103 pp.
                                32

-------
            Figure  3.
NALC Pathfinder Categorization System
LEVEL 0
Land
LEVEL 1
1 .0 Barren or Developed
Land
LEVEL 2
1.1 Exposed Land
1.2 Developed Land

2.0 Woody
2.1 Forest
2.2 Scrub/Shrub

3.0 Herbaceous
3.1 Herbaceous

4.0 Arid
4.1 Arid Vegetation
4.2 Riparian

5.0 Snow/Ice
5.1 Snow/Ice

Water
6.0 Water & Submerged
Land
6.1 Ocean
6.2 Coastal
6.3 Near-Shore
6.4 Inland

Other
7.0 Other
7.1 Cloud
7.2 Shadow
7.3 Missing
7.4 Indeterminable
                   33

-------
               METHODS FOR MANAGING AND DISTRIBUTING

                       NAIiC DATA AND PRODUCTS
     The NALC-Pathfinder products will be identified, indexed and
archived.  The NALC  product characteristics  will  be stored in an
Information Management System  (IMS)  or  "meta"  data set housed at
EROS  Data Center  (EDC).    This  will  allow  identification  and
tracking of NALC products even when archive products are stored at
different physical locations.

     These new products require new identification numbers to be
assigned.  This is due to the variety of Pathfinder products that
will  be   stored  in  the  archive  and  the   variety  of  data
characteristics.   The Information Management  System  (IMS)  will
require additional fields of meta  data beyond the fields specified
in existing systems  (Figure 4).

     The population  of  image scenes in the  archive will include
derivative products created outside of the EDC  environment.  These
products  need  to  be produced   with  NALC  standard  processing
procedures.  They must be of known and  documented quality as per
NALC  standard  quality characteristics.   To be included  in  the
archive, products must be uniform and adhere to NALC standards.


  NALC Data Archive, Management, and Distribution

     EDC  has  responsibility  for the  archive,  management,  and
distribution of  NALC Pathfinder  products  (original  Landsat  MSS
CCTs, geocoded triplicates,  OEMs,  and derivative  products).  EDC
also has this responsibility for other Landsat Pathfinder projects.
Consequently, a  long-term strategy for  data  archive, management,
and distribution must address project-specific needs in the context
of the overall Landsat Pathfinder Program goals.

    Towards  this end, an interim  Information Management System
(IMS)  approach is  being developed  to  facilitate  the indexing,
archiving, and distribution of NALC data  sets.  This approach is
designed to develop a database that is populated by metadata which
describe NALC  data  sets.   It  is  also important that this IMS be
upwardly compatible  with  the Version 0  IMS  at  the Land Processes
Distributed Active Archive Center  (LPDAAC)  for EOSDIS.
                                34

-------
     The  Version 0 IMS  at the LPDAAC will  ultimately be used  to
manage  all  Landsat  Pathfinder  Project  data  (source  data and
derivative products).   This  IMS will be separate  from the current
Global Land Information System (6LIS),  which currently facilitates
query,  browse,  and product  ordering  from the National  Satellite
Land Remote Sensing Data Archive  (NLSRSDA) .  GLIS is linked  to the
NSLRSDA  production database  and  is not  currently structured  to
handle higher level data products, such as those being generated  by
NALC.

     The  types  of metadata being compiled and tracked along with
the  triplicate  data sets are listed in Table 2 and  described  in
Table 3.  These  particular attributes are  considered essential for
describing the  NALC  triplicate  components.   The higher level data
products  which will be produced later  (cluster  data sets, land
cover products, etc.) can be  expected to have different or expanded
metadata  requirements.   Currently,  this metadata  is managed using
a relational  database management system  (RDBMS),  specifically  D-
Base IV.  However, in response to the Peer Review Panel Meeting  in
New  Orleans,  EDC  has  investigated  the IMS  developed by the
University  of  New  Hampshire  for  the  Humid  Tropical Forests
Inventory Project  (HTFIP),  which  is  another Landsat Pathfinder
activity.  The HTFIP IMS  is based on ARC/INFO capabilities, and
provides  graphic  displays  of  data coverage,  browse  images   of
selected  scenes,  and   the   ability   to   query     some  metadata
attributes.

     Upon evaluation of the HTFIP IMS,  EDC  has concluded this
capability  can be developed  for  NALC  in  a more  efficient  manner
using ARC VIEW  capabilities.  The D-Base  IV information  which has
been compiled  would be upwardly  compatible,  so no effort to date
has   been  wasted.     The  following  elements   are functional
capabilities  which an IMS should  support:  Graphical display  of the
extent  of  data  coverage;  geographic  query;  interrogation   of
metadata  attributes;  data richness or  status tracking; viewing  of
selected  browse products; linkages between NALC  triplicates and
their source  in GLIS; and the ability to execute product  orders.

     The  interim  IMS  is not likely to  provide  immediately open
.access  to the  general  public,  but it  should serve  the internal
needs of  EDC  for  product development and  tracking and be portable
to EMSL-LV.   EDC  will provide EMSL with regular,  periodic updates
to  the  metadata.   In addition to  providing   support  to NALC
production and data management activities, the interim IMS will aid
in defining the functional requirements of the Version 0  IMS. EDC
staff are currently investigating IMS requirements for the Landsat
Pathfinder  projects  as part of  a larger effort  to assess the
resources required to  upgrade  GLIS as a phased approach  to the
LPDAAC  Version  0  IMS development.
                                35

-------
  NMiC Data Distribution

     EDC is currently distributing the NALC triplicate products  to
EMSL  on 8mm Exabyte cartridges.   These tapes  contain the data
description record  (DDR) files in ASCII format,  and the image data
in band-sequential  format  (Table 4) .

     The DDRs (Table 5)  contain information describing the image
projection  parameters and UTM  coordinates.   The  order in which
files  are written to tape (Table 6)  is always the same,  except the
number of  files  depends  on the number of scenes used  for each  of
the 1970's  and 1990's components.

     The.image files, in turn, vary in size (i.e., number of bands)
depending  on  the  requirements  for  cloud  composite  and mosaic
products.   A DDR file is written before each image file,  and the
geographic  coordinates are always based on the 1986 scene which was
map-registered.    All  files  (images  and  DEM)   have the  same
dimensions  (NL, NS) ; the MSS images are 8-bit (BYTE) data, whereas
the DEM is  16-bit  (INTEGER*2).

  Query Procedures

     The information on NALC-Pathfinder products in the Information
Management   System   will   be   available   for   a   number    of
characteristics.  The details on a given image may be  queried from
fields set up to  index NALC-triplicate and derivation image product
characteristics.

     Users  would enter  the  IMS via  communication link  from  a
workstation.  The query sequence  can be called up and executed, and
the user could examine the fields by individual  scene.  The fields
could  also be used to  search  the database  for a given  set   of
criteria.  Searches by location,  date, cloud cover, and so on could
be searched to identify  scenes of interest.

     These  and other data fields  (Tables 2 and 3)  are still to be
evaluated.  It may be desirable to add or subtract field types from
the list above to optimize the information content while minimizing
redundant,  unnecessary/  or costly information fields.


  Product Media Available to the User

     The  NALC-Pathfinder products  will be  available  from  EDC.
Products will  be  sold at the cost of duplication.  The media for
distribution will include a  number  of types.   This is due to the
variety of  hardware owned by users and due to the  fact that media
types  change over time.   NALC products such as scene triplicates
will be available on nine track tape, on 8mm tape,  and  on 3480 type
magnetic tape  cartridges.
                                36

-------
     The media type mix may  change  over the life of the program.
This could be caused by a change in industry standards for media or
media  hardware.   It  could also  be caused by  a change  in user
selection of storage devices  or by experience with media that does
not meet the requirements of the project.

     EDC archives the NALC image data on 3480 cartridge media and
only  uses  the  8mm cartridges for  distribution to  EMSL.    EDC
recommends that users verify all 8mm cartridges upon delivery and
that the data be backed-up on  9-track  tapes or  other  media.  The
Exabytes may not be reliable long-term storage media.   On several
occasions EMSL has not been able to read the 8mm cartridges, even
though EDC has been able to successfully run tape mappers on them
prior to shipping.   In addition, there  have  been a number of calls
from other users who have been unable to read these tapes.
  Standard Product Ordering Procedures

     Data orders from the archive will be accomplished by standard
ordering procedures at EROS Data Center. Orders are currently made
by  correspondence,  or  through  electronic  ordering  procedures
available through the EDC's IMS or GLIS query systems.


  Alternative Products and Ordering Procedures

     Cooperators will receive data from EMSL-LV, and return project
related products to EMSL-LV.  Cooperators with EPA may obtain data
from EMSL-LV.
                                37

-------
 S
 O)  Q.
 O  JJ
 9U  O
CL  C
*-*  o
DL  O
~  O)
 03

 S
 03
                                             C/5
                                             O.
                                             CO
                                                 .2
  O


  E
  o.

05-g
"X a
§Q
"O O)
o
J=
05 0
•§ «
O a)
a 8
fcoc
x: o
0?
C3
JC
O
05
CD
e §
o g
as
1?
— CD o
811
UJ < QL
• • *•
                                                38

-------
                          Table 2.

              Interim NALC IMS Metadata Database Schema
Structure for database: C:\PATHFIND\NALC\NALC,
Number of data records:       4
Date of last update   : 04/22/93
Field  Field Name  Type       Width    Dec
                           DBF
    1  TAPEID
    2  LEVEL
    3  PATH
    4  ROW
    5  DECADE
    6  ENTERDT
    7  PROJCODE
    8  FORMAT
    9  RESAMP
   10  SCENE1
   11  CTLPTS1
   12  RMSERR1
   13  ACQDATE1
   14  SUNELEV1
   15  SUNAZIMUT1
   16  SCENE2
   17  CTLPTS2
   18  RMSERR2
   19  ACQDATE2
   20  SUNELEV2
   21  SUNAZIMUT2
   22  SCENES
   23  CTLPTS3
   24  RMSERR3
   25  ACQDATE3
   26  SUNELEV3
   27  SUNAZIMUT3
   28  SCENE4
   29  CTLPTS4
   30  RMSERR4
   31  ACQDATE4
   32  SUNELEV4
   33  SUNAZIMUT4
   34  SOURCE
   35  CONTACT
   36  PHONE
Character      6
Character      1
Numeric        3
Numeric        3
Numeric        2
Date           8
Character      1
Character      4
Character      1
Character     16
Numeric        4
Float          4
Date           8
Numeric        2
Numeric        3
Character     16
Numeric        4
Float          4
Date           8
Numeric        2
Numeric        3
Character     16
Numeric        4
Float          4
Date           8
Numeric        2
Numeric        3
Character     16
Numeric        4
Float          4
Date           8
Numeric        2
Numeric        3
Character     16
Character     3 0
Character     14
Index
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Example
31051
G
16
31
70
04/22/93
U
CCTX
C
8104515240500
97
0.65
09/06/72
47
139

0
0.00
00/00/00
0
0

0
0.00
00/00/00
0
0

0
0.00
00/00/00
0
0
121
TBD
TBD
** Total **
             238
                                39

-------
                                 Table 3.

          Definitions for NALC Interim IMS Database Schema

Field Definitions:

TAPE ID = EDC archive storage location, the tape id serves as a
unique reference
          identity combined with the path, row and decade fields,
          Valid formats are:
            A#### - 1 space, 1 alpha character and 4 numbers.
           AA#### - 2 like alpha characters and 4 numbers.
           ###### - 6 numbers.

LEVEL   = level of processing performed.
          Valid codes are:
            P - processed, G - geocoded/georegistered, M-mosaic,
            C - composite, E - DEM.

PATH    = a World Wide Reference System defined nominal Landsat
satellite track
          (path). All processed NALC data is referenced to WRS2.
          Valid values are:
            1 - 233
ROW     = a World Wide Reference System defined nominal
latitudinal center line
          of a Landsat image. The row indicator represents scene
centers that
          are chosen at 25-second (LS 1-3) and 23.92-second
(LS4/5) increments
          along the orbital track in either direction of the
equator.
          Valid values are:
              1 -  59 = Northern Hemisphere (Descending)
                60    = Equator (Descending)
             61 - 119 = Southern Hemisphere (Descending)
            120 - 122 = Southern Polar Zone (Descending)
            123 - 183•= Southern Hemisphere (Ascending)
               184    = Equator (Ascending)
            185 - 246 = Northern Hemisphere (Ascending)
            247 - 248 = Northern Polar Zone (Ascending)

DECADE  = acquisition decade of set.
          Valid codes are:
           70 - 1970's.
           80 - 1980's.
           90 - 1990's.

ENTERDT = the date the metadata record is added to the data base.
                                40

-------
PROJCODE= map projection, default is 'U'.
          Valid codes are:
            U - UTM or Universal Transverse Mercator.
            S - SOM or Space Oblique Mercator.
            H - Hotine.

FORMAT  = data's current input format.
          Valid codes are:
           CCTX - EDC computer compatible tape in X format.
           EDIP - EDC Enhanced Digital Image Processing System's
format.
           DEM  - EDC digital elevation model format.
           FAST - EOSAT's fast format.

RESAMP  =  resampling technique used to radiometrically process
the data.
           Valid codes are:
             C = cubic convolution.
             N = nearest neighbor.
                 processing.
             B = bilinear.

SCENE1  =  a systematically corrected scene id or the first scene
id used in
           generating a mosaic or composite.

SCENE2-4=  2 to 4 scene ids used to generate a mosaic or
composite.

CTLPTS1 =  total number of geographic control points used for
scene 1.

CTLPTS2-4= total number of geographic control points used for
each scene 2-4.

RMSERR1 =  the average root-mean-square error made using each
scene's control
           points (i.e. RMSERR1 applies directly to SCENE1 and
CTRLPTS1 just as
           ...RMSERR2/3/4 applies directly to their respective
SCENE/CTRLPT
           numbers).

RMSERR2-4=.the average root-mean-square error made using each
like numbered
           scene's control points.

ACQDATE1=  the acquisition date applicable to the first scene id.


ACQDATE2-4= the acquisition date applicable to the like numbered

                                41

-------
scene id.
SUNELEV1 = the elevation angle of the sun above the horizon.
           (Valid values = 0 to 90 degrees)

SUNELEV2-4=the sun elevation applicable to the like numbered
scene id.

SUNAZIMUT1 = the azimuth angle of the sun measured clockwise from
north in
             degrees. (Valid values = 0 to 360 degrees)

SUNAZIMUT2-4=the sun azimuth applicable to the like numbered
scene id.

SOURCE  =  base source used in registration.
           Valid sources are:
             24kUSGS = l:24k USGS topo map.
             SOkMEXC = l:50k Mexico topo map.
             121 = image to image registration.

CONTACT  =  a contact person available offsite whom fully knows
the projects.
            Valid contact is:
              TBD

PHONE     = contact person's office phone number.
            Valid phone number is:
              TBD
                               42

-------
                               Table 4.

                   8mm Tape Mapper for NALC Triplicate
********** Mapper of tape cot mounted on DRIVE 4 **********
       1 RECORDS   3428 BYTES LONG
END OF FILE #1 >»» 1 TOTAL RECORDS.

   25000 RECORDS   5000 BYTES LONG
END OF FILE #2 >»» 25000 TOTAL RECORDS.

       1 RECORDS   3428 BYTES LONG
END OF FILE #3 >»» 1 TOTAL RECORDS.

   25000 RECORDS   5000 BYTES LONG
END OF FILE #4 >»» 25000 TOTAL RECORDS.

       1 RECORDS   3032 BYTES LONG
END OF FILE #5 >»» 1 TOTAL RECORDS.

   20000 RECORDS   5000 BYTES LONG
END OF FILE #6 >»» 20000 TOTAL RECORDS.

       1 RECORDS   3824 BYTES LONG
END OF FILE #7 >»» 1 TOTAL RECORDS.

   30000 RECORDS   5000 BYTES LONG
END OF FILE #8 >»» 30000 TOTAL RECORDS.

       1 RECORDS   1849 BYTES LONG
END OF FILE #9 >»» 1 TOTAL RECORDS.

    5000 RECORDS  10000 BYTES LONG
END OF FILE #10 >»» 5000 TOTAL RECORDS.

**********  END OF TAPE  **********
                                43

-------
                                 Table 5.

                 Example of Data Descriptor Record (DDR)
IMAGE NAME:85068515574x0.trimp
                         NS:5000
                       DATE:10-Feb-93
          NLtSDOO
DTYPE:BYTE
  LAST MODIFIED:
SYSTEM:ieee-std
  PROJ. CODE:(1)UTM
Valid:VALID
   ZONE CODE:15
Valid:VALID
  DATUM CODE:0
Valid:VALID
  PROJ. PARM:
Valid:VALID
  A:   -1.61239334063410E+00
9.99600000000000E-01
  B:   O.OOOOOOOOOOOQOOE+00
O.OOOOOOOOOOOOOOE+00
  C:   5.00000000000000E+05
O.OOOOOOOOOOOOOOE+00
  D:   O.OOOOOOOOOOOOOOE+00
O.OOOOOOOOOOOOOOE+00
  E:   O.OOODOOOOOOOOOOE+00
O.OOOOOOOOOOOOOOE+00
 CORNER COOR:
Valid:VALID
    ULcorner:1.70538000000000E+06
    URcorner:1.70538000000000E+06
               •40544000000000E+06
               •40544000000000E+06
  PROJ. DIST:6.00000000000000E+01
Valid:VALID
 PROJ. UNITS:METERS
Valid:VALID
   INCREMENT:1.OOOOOOOOOOOOOOE+00
Valid:VALID
 MASTER COOR:1       1
         NB:4
       TIME:1643; 01
  LLcorner:1.
  LRcorner:1.
                             2.52200076913181E-01
                            -9.30000000000000E+07
                            O.OOOOOOOOOOOOOOE+00
                            0.OOOOOOOOOOOOOOE+00
                            O.OOOOOOOOOOOOOOE+00
4.44960000000000E+05
7.44900000000000E+05
4.44960000000000E+05
7.44900000000000E+05
6.OOOOOOOOOOOOOOE+01
                                  1.OOOOOOOOOOOOOOE+00
                              44

-------
IMAGE NAME:85068515574x0.trimp

   BAND NO:1

        MINIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID
        MAXIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID

    DATA SOURCE:Landsat 5
    SENSOR TYPE:MSS
CAPT. DIRECTION:descending

           DATE:15-JAN-86
           TIME:1557:4

IMAGE. NAME: 85068515574x0.trimp

   BAND NO:2

        MINIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID
        MAXIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID

    DATA SOURCE:Landsat 5
    SENSOR TYPE:MSS
CAPT. DIRECTIONDescending

           DATE:15-JAN-86
           TIME:1557:4

IMAGE NAME:85068515574x0.trimp

   BAND NO:3

        MINIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID
        MAXIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID

    DATA SOURCE:Landsat  5
    SENSOR TYPE:MSS
CAPT. DIRECTION:descending

           DATE:15-JAN-86
           TIME:1557:4
                                45

-------
IMAGE NAME:85068515574x0.trimp

   BAND NO:4

        MINIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID
        MAXIMUM:0.OOOOOOOOOOOOOOE+00
Valid:INVALID

    DATA SOURCE:Landsat 5
    SENSOR TYPE:MSS
CAPT. DIRECTION:descending

           DATE:15-JAN-86
           TIME:1557:4
                                46

-------
                                   Table 6.

                       Explanation of NALC Tape Mapper
File #1: Data Descriptor Record (DDR)  file which is based on the
     georeferencing parameters for the 1980's data image to map
     registration.  The projection information is the same for the
     1990s and 1970s data.  This is an ASCII file.


File #2: image data for the first 1970s scene used in the
     triplicate.  This contatins four bands of MSS and the
     pixel identity image.  The pixel value used to identify
     this scene can be cross referenced to the Dbase records
     which will follow separately.


File #3.: DDR for the next 1970s scene  (if more than 2 are used)


File #4: the next 1970s scene; four MSS bands and a pixel
     identity image *(as above)


File #5: DDR for the 1980s scene


File #6: 1980s  image data; four MSS bands; no pixel
     identity image required


File #7: DDR for  the 1990s data


File #8: the 1990s  image  data;  four MSS bands, NDVI  image
     resulting  from cloud reduced, compositing, and  the  pixel
     identity image.  The last  two bands  only exist  when
     compositing  is done.  Band 5  is  the  maximum NDVI  image
     resulting  from the multiple inputs.   The pixel  with the
     maximum NDVI value  is used to determine which  input pixel
     DN values  are used to create the  output.  Bnad 6 is the pixel
     identity  image.  These pixel values  are indexed to the scene
     from which the output pixel values were taken.   The pixel
     values correspond  to the scene number in the dBASE record
      (dBASE information isdistributed only to EPA-EMSL).


 File #9: DDR for  the  DEM generated for that Path/Row.   The
     projection parameters  should match those of the MSS data.

 File #10:  DEM data for  the  triplicate; same dimensions as the
     MSS image data,  except that it is INTEGER*2

                                 47

-------
           IDENTIFICATION OF CHANGE DETECTION METHODS
  Introduction

     The goal of NALC change detection methods development research
is to select or develop standard change detection methods for both
pre- and post-categorization delineations of change in land covers.
For the pre-categorization change detection  technique, only those
areas  that have  undergone  significant  spectral change will be
categorized.  The pre-categorization technique will  be utilized as
the  preferred  NALC   change   detection  method.     The   post-
categorization change detection technique  will only be used when
pre-categorization  is  not technically  feasible.  This method will
be a raster-ba^ed Geographic  Information System (GIS) operation and
would require complete (100%) categorized digital database products
for multiple dates.  Because of the  emphasis  on carbon issues, the
NALC-Pathfinder  will place a  special  emphasis on detecting land
cover changes involving forests, which hold more carbon per above-
ground area than other major land cover types.   NALC is interested
in both the losses  and gains in  forest area plus the detection of
forests  which  have been  placed into  a  regrowth  stage due  to
disturbance.
  Change Detection Literature Review

     Early efforts at using Landsat MSS data for the detection of
land cover change were  largely based on visual interpretation of
the    multitemporal    images.       This   is    essentially   a
photointerpretation  process,  using  techniques developed  in the
decades preceding  the launch of Landsat  in  1972  (Shepard 1964).
This method  is  still widely  used for  change detection.  It  is
perhaps the  most direct and  fastest  change  detection technique,
particularly in  the initial stages of a change detection project.

     Following   the  use  of  visual  interpretation  for  change
detection, digital methods began to be employed. Change detection
methods have  been  broadly divided into  either enhancement  (pre-
categorization) or post-categorization methods  (Nelson 1983, Pilon
et al. 1988,  Singh 1989).

     Enhancement   change   detection   techniques   involve   the
transformation  of  two  original images  to  a new single-band  or
multi-band tineategorized  image in which  the areas  of  land cover
change  are readily  detected.  The change-enhanced  data  can  be
further processed by other analytic methods,  such  as by using a
categorizer,  to produce a labeled change detection output product.
Most of the  enhancement techniques are based  on  the  concepts  of
image differencing or image ratioing  (Weismiller et al. 1977, Toll

                                48

-------
et  al.  1980).   Differencing  of vegetation  indices has  been a
popular form  of enhancement change detection  (e.g., Nelson 1983).
It  has been  shown  that ^ image  equalization in  the  data  pre-
processing stage usually improves the results of change detection
(Hall et al.  1991).  Techniques  like band-to-band regressions and
principal  components analysis  have  been used  to simultaneously
perform the image-to-image equalization and the detection of change
areas (Ingram et al. 1981 and other references).

     In  post-categorization  change  detection  two   images  from
different dates are  independently categorized. The area of change
is   then   extracted  through   the   direct   comparison  of   the
categorization  results  (e.g.,  Colwell  and  Weber  1981).   The
advantage  of  post  categorization  change detection  is  that  it
bypasses the  difficulties in change detection associated with the
analysis  of   images  acquired  at  different  times  of year or  by
different sensors.   The disadvantages of the post-categorization
approach include greater computational and labeling requirements,
high sensitivity to the  individual categorization accuracies, plus
the  difficulties  in performing adequate accuracy assessment  on
historic data  sets.

     The  Landsat  Pathfinder  Humid  Tropical  Forest  Inventory
Project, being performed at the University of New Hampshire and the
University of Maryland, uses a post-categorization change detection
method. Because of  the small number of classes being used  (primary
forest, non-forest,  secondary  forest,  water,  cloud,  and shadow),
the  project  is able to achieve  high categorization  and change
detection accuracies.

     Because  all  enhancement  methods  are  based on pixel-wise
operations or scene-wise plus  pixel-wise operations, accuracy in
image registration and co-registration is more critical for these
methods than  for other methods.

  OPTIMIZING  FOR DIGITAL CHANGE DETECTION:

   NALC SCENE SELECTION AND PREPROCESSING

     The basic requirement  for remote sensing land  cover change
detection is  the availability  of two  dates  of imagery upon which
the  same   area of   land  can  be  observed.    Depending  on  the
characteristics of  the  two  data sets,  change detection  can  be
either an easy or  difficult  task.  The  NALC  data  sets have  been
selected and assembled in order to simplify the detection of change
using digital image processing.  Below is a description  of some of
the complicating factors which tend to make change detection more
difficult, with  indications  of  the  design features  of the  NALC
project which address the difficulties.
                                49

-------
     The  accurate detection of land  cover  change using remotely
sensed data can be complicated by the following factors:

1)   Spatial  resolution and spectral bandpass differences between
images acquired with two sensors complicates  the direct comparison
or  the  digital analysis of the data  to  detect  change.   With the
spectral  bandpass differences,  the  detectable land cover classes
for  the  two  dates of imagery may not be comparable.   Land cover
classes  that  are distinct when observed with one  sensor  may be
indistinguishable with a sensor having broader or  fewer numbers of
spectral  bands.    If  there are  substantial spatial  resolution
differences between  the two  input images,  ground features may be
visible in one data set and undetectable in the other.  In the NALC
project  we eliminate  problems  that  could  be introduced  due to
spatial resolution and bandpass differences by working exclusively
with Landsat  MSS, a  sensor  series  that has provided consistent
spatial and spectral characteristics for a twenty-year period.

2)  Variations  in  the  radiometric  response  of  a  sensor  can
complicate change detection by requiring some  form of  image-to-
image equalization prior to  the change  detection or by requiring
the use of a post-classification change detection approach.  This
factor is important to the NALC project because we are working with
data spanning a twenty-year period and collected by  five different
MSS sensors (Landsats 1-5) .  Although each of the MSS sensors were
calibrated before launch, it is known that their response drifted
over time and  that there  is no adequate way  to calibrate the data
to radiance units without extensive ground  based efforts.   It is
clear that the change detection procedures used by the NALC program
must  explicitly   address  the  radiometric  differences  to  be
encountered in performing change detection without calibrated input
data.

3) If clouds are  present in  images  from one or  both dates, it is
impossible to  detect land cover change between  the two dates of
imagery where clouds occur.  To minimize the effects  of clouds, low
cloud cover was one of the primary factors directing the selection
of scenes for  the NALC project.

4) Variations  in  solar irradiance,  solar zenith angle,  and solar
azimuth will  affect  scene brightness levels  and  the  location of
shadows.  The NALC project has attempted to reduce these effects by
selecting scenes  for the three  time  periods which were acquired at
or near the same Julian date in order to  match the  solar conditions
as far as possible.   In general,  high  sun  angles (low  amount of
shadowing) are better than low sun angles for the detection of land
cover change.

5) Variations  in  atmospheric effects (scattering and absorption)
can affect scene  characteristics  sufficiently  that they must be
considered in  evaluating change detection methods.  The NALC data
sets have been visually screened to  remove  scenes  where within-

                                50

-------
scene atmospheric  variations  are obvious.  By  assuming that the
atmospheric effects on the selected scenes are uniform across the
entire  scene  area, it  is possible  to partially  compensate for
scene-to-scene variations in atmospheric effects by scene-to-scene
brightness  equalization.   The  coregistration  of  the NALC MSS
products will also help  to compensate for topographic variations
which alter atmospheric path length.

6) Phenological variations in vegetation result in large changes in
the reflectance patterns  of the land surface.  If images of leaf-on
and leaf-off conditions are compared, whole regions can appear to
have been  "deforested".   The  scene  selection  process of the NALC
project has attempted to  minimize this problem by selecting scenes
from the same time of year for each  of the three  time periods.  In
some cases it was not  possible  to  do this  due to the  lack of
appropriate archive data.

7)  Spatial misregistration  of  images  will  tend  to  reduce the
accuracy of any digital change detection effort.   These  effects are
most severe on the change detection techniques using enhancement.
The NALC data sets are digitally coregistered, with accuracies on
the order  of half  a pixel or less, creating data sets that can be
analyzed for change .with minimal errors due to misregistration.


  NALC Change Detection Strategy

     The basic NALC data  set consists of three  dates of  Landsat MSS
data, from the  early 1970's,  mid-1980's, and early 1990's,  plus
ancillary  data such as digital terrain.   In performing the change
detection  for  the NALC project  two  output  products  will  be
generated: 1)  Land cover change from the early 1990's to the early
1970's; and 2) Land cover change from the  early  1990's  to the mid-
19 80' s.    The  output  products  should show the location  of the
changes as well  as the nature of the change  (e.g., conversion of
woody to herbaceous).
  NALC Change Detection Procedure

     The NALC Landsat MSS triplicates have been selected from the
available archive  of  data held by the USGS  EROS  Data Center.   A
considerable effort  has  been put into selecting  scenes with low
cloud cover and at or near the same date for all  three  time periods
in order to have data sets optimized  for digital change detection.
However,  in the  extreme  southern portions  of  the project area
(southern Mexico,  Central America, and Caribbean Islands)  it was
frequently not possible to find low cloud  cover  scenes at the same
time of year.  Cloud reduction compositing is being used to assist
in  the  development of triplicates which  are matched  to the same
time of year.  However, there are numerous triplicates which will
not be closely matched.

                                51

-------
An example is given below  for Path 21, Row 47 in southern Mexico:
     21/47
     21/47
     22/47
     23/47
92/04/29
84/11/25
74/02/15
78/09/17
LE84357515342xO
LM85026915590xO
LM8157215552500
LM83019615545xO
     In  this circumstance,  it  may be  necessary to use  a post-
classification change detection approach to derive the land cover
change products.

     For the majority of the NALC triplicates,  it has been possible
to  find  data in  the  archive having low cloud  cover  and closely
matched dates for use in the triplicates.  An  example of this type
of triplicate is  given below for Path 45, Row 29 in Oregon:
     45/29
     45/29
     49/29
     48/29
92/08/03
86/08/19
72/09/02
72/07/27
LM85307718130xO
LM85090118114xO
LM8104118265500
LM8100418210500
     In  this  case,  and in  the  case of the majority  of  the NALC
triplicates, it should be possible to use an enhancement technique
to identify areas where change has occurred. Then the  change areas
would  be  labeled using  the same   labeling  procedures  used  in
developing the current  land cover product.

     The first choice of the NALC change detection will be to use
an enhancement method to locate  pixels where land cover change has
occurred.  However, because of  the presence  of triplicates which
contain data from diverse times of the year,  it will be necessary
to have a post-categorization method developed.  We are therefore
developing  a  dual-track research   plan  to   develop  a  suitable
enhancement method  and protocols  for post-categorization  change
detection.
  NALC Change Detection Evaluation Criteria

     NALC has  two groups of criteria:  programmatic  criteria and
technical criteria, for evaluating change detection techniques.

     For  the change detection  procedure using  enhancement,  the
criteria are as follows:

     Programmatic  criteria: 1)  The  method must  be  relatively
sensitive to deforestation and/or loss of vegetation - the primary
target of the NALC project;  2) the method must be  simple in concept
and  have been  successfully tested  widely, so  that  it can  be
accepted  by a wide  spectrum of  application scientists; 3)  the
computation should be straightforward and unambiguous to perform in
                                52

-------
various  computational environments;  and 4)  this method  should
require a minimal amount of expert (human) intervention so that it
can be applied to a large number of scenes.

     Technical criteria:  1)  The method must be  accurate;  2)  the
method  must  be   relatively  insensitive  to sensor  effects  and
atmospheric effect; and 3) the method must work in a wide range of
geographic regions.


     Since the post-categorization method will only be applied to
a limited number of scenes and areas, the speed and simplicity of
the method will no  longer be emphasized.   Instead,  the following
will addressed: 1)  This method must be able to provide detailed and
accurate class to class change information;  2)  this method should
require  minimal  human  intervention  so  to  optimize  automated
processing; and  3)  this  method should provide  highest  possible
categorization accuracy.


  NALC CHANGE DETECTION RESEARCH:

   TECHNICAL WORK PLAN

     This work plan describes the study sites and change detection
methods to be  used in  developing the standard procedures to be used
in the rest of the NALC project.


  Study Sites

     Three diverse areas have been selected for conducting change
detection research for the NALC program.
Washington,  B.C.  - Path.  15,  Row 33:   This  is an  area  that has
undergone substantial urban and suburban expansion during the past
two decades.  Many forest areas surrounding Washington, B.C. have
been converted  to  residential,  commercial,  and industrial sites.
The  exact  magnitude of  these land  cover  changes  has yet  to be
determined.  This Path/Row image scene lies within the Chesapeake
Bay Watershed, which has been  an area of active USEPA Environmental
Monitoring and Assessment Program (EMAP) research. Buring the past
two years a  current  land cover map was produced by EMSL-LV using
recent  Landsat Thematic  Mapper  (TM)  data as  part of  the EMAP
Landscape  Characterization program.   The  NALC project  change
detection research pilot  in  the Washington,  B.C. triplicate will
make use of the EMAP land cover data set as a source of validation
data.   As an  additional validation source,  we will  obtain the
coastal land cover change data for this area produced from Landsat
data by the NOAA Coastwatch-Change Analysis Program  (C-CAP), which
was produced using a pre-categorization change detection method.

                                53

-------
Central  Oregon -  Path. 45, Row 29:   This  scene  covers a forested
transect across the Western flank and crest of the Cascade Range,
plus  arid  areas,  with  irrigated  agriculture  to the  east  of the
Cascades.   The forests  of  the Cascades include some old growth
forests, plus numerous secondary forests which have been disturbed
through  forest harvesting.  All conditions of forestation exist,
ranging from barren to closed canopy forests.


Southern Mexico -  Path.  21,  Row 48:   This  scene  centers  on  a
mountainous tropical forest region, on the Mexico-Guatemala border.
On the Mexican side there has been extensive conversion of forest
for use in agriculture and timber production.  There has been much
less activity  on,the Guatemalan side.  At the northern end of the
scene there are extensive wetland forests which have been cut and
drained.  The area is covered  by detailed 1:250,000 scale Land Use
- Vegetation maps produced  in the  early  1980's by  the  Mexican
government.


  Methods to be Tested and Evaluation Criteria

     In each of the three study areas, photo interpretation will be
used to outline and categorize areas of land cover change.  We will
use  the digitized photointerpretation results  as  a basis  for
comparing the  results  from  the digital change detection methods.
Because of the dual track approach required for change  detection in
the  NALC  project,   the use  of  both  enhancement  and  post-
categorization methods will be investigated.

     The areas of  change delineated by each  of  the investigated
approaches will  be digitally  compared  to the land  cover change
results obtained through photointerpretation of the same scene area
to produce percent accuracy values.  The change detection methods
will be evaluated based primarily on criteria stated above.


  Post-Categorization Change Detection

     Protocols for the  post-categorization change  detection will
involve  the  repeat of the NALC  standard  categorization results,
which are being developed to produce current land cover data sets.
These procedures will be  repeated for the other two time periods.
Then the areas of  change  will be extracted by comparing the early
1970'8 and mid-1980's land cover categorizations with the current
land cover.
                                54

-------
  Enhancement Techniques For Change Detection

     Because most of the NALC triplicates have been optimized for
use in enhancement  change  detection  methods,  an emphasis will be
placed  on experimentation with  these approaches.   Because  of
atmospheric and sensor related radiometric response differences, we
will focus on procedures which have some capacity  to normalize the
radiometry of the data prior to locating change pixels.  Our review
of the literature suggests that three enhancement techniques have
built in capability for normalizing or equalizing the radiometry of
scenes:  Vegetation  index differencing; regression  analysis;  and
principal components analysis.


  Change Detection Through Vegetation Index Differencing

     The   development   of  vegetation  indices   from   spectral
reflectance  values  is based  on the differential absorption and
reflectance  of  energy by vegetation  in the red and near-infrared
portion  of the  electromagnetic  spectrum (Derring  and Haas 1980).
In general,  green vegetation absorbs energy in the red region and
is highly reflective in  the  near-infrared region  (Anderson and
Hanson 1992).  A number of  vegetation indices have been formulated
and utilized for monitoring vegetation change.   Of  these vegetation
indices, NDVI has been used most widely for monitoring terrestrial
vegetation dynamics (Townshend and  Justice 1986, Eidenshink and
Haas  1992,  Tappan  et  al.  1992) .  The  NDVI compensates  for some
radiometric  differences between   images;  however,  it   does  not
completely remove radiometric images that are being compared.  The
difference  in  the  NDVI values  of  two  images in  certain cases
responds to changes in land cover  (Nelson 1982  and  1983, Banner and
Lynham 1981).   Singh (1989)  concluded that NDVI differencing was
among the most  accurate of change detection techniques.

     In our change detection pilot data sets,  the  NDVI values will
be computed  for each date of imagery.  The early 1970's and the
mid-1980's NDVI images will  be subtracted from  the  1990's NDVI
image to create two NDVI difference images.  These  will be compared
to  the  photointerpretation  results  in order  to  calculate  the
accuracy of  the NDVI differencing.approach.


  Change Detection Through Principal Components Analysis

     Principal Components Analysis (PCA) involves the reorientation
of axes of an input data set,  creating  output Principal Components
 (PC) data  sets.  In our case two coregistered MSS images will be
input  as an eight band  or  eight  axis image.   Because  of the
autocorrelation of  the original data,  there is an elongated cloud
of distribution of data located in the axis of each data  set.  The
PCA will orient  the  first axis  of  the output data  through the
central  core of the input data cloud.  The  second  axis will be

                                55

-------
perpendicular to  the  first  and will be directed through the next
major direction of variance in the data set.  The PGA can continue
to establish new  axes, until the number of output axes equals the
number of input axes.  The  first PC normally contains the overall
scene brightness  variations which are in common between all the
input bands.  The second,  third,  and fourth  (and sometimes higher)
PC images frequently contain information on pixels which changed in
reflectance between the two dates of imagery (Byrne et  al. 1980,
Richardson  and  Milne  1983, Fung  and LeDrew 1987).  The last PC
image would be expected to contain random noise that existed in one
image relative to the other.

     In scenes with cloud cover and associated shadowing, it will
be necessary  to  screen out  the cloud and shadow areas  to exclude
them from the PGA. Clouds or shadows which are present in one date
and absent  from  the second date will tend to redirect  the  axes
which would otherwise be established due to reflectance changes due
to land cover change.

     Part  of the PGA algorithm involves  the normalization  or
equalization  of  the input data,  thereby reducing  atmospheric and
sensor radiometric response differences from the output PC images.
This is  one of the characteristics  required  for the NALC change
detection,  making PGA one  of  the  leading  candidates  for  the
standard change detection method for use in the NALC project.

     Color composite images will be made from combinations of the
second, third,  fourth, and  higher PC's.   Areas  of  change will be
identified from  these  color composites and  the  accuracy of these
PGA based change detection areas will be compared with the results
from the photointerpretation.
  Change Detection Through Regression Analysis

     The results of Hall et al.  (1991) indicated that band-to-band
radiometric   rectification  or   equalization,   as   a   distinct
preprocessing step, can substantially improve the accuracy of the
image differencing approach  to  change detection.   The methods of
Hall et al.  relied on 'the manual extraction of pixels which were
consistently  bright  and  dark  pixels from  the image,  to  obtain
digital  number  values  for  use  in  regression analysis.    The
regression line between the bright and dark  pixels was taken to be
the line of "no-change".  Pixels which departed from the regression
line were  taken to be  pixels  in which  some type of  land cover
change occurred.

     We plan to experiment  with a modification of  the approach
described  by Hall et  al. to make  the  procedure  automated,  and
remove the reliance on bright and dark ground targets which occur
in each date of imagery.  For each pair of MSS scenes, four band-
to-band  (e.g.,  band  1  versus  band  1)  scatter diagrams will be

                                56 .

-------
prepared using  a  random subsampling of the image  pixels.   Early
experiments  with  this  approach  indicate  that  in each  scatter
diagram a diagonal axis of  data points  will appear.   This is the
axis of "no-change".   Clouds and cloud shadow present in one image
and absent in the  second  image  will tend  to  form their own axes,
parallel to  the diagram axes  (Figure 5).   Pixels which change in
reflectance between the two  dates due to land cover  change will not
fall on the diagonal scatter diagram axis  (the "no-change" axis).
By determining  the location of the  "no-change"  axis  with linear
regression,  it  is  possible to  calculate  the  orthogonal distance
from the "no-change"  axis  for each pixel in the image, creating an
output  image depicting  the  reflectance  changes  which occurred
between the  two  dates  for a particular band.   By  repeating this
procedure with all four bands,  four output change  images  will be
created from the eight input bands.   As  with the PGA approach, the
cloud and  cloud shadow areas would need to be  excluded from the
regression analysis.

     As with the PGA approach, color composite images will be made
from regression analysis  output images.   With four output change
images there are only two three-band color composite combinations
possible:  1,  2,  and 3 or 1,  2,  and 4.   Areas  of  change  will be
identified from these  color composites  and the  accuracy of these
PGA based change detection areas will be compared with the results
from the photo interpretation.  It may also be possible to utilize
a statistical test such as Chi-square.


  Change Detection Study Time Frame

     NALC  change  detection study is designed to investigate and
select  the  optimal  change detection  techniques  for the  NALC
project. The time  frame for change detection study is as follows:

     September  1992  -  November 1992: Change  detection literature
review.  During this  period,  an  extensive literature  review on
change detection had been conducted. A detailed  internal report on
present change detection  techniques has been accomplished.

     November  1992  -   February 1993:  Commence  the   testing of
techniques with small scale testing. During this period, majority
of  the  present change detection techniques has  been  tested in  a
subscene Chiapas, Mexico.  The NALC project has become conceptually
familiar with many of present change detection techniques. Selected
testing results have been internally reported to NALC.

     February 1993 -  April 1993: NALC change detection  strategy and
procedure  design  and final  refinement.  This  is  done through more
review,  testing  and  personal  contacts   with  change  detection
community.
                                57

-------
     May  1993  -  November  1993:  NALC  pilot  change  detection
experiments.  The  selected change detection  techniques  are to be
tested  in three NALC  pilot  areas.  A draft  report regarding the
result of these experiments will be completed by November 12, 1993.


  Change Detection Correspondences

     During  our process of  soliciting  feasible change detection
techniques for  NALC, the following scientists have shared with us
their knowledge and  experience on change detection.  Many of them
have  kindly  offered  us   their  suggestions  on possible  change
detection techniques for NALC.

     Ford A.  Cross, National Marine Fisheries Service,  (919) 728-
8724.  Cross   is  the  Director  of  the CoastWatch  Change Analysis
Project  (C-CAP), NOAA  Coastal  Ocean Program. The change detection
analysis is primarily  based on a post-classification comparisons.

     Jerry E. Dobson,  Oak Ridge  National Laboratory,  (615)  574-
5937.   Dobson  is  the  principal investigator of  NOAA CoastWatch
Change Analysis Project and designed the change  analysis procedure
for the program.

     Charles  R. Larson, Hughes STX Corporation, EROS Data Center,
(605)  594-6504.   Larson  had performed raw image  differencing
technique to  monitor surface change during the 1991 Persian Gulf
War period.

     Bill Lawrence, Geography, University of Maryland at College
Park, (301) 405-6809.  Lawrence is currently involved in the NASA
Landsat  Pathfinder -  Tropical Deforestation Project  in  Central
Africa. This  project  is planning  to conduct post-classification
change detection on classified image maps obtained by a multiple
threshold technique.

     Richard A. McKinney, Hughes STX Corporation, EROS Data Center,
(605) 594-6500.  McKinney  had conducted NDVI differencing change
detection for continental America for USGS.

     Douglas  Muchoney, The  Nature  Conservancy,  (703)  841-5300.
Muchoney  suggested  that  principal  component  analysis  could
significantly reduce sensor  and atmospheric  effects  from the raw
images,  and  therefore it  would worth  to  test   some  principal
component analysis in  some NALC pilot areas.

     David Peterson,  NASA/Ames Research  Center,  (415)  604-5899.
Peterson had  used  image  regression  technique with chi-square for
detecting changes.
                                58

-------
     Paul M.  Seevers,  Hughes STX Corporation,  EROS  Data Center,
(605) 594-6010.  Seevers had applied a sequential change detection
techniques to the Dallas - Ft. Worth region as well as in the Mt.
St. Helens region.

     Ed Sheffner, TGS Technology, NASA/Ames Research Center, (415)
604-6565.  Sheffner suggested that image regression might serve the
purpose of change detection of NALC, or may be used as a temporal
equalization technique.

     Ashbindu  Singh,  GRID,  United  Nations Environmental Program,
EROS Data Center, (605)  594-6105. Singh has reviewed  (1989) many of
the  present  change detection  techniques  and suggested  that  raw
image  differencing,  NDVI  differencing,  image regression,  and
principal  component differencing produced higher accuracies  for
change detection.

     David Skole,  Complex  System,  University  of New Hampshire,
(603) 862-1792.   Skole is the leader of NASA Landsat Pathfinder -
Tropical Deforestation Project in Southeast Asia. This project will
use change detection methods similar to the Central Africa project
methods.

     John Townshend, Geography, University of Maryland at College
Park. (301) 405 -4050.  Townshend had suggested  to us that we have
more intense  contact  with the scientists  in .the change detection
community.
                          REFERENCES
Anderson, G.  L.,  and Hanson, J. D.,  1992,   Evaluating Hand-Held
Radiometer Derived Vegetation Indices for Estimating Above Ground
Biomass, Geocarto International 1:71-77.

Banner, A.  V.,  and Lynham, T., 1981,   Multitemporal Analysis of
Landsat  Data  for Forest  Cut  Over  Mapping -  a  Trial of  Two
Procedures,  Proceedings  of the 7th  Canadian Symposium on Remote
Sensing,  Canadian Remote  Sensing Society, pp. 233-240.

Byrne, G. F.,  Crapper, P.  F.,  and Mayo, K.  K.,  1980,  Monitoring
Land Cover Change by Principal Component Analysis of Multitemporal
Landsat Data, Remote Sensing of Environment  10:175-184.

Colwell, J.  E., and Weber, F. P., 1981,  Forest Change Detection,
Proceedings  of  the 15th  International Symposium on Remote Sensing
of Environment, Environmental Research  Institute of Michigan, Ann
Arbor, MI, pp.65-99.
                                59

-------
Derring, D. W.  and Haas,  R.  H.,  1980,   Using  Landsat Digital Data
for Estimating  Green Biomass, NASA Technical Memorandum #80727,  21
PP-

Eidenshink,  J.  C., and Haas,  R.  H.,  1992,  Analyzing  Vegetation
Dynamics of Land System With  Satellite Data, Geocarto International
1:53-61.

Fung, T., and LeDrew,  E.,  1987,  Application of Principal Component
Analysis  to  Change  Detection,  Photogrammetric  Engineering  and
Remote Sensing  53:1649-1658.

Hall, F.  G.,  Strebel, D.  E.,  Nickeson J.  E.,  and Goetz, S.  J.,
1991,   Radiometric Rectification:  Toward a  Common Radiometric
Response Among Multidate, Multisensor  Images,  Remote  Sensing  of
Environment 35:11-27.

Ingram, K., Knapp, E., and Robinson, J. W., 1981,  Change Detection
Technique  Development for Improved Urbanized  Area Delineation,
technical memorandum CSC/TM-81/6087, Computer Sciences Corporation,
Silver Springs, MD.

Nelson, R. F., 1982,  Detecting Forest Canopy Change Using  Landsat,
NASA  technical memorandum 83918,  Goddard Space  Flight  Center,
Greenbelt, MD.

Nelson, R. F.,  1983,  Detecting Forest Canopy  Change Due to Insect
Activity Using Landsat MSS, Photogrammetric Engineering  and Remote
Sensing 49:1303-1314.

Pilon,  P.  G.,  Howarth,  P.  J.,  and  Bullock,  R.  A.,   1988,   An
Enhanced Classification Approach to Change Detection in Semi-Arid
Environments,   Photogrammetric  Engineering  and  Remote  Sensing
54:1709-1716.

Richardson,  A.  J.,  and Everitt,  J.  H.,  1992,    Using  Spectral
Vegetation Indices to Estimate  Rangeland Productivity,  Geocarto
International 1:63-77.

Richardson, A.  J., and-Milne, A.  K.,  1983,   Mapping  Fire Burns and
Vegetation  Regeneration   Using  Principal  Component   Analysis,
Proceedings of  IGARSS'83  held  in San Francisco (New York: IEEE),
pp.51-56.

Singh,  A.,  1989,   Digital  Change  Detection Techniques  Using
Remotely-Sensed Data, International  Journal  of Remote  Sensing
10:989-1003.

Shepard, J. R., 1964,  A Concept of Change Detection, Proceedings
of  the   30th  Annual  Meeting   of  the   American Society   of
Photogrammetry, Washington, D.C., pp. 648-651.
                                60

-------
Tappan, G. G., Tyler, D. J.,  and Wehde,  M.  E.,  1992,   Monitoring
Rangeland Dynamics in Senegal with Advanced Very High Resolution
Radiometer Data, Geocarto International 1:87-98.

Townshend, I.  R.  G., and Justice,  C.  O.,1986,    Analysis  of the
Dynamics  of  African Vegetation  Using the  Normalized Difference
Vegetation Index/ International Journal of Remote Sensing 7:1435-
1446.

Toll, D.  L.,  Royal,  J. A.,  and  Davis,  J. B.,  1980,   Urban Area
Up-Date Procedures Using  Landsat Data,    Proceedings  of the Fall
Technical Meeting of the American Society of Photogrammetry, Falls
Church, VA, pp.RS-El-17.

Weismiller, R. A., Kristoof,  S.  J.,  Scholz,  D.  K.,  Anuta,  P. E.,
and  Momen,  S.  A.,  1977,    Change  Detection  in  Coastal  Zone
Environments,  Photogrammetric  Engineering  and  Remote  Sensing
43:1533-1539.
                                 61

-------
•o
 c
 03
 0>


5

 03
 03
P
 CO


 1
 0)
 0>


I
CM
 O


 03
 L.
 OB
 03

*Q
 t_t
•-

 s
CO
 03


 O>



I
 in

 a)
.5   (D

^  t3'5

-------
               NALC QUALITY ASSURANCE AND QUALITY CONTROL

     Quality assurance and quality control  (QA/QC) procedures will
be  developed and  implemented  to insure  and  verify that  NALC
products meet the project's data quality objectives stated herein.
Initial QA/QC procedures have been established and will be tested
in the pilot projects.  The QA/QC  procedures will be evaluated and
revised during the  pilots.  At the completion of the pilots a final
QA/QC guidance document will be prepared.  The principal elements
of the NALC QA/QC include:  Standard operating procedures for all
aspects of the data processing;  quality control checks during data
analysis; implementation  of data tracking forms  to  document all
procedures executed on each MSS  scene; verification and validation
of  spatial   accuracies;  acquisition  of verification data;  and
accuracy assessment of categorizations.


  Standard Operating Procedures

     Each organization participating in NALC is in the process of
developing standard operating procedures and a report documenting
those  procedures.    It is  anticipated and  expected  that  these
reports will  be  revised frequently during the  first  year of the
project.   The NALC components  that  require  standard operating
procedures    include   triplicate   selection   and   production,
georeferencing, image categorization and change detection.


  Quality Control

     Quality  control  procedures  will  include the validation of
initial  image  spectral  quality,  location   and  date  of  image
collection,  and  adequate  georectification.   Both EMSL-LV and the
EROS Data Center  have developed internal quality control checks for
previous projects  that will be applied to NALC.  Aspects of the
quality control procedures cross  over into data analysis tracking
and  spatial  accuracy  validation,  both of which  are  discussed in
more detail after this section.  The importance of quality control
procedures is to establish checks to minimize  and  identify data
degradation at the earliest stage of analysis.   It is important to
document the source of data degradation and probable impacts on the
final products.


  Data Analysis Tracking

     Data analysis tracking includes documentation of all analyses
implemented on the data including quality  control checks.  Forms
designed  to  track  the data can provide  a  means of post-analysis
error evaluation.   A series of these  forms have been developed at
EMSL-LV for the NALC pilots.  An example of  these tracking  forms is
located at the end of  the Image Categorization Section.  All data

                                63

-------
 tracking forms will  be archived  (either  in  digital or hardcopy
 format) ,  by  each  organization  participating  in NALC,  for the
 duration of  the  project and  for a  minimum  of  two years  after
 proj ect completion.
   Spatial Accuracy Verification

    Spatial  accuracy of the MSS georectification includes both a
verification of  each scene  and  a validation of every tenth scene.
The  EROS  Data  Center  performs  a verification  check  of  their
georectification process  on every  scene as part of their internal
quality   control.    They  are  continuing  to perform  the  same
verification process for  NALC.  After  the MSS scenes are received
at EMSL-LV,   every  tenth  scene's  accuracy will  be independently
validated.

    The  EROS Data  Center  uses USGS 7.5 minute quadrangle maps or
the best available maps to  select  control points for image-to-map
georectification.  In the  U.S.,  every ninth map across the scene is
selected for the georectification process.   A total of  25  to 40
maps are available for control point selection per MSS  scene.  One
point per map is selected in both the image and map.   Only those
points for which the.residuals are less than one second (i.e., +/-
80   meters   or   approximately   1.5   pixels)   are   used   for
georectification.  A minimum of 6  control points are required per
scene,   although 10  to  12  points   are  preferred.     For  the
verification process different maps are randomly  selected  and a
total  of  6-7  points  are   randomly  selected.    The  map  and
georectified image coordinates are  statistically compared  for these
points.  Again residuals must be under 1.5 pixels.

     Georectification for NALC involves image-to-image as well as
image-to-map rectification.    The  1980s  MSS  scene  is  first
registered to USGS  maps;  then it  is used as  the "map  base"   for
georectification of the 1970s and 1990s scenes.  The image-to-image
rectification is performed in a different manner than the image to
map, but the verification  process is the same as that stated above.
Initially for the pilot areas,  verification of  every  fifth scene
will be performed for the  image to  image georectification.  If the
spatial  data quality objectives are met for the pilot areas,  then
the number  of scenes that  will  be validated will be  reduced to
every 10th scene.

     The validation begins  by  examining a list  of map control
points used by EROS Data Center to  georectify the scene. This will
prevent  selection  of  a georectification control  point for  the
validation,  i.e.,  use of  georectification control points in  the
validation process would result in a positive bias.  Then a random
selection of at  least four  maps  that were not used by  EDC  in the
rectification process is made.  At  least three points per map that
are identifiable in the  imagery are randomly selected.   The map and

                                64

-------
image  coordinates  for the point are  statistically  compared.   An
RMSE of  +/-1.0  (60m) pixel  has been established as  the spatial
accuracy  data  quality  objective   (DQO)  for  the  image to  map
rectification  and  +/-0.5  pixel  (30m)   for  the  image-to-image
rectification.   Therefore/  in comparing  an  image rectified by
another image  to a map base,  the  total  spatial  error should not
exceed +/-1.5 pixels (i.e.,  +/- 1.0 pixel for the  image to map plus
+/-0.5 pixel for the image-to-image).  If the RMSE is larger than
+/-1.5 pixels, the EROS Data Center will be contacted.  If,  after
evaluating  the georectification process,  it  is  determined  that
there  are  insufficient identifiable  points  to produce  a better
rectification, the scene will  be used,  but it will  be flagged to
identify that  the spatial accuracy is less than the DQO.


Thematic Accuracy Assessment

     The purpose of performing an accuracy assessment of the NALC
categorized data  is  twofold.  General  scientific ethics and EPA
policy  require  that all  final  products  have   a  stated  known
accuracy.  Most importantly for NALC,  the accuracy assessment will
enable  a more precise  estimate  of  the  areal   extent of  each
category; in other words,  an assessment  of categorical accuracy is
more  important  than,  and  prevails  over  the necessity for  an
assessment  of   overall  map  accuracy.     Categorical  accuracy
information will be  critical in the development  of  realistic and
reliable continental carbon  and climate models.   Several methods
have .been presented in the  literature  for improving area estimates
of categorized data  (Card 1982, Hay  1988, Jupp  1989,  Conese and
Maselli 1992,  Czaplewski  and  Catts 1992).  All methods presented by
these  authors use  the  error  matrix derived from an  accuracy
assessment  to  calculate  corrected  or  unbiased  estimates  of
individual  category  areal extent.    Each  of  these authors  also
states that the choice of sample design and sample  size is critical
to the valid use of an error matrix for improving area estima'tes.
The  following  text describes  proposed  methods for conducting  a
pilot accuracy assessment that will result in a statistically valid
error  matrix.    (NOTE:   These methods  are  not   necessarily  the
methods that  will be used  for all NALC  categorization accuracy
assessments.  They are test methods for  the pilot that will enable
the NALC team to  evaluate categorical  error variances,  reasons for
errors, the applicability of  different scale verification sources,
and the overall efficiency and cost effectiveness of the assessment
approaches tested.  Future  pilots based on this initial assessment
will further test modified accuracy assessment procedures.)


Field Verification

     As stated previously, Landsat MSS  data will  be used for the
categorization.   The resolution of these data combined with the
level of categorization indicate that a detailed field verification

                                65

-------
may  not be  required to  perform a  statistically  sound accuracy
assessment.    Therefore,  the verification  effort will  be based
primarily    on    MSS    image  *  interpretation    and    aerial
photointerpretation  with  very limited examination of actual field
sites.  A three-band color composite  of the MSS images will  be used
with ancillary data  as a primary verification source.  While high-
resolution  aerial photographs would be  the optimum verification
source, it is recognized that for some areas of the U.S. and almost
all  of Central America, aerial photographs  from  a compatible time
period will not exist and acquisition of  new aerial photographs is
cost prohibitive.   Whenever possible,  aerial photographs will be
used as the primary verification source.   Field sites will  only be
examined   as  a   validation   check  of   the  MSS/photographic
interpretation sites.  Whenever possible, all interpretations will
be performed by experienced image interpreters or  photointerpreters
who  are not directly involved in the  NALC project.  However, there
may  be some cooperators from Central  America who  will not have the
resources available  for an independent interpretation.  In either
case, care  will  be  taken  to ensure  that the interpreters  do not
have access to the categorized image  as this may  cause bias in the
interpretation.    During  the  pilot projects,   testing   of  the
different  sources   and   scales  of  verification  data  will  be-
evaluated.   In particular, the  Chesapeake Bay pilot area will be
verified with MSS image interpretation, aerial photointerpretation,
and  a field validation.  The results and conclusions  of this effort
will aid  in the  determination of a  final sample scheme for NALC
accuracy assessments.


Sampling Design

     Sampling design includes the choice  of  sampling  scheme, i.e.,
random or systematic, and  identification of an appropriate sample
size.    The  remote  sensing literature states advantages  and
disadvantages of both random and systematic  schemes and is divided
in its support of each scheme (see  Congalton,  1991 for a review of
this literature).  However, it is recognized that the choice of an
appropriate sampling scheme is based on  the goals of the accuracy
assessment.  As stated previously, the goals of  the NALC thematic
accuracy are to state the accuracy of the  mapped categories  as well
as to provide the necessary statistics to improve categorical area
extent estimates.  Prisley and Smith (1987) state that "A crucial
assumption in  this use of  error matrices,"  i.e.,  to improve areal
extent  estimates,   "is  that  the  distribution  of  error  in  the
contingency    table   is   representative   of   the   types   of
misclassification made in the entire area classified." Card (1982)
also states that the sampling design  employed in  the collection of
verification data dictates how the  error  (or confusion) matrix may
be composed to provide conclusions about  the entire population.
                                66

-------
     The  choice  of  a  sampling  scheme  for  the  NALC  accuracy
assessment is based on the following requirements and assumptions:
     1)  Individual  category   accuracies  must   be  determined,
including errors of omission and commission;
     2) the variance of categorical misclassification is unknown,
i.e., no previous pilots have been performed to provide an estimate
of the variances;
     3) the  existence and magnitude  of misclassification spatial
autocorrelation is unknown,  but one is assumed to be present;
     4) the possability of visiting 100% of the verification sites
on the ground is logistically and cost prohibitive;
     5)  the  availability of aerial photographs  for verification
from a  similar  date or  season  as the MSS acquisition  is highly
improbable for large portions of the NALC project•area;
     6)  the  MSS  data has already been acquired,  therefore it is
impossible to acquire simultaneous "in the field" data;
     7)  each MSS scene  will be categorized individually for the
Chesapeake Bay pilot, i.e., each scene will have cluster statistics
generated for it independently of the other scenes within  the pilot
area.   This  will probably result  in different misclassification
variances for each scene;
     8)  and  the result of  the  accuracy assessment must  be  a
statistically valid  error matrix  that may be  effectively used to
better estimate  the area of individual categories.
Sampling Scheme

     While  numerous papers  may be  found that  discuss sampling
schemes in the remote sensing literature,  only two studies to date
(Congalton  1988  and Stehman 1992)  have compared several sampling
schemes using the same categorized remotely sensed data.  Congalton
recommends  the use  of a simple random or  stratified random sample
while Stehman is  a  proponent of systematic sampling schemes.  The
two  papers  do,  however,  have  similar conclusions.   First,  both
authors state  that the choice of sample  scheme  must  be based on
project  objectives.   The objectives  of  the  assessment  and the
statistical approach drive  the choice of  sampling scheme.

     Secondly, both authors indicate that periodicity or spatial
autocorrelation of  misclassification errors can  significantly and
negatively  impact the  precision of a systematic sample.  Stehman
(1992) states that  systematic sampling designs should not be used
if periodicity is suspected, unless the nature of the periodicity
is   understood  well  enough  to  avoid  an unfavorable sampling
interval.   Congalton  (1988)  examined three  images  of differing
landscape pattern types, namely,  forest, range  and agriculture.  In
each case the misclassif ication errors were found to be significant
and  positively spatially  correlated.   The spatial autocorrelation
of NALC categorization errors will be unknown until after the  first
accuracy  assessment  is  performed.    Therefore,  until accuracy
assessment   data  is   acquired   and   estimates  of  spatial

                                67

-------
 autocorrelation may then  be calculated, spatial  autocorrelation
 must  be assumed.

      Another concern in selecting  a  sampling scheme is how well
 that  scheme  estimates  variance.    Stehman  (1992)  states that
 estimation of  error  variance may  be problematic for  systematic
 samples.   Cochran  (1977)_ states  that  there is  "no trustworthy
 method  for estimating  VfVjy)  from the sample data is known" for
 systematic samples.   V(ysy)  is  the variance of  the error mean.
 Rosenfield  (1982)  agrees  with  and   sites  this  statement from
 Cochran.   Congalton  (1988)   states that simple  random sampling
 provided good estimates of  the  mean  error  and variance for each
 image type (forest, agriculture and range).  He further states that
 "stratified  random  sampling also performed well  and should be used
 especially when  it is  necessary to  make  sure  that  small,  but
 important,   areas  are   represented  in   the  sample,"  and that
 systematic samples should only be used with  caution,  i.e.,  in
 situations where  the variance and spatial autocorrelation are well
 understood.

      As stated  previously, the primary objective of the accuracy
 assessment is to provide estimates of  individual category accuracy
 in  the  form of an  error matrix.   The error  matrix must address
 errors  of omission and commission so  that  it  may be used  to
 calculate better  estimates per category.   Given  these constraints,
 the  sample  design  selected  must provide  the  most  precise  and
 unbiased estimator  of  individual class accuracy.   Also, the NALC
 categorization  process  will  cluster  each  Landsat  MSS  scene
 independently.  Therefore,  a between-scene error variance should be
 accounted  for in the pilot  sample design.    Based on the   remote
 sensing and  statistical literature, a stratified random sampling
 scheme will best meet the needs of the  first  pilot and will be used
 for the Chesapeake Bay pilot on a scene by scene basis.  After the
 assessment data is acquired,  a careful  evaluation of scene to scene
 and  categorical  variances will  be made to  determine the most
 appropriate  and  cost effective  sampling scheme  for  future NALC
 accuracy assessments.  Until the variances resulting from the NALC
 categorization procedures are estimated and  understood,  the  choice
 of a  systematic sample may result  in an imprecise and/or  biased
 estimate of  the error.-
Sample Size

     The  number of  samples  to be  selected for  verification is
dependent  on the goals  of the project.   If only  a right-wrong
assessment  is needed  then the binomial  distribution  or  normal
approximation  equation may be used to  calculate  the sample size
(van Genderen and Lock  1977, Rosenfield and Fitzpatrick-Lins 1982).
However,  the  binomial  distribution  does not provide a sufficient
sample  size  to  construct  a  statistically  valid  error  matrix.
                                68

-------
Fitzpatrick-Lins  (1981) and Borella et al.  (1982) used the normal
approximation equation and binomial distribution, respectively, to
calculate  sample  size for  their  accuracy assessments.   In both
cases,  the sample sizes were  insufficient  to construct complete
error matrices.   An  appropriate distribution to calculate sample
size  for  development of  an  error  matrix  is the  multinomial
(Rosenfield   1982).   However,  correct usage of  the multinomial
distribution  requires  information on the expected  error and the
variance  associated  with  that  error.   Since the  categorization
methodology has  not  been  thoroughly  tested prior  to  this pilot
there are  no  variance  estimates available.   Initial calculations
using the multinomial equation presented in Rosenfield (1982) and
the preliminary  results from  a previous  EMSL-LV  categorization
indicate that a sample size between 137 to 163  samples per category
is appropriate.  The equation is depicted below.
     n  = the sample size
     X2 =  the Chi Square distribution
     .6, =  the half width of the desired confidence interval for a
          particular category j.
     a  = the Chi Square distribution tail;  this definition is not
          directly stated  in Rosenfield  (1982), but is surmised
          from the numbers presented
     k  = the number of categories
     E, =  the estimated proportion of  correctly categorized pixels
          for a  particular category j_; this definition is not
          directly stated  in Rosenfield  (1982), but is surmised
          from the text
     However,  the EMSL-LV  data  does not  provide  a  sufficient
measure of error variance and questions arose over the proper use
of  the  multinomial   equation  presented  by  Rosenfield  (1982).
Further   examination   of   the   multinomial    distribution   and
confirmation of  the  multinomial equation  presented by Rosenfield
(1982) will be performed prior  to finalization of the NALC pilot
accuracy  assessment  procedures.   As a minimum,  Hay  (1979)  and
Congalton (1991)  recommend that sample size  per class should be at
least  50  to  test the accuracy  of  determinations,  i.e.,  not only
right versus wrong but examine  the multiple classes  of wrong.

     Another approach used to  determine  sample size  is  an area
weighted  sample.   This may be  an  overall area weighting or area
weighting by  category.  The  overall area  weighting specifies a
total sample size that represents a given percentage  of the entire
population  or  image.    The   categorical   area weighting  is  a
stratification of the overall area weighting wherein the class with
the largest area has  the largest number of  samples again based on
a  given  percentage.    Intuitively,   the   random  or  systematic

                                69

-------
 selection of points from a population should yield similar results
 whether or not the selection  process  is stratified by category.
 Congalton (1988)  recommended  that at  least one  percent  of the
 entire  image be sampled for agricultural, range, and forested image
 subsets.   He also states that his results indicate a larger sample
 size  should  be used if  a stratified  systematic unaligned  or
 systematic sampling scheme is  used.

     Based on  the information  above the NALC Chesapeake Bay pilot
 sample  size will contain no fewer than 50 points per category with
 a  maximum number  of points  per  category  as   determined  by the
 multinomial distribution.   If  the error variances  for individual
 categories or  scenes  are higher  than  anticipated,  additional
 samples will be taken (i.e.,  a double sample) to better define the
 variance.     Based  on  the   final  results   of   the  pilot,   a
 recommendation will be  made  for the appropriate choice of sample
 sizes for future NALC categorization accuracy assessments.
Verification Procedures

     After the verification sites have been selected, MSS Landsat
three-band color composites will be interpreted.  The pixel closest
to  the  accuracy assessment  sample  site will  be  used as  the
verification center point.  A square area consisting of two pixels
per side from the center of the "sample" pixel will be interpreted,
i.e.,  approximately  a 5x5 pixel area.   A draft  form for the MSS
verification  is  displayed  in Table  7.   EMSL-LV  has in  its
possession aerial photographs covering a significant portion of the
Chesapeake Bay watershed.  Any aerial photographs coinciding with
a sample site will be interpreted to evaluate the effect of scale
and as a validation check of  the MSS  interpretations.  Location of
the coordinates on the photographs may be performed by appropriate
scaling, and using maps and analytical stereoscopes.  Table 8 lists
the information that will be  documented  for photointerpreted site.
The  verification forms  should be  filled out  as completely  as
possible for each site.   There will probably  be some areas where
changes have occurred since the date of photograph acquisition or
field visit.  For areas where changes are known to have occurred,
the interpreters should consult with people familiar with the ares.
to collect  up-to-date information.    The source of  this type  of
information should be documented on the verification forms.

     A  field check  will  be performed for four randomly selected
verification points  per  category as a  validation check of  the
interpreted data.  Therefore,  only those points that were included
in both the MSS  and  photo  interpretation will be used  for this
selection.   The  field check may be  performed using one of  two
methods:  walking the site area or  flying over the site at a low
altitude.  Low altitude  aircraft  will  primarily be used in areas
with no access or prohibitively dense vegetation.  Maps, GPS units,
and/or  other  surveying  techniques  may  be  used  to  identify

                                70

-------
locations.  The method of  field checking will depend upon the site
conditions, access,  and  availability  of  resources.   Therefore,
choice  of method  will be based  on  local conditions  and  the
capability of  other  NALC  cooperators.  If  access to any  of  the
sites is prohibitive, then another site will be randomly selected
to replace it  (i.e.,  sampling with replacement).   A verification
form  displayed will  be  used  to  record  as much detailed  site
information as possible and will be compared to the MSS and photo
interpretation forms  for that site.  In addition, 35mm photographic
slides will be  acquired at  field sampling sites.   The field form
level of detail will provide more information than  is necessary for
a validation check.   However, previous experience has shown that
this level of detail is beneficial for evaluating sources of error.
Field crews will consist of at least two individuals per team; at
least one of the team members will have expertise operating GPS and
one with  expertise on the local  biota.   If  there  are significant
differences among the different  scales of verification data,  the
field data and slides  will be used to evaluate possible reasons for
the  differences.   If possible,  a  procedure for  correcting  the
differences will be developed.

     All  individuals  employed  to  perform the  verification  /
validation  of  thematic  accuracy  will  have   expertise  in  the
appropriate areas and whenever possible, will  be  external to the
NALC project.   In addition,  these individuals will not have access
to the categorized MSS images to  reduce the  likelihood of positive
or negative biases.
                                71

-------
Table 7.
            NALC MSS  Interpretation Verification Form

Analyst: 	
Date of  Interpretation:
Verification  Site  Information

MSS Scene ID#: 	    Scene Date:

Verification  Site  ID#:
Scene Coordinates of Verification Site:
Bands and  enhancements  used for image color composite:   Red
Green       Blue
Ancillary Data  (list any maps, photographs, local experts/ etc.)
Interpretive Information


Describe  the  size/  shape/  color/ texture, and  types  of features
within the verification site:
List Appropriate Category Name:
Second choice category name (if applicable)

                                72

-------
Table 8.
        NALC Photographic Interpretation Verification  Form
Analyst:
                              Film type:
Date of Interpretation: 	

Imagery Date: 	

Frame #: 	

Scale: 	    Source: 	

Stereo pairs:  	 yes  	 no

Verification Site Information

MSS Scene ID#:	  Verification Site ID#:

   Photograph Frame #: 	
Method   of   locating  verification   site  on   the   photograph:
Interpretive Information

Describe the  size,  shape,  color, texture, and  types  of features
within the verification site:
List Approprite Category Name:
Second Choice Category Name
  (if applicable):	
                                73

-------
Thematic Accuracy Assessment Reporting
     A  confusion  (error)  matrix and  its  associated parameters,
errors  of commission and  omission  and overall  accuracy,  will be
calculated  as described by  Story and Congalton  (1986).   In the
matrix,  the verification data will  be  compared to the categorized
data  as  columns  and rows,  respectively.   The  cells  within the
matrix  indicate the number  of  pixels categorized  as  category a
through  n that correspond to the verification  sites as labelled
category a  through n.   For instance, in the example error matrix
(Table 9) for the  100 verification  sites examined for category a,
65 were  correctly categorized in the remotely sensed data, 25 were
incorrectly   categorized   as  category  b,   and  10  incorrectly
categorized as category c.  The diagonal from the upper  left to the
lower  right  represents the number of  image   pixels  correctly
categorized  and,  when  divided  by  the  total  number   of  pixels
verified, yields the overall  accuracy of the categorization.  There
are two ways to calculate the accuracy of individual  categories and
they  are termed  user's and producer7 s  accuracies  or  errors  of
commission and omission, respectively.

Table 9.   An Example Error  Matrix  Table
CATEGORIES
Image a
Image b
Image c
Producer7 s
Accuracy
Reference a
65
25
10
65/100 =
0.65
Reference b
5
85
10
85/100 =
0.85
Reference c
10
15
75
75/100 =
0.75
User7 s
Accuracy
65/80 =
0.81
85/125 =
0.68
75/95 =
0.79
225/100. =
0.75
   Sum of major diagonal = 225    Overall  accuracy = 225/300 = 75%
     User7s  accuracy or errors  of  commission may be  defined as
whether  a  categorized pixel within  an  image actually represents
what is  on the ground or not, or how accurately or reliably the
categorized map represents  the area.  Producer7s accuracy or errors
of omission approach  the question of categorized accuracy from an
opposite direction, that is how well a sample site or area can be
represented  by a  categorized  image or map.   The  example  error
matrix demonstrates that the two calculations do not yield the same
results.   Both of these types of  individual category accuracies
will be presented to provide a more  complete  identification of the
types of errors present in the  categorization.   In addition, the
Kappa  coefficient  (Congalton  1991)  will  be calculated.   Kappa
                                74

-------
measures the  relationship  of non-random categorization agreement
versus expected disagreement. It will be used to monitor trends in
categorization reliability from one categorization to another.  The
Kappa  coefficient equals  zero when the  agreement between  the
categorized  data  and  ground  truth  equals  chance  or  random
agreement.  Kappa increases  to one as chance agreement decreases,
and becomes  negative as random assignment  of categories  occurs.
Kappa equal  to  one occurs only when  there  is perfect agreement.
After accuracy assessment calculations are completed an evaluation
will be  performed to  determine  where  the  errors occurred.   In
particular,  comparisons  will be  made between  the error matrices
derived from MSS and aerial photograph interpretation, and between
aerial photointerpretation and field data.   It is essential that
errors identified by the assessment be evaluated to:  address the
adequacy of differing scales of verification data; identify error
sources;  separate categorization errors  from other  errors;  and
correct the errors, if possible.


                             REFERENCES

Borella, H.M., J.E. Estes, C.E. Ezra, J. Scepan, and L.R.  Tinney,
1982,  Image  Analysis  for  Facility  Siting:   A Comparison of Low-
and High-Altitiude Image Interpretability for Land Use/Land Cover
Mapping,  A final report prepared  for Office of Nuclear Regulatory
Research, the U.S. Nuclear Regulatory Commission, NUREG/CR-2861 S-
744-R RE, 61p.

Card, D.H., 1982, Using Known Map  Category Marginal Frequencies to
Improve  Estimates  of  Thematic  Map  Accuracy,    Photogrammetric
Engineering and Remote Sensing 48:431-439.

Cochran, W.G., 1977,  Sampling Techniques,   John Wiley & Sons, New
York, 428 p.

Conese, C. and F. Maselli, 1992,  Use of Error Matrices to Improve
Area Estimates with Maximum  Likelihood Classification Procedures,
Remote Sensing of Environment 40:113-124.

Congalton, R.G.,  1988,  Using Spatial Autocorrelation Analysis to
Explore  the  Errors in Maps  Generated from  Remotely Sensed Data,
Photogrammetric Engineering  and Remote Sensing 54:587-592.

Congalton, R.G.,  1988,  A Comparison of  Sampling  Schemes  Used in
Generating  Error  Matrices  for Assessing  the  Accuracy  of Maps
Generated from Remotely Sensed Data,  Photogrammetric Engineering
and Remote Sensing 54:593-600.

Congalton,  R.G.,  1991,  A Review on Assessing the  Accuracy of
Classifications  of  Remotely Sensed Data,    Remote Sensing of
Environment 37:35-36.
                                75

-------
Czaplewski,  R.L. and  G.P.  Catts,  1992,  Calibration of Remotely
Sensed  Proportion or Area Estimates for Misclassification Error,
Remote  Sensing  of Environment  39:29-43.

Fitzpatrick-Lins, K.,  1981,  Comparison of Sampling Procedures and
Data Analysis for a.  Land-Use and Land-Cover Map,  Photogrammetric
Engineering and Remote Sensing 47:343-351.

Hay, A.M.,  1979, Sampling Designs to Test Land-Use Map Accuracy,
Photogrammetric Engineering  and Remote Sensing 45:529-533.

Hay,  A.M.,  1988,  The  Derivation of  Global  Estimates  from  a
Confusion Matrix,  International Journal  of Remote Sensing 9:1395-
1398.
Jupp,  D.L.B.,  1989,
Confusion  Matrices,
10:1563-1569.
The  Stability  of  Global  Estimates  from
'International  Journal  of  Remote  Sensing
Prisley,  S.P.  and  J.L.  Smith, 1987,  Using Classification Error
Matrices  to Improve the Accuracy  of Weighted Land-Cover Models,
Photogrammetric Engineering and Remote Sensing 53:1259-1263.

Rosenfield,  G.H.  and K.  Fitzpatrick-Lins,  1982,   Sampling  for
Thematic  Map Accuracy Testing,   Photogrammetric Engineering and
Remote Sensing 48:131-137.

Stehman,  S.V., 1992, Comparison of Systematic and Random Sampling
for Estimating the Accuracy of Maps Generated from Remotely Sensed
Data,  Photogrammetric Engineering and Remote Sensing 58:1343-1350.

Story, M. and R.G. Congalton,  1986, Accuracy Assessment:  A User's
Perspective,    Photogrammetric  Engineering  and Remote  Sensing
52:397-399.

Van  Genderen,  J.L.  and B.F.  Lock,   1977,  Testing  Land-Use  Map
Accuracy,  Photogrammetric Engineering and Remote Sensing 43,: 1135-
1137.
                                76

-------
APPENDICES
 77

-------
                        Appendix
I:
NALC Technical Review

 NALC Technical Review Session agenda

 List of Invited Participants

 NALC Review Panel

 Report of the Technical Review Panel, North American
  Landscape Characterization, February, 1993

 Comments on the Report
                            78

-------
5:30 pm
6:00
6:45
7:15
8:00
     NALC TECHNICAL REVIEW SESSION


       JAMES LAWLESS, MODERATOR

    Wednesday, February 17, 1993




           OPENING REMARKS

JAMES LAWLESS, NALC PEER REVIEW CHAIR
         NASA HEADQUARTERS
                     THE "MISSING CARBON" STUDY:
                   IDENTIFYING GAPS IN AVAILABLE DATA

                          ALLAN AUCLAIR
                   SCIENCE AND POLICY ASSOCIATES
                         NALC PROGRAM OVERVIEW

                 ROSS LUNETTA, NALC TECHNICAL DIRECTOR
                         USEPA, EMSL LAS VEGAS
                    NALC USGS/EDC PROGRAM OVERVIEW

             JAMES STURDEVANT, USGE/EDC NALC PROGRAM MANAGER
                        USGS EROS DATA CENTER
                               ADJOURN
                               79

-------
                 NALC-PATHFINDER TECHNICAL SESSION



                       ROSS  LUNETTA,  MODERATOR

                    Thursday,   February 18, 1993
8:30 AM
9:00
9:30
10:00-10:30
                           NALC OVERVIEW

              ROSS  LUNETTA,  NALC TECHNICAL DIRECTOR,
                     US  EPA,  EMSL Las Vegas
                 NALC LANDSAT  MSS DATA ACQUISITIONS

                       JAMES LOVE,  EOSAT
           NALC DATA PURCHASES
        AND PROCESSING PRIORITIES

       JOHN LYON, OHIO STATE UNIVERSITY
                  BREAK
10:30
           USGS NALC OVERVIEW

     JAMES STURDEVANT, EROS DATA CENTER
11:00
IMAGE SELECTION, COMPOSITING AND GEOREFERENCING

         JOHN DWYER, EROS DATA CENTER
11:30
          IMAGE CATEGORIZATION

     JOHN LYON, OHIO STATE UNIVERSITY
                               80

-------
5:30 pm
6:00
     NALC TECHNICAL REVIEW SESSION


       JAMES LAWLESS, MODERATOR

    Wednesday, February 17, 1993




           OPENING REMARKS

JAMES LAWLESS, NALC PEER REVIEW CHAIR
         NASA HEADQUARTERS
                     THE "MISSING CARBON" STUDY:
                   IDENTIFYING GAPS IN AVAILABLE DATA

                          ALLAN AUCLAIR
                   SCIENCE AND POLICY ASSOCIATES
6:45
7:15
                         NALC PROGRAM OVERVIEW

                 ROSS LUNETTA, NALC TECHNICAL DIRECTOR
                         USEPA, EMSL LAS VEGAS
                    NALC USGS/EDC PROGRAM OVERVIEW

             JAMES STURDEVANT, USGE/EDC NALC PROGRAM MANAGER
                        USGS EROS DATA CENTER
8:00
                               ADJOURN
                               79

-------
                  NALC-PATHFINDER TECHNICAL SESSION



                       ROSS  LUNETTA,  MODERATOR

                    Thursday,   February 18, 1993
     AM
9:00
9:30
10:00-10:30
                           NALC OVERVIEW

              ROSS LUNETTA,  NALC TECHNICAL DIRECTOR,
                     US  EPA,  EMSL Las  Vegas
                 NALC LANDSAT MSS  DATA ACQUISITIONS

                       JAMES LOVE,  EOSAT
           NALC DATA PURCHASES
        AND PROCESSING PRIORITIES

       JOHN LYON, OHIO STATE UNIVERSITY
                  BREAK
10:30
           USGS NALC OVERVIEW

     JAMES STURDEVANT, EROS DATA CENTER
11:00
IMAGE SELECTION, COMPOSITING AND GEOREFERENCING

        • JOHN DWYER, EROS DATA CENTER
11:30
          IMAGE CATEGORIZATION

     JOHN LYON, OHIO STATE UNIVERSITY
                               80

-------
5:30 pm
6:00
6:45
7:15
8:00
     NALC TECHNICAL REVIEW SESSION


       JAMES LAWLESS, MODERATOR

    Wednesday, February 17, 1993




           OPENING REMARKS.

JAMES LAWLESS, NALC PEER REVIEW CHAIR
         NASA HEADQUARTERS
                     THE "MISSING CARBON" STUDY:
                   IDENTIFYING GAPS IN AVAILABLE DATA

                          ALLAN AUCLAIR
                   SCIENCE AND POLICY ASSOCIATES
                         NALC PROGRAM OVERVIEW

                 ROSS LUNETTA, NALC TECHNICAL DIRECTOR
                         USEPA, EMSL LAS VEGAS
                    NALC USGS/EDC PROGRAM OVERVIEW

             JAMES STURDEVANT, USGE/EDC NALC PROGRAM MANAGER
                        USGS EROS DATA CENTER
                               ADJOURN
                               79

-------
                  NALC-PATHFINDER TECHNICAL SESSION



                       ROSS LUNETTA,  MODERATOR

                    Thursday,   February 18, 1993
8:30 AM
9:00
9:30
10:00-10:30
                           NALC OVERVIEW

              ROSS  LDNETTA,  NALC TECHNICAL DIRECTOR,
                      US EPA,  EMSL Las Vegas
                 NALC  LANDSAT MSS DATA ACQUISITIONS

                        JAMES  LOVE,  EOSAT
           NALC DATA PURCHASES
        AND PROCESSING PRIORITIES

       JOHN LYON, OHIO STATE UNIVERSITY
                  BREAK
10:30
           USGS NALC OVERVIEW

     JAMES STURDEVANT, EROS DATA CENTER
11:00
IMAGE SELECTION, COMPOSITING AND GEOREFERENCING

         JOHN DWYER, EROS DATA CENTER
11:30
          IMAGE CATEGORIZATION

     JOHN LYON, OHIO STATE UNIVERSITY
                               80

-------
12:00-1:30
1:30 PM
2:00
2:30
                  LUNCH
             DATA MANAGEMENT:
     INDEXING, ARCHIVING AND DISTRIBUTION

          JOHN DWYER, EROS DATA CENTER
         OVERVIEW OF CHANGE DETECTION

       DING YUAN, DESERT RESEARCH INSTITUTE



       ACCURACY ASSESSMENT: QA/QC ISSUES

  LYNN FENSTERMAKER, DESERT RESEARCH INSTITUTE
3:00-3:30                    BREAK
3:30       METHODS DEMONSTRATION AND PILOT PROJECTS
                          OVERVIEW

                 JOHN LYON, OHIO STATE UNIVERSITY
4:00
4:30
5:00
6:00
    CHESAPEAKE BAY WATERSHED PILOT PROJECT

    DORSEY WORTHY, US EPA, EMSL Las Vegas



      GREAT LAKES WATERSHED PILOT STUDY

BERT GUINDON, CANADA CENTRE FOR REMOTE SENSING
          OTTAWA, ONTARIO, CANADA



        STRAWMAN REPORT - CONTENTS/OVERVIEW

                REVIEW PANEL MEMBERS


                    ADJOURN


                     81

-------
8:00 AM
10:30
11:30
1:00
 NALC TECHNICAL REVIEW SESSION

   JAMES LAWLESS, MODERATOR

  Friday, February 19, 1993




    REVXEW PANEL REPORT WRITING

           JAMES LAWLESS
       REVIEW PANEL MEMBERS


   QUESTIONS FOR PARTICIPANTS

      REVIEW PANEL MEMBERS


COMPLETE REVIEW OF DRAFT REPORT

        JAMES LAWLESS


     ADJOURN PANEL REVIEW
                               82

-------
                            LIST OP

                       INVITED PARTICIPANTS

               NALC Technical Review, New Orleans
REVIEW PARTICIPANTS

 Allan Auclair
 Roman Alvarez
 Roberto Bonifaz
 John Dwyer
 Jack Durham
 Lynn Fenstermaker
 James Love
 Ross Lunetta
 John Lyon
 James Sturdevant
 Ridgeway Weerackoon
 Ding Yuan
 Dorsey Worthy
PEER REVIEWERS

 Chuck Dull
 John Estes
 Leonard Gaydos
 James Lawless
 Douglas Muchoney
 David Peterson
 Ed Sheffner
 John Townshend
INVITED OBSERVERS

 Peter Beedlow
 Michael Cairns
 Chris Elvidge
 Chris Justice
 Don Lauer
 Gene Meier
 Joel Morrison
 Lee Mulkey
 Courtney Riordan
 Al Watkins
                               83

-------
                   NALC REVIEW PANEL
James G. Lawless, Review Chair
NASA Headquarters
Earth System Science Division
MS 239-20
Ames Research Center
Moffett Field, CA  94035

PHONE:  0415) 604-5900
FTS: 464-5900
FAX:  (414)  604-1068
E-MAIL:  iglawlessgnasamail
Ed Sheffner, Review Coordinator
TGS Technology
NASA/Ames Research Center
MS 239-20
Ames Research Center
Moffett  Field,  CA  94035

PHONE:  (415) 60A-6565
FTS: 464-6565
FAX:  (415)  604-1088
E-MAIL:  ej sheffner@gaia.arc.nasa.gov
John  Estes,  Reviewer
University of California
Santa Barbara
c/o Al Watkins
Chief USGS/NMD
Reston, VA  22092

PHONE: (703) 648-5752
FAX:  (703) 648-57-92
E-MAIL:  lestes@nasamail
estes@pollux.geog.ucsb.edu
 John Townshend, Reviewer
 University of Maryland
 Department of Geography
 1113 LeFrank Hall
 University of Maryland
 College Park, MD  20742

 PHONE: (301) 405-4050
 FAX: (301) 314-9299
 E-MAIL:  JT59@UMAIL.UMD.EDU
84

-------
David Peterson, Reviewer
NASA
AMES Research Center
MS 239-20
Ames Research Center
Moffett Field, CA  94035

PHONE: (415) 604-5899
FTS: 464-5899
FAX: (415)  604-4680
Leonard Gaydos, Reviewer
USGS
MS 242-4
Ames Research  Center
Moffett Field,  CA   94035

PHONE:  (415)604-6368
FTS: 464-6368
FAX:  (415)  604-4680
E-MAIL:   leng@gala.arc.nasa.gov
.Chuck Dull,  Reviewer
 USDA Forest  Service
 14th & Independence Ave.,  SW
 Washington,  DC  20250

 Phone:  (202)205-1416
 FAX:  (202)205-0861
 E-MAIL:
 Douglas Muchoney,  Reviewer
 The Nature Conservancy
 1850 N. Lynn St.
 Arlington, VA  22209

 PHONE:   (703) 841-5300
 FAX: (703) 841-1283
                           85

-------
                Report  of  the
           Technical  Review  Panel

North  American  Land  Characterization  Project
                February,  1993
                   86

-------
                              Summary
A  technical  review  of the North  American  Landscape Characterization
(NALC)  project,  an element of the Landsat  Pathfinder,  was held in New
Orleans,  Louisiana on  February 17-18,  1993.  The members  of the review
panel emphatically and  unanimously endorse the goals, objectives  and
technical approach  of  the  NALC.  Specific recommendations are made to
improve the  utility and scientific merit of the products  derived during the
course  of  the  project,   to  assure  that  the   project  will  meet  the
information  gathering  requirements  and  performance  specifications of
EPA,  and to establish mechanisms for  integrating the  experience gained,
and  the  data generated  by this  project  into other,  large  scale,  global
change research programs.

1.0   Introduction
      The North American Landscape Characterization (NALC) project  is an
excellent example of  the  type  of  pathfinder envisioned by the  National
Aeronautics  and  Space  Administration  (NASA).    Personnel  from the
Environmental Protection Agency (EPA), United States Geological  Survey
(USGS),  NASA,  Earth  Observation  Satellite Corporation  (EOSAT)  and other
cooperating  organizations  have formed  a team involved in  a scientifically
significant,  technologically  challenging  project. NALC  is  being  conducted
on  a scale  that stretches  our  ability to facilitate science  given  today's
understanding  of  image  acquisition,  archiving,   data  processing,
verification  and  data distribution systems,  and it  is providing the science
community   with  essential  data  products.  The   NALC-Pathfinder's
experience  is  also  providing   lessons  in  the  creation  of large  area,
multitemporal, high spatial  resolution data sets - lessons of great value
in  the development of EOS-DIS and  a -variety of global change research
efforts.  The data sets, archiving,  and techniques  developed can, and will,
make important contributions to  resource  management not only in the
United States and  Canada, but  Mexico,  Central America and  the Caribbean
as well.  The NALC-Pathfinder is important.  The  project is  needed.   The
review panel supports  the effort and believes the  program  will  provide a
tremendous  benefit to  cooperators  throughout  North America.
      The review panel supports, fundamentally, the goals, objectives, and
methods of  the  NALC.   As with  all reviews, there  are areas  where the
panel members feel improvements  can be made.   The  panel applauds the
NALC-Pathfinder's  efforts  in technique  development, and, while  comments
                                 87

-------
directed  toward  specific techniques  suggested or  recommended by  the
NALC  staff are presented,  the  comments are intended to  call attention to
ways in which  the final products of the NALC may be  improved in an area
where  continuing research is  urgently required.
      The  goal of the  NALC  program  is to provide  scientists  with  the
information they require  to conduct global change research. The  Pathfinder
specifically focuses on  the  development  of techniques and  methods which
can  be applied to  the  development  of standard land  cover categorization
products in support of global change research efforts  of EPA  Laboratories
in  Corvallis, Oregon, and  Athens, Georgia.  From a science perspective,  the
Pathfinder seeks to  reduce the uncertainty  in the  area term  for carbon
stocks  and  atmospheric trace  gas  emissions  associated  with  North
American  land  cover  type classes.   To accomplish this the  NALC program
will produce  hard copy and digital  image products  from Landsat  MSS data
suitable to inventory  carbon stocks,  determine location of carbon sources
and  sinks  and  identify areas where  the  potential  exists  for carbon
sequestration.
      The  importance of  the  land cover  type categorization  of the NALC-
Pathfinder  should not be  underestimated.   If we take  the equation:
(i_ci)
           A(LC1) + C(LC1)  = ^ota' Glob3' Carbon
             where:
                 A(LC1) = Area of land cover class 1, and
                 C(LC1) - Carbon per unit area for land cover class 1,

the reviewers  believe  strongly  that the uncertainty  present in  the area
term of the equation  is at least as large, if not larger (and much  harder to
address,)  than the current uncertainty in our knowledge of carbon  stock
per unit area for given land cover  classes.   From that perspective,  the
NALC
project contributes toward meeting the goals of global change science.
      Technique  development  initiated within  the scope  of  the  NALC is
also significant.  There  is  a need to develop advanced feature  extraction
procedures for the processing of earth observing  satellite data.  The NALC
is  addressing aggressively  the  issues of categorization,  change  detection,
and quality assurance/quality control (QA/QC) development for large area
cover type mapping.   The  development of a  land cover  type product,  for
example,  is  very much a research area, and any approach taken  is subject
                                 88

-------
NALC Review
Panel Report 2/93

to criticism.
     The review panel received oral presentations on the NALC based on a
draft document, the  "Landsat Pathfinder  Technical  Work  Plan"  prepared
by  the staff  of the  EPA's  Environmental  Monitoring Systems Laboratory.
The findings  of  the review  panel  are presented  in  detail  below.  The
organization  of this  report follows,  in  general,  the  organization  of  the
draft document. Specific suggestions and  recommendations are  made for
data acquisition and processing,  archiving and  analysis  techniques, and
quality control/quality assurance.  Comments are  also directed  toward
project management  and  the  criteria for success.  The report concludes
with a  brief summary and  recommendations. This report is a  response to
the draft  document and the material presented during the  review.  It is  not
intended as a summary of the NALC.

2.0   Goals  and   Objectives
     The stated  goal of  the NALC,    "...assessing terrestrial biosphere
management  options  as  they  influence  land  cover,  carbon  pools,
atmospheric  trace  gases,   and  feedbacks to atmospheric conditions...,"
while  broad,  is made tractable  by the  careful selection and  prioritization
of areas  for  study that address the uncertainties  associated with  sources
and sinks of  carbon in North  America. The proposal by the NALC team to
address  the  goal   through  the  use  of  "triplet"  Landsat  Multispectral
Scanner System (MSS) data sets,  is appropriate and probably essential  for
cost effectiveness.  Further,  the  success  of the  program  is enhanced
through the  leveraging of the  resources  of other  agencies,  universities and
the private sector.  Compiling  the  necessary MSS data  sets  and  initiating
the effort through   pilot  study  areas  to test the  working   scientific
hypotheses   and  to  develop  and  test methodologies are  appropriate
objectives. The use of Landsat  images  from  three  dates  will provide
information on  the current distribution  of  the land  cover  types,  rates  of
change within and among  the land cover types, and trends in  the rates of
change.  Such  information  will  be  of  value  not  only  to EPA  and  the
collaborators  in the NALC but to the US Global Change  Research  Program
as  a   whole.  The  involvement of  cooperators   outside  the  EPA  takes
advantage  of the  best skills  of  the  nation and  helps  assure successful
accomplishment of the objectives  of  the  program  in  a timely  and  cost
                                  89

-------
effective manner.

3.0   Image   Categorization   and   Development   of   Land  Cover
Products
      Two key products will provide the bases  for  the  applications - the
geographically  registered,   multi-epoch,   multispectral   data  being
assembled at EDC, and the clustered data.  NALC is  on target  to realize the
creation  of  both.   Specifically,  its  proposal  to  test alternative  clustering
methodologies and implementations  is a good  approach.  The  review panel
recommends   that  these  tests   be  conducted  using  triplets from
representative environments and  that  measures of success  be established.
The  success criteria  might include  ability to  separate classes of interest,
minimization of the border class  problem, and  ease in labeling.
      No categorization  technique  can be successful without substantial
involvement  of  human   interpreters.   NALC   is  correct  in  stressing
techniques that rely  on  human  expertise  and judgement  in   the  labeling
stage  and  in  trying  to  minimize  the  subjectivity  that  comes  with
techniques that require  selection of  training  fields,  numbers of  clusters,
etc.
      The panel  is concerned about  the qualifications  of  the cooperators
performing the  cluster labeling, and  it recommends that the program make
sure that the  individuals  performing  that function  are  adequately  trained
in  multispectral categorization.   Supplying  the cooperators with  measures
of inter-cluster separability,  divergence  and variance  is  insufficient  if
the information cannot  be used intelligently.   Experience indicates that
spectral  plots  are very useful  in  visualizing the categorization  process.
In  that  regard,  selection  of  software  that  eases  class   labeling  is
absolutely essential.   It should  be  possible to highlight  classes  one  at  a
time,  or  in groups, with easily selected color codes!  Systems that allow
viewing of the multispectral data while labeling are especially helpful.
      The selection of a categorization  system is  obviously important and
difficult.  While the proposed  system appears  to be  one  that  is achievable
with  Landsat MSS data, it does  not seem  to  be "ecologically oriented" as
presented. Since measurement of closed forest canopy and  its change with
respect to clearing and regrowth  is  critical to the Global  Change Research
Program, a closed forest class should be included.   There also appears  to
be a  gap with  respect to  savanna or scrub/shrub categories.   Since NALC
data can be interpreted in concert  with ancillary data,  more  consideration
                                  90

-------
NALC Review
Panel Report 2/93

should  be  given  to  how  such  additional information can  aid  in  the
categorization  or  refinement.    There  is  also  the   problem  of clear
definition  and procedure.   How is  an important phenomenon like forest
regrowth handled?  Is a patch that has just  been cleared "barren,"  or is  it
"herbaceous?"   Due  to  seasonality  inherent in the data,  such a cleared
patch  may appear barren one month, herbaceous the next.  At which stage
does it  get classified  as woody?  In an evergreen  forest (as in the Pacific
Northwest),  does the  regrowth  start out  "deciduous" or. "mixed?"  These
stages,  and how they are handled  in the categorization, are  important to
consider carefully  because maps and  statistics on land cover change will
result from a coarse  temporal sampling of the data over 20 years.  Much
work  needs to  be  done,  prior to  actual   labeling,  in  categorization,
standardization,  description of and  developing application  protocols, and
in   documenting  clear   guidelines  for  cooperators   to  apply  to  the
categorization system.   The  present  categorization system would  benefit
from additional  review  and testing.    In particular,  a  review  of results of
other  categorizations  applied to MSS  data and like techniques is  needed.
There are robust examples from  which  to choose.
     Strong  consideration should  be  given   to  preparation  of  several
categorizations for specific science  users.  Though science  users  of NALC
have been consulted with regard to the  categorization  system, it may be
feasible to  achieve   finer categorizations  for specific  objectives  (like
tracking  forest  regrowth  or  burns) if  multiple  categorizations  was
offered.
     The NALC is creating a data  base of  wide applicability. In addition to
the specific  land  cover  type and land  cover change  products  that will
emerge from the NALC  to meet the science and operational  requirements
of   EPA,   the  variety of  other  potential  products  is  quite  large. These
products  may  include various  image enhancements  to   expedite  photo
interpretation  or  image  mosaicking,  vegetation indices,  categorizations
optimized  for specific  purposes (such  as  identifying stages  of  forest
regrowth),  or   categorizations   mapping  different  classes   than  those
proposed  by NALC.   Furthermore,  a  variety  of measurements  related to
patch  shape and size,  as well  as spatial  autocorrelation,  may be
calculated  based  on  either  the  multispectral   or   categorized  data.
Combining NALC data with other sources, such  as AVHRR time-series, TM,
                                  91

-------
 SPOT, or even digital orthophotos, may also prove valuable.
      The  ability to  use NALC  data in a geographic  information system
 (GIS) is quite  promising, both for enhancing  the  ability to achieve desired
 categorizations, and as  a way of quantifying  measures  over specific  areas
 (such  as  forest districts) or  in association  with other  factors  (such as
 steep slopes).  While these may be beyond the needs or  ability of the  NALC
 program or  EPA at present, it  should be  noted that  the  program  will
 provide a solid foundation for others to  work  from.

 4.0  Methods  for  Managing  and  Distribution of  NALC  Data
 4.1 MSS Data Processing
      Preprocessing  of MSS  data ceased at the Norman OK ground station
 after September  19, 1992.   It  is uncertain that all MSS  scenes acquired in
 1992 during  the acquisition  window  were  pre-processed  prior to  the
 failure.   It is  also uncertain  if (and when)  EOSAT will  bring the capability
 for MSS pre-processing  back  on-line, and/or if any  of  the foreign ground
 stations  retain  the  capability.  Given the  current  status  of  MSS  pre-
 processing, arid the  ongoing  effort by EPA to  bring the Canadians  in to the
 NALC  project as a  scientific collaborator, EPA is encouraged  to  inquire of
 CCRS:
      1. the extent of MSS coverage of Canada for 91-92
      2. if  the  Canadian  Ground Station  at Prince  Albert has maintained the
           capability to  pre-process  MSS data, and
      3. if the HDT's from Norman can be pre-processed by  CCRS.
           EPA is encouraged to determine, as soon as  possible, if all MSS
           data required for  NALC has been acquired  and  pre-processed,
           and to make any unfulfilled  needs  for  MSS data acquisition or
           pre-processing known to the Landsat  program  managers at
           NASA HQ.

4.2 Triplet Processing
      Current  output of triplets at EDC  is about one  every 1.5 days  on a
 multi-tasking  system.   Although the  rate  of data  processing  may be
adequate  to  meet the applications  schedule  as described, increasing  the
throughput rate would be advantageous as it would  allow for pilot testing
and technique evaluation sooner and in a broader  ranger of environments.
                                  92

-------
NALC Review
Panel Report 2/93

4.3 IMS
      Procedures  for  comprehensive  collection  of  ancillary  meta-data
describing the images has  been put in  place by EDC and is ready for placing
these  data in an  Information Management System.   The panel recommends
that the IMS  is implemented as soon  as possible and  that it is not treated
as a  retrospective exercise after image  preprocessing is  complete.
      The  project  would  benefit   now  from  the  availability  of  an
operational IMS so  that the  progress of  the  project in  developing triplets
with know characteristics can be   assessed   comprehensively.  In  the
absence  of an  IMS, project management, currently, has to rely  on overly
qualitative  assessments of many  important image characteristics.
      The experience of the Landsat  Pathfinder in  the  humid tropics in  the
construction of an  IMS  should  be used to help  guide the development of the
NALC-Pathfinder IMS.

4.4 Data Standards
      Compilation of  land cover and land cover change standard products to
repopulate the  data  archive must  be  closely 'monitored and  analyzed with
the cooperation  of the  appropriate   land  management  organization.
Standards for  the  data  products,  must also  be  developed  in  close
coordination with  the Federal  Geographic  Data Committee.

4.5 Field  Inventory/Field Data  Management
      Field inventory should  include  the  consistent  and  proper  use  of
appropriate field forms and data collection and management techniques. A
plan for  management  of  field data information  should be  developed and
implemented. Sources  of  guidance  for field  inventory and  field information
management  are the  US Forest Service and The Nature  Conservancy.

4.6 Sampling Allocation/Sampling Frame Development
      Field verification methods  have  been .developed  by  the NALC-
Pathfinder, but methods  for  guiding  field  evaluations for cluster  labeling
activities  have  been neglected.  Even  when  analysts with  knowledge  of
local  conditions   are  available,  labeling  benefits  from  field  work.  A
sampling frame based  on  cluster categories should be developed and field
sites  selected so  as to  maximize observations  along  gradients.  Sampling
                                 93

-------
should  be nominally proportional to  class or  strata area  while  keeping
potential  access  restrictions  in  mind.  Collection  of  verification  data
should be made concurrently.

4.7 Ancillary Data - Biotic Community
      Biotic community  categories  and  scales need to  be  defined  to
evaluate their utility. The most appropriate  use  of this  type  of  data  needs
to be defined. Biotic community data  may be useful, but research needs to
be  performed  on data availability  and integration techniques.

5.0  Change  Detection  Methodology
      The panel  agrees  strongly that change  detection  is  an important
activity  that can aid in the characterization and monitoring  of  land cover.
Retrospective  change  detection  provides timely estimates  on  the extent
and  type  of  change  as  a  prelude  to  monitoring.     The  panel's
recommendation to refine the change detection methods  primarily  address
technique,  rather  than  program.    Specific areas of   concern  are
multitemporal  radiometric  normalization,  atmospheric  correction,  change
detection  techniques,   test  site   activities,  and   most importantly,
reconciling program  objectives with change  detection  methods and change
categories.
      The science requirements  of NALC for  change detection  information
have  not been  developed and/or reported.    The  objectives  of  change
detection  are  not  clearly developed.   Change  categories have not been
defined  and  reconciled  with the  land cover categorization  scheme.   In
addition,  the  means of  incorporating change  detection  information  into
categorization,   categorization,   stratification  and   sampling   frame
development and verification need  to be further developed.
      It  is not apparent that change detection techniques,  applications and
results  have  been  researched   fully.   The  categorization  of  change
detection  techniques presented to the reviewers may  not be appropriate.
It  does  not  convey  balanced information  on  categories  of "techniques.
      Generally accepted categories of change detection techniques are:
      •  Spectral-temporal change detection (layered-temporal)
      •  Principal components analysis
      •  Image differencing (image  ratioing  inclusive)
      •  Post-categorization change  detection  differencing
Vegetation index differencing  (VID),  the  technique selected  by NALC, is
                                 94

-------
NALC Review
Panel Report 2/93

most   appropriately  a  sub-group  of  image  differencing  or  post
categorization   change  set-differencing  since  derived,  single  data
categorizations  (VI) are subtracted  on  a pixel-by-pixel basis.  At  best, VID
belongs at the same level as the other categories.
      Change detection  method  development  in  NALC  has not  been
adequately undertaken (or documented).  VID was apparently chosen prior
to testing.   It has several serious flaws and other techniques  should have
been  included in  testing and analysis.
      One problem with VID is the possibility that  a  number of land cover
types, given  their  natural  variability  and  high  level  of  abstraction  as
classes, will  map  across  virtually the entire range  of Vis. This  can be the
result  of  either ecological  factors  which  vary  across  the  landscape  and
control biotic  parameters,  e.g.   leaf  area  development,  or  seasonal
variation  in  the   leaf  development of  many  communities as varied  as
"deciduous"  or "cultivated land, or  both.  VI, therefore, may have difficulty
distinguishing corn  fields from deciduous forest  -  a potential  confusion
between  land cover types of some  significance.
      The need  and  methods  for  performing  multitemporal   equalization
(normalization) and atmospheric corrections are  in  question.  There are
techniques such as  principal components analysis (PCA or  Karhunen-Loeve
Transformation)   which  obviate   normalization  and  provide   useful
information   on  sensor  calibration  and  atmospheric  effects.    Image
differencing   would  characterize  the per-band  distribution  of  inter-image
change, thereby providing information on change due to  factors other than
object (feature).
      Vis  are  particularly  sensitive  to  variations  in  atmospheric
conditions. These indices are typically  based on  the  contrasting  biological
influences   on   the  near-infrared   versus   red   reflectance   patterns
(scattering  versus  absorption,  respectively).  Given  Rayleigh  and  Mei
scattering   effects,  particularly   strong  in  the  visible   (red),  the
atmospheric  effects  are  manifested  largely  through  the  red  band
(denominator).  No VI  can  correct  for  atmospheric effects since  they are
additive  not multiplicative.  The project should carefully  re-consider the
atmospheric correction issue.
      It  is  recommended  that  extensive  testing  of  change  detection
techniques  be performed at each  of  the pilot sites - especially  in areas
                                  95

-------
where prior research  data are available such as the Virginia Coast Reserve
LTER which is  within  the Chesapeake Bay  Pilot area.
      The panel recommends that NALC cooperate with  the  larger  technical
science  community,  including   experts  at  EDC,  for  evaluation  and
development  of  change detection  techniques   and methods.   Change
detection  would  be  significantly  helped  by determining,  with the  aid  of
the  scientific  community,  which  land  cover  transitions  are  most
important  and  which  have  little  impact  on the scientific objectives.   This
should lead to  substantial simplification of this part of NALC.
      There is also  a need  to  look at the effect  of  the  long temporal
windows anticipated for  some  areas.  Because of  cloud cover, triplets will
be  created  using  data obtained  at  different tinies in the  growing season.
The inevitable  variation  in  phenology and vegetation condition may make
characterization of condition difficult and  obscure changes  in  land  cover.
Efforts  should be directed toward  developing   techniques that  address
these  anticipated  problems.
      There should also  be a closer tie with the  GCRP program  scientists
to  ensure  that change  detection  products  meet their specific,  scientific
goals.

6.0  NALC  Management.  Scheduling  and  Coordination
      Based  upon   discussions  at  the  management  overview  and
presentation by Allan  Auclair,  it  appears that  the primary  focus  of NALC
should address and  identify fluctuations  in  and  to  better  understand the
carbon flux in  the  temperate/boreal  forest.    Preliminary  conclusions
suggest the temperate/boreal forest  are  not at  climatic equilibrium or  at
carbon  steady-state  and appear to  drive the net biospheric  flux of the
carbon cycle.   Suggested research and  policy  applications  included further
action to   acquire  and  verify data  on  forest depletions  and  inventory
available databases.   Therefore  the development of a terrestrial  carbon
monitoring  system to  track  sinks, sources  and  net carbon  budget annually,
with subsequent  development  of criteria to regulate terrestrial  biospheric
carbon  fluxes,  is  predicated  upon change detection analysis, specifically
forest change  detection and monitoring.
      The  NALC-Pathfinder primary objectives   include;   acquisition  of
current observations  for North America to complement existing  historical
data sets;  to provide easy  access and integration with  past, present,  and
future  data sets; and  derive  land  cover change  output  products using
                                  96
1-0

-------
NALC Review
Panel Report 2/93

standardized  data sets and standardized  methods.  -Close coordination with
major land management organization will be  required to implement  these
objectives  most  effectively.
     The NALC-pathfinder program  has  the  potential  to  provide a wealth
of needed  information to land management  agencies such as  the  USDA
Forest  Service,  BLM, National Park  Service,  TVA,  state agencies and
private individual  land managers.  Their cooperation and participation will
be  instrumental,  if  not absolutely essential,  in  the  development  of
reliable  and  credible  outputs  for the NALC-Pathfinder program  in support
of global change  programs.   Use of existing  databases established by the
organizations will be  necessary to support and validate output products.
These  organizations  also  comprise the  primary  users,  or clients,   the
program  will  support.   Proposed benefits of  the NALC-Pathfinder program
must be  coincident with the data needs of these organization to  ensure the
success  of this interagency, cooperative  program.
     The EPA, USDA,  NASA, & DOI  have already  contributed  substantial
efforts  to support the USGCRP.  Data  acquisitions and  production  of co
registered  MSS triplets will  provide  a  national  archive of tremendous
utility and importance to the  global change research community. However,
the data  base must also address the needs, and meet the requirements,  of
those  agencies  having  legislative  authority  to   affect  the change
necessary to "regulate" terrestrial biospheric  carbon  fluxes.
     Data  acquisition  and assembly of  MSS  triplets is  a substantial task
which will  continue  to  require close  cooperation between EPA,  NASA,
EOSAT,  and  EDC.   Production of the land cover and land cover change
products  is  a major  task requiring  substantial cooperation  between the
aforementioned land  management  organizations.   Other  agencies such as
ASCS,  SCS,  F&WS,  FAS  and international forestry  and  conservation
organizations (to  name  a few)  should support, and receive  benefit  from,
this  program.   Standard  algorithms  and methods  manuals  should  be
provided  to   these   organizations  along with  the  MSS  triplets  and
appropriate technical and financial  support.
     Given   the  complexities   of   interagency,   interdisciplinary
organizations, the NALC  is being managed with a  high degree of success.
The  project  has faced  significant  logistical and research tasks  that  have
necessitated  a flexible,  innovative approach  in project management. This
                                  97
11

-------
quality  was well demonstrated in  the review by  the  effective  way  task
managers responded to questions from the review  panel.
      An area where strengthening  of the  management structure would be
beneficial is in closer coordination  between  EPA  scientists,  for whom the
data  sets  are being  created,  and  the  remainder of the  project.    For
example, representatives  of  this  community  participate  in  the  NALC
Science Working  Group, but  closer  involvement  in the  development of
technical aspects  of the project,  possibly through  the   NALC-Pathfinder
Technical Work Group,  is desirable. Closer involvement is  required so  that
the  technical  development  of  the   project can  respond   to  improved
specification  of  scientific   requirements  and   to   revised  scientific
priorities.   Specifically, the  review  panel recommends  that the NALC-
Pathfinder Technical Working Group is expanded to  include representatives
of EPA  and other global change science groups.
      It  is expected that it will  be necessary  for the Technical Work  Group
to maintain  a relatively  high frequency of meetings to manage  the  large
range of technical  developments anticipated  in the work plan of NALC-
Pathfinder.
      The review panel also  recommends  that  those responsible in  NASA
and  other  agencies  for the  overall   management of  Pathfinder projects
develop active mechanisms to compile  and  analyze the experience of  NALC-
Pathfinder,  and other  Pathfinder  projects, so  that EOS-DIS will benefit
from  the Pathfinder  activities.    One possible mechanism  might  be  to
create  an   overall  Pathfinder Science   Committee   containing agency
managers,  chairs  of the individual  science  teams  and managers of  the
pathfinder  projects.
      Another  area  where the  input  of  the  EPA science  community is
especially desirable at present  is in the sequence  of areas to be analyzed
by the  project   It  would appear  that by modifying  the  order in  which
different regions  are  analyzed, a  more timely  response  to  scientific
priorities could  be achieved.   For  example,  an  earlier  analysis of  the
Pacific  Northwest than  that  currently  planned,  would seem to be needed.
Also,  the order in which  areas  are analyzed for testing and development of
methodology should be  reconsidered so that a wider range  of environments
is examined than those currently proposed.   It is  therefore  recommended
that  a  review of  the  sequence  of data  processing  and  analysis be
implemented.
                                 98
1 2

-------
NALC Review
Panel Report 2/93

7.0 Quality  Assurance  and  Quality  Control
     QA/QC in the NALC is  in the developmental stage.  The review panel
is  pleased  to  see  that this topic is  being addressed  and that  NALC
participants  are  researching  cost efficient methods  for implementing
QA/QC  procedures  in  all aspects of the Pathfinder.  One major  success
criteria, the  attainment  of  the  85%/85%  goal  (85% correct/85% of  the
time)  for  categorization  accuracy  can only  be  achieved  if innovative
statistical  spatial sampling  procedures  are  implemented.   Other  QA/QC
activities,  while challenging,  are somewhat more straight forward  (e.g.
cloud  cover  estimation,  RMS,  geometric registration  verification.)
     The purpose  of the  quality assessment  (particularly the  accuracy
assessment) effort  is uncertain,  both  as  to  its  intended uses  and  its
motivation.  Since such efforts can be very costly and time consuming, the
effort should be designed to  provide more  than a "figure  of merit"  such  as
class-by-class  categorization  accuracy.   This  depends  on  whether,  for
example,  its  motivation  is  to  improve area proportion  estimates  or  to
provide "hard"  data  for improved  spatial mapping  using  the  "soft" remote
sensing data (e.g.,  spatial  statistics or geostatistical uses.)   The latter
motivation would imply a design  which avoids  introducing too much bias
into the accuracy  assessment  sample  through  use  of multistage  sampling
techniques.  To  the  degree  this is done,  and  the assessment data  are
preserved, the continued re-use  of these  data  by the  broader  scientific
community will be assured.
     The panel believes that a good start has  been  made, but more still
needs  to be  done.   The  choice  of 85%/85%  seems  to be  arbitrary  and the
consequences  of failing  to  meet  it should be developed from a  science
perspective.  Examination of spatial  sampling that  have been  accomplished
in  projects such as those done by the United States Geological Survey, and
the University of California Santa  Barbara,  need to  be examined.   Several
panel  members feel  that the  project  should   look   at  a  stratified,
systematic,  unaligned  sampling  strategy with double sampling to,  as best
as  practical,  cover  under-represented categories.   In  addition  to  the
procedures discussed, the panel believes that Pathfinder project personnel
should consider the use of  independent analyst interpreter verification  of
the accuracy of sample locations  directly from the  MSS imagery employed
in  the  categorization process.  Done with  proper  care and  diligence, this
                                 99
                                                                     13

-------
should  not introduce  significant  bias  into  the  verification process,  and
studies  have shown it to be an effective and efficient method of accuracy
verification.
      Use  of qualified  interpreters  analyzing  the  MSS  data  does  not
necessarily rule out the use of local expertise but can make such use more
effective and cost efficient  by focusing  attention  on categories that  are
more difficult to  verify.
      The panel  questions whether the data form categories  proposed  are
realistic using   photointerpretation  when   land  cover  goals  are more
modest. The use of photointerpretation as a  basis for accuracy assessment
may  be  fraught with  logistical  difficulties.    In  addition,  the  goals
described in  the  intended interpretation data form may  be  unrealistic, e.g.,
species  types.    The  procedure  is  probably  sufficient  for the  limited
purposes of  the  categorization scheme.
      The panel  recommends that  efforts continue on the  development of
cost-effective,    large-area   categorization   verification   sampling
procedures.  It also  recommends that the  panel, or some  other independent
review board,  review  the procedures developed  prior to  implementation.

8.0  Global  Change  Science and Methodology
      To achieve a  balance  between meeting global  change science goals
and  meeting  the technical  goals  of NALC, the  program must strike  a
reasonable  balance between the  diverse needs  of  a  changing  scientific
customer base and the desire  to standardize procedures.   Land cover type
and  change in  land  cover,  features  o.f the landscape  resolvable  with
Landsat MSS, have  been established as the  data needs by  NALC's primary
customers  -  the  studies  being  conducted  at  the Corvallis and  Athens EPA
labs  on carbon cycle questions.  The NALC team  should continue to work
with  these  two  EPA  labs  to arrive  at  an appropriate  and  translatable
scheme for categorizing, the  data  set into  land  cover type classes.  This
cooperation  is the  basis  for the  scientific value of tracing a  20-year
change  in land cover for North America.   Once it has addressed  the needs
of its primary  customers,  the data  set  developed by the NALC will be
suitable  for  applications pilot projects  throughout  North America.  The
pilots  projects  will help  maintain  the  balance  between  science   and
technical   objectives  by  evaluating  the   trade-offs   between  desired
information and its dependence on  particular  techniques.
                                100                                  14

-------
NALC Review
Panel Report 2/93

9.0 Measures  of  Success
      Listed  below  are  four criteria  for  success that  the  review  panel
members believe represent the goals and objectives of the NALC  program:
      1. Provide data, currently  unavailable, to drive  existing  EPA land
           processes  models, which,  in turn, will help formulate  agency
           policy decisions.
      2. Compilation of a well designed  standardized  set  of data  products,
           including  wall-to-wall triplet  raw MSS  and ancillary  data,
           land cover images and statistics, tabular  data, and evaluation
           of land cover change.  This data set is designed for, and must
           be made readily available  (as evaluated by physical  access and
           cost) to the science  and application  user  community. As such,
           it will have the  very-real  potential  of  providing return on
           investment will beyond  anything currently envisioned.   The
           degree to which  the  data base is acquired  and used  in
           meaningful  scientific investigations  will  measure  how well
           this  is accomplished.
      3. Scientific  and technical  publications in the peer reviewed
           literature.
      4. Concrete examples  of how high  spatial  resolution (MSS) data
           improve our understanding  of land cover change  and  its effects
           on global change phenomena versus other  approaches,  e.g. point
           data, traditional  mapped data, and  coarse resolution data.

10.0  Conclusions  and  Recommendations
      The NALC is truly  a "pathfinder." It is developing a  data  base and an
operational use  for remotely sensed data that  will  meet  the needs .of EPA
and that will  serve  as a model for other -global change research efforts.
The review  panel  concurs with  the goals, objectives  and approach of the
program,  and  presents  the  following  recommendations to  improve  the
efficiency of  the data  processing and  the robustness  of the  results.

10.1  General  Considerations
      •  The  direct links  between  specific science  questions on carbon
cycles  and   the  information  needed  to solve  them, or  improve  EPA's
understanding with the  NALC  technical  approach and  goals, need  more
                                 101
1 5

-------
clarity  and definition.
     •  The  purpose of some  activities, e.g.,  accuracy assessment,   is
unclear and difficult to  tie  to  specific science goals.
     •  The  scientific  arguments  as  to  why  the specific  land  cover
categorization   scheme  will  or  will  not  satisfy  important  gaps   in
understanding  and predicting carbon  cycle questions  are not  sufficiently
made and  need better definition.
     *  The  relationship between the  schedule for data delivery  and the
applications  projects  with  major  scientific  uncertainties  suffers  from
the lack of a clear scientific rationale.
     •  NALC's efforts are  a  key  driving  force  in the  overall  Landsat
Pathfinder  program.   As such, the potential  customer or user  base is likely
much larger even  within  EPA  than the current  EPA base.   The  NALC
management clearly  recognizes this and has  been motivated  to remain
sensitive  to   potential  changing  information  needs  by  seeking   to
standardize products,  to  retain useful  analyses,  such  as clustering,  at
levels  that permit reinterpretations, and by seeking  to  retain  accuracy
assessment data together with the Landsat  data, all archived  at  the  EROS
Data Center.  The team is attempting to define  an  information  management
systems for  the NALC data base  that  will  allow  access to the greater
scientific community.    EPA should  proceed with  caution,  once  the  NALC
technical procedures are set,  in deviating  significantly from  their path  if
needed  or  requested  by other science users.  Different  users  may want  to
alter the priorities on scheduling.   This  should be weighed,  with caution,
against  the need  to  complete the wall-to-wall,  3-epoch data  set and the
needs of NALC's original EPA Global  Change customer base.   EPA could
avail  itself,  as  NALC  has  already  begun   to do,  of  collaborative
opportunities for "application  projects"  so as to  leverage  their  significant
NALC effort.
     ;  NASA,  and  the  other   agencies participating  in  the  Pathfinder
Program, should develop  mechanisms to assimilate the  NALC-Pathfinder
experience, and that of other Pathfinder projects,  in the  development  of
EOS-DIS.
     •  One  area, where the  input of  the EPA  science  community  is
especially desirable at  present is in  the  sequence  of areas to be analyzed
by  the  project.  It  would appear that  by  modifying the  order in  which
different regions  are  analyzed, a  more  timely   response   to  scientific
priorities could  be  achieved.    For example an  earlier  analysis of the
                                 102
1 6

-------
NALC Review
Panel Report 2/93

Pacific Northwest,  than that currently planned,  would seem to  be needed.
      • The NALC should specify criteria  for success so that the progress
and accomplishments  of the project are readily apparent.

10.2 Technical Recommendations
•  Categorization:
      1. Measures of success should be established for testing
           categorization techniques,  e.g.  ability  to recognize categories
           of interest,  ease of labeling,  etc.
      2. Categorization tests should be conducted using triplets from
           representative  environments.
      3. Alternative  categorization techniques should be tested  before
           a categorization methodology  is chosen  for the duration  of
           the program.
•  Labeling  of categorized pixels is  crucial  to the success  of the program.
      The  individuals performing the labeling must be  trained adequately
      in  multispectral categorization  techniques.
•  Immediate  implementation of an  IMS  will greatly improve the efficiency
      of the program and its chances to  meet all its objectives.
•  Increasing  throughput of compilation   of  the  triplets  would  allow  for
      pilot  testing  and technique evaluation sooner and in  a broader range
      of  environments.
•  Efforts  should be taken immediately to  assure all  required MSS scenes
      were collected and preprocessed, or that a  mechanism exists to
      complete the  preprocessing.
•  Compilation  of  land  cover  and  land  cover change standard  products
      should  be  monitored and  analyzed with  the  cooperation  of  the
      appropriate  land  management organizations
•  Field  inventories should  include  the consistent and proper use  of field
      forms and data collection and management  techniques.
•  A sampling  frame  approach  for cluster  labeling  activities  should  be
      developed and  implemented.
•  Change detection  methodology:
      1.  Evaluation  of  alternative  change  detection  methods  should  be
           performed.  The cooperation of the appropriate elements of the
           science  community should be  sought  for that  effort.
                                 103
1 7

-------
    2.  Extensive testing  of change  detection techniques  should  be
         performed at each of  the  pilot sites.
    3.  Explicit  procedures  for  change  detection  given the  temporal
         variation in the triplets should be developed and evaluated.
Membership  in  the  NALC-Pathfinder Technical Working  Group  should
    include  the  science  community within  EPA  and  other  research
    programs directed toward global  change.
QA/QC  needs to be fully developed and evaluated. A rationale for the
    selection  of  the "85/85" criterion  should  be  developed based on
    science   requirements.  Evaluation  of  other  QA/QC  should be
    attempted.
Efforts  should  continue  toward  the  development  of  cost-effective
    categorization  verification  sampling procedures.  The  procedures
    must be  reviewed prior  to  implementation.
                             104
                                                                 1 8

-------
NALC Review
Panel Report 2/93
                       Panel  Review Members
Dr. Chuck Dull
Dr. John Estes
Dr. Leonard Gaydos
Dr. James G. Lawless
(Chair)

Mr. Douglass Muchoney
Mr. David  Peterson
Mr. Edwin Sheffner
Dr. David Skole
Dr. John Townshend
US Forest Service
Washington,  D.C.

Dept. of Geography
University of California, Santa  Barbara

USGS, Ames Research Center
Moffett  Field,  California

NASA Headquarters
Washington,  D.C.

The  Nature Conservancy
Arlington, Virginia

NASA Ames  Research Center
Moffett  Field,  California

TGS Technology
Ames Research Center

CSRC/EOS
University of New Hampshire, Durham, N.H.

Dept. of Geography
University of Maryland,  College Park,  MD.
                               105
                                                                    1 9

-------
        NORTH AMERICAN LANDSCAPE CHARACTERIZATION (NALC) PROJECT

             RESPONSE TO THE NALC PANEL REVIEW REPORT
  The organization of this response follows the format provided in the NALC Panel
Review report.  The order is patterned after the format of the NALC Technical Work
Plan and order of presentations at the NALC Technical Review session in  New
Orleans.

  The  responses to comments by the Technical Review Panel specifically address
those  areas  identified as requiring  additional attention.   The objectives of the
responses are to identify how we intend to act upon the comments.
3.0 IMAGE CATEGORIZATION AND DEVELOPMENT OF LAND COVER PRODUCTS

  In response to comments in this section of the Panel Report, we intend to develop
measures of success, establish them programmatically, and characterize the methods
to evaluate them as to their appropriateness. These issues include addressing the
separation  of classes, the border class problem, and methods to ease the labor
involved in labeling.

  It is desirable  to devote further attention to  the qualifications  of associated-
researchers as related to capabilities in  computer categorization and labelling of
classes resulting from analysis  of Landsat data.  This  will help assure their
performance and adherence to standard procedures.   We  will  assist  them, by
developing routines which supply spectral plots and  which will highlight individual
classes to further assure high quality work in labeling land cover classes.  Attention
will be devoted to better describing the classification system classes, including closed
forest and savanna  or shrub/scrub classes.  We  will address and clarify methods to
handle the  question of land cover condition or type following harvest.

  Further efforts  will examine the use of ancillary data along with NALC products,
such as AVHRR time series, TM, SPOT and/or digital orthophoto data.  These efforts
will be conducted directly to insure their characterization early on.

  These efforts will also entail further review of literature as to' previous results, levels
of accuracy achieved,  and  further documentation on background  of the  criteria
selected for the NALC Project.
                                    106

-------
4.0 METHODS FOR MANAGING AND DISTRIBUTION OF NALC DATA

  In 1992 over five hundred and sixty MSS scenes were acquired by EOSAT and pre-
processed in support of the project.  These scenes along with those in the archive are
being identified and scheduled for triplicate production. To augment this source, EPA
and USGS are in contact with the Canada Centre for Remote Sensing. The goal is to
determine the extent of 1990/91/92 MSS  data that have been acquired from
Canadian sources, and that can be used for NALC.   We will also determine their
capabilities to pre-process  EOSAT  HDT-R  products using their HDT-A processing
stream.

  As part of these efforts we will evaluate options to increase the throughput of MSS
scenes through the processing procedures by the EROS Data Center.

  The NALC Information Management System (IMS) is currently being implemented.
The characteristics  will  be evaluated with a perspective of serving the needs of
production level data and metadata management, and later requirements important to
public access and browsing. Work has been conducted to learn from the experience
of the  IMS of Landsat Pathfinder on the Humid Tropics (HTRIP), and see how this
Arc/Info based approach may be of assistance.

  Data processing standards  will  be  further coordinated  with  interested  federal
agencies.  This will take the form of further consultations, and inclusion of additional
personnel in the Technical Work Group.

  Field and aerial photo sampling techniques and data record sheet will be  further
developed.   This will  be accomplished  with an  eye toward  assisting  in  the
field/laboratory labeling of categorized images, and towards the capture of a variety
of land cover conditions during field studies.
5.0 CHANGE DETECTION METHODOLOGY

  The approach to detection of change in land covers will be refined following the
comments provided by the Panel.  This effort will include evaluating other change
detection approaches, such as principal components analysis, regression-Chi square
analysis, image difference methods, and post-classification change detection methods.
The  methods and approach will be evaluated on a variety of sites representing the
variability in land covers to be encountered during the Project. The determination of
methods to be used will be made with a focus on the final products and science
issues related to detection of change.
                                   107

-------
  In particular, we will focus on relating change detection methods to identifying land
cover changes important to the science objectives, and insure  there is a fully
reconciled land cover classification system for change detection. This effort will be
conducted in cooperation with the science clients within the Agency and others. This
effort will involve additional, exhaustive evaluations of available methods. This work
will be conducted with the assistance of researchers actively involved in applications.
The experience of governmental, industrial and academic personnel will be tapped.
  The issues of image pre-processing with regard to systematic corrections and other
radiometric corrections will be re-visited. Issues will include atmospheric corrections,
sun angle corrections, and the long temporal windows encountered in analysis of
some scenes.
6.0 NALC MANAGEMENT, SCHEDULING AND COORDINATION

  Further cooperation will be fostered with Agency science clients and with other land
management agencies that have an opportunity to respond to the results through
modification in management practices.  Cooperation with the USGS,  DOI agencies,
USDA agencies, TVA and others is envisioned. It is anticipated that vehicle for such
cooperation will be meetings with interested parties and membership on the Technical
Work Group.

  An additional goal of this cooperation will be to identify existing databases held by
these groups. Use of such ancillary data may be of great value to the  work  of NALC,
and its scientific clients.

  It is also necessary to further develop the links with the Agency science clients.  It
is anticipated that attendance of more clients at the Technical Work Group  sessions
would assist this process.  Further visits to Agency Laboratories, as  a follow on to
previous visits, will also help to further appraise clients  of opportunities, project
progress, and allow feedback on requirements related to scientific objectives.

  Stronger links will  be forged with the image processing and image  application
communities.  The Project can use their input to assure high quality results and assure
acceptable standard methods for data processing. This effort is on-going  with the
advent of  talks on the Project presented  to university audiences, non-profit and
industrial groups, and additional federal audiences. Linkages have been developed to
date with Agency scientist including EMAP personnel, USGS Center for Inter-American
Mineral Resource Investigations (CIMRI) scientists, USGS Water Division scientists,
USFWS Gap Project personnel and others.
                                    108

-------
7.0 QUALITY ASSURANCE AND QUALITY CONTROL (QA/QC)

  The need for quality assurance procedures and quality control  mechanisms is
mandated by Agency policy.  It is an important aspect of the Project due to the high
level of interest in accuracy assessments, and methods are also an important result
of the work.

  The QA/QC plan is being further developed to track the production and document
the quality of triplicate products. The quality goals have established based on Agency
policies and the work is on-going. The procedures are currently being reviewed to
determine their adequacy.
  The determination as the categorization accuracy is an important Agency QA/QC
goal.  It is also important to the user community and to the Technical Review Panel.
We are currently developing the protocol to be applied.  Recommendations from the
Panel such as reconciliation of science objectives and accuracy assessment objectives
are being implemented. We are also re-evaluating the categorization accuracy quality
objectives to further ground them in scientific experience.

  The land categorization methods and survey sampling methods to accomplish the
accuracy assessment are undergoing  further study and exhaustive documentation.
This effort will be thoroughly grounded in theory and  literature, and  draw upon
experience in the government and in other research arenas.  The approaches  of
stratified  systematic sampling, and  double  sampling  will be  further evaluated.
Methods for systematic sampling in general require further attention, as does the level,
of effort devoted to each  MSS scene, and the use of a combination  of photo
interpretation and ground sampling  methods.

  Further  attention  will be devoted to developing photo interpretation sheets for
accuracy assessments. The issue of  independent photo interpretive assessment of
accuracy will be addressed.  Photo interpretive keys will be developed to guide the
interpreters in their assessments. The  photo interpretative keys will also be optimized
recognize the difference in detail between photos and the land cover categories.

  The Panel will have the opportunity to review these  procedures during the  next
Technical  Work Group meeting in May 1993 and in the next draft of the Technical
Work Plan appearing at about the same time.
8.0 GLOBAL CHANGE SCIENCE AND METHODOLOGY

  In additional to the comments addressed above, the Project will develop procedures
to further communicate the elements of the work and further the dialogue between
the user community and NALC personnel.
                                    109

-------
          9.0 MEASURES OF SUCCESS

            The measures of success articulated in the Technical Panel Report are particularly
          clear, and germane to the effort. These measures include: NALC products and results
          will augment EPA inventory and modeling efforts; the standard NALC products will be
          a valuable resource  for the general scientific community;  the  characteristics of
          products and result of NALC will be disseminated in the peer reviewed literature;  and
          the results  will  demonstrate the value of area-based  sampling as  compared to
          traditional sampling approaches alone.
          10.0 CONCLUSIONS AND RECOMMENDATIONS

            As addressed previously, the NALC Project will continue to refine the link between
          the specific scientific questions that drive the Project and the technical approaches
          and goals.
                                             110
.

-------
          Appendix   II:



Agreement Between INEGI and USEPA
              111

-------
          UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

                  OFFICE OF RESEARCH AND DEVELOPMENT
           ENVIRONMENTAL MONITORING SYSTEMS LABORATORY-LAS VEGAS
                          P.O. BOX 93478
                     LAS VEGAS. NEVADA 89193-3478
                      (702/798-2100 - FTS 545-2100)
                          MAR 24 I393
Nestor Duch Gary
Director de Geographia
Institute Nacional de Estadistica, Geografia e  Inforiaatica
Heroe de Nacozari No. 2301
Puerto 7 Anexo
C.P. 20290 CD, Industrial
Aguascalientes, Mexico

Dear Mr. Duch:

     The Office of Research and Development of  the  U.S.
Environmental Protection Agency has undertaken  a  project  known as
the North American Landscape Characterization  (NALC)  as part of
its Global Change Research Program.  The  objective  of this
project is to produce land cover  and land cover change datasets
for use in the inventory of terrestrial carbon  stocks and for
estimating the flux  of  carbon which has occurred  due  to land
cover change.  Project  areas include Central America,  Mexico,
Caribbean Islands, and  the U.S.A.

     The NALC land cover and land cover change  data sets  are to
be derived from Landsat MSS "triplicates" which are currently
being assembled for  the project areas  by  the U.S. Geological
Survey, EROS Data Center  (EDC).   The triplicates  consist  of three
dates of georeferenced  Landsat MSS data,  with  scenes  from the
early 1970's, mid-1980's, and early 1990's.  Where  available we
are also adding digital terrain model  (DTM) data  and  digital
ecoregion or biotic  community data sets.   To cover  all of Mexico
will require the assembly and analysis of 99 of these
triplicates.

     Our primary collaborator  in  the U.S.A.  is the  EROS Data
Center.  This is a very natural collaboration  for several
reasons.  EDC is an  operational arm of the USGS National  Mapping
Division, which has  responsibility for coordinating national
mapping projects and producing national map products.  In
addition, the federal archive  of  Landsat  MSS data resides with
EDC.

     The Data Center is assembling the NALC triplicates in
support of the NALC  project.   The topographic  maps  and digital
terrain datasets for the  USA  triplicates  are being provided by
EDC as one of their  contributions to the  "project.  EDC will
establish  an  open  archive for  the NALC datasets,  which will be
available  to  the wider scientific community.   While the specific
                              112

-------
details on distribution media and costs are still to be worked
out, the intent is that the NALC datasets will be distributed by
EDC at the marginal cost of filling individual orders.

     Our relationship with the EROS Data Center is very
productive and represents a long term commitment of EPA and USGS
to work together in producing high quality land cover and land
cover change datasets.  We would like to establish a similar
relationship with the Institute Nacional de Estadistica Geografia
e Informatica (INEGI).  This is in recognition of the special
role INEGI plays in producing and distributing geographic
datasets for Mexico.  As the Technical Director of the NALC
project, I am able to commit .to the following terms of agreement,
which can be the basis for am expanding collaboration between the
EPA's Environmental Monitoring Systems Laboratory - Las Vegas
(EMSL-LV) and INEGI.  All data distribution commitments by EMSL-
LV are subject to the availability of NALC project funds.


DATA EXCHANGE:

1)  EMSL-LV will provide a complete set of the georeferenced
Landsat MSS triplicates and ancillary data sets to INEGI in
digital form.

2)  INEGI will provide the topographic maps and digital terrain
model data sets required for the assembly of the Landsat MSS
triplicates for Mexico.

3)  The objective of  the NALC is to derive land cover and land
cover change datasets for large portions of North America.  EPA
will invite INEGI participation in the analysis of the Mexican
data sets.  Particular attention will be paid to producing output
products which are compatible with existing INEGI land cover maps
for Mexico.

4)  INEGI is free to  use the NALC data sets they receive for
other projects.


DATA DISTRIBUTION:

1)  No copyright or trade secret restrictions will constrict the
distribution of the NALC Landsat data sets.  Any holder of those
data sets may utilize, copy, and distribute those data sets.  At
INEGI's request, restrictions will be placed on the distribution
of INEGI provided DTM data sets of Mexico.

2)  The central archive and distribution center for the NALC
project  is the USGS EROS Data Center.  In determining a price for
the data sets, EDC plans to simply recover the costs associated
with filling a specific request.

                               113

-------
3)  EMSL-LV invites INEGI to become an archive and distribution
center for NALC data sets of Mexico.  INEGI is free to set their
own pricing and distribution system.
TERMS FOR USE -OF THE DTM DATA:

     The DTM data will be coregistered to the Landsat MSS scenes
and these combined data sets will be archived and distributed
from EROS Data Center.  Based on our agreement with EDC terms
set by INEGI for the distribution of the Mexican DTM data sets
will be respected.  In working with EDC on this issue it would be
possible to restrict access of the Mexican DTM data sets to non-
commercial applications.  This would involve the submission of
signed statements from any organization requesting the data sets,
attesting to their non-commercial status and non-commercial
applications for the DTM data of Mexico.
                            Sincerely,

                                   s.
                            Ross S. Lunetta
                            NALC - Technical Director
                                114

-------
Institute  National  de Estadlstica,
      Geograiia e Informaiica
DIRECCION GENERAL DE GEOGRAEEA


A.93.72.082


April 5, 1993
       Mr. Ross S. Lunette
       Technical Director - NALC
       U.S. Environmental Protection Agency
       En.vironmCT.tal Monitoring Systems Laboratory
       944 E. Harmon Street
       Las Vegas, Nevada 89193
       ILS-A.
       Dear Mr, Lunetta,
              Thank you for your letter of March. 23,1993 regarding the EPA-USGS North American
       Landscape  Characterization (NALQ, which is a- component of the- ELS. Global Change
       Research. Program.  In your letter  yon have proposed an exchange of data, with INEGI
       providing map and digital elevation data for nse in producing" the basic NALC data sets for
       Mexico.  To. addition, yon have invited the participation of INEGI in the conduct of the NALC
       project in Mexico.

              In recognition of the scientific  merit of the NALC project and its close match to
       several of INEGTs technical objectives for the 1990's, I am pleased to inform you that INEGI
       agrees to the data, exchange outlined in your letter.
              Specifically, INEGI agrees to provide:
        1) The digital terrain models (DTMs) for an of Mexico. Please note mat these data sets are
              being provided Sor non-commercial use only, under terms-described later in this letter
              agreement.
                                              115

-------
2) Topographic maps at 1:50,000 and 1:250,000 scales for all of Mexico.
3) Land use / land cover maps of Mexico at 1:250,000 scale.
4) Selected aerial photographic prints of Mexico from existing archives held by -INEGL
       Tlie data sets -to be provided to INEGI by EPA include:

1) The NALC Landsat MSS data sets of Mexico for three time periods (1970's, 1980's, and
       1990*s) on the form of georefererieed and coregistered triplicates.

2) The DTM data for the Mexican MSS triplicates, resampled to the MSS spatial resolution
       and clipped to the MSS scene coverages.

3) The digital land cover and land cover change data sets produced through the analysis of
       the NALC triplicates of Mexico.
       Regarding the DTM data, 3NEGI is providing these data sets for use in the scientific
research to "be conducted with the NALC data sets.  In order to preserve the  commercial
markets that INEGI lias established for these data sets, we request that the EPA and USGS
restrict the access of these  data sets to government entities, universities.,  and non-profit
organization.  We request that the users of the DTM data sign agreements binding- their usep
of the data to noa-K^rnrnercial projects and preventing 'the unauthorized dispersion of the data.
Attached for  your consideration are  draft copies of the style of apjph'cafion and agreement
which we would like to have executed by each group receiving the Mexican DTM data.  It
is  our understanding  that the USGS EROS  Data Center has agreed to administer this
 application and agreement system  and that INEGI  win be provided with  copies of 1he
 completed forms on a periodic basis.
        INEGI frirenffe to maintain art archive of the NALC data sets, which wJH be used for
 our internal projects. We expect to also establish a system for distributing these data sets to
 government entities and scientific investigators in Mexico, in response to data requests.  We
 recognize that rnese data sets will also be available from the USGS EROS Data Center.

        INEGI is the principal agency in Mexico responsible for the production, maintenance,
 and distribution of cartographic and spatial data sets. As such, we are pleased to accept your
 invitation for INEGI participation in the NALC project for Mexico.  I believe that INEGTs
 participation win mate a valuable contribution to the NALC project  Please top us informed
 of the meetings, reviews, and other activities for the NALC project so that our staff has an
            to attend..
                                         116

-------
      At the present time our staff has commenced the assembly of the topographic maps,
land use - land cover maps, and DTM data sets you have requested for the NALC project
We look forward to collaborating with. EPA on the NALC project and on other research
projects in the coming years.
                                          Sincerely,
                                         DIRECTOR
                                      LIC. NESTOR DUCH GARY
                                      117

-------
Bear Investigator,


       The  data  you  have requested of Mexico from the  North American Landscape
Characterization (NALQ  project contains digital elevation data which  was provided by
Mexico's National Infinite for Statistics, Geography, and Informatics (INEGI)*  These data
sets have been provided by INEGI for use in scientific investigations and projects conducted
by government agencies, universities, and non-profit organizations.  These data, provided at
the marginal costs of distribution, are not to be used for commercial applications.  These data
sets can be purchased from INEGI by any group wishing to use them in a commercial project-

       In order to preserve the commercial markets INEGI has developed for these data sets,
we request that you complete the foUowing application and  agreement.  Your cooperation in
abiding by the terras outlined in the agreement win ensure the continued availability of these
data, sets for scientific research.

       The Landsat MSS  data sets that yon have requested of Mexico  are not restricted in
their Tise or distribution-

       Please  send th& completed application and agreement forms to the following address:

                    Customer Services
                    USGS EROS Data Center
                    Mundt Federal Building
                    Sioux Falls, South Dakota 57198
                    U.S-A-

                    TeL  605-594-6961

                    No facsimiles of the forms will be accepted.
To obtain the Mexican. DIM data for use in a commercial project, please contact INEGI at
the following address:


                    Servicio a Usarios
                    1NEGEI, Direccion de Integracion y Analisis
                    Subdireccion de Comercializacion
                    Heroe de Nacozari 2301 Pueita 11
                    CoL del Parqne OP. 20290
                    Aguascalientes, AGS
                    MEXICO

                    TeL 52-49-181948
                    FAX 52-49-180739
                                         118

-------
APPLICATION FOR USE OF THE DIGITAL TEKRAM DATA OF MEXICO




Applicant Name (Last, First, MJ):



Title:



Organization:



Division:



Mailing Address:
 TeO.epi.one:
 Fax:
 Electronic Mail:
 Title of Reseatca:


 Purpose of Usage:
                                        119

-------
AGREEMENT FOR USE OF THE DIGITAL TERRAIN DATA OF MEXICO

The digital terrain data you have requested of Mexico has been provided by the Mexico's
National Institute for Statistics, Geography., and Informatics
(INEGI) for use in non-commercial scientific research.

The researcher agrees to the following conditions for use of the Mexican digital terrain
1) The data -will not be used for commercial purposes.

2) The source of the data (ENEG3) mQ. be acknowledged in any derived data-product, reports,
       or publications.

3) .The data win not be distributed to third parties. Investigators affiliated with, your project
       who •wish to obtain the digital t^tr-miy data must complete their own application and
       agreement forms-

4) The undersigned agree to take appropriate measures to restrict the transfer of the data to
       commercial entities.
RESEARCHER

Name
ORGANIZATION

Name of Superintendent
 Signature / Date


 Address
Signature / Date


Address
                                       120

-------
                     United States Department of the interior
                                    GEOLOGICAL SURVEY
                                      EROS Data Center
                                 Sioux Falls, South Dakota 57198
INR£PLYR£FERTO:  OC  6-4
June 8,  1993
       Mr.  Nestor Duch  Gary
       Director  d Geographia
       Institute Nacional  de  Estadistica,.Geografia e Informatica
       Heroe  de  Nacozari No.  2301
       Puerto 7  Anexo
       C.P. 20290 CD,  Industrial
       Aguascalientes,  Mexico

       Dear Mr.  Duch:

       We  are pleased  to learn  of your participation in the North American Landscape
       Characterization (NALC)  Program and your contribution of several data sets,
       including digital terrain models  (DTM's) for all of Mexico.  We understand
       that Mexico's National  Institute  for Statistics, Geography, and Informatics
       (INEGI) is providing DTM's for use in the scientific research to be conducted
       with the  NALC data  sets.  Per your request, the Environmental Protection
       Agency and the  U.S. Geological Survey (USGS) will restrict the access of these
       data sets to government  entities, universities, and non-profit organizations.
       This will be administered by the  USGS EROS Data Center as follows:

         —  A user requesting  Mexican DTM data will be sent three items (enclosed):

             •  a letter explaining data use restrictions,

             •  an application  for use of the data, and

             •  an agreement  restricting use and distribution of the data.

         —  Upon receipt  of  both completed forms and proper payment, the Data Center
             will fill  the order.

         —  The Data  Center  will provide copies of completed application and
             agreement forms  to INEGI on a quarterly basis.
                                          121

-------
Mr. Nestor Ouch Gary                   .       ,                              2
We hope that this procedure meets'with  your approval.  We  look  forward to
collaborating with INEGI on the NALC project and  on other  research projects in
the coming years.
Enclosures
cc: R. Lunetta, EPA
                                    Sincerely,
                                                  /
                                 /AjDonald T.  Lauer
                                    Chief, EROS Data Center
                                    122

-------
Dear Investigator,

The data you have requested of Mexico from the North American Landscape
Characterization  (NALC) project contains digital elevation data which was-
provided by Mexico's National Institute for Statistics, Geography, and
Informatics (INEGI).  These data sets have been provided by INEGI for use
scientific investigations and projects conducted by government agencies,
universities,  and non-profit organizations.  These data, provided at the
marginal costs of distribution, are not to be used for commercial
applications.   These data sets can be purchased from INEGI -by any group
wishing to use them in a commercial project.
In order to preserve the commercial markets INEGI has developed for these data
The Landsat MSS data sets that you have requested of Mexico are not restricted
in their use or distribution.

Please send the completed application and agreement forms to the following
address:

            Customer Services
            USGS EROS Data Center
            Mundt Federal Building
            Sioux Falls, South Dakota  57198
            USA

            Tel. 605-594-6151

            No facsimiles of the forms will be accepted.
To obtain the Mexican DTM data for use
INEGI at the following address:
in a commercial  project,  please contact
            Servicio a Usarios
            INEGI, Direccion de Integracion y Analisis
            Subdireccion de Comercializacion
            Heroe de Nacozari 2301 Puerta 11
            Col. del Parque C.P. 20290
            Aguascalientes, AGS
            MEXICO

            Tel. 52-49-181948
            Fax  52-49-180739

                                    123

-------
APPLICATION FOR USE OF THE DIGITAL TERRAIN DATA OF MEXICO


Applicant Name (Last, First, M.I.):


Title:


Organization:


Division:


Mailing Address:
Telephone:
Fax:
Electronic Mail:
Ti tl e of "Research:
Purpose of Usage:
Latitude and Longitude Coverage Limits of Data Requested:
                                    124

-------
AGREEMENT FOR USE OF THE DIGITAL TERRAIN DATA OF MEXICO

The digital terrain data you have requested of Mexico has been provided by
Mexico's National Institute for Statistics, Geography and Informatics (INEGI)
for use in non-commercial scientific research.

The researcher agrees to the following conditions for use of the Mexican
digital terrain data:

1)  The data will not be used for commercial purposes.

2)  The sources of the data (INEGI) will be acknowledged in any derived data
    product, reports, or publications.

3)  The data will not be distributed to third parties.  Investigators
    affiliated with your project who wish to obtain the digital terrain data
    must complete their own application and agreement forms.

4)  The undersigned agree to take appropriate measures to restrict the
    transfer of the data to commercial entities.                        .  v
RESEARCHER

Name
ORGANIZATION

Name of Superintendent
Signature/Date


Address
Signature/Date


Address
                                  125

-------
          UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

           cwv/m^K,  OFFICE OF RESEARCH AND DEVELOPMENT
           ENVIRONMENTAL MONITORING SYSTEMS LABORATORY-LAS VEGAS
                          P.O. BOX 93478
                     LAS VEGAS. NEVADA 89193-3478
                      (702/798-2100 - FTS 545-21 OO)
                         JUN I 6 I993
Mr. Nestor Duch Gary
Director General de Geografia
Institute Nacional Estadistica, Geografia e Informatica
Heroe de Nacozari N. 2301
Puerto 7 Anexo
C.P. 20290 CD, Industrial
Aguascalientes, ACS.  Mexico

Dear Mr. Duch:

     Thank you for your  letter  of April  5, 1993  regarding the
data exchange agreement  between the  Institute  Nacional
Estadistica, Geografia e Informatica (INEGI) and the
Environmental Monitoring Systems Laboratory -  Las Vegas  (EMSL-
LV).  The data exchange  is being conducted in  support of the
North American Landscape Characterization  (NALC)  project,  which
is  a component of the U.S. Global Change Research Program.

     In accordance with  our data exchange agreement, enclosed
please find  19 NALC triplicate  data  sets for Mexico.  We will
continue to  ship triplicate data sets,  in increments of  12 or
greater, until all 102 path/rows for Mexico are  provided to
INEGI.

     Also in accordance  with your reguest, the Mexico Digital
Terrain Data provided by INEGI  will  be  restricted to
noncommercial users of NALC triplicate  data sets.  The U.S.
Geological Survey  (USGS),  EROS  Data  Center  (EDC)  has provided you
with correspondence and  a  copy  of the forms "Application or
Agreement for the use of the Digital Terrain  Data of Mexico", to
comply with  your distribution restrictions.

     To meet the current NALC triplicate data  set assembly
schedule we  would appreciate receipt of Digital  Terrain Data for
the following coordinates  by July  12, 1993.

                  N 27°, W 99° to N 23°,  W 96°
                  N 33°, W 117° to  N 16°, W 99°

We request that the digital  terrain data for  the remaining
locations  in Mexico be  provided by August  13,  1993.
                             126

-------
      I have also enclosed a letter request from Ms. Susan K.
Jenson of the EDO.  Ms. Jenson would like permission from INEGI
to produce a 15 arc-second generalized elevation data set for
Mexico.  This data set would complement similar efforts completed
for the United States and Canada.  If you would like additional
details please contact Ms. Jenson directly at  (605)594-6011.

     My current travel plans provide an opportunity for me to
visit Aguascalientes on September 30, 1993.  If this is a good
tir.c for you, I would appreciate the opportunity to again visit
INEGI.  During the visit I could brief you and your staff on the
status of the NALC project.  Also, I am interested in the
progress of your facility upgrades and potential utilization of
NALC data sets.

      I would like to invite you or member(s) of your staff to
attend the next "NALC Technical Work Group" at EMSL - Las Vegas,
Nevada on December 8-9, 1993.  We could also use the opportunity
to tour our facility and brief you on the environmental
monitoring research activities at EMSL-LV.

     Thank you for your interest and collaboration on the NALC
project.  I look forward to seeing you on September 30, 1993, in
Aguascalientes.
Enclosures
                               Sincerely,
                                 f^fp&
                               Ross S. Lunetta
                               NALC Technical Director
CC:
Lauer,  USGS/EDC
Sturdevant,  USGS/EDC
Jenson,  USGS/EDC
                             127

-------
                    Appendix  Illi



Path/Row Data Acquisition List, Triplicate Assembly Priorities
                            128

-------
              NALC MSS DATA ACQUISITIONS BY PATH/ROW

{* =PATH/ROWS PURCHASED FROM EOSAT USING EDC BROKERAGE AGREEMENT)
Caribbean
48 scenes.
1/49-50-51-52
2/48
4/47-48
5/47-48
6/47
7/46-47
8/46-47-48
9/45-46-47
10/44-45-46-47
11/43-44-45-46-47-48
12/43 -45-46-47-48
13/42-43-44-45-46
14/41-42 -44-45
15/44-45
16/44-45
17/44-45
            *48
            *47
            *47
            *47,48
            *45,46,47
            *44,45,46,47
            *44,45,46,47
            *43,45,46,48
            *45
            *45
Central America

10/54-55
11/54-55
12/53-54-55
13/54-55
14/53-54
15/50-51-52-53
16/49-50-51-52-53
17/49-50-51-52
18/49-50-51
19/48-49-50-51
20/50
      32 scenes.



            *55

            *53
            *52,53
            *51,52

            *49,50,51
            *50
Chesapeake Bay Watershed
               18 scenes.
14/30-31-32-33-34
15/30-31-32-33-34
16/30-31-32-33-34
17/31-32  -34
             *30,31,32,33,34
             *30,31,32,33,34
             *30,31,32,33,34
             *31,32,34
                                129

-------
Mexico
119 scenes.
18/45-46
19/45-46-47
20/45-46-47-48-49
21/46-47-48-49-50
22/47-48-49
23/47-48-49
24/46-47-48-49
25/45-46-47-48-49
26/42-43-44-45-46-47-48
27/42-43-44-45-46-47-48
28/40-41-42-43-44-45-46-47
29/39-40-41-42-43-44-45-46-47
30/40-41-42-43-44-45-46-47
31/39-40-41-42-43-44
32/38-39-40-41-42-43
33/38-39-40-41-42-43-44
34/38-39-40-41-42-43-44
35/38-39-40-41-42-43
36/38-39-40-41-42
37/38-39-40-41
38/37-38-39-40
39/37-38-39
40/37-38

*46
*45,46,47
*46, 47,48, 50
*47,49
*48,49
*48,49
*46,47,48

*45,46,47
*46,47

*40, 41, 42, 43, 44, 45, 46
*40, 41, 42,43
*39,40,41,42,43
*39,40,41,42

*39
*39,40,41,42
*40,41
*38
*38

Brazilian MSS Acquisitions
                     83 scenes.
214/64-65-66-67
215/64-65-66-67-68-69   -71-72-73-74
216/63-64   -69-70-71-72-73-74-75-76
217/63   -70-71-72-73-74-75-76
218/62-63   -70-71-72-73-74-75-76
219/74-75-76-77
220/74-75-76-77-78-79-80-81
221/71   -74-75-76-77-78-79-80-81-82-83
222/74-75-76-77-78-79-80-81-82-83
223/75-76-77-78-79-80
224/75-76-77

   (* sPATH/ROWS PURCHASED FROM EOSAT USING EDC BROKERAGE AGREEMENT)
                                130

-------

Hawaii

62/46-47
63/46-47
64/45-46
65/45
66/45
     8 scenes.

Alaska

54/21-
55/21-
56/20-
57/19-
58/19-
59/18-
60/19
61/18
62/17-
63/17-
64/16-
65/15-
66/14-
67/13-
68/12-
69/11-
70/11-
71/11-
72/11-
73/11-
74/10-
75/10-
76/10-
77/10-
78/10-
79/10
80/10
81/10
82/10
83/12
     180  scenes.
22
22
21
20-21
20
19-20
18
18
17-
16-
15-
14-
13-
12-
•12-
•12-
•12-
•12-
•11-
•11-
•11-
•11-
-11-
•11-
•11
-11
•11
•13
18
17-18
16-17-18
15-16-17-18
•14-15-16-17-18
-13-14-15-16-17-18-19-20
•13-14-15-16-17-18-19-20
•13-14-15-16-17-18-19-20
•13-14-15-16-17-18-19-20-21
•13-14-15-16-17-18-19-20-21
•12-13-14-15-16-17-18-19-
•12-13-14-15-16-17-18-19-
•12-13-14-15-16-17-18-19-
•12-13-14-15-16-17-18-
•12-13-14-15-16-17-18-
-12-13-14-15-16-17-18-
•12-13-14-15
•12-13-14-15
•12-13-14
•14
       *19
       *16
-21-22
-22
-22
-23
-23
       *12,13,14,15,16,17
   (* =PATH/ROWS PURCHASED FROM EOSAT USING EDC BROKERAGE AGREEMENT)

                                131

-------

Western US
          133 scenes.
33/33-
34/31-
35/31-
36/29-
37/29-
38/27-
39/27-
40/27-
41/26-
42/26-
43/26-
44/26-
45/26-
46/26-
47/26-
48/26-
34-35-
32-33-
32-33-
30-31-
30-31-
28-29-
28-29-
28-29-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28-
27
•36-37
34-35-
34-35-
32-33-
32-33-
30-31-
•30-31-
30-31-
29-30-
29-30-
29-30-
29-30-
29-30-
29-30-
29-30
36-37
36-37
34-35-36-37
34-35-36-37
32-33-34-35-36
32-33-34-35-36
32-33-34-35-36
31-32-33-34-35-36-37
31-32-33-34-35-36
31-32-33-34-35-36
31-32-33-34-35
31-32-33
31-32
  *34
  *31,35,36,37,38
*31,32,33,34,35,36,37,38

  *37
  *30,33,34,35
  *35,36,37
  *27,34,36
  *28,29,30,31,32,33
Eastern and Southern US
                      141  scenes.
06/38
10/29
11/27-
12/27-
13/28-
14/29-
15/29-
16/35-
17/30-
18/31-
19/31-
20/29-
21/28-
22/27-
23/27-
24/27-
25/26-
28-29-  -31
•28-29-30-31
•29-30-31-32
• 35-36
• 35-36-37  -41-42-43
•36-37-38-39-40-41-42
 33-  35-36-37-38-39-40-41
•32-33-34-35-36-37-38-39
•32-33-34-35-36-37-38-39
•30-31-32-33-34-35-36-37-38-39
•29-30-31-32-33-34-35-36-37-38-39-40
•28-29-30-31-32-33-34-35-36-37-38-39-40
•28-29-30-31-32-33-34-35-36-37-38-39-40
•28-29-30-31-32-33-34-35-36-37-38-39
•27-28-29-30-31-32-33-34-35-36-37-38-39-40
 (* =PATH/ROWS PURCHASED FROM EOSAT USING EDC BROKERAGE AGREEMENT)

                                132

-------

Midwest and Great Plains
                       122  scenes
26/27
27/27
28/26
29/26
30/26
31/26
32/26
33/26
34/26
35/26
36/26
37/26
38/26
39/26
40/26
-28-29-30-31-32-33-34-35-36-37-38-39-40-41
•28-29-30-31-32-33-34-35-36-37-38-39-40-41
-27-28-29-30-31-32-33-34-35-36-37-38-39
-27-28-29-30-31-32-33-34-35-36-37-38
•27-28-29-30-31-32-33-34-35-36-37-38-39
-27-28-29-30-31-32-33-34-35-36-37-38
-27-28-29-30-31-32-33-34-35-36-37
-27-28-29-30-31-32
-27-28-29-30
-27-28-29-30
-27-28
-27-28
TOTAL NON-BRAZILIAN NALC SCENES  =  801
   (* =PATH/ROWS PURCHASED FROM EOSAT USING EDC BROKERAGE AGREEMENT)
                                133

-------
     EPA 1992-1993 MSS  DATA TRIPLICATE PRIORITIES BY PATH/ROW

Priorities are indicated by numbers following the "#"  symbol.
                            1993 efforts
#1  Southeastern Mexico   39   scenes.
18/45-46
19/45-46-47
20/45-46-47-48-49
21/46-47-48-49-50
22/47-48-49
23/47-48-49
24/46-47-48-49
25/45-46-47-48-49
26/44-45-46-47-48
27/45-46-47-48
#1  Chesapeake Bay Watershed     21 scenes,

14/30-31-32-33-34
15/30-31-32-33-34
16/30-31-32-33-34
17/31-32-33-34
18/33-34
#1  Oregon Transect            2 scenes,

45/29
46/29
Priority total  =  62  scenes
                                134

-------
#2  Central America

10/54-55
11/54-55
12/53-54-55
13/54-55
14/53-54
15/50-51-52-53
16/49-50-51-52-53
17/49-50-51-52
18/49-50-51
19/48-49-50-51
20/50
        32  scenes.
 #2   Caribbean
  48 scenes,
 1/49-50-51-52
 2/48
 4/47-48
 5/47-48
 6/47
 7/46-47
 8/46-47-48
 9/45-46-47
 10/44-45-46-47
 11/43-44-45-46-47-48
 12/43 -45-46-47-48
 13/42-43-44-45-46
 14/41-42 -44-45
 15/44-45
 16/44-45
 17/44-45
 #2  Hawaii
8 scenes.
 62/46-47
 63/46-47
 64/45-46
 65/45
 66/45
  Priority total =88 scenes,
                                 135

-------
$3 US  Great Lakes Watershed
        35 scenes.
15/29
17/30
18/31-32
19/31-32
20/29-30-31-32
21/28-29-30-31
22/27-28-29-30-31
23/27-28-29-30-31
24/27-28-29-30
25/26-27-28
26/27-28
27/27-28

(note that some triplicates made for the Chesapeake Bay Watershed
also include portions of the US Great Lakes Watershed).
#3  US/Mexico Border
20 scenes.
26/42-43
27/42
28/40-41
29/40
30/40
31/40
32/39
33/38-39
34/38
35/38
36/38
37/38
38/38
39/37-38
40/37-38
#3  Southeastern US

16/36-37
17/35-36-37
18/35-36-37 -39
19/34-35-36-37
20/34-35-36-37
21/35-36

Priority total =  74 scenes.
    19 scenes,
                               136

-------
#4
Mexico Forested Regions   37 scenes,
28/46-47
29/47
30/45-46-47
31/44
32/40-41-42-43
33/40-41-42-43-44
34/39-40-41-42-43-44
35/39-40-41-42-43
36/39-40-41-42
37/39-40-41
38/39-40
39/39
#4  Western USA
               130 scenes.
31/39
33/33-34-35-36-37
34/31-32-33-34-35-36-37
35/31-32-33-34-35-36-37
36/29-30-31-32-33-34-35-36-37
37/29-30-31-32-33-34-35-36-37
38/29-30-31-32-33-34-35-36-37
39/29-30-31-32-33-34-35-36
40/26-27-28-29-30-31-32-33-34-35-36
41/26-27-28-29-30-31-32-33-34-35-36-37
42/26-27-28-29-30-31-32-33-34-35-36
43/26-27-28-29-30-31-32-33-34-35-36
44/26-27-28-29-30-31-32-33-34-35
45/26-27-28  -30-31-32-33
46/26-27-28  -30-31-32
47/26-27-28-29-30
48/26-27
 Priority total = 167 scenes
 Year total  =   391 scenes.
                                137

-------
                           1994  efforts
#5
Alaska
 180 scenes.
54/21-
55/21-
56/20-
57/19-
58/19-
59/18-
60/19
61/18
62/17-
63/17-
64/16-
65/15-
66/14-
67/13-
68/12-
69/11-
70/11-
71/11-
72/11-
73/11-
74/10-
75/10-
76/10-
77/10-
78/10-
79/10-
80/10-
81/10-
82/10-
83/12-
 22
 22
 21
 20-21
 20
 19-20
 18
 18
 17-18
 16-17-18
 15-16-17-
 14-15-16-
 13-14-15-
 12-13-14-
 12-13-14-
 12-13-14-
 12-13-14-
 12-13-14-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 11-12-13-
 13-14
18
17-18
16-17-
15-16-
15-16-
15-16-
15-16-
15-16-
14-15-
14-15-
14-15-
14-15-
14-15-
14-15-
14-15
14-15
14
18
17-18-19-20
17-18-19-20
17-18-19-20
17-18-19-20-21
17-18-19-20-21
16-17-18-19-  -21-22
16-17-18-19-  -22
16-17-18-19-  -22
16-17-18-     -23
16-17-18-     -23
16-17-18
                                138

-------
#6  Eastern and Southeastern  US    88  scenes,

06/38
10/29
11/27-28-29 -31
12/27-28-29-30-31
13/28-29-30-31-32
14/29- 35-36
15/35-36-37 -41-42-43
16/35- 38-39-40-41-42
17/38-39-40-41
18/38
19/33 -38-39
20/33 -38-39
21/32-33-34 -37-38-39-40
22/32-33-34-35-36-37-38-39-40
23/32-33-34-35-36-37-38-39-40
24/31-32-33-34-35-36-37-38-39
25/29-30-31-32-33-34-35-36-37-38-39-40
#7 Non-Forested  Mexico   19  scenes,

27/43-44
28/42-43-44-45
29/41-42-43-44-45-46
30/41-42-43-44
31/41-42-43
#8  Midwest and Great Plains
                           123 scenes.
26/29-
27/29-
28/26-
29/26-
30/26-
31/26-
32/26-
33/26-
34/26-
35/26-
36/26-
37/26-
38/26-
39/26-
30-31-
30-31-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28-
27-28
27-28
27-28
27-28
32-33
32-33
29-30
29-30
29-30
29-30
29-30
29-30
29-30
29-30
-34-35-36-37-38-39-40-41
-34-35-36-37-38-39-40-41
-31-32-33-34-35-36-37-38-39
-31-32-33-34-35-36-37-38-39
-31-32-33-34-35-36-37-38-39
-31-32-33-34-35-36-37-38
-31-32-33-34-35-36-37-38
-31-32
Year total = 410 scenes.

Grand total  =  801 scenes.
                               139

-------
                      Appendix   IV:






Triplicate Image Selection for all Path/Rows




 Caribbean



 Central America



 Chesapeake Bay Watershed



 Mexico



 Hawaii



 Alaska



 Western US



 Eastern and Southern US



 Midwest and Great Plains US
                             140

-------
     The following  lists provide  suggested triplicate images for
production.  The entries are  listed by region,  and identified by
the path  / row  scene identification  numbers for  Landsat World
Reference System 2  (WRS-2).   Also  listed  are the selected scene
dates, Landsat scene  identification numbers, and  for 70's epoch
scenes the  WRS-1 identifiers  that  are- correlated to  the WRS-2
scenes are provided.

     Some comments are included.  They include:  suggestions as to
making Reduced Cloud Composites, "Composite" or "Composite with" or
"Composite 3" when a third  image may be necessary; suggestions for
EDC to examine the images again, "EROS Check"; suggestions that a
second WRS-1  scene  may be unnecessary,  "Second WRS-1  scene not
required ?";  indications that the microfiche image is unsuitable to
judge  image  quality,   "fiche  missing" or  "bad  fiche"  or  "black
fiche" or  "faded fiche"; or  suggestions to  check  for available
scenes using the  EDC metadata  system,  "GLIS all 90's" or "GLIS for
ID".
                               141

-------
                NALC MSS Triplicates - Caribbean
Path 01  Row 49
     01/49
92/04/01

86/04/25
86/03/24

72/10/14
                              LM84354713304xO

                              LM85078513512xO
                              LM85075313521xO
LM8108314000500
     Second WRS-1 not required?
Path 01  Row 50

               92/09/08
               92/07/22

               86/01/19
               86/01/03

     251/50    73/03/24
               LM84370713361xO
               LM84365913345xO

               LM85068913535xO
               LM85067313541xO
               LM8124413553500
     Second WRS-1 not required?
Path 01  Row 51

               92/04/17

               86/01/19
               86/01/03

     251/51    73/03/24
               LM84356313320xO

               LM85068913542xO
               LM85067313544xO
               LM8124413560500
     Second WRS-1 not  required?
                    Composite with
                    Composite with
                    Composite with
                    Composite with
Path  01  Row 52

                92/04/17

                86/01/19
                86/01/03

      251/52     75/12/18
               LM84356313322xO

               LM85068913544xO
               LM85067313550xO

               LM8233013415500
                    Composite with
      Second WRS-1 not required?
                                142

-------
Path 02  Row 48
     01/48
92/03/07

86/03/15
86/01/10

73/02/17
                              LM84352213361xO

                              LM85074413580xO
                              LM85068013593xO
LM8120914001500
     Second WRS-1 not required?
                    Composite with
Path 04  Row 47
     04/47
92/08/12

85/01/21

73/03/10
73/01/15
LM84368013525xO

LM85032614142xO

LM8123014171500
LM8117614162500
Composite with
     Second WRS-1 not required?
Path 04  Row 48
     04/48
92/08/12
92/06/25

83/02/09

73/03/10
LM84368013531xO
LM84363213514xO

LM84020814134xO

LM8123014174500
     Second WRS-1 not required?
                                                  Composite with
Path 05  Row 47
               92/05/31

               83/06/24
               LM84360713564xO

               LM84034314200xO
     04/47
     05/47
73/03/10
73/01/15

78/03/30
LM8123014171500
LM8117614162500

LM82116313442xO
Composite with
                               143

-------
Path 05 Row 48
     04/48

     05/48


Path. 06 Row 47
     06/47
92/08/19

85/01/12

73/03/10

78/03/30
92/04/20
92/04/04
92/03/03

86/01/06

79/01/31
LM84368713593xO

LM85031714205xO

LM8123014174500

LM82116313445x0
LM84356614012xO
LM84355014005x0
LM84351814003xO

LM85067614235xO

LM82152414062xO
     Second WRS-1 not required?


Path 07 Row 46

               No 1990's

               86/01/13        LM85074714281xO

     07/46     79/02/01        LM82152414052xO

     Second WRS-1 not required?
Path 07 Row 47
      07/47
92/05/13

85/05/02

73/12/08
LM84358914082xO

LM85042714324xO

LM8150314311500
                                                  Composite 3
     Second WRS-1  not required?
                                144

-------
Path 08 Row 46
     08/46

     09/46


Path 08 Row 47
     08/47

     09/47


Path 08 Row 48
     08/48
92/05/04

86/07/31

72/08/28

75/02/06
92/05/20
92/05/04

86/01/20

73/12/09

75/02/06
92/05/20

86/03/09

73/12/09
LM84358014135xO

LM85088214303xO

LM8103614384500

LM8201514332500
LM84359614145x0
LM84358014141xO

LM85069014360xO

LM8150414365500

LM8201514334500
LM84359614151xO

LM85073814352xO

LM8150414372500
     Second WRS-1 not required.
Path 09  Row 45
     09/45
92/07/30

86/01/11

78/09/03
LM84366714223xO

LM85068114413xO

LM83018214333xO
                                                  Composite with
     Second WRS-1 not required?
                               145

-------
Path  09 Row 46
     09/46



     10/46
 92/06/12




 86/01/11




 75/02/06




 74/01/16
LM84363514215xO




LM85068114420xO




LM8201514332500




LM8154214470500
Path 09 Row 47
     09/47




     10/47






Path 10  Row 44
     11/44
 92/08/15




 86/09/08




 74/02/06




 74/01/16






t




 92/09/07




 86/08/14




 76/03/28
LM84368314235xO




LM85092114354xO




LM8201514334500




LM8154214472500










LM84370614293xO




LM85089614413xO




LM8243114393500
     Second WRS-1 not required?
Path 10  Row 45
     10/45
92/05/02




85/06/08




78/11/15
LM84357814254xO




LM85046414502xO




LM83025514394xO
     Second WRS-1 not required?
                               146

-------
Path 10 Row 46
     11/46
92/06/03
92/05/02

85/06/08

76/03/28
LM84361014272xO
LM84357814260xO

LM85046414505xO

LM8243114402500
     Second WRS-1 not required.
Path 10 Row 47
     10/47
92/03/31

85/06/08

74/01/16
LM84354614253xO

LM85056414511xO

LM8154214472500
     Second WRS-1 not  required?


Path 11  Row  43

               DOUBLET !!!

               86/03/14        LM85074314514xO

     12/43     73/04/05        LM8125615013500

     Second WRS-1 not  required?
 Path 11   Row 44
      11/44
 92/04/23

 85/04/28

 76/03/28
 LM84356914311xO

 LM85042314561xO

 LM8243114393500
      Second WRS-1 not required?
                                                  Composite with
                                147

-------
Path 11 Row 45
     12/45
Path. 11 Row 46
     11/46

     12/46


Path 11 Row 47
      12/47
          92/08/29

          85/02/07

          80/06/24




          92/04/23

          86/02/10

          76/03/28

          73/02/10




          .92/08/13

          87/05/04

          73/02/10
LM84369714354x0

LM85034314565xO

LM82198014482xQ




LM84356914320xO

LM85071114534xO

LM8243114402500

LM8120215023500




LM843 681143 60x0

IiM85115914504xO

LM8120215023500
      Second WRS-1 not required?
 Path 11 Row 48
                92/02/19
                92/03/22

                84/08/15
                         LM84350514324xO
                         LM84353714315XO
                         LM850167145572xO
                79/02/15
                78/11/17
12/48


Second WRS-1 not required?
 LM83034714521xO
 LM83025714520xO
                     Composite with
OR
EROS CHECK
                                148

-------
Path 12  Row 43
     12/43
92/03/29

86/01/16

73/04/05
LM84354414361xO

LM85068614591xO

LM8125615013500
     Second WRS-1 not required?
Path 12 Row 45
     12/45

     13/45


Path 12 Row 46
     12/46

     13/46


Path 12 Row 47
     12/47

     13/47
92/03/29
92/03/13

86/02/17

80/06/24

78/12/24
92/03/29

86/02/17

73/02/10

73/01/06




92/07/19

86/07/27

73/02/10

73/01/06
LM84354414370xO
LM84352814365xO

LM85071814591xO

LM82198014482xO

LM83029414563xO
LM84354414372xO

LM85071814594xO

LM8120215023500

LM8116715075500
LM84365614413xO

LM85087814554xO

LM8120215025500

LM8116715081500
                                                  Composite with
                                149

-------
Path 12 Row 48
               92/04/30

               87/02/04
                         LM84357614391xO

                         LM85107014541xO
               79/02/15
               78/10/30
12/48


Second WRS-1 not required?
LM83034714521xO
LM83023914522xO
Path 13 Row 42
     14/41

     14/42


Path 13 Row 43
     14/43
          92/03/20

          86/04/29

          72/10/09

          73/03/20




          92/03/20

          86/04/29

          73/03/20
LM84353514415xO

LM85078915024xO

LM8107815111500

LM8124015123500




LM84353514422xO

LM85078915030xO

LM8124015130500
     Second WRS-1 not required?
Path 13 Row 44
     13/44
          92/02/25

          86/04/29

          72/09/02
LM85291715023xO

LM85078915033xO

LM8104115063500
OR
EROS Check
     Second WRS-1  not required?
                                150

-------
Path 13 Row 45
     13/45



     14/45






Path 13 Row 46
     13/46
92/02/25




86/02/24




78/12/24




75/06/26











92/02/25




86/02/24




78/12/24
LM85291715030x0




LM85072515052xO




LM83029414563xO




LM8506814501500











LM85291715032xO




LM85072515054xO




LM83029414563xO
     Second WRS-1 not required?
Path 14  Row 41
     15/41
92/03/19



85/05/03



73/04/26
LM85294015073xO




LM85042815133x0




LM8127715174500
     Second WRS-1 not required?
Path 14 Row 42
     14/42



     15/42
92/03/19




86/04/20




73/03/20




75/02/03
LM85294015080xO




LM85078015090xO




LM8124015123500




LM8192515015500
                               151

-------
Path 14 Row 44
     14/44

     15/44


Path 14 Row 45
     14/45

     15/45


Path 15 Row 44
     15/44

     16/44
92/08/02
92/07/01

86/09/27
86/09/11

75/06/26

80/09/07
92/08/02
92/07/01

87/09/30

75/06/26

76/04/01
92/03/10

86/02/22

80/09/07

72/08/18
LM84367014531xO
LM84363814521xO

LM85092415051xO
LM85092415051xO

LM8506814501500

LM82205515061xO
LM84367014533xO
LM84363814523xO

LM85130815121xO

LM8506814501500

LM8243515025500
LM85293115150xO

LM85072315172xO

LM82205515061xO

LM8102615235500
                                                  Composite with
                                                  Composite with
                                                  Composite with
Path 15 Row 45
     16/44
92/07/24

86/08/01

72/08/18
LM84366114592xO

LM85088315132xO

LM8102615235500
     Second WRS-1 not required?
                               152

-------
Path 16 Row 44
     16/44

     17/44


Path 16 Row 45
     17/45
92/05/28

86/04/18

72/08/18

79/02/02
92/07/15
92/06/13

85/03/30
85/03/14

73/07/09
LM84360415032xO

LM85077815222xO

LM8102615235500

LM83033415191xO
                              LM84365215052xO
                              LM84362015041xO

                              LM85039415274xO
                              LM85037815275xO
LM8135115301500
     Second WRS-1 not required.
Path 17 Row 44
     18/44
92/09/08

86/07/30

72/10/31
LM84370715124xO

LM85088115252xO

LM8110015354500
     Second WRS-1 not required?
Path 17 Row 45
     18/45
92/03/08

86/07/30

78/11/05
LM85292915274xO

LM85088115254xO

LM83024515253xO
                    Composite with
                    Composite with
     Second WRS-1 not required?
                               153

-------
             NALC MSS Triplicates - Central American

Path 10  Row 54      To be selected

Path 10  Row 55      To be selected

Path 11  Row 54

               92/05/25       LM84360114362xO

               85/02/07       LM85034315003xO

     11/54     No low cloud data.

     12/54     74/02/23       LM8158015004500
Path 11   Row 55
                92/04/15
                92/02/27

                85/02/23
                85/02/07
               LE85296714542xO
               LE85291914543xO

               LM85035915005xO
               LM85034315005xO
      11/55

      12/55


 Path 12 Row 53
No low cloud 1970's data.

No low cloud 1970's data.
                92/04/22

                86/02/01
                86/01/16
               LE85297414594xO

               LM85068615031xO
               LM85068615031xO
      12/53
No low cloud 1970's data.
 Path 12 Row 54
      12/54

      13/54
 92/04/22

 87/01/19
 87/01/03

 74/02/23

 74/02/24
LE85297415000xO

LM85105414560xO
LM85103814553x0

LM8158015004500

LM8158115063500
                   Composite with
                   Composite with
                    Composite with
                                                   Composite with
                                 154

-------
Path 12 Row 55
     12/55

     13/55


Path 13 Row 54
92/04/22       LE85297415003xO

85/03/18       LM85038215070xO

No low cloud 1970'a data.

73/03/19       LM8123915121500
     13/54

     14/54


Path 13 Row  55
      13/55
92/03/28
92/03/12

87/03/15

74/02/24

74/01/20
92/03/12

86/03/28

73/03/19
LE85294915063xO
LE85293315063xO

LM85110915041xO

LM8158115063500

LM8154615130500
LE85293315070xO

LM85075715084xO

LM8123915121500
      Second WRS-1 not required?
 Path 14  Row 53
      15/53
 92/05/14
 92/03/27

 87/02/18
 87/02/02

 74/05/09
LE84359014541xO
LE84354214523xO

LM85108415091xO
LM85106815083xO

LM8165515160500
                                                  Composite with
                                                   Composite with
                                                   Composite with
      Second WRS-1 not required???
                                155

-------
Path. 14 Row 54
     14/54

     15/54
92/04/20
92/03/03

87/02/18
87/02/02
87/01/01

79/01/30

79/01/22
LE85297215123xO
LE85292415124xO

LM85108415093xO
LM85106815085xO
LM85103615075xO

LM83033115061xO

LM82146114582xO
                                                  Composite 3
Path 15 Row 50
      16/50
92/08/09

86/02/22
86/02/06

78/04/01
                               LE84367715015xO

                               LM85072315194xO
                               LM85070715200xO
LM83002715141x0
      Second WRS-1 not required.
Path 15  Row 51
      16/51
92/05/05

87/10/07

78/04/01
                               LE84358114591xO

                               LM85131515205xO

                               LM83002715144xO
      Second WRS-1 not required.
 Path 15 Row 52
      16/52
 92/05/05

 85/04/24
 85/04/08

 78/04/01
                               LE84358114593xO

                               LM85041915241xO
                               LM85040315242xO
                               LM83002715150x0
      Second WRS-1 not required.
                    Composite with
                     Composite with
                                156

-------
Path 15 Row 53
     15/53

     16/53


Path 16 Row 49
     16/49

     17/49


Path 16 Row 50
     16/50

     17/50


Path 16 Row 51
     16/51

     17/51
92/05/13

86/02/06

74/05/09

75/03/03




92/04/02

86/01/28

78/04/19

78/04/02
92/04/02
92/03/01

86/01/28

78/04/01

76/03/16
92/04/26
92/03/01

86/01/28

78/04/01

78/04/02
LE85299515180xO

LM85072315205xO

LM8165515160500

LM8204015163500
LE85295415230xO

LM85069815260xO

LM83004515140xO

LM83002815193xO
LM85295415232xO
LM85292215233xO

LM85069815262xO

LM83002715141xO

LM8241915165500
LM84357215045xO
LM85292215235xO

LM85069815265xO

LM83002715144xO

LM83002815202xO
                                                  Composite with
                                                  Composite with
                               157

-------
Path 16 Row 52
     16/52

     17/52


Path 16 Row 53
     16/53

     17/53


Path 17  Row 49
92/03/01

86/04/02

78/04/01

78/04/02




92/03/01

86/03/01

75/03/03

78/04/02
      17/49

      18/49


 Path 17 Row 50
 92/04/01
 92/03/08

 86/03/08

 78/04/02

 73/12/19
      17/50

      18/50
 92/03/08

 85/02/01

 76/03/16

 75/03/23
LE85292215242xO

LM85076215255xO

LM83002715150xO

LM83002815205xO




LE85292215244xO

LM85073015265xO

LM8204015163500

LM83002815211XO
LE84354715093XO
LE85292915292xO

LM85073715312xO

LM83002815193xO

LM8151415345500
 LE85292915294xO

 LM85033715355xO

 LM8241915165500

 LM8206015263500
                                                   Composite with
                                 158

-------
Path 17 Row 51
     17/51

     18/51


Path 17 Row 52
     17/52
92/04/01
92/03/08

85/02/01

78/04/02

72/12/24
92/04/01

86/02/20

78/04/02
LE84354715102xO
LE85292915300xO

LM85033715362xO

LM83002815202xO

LM8115415385500
LE84354715104xO

LM85072115325xO

LM83002815205xO
                                                  Composite with
     Second WRS-1 not  required?
Path 18 Row 49
      18/49

      19/49.


 Path 18  Row 50
      18/50

      19/50
92/03/23

86/03/15

73/12/19

78/12/21




92/03/23

86/03/15

75/03/23

74/03/02
LE84353815153xO

LM85074415372xO

LM8151415345500

LM82142915180xO




LE84353815160xO

LM85074415375xO

LM8206015263500

LM8158715391500
                                159

-------
Path 18 Row 51
     18/51

     19/51


Path 19 Row 48
92/03/23

86/03/31

72/12/24

75/03/06
                92/06/02
                92/03/30

                86/08/29
                86/07/28

     20/48      75/03/25
LE84353815162xO

LM85076015375xO

LM8115415385500

LM8204315323500
               LM84360915234xO
               LM84354515213xO

               LM85091115381xO
               LM85087915392xO
               LM8206215371500
     Second WRS-1 not required.
Path  19 Row 49
      20/49
92/05/01
92/03/30

87/12/14

75/03/25
LE84357715225xO
LE84354515215xO

LM84197715394xO

LM8206215373500
      Second WRS-1 not required.
 Path 19 Row 50
      20/50
 92/05/01

 86/04/07

 78/12/31
 LE84357715231xO

 LM85076715433xO

 LM83030115385xO
                    Composite with
                    Composite with
                                                   Composite with
      Second WRS-1 not required.
                                160

-------
Path 19 Row 51
     20/51
92/05/01

86/03/22

73/02/18
LE84357715234xO

LM85075115441xO

LM8121015503500
     Second WRS-1 not required?
Path 20 Row 50
     21/50
92/03/21
92/04/22

86/03/13

76/03/02
LE84353615282xO
LE84356815291xO

LM85074215501xO

LM8240515402500
                                                  Composite with
     Second WRS-1 not required?
                               161

-------
         NALC MSS Triplicates - Chesapeake Bay Watershed
Path 14  Row 30
     15/30
91/09/09




86/08/26




73/07/07
LM85274815025xO




LM85090815001xO




LM8134915123500
     Second WRS-1 not required?
Path 14  Row 31
     15/31
Path 14  Row 32
      15/32
91/06/21




87/06/10




73/07/07




73/07/24





I




91/06/21




87/06/10




73/08/30
LM85266815022x0




LM85119615035xO




LM8134915125500




LM8136615065500










LM85266815025xO




LM85119615041xO




LM8140315123500
      Second WRS-1 not required?
 Path 14  Row 33
      15/33
 91/09/09




 87/06/10




 73/08/30




 73/08/29
 LM85274815040xO




 LM85119615044xO




 LM8140315125500




 LM8140215071500
                                162

-------
Path 14  Row 34
     15/34
91/09/09



85/06/20




73/08/12
LM85274815043xO




LM85047615104xO




LM8138515134500
     Second WRS-1 not required?
Path 15  Row 30
     16/30



     17/30
90/09/13




87/06/17




73/08/30




72/08/19
LM85238715050xO




LM85120315095xO




LM8140315114500




LM8102715233500
Path 15  Row 31
     16/31
90/08/12




87/06/17




73/07/08
LM85235515053xO




LM85120315101xO




LM8135015183500
     Second WRS-1 not required?
Path 15  Row 32
     16/32
90/07/27




87/06/17




73/07/03
LM85233915060xO




LM85120315104xO




LM8135015190500
     Second WRS-1 not required?
                               163

-------
Path 15  Row 33



               90/06/25




               87/06/17




     16/33     73/07/08
LM85230715063xO




LM85120315110xO




LM8135015192500
     Second WRS-1 not required?
Path 15  Row 34



               90/06/25




               85/08/14




     16/34     73/08/31
LM85230715070xO




LM85053115163xO




LM8140415190500
     Second WRS-1 not  required?
Path 16  Row 30



               91/08/22




               86/07/23




     17/30     74/07/22
LM85273015150xO




LM85087415135xO




LM8172915153500
Path  16   Row 31



                91/08/07




                86/09/09




      17/31     72/09/06
LM85274615154xO




LM85092215123xO




LM8104515240500
                                164

-------
Path 16  Row 32
     17/32
91/09/07



86/09/09




72/09/06
LM85274615160xO




LM85092215125xO




LM8104515243500
     Second WRS-1 not required?
Path 16  Row 33
     17/33
91/09/07




86/09/09



76/09/12
LM85274615162xO




LM85092215132xO




LM8259915045500
     Second WRS-1 not required?
Path 16  Row 34
     17/34
90/08/30




87/06/08




75/08/22
LM85234615130xO




LM85119415172xO




LM8512514593500
     Second WRS-1 not required?
Path 17  Row 31
     18/31
92/06/10




87/06/15




75/06/04
LM85265715204xO




LM85120115223xO




LM8215015184500
                                165

-------
Path. 17  Row 32




               91/08/13




               85/08/12




     18/32     75/06/21
LM85272115215xO




LM85052915280xO




LM8215015190500
Path 17  Row 34



               92/09/14



               86/07/14



     18/34     73/09/02
LM85275315230xO




LM85086515215xO




LM8140615303500
                                166

-------
                  NALC MSS Triplicates - Mexico
Path 18  Row 45
      To be selected
Path 18  Row 46
      To be selected
Path 19  Row 45
     20/45
90/11/04

86/07/28
86/07/12

76/01/07
                              LM84303315422xO

                              LM85087915381xO
                              LM85086315384xO
LM8235015333500
     Second WRS-1 not required?
Path 19  Row 46
     20/46
92/05/01

86/12/19
86/12/03

76/02/12
                              LE84357715214xO

                              LM85102315351xO
                              LM85100715344xO
LM8238615332500
     Second WRS-1 not required?
Path 19  Row 47
                    EROS CHECK

                    Composite
                    Composite
     20/47
90/11/20

84/11/11

75/12/02
LM84304915424xO

LM85025515463xO

LM8231415344500
     Second WRS-1 not  required?
                                167

-------
Path 20  Row 45
     21/45

     22/45


Path 20  Row 46
     21/46

     22/46
90/04/17

86/03/13

74/03/04

75/11/16




92/05/08

86/03/13

76/04/07

76/02/14
72/08/24
LM84283215531xO

LM85074215481xO

LM8158915483500

LM8229815453500




LE84358415281xO

LM85074215483xO

LM8244115373500

LM8238815445500
LM8103215585500
OR
EROS CHECK
Path 20  Row 47
     21/47

     22/47


Path 20  Row 48
90/04/17

86/04/14
86/03/13

78/01/15

74/02/15
     21/48
90/04/17

86/04/14

74/02/14
LM84283215540xO

LM85077415482xO
LM85074215490XO

LM82108915121xO

LM8103215592500
LM84283215542xO

LM85077415484xO

LM8157115500500
                                                  Composite
     Second WRS-1 not required?
                               168

-------
Path 20  Row 49

     No 90's data?       Doublet not triplet.

               86/03/13       LM85074215495xO

     21/49     79/02/15       LM83033815441xO
Path 21  Row 46
     To be selected
Path 21  Row 47
     22/47

     23/47



Path 21  Row 48
92/08/27
90/12/20

84/11/25

74/02/15

78/09/17
80/11/17
     22/48

     23/48
92/08/27

84/11/25

74/02/15

75/12/05
LE85310115512xO
LM84307915540xO

LM85026915590xO

LM8157215552500

LM83019615545xO
LM83098815410xO
LE85310115515xO

LM85026915592xO

LM8157215554500

LM8231715522500
                                                  OR
                                                  EROS CHECK
OR
EROS CHECK
                               169

-------
Path 21  Row 49
     22/49

     23/49


Path 21  Row 50
90/12/20

86/01/15

74/02/15

75/12/05
     22/50
90/10/01
90/08/06

86/01/15

74/02/15
LE84307915545xO

LM85068515571xO

LM8157215561500

LM8231715524500
LM84299915574xO
LM85234915501xO

LM85068515574xO

LM8157215563500
                                                  EROS CHECK
     Second WRS-1 not required?
Path 22  Row 47
     23/47

     24/47


Path 22  Row 48
92/04/20

86/04/12

78/09/17

72/12/30
     23/48

     24/48
90/11/25
90/11/09

84/01/09

80/11/17

76/05/16
LE84356615401xO

LM85077216004xO

LM83019615545xO

LM8116016112500
LM84305416013xO
LM84303816015xO

LM84054216040xO

LM83098815412xO

LM8248015542500
                                                  Composite
                               170

-------
Path 22  Row 49
     23/49

     24/49
92/04/04
92/03/03

86/03/11

75/12/05

73/01/17
LE84355015403xO
LE84351815401xO

LM85074016021xO

LM8231715524500

LM8117816115500
                                                  Composite
Path 23  Row 47
     24/47

     25/47
90/12/18

86/03/18

72/12/30

76/04/11
76/03/24
LE84307716063xO

LM85074716072xO

LM8116016112500

LM8244516004500
LM8242716012500
                                                  EROS CHECK
Composite
Path 23  Row 48
     24/48

     25/48
90/12/18

86/05/05
86/04/03

76/05/16

79/02/28
LE84307716065xO

LM85079516064xO
LM85076316073xO

LM8248015542500

LM83036016063xO
                                                  EROS CHECK

                                                  Composite
Path 23  Row 49
     24/49


     25/49
92/04/11

86/04/19

73/01/17
72/12/30

73/11/20
LE84355715465xO

LM85077916073xO

LM8117816115500
LM8116016121500

LM8148516153500
Composite
                               171

-------
Path  24   Row 46
     To be selected
Path  24   Row 47
     25/47

     26/47


Path 24  Row 48
90/08/27

87/08/19
87/08/03

73/02/05

73/05/25
     25/48
     26/48
92/04/18
92/04/02

86/04/10
86/03/25

79/03/09
79/02/28

73/01/19
73/05/25
Path 24  Row 49

               92/04/18

               86/03/25

     25/49     73/11/20

     26/49     74/02/01


Path 25  Row 45

               89/10/10

               86/01/27

     27/45     72/11/09
LM85237016073xO

LM85126616134xO
LM85125016132xO

LM8119716172500

LM8130616231500
LE84356415530xO
LE84354815522xO

LM85077016133xO
LM85075416135xO

LM82150715551xO
LM83036016063xO

LM8118016231500
LM8130616233500
               LE84356415532xO

               LM85075416141xO

               LM8148516153500

               LM8155816192500




               LM84265916261xO

               LM85069716200xO

               LM8110916275500
                                                   Composite
Composite


Composite


Composite
OR
EROS CHECK
                               172

-------
Path 25  Row 46
     26/46


     27/46
89/11/11

84/12/23

73/05/25
73/03/14

73/04/20
LM84267516263xO

LM85029716233xO

LM8130616224500
LM8123416230500

LM8127116284500
Composite
Path 25  Row 47
     26/47

     27/47


Path 25  Row 48
92/05/27

86/04/17
86/03/16

73/05/25

73/02/27
     26/48


     27/48


Path 25  Row 49
90/12/16

86/01/27

73/01/19
73/05/25

79/02/03
     26/49

     27/49
90/08/02

84/12/23

74/02/01

74/02/20
LE84360316001xO

LM85077716191xO
LM85074516195xO

LM8130616231500

LM8119916285500
LE84307516192xO

LM85069716212xO

LM8118016231500
LM8130616233500

LM82147316050xO
LM85234516143xO

LM85029716244xO

LM8155816192500

LM8157716245500
                                                  Composite
                                                  EROS CHECK
OR
EROS CHECK
                                173

-------
Path 26  Row 42
     28/42
92/07/29
92/06/27

87/08/17

73/09/12
LE85312016200xO
LE85304016211xO

LM85126416240xO

LM8141616310500
     Second WRS-1 not required.
Path 26  Row 43
     28/43
90/10/20

85/10/30

74/09/07
LM84301816252xO

LM85060816271xO

LM8177616221500
     Second WRS-1 not required.
Path 26  Row 44
     27/44

     28/44


Path 26  Row 45
     27/45

     28/45
90/10/20

85/10/30

72/11/09

73/05/27




90/10/20

85/10/30

72/11/09

73/05/27
LM84301816254xO

LM85060816273xO

LM8110916273500

LM8130816332500




LM84301816260xO

LM85060816275xO

LM8110916275500

LM8130816334500
                                                  Composite
                                                  EROS CHECK
                                174

-------
Path 26  Row 46
     27/46



     28/46






Path 26  Row 47
     27/47




     28/47






Path 26  Row 48
     27/48




     28/48










Path 27  Row 42
 90/11/21




 85/10/30




 73/04/20




 73/11/23





r




 89/03/07




 85/01/31




 73/02/07




 73/11/23






J




 92/05/18




 86/04/24




 79/02/03




 73/03/16
LM84305016253xO




LM85060816282xO




LM8127116284500




LM8148816312500
LM84242616325xO




LM85033616302xO




LM8119916285500




LM8148816315500











LE84359416063xO




LM85078416253xO




LM82147316050xO




LM8123616352500
Use as much  of  73/03/16  as possible.
     28/42



     29/42
 91/07/18




 86/07/04




 73/09/12




 74/03/30
LE85269516273xO




LM85085516271xO




LM8141616310500




LM8161516324500
                               175

-------
Path. 27  Row 43
     28/43

     29/43


Path 27  Row 44
92/04/07

86/05/01
86/03/14

73/05/27

73/03/17
     28/44

     29/44


Path 27  Row 45
92/05/09
92/04/07

86/03/14

73/05/27

73/05/10
      28/45

      29/45
92/05/09

83/05/17

73/05/27

76/03/28
Path 27   Row 46
      28/46

      29/46
92/05/25
92/05/09

86/03/14

73/04/21

76/03/28
LE84355316090xO

LM85079116293xO
LM85074316304xO

LM8130816325500

LM8123716390500
LE84358516104xO
LE84355316093xO

LM85074316310xO

LM8130816332500

LM8129116391500
LE84358516110xO

LM84030516350xO

LM8130816334500

LM8243116231500
LE84360116120xO
LE84358516113xO

LM85074316315xO

LM8127216343500

LM8243116234500
                                                  Composite
                                                   Composite
                                                   Composite
                                176

-------
Path. 27  Row 47
     28/47



     29/47






Path 27  Row 48
     28/48




     29/48
92/09/06




86/03/14




73/11/23




76/03/28






J




DOUBLET




86/03/14




73/03/16




76/03/28
Path  28   Row 40
      30/40
 92/06/25




 86/07/27




 73/05/29
LE85311116282xO




LM85074316321xO




LM8148816315500




LM8243116240500
LM85074316324xO




LM8123616352500




LM8243116243500











LE85303816324xO




LM85087816315xO




LM8131016431500
      Second WRS-1 not required?
 Path 28  Row 41
      29/41



      30/41
 92/06/25




 86/07/27




 74/09/08




 73/05/29
 LE85303816331xO




 LM85087816321x0




 LM8177716271500




 LM8131016433500
                                177

-------
Path 28  Row 42
     29/42




     30/42






Path 28  Row 43
     29/43



     30/43






Path 28  Row 44
     29/44




     30/44
90/03/16




83/04/22




74/03/20




76/04/16






I




90/03/16




83/04/22




73/03/17




74/06/29






t




92/04/30




83/04/22




73/05/10




73/03/18
Path 28  Row 45
     29/45




     30/45
 90/03/16




 86/04/06




 76/03/28




 76/03/20
LM85220616305xO




LM84028016401xO




LM8161516324500




LM8245016271500











LM85220616311xO




LM84028016403xO




LM8123716390500




LM8170616361500










LE84357616162xO




LM84028016410xO




LM8129116391500




LM8123816451500










LM85220616320xO




LM85076616371xO




LM8243116231500




LM8533616053500
                                178

-------
Path 28  Row 46
     29/46

     30/46



Path 28  Row 47
 92/04/30

 86/04/06

 76/03/28

 73/05/11
 76/03/20
     29/47

     30/47


Path 29  Row 39
     31/39
 90/03/16

 86/04/06

 76/03/28

 76/03/20


)

 91/07/16

 85/07/31

 73/07/05
LE84357616171xO

LM85076616373xO

LM8243116234500

LM8129216455500
LM8533616055500
OR
EROS CHECK
LM85220616325xO

LM85076616375xO

LM8243116240500

LM8533616062500




LE85269316384XO

LM85051716451xO

LM8134716480500
     Second WRS-1 not required?
Path 29  Row 40
     31/40
92/04/21

86/03/12

75/03/18
LE84356716204xO

LM85074116415xO

LM8205516370500
     Second WRS-1 not required?
                               179

-------
Path 29  Row 41
     31/41
Path 29  Row 42
     31/42



     32/42
92/04/21




86/03/12




73/12/14





i




92/04/21




86/04/14




76/09/08




76/05/06
Path  29   Row 43
      30/43




      31/43






 Path 29  Row 44
      30/44




      31/44
 92/04/21




 86/04/13




 73/07/22




 72/08/15






 I




 92/04/21




 86/04/1-3




 73/03/18




 76/03/21
LE84356716211xO




LM85074116421xO




LM8150916461500










LE84356716213xO




LM85077316420xO




LM8259516290500




LM8247016380500










LE84356716220xO




LM85077316422xO




LM8136416434500




LM8102316490500










LE84356716222xO




LM85077316424x0




LM8123816451500




LM8533716104500
                                180

-------
Path 29  Row 45
     30/45



     31/45
92/04/21




86/03/12




76/03/20




76/03/21
Path 29  Row 46




               92/04/21




               86/03/12




     30/46     76/03/20




     31/46     73/02/11






Path 29  Row 47




               92/04/21




               86/03/12




     30/47     76/03/20




     31/47     72/12/19
LE84356716224xO




LM85074116435xO




LM8533616053500




LM8533716111500
               LE84356716231xO




               LM85074116441xO




               LM8533616055500




               LM8120316513500
               LE84356716233xO




               LM85074116444xO




               LM8533616062500




               LM8114916514500
Path 30  Row 40
     32/40
92/04/28




86/04/20




75/06/17
LE84357416271xO




LM85078016471xO




LM8214616424500
     Second WRS-1 not required?
                               181

-------
Path 30  Row 41
     32/41
92/04/28




86/04/20




73/03/02
LE84357416273xO




LM85078016473xO




LM8122216552500
     Second WRS-1 not required?
Path 30  Row 42
     32/42
92/04/28




86/04/20




76/05/06
LE84357416275xO




LM85078016480xO




LM8247016380500
     Second WRS-1 not required?
Path 30  Row 43
     32/43
92/05/14




86/04/20




76/05/06
LE84359016290xO




LM85078016482xO




LM8247016383500
     Second WRS-1 not required?
Path 30  Row 44
     32/44
90/05/25




86/04/20




73/03/02
LM84287016535xO




LM85078016485xO




LM8122216563500
     Second WRS-1 not required?
                               182

-------
Path 30  Row 45
     32/45
 92/05/14
 92/04/28

 86/04/20

 73/03/02
LE84359016295xO
LE84357416291xO

LM85078016491xO

LM8122216570500
                                                   Composite
     Second WRS-1 not required?
Path 30  Row 46
     32/46
 92/05/30
 92/05/14

 86/04/20

 73/03/02
LE84360616305xO
LE84359016301xO

LM85078016493xO

LM8122216572500
                                                  Composite
     Second WRS-1 not required?
Path 30  Row 47
     31/47

     32/47


Path 31  Row 39
     33/39
DOUBLET

85/05/19

72/12/19

76/05/24


)

92/06/30

86/06/14

72/07/30
LM85044416545xO

LM8114916514500

LM8248816394500




LE85304316505xO

LM85083516512xO

LM8100716590500
     Second WRS-1 not required?
                               183

-------
Path 31  Row 40
     33/40
92/05/21

87/06/17

74/05/27
LE84359716342xO

LM85120316524xO

LM8167316532500
     Second WRS-1 not required?
Path. 31  Row 41
     33/41
92/05/21

85/05/10

75/05/13
LE84359716344xO

LM85043516583xO

LM8211116484500
     Second WRS-1 not required?
Path 31  Row 42
     33/42
92/04/19

85/05/10

76/03/23
LE84356516335xO

LM85043516590xO

LM8533916211500
     Second WRS-1 not  required?
Path 31  Row 43
                92/05/05
                92/04/19
               LE84358116345xO
               LE84356516341xO
                    Composite
                86/03/10        LM85073916552xO

      33/43      76/03/23        LM8533916214500

      Second WRS-1  not requied?
                                184

-------
Path 31  Row 44
     33/44
92/05/05
92/04/19

86/03/10

73/03/03
LE84358116351xO
LE84356516344xO

LM85073916555xO

LM8122317022500
     Second WRS-1 not required?
Path 32  Row 38
     34/38
92/06/29

85/07/04

75/06/28
LE84363616405xO

LM85049017033xO

LM8507016420500
     Second WRS-1 not required?
Path.32  Row 39
     34/39

     35/39
92/06/05

86/05/04

73/06/02

73/06/03
LE85301816572xO

LM85079416585xO

LM8131417053500

LM8131517112500
                                                  Composite
Path 32  Row 40
     34/40

     35/40
92/04/10

86/04/18

73/06/02

73/06/21
75/06/20
LE84355616385xO

LM85077816593xO

LM8131417060500

LM8133317113500
LM8214917000500
OR
EROS CHECK
                               185

-------
Path 32  Row 41
     34/41



     35/41
92/04/10




86/04/02




73/06/02




73/06/03
Path 32  Row 42



               92/04/10




               86/04/02




     34/42     73/06/02




     35/42   ' 75/07/08
Path 32  Row 43
      34/43




      35/43
92/04/10




84/04/04




73/04/09




73/02/15
LE84355616392xO




LM85076217002xO




LM8131417062500




LM8131517121500











LE84355616394xO




LM85076217004xO




LM8131417065500




LM8216717004500











LE84355616401xO




LM84062817022xO




LM8126017074500




LM8120717131500
Path 33   Row 38
      35/38




      36/38
 92/06/04




 86/07/30




 73/06/21




 74/06/17
 LE84361116462xO




 LM85088117015x0




 LM8133317104500




 LM8169417090500
                                186

-------
Path 33  Row 39
     35/39



     36/39






Path 33  Row 40
     35/40



     36/40






Path 33  Row 41
     35/41



     36/41
 92/04/01




 86/03/24




 74/04/05




 73/05/17






)




 92/04/01




 86/03/24




 75/06/20




 73/04/11











 92/04/01




 86/03/24




 73/06/03




 73/03/24
LE84354716442xO




LM85075317055xO




LM8162117054500




LM8129817171500











LE84354716445xO




LM85075317062xO




LM8214917000500




LM8126217175500











LE84354716451xO




LM85075317064xO




LM8131517121500




LM8124417182500
Path 33  Row 42
     35/42




     36/42
92/05/03




86/03/24




73/01/28




73/03/24
LE84357916464xO




LM85075317071xO




LM8118917123500




LM8124417184500
                               187

-------
Path 33  Row 43
     35/43
92/06/04




86/03/24




73/02/15
LE84361116482xO




LM85075317073xO




LM8120717131500
     Second WRS-1 not required.
Path 33  Row 44
     36/44
92/05/03




86/03/24




73/04/11
LE84357916473xO




LM85075317075xO




LM8126217193500
     Second WRS-1 not  required?
Path 34  Row 38
     36/38



     37/38
92/06/11




86/05/18




74/06/17




73/06/05
LE85301617092xO




LM85080817102xO




LM8169417090500




LM8131717222500
Path 34   Row 39
      36/39




      37/39
92/06/11




83/06/03




73/05/17




73/06/05
LE84361816530xO




LM84032217155xO




LM8129817171500




LM8131717224500
                                188

-------
Path 34  Row 40
     36/40




     37/40






Path 34  Row 41
     36/41



     37/41






Path 34  Row 42
92/06/11




83/06/03




73/04/11




73/04/12










92/04/24



86/03/15



73/03/24



73/04/12
     36/42




     37/42
92/04/24




86/03/15




73/03/24




73/03/25
LE84361816533xO




LM84030617161xO




LM8126217175500




LM8126317234500
LE84357016521xO




LM85074417130xO




LM8124417182500




LM8126317240500











LE84357016523xO




LM85074417133xO




LM8124417184500




LM8124517243500
Path 34  Row 43
     36/43




     37/43
92/04/24




86/03/15




73/04/11




73/03/25
LE84357016530xO




LM85074417135xO




LM8126217191500




LM8124517245500
                               189

-------
Path 34  Row 44
     36/44
Path 35  Row 38
     37/38




     38/38






Path 35  Row 39
92/04/24




86/03/15




73/04/11






5




92/06/02




86/06/10




73/06/05




74/06/01
     37/39




     38/39
 92/06/02




 83/04/23




 73/06/05




 73/06/06
LE84357016532xO




LM85074417141xO




LM8126217193500










LE84360916583xO




LM85083117155xO




LM8131717222500




LM8167817210500










LE84360916590xO




LM84028117222xO




LM8131717224500




LM8131817283500
Path 35  Row 40
      37/40



      38/40
 92/05/01




 83/04/23




 73/04/12




 75/03/16
LE84357716581xO




LM84028117224xO




LM8126317234500




LM8196617111500
                                190

-------
Path 35  Row 41
     37/41



     38/41






Path 35  Row 42
     37/42



     38/42






Path 35  Row 43
92/05/01




83/04/23




73/04/12




73/05/09






I





92/05/01




86/04/23




73/05/18




73/05/01
     37/43




     38/43
 92/05/01




 86/04/23




 73/03/25




 74/11/28
LE84357716583xO




LM84028117230xO




LM8126317240500




LM8130017293500











LE84357716590xO




LM85078317185xO




LM8129917241500




LM8128217301500











LE84357716592xO




LM85078317191xO




LM8124517245500




LM8185817170500
 Path  36   Row 38
      38/38




      39/38
 92/10/07




 86/09/05




 74/06/01




 72/09/28
LE85314217194xO




LM85091817190xO




LM8167817210500




LM8106717330500
                                191

-------
Path. 36  Row 39
     38/39




     39/39






Path 36  Row 40
     38/40




     39/40
92/05/08




83/04/14




73/06/06




73/05/20






)




92/05/08




83/04/14




72/08/22




72/09/10
Path  36   Row 41
      38/41



      39/41






 Path 36  Row 42
      38/42
 92/05/24




 87/07/22




 72/08/22




 72/09/10





»




 92/05/24




 87/07/22




 73/05/01
LE84358417042xO




LM84027217282xO




LM8131817283500




LM8130117342500










LE84358417044XO




LM84027217285xO




LM8103017280500




LM8104917340500










LE84360017054xO




LM85123817244x0




LM8103Q17283500




LM8104917342500










LE84360017060xO




LM85123817250xO




LM8128217301500
      Second WRS-1 not required.
                                192

-------
Path 37  Row 38




               92/10/14




               86/10/14




     39/38     72/10/16




     40/38     72/09/29






?ath 37  Row 39




               92/05/31




               87/07/13




     39/39     73/05/20




     40/39     73/06/08
Path 37  Row 40
     39/40



     40/40
92/05/15




83/05/07




72/09/10




73/04/15
               LE85314917255xO




               LM85095717240xO




               LM8108517332500




               LM8106817385500
LE84360717112xO




LM85122917295xO




LM8130117342500




LM8132017395500











LE84359117111xO




LM84029517350xO




LM8104917340500




LM8126617405500
Path 37  Row 41




               92/04/29




               86/04/21




     39/41     76/03/20




     40/41     76/03/21
               LE84357517105xO




               LM85078117305xO




               LM8242317191500




               LM8242417245500
                                193

-------
Path 38  Row 37
     40/37

     41/37


Path 38  Row 38
     40/38

     41/38
92/08/02

86/09/03

72/09/29

72/09/30


>

92/07/17

86/06/15

73/06/08

73/07/15
Path  38  Row 39
      40/31


      41/39


 Path 39  Row 40


 Path 39  Row 37
 92/04/20

 87/07/20

 73/06/08
 73/04/15

 76/03/31
LE85307617325xO

LM85091617310xO

LM8106817382500

LM8106917441500




LE85306017333xO

LM85083617341xO

LM8132017393500

LM8135717445500




LE84356617160XO

LM85123617361xO

LM8132017395500
LM8126617402500

LM8534717052500
      To be selected
      41/37

      42/37
 92/06/30

 86/06/06

 73/06/09

 73/05/23
 LE84363717235xO

 LM85082717402xO

 LM8132117445500

 LM8130417504500
OR
EROS CHECK
                                194

-------
Path 39  Row 38
     41/38




     42/38






Path 39  Row 39
     41/39
92/06/30




86/05/05




73/05/22




73/06/10






)




DOUBLET




92/06/30




76/06/11
Path 40  Row 37
      42/37




      43/37






 Path 40  Row 38
 92/08/16




 87/09/04




 72/09/13




 73/09/27
LE84363717241xO




LM85079517414xO




LM8130317453500




LM8132217510500
LE84363717244xO




LM8534717052500











LE85309017450xO




LM85128217490xO




LM8105217495500




LM8143117543500
      To be selected
                                195

-------
                 NALC MSS Triplicates - Hawaii
Path 62  Row 46

               No good 90'8 data.

               No 80's data.

     67/46     73/02/11       LM8120320180500
Path. 62  Row 47
     67/47
90/07/28

No 80's data.

73/02/11
75/07/22
                               LM84293420113x0
LM8120320182500
LM8218120061500
OR
Check fiche
Path 63  Row 46
                92/07/24
                92/05/21
               LM84366119563xO
               LM84359719542xO
                No 1980's data.

      67/46      73/02/11       LM8120320180500
                    Composite with
Path 63   Row 47
     67/47
92/08/09
92/07/24

No 1980's data

73/02/11
75/07/22
                               LM84367719572xO
                               LM84366119565xO
LM8120320182500
LM8218120061500
                    Composite with
OR
Check fiche
                                196

-------
Path 64  Row 45
               92/07/15
               92/06/29
               LM84365220020xO
               LM84363620013xO
 Composite with
     68/45
No 1980'8 data

73/01/25       LM8118620230500
Path 64  Row 46

               92/07/15       LM84365220022xO

               No 1980's data

     68/46     LM
Path 65  Row 45
               92/08/07
               92/07/22
               92/07/06
               92/06/20
Path 66  Row 45
               92/07/29
               92/06/27

               No 1980's data
               LM84367520085xO
               LM84365920082xO
               LM84364320075xO
               LM84362720072xO
               LM84366620144xO
               LM84363420135xO
Composite 3 of 4?
 Composite with
                               197

-------
                 NALC MSS Triplicates - Alaska
Path 54  Row 21
     59/21



     60/21
92/07/01




85/07/30




74/09/02




73/07/16
LM85304419055xO




LM85051619123xO




LM8177119111500




LM8135819264500
Path 54  Row 22



               GLIS ALL 90'8



               85/07/30       LM85051619125xO




     59/22     GLIS




     60/22     GLIS






Path 55  Row 21



               GLIS ALL 90's



               86/08/09       LM85089119104xO




     60/21     73/07/16       LM8135819264500




     Second WRS-1 not required?




Path 55  Row 22



               GLIS ALL 90's



               86/09/26       LM85093919092xO




     60/22     GLIS




     Second WRS-1 not required?
                                198

-------
Path 56  Row 20

               GLIS ALL 90's

               86/09/17       LM85093019150xO

     61/20     74/09/22       LM8179119212500

     Second WRS-1 not required?


Path 56  Row 21

               GLIS ALL 90's

               86/09/17       LM85093019153xO

     61/21     GLIS at 0-50% cloud level

     Second WRS-1 not required?
Path 57  Row 19
     61/19

     62/19


Path 57  Row 20
GLIS ALL 90's

86/06/04
84/08/17

GLIS

GLIS 0-50%
                                        OR
               GLIS ALL 90's

               86/06/04
               86/10/10
                         OR
     61/20     GLIS 0-50%

     62/20     74/07/31       LM8173819291500
                               199

-------
Path 57  Row 21
     61/21



     62/21






Path. 58  Row 19
GLIS ALL 90 's




86/06/04       GLIS  for ID




GLIS 0-50%




GLIS
     63/19




     64/19






Path 58  Row 20
     63/20
92/06/27




86/06/27




72/08/11




GLIS  0-50%






)




92/06/27




86/09/15




74/09/06
LM85304019295xO




LM85084819295xO




LM8101919430500
LM85304019301xO




LM85092819273xO




LM8177519333500
     Second WRS-1 not required?
Path 59  Row 18
     64/18



     65/18
GLIS ALL  90's



86/09/06        LM85091919331xO




72/08/12        LM8102019480500




GLIS
check fiche
                                200

-------
Path 59  Row 19




               GLIS ALL 90's




               86/09/06       LM85091919333xO




     64/19     GLIS 0-50%




     65/19     72/09/18       LM8105719542500






Path 59  Row 20




               GLIS ALL 90'8




               86/09/06       GLIS for ID




     64/20     GLIS




     65/20     GLIS
Path 60  Row 19




               92/06/25




               86/06/25




     65/19     72/09/18
               LM85303819421xO




               LM85084619422xO




               LM8105719542500
     Second WRS-1 not required?
Path 61  Row 18
     66/18




     67/18
92/09/04




84/08/13




73/07/22




73/07/23
LM85310919470xO




LM85016519542xO




LM8136419595500




LM8136520053500
                               201

-------
Path 62  Row 17
     To be selected
Path 62  Row 18
     67/18



     68/18






Path 63  Row 17
92/08/10




86/09/11




73/07/23




72/09/21
LM85308419533xO




LM85092419513xO




LM8136520053500




LM8106020111500
     68/17




     69/17






Path 63  Row  18
GLIS ALL 90 's




84/09/12       LM85019520065xO




GLIS



74/07/02       LM8170920090500
      68/18




      69/18
GLIS ALL 90's




84/09/12       LM85019520071xO




72/09/21       LM8106020111500




72/09/22       LM8106120165500
Path 64   Row 16



               GLISS ALL 90's



                85/06/02       LM85045820123xO




      70/16     76/06/13       LM8250820042500




      Second WRS-1 not required?
                                    check fiche
                                202

-------
Path 64  Row 17




               GLIS ALL 90's




               85/06/02       LM85045820130xO




     70/17     74/06/15       LM8169220152500




     Second WRS-1 not required?






Path 64  Row 18




               GLIS ALL 90's




               86/09/09       LM85092220040xO




     70/18     73/09/18       LM8142220215500




     Second WRS-1 not required?




Path 65  Row 15




               92/06/28       LM85304120113xO




               86/06/28       LM85084920113xO




     71/15     72/09/24       LM8106320271500




     Second WRS-1 not required?
Path 65  Row 16
     71/16
92/06/28



86/06/28



75/07/08
LM85304120115xO




LM85084920120xO




LM8216720162500
                                   check fiche
     Second WRS-1 not required?
                               203

-------
Path. 65  Row  17
     71/17
92/06/28

86/06/28

76/07/20
GLIS for ID

LM85084920122xO

LM8254520093500
     Second WRS-1 not required?
Path 65  Row 18
     71/18
92/06/12

86/06/28

76/06/14
LM85302520125xO

GLIS for ID

LM8250920105500
     Second WRS-1 not required?
Path. 66  Row 14
     72/14

     73/14


Path 66  Row 15
     72/15

     73/15
92/07/05

86/07/05

76/08/26

72/08/21




92/07/05

86/08/22
86/09/07

76/08/26

72/08/21
GLIS for ID

LM85085620171xO

LM8258220125500

LM8102920381500




LM85304820173xO

LM85090420155xO
LM85092020151xO

LM8258220131500

LM8102920383500
                                                       OR
                               204

-------
Path 66  Row 16
     72/16
92/07/05

86/07/05

73/07/10
GLIS for ID

LM85085620175xO

LM8135220333500
     Second WRS-1 not required?
Path 66  Row 17
     72/17
92/07/05
     _/
86/07/05

75/07/09
LM85304820182xO

,LM85085620182xO

LM8216820223500
     Second WRS-1 not required?
Path 66  Row 18
               92/07/05

               86/07/05
               85/07/18
               GLIS for ID

               OR
     72/18     75/07/09       GLIS for ID and X or N

     Second WRS-1 not required?
Path 67  Row 13
     73/13

     74/13
92/08/29

84/09/24

76/08/27

76/07/23
LM85310320221xO

LM85020720295xO

LM8258320181500

LM8254820245500
                                205

-------
Path 67  Row 14
     73/14



     74/14






Path. 67  Row 15
 GLIS ALL  90' s




 84/09/24        LM85020720302xO




 72/08/21        LM8102920381500




 76/07/23        LM8254820252500
     73/15




     74/15






Path 67  Row 16
     73/16



     74/16






Path 67  Row 17
     73/17
 92/08/28




 86/09/14




 72/08/21




 76/08/01










 92/06/10




 86/09/14




 75/07/10




 76/08/01





r




 92/06/10




 86/07/28




 75/07/10
LM85310020230xO




LM85092720211xO




LM8102920383500




LM8547019553500











LM85302320242xO




LM85092720213xO




LM8216920275500




LM8547019560500











LM85302320245xO




LM8 5087920234x0




LM8216920281500
     Second WRS-1 not required?
                                206

-------
Path 67  Row 18
     73/18
92/08/29

86/08/13
86/07/28

75/07/09
                              LM85310320241xO

                              LM85089520233xO
                              LM85087920240xO
                    OR
GLIS for ID and X or N
     Second WRS-1 not required?
Path 68  Row 12
     74/12

     75/12


Path 68  Row 13
92/06/01

86/07/03
86/06/01

72/08/22

76/07/06
     74/13

     75/13


Path 68  Row  14
92/07/03

86/07/03
85/08/01

76/07/23

76/08/11
     74/14

     75/14
92/07/03

85/08/01

76/07/23

76/07/06
LM85301420291xO

LM85085420284xO
LM85082220294xO

LM8103020430500

LM8253120304500
LM85304620291xO

LM85085420291xO
LM85051820355xO

LM8254820245500

LM8256720300500
LM85304620293xO

LM85051820361xO

LM8254820252500

LM8253120313500
                                                  OR
                                                  OR
                                207

-------
Path 68  Row 15
     74/15

     75/15


Path 68  Row 16
     74/16

     75/16


Path 68  Row 17
     74/17

     75/17


Path 68  Row 18
92/07/03

86/09/05

76/08/01

76/07/06




92/07/03

86/07/03
86/09/05

76/08/01

76/07/06


r

92/07/03

86/09/05

75/07/11

76/07/06
     74/18
 92/07/03

 86/09/05
 84/08/14

 73/08/17
LM85304620295xO

LM85091820274xO

LM8547019553500

LM8253120320500




LM85304620302xO

LM85085420302xO     OR
LM85091820280xO

LM8547019560500

LM8253120322500




LM85304620304xO

LM85091820282xO

LM8217020340500

LM8253120325500




LM85304620311xO

LM85091820285xO     OR
LM85016620374xO

LM8139020452500
                                208

-------
Path 69  Row 11



               GLIS ALL 90's



               85/07/07       LM85049320412xO




     76/11     74/06/21       LM8169820470500




     Second WRS-1 not required?






Path 69  Row 12



               GLIS ALL 90's



               85/07/07       LM85049320414xO




     76/12     74/09/01       LM8177020450500




     Second WRS-1 not required?
Path  69  Row  13
      75/13



     ,.76/13






 Path 69   Row 14
      75/14




      76/14
92/09/04




84/09/22




76/08/11




75/08/18






t



92/09/04




84/09/22




76/07/06




76/09/17
LM84370320214xO




LM85020520422xO




LM8256720300500




LM8220820425500











LM84370320220xO




LM85020520424xO




LM8253120313500




LM8260420351500
check fiche
                                209

-------
Path 69  Row 15
     75/15




     76/15






Path. 69  Row 16
     75/16




     76/16






Path 69  Row 17
     75/17



     76/17
92/09/04




84/09/22




76/07/06




GLIS 0-50%











92/06/08




84/06/18




76/07/06




74/07/27






r




92/09/12




86/09/12




76/07/06




74/07/27
LM84370320223xO




LM85020520430xO




LM8253120320500
LM85302120365xO




LM85010920414xO




LM8253120322500




LM8173420480500










LM85311720355xO




LM85092520342xO




LM8253120325500




LM8173420482500
Path 69  Row 18



               92/06/08




               84/06/18




     75/18     GLIS  0-50%




     Second WRS-1  not  required?
                LM85302120374xO




                LM85010920423xO
                                                   check  fiche
                                210

-------
Path 69  Row 19




               GLIS ALL 90'8




               86/07/26       LM85087720365xO




     75/19     GLIS 0-50%




     Second WRS-1 not required?






Path 69  Row 20




               GLIS ALL 90's




               86/07/26       LM85087720372xO




     75/20     74/09/18       LM8178720415500




     Second WRS-1 not required?
Path 70  Row 11
     77/11
92/07/17



86/07/01



73/08/02
LM85306020403xO




LM85085220404xO




LM8137520595500
     Second WRS-1 not required?
Path 70  Row 12
     77/12
92/07/17




86/07/01




73/08/02
LM85306020405xO




LM85085220411xO




LM8137521002500
     Second WRS-1 not  required?
                                211

-------
Path 70  Row 13
     76/13



     77/13






Path 70  Row 14
     76/14



     77/14






Path 70  Row 15
     76/15




     77/15






Path 70  Row  16
92/07/17




86/07/17




75/08/18




76/09/18






I




92/07/01




84/08/12




76/09/17




74/09/02










92/08/02




84/08/12




76/06/01




74/09/02
LM85306020412xO




LM85086820410xO




LM8220820425500




LM8260520403500











LM85304420415x0




LM85016420483xO




LM8260420351500




LM8177120513500










LM85307620415xO




LM85016420485xO




LM8249620384500




LM8177120515500
     76/16




     77/16
 GLIS ALL 90'a



 86/07/01       LM85085220424xO




 74/07/27       LM8173420480500




 72/08/25       LM8103321020500
                                                  check fiche
                                212

-------
Path 70  Row 17
     76/17




     77/17
92/09/03



85/09/16




74/07/27




72/08/25
LM85310820421xO




LM85056420491xO




LM8173420482500




LM8103321022500
Path 70  Row 18
     76/18



     77/18






Path 70  Row 19
GLIS ALL 90'-s




85/09/16       LM85056420493xO




GLIS 0-50%




72/08/25       LM8103321025500
     76/19
92/06/07




84/08/12




74/07/27
LM84361420273xO




LM85016420502xO




LM8173420491500
     Second WRS-1 not required?






Path 70  Row 20




               GLIS ALL  90 's




               84/08/12       LM85016420505xO




     76/20     74/07/27       LM8173420494500




     Second WRS-1 not required?
                                213

-------
Path 71  Row 11
     78/11
 92/06/30

 86/07/08

 76/08/14
LM84363720311xO

LM85085920464xO

LM8257020462500
     Second WRS-1 not required?
Path 71  Row 12
               92/08/01
               92/07/16
                LM84366920324xO
                LM84365320321xO
               86/07/08

     78/12     GLIS 0-50%

     Second WRS-1 not required?
                LM85085920471xO
Path 71  Row 13
     77/13

     78/13


Path 71  Row 14
     77/14

     78/14
 92/08/01

 85/07/21

 76/09/18

 74/09/03


t

 92/08/01

 85/07/21

 74/09/02

 75/09/25
LM84366920330xO

LM85050720542xO

LM8260520403500

LM8177220565500




LM84366920332xO

LM85050720545xO

LM8177120513500

LM8224620542500
                    Composite with
                               214

-------
Path 71  Row 15
     77/15



     78/15






Path 71  Row 16
92/08/01




85/07/21




74/09/02




74/09/03
LM84366920335xO




LM85050720551xO




LM8177120515500




LM8177220574500
     77/16




     78/16
GLIS ALL 90 's




85/07/05       LM85049120554xO




72/08/25       LM8103321020500




73/07/16       LM8135821075500
Path 71  Row 17
      77/17




      78/17






 Path 71  Row 18
GLIS ALL 90's



84/09/04       LM85018720561xO




72/08/25       LM8103321022500




74/09/03       LM8177220583500
      77/18




      78/17
 92/06/30




 GLIS  80's




 76/06/20




 73/07/16
 LM84363720340xO








 LM8251520451500




 LM8135821082500
                                                   CHECK FICHE
                                215

-------
Path. 71  Row 19
     77/19




     78/19






Path, 71  Row 20
     77/20




     78/20
GLIS ALL  90 's




85/07/05        LM85049120565xO




76/07/08        LM8253320450500




76/07/09        LM8253420504500






)




GLIS ALL  90's



84/09/04        LM85018720572xO




76/07/08        LM8253320453500




76/07/09        LM8253420511500
Path 72  Row 11
     79/11
 92/07/23




 86/07/15




 74/09/04
LM84366020381xO




LM85086620524xO




LM8177321014500
     Second WRS-1 not  required?
Path 72  Row 12



               92/06/05




               85/06/10




     79/12     GLIS  0-50%




     Second  WRS-1  not  required?
                LM84361220370xO




                LM85046621002xO
                                216

-------
Path 72  Row 13
     79/13
92/06/05




86/06/13




73/06/29
LM84361220372xO




LM85083420543xO




LM8134121123500
     Second WRS-1 not required?
Path 72  Row 14
     78/14




     79/14






Path 72  Row 15
92/06/05




86/06/13




75/09/25




73/06/29
LM84361220375xO




LM85083420545xO




LM8224620542500




LM8134121130500
     78/15




     79/15






Path 72  Row 16
GLIS ALL 90's




86/06/13       LM85083420552xO




73/07/16       LM8135821073500




73/06/29       LM8134121132500
      78/16




      79/16
92/06/05




86/06/13




73/07/16




73/06/29
LM84361220383xO




LM85083420554xO




LM8135821075500'




LM8134121135500
                                217

-------
Path 72  Row 17
     78/17



     79/17






Path 72  Row 18
92/06/05




86/06/13




73/07/16




73/06/29
LM84361220390xO




LM85083420560xO




LM8135821082500




LM8134121141500
     78/18




     79/18






Path 72  Row 19
     78/19




     79/19
GLIS ALL  90's




86/05/28        LM85081820570xO




76/07/09        LM8253420502500




76/07/10        LM8253520560500






)




GLIS ALL  90's




86/05/28        LM85081820572xO




76/07/09        LM8253420504500




GLIS  0-50%
Path 72  Row 20
      78/20




      79/20






Path 72   Row 21
 GLIS ALL 90's



 86/05/28       LM85081820575xO




 76/07/09       LM8253420511500




 74/07/30       LM8173721064500






      To be Selected
                                218

-------
Path 73  Row 11
     80/11




     81/11
92/08/31




85/08/04




74/06/25




76/07/12
LM84369920453xO




LM85052121055xO




LM8170221095500




LM8253721043500
Path 73  Row 12




               GLIS ALL 90's




               85/08/04       LM85052121061xO




     80/12     74/09/05       LM8177421074500




     Second WRS-1 not required?






Path 73  Row 13




               GLIS ALL 90's




               86/08/23       LM85090520581xO




     80/13     73/06/30       LM8134221182500




     Second WRS-1 not required?






Path 73  Row 14




               GLIS ALL 90's




               85/07/19       LM85050521071xO




     79/14     73/06/29       LM8134121130500




     80/14     76/07/11       LM8253621000500
                                219

-------
Path 73  Row 15



               GLIS ALL 90 's



               84/09/02       LM85018521074xO




     79/15     74/09/04       LM8177321032500




     80/15     75/09/09       LM8223021063500






Path 73  Row 16



               GLIS ALL 90's



               84/09/02       LM85018521081xO




     79/16     74/09/04       LM8177321034500




     80/16     72/08/10       LM8101821191500
Path 73  Row  17
     79/17




     80/17






Path 73   Row 1.8
      79/18




      80/18
GLIS ALL  90's




84/09/02        LM85018521083xO




73/06/29        LM8134121141500




72/08/10        LM8101821193500







3




GLIS ALL  90's




84/09/02        LM85018521090xO




76/07/10        LM8253520560500




74/08/18        LM8175621110500
                                220

-------
Path 73  Row 19




               GLIS ALL 90's



               84/09/02       LM85018521092xO




     79/19     GLIS 0-50%




     80/19     GLIS Q-50%






Path 73  Row 20




               GLIS ALL 90's



               84/09/02       LM85018521094xO




     79/20     74/07/30       LM8173721064500




     80/20     72/09/15       LM8105421205500






Path 73  Row 21



               GLIS ALL 90's



               84/09/02       LM85018521101xO




     79/21     GLIS 0-50%




     80/21     GLIS 0-50%






Path 74  Row 10



               GLIS ALL  90's



               85/08/27       LM85054421112xO




     81/10     74/09/06       LM8177521124500




     82/10     GLIS
                                221

-------
Path 74   Row 11
      81/11



      82/11






Path  74  Row 12
 GLIS ALL 90's




 85/08/27       LM85054421115xO




 76/07/12       LM8253721043500




 GLIS
     81/12




     82/12






Path 74  Row 13
     81/13
 92/09/07




 85/07/26




 74/09/06




 GLIS






5




 92/09/07




 85/07/26




 76/07/12
LM84370620521xO




LM85051221123xO




LM8177521133500
LM84370620524xO




LM85051221125xO




LM8253721052500
     Second WRS-1 not required?
Path 74  Row 14
     81/14
92/09/07




85/07/26




76/07/12
LM84370620530xO




LM85051221132xO




LM8253721055500
     Second WRS-1 not required?
                               222

-------
Path 74  Row 15
     81/15
92/09/07




85/07/26




72/08/29
LM84370620533xO




LM85051221134xO




LM8103721243500
     Second WRS-1 not required?
Path 74  Row 16
     80/16




     81/16






Path 74  Row 17
     80/17




     81/17






Path 74  Row  18
92/09/07




85/07/26




72/08/10




75/09/28






r




92/06/03




86/05/26




72/08/10




GLIS  0-50%
LM84370620535xO




LM85051221141xO




LM8101821191500




LM8224921123500
LM84361020512xO




LM85081621090xO




LM8101821193500
     80/18




     81/18
 GLIS  ALL 90's




 86/05/26       LM85081621092xO




 74/08/18       LM8175621110500




 75/08/05       LM8219521140500
                                223

-------
Path. 74  Row 19
     80/19



     81/19
92/08/06




GLIS 80's




74/08/18




75/08/05
LM84367420541xO








GLIS for ID and X or N




LM8219521142500
Path 74  Row 21     To be  selected
Path. 74  Row 22      To be  selected
Path. 75  Row 10



               GLIS  ALL  90's



               86/07/04        LM85085521111xO




     82/10     GLIS



     83/10     74/07/16        LM8172321260500
Path. 75  Row 11
      82/11



      83/11
92/07/28




85/08/02




GLIS



74/07/16
LM84366520565xO




LM85051921181xO








LM8172321262500
                                224

-------
Path 75  Row 12
     82/12

     83/12
92/07/28

85/07/17

GLIS

76/08/01
LM84366520571xO

LM85050321184xO



LM8255721155500
check fiche
Path 75  Row 13

               GLIS ALL 90's

               86/07/04       LM85085521122xO

     82/13     GLIS

     Second WRS-1 not required?
Path 75  Row 14

               92/07/12

               86/07/04

     82/14     GLIS

     Second WRS-1 not  required?
               LM84364920573xO

               LM85085521124xO
Path 75  Row 15
                92/07/28
                92/07/12
               LM84366520582xO
               LM84364920575xO
                    Composite with
                85/08/02

      82/15      GLIS

      Second WRS-1 not required?
               LM85051921195xO
                                225

-------
Path. 75  Row IS



                GLZS ALH 9d's



                85/08/02       LM850.51921201xO




     82/16      GLIS




     Second  WRS-1 not required?
Path 75  Row IT
     81/17




     82/17






Path 75  Row 18
          CLOTS ALL 50's




          85/0 8/02       LM8 5 0 51921204x0




          GLXS a-50%




          GILXS
     81/18




     82/18






Path 75  Row
          GLIS ALE.




          85/08/02




          75/08/05




          GIrTS
                               LM8 50 5192 12 10x0




                               LM8219521140500
81/19




82/19
                     ALL 9:0 's



                a4/09/16       LM8 5 019 9 2 12 14x0




                7e/a5/0€       LM8 2 531210 62 5 00.




                GLIS
                                                    check fiehe
Path 75  Row 22     To be selected

-------
Path 76  Row 10
     83/10



     84/10






Path 76  Row 11
GLIS ALL 90's




86/08/12       LM85089421160xO




74/07/16       LM8172321260500




74/08/22       LM8176021302500
     83/11




     84/11
92/06/17




84/09/23




74/07/16




74/08/22
LM84362421014xO




LM85020621244xO




LM8172321262500




LM8176021305500
Path 76  Row  12




               GLIS ALL  90's



               86/09/13        LM85092621154xO




     83/12     74/09/08        LM8177721245500




     84/12     72/09/19        LM8105821403500






Path 76  Row  13




               GLIS ALL  90's



               84/06/03        LM85009421233xO




     83/13     74/08/03        LM8174121264500




     84/13     75/07/21        LM8218021293500
                                    check  fiche
                                227

-------
Path. 76  Row 14

               GLIS ALL 90 's

               84/08/06       LM85015821252xO

     83/14     74/08/03       LM8174121270500

     Second WRS-1 not required?
Path 76  Row 15
               GLIS ALL 90's

               85/05/21
               84/06/03
               LM85044621262xO
               LM85009421242xO
     83/15     76/08/19       LM8257521163500

     Second WRS-1 not required?
Path. 76  Row 16
     83/16
92/06/17

85/07/24

76/08/19
LM84362421034xO

LM85051021263xO

LM8257521170500
     Second WRS-1 not  required?
Path 76  Row  17
     82/17

     83/17
92/06/17

85/07/24

GLIS

75/06/14
LM84362421040xO

LM85051021265xO



LM8214321254500
                    OR
                                228

-------
Path 76  Row 18




               GLIS ALL 90's




               84/06/03       LM85009421253xO




     82/18     GLIS




     83/18     GLIS 0-50%






Path 76  Row 19




               GLIS ALL 90' 8,




               84/06/03       LM85009421255xO




     81/19     75/08/05       LM8219521142500




     82/19     GLIS






Path 76  Row 22     To be selected






Path 77  Row 10



               GLIS ALL 90's



               84/06/26       LM85011721285xO




     84/10     74/08/22       LM8176021302500




     85/10     73/06/17       LM8132921455500






Path 77  Row 11



               GLIS ALL  90's



               84/08/29       LM85018121305xO




     84/11     74/08/22       LM8176021305500




     85/11     75/07/22       LM8218121342500
                                229

-------
Path 77  Row 12
     84/12



     85/12






Path 77  Row 13
     84/13



     85/13
92/09/28




84/08/29




72/09/19




73/06/17






I




92/09/28




86/07/02




75/07/21




75/07/22
LM84372721112xO




LM85018121311xO




LM8105821403500




LM8132921464500











LM84372721114xO




LM85085321245xO




LM8218021293500




LM8218121351500
                                                  check fiche
Path 77  Row 14



               GLIS ALL 90's



               86/07/02       LM85085321251xO




     84/14     75/07/03       LM8216221301500




     Second WRS-1 not  required?






Path 77  Row 15



               GLIS ALL 90's



               86/07/02       LM85085321253xO




     84/15     75/07/03       LM8216221304500




     Second WRS-1 not  required?
                                230

-------
Path. 77  Row 16




               GLIS ALL 90 's



               86/07/02       LM85085321260xO




     84/16     76/06/27       LM8252221242500




     Second WRS-1 not required?






Path 77  Row 17



               GLIS ALL 90's



               84/08/29       LM85018121331xO




     84/17     75/07/21       LM8218021311500




     Second WRS-1 not required?
Path 77   Row 18
      84/18
92/06/24




84/06/26




75/08/08
LM84363121105x0




LM85011721320xO




LM8219821311500
      Second WRS-1 not required?
 Path 77   Row 23
 Path 78  Row 10
      85/10




      86/10
     To be selected











GLIS ALL 90's



84/07/03       LM85012421351xO




73/06/17       LM8132921455500




GLIS
                                231

-------
Path. 78  Row 11
     85/11


     86/13




Path 78  Row 12
92/06/15


84/07/03
   -.

73/06/17


GLIS
LM84362221140xO


LM85012421353xO


LM8132921462500
     85/12


     86/12




Path. 78  Row 13
GLIS ALL 90's


85/07/06       LM85049221372xO


73/06/17       LM8132921464500


GLIS
     85/13


     86/13




Path 78  Row 14
92/09/03


85/07/06


75/07/22


GLIS
LM84370221172xO


LM85049221374xO


LM8218121351500
     85/14


     86/14
GLIS ALL 90's


85/07/22       LM85050821380xO


75/07/04       LM8216321355500



GLIS
                                232

-------
Path 78  Row 15




               GLIS ALL 90's



               84/06/17       LM85010821365xO




     85/15     73/06/17       GLIS for ID and X or N




     Second WRS-1 not required?






Path 78  Row 16




               GLIS ALL 90's




               84/06/17       LM85010821372xO




     85/16     76/06/10       LM8250521303500




     Second WRS-1 not required?






Path 78  Row 17




               GLIS ALL 90's



               84/06/01       LM85009221373xO




     85/17     76/06/10       LM8250521305500




     Second WRS-1 not required?






Path 78  Row 18



               GLIS ALL 90's



               84/06/01       LM85009221375xO




     85/18     76/06/10       LM8250521312500




     Second WRS-1 not required?






Path 78  Row 23     To be selected
                                233

-------
Path 79  Row 10




               GLIS ALL 90'8




               86/06/30       LM85085121360xO




     86/10     GLIS




     87/10     75/07/06       LM8216521454500
Path 79  Row 11




               92/07/08




               86/06/14




     86/11     GLIS




     87/11     75/07/06




Path 79  Row 12




               92/06/06




               86/06/14




     86/12     GLIS




     87/12     75/07/06
Path 79  Row 13
     86/13




     87/13
92/06/06




86/06/14




GLIS




74/07/02
LM84364521205xO




LM85083521365xO








LM8216521461500








LM84361321202xO




LM85083521372xO








LM8216521463500










LM84361321204xO




LM85083521374xO








LM8170921504500
                               234

-------
Path 79  Row 14
     86/14




     87/14






Path 79  Row 15
     86/15




     87/15
92/06/06




86/06/30




GLIS




74/07/02











92/06/06




86/06/30




GLIS




74/07/02
LM84361321210xO




LM85085121374xO








LM8170921510500
LM84361321213xO




LM8508512l380xO








LM8170921513500     check X or N
Path 79  Row 16




               92/06/06




               84/08/27




     86/16     GLIS




     Second WRS-1 not required?
               LM84361321215xO




               LM85017921451xO
Path 79  Row 17




               92/09/26




               84/09/28




     86/17     GLIS




     Second WRS-1 not required?
               LM84372521254xO




               LM85021121454xO
                               235

-------
Path 79  Row 18



               GLIS ALL 90's



               84/09/28       LM85021121460xO




     86/18     GLIS



     Second WRS-1 not required?
Path 80  Row 10
     87/10



     88/10






Path 80  Row 11
      87/11



      88/11






Path 80   Row 12
      87/12




      88/12
92/07/15




84/07/01




75/07/06




76/08/06











92/07/15




86/07/07




75/07/06




74/07/03





>




92/07/31




86/07/07




75/07/06




GLIS  0-50%
LM84365221265xO




LM85012221473xO




LM8216521454500




LM8256221433500











LM84365221272xO




LM85085821422xO




LM8216521461500




LM8171021553500










LM84366821281xO




LM85085821424xO




LM8216521463500
                                236

-------
Path 80  Row 13
     87/13




     88/13






Path 80  Row 14
     87/14




     88/14






Path 80  Row 15
               92/07/31




               84/07/01




               74/67/02




               75/07/25






               t




               92/07/31




               84/07/01




               74/07/02




               74/07/03
     87/15




     88/15
               92/07/31




               84/07/01




               74/07/02




               74/07/03
LM84366821284xO




LM85012221484xO




LM8170921504500




LM8218421522500
LM84366821290xO




LM85012221490xO




LM8170921510500




LM8171021565500











LM84366821292xO




LM85012221493xO




LM8170921513500




LM8171021571500
                               LM84364321325xO
Path 81  Row 10



               92/07/06




               GLIS



     89/10     76/08/25       LM8258121484500




     Second. WRS-1 not required?
                                                   check X or N
                                237

-------
Path 81  Row 11
     89/11
 92/07/06

 84/06/06

 72/08/01
LM84364321331xO

LM85009721534xO

LM8100922083500
     Second WRS-1 not required?
Path 81  Row 12
     88/12

     89/12


Path 81  Row 13
     88/13

     89/13


Path 81  Row 14
 92/07/06

 84/06/06

 GLIS  0-50%

 72/08/01


!

 92/06/20

 84/06/06

 75/07/25

 72/08/01
     88/14

     89/14
 92/07/06
 92/07/22

 84/06/06

 74/07/03

 72/08/01
LM84364321334xO

LM85009721540xO



LM8100922090500




LM84362721333xO

GLIS for ID

LM8218421522500

LM8100922092500
LM84364321334xO
LM84365921345xO

GLIS for ID

LM8171021565500

LM8100922095500
                                                  Composite with
                               238

-------
Path 81  Row 15



               GLIS ALL 90's




               84/06/06       GLIS for ID




     88/15     74/07/03       LM8171021571500




     Second WRS-1 not required?






Path 82  Row 10



               GLIS ALL 90's



               84/09/01       LM85018422012xO




     90/10     72/08/02       LM8101022135500




     Second WRS-1 not required?
Path  82   Row 11
      90/11
92/07/29




84/06/29




75/08/14
LM4366621401xO




LM85012022001xO




LM8220422022500
      Second WRS-1 not required?
 Path 82  Row 12
      89/12



      90/12
92/07/29




84/06/29




72/08/01




72/08/02
LM84366621403xO




LM85012022004xO




LM8100922090500




LM8101022144500
                                239

-------
Path. 82  Row 13
     89/12

     90/12


Path 82  Row 14
92/07/13

84/07/15

72/08/01

72/08/02
     89/14

     90/14


Path 83  Row 12
92/07/29
92/07/13

84/07/15

72/08/01

72/08/02
     90/12

     91/12
92/07/04

84/09/08

72/08/02

73/06/23
LM84365021402xO

LM85013622011xO

LM8100922090500

LM8101022144500
LM84366621412xO
LM84365021405xO

LM85013622014xO

LM8100922095500

LM8101022153500
LM84364121455xO

LM85019122082xO

LM8101022144500

LM8133522210500
Path 83  Row 13

               GLIS ALL  90's

               84/09/08        LM85019122085xO

     90/13     74/08/10        LM8174822063500

     Second WRS-1 not  required?
                                                  Composite with
                                240

-------
Path 83  Row 14




               GLIS ALL 90's




               84/06/04       LM85009522071xO




     90/14     72/08/02       LM8101022153500
                               241

-------
             NALC MSS Triplicates - Western USA
Path 33  Row 33
     35/33



     36/33






Path 33  Row 34
     35/34



     36/34






Path 33  Row 35
     35/35



     36/35






Path 33  Row 36
     35/36




     36/36
92/07/06




86/07/30




74/07/04




75/06/30






I




92/07/06




86/06/28




74/07/04




75/06/30










92/07/06




86/06/12




75/06/29




74/06/17










92/08/07




86/07/30




74/07/04




72/08/02
LM84364316452xO




LM85088116595xO




LM8171117004500




LM8507216511500











LM84364316454xO




LM85084917012xO




LM8171117010500




LM8507216514500










LM84364316461xO




LM85083317021xO




LM8507116462500




LM8169417074500










LM84367516473xO




LM85088117010xO




LM8171117015500




LM8101017145500
                               242

-------
Path 33  Row 37
     35/37

     36/37


Path 34  Row 31
 92/08/07

 86/07/30

 73/06/21

 72/08/02
     36/31

     37/31


Path 34  Row 32
 92/07/21
 92/07/05

 85/07/02

 74/07/23

 75/07/01
     36/32

     37/32


Path 34  Row 33
     36/33

     37/33
 92/07/05

 85/06/16

 73/07/10

 75/07/01


\

 92/07/05

 83/07/05

 75/06/30

 75/07/01
LM84367516475xO

LM85088117013xO

LM8133317102500

LM8101017152500
LM85306417055xO
LM84304817061xO

LM85048817130xO

LM8173017045500

LM8507316561500
LM85304817063xO

LM85047217133xO

LM8135217134500

LM8507316563500




LM85304817065xO

LM84035417134xO

LM8507216511500

LM8507316570500
                                                  Composite with
fiche missing
                               243

-------
Path 34  Row 34
     36/34



     37/34






Path 34  Row 35
     36/35




     37/35






Path 34  Row 36
     36/36




     37/36






Path 34  Row 37
92/07/05




83/07/05




75/06/30




75/07/01











92/09/07




86/08/22




75/09/19




73/09/21










92/09/07




86/08/22




72/08/02




73/08/16
     36/37




     37/37
92/07/05




85/07/02




73/06/22




75/06/22
LM85304817072xO




LM84035417140xO




LM8507216514500




LM8507316572500











LM85311217065xO




LM85090417060xO




LM8224017022500




LM8142517193500










LM85311217071xO




LM85090417062xO




LM8101017145500




LM8138917204500










LM85304817083xO




LM85048817152xO




LM8133417160500




LM8215117101500
                                244

-------
Path 35  Row 31
     37/31

     38/31


Path 35  Row 32
     37/32


     38/32
 92/06/02

 85/06/07

 75/07/01

 73/06/06


i

 92/08/13

 86/07/28

 74/08/11
 75/07/01

 74/08/30
Path 35  Row 33

               92/06/18

               86/07/28

     37/33     75/07/01

     38/33     73/06/06


Path 35  Row 34

               92/06/18

               85/06/23

     37/34     75/07/01

     38/34     75/06/05
LM84360916555xO

LM85046317192xO

LM8507316561500

LM8131817251500
LM85308717121xO

LM85087917115xO

LM8174917102500• OR (Fiche missing)
LM8507316563500

LM8176817152500
               LM84362516570x0

               LM85087917122xO

               LM8507316570500

               LM8131817260500




               LM84362516572xO

               LM85046317203xO

               LM8507316572500

               LM8213417143500
                               245

-------
Path 35  Row 35
     37/35




     38/35






Path 35  Row 36
     37/36




     38/36






Path 35  Row 37
     37/37




     38/37






Path 36  Row 29
     38/29




     39/29
 92/06/18




 85/06/23




 73/06/05




 73/06/06











 92/06/02




 85/07/09




 73/06/05




 75/06/05






r




 92/06/02




 86/06/10




 73/06/05




 74/06/01






)




 92/07/27




 86/07/19




 74/07/25




 74/07/26
LM84362516575xO




LM85047917205xO




LM8131717210500




LM8131817265500











LM84360916575xO




LM85049517211xO




LM8131717213500




LM8213417152500











LM84360916581xO




LM85083117153xO




LM8131717215500




LM8167817203500











LM84366417025xO




LM85087017171xO




LM8173217153500




LM8173317211500
                                246

-------
Path 36  Row 30




               92/07/27




               86/07/19




     38/30     73/07/30




     39/30     72/08/05






Path 36  Row 31




               92/08/12




               86/08/17




     38/31     74/07/25




     39/31     72/08/05






Path 36  Row 32




               92/08/28




               85/08/17




     38/32     74/08/30




     39/32     73/09/05






Path 36  Row 33




               92/08/20




               86/08/04




     38/33     73/09/04




     39/33     73/09/05
LM84366417032xO




LM85087017173xO




LM8137217240500




LM8101317294500











LM84368017041xO




LM85053417250xO




LM8173217162500




LM8101317300500











LM84369617050xO




LM85053417252xO




LM8176817152500




LM8140917300500











LM85309417184xO




LM85088617181xO




LM8140817244500




LM8140917303500
fiche missing
                               247

-------
Path 36  Row 34




               92/08/28




               86/09/05




     38/34     73/09/04




     39/34     74/08/31






Path 36  Row 35




               92/09/05




               86/09/05




    , 38/35     73/09/04




     39/35     72/09/10
Path 36  Row 36
     38/36




     39/36






Path 36  Row 37
     38/37




     39/37
 92/07/19




 85/06/30




 74/07/25




 75/06/24





r




 92/07/19




 85/06/14




 75/06/23




 73/06/25
LM84369617055xO




LM85091817172xO




LM8140817251500




LM8176917215500










LM85311017191xO




LM85091817175xO




LM8140817253500




LM8104917315500











LM85306217202xO




LM85048617272xO




LM8173217182500




LM8215317211500










LM85306217204xO




LM85047017275xO




LM8215217155500




LM8133717332500
                                    fiche missing
fiche missing
                               248

-------
Path 37  Row 29




               92/08/19




               86/08/27




     39/29     72/09/05




     40/29     73/08/01






Path 37  Row 30




               92/08/27




               86/08/27




     39/30     72/08/05




     40/30     72/08/06






Path 37  Row 31




               92/07/26




               85/08/08




     39/31     72/08/05




     40/31     72/08/06






Path 37  Row 32




               92/07/26




               85/08/08




     39/32     73/09/05




     40/32     72/08/06
LM84368717095xO




LM85090917220xO




LM8140917285500




LM8137417351500
LM85310117233xO




LM85090917222xO




LM8101317294500




LM8101417352500
LM85306917242xO




LM85052517312xO




LM8101317300500




LM8101417355500
LM85306917245xO




LM85052517312xO




LM8140917300500




LM8101417361500
                               249

-------
Path 37  Row 33
     39/33



     40/33






Path. 37  Row 34
     39/34



     40/34






Path 37  Row 35
     39/35



     40/35






Path 37  Row. 36
     39/36




     40/36
 92/08/27




 85/08/24




 73/09/05




 72/08/06






t




 92/08/27




 85/08/24




 74/08/31




 72/08/06










 92/08/27




 85/08/24




 72/09/10




 72/08/24










 92/08/27




 86/07/26




 72/08/23




 72/08/24
LM85310117244xO




LM85054117315xO




LM8140917303500




LM8101417364500











LM85310117251xO




LM85054117322xO




LM8176917215500




LM8101417370500











LM85310117253xO




LM85054117324xO




LM8104917315500




LM8103217373500











LM85310117255xO




LM85087717255xO




LM8103117322500




LM8103217375500
                               250

-------
Path 37  Row 37




               92/07/26




               86/07/26




     39/37     73/06/25




     40/37     75/07/31






Path 38  Row 27




               92/08/02




               86/08/02




     40/27     73/08/01




     41/27     72/08/25






Path 38  Row 28




               92/07/17




               86/08/02




     40/28     73/08/01




     41/28     75/08/10






Path 38  Row 29




               92/08/26




               86/08/02




     40/29     73/08/01




     41/29     72/08/07
LM85306917265xO




LM85087717262xO




LM8133717332500




LM8219017265500











LM85307617285xO




LM85088417281xO




LM8137417342500




LM8103317400500











LM85306017293xO




LM85088417284xO




LM8137417344500




LM8511317152500











LM84369417161xO




LM85088417290xO




LM8137417351500




LM8101517404500
                               251

-------
Path 38  Row 30
     40/30



     41/30






Path 38  Row 31
     40/31



     41/31






Path 38  Row 32
     40/32



     41/32






Path 38  Row 33
     40/33




     41/33
92/08/26




85/08/15




72/08/06




72/08/07










92/07/17




86/08/02




72/08/06




72/08/07





i




92/07/17




86/08/02




72/08/06




72/08/07






5




92/08/26




85/08/15




72/08/06




72/08/07
LM84369417163xO




LM85053217370xO




LM8101417352500




LM8101517410500










LM85306017304xO




LM85088417295xO




LM8101417355500




LM8101517413500










LM85306017311xO




LM85088417301xO




LM8101417361500




LM8101517415500










LM84369417174xO




LM85053217381xO




LM8101417364500




LM8101517422500
                                252

-------
Path 38  Row 34
     40/34



     41/34






Path 38  Row 35
     40/35




     41/35






Path 38  Row 36
     40/36




     41/36






Path 39  Row 27
     42/27
 92/08/18




 86/08/02




 72/08/06




 72/08/07











 92/07/01




 86/07/17




 74/07/09




 74/06/22











 92/08/02




 85/08/15




 72/08/24




 72/08/07






r




 91/07/06




 85/06/19




 74/06/23
LM85309217312xO




LM85088417310xO




LM8101417370500




LM8101517424500
LM85304417323xO




LM85086817320xO




LM8171617300500




LM8169917362500
LM85307617323xO




LM85053217392xO




LM8103217375500




LM8101517433500











LM85268317353xO




LM85047517422xO




LM8170017384500
     Second WRS-1 not required?
                               253

-------
Path 39  Row 28
     42/28
91/07/22




85/07/21




73/07/16
LM85269917361xO




LM85050717424xO




LM8135817462500
     Second WRS-1 not required?
Path 39  Row 29
     42/29
92/08/09




85/07/05




75/08/11
LM85308317355xO




LM85049117431xO




LM8511417213500
     Second WRS-1 not required?
Path 39  Row 30
     41/30



     42/30






Path 39  Row 31
     41/31



     42/31
92/07/16




85/07/05




72/08/07




73/07/16










92/07/16




86/06/22




74/06/22




74/06/23
LM84365317213xO




LM85049117433xO




LM8101517410500




LM8135817471500










LM84365317215xO




LM85084317372xO




LM8169917344500




LM8170017402500
                               254

-------
Path 39  Row 32
     41/32




     42/32
92/06/22




85/06/19




73/06/27




74/06/23
Path 39  Row 33




               92/08/01




               85/08/22




     41/33     72/08/07




     42/33     74/08/16






Path 39  Row 34




               92/08/09




               86/08/09




     41/34     72/08/07




     42/34     73/07/16






Path 39  Row 35




               92/09/10




               86/09/10




     41/35     72/08/07




     42/35     72/09/13
LM85303517373xO




LM85047517442xO




LM8133917423500




LM8170017404500
               LM84366917231xO




               LM85053917442xO




               LM8101517422500




               LM8175417392500











               LM85308317375xO




               LM85089117370xO




               LM8101517424500




               LM8135817485500











               LM85311517374xO




               LM85092317361xO




               LM8101517431500




               LM810521749050Q
                               255

-------
Path 39  Row 36
     41/36

     42/36
          92/09/10

          86/09/10

          72/08/07

          72/09/13
LM85311517380xO

LM85092317363xO

LM8101517433500

LM8105217493500
Path 40  Row 27
     43/27
          92/07/31

          86/07/31

          72/08/27
LM85307417412xO

LM85088217404xO

LM8103517513500
     Second WRS-1 not required?
Path 40  Row 28
     43/28
          92/07/15

          86/07/31

          73/07/17
LM85305817415xO

LM85088217410xO

LM8135917521500
     Second WRS-1 not required?
Path 40  Row 29
               92/07/31

               86/07/31
                         LM85307417420xO

                         LM85088217413xO
               75/07/25
               74/07/12
43/29


Second WRS-1 not required?
LM8509717281500
LM8171917443500
OR
fiche missing
                                256

-------
Path 40  Row 30
     43/30
92/07/31




86/07/31




74/07/12
LM85307417423xO




LM85088217415xO




LM8171917450500
     Second WRS-1 not required?
Path 40  Row 31
     43/31
92/07/07




86/07/31




74/07/12
LM84364417275xO




LM85088217422xO




LM8171917452500
     Second WRS-1 not required?
Path 40  Row 32
     43/32
92/07/23




86/07/31




74/07/12
LM84366017284x0




LM85088217424xO




LM8171917455500
     Second WRS-1 not required?
Path 40  Row 33
     43/33
92/07/31




86/07/31




74/07/12
LM85307417434xO




LM85088217430xO




LM8171917461500
     Second WRS-1 not required?
                               257

-------
Path 40  Row 34
     43/34
92/07/23

86/07/31

74/07/12
LM84366017293xO

LM85088217433xO

LM8171917464500
     Second WRS-1 not required?
Path 40  Row 35
     42/35

     43/35
92/06/05

86/06/13

74/06/23

73/06/29
LM84361217282xO

LM85083417453xO

LM8170017420500

LM8134117551500
Path 40  Row 36
     42/36

     43/36


Path 41  Row 26
92/06/21

85/06/10

74/06/23

73/06/29
LM84362817291xO

LM85046617521xO

LM8170017422500

LM8134117554500
                92/07/22       LM85306517471xO
                GLIS  for composite

                86/08/07       LM85088917461xO

     44/26      72/08/10       LM8101817565500

     Second WRS-1 not  required?
                                   Composite with
                                258

-------
Path 41  Row 27
     44/27
92/07/22
92/06/20

86/08/07

72/08/28
LM85306517473xO
LM85303317480xO

LM85088917461xO

LM8103617571500
                                                  Composite with
     Second WRS-1 not required?
Path 41  Row 28
     44/28
92/07/22
92/07/14
92/06/20

86/07/19

74/07/13
LM85306517480xO  Composite 2 of 3
LM84365117330x0
LM85303317482xO

LM85050517550xO

LM8172017495500
     Second WRS-1 not required?
Path 41  Row 29
     44/29
92/07/14

85/07/03

74/07/13
LM84365117332xO

LM85048917553xO

LM8172017502500
     Second WRS-1 not required?
Path 41  Row 30
     44/30
92/07/14

85/07/03

74/07/13
LM84365117335xO

LM85048917555xO

LM8172017504500
     Second WRS-1 not required?
                               259

-------
Path. 41  Row 31
     44/31
92/07/14




85/07/03




74/07/13
LM84365117341xO




LM85048917562xO




LM8172017511500
     Second WRS-1 not required?
Path 41  Row 32
     44/32
92/07/22




85/07/03




74/07/13
LM85306517493xO




LM85048917564xO




LM8172017513500
     Second WRS-1 not required?
Path 41  Row 33
     44/33
92/07/22




85/07/03




73/06/30
LM85306517500xO




LM85048917570xO




LM8134218001500
     Second WRS-1 not required?
Path 41  Row 34
     44/34
92/07/22




85/07/03




73/06/30
LM85306517502xO




LM85048917573xO




LM8134218003500
     Second WRS-1 not required?
                               260

-------
Path 41  Row  35
     44/35
92/07/22




85/07/03




73/06/30
LM85306517505xO




LM85048917575xO




LM8134218010500
     Second WRS-1 no required?
Path 41  Row 36
     44/36
92/07/06




85/07/03




74/06/25
LM85304917512xO




LM85048917582xO




LM8170217535500
     Second WRS-1 not required?
Path 41  Row 37
     44/37
92/07/06




85/07/03




74/06/25
LM85304917515xO




LM85048917584xO




LM8170217541500
     Second WRS-1 not required?
Path 42  Row 26
     45/26




     46/26
92/08/14




85/07/26




74/07/14




75/07/10
LM85308817530xO




LM85051218002xO




LM8172117544500




LM8508217450500
                               261

-------
Path 42  Row 27
     45/27

     46/27
92/06/27

85/07/10

74/07/14

75/07/10
73/08/07
Path 42  Row 28
     45/28

     46/28
90/07/08

85/07/10

74/06/26

75/07/10
73/08/07
72/08/12
Path 42  Row 29
     45/29

     46/29


Path 42  Row 30
     45/30

     46/30
92/07/29

85/06/24

72/08/11

72/08/12


)

92/07/13

86/07/13

74/06/26

74/06/27
LM85304017540xO

LM85049618005xO

LM8172117551500

LM8508217453500
LM8138018084500
LM85232017512xO

LM85049618011xO

LM8170317561500

LM8508217455500
LM8138018090500
LM8102018085500
LM85307217543xO

LM85048018014xO

LM8101918035500

LM8102018092500




LM85305617550xO

LM85086417545xO

LM8170317570500

LM8170418024500
OR f. missing
fiche missing
OR
OR
all fiche missing
                                262

-------
Path 42  Row 31
     45/31

     46/31


Path 42  Row 32
     45/32


     46/32


Path 42  Row 33
 92/07/21

 86/07/13

 74/06/26

 74/06/27


>

 92/07/29

 86/07/29

 73/08/06
 74/06/26

 72/07/25
     45/33

     46/33


Path 42  Row 34
     45/34

     46/34
 92/07/29

 86/07/29

 75/08/05

 72/07/25


t

 92/07/29

 86/07/29

 75/08/05

 72/07/25
LM84365817404xO

LM85086417552xO

LM8170317572500

LM8170418031500
LM85307217554xO

LM85088017551xO

LM8137918050500
LM8170317575500

LM8100218125500
LM85307217560xO

LM85088017553xO

LM8219517541500

LM8100218131500
LM85307217563xO

LM85088017555xO

LM8219517541500

LM8100218134500
OR £. missing
fiche missing
                               263

-------
Path. 42  Row 35
     45/35

     46/35


Path 42  Row 36
     45/36
92/06/19

85/06/16

74/06/26

74/07/15




92/08/30

85/09/12

74/08/19
LM84362617411xO

LM84106617583xO

LM8170317590500

LM8172218041500




LM85310417565xO

LM85056018034xO

LM8175717574500
     Second WRS-1 not required?
Path 43  Row 26
     46/26

     47/26


Path 43  Row 27
     46/27


     47/27
90/07/15

85/08/18

75/07/01

73/08/08


r

90/07/15

86/07/20

75/07/10
73/08/07

73/08/08
LM85232717564xO

LM85053518062xO

LM8508217450500

LM8138118140500




LM85232717571xO

LM85087117594xO

LM8508217453500
LM8138018084500

LM8138118142500
OR
fiche missing
                                264

-------
Path 43  Row 28
     46/28
     47/28
92/07/28

86/07/20

72/08/12
73/08/07
75/07/10

73/08/08
LM84366517455xO

LM85087118000xO

LM8102018085500
LM8138018090500
LM8508217455500

LM8138118145500
OR
OR
fiche missing
Path 43  Row 29
     46/29


     47/29


Path 43  Row 30
92/07/28

86/07/20

72/08/12
73/08/07

73/08/08
     46/30

     47/30


Path 43  Row 31
     46/31

     47/31
92/08/05

86/08/05

73/08/07

73/08/08




92/08/13

86/08/21

73/08/07

73/08/08
LM84366517461xO

LM85087118002xO

LM8102018092500
LM8138018093500

LM8138118151500
LM85307918010xO

LM85088718001xO

LM8138018095500

LM8138118154500
LM84368117473xO

LM85090318000xO

LM8138018102500

LM8138118160500
OR
fiche missing
                               265

-------
Path 43  Row 32
     46/32



     47/32






Path 43  Row 33
     46/33



     47/33






Path 43  Row 34
     46/34



     47/34






Path 43  Row 35
     46/35
 92/07/28




 86/07/20




 72/07/25




 73/07/26






J




 92/08/05




 86/08/05




 72/07/25




 72/07/26






t




 92/08/05




 85/08/05




 72/07/25




 74/06/28











 92/09/06




 85/08/18




 72/09/17
LM84366517472xO




LM85087118014xO




LM8100218125500




LM8100318170500











LM85307918021xO




LM85088718013xO




LM8100218131500




LM8100318173500











LM85307918023xO




LM85051918094xO




LM8100218134500




LM8170518100500











LM85311118023xO




LM85053518095xO




LM8102018115500
     Second WRS-1 not  required?
                                266

-------
Path 43  Row  36
     46/36
92/09/06




85/08/18




72/09/17
LM85311118025xO




LM85053518101xO




LM8105618123500
     Second WRS-1 not required?
check fiche
Path 44  Row 26




               92/07/19




               85/07/08




     47/26     73/08/08




     48/26     72/07/27






Path 44  Row 27




               91/08/02




               85/08/25




     47/27     73/08/08




     48/27     72/07/27






Path 44  Row 28




               92/08/20




               86/08/12




     47/28     73/08/08




     48/28     72/07/27
               LM84365617510xO




               LM85049418125xO




               LM8138118140500




               LM8100418194500










               LM84330417591xO




               LM85054218125xO




               LM8138118142500




               LM8100418201500











               LM84368817524xO




               LM85089418052xO




               LM8138118145500




               LM8100418203500
                               267

-------
Path 44  Row 29
     47/29




     48/29
92/08/20




85/08/25




73/08/08




72/07/27
LM84368817530xO




LM85054218133xO




LM8138118151500




LM8100418210500
Path. 44  Row 30
     47/30



     48/30






Path 44  Row 31
     47/31



     48/31






Path 44   Row 32
92/08/28




86/08/12




73/08/08




72/07/27










92/08/04




85/08/09




73/08/08




72/07/27
      47/32




      48/32
92/08/04




86/08/12




73/07/26




72/07/27
LM85310218065xO




LM85089418061xO




LM8138118154500




LM8100418212500










LM84367217532xO




LM85052618143xO




LM8138118160500




LM8100418215500










LM84367217535xO




LM85089418070xO




LM8100318170500




LM8100418221500
                                268

-------
Path 44  Row 33
     47/33



     48/33






Path 44  Row 34
     47/34
 92/08/04




 86/08/12




 72/07/26




 73/07/22






I




 92/08/04




 85/07/08




 74/06/28
LM84367217541xO




LM85089418072xO




LM8100318173500




LM8136418225500











LM84367217544x0




LM85049418160xO




LM8170518100500
     Second WRS-1 not required?
Path 44  Row 35
     47/35
92/09/13




84/09/07




72/08/13
LM85311818083xO




LM85019018163xO




LM8102118174500
     Second WRS-1 not required?
Path 45 Row 26
     49/26




     48/26
92/08/11




86/08/19




73/09/15




72/07/27
LM84367917575xO




LM85090118103xO




LM8141918244500




LM8100418194500
Bad fiche
                               269

-------
Path 45 Row 27
     49/27



     48/27






Path 45 Row 28
     49/28




     48/28






Path 45 Row 29
     49/29




     48/29






Path 45 Row 30
     49/30



     48/30
92/08/03




86/08/19




73/09/15




72/07/27











92/08/03




86/08/19




72/09/02




72/07/27










92/08/03




86/08/19




72/09/02




72/07/27










92/08/03




86/08/19




72/09/02




72/07/27
LM85307718121xO




LM85090118105xO




LM8141918251500




LM8100418201500











LM85307718123xO




LM85090118112xO




LM8104118262500




LM8100418203500











LM85307718130xO




LM85090118114xO




LM8104118265500




LM8100418210500











LM85307718132xO




LM85090118120xO




LM8104118271500




LM8100418212500
                                270

-------
 Path 45  Row 31




                92/08/03




                86/08/19




      49/31      73/08/28




      48/31      72/07/27






 Path 45  Row 32




                92/09/04




                86/08/19




      49/32      73/08/28




      48/32      72/07/27






Path  45  Row 33




                92/09/04




                86/08/19




      49/33      73/08/28




      48/33      73/07/22






Path 46 Row 26




                92/08/18




                85/08/23




     49/26     73/09/16




     50/26     73/09/16
 LM8530771834xO




 LM85090118123xO




 LM8140118271500




 LM8100418215500











 LM85310918134xO




 LM85090118125xO




 LM8140118273500




 LM8100418221500











 LM85310918140xO




 LM85090118132xO




 LM8140118280500




 LM8136418225500











 LM84368618041xO




LM85054018244xO




LM8141918244500




LM8142018303500
Fiche missing
Fiche missing
Fiche missing



EROS Check Film
                               271

-------
Path 46 Row 27




               92/08/10




               85/08/23




     49/27     73/09/15




     50/27     72/07/27






Path 46 Row 28



               92/08/10




               85/08/23




     49/28     72/09/02




     50/28     72/07/29
Path 46 Row 29
     49/29



     50/29






Path 46 Row 30
     49/30




     50/30
92/08/26




85/09/24




72/09/02




72/07/29










92/08/10




86/08/10




72/09/02




72/07/29
LM85308418181xO




LM85054018251xO




LM8141918251500




LM8100418201500










LM85308418184xO




LM85054018253xO




LM8104118262500




LM8100618315500











LM85310018185xO




LM85057218253xO




LM8104118265500




LM8100618322500










LM85308418193xO




LM85089218183xO




LM8104118271500




LM8100618324500
                                272

-------
Path 46 Row 31
     49/31


     50/31



Path 46 Row 32
     49/32
92/08/10


84/07/03


73/07/23


73/07/24






92/07/09


84/07/03


73/08/28
LM85308418195xO


LM85012418254xO


LM8136518274500


LM8136618332500
LM85305218204xO


LM85015618270xO


LM8140118273500
     Second WRS-1 not required?
Path 47 Row 26
     50/26


     51/26



Path 47 Row 27
     50/27


     51/27
92/07/16


86/08/01


73/09/16


74/09/12







92/07/16
      c

86/08/01


75/07/23


74/09/12
LM85305918242xO


LM85088318233xO


LM8142018303500


LM8178118265500






LM85305918244xO


LM85088318235xO


LM8218218201500


LM8178118272500
Black fiche
Fiche missing
                               273

-------
Path 47 Row 28
     50/28



     51/28






Path 47 Row 29
     50/29
92/07/16




86/05/29




75/07/23




74/09/12











92/07/16




86/05/29




72/07/29
LM85305918251xO




LM85081918262xO




LM8218218204500




LM8178118274500











LM85305918253xO




LM85081918265xO




LM8100618322500
     Second WRS-1 not required?
Path 47 Row 30
     50/30
92/07/16




86/06/30




72/07/29
LM85305918260xO




LM85305918260xO




LM8100618324500
     Second WRS-1 not required?
Path 48 Row 26
     51/26



     52/26
92/09/09




86/08/08




74/09/12




74/09/13
LM85311418294xO




LM85089018293xO




LM8178118265500




LM8178218324500
                               274

-------
Path 48 Row 27
     51/27
92/09/09




84/07/17




74/09/12
LM85311418301xO




LM85047418380xO




LM8178118272500
     Second WRS-1 not required?
                               275

-------
          NALC HSS Triplicates - Eastern and Southern US
Path 6  Row 38

Path 10  Row 29
     To be selected
     11/29
91/07/11
91/06/25

86/08/14

75/07/02
LM85268814370xO
LM85267214370xO

LM85089614353xO

LM8216114374500
     Second WRS-1 not required?
Path 11  Row 27
     11/27

     12/27


Path 11  Row 28
     11/28

     12/28


Path 11  Row 29
     11/29

     12/29
91/06/16

85/07/01

74/08/21

76/06/09


J

91/06/16

85/07/01

76/08/19

73/07/22


)

91/07/02

85/07/01

75/07/02

74/07/17
LM85266314421xO

LM85048714492xO

LM8175914384500

LM8250414361500




LM85266314423xO

LM85048714494xO

LM8257514285500

LM8136414541500




LM85267914431xO

LM85048714500xO

LM8216114374500

LM8172414463500
                                                  Composite with
                                276

-------
Path 11  Row  31




               92/08/21




               86/08/05




     12/31     74/07/17






Path 12  Row  27




               91/06/07




               86/07/27




     12/27     76/06/09




     13/27     73/07/23






Path 12  Row  28




               91/06/07




               85/06/22




     12/28     73/07/22




     13/28     73/07/23






Path 12  Row  29




               91/06/07




               85/06/22




     12/29     74/07/17




     13/29     73/07/23
LM85309514432xO




LM85088714425xO




LM8172414472500











LM85265414481xO




LM85087814475xO




LM8250414361500




LM8136514593500
LM85265414484xO




LM85047814555xO




LM8136414541500




LM8136514595500











LM85265414490xO




LM85047814562xO




LM8172414463500




LM8136515002500
                               277

-------
Path 12  Row 30
     13/30
92/08/12




86/09/13




73/07/23
LM85308614492xO




LM85092614471xO




LM8136515004500
     Second WRS-1 not required?
Path 12  Row 31
     12/31



     13/31






Path 13  Row 28
     14/28
91/06/07




85/08/09




74/07/17




73/07/23






(




91/06/14




84/06/10




75/06/26
LM85265414495xO




LM85052614564xO




LM8172414472500




LM8136515011500










LM85266114545xO




LM85010115001xO




LM8506814431500
     Second WRS-1 not required?
Path 13  Row 29
     14/29
 92/06/16




 84/06/10




 74/06/13
LM85302914554xO




LM85010115004xO




LM8169014591500
     Second WRS-1 not required?
                               278

-------
Path 13  Row 30
     14/30
92/09/20




85/09/17




73/08/29
LM85312514545x0




LM85056515020xO




LM8140215060500
     Second WRS-1 not required?
Path 13  Row 31
     14/31
91/07/16




84/06/10




74/06/13
LM85269314563xO




LM85010115013xO




LM8169015000500
     Second WRS-1 not required?
Path 13  Row 32
     14/32
91/07/16



84/06/10



74/06/13
LM85269314570x0




LM85010115015xO




LM8169015002500
     Second WRS-1 not required?
Path 14  Row 29
     15/29
91/08/08




86/08/26




76/06/03
LM85271615021xO




LM85090814595xO




LM8541114274500
     Second WRS-1 not required?
                               279

-------
Path 14  Row 35




               91/07/23




               86/07/09




     15/35     75/06/18
               LM85274815045xO




               LM85086015040xO




               LM8214715031500
     Second WRS-1 not required?
Path 14  Row 36
     15/36
90/09/06




86/09/27




73/08/30
LM85238015011xO




LM85094015013xO




LM8140315141500
Path 15  Row 29
     16/29



     17/29
92/07/16




85/07/13




75/06/10




72/08/19
LM85305915075xO




LM85049915144xO




LM8505214560500




LM8102715231500
Path 15  Row 35
     15/35



     16/35
91/10/18




85/09/15




73/08/30




72/10/11
LM85278715112xO




LM85056315163xO




LM8140315134500




LM8108015201500
                               280

-------
Path 15  Row 36
     15/36

     16/36
91/10/18

85/09/15

73/08/30

72/10/11
LM85278715114xO

LM85056315165xO

LM8140315141500

LM8108015203500
Path 15  Row 37
     16/37
91/10/18

85/08/14

72/10/11
LM85278715121xO

LM85053115171xO

LM8108015210500
     Second WRS-1 not required?
Path 15  Row 41
     16/41
92/05/13

86/04/27

73/04/09
LM85299515132xO

LM8507871544xO

LM8126015233500
     Second WRS-1 not required?
Path 15  Row 42
     16/42
91/03/08
91/02/20

86/04/27

73/03/22
LM85256315111xO
LM85254715105xO

LM85078715151xO

LM8124215240500
     Second WRS-1 not required?
                                                  Composite with
                               281

-------
Path 15  Row 43
     16/43
92/05/13



86/04/27



75/02/13
LM85299515140xO




LM85078715153xO




LM8202215122500
     Second WRS-1 not required?
Path 16  Row 35
     17/35
91/10/09




85/09/06




75/08/22
LM85277815172xO




LM85055415224xO




LM8512514595500
     Second WRS-1 not required?
Path 16  Row 36
     17/36
90/10/06




85/09/06




72/10/12
LM85241015132xO




LM85055415231xO




LM8108115262500
     Second WRS-1 not  required?
Path 16  Row 37
      17/37
92/10/06




86/09/25




72/10/12
LM85241015134xO




LM85093815142xO




LM8108115264500
      Second WRS-1 not required?
                                282

-------
Path  16   Row 38
                91/08/06

                86/08/24
               LM85271415180xO

               LM85090615155xO
     17/38


     Second WRS-1 not  required?
72/08/19
74/08/27
LM8102715265500
LM8176515173500
OR (EROS Check)
Path 16  Row 39

               92/05/04

               85/05/17

     17/39     74/05/11
               LM85298615184XO

               LM85044215250xO

               LM8165715213500
     Second WRS-1 not required?
Path 16  Row 40

               92/03/17

               86/04/02

     17/40     73/04/10
               LM85293815193xO

               LM85076215210xO

               LM8126115285500
     Second WRS-1 not required?
Path 16  Row 41

               92/03/17

               86/04/18

     16/41     73/04/09

     17/41     73/04/28
               LM85293815200xO

               LM85077815211xO

               LM8126015233500

               LM8127915291500
                               283

-------
Path. 16  Row 42

               92/05/04

               86/04/18

     16/42     73/03/22

     17/42     75/04/09
               LM85298615195xO

               LM85077815213xO

               LM8122415240500

               LM8207715173500
Path 17  Row 30

               92/06/28

               85/07/27

     18/30     75/06/30

     19/30     74/07/06
               LM85304115204xO

               LM85051315272XO

               LM8507215064500

               LM8171315273500
Path 17  Row 33
     18/33

     19/33
92/10/02

85/09/29

76/09/13

73/09/03
LM85313715203xO

LM85057715275xO

LM8260015103500

LM8140715355500
Path 17  Row 35
     18/35

     19/35
92/10/02

84/10/12

72/09/07

73/10/27
73/07/29
LM85313715211xO

LM85022515294xO

LM8104615313500

LM8146115352500
LM8137115371500
OR (if too late)
                               284

-------
 Path  17   Row 36




                91/09/30




                86/10/18




      18/36      72/09/07




      19/36      73/09/03
 LM85276915235xO




 LM85096115192xO




 LM8104615315500




 LM8140715370500
Path 17  Row  37




               90/09/27




               86/10/18




     18/37     74/10/21




     19/37     74/10/04






Path 17  Row  38




               90/10/29




               86/10/18




     18/38     72/10/31




     19/38     74/10/04






Path 17  Row  39




               91/10/16




               84/10/12




     18/39     72/10/13




     19/39     72/10/14
LM85240115200xO




LM85096115195xO




LM8182015212500




LM8183915262500











LM85243315201xO




LM85096115201xO




LM8110015331500




LM8180315273500











LM85278515252xO




LM85022515311xO




LM8108215332500




LM8108315390500
                               285

-------
Path 17  Row 40
     18/40
91/10/16




84/10/12




74/10/21
LM85278515254xO




LM85022515314xO




LM8182015223500
     Second WRS-1 not required?
Path 17  Row 41
     18/40
91/10/16




86/10/18




73/11/13
LM85278515260xO




LM85096115212xO




LM8147815315500
     Second WRS-1 not required?
Path 18  Row 31
     19/31



     20/31






Path 18  Row 32
     19/32



     20/32
92/06/03




86/06/03




74/07/06




74/07/07





>




90/08/17




85/07/18




73/09/03




73/09/04
LM85301615274xO




LM85082415281xO




LM8171315275500




LM8171415334500
LM85236015243xO




LM85050415343xO




LM8140715352500




LM8140815410500
                               286

-------
Path 18  Row 33
     19/33

     20/33
91/09/21

84/09/17

73/09/03

73/09/04
LM85276015285xO

LM85020015350xO

LM8140715355500

LM8140815413500
Path 18  Row 34

               91/09/21

               84/09/17

     19/34     73/09/03

     20/34     74/10/05
               LM85276015292xO

               LM85020015352xO

               LM8140715361500

               LM8180415313500
Path 18  Row 35

               90/10/10

               84/10/03

     19/35     73/10/27
               73/07/29
     20/35
72/10/15
LM85242415251xO

LM85021615355xO

LM8146115352500
LM8137115371500

LM8108415431500
                                   OR (if too late)
Path 18  Row 36

               90/10/20

               84/10/03

     19/36     73/09/03

     20/36     72/10/15
               LM85242415254xO

               LM85021615361xO

               LM8140715370500

               LM8108415433500
                               287

-------
Path 18  Row 37
     19/37

     20/37
90/10/20

85/10/06

74/10/04

72/10/15
LM85242415260xO

LM85058415353xO

LM8180315271500

LM8104815434500
Path 18  Row 38
     19/38

     20/38
90/10/20

85/10/06

74/10/04

74/10/23
LM85242415263xO

LM85058415360xO

LM8180315273500

LM8182215331500
Path. 18  Row 39
     19/39

     20/39
90/09/18

85/10/06

72/10/14

74/10/23
LM85239215271xO

LM85058415362xO

LM8108315390500

LM8182215333500
Path 19  Row 31
     20/31

     21/31
91/06/08

85/06/23

74/07/07

73/06/07
75/07/30
LM85265515331xO

LM85047915402xO

LM8171415334500

LM8131915474500
LM8218915352500
OR (EROS Check)
                               288

-------
Path 19  Row 32
     20/32

     21/32
92/08/29

86/08/29
86/08/13

73/09/04

73/07/13
LM85310315330xO

LM85091115315xO
LM85089515322xO

LM8140815410500

LM8135515474500
                                                   Composite with
Path 19  Row 33

              '92/08/29

               84/09/08

     20/33     73/09/04

     21/33     74/10/06
Path 19  Row 34
     20/34

     21/34


Path 19  Row 35
     20/35

     21/35
90/09/25

84/09/08

74/10/05

74/10/06




92/09/30

86/09/14

72/10/15

74/10/06
               LM85310315333xO

               LM85019115,411xO

               LM8140815413500

               LM8180515365500
LM85239915311xO

LM85019115413xO

LM8180415313500

LM8180515372500
LM85313515334xO

LM85092715323xO

LM8108415431500

LM8180515374500
                               289

-------
Path 19  Row 36
     20/36



     21/36
91/09/28




84/09/08




72/10/15




74/10/06
LM85276715361xO




LM85019115422xO




LM8108415433500




LM8180515381500
Path 19  Row 37



               91/09/28




               83/10/16




     20/37     72/10/15




     21/37     74/10/06
               LM85276715364xO




               LM84045715421xO




               LM8108415440500




               LM8180515383500
Path 19  Row 38
     20/38



     21/38
91/10/14




86/10/16




74/10/23




72/10/16
LM85278315371xO




LM85095915324xO




LM8182215331500




LM8108515501500
Path 19  Row 39




               91/10/14




               86/10/16




     20/39     74/10/23




     21/39     72/10/16
               LM85278315374xO




               LM85095915330xO




               LM8182215333500




               LM8106715501500
                               290

-------
Path 20  Row 29




               92/08/20



               86/08/20




     21/29     75/08/08




     22/29     75/07/31
               LM85309415381xO




               LM85091215371xO




               LM8511115211500




               LM8219015401500
Path 20  Row 30




               92/09/05




               86/08/20




     21/30     75/08/08




     22/30     75/07/31
               LM85311015382xO




               LM85090215373xO




               LM8511115214500




               LM8219015404500
Path 20  Row 31




               90/08/31




               85/08/09




     21/31     75/07/30




     22/31     75/07/31
Path 20  Row 32
     21/32




     22/32
91/08/02




86/08/04




73/07/13




74/07/09
LM85237415362xO




LM84112015402xO




LM8218915352500




LM8219015410500











LM85271015402xO




LM85088615385xO




LM8135515474500




LM8171615453500
                               291

-------

Path 20  Row 33
     21/33




     22/33
91/08/02




86/08/04




75/08/08




76/07/25
LM85271015404xO




LM85088615392xO




LM8511115225500




LM8255015345500
Path 20  Row 34
     21/34




     22/34
92/10/07




86/10/07




74/10/06




76/08/30
LM85314215391xO




LM85095015373xO




LM8180515372500




LM8258615341500
Path 20  Row 35
     21/35




     22/35
90/09/16




86/10/07




74/10/06




75/10/20
LM85239015375xO




LM85095015375xO




LM8180515374500




LM8518415251500
Path 20  Row 36
     21/36




     22/36
90/09/16




86/07/19




74/10/06




75/10/02
LM85239015382xO




LM85087015410xO




LM8180515381500




LM8516615264500
                                292

-------
Path 20  Row 37
     21/37

     22/37
90/09/16

86/09/21

74/10/06

74/10/07
LM85229415390xO

LM85093415391xO

LM8180515383500

LM8180615442500
Path 20  Row 38
     21/38

     22/38
90/09/16

84/10/01

74/09/18

74/10/07
LM85231015393xO

LM85021415492xO

LM8178715100500

LM8180615444500
Path 20  Row 39
     21/39

     22/39
90/09/16

84/10/01

72/10/16

74/10/07
LM85239015393xO

LM85021415495xO

LM8108515503500

LM8180615451500
Path 21  Row 28
     22/28

     23/28
91/07/08

86/06/24
86/06/08

75/07/31

74/07/10
75/06/26
LM85268515444xO

LM85084515445xO
LM85082915453xO

LM8219015395500

LM8171715493500
LM8215515460500
                                                  OR
OR (EROS Check)
                               293

-------
Path 21  Row 29
     22/29



     23/29
91/08/09




85/08/08




75/07/31




75/08/01
LM85271715452xO




LM85052515513xO




LM8219015401500




LM8219115460500
Path 21  Row 30
     22/30




     23/30






Path 21  Row 31
     22/31



     23/31
91/08/25




85/07/23




75/07/31




75/08/01










91/08/25




85/07/23




75/08/27




74/08/15
LM85273315455xO




LM85050915521xO




LM8219015404500




LM8219115462500










LM85273315462xO




LM85050915523xO




LM8513015264500




LM8175315493500
Path 21  Row  32
     22/32




     23/32
91/06/06




84/06/02




73/06/08




73/06/09
LM85265315455xO




LM85009315511xO




LM8132015534500




LM8132115593500
                                294

-------
Path 21  Row 33

               92/09/28

               85/09/01

     22/33     75/10/02

     23/33     72/09/30
LM85313315452xO

LM84114315465xO

LM8516615252500

LM8106915591500
Path 21  Row 34

               92/09/28

               86/09/28

     22/34     76/08/30

     23/34     72/09/30
LM85313315454xO

LM85094115440xO

LM8258615341500

LM8106915594500
Path 21  Row 35

               91/09/26

               86/09/28

     22/35     75/10/20

     23/35     72/09/12



Path 21  Row 36

               91/09/26

               86/09/28

     22/36     75/10/02

     23/36     73/08/20
LM85276515481xO

LM85094115442xO

LM8518415251500

LM8105116000500
Faded fiche
(EROS Check)
LM85276515483xO

LM85094115445xO

LM8226215421500

LM8139316001500
Black fiche
(EROS check)
                               295

-------
Path 21  Row 37
     22/37

     23/37
91/09/26

86/09/29

74/10/07

74/09/20
LM85276515490xO

LM85094115451xO

LM8180615442500

LM8178915503500
Path 21  Row 38
     22/38

     23/38



Path 21  Row 39
91/09/26

86/09/28

74/10/07

72/09/12
     22/39

     23/39
92/10/14

84/09/06

74/10/07

74/10/08
LM85276515492xO

LM85094115453xO

LM8180615444500

LM8105116012500
LM85314915472xO

LM85018915560xO

LM8180615451500

LM8180715505500
Faded Fiche
(EROS Check)
Path 21  Row 40
     22/40

     23/40
92/10/14

85/09/09

75/10/11

74/10/08
LM85314915474xO

LM85055715554xO

LM8226215435500

LM8180715512500
                                296

-------
Path 22  Row 27

               91/07/15

               86/08/18

     23/27     75/06/26

     24/27     73/07/16
               LM85269215503xO

               LM85090015485xO

               LM8215515454500

               LM8135816024500
Path 22  Row 28

               91/07/15

               85/06/28

     23/28     74/07/10
               75/06/26
     24/28
74/07/11
LM85269215510xO

LM85048415574xO

LM8171715493500
LM8215515460500

LM8171815551500
                                   OR  (EROS Check)
Path 22  Row 29

               91/07/15

               85/06/28

     23/29     73/06/09

     24/29     73/06/10


Path 22  Row 30

               91/07/15

               85/06/28

     23/30     73/06/09

     24/30     73/06/10
               LM85269215512xO

               LM85048415580xO

               LM8132115581500

               LM8132216042500




               LM85269215514xO

               LM85048415583xO

               LM8132115584500

               LM8132216042500
                               297

-------
Path 22  Row 31
     23/31




     24/31
91/07/15




85/07/22




73/06/09




73/06/10
LM85269215521xO




LM84110215530xO




LM8132115590500




LM8132216045500
Path 22  Row 32
     23/32




     24/32
92/10/05




85/10/02




75/10/12




74/09/21
LM85314015505xO




LM85058015582xO




LM8226315461500




LM8179015541500
Path 22  Row 33
     23/33




     24/33






Path 22  Row 34
     23/34




     24/34
92/10/05




85/10/02




74/10/08




74/10/09






[




92/10/05




85/09/16




72/09/30




74/10/09
LM85314015511xO




LM85058015584xO




LM8180715482500




LM8180815541500











LM85314015514xO




LM85056415592xO




LM8106915594500




LM8180815543500
                                298

-------
Path 22  Row 35
                92/10/05

                85/09/16

     23/35      72/09/12

     24/35      72/10/01
                75/09/07
LM85314015520xO

LM85056415594xO

LM8105116000500

LM8107016055500
LM8222815533500
Prime
Alternate
Path 22  Row 36

               91/07/31

               86/07/17

     23/36     73/08/20

     24/36     74/07/29
LM85270815542xO

LM85086815533xO

LM8139316001500

LM8173615575500
Black fiche
(EROS Check)
Path 22  Row 37

               92/10/05

               86/10/21

     23/37     72/09/12

     24/37     75/10/13


Path 22  Row 38

               92/10/05

               85/08/31

     23/38     74/10/08

     24/38     75/10/13
LM85314015525xO

LM85096415504xO

LM8105116005500

LM8226415540500




LM85314015531xO

LM85054816011xO

LM8180715503500

LM8226415543500
                               299

-------
Path. 22  Row 39
     23/39




     24/39
92/10/05




85/08/31




74/10/08




72/10/01
LM85314015534xO




LM85054816013xO




LM8180715505500




LM8107016073500
Path 22  Row 40
     23/40




     24/40
92/10/05




85/08/31




74/10/08




74/10/09
LM85314015540xO




LM85054816015xO




LM8180715512500




LM8180815570500
Path 23  Row 27
     24/27




     25/27






Path 23  Row 28
     24/28




     25/28
91/06/04




85/06/03




73/07/16




74/06/24






(




91/06/04




86/08/09




74/07/11




75/07/16
LM85265115561xO




LM85045916033xO




LM813581602450d




LM8170116010500










LM85265115564x0




LM85089115554xO




LM8171815551500




LM8217515572500
                                300

-------
Path 23  Row  29




               91/06/04




               85/06/03




     24/29     74/07/11




     25/29     73/07/17
LM85265115570xO




LM85045916042xO




LM8171815554500




LM8135916091500
Path 23  Row 30




               91/06/04




               85/06/03




     24/30     73/06/10




     25/30     73/06/11
LM85265115572xO




LM85045916044xO




LM8132216042500




LM8132316100500
Path 23  Row 31




               92/09/10




               85/09/07




     24/31     75/09/07




     25/31     72/10/02






Path 23  Row 32




               90/09/05




               84/09/20




     24/32     74/09/21




     25/32     72/10/02
LM85311515571xO




LM85055516042xO




LM8222815515500




LM8107116095500











LM85237915551xO




LM85020316053xO




LM8179015541500




LM8107116102500
                               301

-------
Path 23  Row 33
     24/33

     25/33
90/09/05

84/09/20

74/10/09

72/10/02
LM85237915554xO

LM85020316060xO

LM8180815541500

LM8107116104500
Path 23  Row 34
     24/34

     25/34
90/09/05

84/09/20

74/10/09

72/10/02
LM85237915560xO

LM85020316062xO

LM8180815543500

LM8107116111500
Path 23  Row 35
     24/35


     25/35


Path 23  Row 36
90/09/05

84/09/20

75/09/07
73/08/21

72/10/02
     24/36

     25/36
92/10/12

84/09/20

74/07/29

72/10/02
LM85237915563xO

LM85020316064xO

LM8222815533500
LM8139416053500

LM8107116113500
LM85314715583xO

LM85020316071xO

LM8173615575500

LM8107116120500
OR (EROS Check)
                                302

-------
Path 23  Row 37

               92/10/12

               84/09/20

     24/37     75/10/13

     25/37     72/10/02
               LM85314715585xO

               LM85020316073xO

               LM8226415540500

               LM8107116122500
Path 23  Row 38

              , 91/10/10

               86/08/25

     24/38     75/10/13

     25/38     72/10/02
               LM85277916015xO

               LM85090715590xO

               LM8226415543500

               LM8107116125500
Path 23  Row 39

               92/10/12

               86/08/25

     24/39     72/10/01
               75/09/25
     25/39
74/10/10
Path 23  Row 40

               92/10/12

               86/08/25
               84/10/06

     24/40     74/10/09

     25/40     74/10/10
LM85314715594xO

LM85090715593xO

LM8107016073500
LM8224615550500

LM8180916022500
               LM85314716001xO

               LM85090715595xO
               LM85021916084xO

               LM8180815570500

               LM8180916025500
                                   OR (EROS Check)
                    OR (EROS Check)
                               303

-------
                                                                     •.-pa-
     24  Row 27
     26/27



     27/27
 92/06/29




 85/08/13




 74/06/25




 76/06/06
LM85304216025xO




LM85053016092xO




LM8170216065500




LM8250116022500
Path 24  Row 28
     26/28



     27/28






Path 24  Row 29
     26/29



     27/29






Path 24  Row 30
     26/30




     27/30
 90/06/24




 85/07/12




 75/07/17




 74/06/26






)




 91/08/14




 86/08/16




 73/08/05




 73/09/11






)




 90/09/04




 86/09/01




 72/08/28




 72/08/29
LM85230616001xO




LM85049816100xO




LM8217616030500




LM8170316125500










LM85272216035xO




LM85089816020xO




LM8137816144500




LM8141516195500










LM84297216090xO




LM85091416015xO




LM8103616152500




LM8103716210500
                               304

-------
Path 24  Row  31
     25/31




     26/31
92/10/03       LM85313816025xO




86/09/01       LM85091416021xO



GLIS for scene




74/09/05       LM8177416055500
Path 24  Row 32
     25/32




     26/32
92/10/03




85/10/16




72/10/02




73/10/16
LM85313816032xO




LM85059416103xO




LM8107116102500




LM8145016142500
Path 24  Row 33




               92/10/03




               85/10/16




     25/33     72/10/02




     26/33     73/10/16






Path 24  Row 34




               91/10/17




               85/10/16




     25/34     72/10/02




     26/34     72/08/28
               LM85313816034xO




               LM850594l6110xO




               LM8107116104500




               LM8145016145500











               LM85278616063xO




               LM85059416112xO




               LM8107116111500




               LM8103616165500
                               305

-------
Path 24  Row 35
     25/35



     26/35
92/10/03




85/10/16




72/10/02




74/07/31
LM85313816043xO




LM85059416114xO




LM8107116113500




LM8173816085500
Path 24  Row 36
     25/36




     26/36
92/10/19




85/10/16




72/10/02




75/09/27
LM85315416044xO




LM85059416121xO




LM8107116120500




LM8224816051500
Path 24  Row 37
     25/37




     26/37






Path 24  Row 38
     25/38



     26/38
92/10/19




86/10/19




72/10/02




75/09/27






!




92/10/19




86/09/17




72/10/02




72/10/03
LM85315416050xO




LM85096216030xO




LM8107116122500




LM8224816054500










LM85315416053xO




LM85093016043xO




LM8107116125500




LM8107216183500
                               306

-------
Path 24  Row  39




               92/10/19




               86/10/19




     25/39     74/10/10




     26/39     72/10/03
LM85315416055xO




LM85096216035xO




LM8180916022500




LM8107216190500
Path 25  Row 26




               92/06/20




               86/07/06




     26/26     76/06/05




     27/26     76/07/03
LM85303316084xO




LM85085716083xO




LM8250015561500




LM8544115325500
Path 25  Row 27




               91/08/05




               86/07/22




     26/27     74/06/25




     27/27     76/06/05






Path 25  Row 28




               91/08/05




               86/07/22




     26/28     75/07/17




     27/28     74/06/26.
LM85271316092xO




LM85087316082xO




LM8170216065500




LM8250116022500











LM85271316094xO




LM85087316084xO




LM8217616030500




LM8170316125500
                               307

-------
Path 25  Row 29




               92/09/0a




               86/09/08




     26/29     7/3/0 8/Q5




     27/29     73/09/11
               LM85311316085xO




               LM85092116072xO




               LM8137816144500




               LM8141516195500
Path 25  Row 30




               92/09/08




               86/09/08




     26/30     72/08/28




     27/30     73/09/11






Path 25  Row 31



               92/09/08




               86/09/08




     26/31     74/09/05




     27/31     72/08/29
Path 25  Row 32
     26/32




     27/32
92/09/24




86/09/08




73/08/05




74/08/19
LM85311316091xO




LM85092116074x0




LM8103616152500




LM8141516202500










LM85311316093xO




LM85092116081xO




LM8177416055500




LM8103716213500










LM85312916094xO




LM85092116083xO




LM8137816160500




LM8175716124500
                                308

-------
 Path 25  Row 33




                92/09/24




                84/09/18




      26/33      73/10/16




      27/33      72/10/04
 LM85312916100xO




 LM85020116182xO




 LM8145016145500




 LM8107316221500
Path 25  Row 34




               92/09/24




               84/09/18




     26/34     72/08/28




     27/34     72/10/04
LM85312916100xO




LM85020116182xO




LM8103616165500




LM8107316224500
Path 25  Row 35




               90/08/26




               85/08/28




     26/35     74/07/31




     27/35     72/08/11






Path 25  Row 36




               91/10/08




               84/07/08




     26/36     75/09/27




     27/36     72/10/04
LM84296316172xO




LM84113916122xO




LM8173816085500




LM8101916225500











LM85277716132xO




LM84072316143xO




LM8224816051500




LM8107316233500
                               309

-------
Path. 25  Row 37



               91/10/08



               8.4/07/08



     26/37     75/03/27



     27/37     72/10/04
LM85277716135xO



LM84072316145xO



LM8224816051500



LM8107316235500
Path 25  Row 38



               92/07/06



               814/0 9/18



     26/38     74/07/13



     27/38     73/07/19





Path 25  Row 39



               92/07/06



               85/06/01



     26/39     74/07/13



     27/39     74/06/26




Path 25  Row 40



               92/07/06



               85/06/01



     26/40     76/06/23



     27/40     74/06/26
LM85304916132xO



LM85020116202xO



LM8172016105500



LM8136116243500









LM85304916134x0
             , .V '"


LM85045716204xO



LM8172016111500



LM8170316173500









LM85304916140xO



LM85045716210xO



LM8251816014500



LM8170316175500
                                310

-------
             NALC MSS Triplicates - Midwest and Great Plains
Path 26  Row 27
     28/27
 91/08/12




 86/07/29




 72/08/12
LM85272016153xO




LM85088016141xO




LM8102016252500
     Second WRS-1 not required?
Path 26  Row 28
     28/28
 91/08/28




 85/09/12




 72/08/12
LM85273616160xO




LM85056016214xO




LM8102016255500
     Second WRS-1 not required?
Path 26  Row 29
     27/29




     28/29






Path 26  Row 30
     27/30




     28/30
 90/08/25




 85/09/12




 73/09/11




 74/09/25






)




 92/08/30




 85/07/10




 73/09/11




 74/09/25
LM85236816124xQ




LM85056016220xO




LM8141516195500




LM8179416154500











LM85310416153xO




LM85049616231xO




LM8141516202500




LM8179416160500
                               311

-------
Path 26  Row 31
     27/31



     28/31
92/08/30




84/08/08




72/08/29




74/09/25
LM85310416155xO




LM85016016231xO




LM8103716213500




LM8179416163500
Path 26  Row 32
     27/32



     28/32






Path 26  Row 33
     27/33



     28/33
92/10/01




86/10/17




74/08/19




74/09/25






I




92/10/01




86/10/17




72/10/14




74/09/25
LM85313616154xO




LM85096016133xO




LM8175716124500




LM817941616.5500










LM85313616160xO




LM85096016135xO




LM8107316221500




LM8179416165500
Path 26  Row  34
     27/34




     28/34
 92/10/01




 86/10/17




 72/10/04




 72/09/17
LM85313616163xO




LM85096016142xO




LM8107316224500




LM8105616282500
                                312

-------
Path 26  Row 35

               90/08/25

               84/08/24

     27/35     72/08/11

     28/35     73/08/25



Path 26  Row 36

               90/06/06

               85/06/08

     27/36     75/06/12

     28/36     74/08/20
Path 26  Row 37
     27/37

     28/37


Path 26  Row 38
     27/38

     28/38
 91/06/25

 85/06/08

 75/06/12

 74/06/27


J

 91/06/25

 84/07/15

 73/07/19

 74/06/27
                LM85236816150xO

                LM85017616251xO

                LM8101916225500

                LM8139816282500
LM85228816154xO

LM85046416254xO

LM8214116122500

LM8175816200500




LM85267216190xO

LM85046416260xO

LM8214116124500

LM8170416222500




LM85267216192xO

LM84073016213xO

LM8136116243500

LM8170416225500
                    Black fiche
                    (EROS Check)
                               313

-------
Path 26  Row 39
     27/39

     28/39
90/07/08

84/07/15

74/06/26

74/06/27
LM85232016165xO

LM84073016215xO

LM8170316173500

LM8170416231500
Path 26  Row 40
     27/40

     28/40


Path 26  Row 41
     28/41
 92/10/01

 86/10/17

 76/09/22

 75/10/17


L

 92/10/01

 86/10/17

 73/10/18
LM85313616185xO

LM85096016164xO

LM9260916045500

LM8226816181500




LM85313616192xO

LM85096016170xO

LM8145216293500
     Second WRS-1 not required?
Path 27  Row 27
     29/27
 91/06/16

 84/06/28

 73/07/03
LM85266316210xO

LM85011916264xO

LM8134516313500
EROS Check
Black fiche
     Second WRS-1 not  required?
                                314

-------
 Path  27   Row 28
                91/06/16

                86/06/02
                LM85266316212xO

                LM85082316224xO
     29/28


     Second WRS-1  not  required?
 73/07/03
 74/06/28
 LM8134516320500
 LM8170516242500
OR (EROS Check)
Path 27  Row  29
     29/29
92/09/22

84/09/16

74/09/26
LM85312716205xO

LM85019916290xO

LM8179516212500
     Second WRS-1 not required?
Path 27  Row 30
     29/30
92/09/22

84/10/02

74/09/26
LM85312716212xO

LM85021516293xO

LM8179516214500
     Second WRS-1 not required?
Path 27  Row 31
     29/31
92/09/22

84/10/02

74/09/26
LM85312716214xO

LM85021516295xO

LM8179516221500
     Second WRS-1 not required?
                               315

-------
Path 27  Row 32
     29/32
92/09/22




85/08/10




74/09/26
LM85312716221xO




LM84112116235xO




LM8179516223500
     Second WRS-1 not required?
Path 27  Row 33
     29/33
92/09/22




84/08/15




73/08/26
LM85312716223xO




LM85016716302xO




LM8139916332500
     Second WRS-1 not required?
Path 27  Row 34
     29/34
92/09/22




85/10/05




74/09/08
LM85312716225xO




LM85058316300xO




LM8177716241500
     Second WRS-01 not required?
Path 27  Row 35     To be selected
                               316

-------
Path 27  Row 36     To be selected
Path 27  Row 37     To be selected
Path 27  Row 38     To be selected
Path 27  Row 39     To be selected
                               317

-------
Path 27  Row 40     To be selected
Path 27  Row 41     To be selected
Path 28  Row 26
     30/26

     31/26
90/09/08

85/07/08

73/07/22

76/06/10
LM85238216235xO

LM85049416335xO

LM8136416364500

LM8250516245500
Path 28  Row 27

               90/09/08
               90/07/06

               85/07/08

     30/27     74/07/17

     31/27     74/08/05
               LM85238216241xO
               LM85231816243xO

               LM85049416342xO

               LM8172416290500

               LM8174316340500
                    OR
                               318

-------
Path 28  Row 28
     30/28
91/08/26




85/08/25




74/07/17
LM85273416282xO




LM85054216342xO




LM8172416292500
     Second WRS-1 not required?
Path 28  Row 29




               91/08/26




               85/09/26




     30/29     73/09/14
               LM85273416285xO




               LM85057416342xO




               LM8141816370500
     Second WRS-1 not required?
Path 28  Row 30
     30/30
92/09/29




85/09/26




73/09/14
LM85313416272xO




LM85057416344xO




LM8140016375500
     Second WRS-1 not required?
Path 28  Row 31




               92/08/28




               84/08/14




     30/31     73/08/27
               LM85310216282xO




               LM84076016305xO




               LM8140016381500
     Second WRS-1 not required?
                               319

-------
Path 28  Row 32



               92/08/28



               84/08/14



     30/32     72/08/14
LM85310216284xO




LM84076016311x0




LM8102216384500
     Second WRS-1 not required?
Path 28  Row 33




               92/08/28




               86/09/13




     30/33     74/08/04
LM85310216290xO




LM85092616272xO




LM8174216305500
     Second WRS-1 not required?
Path 28  Row 34




               92/08/28




               84/09/07




     30/34     72/09/19
LM85310216293xO




LM85019016371xO




LM8105816395500
     Second WRS-1 not required?
Path 28  Row 35



               92/08/28




               86/08/28




     30/35     72/09/19
LM85310216295xO




LM85091016284xO




LM8105816401500
     Second WRS-1 not  required?
                                320

-------
Path 28   Row 36




               92/08/28




               85/08/09




     29/36     72/07/26




     30/36     74/07/17






Path 28   Row 37




               92/08/28




               85/08/09




     29/37     73/07/21




     30/37     72/09/19
LM85310216302xO




LM85052616374xO




LM810.0316355500




LM8172416324500
LM85310216304xO




LM85052616380xO




LM8136316353500




LM8105816410500
Path 28  Row 38




               92/09/29




               84/09/07




     29/38     72/08/31




     30/38     74/10/15
LM85313416303xO




LM85019016385xO




LM8103916355500




LM8181416304500
Path 28  Row 39




               92/08/28




               86/07/27




     29/39     72/08/31




     30/39     72/08/14
LM85310216313xO




LM85087816313xO




LM8103916361500




LM8102216414500
                               321

-------
Path 29  Row 26
     31/26



     32/26
90/08/30




85/07/31




76/06/10




74/08/06
LM85237316300xO




LM85051716395xO




LM8250516245500




LM8174416392500
Path 29  Row 27
     31/27




     32/27
90/07/13




86/07/18




74/08/05




74/08/06
LM85232516304xO




LM85086916331xO




LM8174316340500




LM8174416394500
Path 29  Row 28
     31/28




     32/28
90/07/13




84/07/20




73/08/28




72/07/29
LM85232516310xO




LM84073516360xO




LM8140116424500




LM8100616484500
Path 29  Row 29
     31/29




     32/29
90/07/13




84/08/13




73/07/05




72/07/29
LM85232516313xO




LM85016516410xO




LM8134716435500




LM8100616490500
                               322

-------
Path 29  Row 30




               90/08/30




               84/08/29




     31/30     73/08/28
               LM85237316313xO




               LM85018116414xO




               LM8140116433500
     Second WRS-1 not required?
Path 29  Row 31




               91/07/16




               84/08/13




     31/31     74/07/18
               LM85269316352xO




               LM85016516415xO




               LM8172516362500
     Second WRS-1 not required?
Path 29  Row 32
     31/32
90/08/30




84/08/29




73/08/10
LM85237316322xO




LM85018116423xO




LM8138316444500
     Second WRS-1 not required?
Path 29  Row 33




               92/08/19




               86/09/20




     31/33     73/08/28
               LM85309316352xO




               LM85093316332xO




               LM8140116444500
     Second WRS-1 not required?
                               323

-------
Path 29  Row 34



               92/07/18



               86/08/03



     31/34     73/07/05
               LM85306116362xO




               LM85088516352xO




               LM8134716455500
     Second WRS-1 not required?
Path 29  Row 35




               91/08/01




               86/07/18




     31/35     73/07/05
               LM85270916371xO




               LM85086916362xO




               LM8134716462500
     Second WRS-1 not required?
Path 29  Row 36



               92/08/03




               86/07/18




     31/36     75/08/09
               LM85307716365xO




               LM85086916364xO




               LM8219916343500
     Second WRS-1 not required?
Path 29  Row 37
     30/37




     31/37
92/08/03




85/07/31




72/09/19




75/08/09
LM85307716371xO




LM85051716442xO




LM8105816410500




LM8219916350500
                               324

-------
Path  29   Row 38




               92/08/03




               86/07/18




      30/38     75/08/08




      31/38     72/07/28
LM85307716374xO




LM85086916373xO




LM8219816294500




LM8100516470500
Path 30  Row 26




               92/08/10




               85/08/07




     32/26     74/08/06




     33/26     72/07/30






Path 30  Row 27




               92/08/10




               84/08/12




     32/27     74/08/06




     33/27     73/08/30
LM85308416390xO




LM85052416460xO




LM8174416392500




LM8100716533500











LM85308416392xO




LM84075816413xO




LM8174416394500




LM8140316534500
Path 30  Row 28




               92/08/10




               86/07/25




     32/28     72/07/29




     33/28     73/09/17
LM85308416394xO




LM85087616393xO




LM8100616484500




LM8142116534500
                               325

-------
Path 30  Row 29
     32/29




     33/29
91/08/24




84/08/12




72/07/29




72/09/04
LM85273216411xO




LM84075816422xO




LM8100616490500




LM8104316550500
Path 30  Row 30
     32/30




     33/30






Path 30  Row 31
     32/31
91/08/24




84/08/12




72/07/29




72/08/17










90./09/14




84/08/28




72/09/21
LM85273216413xO




LM84075816424xO




LM8100616493500




LM8102516551500











LM84298216461xO




LM84077416431xO




LM8106016500500
     Second WRS-1 not required?
Path 30  Row 32
     32/32
92/09/27




86/09/27




72/09/21
LM85313216403xO




LM85094016385xO




LM8106016503500
     Second WRS-1 not  required?
                                326

-------
Path 30  Row 33




               92/09/27




               86/09/27




     32/33     75/09/24
               LM85313216405xO




               LM85094016391xO




               LM8515816231500
     Second WRS-1 not required?
Path 30  Row 34
     32/34
92/09/27




86/09/27




74/10/17
LM85313216412xO




LM85094016394xO




LM8181616403500
     Second WRS-1 not required?
Path 30  Row 35
     32/35
92/09/27




86/09/11




74/10/17
LM85313216414xO




LM85092416403xO




LM8181616405500
     Second WRS-1 not required?
Path 30  Row 36




               92/09/27




               86/09/11




     32/36     72/10/09
               LM85313216421xO




               LM85092416410xO




               LM8107816522500
     Second WRS-1 not required?
                               327

-------
Path 30  Row 37
     32/37
92/10/13




86/10/13




72/10/09
LM85314816421xO




LM85095616402xO




LM8107816524500
     Second WRS-1 not required?
Path. 30  Row 38
     32/38
92/10/13




86/10/13




72/10/09
LM85314816423xO




LM85095616404xO




LM8107816531500
     Second WRS-1 not required?
Path 30  Row 39
     32/39
92/10/13




86/10/13




72/10/09
LM85314816430xO




LM85095616411xO




LM8107816533500
     Second WRS-1 not required?
Path 31  Row 26
     33/26




     34/26
91/08/31




84/08/11




72/07/30




72/08/18
LM85273916461xO




LM85016316521xO




LM8100716533500




LM8102616592500
                                328

-------
Path  31   Row 27




               91/08/31




               84/08/11




      33/27     73/08/30




      34/27     72/08/18






Path  31   Row 28




               91/08/31




               84/08/11




      33/28     73/09/17




      34/28     72/08/18






Path  31   Row 29




               92/09/02




               84/08/27




      33/29     72/09/04




      34/29     73/10/06
LM85273916463xO




LM85016316523xO




LM8140316534500




LM8102616594500
LM85273916470xO




LM85016316530xO




LM8142116534500




LM8102617001500
LM85310716460xO




LM85017916534xO




LM8104316550500




LM8144016592500
Path 31  Row 30




               92/09/02




               84/08/27




     33/30     72/08/17




     34/30     73/10/06
LM85310716462xO




LM85017916541xO




LM8102516551500




LM8144016595500
                               329

-------
Path 31  Row 31
     33/31



     34/31
92/08/01




84/08/11




73/08/12




73/07/26
LM85307516471xO




LM85016316541x6




LM8138516554500




LM8136817014500
Path 31  Row 32
     33/32




     34/32






Path 31  Row 33
     33/33




     34/33
 90/08/28




 84/08/11




 72/08/17




 73/07/26






1




 92/08/01




 86/07/16




 73/07/25




 72/08/18
LM85237116444xO




LM85016316543xO




LM8102516560500




LM8136817020500











LM85307516480xO




LM85086716480xO




LM8136716564500




LM8102617021500
Path 31  Row 34
     33/34
 92/08/01




 86/07/16




 73/07/07
LM85307516483xO




LM85086716482xO




LM8134916572500
     Second WRS-1 not required?
                               330

-------
Path 31  Row 35
     33/35
90/09/13




84/08/27




72/09/22
LM85238716455xO




LM85017916560xO




LM8106116573500
     Second WRS-1 not required?
Path 31  Row 36
     33/36
90/08/28




84/08/27




72/09/22
LM85237116462xO




LM85017916563xO




LM8106116575500
     Second WRS-1 not required?
Path 31  Row 37
     33/37
92/06/30



86/08/01



72/07/30
LM85304316500xO




LM85088316490xO




LM8100716581500
     Second WRS-1 not required?
Path 31  Row 38
     33/38
92/07/16




86/07/16




72/07/30
LM85305916501xO




LM85086716500xO




LM8100716584500
     Second WRS-1 not required?
                               331

-------
Path. 32  Row 26




               92/08/08




               84/07/01




     34/26     72/08/18




     35/26     74/07/22






Path, 32  Row 27




               91/08/22




               84/08/10




     34/27     72/08/18




     35/27     74/07/22
LM85308216512xO




LM85012216571xO




LM8102616592500




LM8172916570500











LM85273016524xO




LM84075616535xO




LM8102616594500




LM8172916573500
Path. 32  Row 28



               91/08/22




               84/08/10




     34/28     72/08/18




     35/28     74/07/04
LM85273016530xO




LM84075616542xO




LM8102617001500




LM8171116583500
Path 32  Row 29



               92/09/09




               85/09/06




     34/29     73/10/06




     35/29     72/10/12
LM85311416520xO




LM85055416591xO




LM8144016592500




LM813511706450
                                332

-------
Path 32  Row  30




               92/09/09




               85/09/06




     34/30     73/10/06




     35/30     72/08/19
Path 32  Row 31
     34/31




     35/31
92/09/09




85/09/06




73/10/06




72/08/19
               LM85311416523xO




               LM85055416594xO




               LM8136817011500




               LM8102717065500
LM85311416525xO




LM85055417000xO




LM8144017001500




LM8102717072500
Path 32  Row 32




               92/09/09




               85/09/06




     34/32     73/10/06




     35/32     72/08/01
               LM85311416531xO




               LM85055417003xO




               LM81440170045QO




               LM8100917073500
Path 32  Row 33




               92/09/09




               85/09/06




     34/33     72/08/18




     35/33     74/07/04
               LM85311416534xO




               LM85055417005xO




               LM8102617021500




               LM8171117004500
                               333

-------
Path 32  Row 34
     34/34



     35/34






Path 32  Row 35
     34/35
92/09/09




85/09/06




72/07/31




75/07/26










92/09/09




85/09/06




74/10/19
LM85311416540xO




LM85055417011xO




LM8100817024500




LM8218516570500











LM85311416543xO




LM85093816525xO




LM8181816522500
     Second WRS-1 not required?
Path 32  Row 36




               92/09/25




               86/09/25




     34/36     74/10/19
               LM85313016543xO




               LM85093816525xO




               LM8181816524500
     Second WRS-1 not required?
Path 32  Row 37
     34/37
92/09/09




85/09/06




74/10/19
LM85311416551xO




LM85055417023xO




LM8181816531500
     Second WRS-1 not required?
                               334

-------
Path 33  Row  26




               90/09/11




               84/08/25




     35/26     74/07/22




     36/26     74/07/23
               LM85238516544xO




               LM85017717045xO




               LM8172916570500




               LM8173017025500
Path 33  Row 27




               90/09/11




               84/08/25




     35/27     74/07/22




     36/27     74/09/15
               LM85238516550xO




               LM85017717051xO




               LM8172916573500




               LM8178417011500
Path 33  Row 28
     35/28




     36/28
92/08/15




86/08/15




74/07/04




73/07/10
LM85308916584xO




LM85089716574xO




LM8171116583500




LM8135217120500
Path 33  Row 29
     35/29




     36/29
92/08/15




86/08/15




73/07/09




74/07/05
LM85308916590xO




LM85089716580xO




LM8135117064500




LM8171217044500
                               335

-------
Path 33  Row 30




               92/08/15




               86/08/15




     35/30     72/08/19




     36/30     74/07/05
               LM85308916590xO




               LM85089716580xO




               LM8102717065500




               LM8171217050500
Path 33  Row 31
     35/31



     36/31
92/08/15




86/08/15




72/08/19




73/08/15
LM85308916592xO




LM85089716583xO




LM8102717072500




LM8138817125500
Path 33  Row 32



               92/08/15




               86/08/15




     35/32     72/08/01



     36/32     73/08/15
               LM85308916595xO




               LM85089716585xO




               LM8100917073500




               LM8138817131500
Path 34  Row 26
     36/26




     37/26
92/07/21




85/07/02




74/07/23




75/07/10
LM85306417040xO




LM85048817110xO




LM8173017025500




LM8216917053500
                               336

-------
Path 34   Row 27




               90/09/18




               84/09/17




     36/27     75/09/15




     37/27     74/09/16
Path 34  Row 28
     36/28




     37/28
90/09/18




84/09/17




73/07/10




72/09/08
               LM85239217011xO




               LM85020017113xO




               LM8178417011500




               LM8178517065500
LM85239217014xO




LM85020017115xO




LM8135217120500




LM8104717173500
Path 34  Row 29
     36/29




     37/29
91/09/05




84/09/17




75/09/01




72/09/08
LM85274417060xO




LM85020017122xO




LM8222216595500




LM8104717175500
Path 34  Row 30




               91/09/21




               84/09/17




     36/30     75/09/01




     37/30     72/09/08
               LM85276017063xO




               LM85020017124xO




               LM8222217002500




               LM8104717182500
                               337

-------
Path. 35  Row 26
     37/26




     38/26
92/08/13




86/07/28




75/07/10




73/07/12
LM85308717095xO




LM85087917093xO




LM8216917053500




LM8135417224500
Path 35  Row 27
     37/27




     38/27
92/08/13




86/08/29




74/09/16




73/07/12
LM85308717101xO




LM85091117084xO




LM8178517065500




LM8135417230500
Path 35  Row 28
     37/28




     38/28






Path 35  Row 29
      37/29




      38/29
92/08/13




86/08/29




72/09/08




74/07/25






>




92/08/29




86/08/29




72/09/08




74/09/17
LM85308717104xO




LM85091117091xO




LM8104717173500




LM8173217150500











LM85310317104xO




LM85091117093xO




LM8104717175500




LM8178617132500
                                338

-------
Path 35   Row 30




               92/09/30




               86/10/16




     37/30     72/09/08




     38/30     74/09/17
LM85313517103xO




LM85095917082xO




LM8104717182500




LM8178617135500
Path 36  Row 26




               91/08/18




               84/08/14




     38/26     73/07/12




     39/26     73/09/05
LM85272617170xO




LM85016617231xO




LM8135417224500




LM8140917273500
Path 36  Row 27




               91/08/18




               84/08/14




     38/27     73/07/12




     39/27     73/09/05






Path 36  Row 28




               91/08/18




               86/08/20




     38/28     74/07/25




     39/28     72/08/05
LM85272617172xO




LM85016617233xO




LM8135417230500




LM8140917280500











LM85272617174xO




LM85090217154xO




LM8173217150500




LM8101317285500
                               339

-------
Path 37  Row 26
     40/26
91/07/24




84/07/20




75/08/09
LM85270117230xO




LM85014117283xO




LM8511217085500
     Second WRS-1 not required?
Path 37  Row 27
     40/27
92/07/26




86/08/27




73/09/06
LM85306917225xO




LM85090917211xO




LM8141017334500
     Second WRS-1 not required?
Path 37  Row 28
     39/28



     40/28






Path 38  Row 26
     41/26
91/07/24




86/08/27




72/08/05




73/08/01










91/07/15




84/07/19




72/08/07
LM85270117235xO




LM85090917213xO




LM8101317285500




LM8137417344500











LM85269217290xO




LM84073417305xO




LM8101517392500
     Second WRS-1 not required?
                                340

-------
Path 39  Row 26
     42/26
92/08/09




85/07/21




74/07/29
LM85308317344xO




LM85050717415xO




LM8173617370500
     Second WRS-1 not required?
Path 40  Row 26
     43/26
92/07/31




85/07/28




72/08/27
LM85307417405xO




LM85051417480xO




LM8103517511500
     Second WRS-1 not required?
                              341

-------
     These  lists  of   triplicates   may  change  in  number  and
characteristics  over  time.    This  may  be  due  to  data  being
•unavailable,  or magnetic  tapes  may have  failed  over time  in
storage, or for a variety of other reasons.
                                342

-------
     Questions related to  the  characteristics  of the triplicates
may be  directed  to User Services  at  the EROS Data  Center.   The
phone number and address are provided in the overview on page 2.
                               343

-------
              Appendix    V:



Interagency Agreement (IAG)  Documentation
              344

-------
               United States Environmental Protection Agency
                       Washington, DC 2O460
                interagency Agreement/
                       Amendment
                 Part 1 - General Information
  6. Name and Address of EPA Organization
    US Environmental Protection Agency
    EMSL-LV, AMD/AMS
    944 E.Harmon  Ave.,  P.O. Box 93478
    Las Vegas, NV 89193-3478
  8. Project Title
                          1. EPA IAG Identification Number
                            DW14934750-OI-4
                          2. Other Agency IAG ID Number (if known)
                          3. Type of Action
                            Augmentation  INCR.
                          7. Name and Address of Other Agency
                            Interior Department  of the
                            U.S. Geological Survey
                            EROS Data Center
                            Sioux  Falls, SD  57198
4. Funding Location by
  Region

    XI
5. Program Office
  Abbreviation
             EROS Data  Center:  AVHRR Data/Product Support
  9. EPA Project Officer (Name, Address, Telephone Number)
    Lunetta,  Ross  S.
    EMSL-LV,  AMD/AMS
    Las Vegas, NV  89193-3478
    (702)  798-2175
  11. Project Period
   06/01/90 to 09/30/94
 13. Scope of Work (Attach additional sheets, as needed)"
        SEE ATTACHED REVISED SOW
                          10. Other Agency Project Officer (Name, Address, Telephone Number)	~
                            Lauer, Donald T., Department  of the Interior
                            U.S. Geological Survey, EROS  Data  Center
                            Science  and Applications Branch
                            Sioux Falls, SD  57198
                            (605)  594-6511
                          12. Budget Period
                            10/01/91 to 09/30/93
        EPA GRANTS SPECIALIST FOR THIS IAG
        IS FELICIA TARVER,  202-260-4392
 14. Statutory Authority for Both Transfer of Funds and Project Activities
     Economy Act 1932  Amended (31 USC 1535)
     Clean Air Act As  Amended Sec. 327 (b)  (6)
                                                           15. Other Agency Type
                                                                                     Federal
                                Previous Amount
                                                       Amount This Action
 16. EPA Amount
 17. EPA In-Kind Amount
 18. Other Agency Amount
 19. Other Agency In-Kind Amount
20. Total Project Cost
   Fiscal Information
 Program Elemenl
                                     Doc. Control No.
                                                                                Obligation/Deobligation Amt

                                                                                    10,000
683/40107
                             3CC926J011
EPA Form 1610-1 (Rev. 10-88)  Previous editions are obsolete.
                                                 345
                                                                 Page 1 of 5

-------
13.   Scope of Work
     The U.S.  Geological Survey,  EROS Data Center (EDC)  and U.S.
Environmental  Protection Agency (EMSL-LV) , will jointly participate
in the production and evaluation of Advanced Very High Resolution
Radiometer (AVHRR)- derived vegetation greeness maps and subsequent
land cover mapping efforts for the 1992 growing season.   Also, EDC and
EMSL-LV will work jointly on the Landsat MSS North American Landscape
Characterization Pathfinder (NALC) - Pathfinder Interagency Global
Change project.
     Since 1990, EDC has been providing EMSL-LV digital products of the
AVHRR-derived data from the Conterminous U.S. and Alaska greeness mapping
program.  This program is continuing in 1992 with the production of
greeness products of 1992 and 1989 AVHRR data from the Conterminous U.S.
and 1992 AVHRR data of Alaska.  EDC and the Canada Center for Remote
Sensing have begun developing prototype greeness data sets of North
America.  EDC will provide EMSL-LV all interim and final products of
North America as they become available.
     EDC- will continue to refine the AVHRR-derived land characterization
data base of the Conterminous U.S. and provide EMSL-LV with improved  land
cover data sets.  EDC will begin development of the North American land
characterization data base and provide EMSL-LV with data as they become
available.
     Cooperation between EMSL-LV and EDC  on the NALC-Pathfinder project
is anticipated  to continue well into the  decade.  The areas of  NALC       :
collaboration  include:  1) Landsat MSS 1992 data acquisition  for North
America, 2) cloud-free MSS composite data processing, 3) the  assembly    ;
of multiple date MSS coregistered data sets, 4) data  index and  archiving, \
and 5)  the development  of MSS  land cover  and change detection data analysis
techniques.  EDC  will collaborate on all  of the above tasks with EMSL-LV. •
However,  EDC  will  concentrate on the  assemblage of MSS multi-date scenes ;
coregistered data sets, image  compositing, data  index, and  archiving,;; and ;
data distribution.                              ,         ,         :>: :'.<     I
                                                                          f
I     Budget  (FY92/3)                                                      :
     	                              ... •                      '. »
     The following  is the  total FY92/3  budget  under this  IAG.  This  is the
final  budget  distribution  that reflect changes resulting  from the  release
of Landsat MSS data distribution rights  from EOSAT Corporation.
         Funding Elements
  FY92
                                                                    FY93
      Personnel (1992 AVHRR Support)  	
      Personnel (NALC Image Compositing)  	•	
      NALC Travel Support 	
      EPA NALC Graphics/Data Support 	
      NALC MSS Triplicate Assembly	
      NALC MSS Data Acq. Brokerage Assistance ......
      NALC MSS Data Acquisition 	
      Brazil MSS Data Acquisition (OEPER) 	
      EPA Contribution Access to SPOT(pan.)
       Interagency Acquisition  (OEPER/EMAP) . .	
      EPA Contribution to EROS Brazilian Retrieval
       of Landsat MSS from decaying HDDT's (OEPER)..
      Purchase of approx. 375 Landsat for Selected
       Locations in S.E. Asia by David Skole (OEPER)
      EDC Procurement of Mexico & C. Am. Topo Maps  .
$
$
$
$
   39. OK
   39. OK
    9. OK
   19. OK
$ 390.OK
$ 147.5K
$ 152.OK
$  16.6K
-0-
-0-
-0-
-0-
-0-
-0-
-0-
-0-
$  15. OK   $ 10. OK

$  20.OK    -0-

$  30.OK    -0-
$   5.225K  -0-
                               346
                                      TOTAL FUNDING
$ 882.325K  $  10.OK

-------
                        Part II - Approved Budget
                                                                                  EPA IAG Identification Number
                     22. Budget Categories
                                             ttemization of
                                             This Action
Itemization of Total Project
  Estimated Cost to Date
  (c)
                                                                        JOL
  (f)Prc
                                                                   io,nnn
  (i)
                                                                                                  917.32.
                                      $
  (k) Total

     (EPA Share  IQQ.00%)   (Other Agency Share
                                             Q_
                                0.00%)
                            :d, purchased,
                                                                   10.000
     (Identify all equipment costing $1,000 or more)
                                                             or rented with EPA funds?
                                                                      Yes
                                                                                                        No
 24. Are any of these funds being used on extramural agreements?  (See hem 22f)
 Type of Extramural Agreement
                                                                          X
                                                                              Yes
                                                                                            No
 Contractor/Recipient Name (if known)
                                  Grant
                                                  Cooperative Agreement
                     Total Extramural Amount Under This Project

                                  10,000
                                                          Procurement (Includes Small Purchase Order)
                                                                                  Percent Funded by EPA (if known)

                                                                                                   100.00
 25.
           Funds-Out Agreement
                                 Part III - Funding Methods and Billing Instructions
                  Disbursement Agreement
jr   Repayment
                                        '(Note: EPA Agency Location Code (ALC) - 68010727)
                                                                                      SF1081 °(SF 108°and
                                           Monthly
                                          Quarterly
                                                                                   Upon Completion of Work
                        Advance

26.

           Funds-ln Agreement

                 Reimbursement Agreement
                         Repayment

                         Advance
                 Allocation Transfer-In
Other Agency's LAG Identification Number
Other Agency's Billing Address  (Include Agency Location Code
                           or Station Symbol Number)
EPA Form 161O-1 (Rev. 10-88)
                                                             EPA Program Office Allowance Holder/Responsibility Center Number
                                        Other Agency's Billing Instructions and Frequency
                                                           347
                                                                                                            Page 2 of 5

-------
                    Part IV - Acceptance Conditions
                           EPA IAG Identification Number
                            DW14934750-01-4
27, Ganotal Conditions
   The other agency covenants and agrees that It will expedftiously initiate and complete the project for which funds
   nave been awarded under this agreement.
,58, Sp«e«al Conditions (Attach additional shoots if needed)



    See Attachment  A:   DBE Special Condition for  Interagency Agreement
                                        Part V - Offer and Acceptance
 Note:  1)   For Funds-out actions, the agreement/amendment must be signed by the other agency official in duplicate
           and one original returned to the Grants Administration Division for Headquarters agreements or to the
           appropriate EPA Regional IAG administration office within 3 calendar weeks after receipt or  within any
           extension of time as may be granted by EPA. The agreement/amendment must be forwarded to the
           address cited in Item 29 after acceptance signature.

           Receipt of a written refusal or failure to return the properly executed document within the prescribed time
           may result in the withdrawal of the offer by EPA. Any change to the agreement/amendment by the other
           agency subsequent to the document being signed by the EPA Action Official, which the Action Official
           determines to materially alter the agreement/amendment, shall void the agreement/amendment.

       2)   For  Funds-in  actions,  the  other  agency  will  initiate  the   action  and  forward  two  original
           agreements/amendments  to   the   appropriate    EPA   program   office   for   signatu're.   The
           agreements/amendments will then be  forwarded to the appropnate EPA IAG administration office for
           acceptance signature on behalf of the EPA. One original copy will  be returned to the other agency after
           acceptance.
     EPA tAG Administration Office (for administrative assistance)
                      EPA Program Office (for technical assistance)
 29, Organization/Address

  U.S. EPA

  Grants  Information and Analysis  Branch

  Grants  Administration  Division  (PM-216F)

  401 M.  Street, SW

  Washington,  DC   20460
           30. Organization/Address
           U.S. EPA, EMSL-LV,  AMD/AMS

           P.O. Box 93478

           Las Vegas, NV  89193-3478
                                                Certification
    All signers certify, that the statements made on this form and all attachments thereto are true, accurate, and
    complete. Signers acknowledge that any knowingly false or misleading statement may be punishable by fine or
    imprisonment or both under applicable law.
                          Decision Official on Behalf of the Environmental Protection Agency Program Office
 31 Signature
Typed Name and Title
 Wayne N.  Marchant

 Director, EMSL-LV
Date
                                Action Official on Behalf of the Environmental Protection Agency
 32. Signature
Typed Name and Title

W.  Scott  McMoran, Chief

Grants Info. &  Analysis Branch
Date
                                       Authorizing Official on Behalf of the Other Agency
 33, Signature
Typed Name and "Title
 Allen H.  Watkins, Chief

 National  Mapping Division
Date
 EPA Form 1610-1 (Rev. 10-88)
                                                   348
                                                    Page 3 of 5

-------
                                            DW14934750-01-4
DEE Interagency Agreements Special Conditions

     EPA's policy  requires at least  8% of its Federal  funding for
     prime  and subcontracts  be  awarded to businesses  or other
     organizations owned or controlled by socially and economically
     disadvantaged individuals.

          As a recipient of monies under this IAG, the Department
          of  the  Interior must  ensure  to  the  fullest extent
          possible  that  at  least  8%   of   funds for  prime  or
          subcontracts   and   subgrants   for  services  are  made
          available to businesses owned or controlled  by  socially
          and economically disadvantaged individuals, women-owned
          businesses,   and   Historically   Black  Colleges   and
          Universities.  (DBE)

     The Department of Interior must submit a report to EPA showing
     the  total  extramural  funds awarded   and  the  amount  and
     percentage of extramural funds awarded to DBEs by November 15,
     1992.  Reports should be  submitted to:

               Office of Small and Disadvantaged
                 Business Utilization (A-149C)
               U.S. Environmental Protection Agency
               401 M Street,  S.W.
               Washington, D.C.  20460
                             349

-------
                                                                                           Page 1 of 3
Untied State* Environmental Protection Agency
Washington DC 20460
.Q. Interagency Agreement /
EPA Amendment
Part 1 - General Information
6, Nam* and Addre»» of EPA Organization
EMSL-LV, AMD/AMS
944 E. HARMON AVE.
P.O. BOX 93478
LAS VEGAS, NV 89193-3478
1 . EPA / IAG Identification Number
DW1 4935881 -01-0
2. Other Agency ID Number
3. Type of Action
NEW
7. Name and Address of Other Agency
INTERIOR DEPARTMENT OF THE
USGS, EROS DATA CENTER
MUNDT FEDERAL BUILDING
SIOUX FALLS, SD 571 98
4. Funding Location
by Region
XI
5. Program Office
Abbreviation
EMSL-LV

8. Proiwt TMle EROS DATA CENTER AVHRR/NALC/PRODUCT SUPPORT
9. EPA Project Officer (Name, Address, Telephone Number)
ROSS S. LUNETTA
EMSL-LV. AMD/AMS
P.O. BOX 93478
LAS VEGAS, NV 89193-3478
(702)798-2175
11. Project Period
10/01/921009/30/94
10. Other Agency Project Officer (Name, Address, Telephone Number)
DONALD T. LAUER, DOI, USGS
EROS DATA CENTER, MUNDT FEDERAL BLDG.
SCIENCE & APPLICATIONS BRANCH
SIOUX FALLS, SD 57198
(605)594-6511
1 2. Budget Period
10/01/921009/30/94
13. Scope of Work (Attach additional sheets, »» needed)
            See Attached SOW
           THE EPA GRANTS SPECIALIST FOR THIS
           IAG IS FELICIA TARVER, 202-260-4392
CGMIl!
                                                                              VERi
14. Statutory Authority for Both Transfer of Funds and Project Activities 1 5. Other Agency Type
ECONOMY ACT OF 1 932 AS AMENDED FEDERAL
FUNDS
1 6. EPA Amount
17. EPA In-Kind Amount
18. Other Agency Amount
19. Other Agency In-Kind Amount
20. Total Project Cost
PREVIOUS AMOUNT
0
0
0
0
0
AMOUNT THIS ACTION
416,000
0
0
0
416,000
21. Fiscal Information
Program Element
CNRA1E
FY
93
Appropriation
683/40107
Doc. Control No.
JRS006
Account Number
3CNR26J012
AMENDED TOTAL






Object Class
25.71

Obligalion/Deoblig. Amt.
416,000
                                                   350

-------
                                                 DW14935881-01-0
 13.  Scope of Work
     The U.S. Geological  Survey,  EROS  Data  Center  (EDC)  and U.S.
 Environmental Protection  Agency  {EMSL-LV),  will  jointly  participate
 in the production and evaluation  of Advanced Very  High Resolution
 Radiometer  (AVHRR)- derived  vegetation greenness maps for  the 1993
 growing season and subsequent  land cover mapping efforts for the
 1992 and 1993 growing seasons.  Also,  EDC and  EMSL-LV will work
 jointly on the Landsat MSS North  American Landscape Characterization
 Pathfinder  (NALC) - Pathfinder Interagency  Global  Change project..

     Since 1990, EDC has  been providing EMSL-LV  digital  products
 of the AVHRR-derived data from the Conterminous  U.S. and Alaska
 greenness mapping program.   This  program is continuing in  1993  with
 the production of greenness  products of 1992 and 1993 AVHRR data
 from the Conterminous U.S. and Alaska.  EDC and  the Canada Center
 for Remote Sensing have developed prototype greenness data sets of
 North America.  The production of greeness  data  sets for North
 America will continue for the 1993 growing  season.  EDC  will provide
 EMSL-LV all interim and final products of North  America  as they
 become available.

     EDC will continue to refine the AVHRR-derived land
 characterization data base of the Conterminous U.S. and  provide
 EMSL-LV wxth improved land cover data  sets.  EDC will continue
 development of the North  American land characterization  1992  data
 base and provide EMSL-LV  with data as they  become available.

     Cooperation between  EMSL-LV and EDC on the NALC-Pathfinder
 project is anticipated to continue well into the decade.   The areas
 of NALC collaboration include:  l) cloud-free MSS -composite  data
 processing, 2)  the assembly  of multiple date MSS' coregistered data
 sets, 3)  data index and archiving, and 4)   the development of MSS
 land cover and change detection data analysis techniques.   EDC  will
 collaborate on all of the above tasks with EMSL-LV.  However,   EDC
will concentrate on the assemblage of MSS multi-date scenes
 coregistered data.sets,  image compbsiting,  data  index and  archiving
 and data distribution.


     Budget (FY93)

     The following is the total FY93 buJ.~ '  under this IAG.
        Funding Element
                                                  Total FY93 Funds
1993 AVHRR Support
Personnel (Image Compositing)
NALC Travel Support
EPA Graphics/Data Support
MSS Multidate Scene Coregistration
Pecora Symposium EPA Sponsorship
$
$
$
$
$
$
45. OK
45. OK
10. OK
15. OK
291. OK
10. OK
                              TOTAL FY92 FUNDING
$ 416.OK
                                    351

-------
                                                                                                                  Paga 2 of 3
Part II - Approved Budget
22. Budget Categories
(*) Personnel
(b) Fringe Benefit*
(c) Travel
(d) Equipment
{«} Supplies
(0 Procurement / Assistance
(g) Construction
(h) other Graphics Support
(I) Total Direct Charges
(1) Indirect Costs: Rate 0.00% Base $ 0.
(k) Total:
(EPA Share: 1 00.00%) (Other Agency Share 0.00%)
Itemization of
This Action
381,000
0
10,000
0
0
10,000
0
15,-ODO
416,000
0
416,000
EPA IAG Identification Number
DW1 4935881 -01-0
Hemizatlon of Total Project
Estimated Cost to Date
381,000
0
1 0,000
0
0
10,000
0
15,000
416,000
0
416,000
23. Is Equipment authorized to be furnished by EPA or leased, purchased, or rented with EPA funds?
 (Identify all equipment costing S10OO or more.)
                                                                                              Yes
                                                                                                      No
24. Are any of these funds being used on extramural agreements? (See Item 221.)   [x"l Yes
                                                                                   No
 Type of extramural agreement
                               Grant  I   I Cooperative Agreement
                                                                             Procurement (includes Small Purchase Order)
 Contractor / Recipient  Name  (if known)

  TBD

 (Pecora  Symposium)
                                              Total Extramural Amount under this Project

                                                               10,000
                                              Percent Funded by EPA  (if known)

                                                           100.00
                                   Part III - Funding  Methods and Billing Instructions
25.
   |X] Funds-Out Agreement     (Note: EPA Agency Location Code .(ALC) - 68010727)


        |XJ Disbursement Agreement


              IX | Repayment  Request for rep^  tent of actual costs must be itemized on SF-1080 and submitted to the Financial Management
                             Office, Cincinnati, OH 45268:

                                      I	I Monthly     |X | Quarterly    |	| Upon Completion of Work

                             Only available for'use by Federal agencies on working capital fund or with appropriate justification of need for
              C"] Advance    this type of payment method. Unexpended funds at completion of work will be returned to EPA. Quarterly cost
                             reports will be forwarded to the Financial Management Center, EPA, Cincinnati, OH 45268.
                          Used to transfer obligational authority or transfer of function between Federal agencies. Must receive prior
        I—I Allocation     approval by the Office of the Comptroller, Budget Division, Budget Formulation  and Control Branch, EPA Hdqtrs.
        I	I Tr.nsfcr nut   Forward appropriate reports to the Financial Reports and Analysis Branch, Financial Management Division, PM-
                          226F, EPA, Washington, DC 20460.                                          	
26.
   D
       Funds-ln Agreement
      I _ | Reimbursement Agreement
                                       B
Repayment
Advance
      I _ I
          Allocation Transfer-In
 Other Agency's IAG Identification Number
                                                                          EPA Program Office Allowance Holder/Resp. Center No.
                                                                              26J
 Other Agency's Billing Address (Include ALC or Station Symbol Number)
                                                                          Other Agency's Billing Instruction and Frequency
                                                                352

-------
                     Part IV - Acceptance Conditions
                             EPA IAG Identification Number

                             DW14935881-01-0
 27. General Conditions
    The other agency covenants and agrees that it will expeditiously initiate and complete the project for which funds
    have been awarded under this agreement.
 28. Special Conditions (Attach additional sheets if needed)
     .See Attached:      DBE Special Condition for Interagency Agreement
                                        Part V - Offer and Acceptance
 Note: 1)   For Funds-out actions, the agreement/amendment must be signed by the other agency official in duplicate
           and one original  returned to the Grants Administration !?:---:- ?-.  '      idquarters agreements or to the
           appropriate EPA  Regional IAG administration office within 3 calendar weeks after receipt or within any
           extension of time as may be  granted by EPA. The agreement/amendment must be forwarded to the
           address cited in Item 29 after acceptance signature.

           Receipt of a written refusal or failure to return the properly executed document within the prescribed time
           may result in the  withdrawal of the offer by EPA. Any change to the agreement/amendment by the other
           agency subsequent to the document being signed by the EPA Action Official, which the  Action Official
           determines to materially alter the agreement/amendment, shall void the agreement/amendment.

       2)   For  Funds-in  actions,   the  other  agency  will   initiate  the  action  and   forward  two   original
           agreements/amendments  to   the   appropriate   EPA   program   office   for   signature.   The
           agreements/amendments will then be forwarded to the  appropriate  EPA IAG  administration office for
        ,   acceptance signature on behalf of the EPA. One original copy will be  returned to the other agency after
           acceptance.                                                                        y   y
     EPA IAG Administration Office (for administrative assistance)
                       .EPA Program Office (for technical assistance)
 29. Organization/Addres
    U.S.  EPA

    Grants  Information and  Analysis Branch

    Grants  Administration Division (PM-216F)
    401  M Street,  SW

    Washington, DC  20460


                •••'.c .«t' ,  £c ,.p •'-, X-T-^ -~ - ,.„ ^r-,—-^.;
           30. Organization/Address

              U.S.  EPA,  EMSL-LV,  AMD/AMS

              P.O.  Box 93478

              Las Vegas,  NV 89193-3478
                                                Certification
    All signers certify that the  statements made on this form and all  attachments thereto are true, accurate, and
    complete. Signers acknowledge that any knowingly false or misleading statement may be punishable by fine or
    imprisonment or both under applicable law.
                         Decision Official on Behalf of the Environmental Protection Agency Program Office
31 Signature
                                           Typed Name and Title
                                             Wayne  N. Marchant

                                             Director,  EMSL-LV
                                           Date
32. Sigr-.ature
                                Action Official on Behalf of the Environmental Protection Agency
             /*--.-/.'/> A
Typed Name and Title
  W. Scott  McMoran, Chief

  Grants Info.  &  Analysis  Branch
                                                                                      Date
/
                                      Authorizing Official on Behalf of the Other Agency
 ;3. SignalL/e
Typed Name and Title
  Allen H.  Watkins, Chief

  National  Mapping Division
                                                                                      Date
EPA Form 1610-1 (Rev. 10-88)
                                                          353
                                                                                               Page 3 of 5

-------
                                            DW14935881-01-0
DBE Interagency Agreements Special Conditions

     EPA's policy  requires at least  8% of its Federal funding for
     prime  and subcontracts  be  awarded to  businesses  or other
     organizations owned or controlled by socially and economically
     disadvantaged individuals.

          As a recipient of monies under this IAG, the Department
          of  the  Interior must  ensure  to  the  fullest  extent
          possible  that  at  least  8%   of   funds for   prime  or
          subcontracts   and   subgrants   for   services  are  made
          available to businesses owned or controlled by socially
          and economically disadvantaged individuals, women-owned
          businesses,   and   Historically   Black  Colleges   and
          Universities.  (DBE)

     The Department  of the Interior must submit a report to EPA
     showing the total extramural funds  awarded  and the amount and
     percentage of extramural  funds awarded to DBEs by November 15,
     1993.  Reports should be submitted to:

               Office  of Small and Disadvantaged
                 Business Utilization (A-149C)
               U.S. Environmental Protection Agency
               401 M Street,  S.W.
               Washington, D.C.  20460
                               354

-------
        EPA-EOSAT 1992  LANDSAT MSB DATA ACQUISITION AGREEMENT


     To satisfy EPA's requirement  to  have MSS data  acquired,  and
 permit EPA the flexibility to purchase multiple  scenes  of cloud-
 prone locations,  EPA proposes the following conditions for 1992
 MSS  data collection:

   o    The  EPA's   1992   Landsat  MSS data  purchases   will   be
 accomplished through the EROS Data Center (EDC) Federal Government
 Brokerage Agreement.   The EPA path /  row and  acquisition windows
 are  being provided  to  EDC  for transmittal to EOSAT.

   o  EOSAT  will utilize its  best efforts to  provide MSS data
 collects to meet EPA requirements.  This includes negotiating with
 foreign  receiving stations to provide MSS data  for EPA's  projects.

   o EOSAT agrees to collect Landsat MSS data on a continual basis
 for  the  conterminous U.S., Mexico, and portions of Central America
 and  the  Caribbean  that are  within the collection  range of  the
 Norman,  Oklahoma receiving  station,  starting from  the  time  the
 Norman  station becomes operational, until  the conclusion of  the
 1992  growing season.

   o EPA agrees to purchase a minimum of one  MSS  scene  at 30% or
 less  cloud  cover  for  each  of  the  EPA specified  path  / rows
 collected by the Norman receiving  station during the 1992  growing
 season.   Cloud cover determinations will be based on  the  percent
 cloud cover over land areas.   Cloud cover will be determined by  the
 visual inspection of image film products at  EROS  Data Center  (EDC)
 by EPA  or EPA's designated representative at EDC.   Cirrus clouds
 will  be  included in the percent cloud  cover  determinations.

   o   For those path / row  scenes requested by EPA that  require
 TDRSS acquisitions, EPA agrees to purchase all  MSS data having  30%
 or less  cloud cover, collected within EPA  specified  acquisition
 windows.  EPA has the option to close the TDRSS  acquisition windows
 for  specific path / rows by  notifying EOSAT at  least seven days
 prior to the next scheduled acquisition. Notification  will be  via
 FAX by EPA or  its EDC   designate.

   o  For  the  EPA specified path /  rows, EPA may purchase any MSS
 scene meeting the  cloud  cover specification  of  30%  or less  at
 $1,000.00  per  initial  scene  and  $500.00 per   scene  for  all
 subsequent scenes of the same path/row. If  for a given path / row
 there is  no scene meeting  the 30% cloud cover  specification, EPA
may,   at its option,  purchase any MSS scenes collected within EPA's
 specified  acquisition window at $500.00 per scene.

   o  EPA intends to purchase approximately $ 520,000.00 of Landsat
MSS data by December 31, 1992.  All data purchases are subject to
the availability of government funds and  the availability of MSS
data  for the EPA .specified path / rows and acquisition windows.

                              355

-------
                        Appendix   VI:

   Mexico  Worksheet  for Acquisition  and  Determining  Temporal
Windows for Acquisition of Remotely Sensed Imagery
                               356

-------
                             Mexico Worksheet
                               (by Path/ Row)
                              January 31, 1992
 Path/Re
 24/47
 24/48
 24/49
 25/46
 25/47
 25/48
 25/49
 26/43
 26/44
 26/45
 26/46
 26/47
 26/48
 27/42
 27/43
 27/44
 27/45
 27/46
 27/47
 27/48
 28/41.
 28/42
 28/43
 28/44
 28/45
 28/46
 28/47
29/40
29/41
29/42
29/43
29/44
29/45
29/46
29/47
30/40
30/41
30/42
30/43
30/44-
30/45
30/46
31/40
Phenological Window
Jul-Sep
N/A
N/A
N/A
N/A
Jul-Sep
N/A
N/A
Jun-Aug
N/A
Jun-Aug
Jun-Aug
Jun-Aug
N/A
N/A
Jun-Aug
Jul-Sep
Jul-Sep
Jun-Aug
Jun-Aug
N/A
N/A
N/A
N/A
Jun-Aug
Jun-Aug
Jun-Aug
N/A
N/A
N/A
N/A
Jul-Sep
Jul-Sep
Jul-Sep
Jul-Sep
N/A
N/A
N/A
N/A
     N/A
N/A
N/A
Jul-Sep

                 357
Cloud Window
Mar-May
Feb-Apr
Jan-Mar
N/A
Feb-Apr
Feb-Apr
Feb-Apr
N/A
Jan-Mar
N/A
Feb-Apr
Jan-Mar
Jan-Mar
Jun-Aug
N/A
N/A
Jan-Mar
Jan-Mar
Jan-Mar
Jan-Mar
Jun-Sep
N/A
Apr-Jun
Mar-May
Feb-Apr
Mar-May
Feb-Apr
N/A
N/A
Jun-Aug
Apr-Jun
Mar-May
Mar-May
Mar-May
Feb-Apr
N/A
Apr-Jun
Apr-Jun
Apr-Jun
      Mar-May
Feb-Apr .
Mar-May
Mar-May
Date of'86 Coverage
N/A
N/A
Mar/Apr
N/A
Mar/Apr
Mar/Apr
Mar/Apr/May
N/A
N/A
N/A
May
Mar/Apr
Mar/Apr
N/A
Mar
Mar
Mar
Mar
Mar
Mar
Apr
Apr
Apr
Apr
Apr
Apr
Apr
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
     Mar
Mar
Mar
N/A

-------
31/41
31/42
31/43
31/44
'32/39
*32/40
•32/41
•32/42
"32/43
33/39
33/40
33/41
33/42
3a'43
33/44
34/39
34/40
34/41
34/42
34/43
34/44
35/39
35/40
35/41
35/42
35/43
36/39
36/40
36/41
36/42
•Jul-Sep
N/A
N/A
N/A
      N/A
      N/A
      N/A
      N/A
      N/A
Jul-Sep
Jul-Sep
N/A
Jul-Sep
Jul-Sep
N/A
N/A
N/A
Jul-Sep
Jul-Sep
N7A
N/A
N/A
A«g-0ct
Aug-Oct
N/A
N/A
N/A
N/A
N/A
N/A
Apr-Jun
Apr-Jun
Jan-Mar
Nov-Jan
      Fob-Apr
      Fob-Apr
      Fcb-Apr
      Feb-Apr
      No v-Fob
Apr-Jun
Apr-Jun
Feb-Apr
Apr-Jun
Apr-Jun
Mar-May
Mar-May
Mar-May
Mar-May
Mar-May
Mar-May
Mar-May
Apr-Jun
Apr-Jun
Apr-Jun
Apr-Jun
Apr-Jun
Apr-Jun
Apr-Jun
Apr-Jun
May-Jul
Mar
Mar
Mar
Mai-
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
Mar
N/A
N/A
N/A
Apr
Apr
Mar
Mar
Mar
Mar
     Apr
     A p r
     Apt-
     Apr
     Apr
* Scene exists for March 1991, with less than 10 % clouds.
                                    358

-------
  O

  rH
  O
 CJ



 rH
W

O
d
a
Oj
rH
O
a
 be
 a
 CD
Q
            bJO
            03
            a
            CD
            02
            a

           CO
            
-------
                                                               PI' •'
 CD
 co

 a
 CD
 co
.S   8
 o   be
 a   d

 8   g
 rH  TH

 be  2

 S   S
•rH   O
 CO   JH
 rj   O

 CO   B
 o>   d
•rH   £1
13   H

3*8

^   «
^3  .3

•8£
 §  .2
 PI   s
•rt   cr1


 §   s
 CJ   CD

"S  r£

 §1
             cd
 a
 cfl
 •M
 cd
 ft  oj
                       rH

                       CD
                           CD
                       CO
                       o

             cd
 «5  .-S


-23

 §   ^

 03   2
 CJ  rM

.cd  ^

     CD
     CJ


 ™   CD

     S3

    CS
 CO   fH

 JH  * ^

 0   >>
•-C  ^


 S   d1
•rH   H

 b   O

 g  4^

rZ   ^

 S   S


 2   S


 ^rS

 §  ^

EH   CD
                                      cd
                                      o
rH

CJ


CD

rH

cd


CO
                                     rO
                                      cd
                                     •rH
                                      rH
                                      Cd 4J
                                      >  o


                                     •5-1
                                      CJ  P
                                     •rH  rH
                                     r£3  P.



                                      £  s
                                      CD r-H
                                      ti  r^
                                     •rH  CJ


                                      S '£
                                      ^  b
                                      CD  cd

                                     •8  ft
                                     •-d  cd
                                          CO
                                          CO
                                      a  ^
                                      I  8
                                      cd  co

                                      ,2  03
                              PH -4.3


                                  O
                                            CD
                                                    cd
 cd


 cr

13

 rt
 cd



 >
•rH
•4J
•rH

 d
-4J

 a
•rH

 fl   rH
 O   CD

 co   £3

 CD   H

r^H   ^

 S'S

     a
     o
    'rH
     CO




 §1
r3   »

 S   ^
 CD
              ^^  -4-3
              ! **  -
              "^^^"  t^^^j

              a  o
              CD

              ?-»  r
              rH  r

              d  •

              O  "co
                                                        CO
o
cd
                   36Q

-------
CJ
cd
o
rH
PH
PH
CD
CJ
fl
cd
•rH
rH
cd
>
f-j
CO
•rH
4-3
cd
Qd
r—H
CO

a
cd
.
cd
rH
O
PH
8
CD
4-3
CD
O
be
a
•rH
An understand

rj
M
O •
_ ^^j
• ^^
CO
CO
•rH
a
o
4_3

13
•rH
CJ
d
rH
CO
•rH
4-3
CO
CD
gion of inter
relating to a re;
planning.
CD
•"^
2
rj
0
CO
CD
r—H
rQ
cd
• »»4
V pll 1^
rH
cd
^
rH
cd
r-H
CJ
• rH
rH
cd
PH
cd
o
CD
The significanc






rA
QO
CD
Id
rH
v^s
CO
a
o
•rH
*CO
•rH
CJ
cd
CD
O
4-3
incorporated in

O'
/i\
vU
CO
^^
• rH
CJ
PH
CO
cd
CO
• rH
m
CO
Q
s — '
a
CD
.CO
CO
rH
0
PH
A Decision Sup
rH
CD
cd
^>i
£
0
• rH
CO
•rH
Q
CD
>4>
•^^^
CD
^
r—H
O
a
•rH
0
.s designed t
application too]









CO
CO
v^
CD
CJ
O
rH
_j
rH
0
•rH
r— <
O
CO
CD
interactively in
CD
Id
• £H
PH
O
rH
PH
PH
rrt
"•w
CD
r~{
-^
be
a
• rH

"cd
rH
O
PH
rH
O
CJ
a
•rH
CO
Q
r—H
cd
rH
o
^^
ft
a
CD
r^^i
^^^i
0
•rH
"cd
p,
•V
^
o
rH
PH
rS
3
0
^
CD"

rH
CD
4-3
«_4
rH
•rH
CD

•rH
•rH
^
•rH
a
cd
-j_j
parameters, wii
•
CO
^CD
ho
CD
4-3
cd
rH
4-3
CO
rH
O
•rH
4-3
•rH
CO
•rH
cr
CJ
cd
t» »\
be
— »
r determinin
a framework fo:
361

-------
None
 Jone
 None
              Phenological "Seasons"
MSS Cloud Cover


                Temporal Continuity
             (1986 Existing MSS Coverage)
                Other Considerations
            (Spatial Continuity, Priorities)
                         362

-------
                 Appendix  VII:



NALC / GCRP Coordination / Collaboration Documents
                           363

-------
          NALC / 6CRF  Coordination / Collaboration

                         Mexico
NALC Task:

Delivery of standard  and modified NALC products focused on humid
tropical forested regions, in particular Mexico

GCRP Task:

Characterize changing land cover and land use and forest types in
Southeastern Mexico

GCRP Contact:

Point of contact is Michael A. Cairns, EPA-ERL-C
   Phones:  Cairns, 430-4777 or 503-754-4777 FAX -4799

GCRP Goals:

A goal of ERL-C is assess terrestrial biosphere management options
as  they  influence land cover,  carbon  pools  and  feedback  to
atmospheric conditions.  To assist these efforts it is desirable to
supply good  land  cover and land use  (LC/LU)   images  and tabular
data  for  areas of  interest.    These data  will be  useful  in
calculations of carbon pool, change in the pool or carbon flux, and
for model calibration  and verification activities.  The current and
past condition, and change  in these  carbon pools are significant
information for the charge of  evaluating natural and human-induced
atmospheric  conditions  and their affects  on ecosystems.   This
effort can support the understanding of feedbacks of change in land
surface characteristics  on atmospheric conditions.

Estimates of total land cover and change in cover will be useful in
calculations of carbon pool, change in the pool or carbon flux, and
for model calibration and verification  activities.  Quantification
of current and past condition,  and change  in these carbon pools is
significant in evaluating natural and human-induced  influences on
forests.  This effort also supports the understanding of feedbacks
from change in terrestrial landscapes on atmospheric conditions, as
well as having utility in planning of substitutes for fossil fuels.

Use of  remote sensor data  and the objectives  of North American
Landscape  Characterization  (NALC)  can supply the  LC/LU  data and
other  products.    LC/LU  data  can  be  linked  with  management
information to obtain  variables and develop quantities for land use
descriptors  including:  total  forest  area,  commercial  forests,
agroforest  systems, pastures,  total area  of  agricultural lands,
urban areas, park and preserves, and other  land cover/land uses.

Early work on data sets have been  slated for the  Southern and
Southeastern  portions of  Mexico.    Coordination with ERL-C has
prioritized  the states  of  Chiapas,  Quintana Roo, Campeche,  and
Tabasco for early data acquisition and analyses.

                                 364

-------
 NALC Applications (Standard Products):

 The standard products that apply to this work element would be land
 cover/land use  images and tabular data,  and evaluations of change.


 NALC Applications (Modified Products):

 NALC data  products and  analyses  will  provide  assistance  with
 characterizing  and  quantifying  certain  variables.    These may
 include optimization of images for detection of land use conditions
 such  as   agroforestry,   natural  regeneration,   and   increase
 sensitivity to forest  species composition.   Analysis of  change
 induced  by fire,  pestilence,  and weather  is  necessary,  and the
 management practices subseguent to the event would be of interest.
 General land use  practices  may be derived from NALC remote  sensor
 data.


 Scheduling for  Deliveries of Standard Products:

 Current  NALC  scheduling  calls   for  the  above  products  to be
 delivered   in  calendar years 1995  for  Mexico,   the  Caribbean,
 Chesapeake Bay Watershed,  and Central America.  Alaska and Western
 US  areas will be available in 1996.  In 1997  the midwest and Great
 Plains and Eastern-Southern US will be available.


 Potential  Schedule for Modified NALC Products:

 The above  issues may be addressed in preliminary tests of  the NALC
 program.  The levels of sensitivity can be  evaluated, and suggested
 modified  products could be generated in  support of  GCRP.   The
 timeframe  would be the same  from  the standard products.


 Proposed Coordination Action/Schedule:

Much  of  this  effort  will  involve production of  data sets and
 standard and modified products by Cooperators.  Hence,  it  would be
desirable  to  collaborate  on Cooperator tasks  to  insure that the
products   of  interest  are  addressed  early  in  the   delivery
timeframes, and  that  modified products of  interest to GCRP will
generated.  In the case of the Mexico effort,  it would be desirable
to  meet with the  Cooperators  soon to develop plans.   Potential
cooperators  include  the Institute  of  Geography  at the National
Autonomous University of Mexico.
                              365

-------
         NALC / GCRP  Coordination / Collaboration

                          Forests

NALC Task:

Delivery of standard and modified NALC products focused on forested
regions of the US

GCRP Task:

Potential of global forested  and agroforested lands to sequester
and conserve carbon


GCRP Contact:

Points of contacts  are Robert Dixon and Jack Winjum, EPA-ERL-C
   Phones:  Dixon,  430-4772 or 503-754-4772 FAX -4799


GCRP Goals:

The goal of ERL-C is to evaluate global forest management practices
and  to  determine   the  potential  for  management  options  for
conservation of carbon and sequestration in tropical, temperate and
boreal forests.     These efforts  can  be  facilitated by supplying
good land  cover  and land use  (LC/LU)  images and tabular data for
areas  of interest.   These  efforts would  include evaluations of
change in  land cover due to  natural  phenomena and through human
activities.    These  data  can  be combined  with  information  on
management  practices, and  be used to study  the  forest related
management options  in a given region.
Estimates  of land  cover  and change  in  cover will  be  useful in
calculations of carbon pool,  change in the pool or carbon flux, and
for model calibration and verification activities.  Quantification
of current and past condition, and change in these carbon pools is
significant  information for  evaluating natural and  human-induced
influences on forests.  This effort also supports the  understanding
of feedbacks from change  in terrestrial  landscapes on atmospheric
conditions,  as  well as having utility in planning of substitutes
for fossil fuels.

Us© of remote  sensor  data and the  objectives  of North American
Landscape  Characterization (NALC) can supply  the LC/LU data and
other  products.   Early work  on data sets has  been  slated for
portions of  the Pacific Northwest,  Chesapeake Bay and for a  study
area  in Alaska.   These  efforts  will address:  a)  capability to
detect land  cover classes of  interest,  and b)  ability  to detect
change in land cover and quantify  that change  at  the scale of
interest.                     „__
                              3oo

-------
Later  needs for  satellite derived  data  include the  greater US
Pacific  Northwest region  with emphasis  on forested  areas, the
Canadian Pacific Northwest regions,  and the trans-boreal portions
of Canada and Alaska.


NALC Applications  (Standard Products):

The standard products that apply to this work element would  be land
cover/land use images and tabular data,  and  evaluations of  change.


NALC Applications  (Modified Products):

Potential   areas   of  NALC  data   analysis  assistance   include
optimization of images for detection  of land use conditions  such as
agroforestry, natural regeneration,   and  increase sensitivity to
forest species composition.  Analysis of  change induced by  fire,
pestilence, and weather is necessary, and  the management practices
subsequent  to the event would be of interest.   General land use
practices may be derived from remote sensor data.


Scheduling for Deliveries of Standard Products:

Current  NALC scheduling  calls for the  above  products  to  be
delivered  in calendar  years   1995  for  Mexico,  the  Caribbean,
Chesapeake Bay Watershed, and Central America.   Alaska and Western
US areas will be available in 1996.   In  1997 the midwest and  Great
Plains and Eastern-Southern US will  be available.


Potential Schedule for Modified NALC Products:

The above issues may be addressed in  preliminary tests of the NALC
program.  The levels  of sensitivity can be  evaluated,  and suggested
modified products  could be  generated in support of GCRP.   The
timeframe would be the same from the standard products.


Proposed Coordination Action/Schedule:

Much of  this effort will involve  production  of data sets and
standard and modified products  by Cooperators.   Hence, it would be
desirable to collaborate on  Cooperator tasks to  insure that the
products  of  interest   are  addressed  early   in  the  delivery
timeframes, and that modified  products on interest  to  GCRP will
generated.
                              367

-------
            NALC / GCRP  Coordination / Collaboration

                           Tropical Forests
NALC Task:
Delivery  of  standard  and  modified  NALC  products  focused  on
tropical regions

GCRP Task:

Characterize changing land cover and land use and forest types in
humid tropical regions of interest in the NALC area of operations
including Mexico, the Caribbean and Central America

GCRP Contact:

The point of contact is Michael A.  Cairns, EPA-ERL-C.  A number of
other EPA groups and NASA, and others are interested  in the results
of this effort.
    Phones:  Cairns, 430-4777 or 503-754-4777 FAX -4799


GCRP Goal:

The goal is to determine the quantity of tropical forests and the
rate of conversion to other land uses.  These data are valuable for
estimating  the current  abundance  of  carbon  sequestered  in  the
regions,  and  allow  better  model  inputs  and  calibration  and
verification.  The determination of change in land cover and land
use (LC/LU) will also assist in the estimation of carbon pool and
the flux or dynamics of the pool.

Use of remote  sensor data will  supply  a number of  useful data or
information^   These would include LC/LU, detection of  change in
LC/LU, and potential use of remote  sensor data as input to habitat
or resource quality models.

Pilot studies and production of standard NALC product will occur on
tropical forest areas within North America.   A separate effort in
Brazil  will  be  conducted under  the  Deforestation  Monitoring
program.  The regions for early attention will include the states
of Southern or Southeast Mexico,  the  Atlantic Forest  region of
Brazil, the island of Hispaniola, and other Caribbean islands.

Early work on data sets have been  slated  for the Southern  and
Southeastern  portions  of Mexico.    Coordination  with  ERL-C  has
prioritized the  states of Chiapas,  Quintana  Roo,  Campeche,  and
Tabasco for early data acquisition and analyses.
                             368

-------
NALC Applications  (Standard Products):

The standard products that apply to this work element would be land
cover/land use images and tabular data,  and evaluations of change.
Unprocessed data and standard image products may also be useful.
NALC Applications  (Modified Products):

NALC  data products  and  analyses will  provide  assistance  with
characterizing  and  quantifying  certain  variables.   These  may
include optimization of images for detection of  change in land use
conditions  such as  forest harvesting  and  forest  regeneration.
Analysis  of  change induced by fire, pestilence, and  weather is
necessary, and  the management practices subsequent to  the event
would be of interest.


Scheduling for Deliveries of Standard Products:

Current  NALC  scheduling  calls   for  the  above  products  to  be
delivered  in  calendar  years  1995  for Mexico,  the  Caribbean,
Chesapeake Bay Watershed, and Central America.   Alaska and Western
US areas will be available in 1996.   In  1997  the midwest and Great
Plains and Eastern-Southern US will be available.


Potential Schedule for Modified NALC Products:

The above issues may be addressed in  preliminary tests of the NALC
program.  The levels of sensitivity can be evaluated, and suggested
modified products  could  be generated the data production work by
Cooperators in support of  GCRP.   The timeframe would be the same
from the standard products.


Proposed Coordination Action/Schedule:

Early work has been slated  for this activity in Brazil, Mexico, the
Caribbean,  and  Central  America.   It would  be  desirable  to
communicate  on  a  regular  basis,  to  facilitate  exchange  of
information.


Much  of this  effort will  involve production  of data  sets  and
standard and modified products by Cooperators.   Hence, it would be
desirable to collaborate  on Cooperator contracting issues to insure
that the products of interest are addressed early in the delivery
timeframes,  and  that  modified  products   of   interest  to  GCRP
researchers will be generated.
                              369

-------
              NALC / GCRP  Coordination / Collaboration



                 Atmospheric Trace Gases and Land Cover

NALC Task:

Delivery of  standard and modified NALC products  focused  on land
cover and land cover changes for the Southeastern US.

GCRP Task:

Characterize changing agricultural  land  cover  types and how they
may influence atmospheric trace gases

GCRP Contact:

Point of contact is Lee Mulkey, EPA-ERL-A


GCRP Goals:

A goal of ERL-A is  assess terrestrial biosphere management options
as  they influence  land  cover, atmospheric  trace gases  such  as
carbon and methane,  and  feedbacks  to  atmospheric conditions.   To
assist these efforts it is desirable to supply good  land cover and
land cover change  images and tabular  data for  areas of interest.
These data will be  useful in calculations of carbon pool, change in
the pool or  carbon flux, evaluations  of  sources/sinks of methane
and  nitrous  oxide,  and   for  model calibration  and verification
activities.  The current and past  condition, and change in these
land cover are significant information for the charge of evaluating
natural and human-induced atmospheric  conditions  and their affects
on  ecosystems.   This effort  can  support the  understanding  of
feedbacks of change in land surface characteristics  on atmospheric
conditions.


Use  of  remote sensor data and the objectives of North American
Landscape Characterization (NALC)  can supply the land  cover and
change in land cover data and other products.   These data can be
linked with management information to  obtain variables and develop
quantities   for   land   use   characteristics   including:   total
agricultural area and change in area,  commercial  farms and general
crop types, agroforest systems, conservation reserve areas, urban
areas, park and preserves, and other land  cover/land uses.

Early work on data  sets have been slated for  the Southern and
Southeastern  portion of  the US.    Coordination  with  ERL-A will
prioritize the states  of interest  for early data acquisition and
analyses.

                               370

-------
NALC Applications  (Standard Products):

The standard products that apply to this work element would be land
cover/land use images and tabular data,  and evaluations of change.


NALC Applications  (Modified Products):

NALC  data products  and  analyses will  provide  assistance With
characterizing  and  quantifying  certain  variables.   These  may
include  optimization  of  images  for  detection  of land cover
conditions such  as agricultural areas and change in these  areas,
conservation research program areas,  and wetland areas.   Analysis
of change induced  by land cover conversion, fire, pestilence, and
weather is necessary, and the land cover change subsequent  to the
event would be of  interest.


Scheduling for Deliveries of Standard Products:

Current  NALC  scheduling  calls  for  the  above  products  to  be
delivered in calendar year 1997, when the midwest and Great  Plains
and Eastern-Southern US will be available.


Potential Schedule for Modified NALC Products:

The above issues may be addressed in  preliminary tests of  the NALC
program.  The levels of sensitivity can be evaluated,  and suggested
modified  products  could be  generated in  support  of GCRP.   The
timeframe would be the same from the standard products.


Proposed Coordination Action/Schedule:

Much of  this  effort  will  involve  production of  data sets  and
standard and modified products by Cooperators.  Hence, it would be
desirable to  collaborate on Cooperator tasks to  insure  that the
products  of  interest   are   addressed  early  in  the  delivery
timeframes,  and  that  modified products of  interest  to GCRP will
generated.
                             371

-------
           NALC / GCRP  Coordination / Collaboration

             Remote Sensor Methods Development for
               Identification of Methane Sources
NALC Task:
Delivery of standard and modified NALC products focused on wetland
and agricultural  regions of selected areas of the US  and North
America
GCRP Task:

Calibration and verification of models of atmospheric trace gases
in particular methane using detail land cover and land cover change
data


GCRP Contact:

Point of contacts are Lee Mulkey and Richard Zepp, EPA-ERL-A
GCRP Goals:

The goal  of ERL-A  is  to evaluate the  influence  of agricultural
inputs, outputs and wastes that alter atmospheric concentrations of
trace gases, and study  how these gas flux influence global climatic
conditions and evaluate feedbacks to climate.  These  efforts can be
facilitated  by supplying good  land  cover data,  NDVI  or biomass
images, and  land cover change data for areas of interest.  These
data could assist in evaluating the role of change in agricultural
or wetland management practices as they influence natural phenomena
such as atmospheric and terrestrial fluxes of methane.


Estimates  of land  cover  and change  in cover will be  useful in
calculations  of  methane  flux,  and  for model  calibration  and
verification activities.  Quantification of current and past land
cover condition,  and change  in these  conditions is significant in
evaluating natural and human-induced influences on landscape.  This
effort also supports the understanding of feedbacks  from change in
terrestrial landscapes to atmospheric  conditions, as well as having
utility in planning of alternative management practices.

                              372

-------
Use  of remote  sensor  data and the  objectives  of North American
Landscape Characterization (NALC) can  supply  the land cover  and
change in land  cover data, and other products.   Early work  on data
sets  has been  slated  for portions of  the  US  and North America.
These efforts will address:  a)  capability to  detect land cover
classes of interest, and b) ability to detect change in land covers
such  as agricultural type and wetlands.
NALC Applications  (Standard Products):

The standard products that apply to this work element would be  land
cover  images, tabular data, and evaluations of  land  cover  change.
NALC Applications  (Modified Products):

Potential   areas  of  NALC   data   analysis  assistance  include
optimization of images for detection of land cover conditions such
as  agricultural crop  composition  and wetland  growth structure
(shrub vs. tree wetlands) and pattern.   Analysis  of change induced
by  fire  and weather is necessary,  and regeneration subsequent to
the event would be  of  interest.


Scheduling  for Deliveries of  Standard Products:

Current  NALC  scheduling  calls  for  the  above products  to  be
delivered in calendar  year  1997  for the midwest and Great Plains
and Eastern-Southern US.  .


Potential Schedule  for Modified NALC Products:


The above issues may be addressed in preliminary  tests  of the NALC
program.  The levels of sensitivity can be evaluated, and suggested
modified  products  could be  generated in  support  of  GCRP.   The
timeframe would be  the same from the standard products.


Proposed Coordination Action/Schedule:

Much of  this  effort  will  involve  production of  data sets  and
standard and modified products by Cooperators.  Hence,  it would be
desirable to collaborate on Cooperator tasks to insure  that the
products  of   interest   are   addressed  early  in   the  delivery
timeframes,  and  that  modified products on  interest to GCRP  will
generated.
                               373

-------
             NALC / GCRP  Coordination / Collaboration

                Trans-Boreal/ Former Soviet Union
NALC Task:
Delivery of standard and modified  NALC products focused on pilot
efforts to characterise land cover  /  land use in Alaska.  Identify
sources to  obtain additional remote sensor-derived data  on land
cover or NDVI for areas in the former Soviet Union
GCRP TasX:

Characterize the quantity and variety of land cover in the former
Soviet Union :and Trans-Boreal ecosystem, and determine change, and
utilize the data in .modeling of carbon and methane flux


GCRP Contact:

The points of contact are Jeff Lee of ERL-C, and Chris Elvidge of
ORD/DRI.   There are  a  variety of  parties  involved in  the data
generation and in development of products.  Their are also a number
of end  users of the  data.   In particular, Ted  Vinson  of Oregon
State University would be interested in land cover information of
the former Soviet Union.  This would assist his data collection and
modeling efforts related to carbon budget.
      Phones:  Lee, 430-4634 or 503-754-4634 FAX -4799
GCRP Goals:

A number  of parties  are involved  in the  Pathfinder activities
related to boreal forests.  A goal is to evaluate the carbon budget
of these  areas, because they sequester considerable  carbon and
potentially  can  be  a  large  source  and/or  sink  depending  on
management and climate conditions.  The inventory these land covers
and use results in modeling  of  carbon pool and  flux requires a
remote sensor experiment-

Use of  remote  sensor data can  potentially supply  location and
extent of land covers.  Use  of multiple date images will allow the
evaluation of change in land covers such as deforestation.  These
analyses will also  supply information on natural  events that can
forest  morbidity or  mortality.    Event  include  fires,  certain
diseases,  windthrow from tornados or hurricane, volcanic eruption,
and other events.

The Trans-Boreal effort is  a continuing  one and  products  will
ultimately  proiTide detail  valuable  to  global   change  analysis
efforts.   Initial products  include  AVHRR bi-weekly composites of
selected areas,  and region area composites by year.

                             374

-------
NALC Applications (Standard Products):

The standard products that apply to this work element would be land
cover/land use images and tabular data,  and evaluations of change.
Unprocessed data and  standard  image  products may also be useful.
These products  would be used  to  study the  capability  of  remote
sensor  input,  and would utilize NALC  products from Alaska  and
Boreal Canada.


NALC Applications (Modified Products):

NALC  data products  and analyses  will  provide assistance  with
characterizing  and  quantifying  certain  variables.    These  may
include optimization of images  for detection of change  in extent of
land covers due to forest harvesting.  Analysis of change induced
by fire, pestilence, and weather is necessary, and the management
practices subsequent to the event would be of interest.


Scheduling for Deliveries of Standard Products:

Current  NALC  scheduling calls  for  the  above products  to  be
delivered beginning  in calendar year  1995 for pilot studies  of
interest.  Data for Alaska will be available in the 1995 timeframe,
also.


Potential Schedule for Modified NALC Products:

The above issues may be addressed in  preliminary tests of the NALC
program.  The levels of sensitivity can be evaluated, and suggested
modified products could  be generated the data production work by
Cooperators in support of GCRP.  The timeframe would  be the same
from the standard products.

Products  from  other  Landsat  Pathfinder experiments  or  AVHRR
Pathfinder experiment will become available, and will be obtained
the facilitate the evaluation  of conditions  in the former Soviet
Union.


Proposed Coordination Action/Schedule:

It would be desirable  to  coordinate  activities with the  other
Pathfinder and Trans-Boreal  initiative  to insure early  access to
products of interest.
                               375

-------
            NALC / GCRP  Coordination / Collaboration

                    Rice Growing Areas
NALC Task:

Delivery of standard and  modified NALC products focused on  rice
growing areas of the US, and identify points of contact to obtain
additional remote sensor-derived data on rice growing areas


GCRP Task:

Characterize  the quantity  of  rice  growing lands  and  determine
change, and  utilize the  data  in modeling of carbon  and methane
flux.
GCRP Contact:

The point of contact is Daniel Marks,  EPA-ERL-C.  Chris Elvidge of
ORD/DRI will coordinate this activity to provide linkages with non-
NALC remote sensing programs.

      Phones:  Marks, 430-4634 or 503-754-4634=FAX -4799
GCRP Goals:

The goal is to estimate the extent and the change over time of rice
growing  lands.   The information will  assist the  evaluation  of
carbon pool and flux, and the potential for generation of methane
from cropped anaerobic soils.   These and other data  can be helpful
in analysis of management activities and may potentially assist in
development and implementation of policies.
The remote sensor data can supply data on the location and extent
of rice lands,  and multiple date analyses will yield information on
change.   A goal of the program  is  to evaluate US  rice growing
areas, and  develop  methods to supply  information.   Later,  these
techniques can be used to evaluate other rice growing  lands.  It is
also  a goal of  the Landsat Pathfinder  program is to  develop  a
Multispectral  Scanner (MSS) data  set for  Southeast Asia.   The
format will  be similar to  that  of the  North  American Landscape
Characterization  (NALC)  program.   From  this effort,  it  will be
possible to estimate the rice lands of Southeast Asia and provide
tabular and imagery summarizes of the data.

                              376

-------
NALC Applications  (Standard Products):

The standard products that apply to this work element would be land
cover/land use images and tabular data,  and evaluations  of change.
Unprocessed data and standard image products may also be  useful.
NALC Applications  (Modified Products):

NALC  data  products  and  analyses will  provide  assistance with
characterizing  and  quantifying  certain  variables.    These  may
include optimization of images for detection of change in extent of
rice land and change in management.   Analysis  of change  induced by
fire,  pestilence,  and weather is necessary,   and the management
practices subsequent to the event would be of  interest.


Scheduling  for Deliveries of  Standard Products:

Current  NALC  scheduling calls  for  the  above  products  to  be
delivered in calendar  years  1996  for the Western US such as rice
growing areas of California.   In 1997 the  midwest  and Great  Plains
and Eastern-Southern US  will be  available, and  supply detail on
rice growing areas in Louisiana, Arkansas and  other states.
Potential Schedule for Modified NALC Products:

The above issues may be addressed in preliminary tests of the NALC
program.  The levels of sensitivity can be evaluated, and suggested
modified products could be  generated  the data production work by
Cooperators  in support of ,GCRP.   The  timeframe would be the same
from the standard products.


Proposed Coordination Action/Schedule:

Much  of this  effort  will  involve  production of  data  sets  and
standard and modified products by Cooperators.  Hence, it would be
desirable to collaborate on Cooperator contracting issues to insure
that the products of interest are addressed early in the delivery
timeframes,  and  that  modified  products of  interest  to  GCRP
researchers will be generated.
                              377

-------
                         Appendix VIII:
NALC Product Refinement Workshop
 Agenda
 Participants
 Workshop Summary
 Letter Reports
  Houghton
  Lawrence
  Salas
                           378

-------
    NORTH AMERICAN LANDSCAPE CHARACTERIZATION (NALC) - PATHFINDER

                    PRODUCT REFINEMENT WORKSHOP
                           April 27, 1993
                           Las Vegas, NV
8:00 AM
8:15 AM
9:00 AM
9:30 AM
10:00
-10:20 AM

10:20 AM
11:00 AM
11:40  AM
- 1:00 PM

1:00
- 1:30 PM
                WELCOME
   Eugene Meier - U.S. EPA, EMSL-LV

INTRODUCTION & NALC OVERVIEW PRESENTATION
 Ross Lunetta - NALC Technical Director

U.S. EPA GCRP Land Cover Data Requirements
Peter Beedlow - ERL/Corvallis Perspective

U.S. EPA GCRP Land Cover Data Requirements
   Lee Mulkey - ERL/Athens Perspective

                 BREAK
  Tropical Deforestation - Data Base Products
  William Lawrence - University of Maryland
  William Salas - University of New Hampshire

  Carbon Modeling - Model Input Requirements
  Richard Houghton - Woods Hole Research Center

                 LUNCH
  EMSL-LV Remote Sensing Laboratory Tour
                       DISCUSSION SESSIONS
1:40 PM
     NALC MSS LAND COVER PRODUCTS
2:50 PM
   NALC MSS LAND COVER CHANGE PRODUCTS
4:00 PM
      NALC DATA QUALITY OBJECTIVES
5:00 PM
      PARTICIPANTS CLOSING REMARKS
5:30 PM
                              ADJOURN

                               379

-------
  NORTH AMERICAN LANDSCAPE CHARACTERIZATION (NALC) - PATHFINDER
                   PRODUCT REFINEMENT WORKSHOP
                          April 27, 1993
                          Las Vegas, NV
Workshop Participants:

    Name
Peter Beedlow
Richard Houghton
William Keith
William Lawrence
Barbara Levinson
Ross Lunetta
John Lyon
Gene Meier
Lee Mulkey
William Salas
James Sturdevant
Hal Walker
Dorsey Worthy
  Affiliation
EPA/ERC-Covallis
Woods Hole Research Center
EPA/OMMSQA-HQ
University of Maryland
EPA/OEPER-HQ
EPA/EMSL-Las Vegas
OSU/EMSL-Las Vegas
EPA/EMSL-Las Vegas
EPA/ERL-Athens
University of New Hampshire
USGS/EROS Data Center
EPA/ERL-Narragansett
EPA/EMSL-Las Vegas
                               380

-------
                   GLOBAL CHANGE SCIENCE ISSUES
                 NALC PRODUCT REFINEMENT WORKSHOP

                 Las Vegas, NV,  April 27, 1993
                       WORKSHOP SUMMARY

     The  workshop was  organized  to  address  several  questions
advanced by the Technical Review Panel.  These questions focused on
the use of the NALC data products by the science community.

     One question was  how to optimize  the  match between science
issues and the products that NALC will supply in support of these
Global Change Research Program issues.  A second question or effort
was how to define products better with help from Agency and other
global  change scientists,  and  thereby  to  further  optimize  the
products for science applications.

     The major Agency  science  clients  and other  global  change
scientists  presented their  research program  related to  global
climate change, and identified how NALC  products will assist these
efforts.  The global change scientists have provided  their response
to the project and this information is provided here.


  Laboratory Science Issues

     The Environmental Research Laboratory  (ERL)  in Corvallis is
concerned with several programs related with  global climate change.
In  particular, the  assessment  of  forest  management  practices
envisions .the use of  NALC products.   The  goal is  to  identify
strategies for  the  sequestration  of carbon  in forest land cover
through improved forest management practices.  The initial area of
activity is  in southern  Mexico,  with  later  work  in Brazil  and
perhaps the Caribbean.   The plan is that NALC will provide a land
cover baseline and the  land cover  change  over  the  last  twenty
years.

     An  additional  area of  effort  is  the  forest  sector  carbon
budget.   This is  focused on the western  and  eastern US,  Mexico,
Brazil and the  former Soviet Union.   In the US and in the former
Soviet Union good forestry statistics are available  for estimation
purposes.  In Mexico and in Brazil these statistics are rare,  and
in Mexico  it is  hoped  that NALC  will  help provide  the  spatial
distribution  of  land  cover  types  for  forest  carbon  budget
estimates.
                               381

-------
     A third area of interest is the program of modeling the global
redistribution of vegetation due to climate  change.  These process
models  require  baseline  information  on  the  distribution  of
vegetation and NALC data  can potentially  be used to initiate the
modeling scenarios.  NALC data  may also be useful for verification
or validation of  the results of climate and vegetation distribution
scenarios.

     In general,  NALC  data can be  useful to predict  the spatial
distribution of vegetation or plant assemblages.   In these efforts,
NALC data along with known variances or uncertainties can be used
with carbon density estimates for each land cover type of interest.
The combination of spatial distribution of land cover from NALC and
carbon density of land cover types from field experiments, along
with  uncertainties   of  each  estimates,  will  help   allow  a
determination  of  spatial distribution of  carbon  and a  carbon
inventory.  Use  of NALC land cover change data along with carbon
density data will  potentially  allow estimates of carbon flux over
the last twenty years.


     The  Environmental Research  Laboratory in  Athens  suggested
several areas where NALC data may be useful. NALC land cover data
may  be  useful   in  their efforts to  model  the  release  and
sequestration of carbon and other atmospheric trace  gases in soils
(methane, nitrogen compounds).  Land  cover could be important in
determining the  spatial distribution  of managed land cover types
and their soils,  and help fix these quantities  for use as input to
models.

     A second  area of  effort would be in their determinations of
greenhouse gas budgets (GHG) and how  they are influenced by land
use changes and the land use management  activities.  In this case,
the change  of land cover  types  as determined  from NALC products
could be useful in better estimating the area that these processes
occur.


     The  Agency  and  other global change scientists discussed the
general value of Multispectral Scanner Data  (MSS).   It is clear
that MSS data are valuable as compared to  coarser resolution data,
as  based  on spatial  resolution,  spectral resolution,  and the
twenty-year digital data archive.  In  particular, the estimates of
deforestation  in Mexico and Central America are  a valuable use of.
MSS and NALC data products, due  to the  lack of organized forestry
statistics.

     A number of visiting global change researchers discussed their
work and addressed the potential contributions of NALC products to
science   issues.    Their  inputs  are  included throughout  this
document.   In  particular,  they agreed upon  the potential value of
NALC products.   NALC products  can be  used as a baseline and serve
a  valuable role  as  there are no  data sets  available for these
researchers  to use.  It will also supply  data with  known accuracy

                             382

-------
 characteristics,  and this is  valuable  for making better  quality
 estimates  of  carbon stocks.


      Comments on  specific  elements  of the  NALC technical plan  and
 the data products are presented  below.
  Land Cover Products

     Land  cover category  products will  be valuable  for carbon
 inventory and flux calculations.  The NALC products will supply the
 general  land cover types.  Carbon  density estimates  for each  land
 cover type can be developed  from field data.  Carbon density  data
 and  NALC products  on  land  cover  can  be  used to  integrate or
 inventory  carbon  stocks.    Land  cover  types  will provide  the
 mechanism to distribute the  carbon and calculate stocks.

     A further advantage is the definition of the current condition
 or current boundaries.   This definition of baseline is important
 for  monitoring  change,  for  setting initial model conditions and
 verifying model results.

     In  terms  of  land  cover,  it  is desirable to determine  the
 antecedent forest  types,  the timing of the event such as forest
 harvesting or  a forest  fire,  and the current  condition  such as
 whether  the site was reforested.

     From a regional  forest  carbon budget viewpoint, it would be
 desirable to use the NALC land cover woody class to compare forest
 source/sink conditions in the US.   For example,  it  may be desirable
 to examine  the carbon  sink  conditions in  the  eastern deciduous
 forest and in the southeastern forests, as  compared to the carbon
 source conditions in the Pacific Northwest.

     A number of scientists were supportive  of the  availability of
 "raw" or slightly processed  data  sets  as  they  come from the  EROS
 Data Center (EDC).  This may be  a very useful  product and may be
 favored  by certain application scientists.

     The land cover categorization scheme was  deemed adequate to
 address the needs of carbon cycling researchers.  It was emphasized
 that  the  land  cover  categorization  scheme was  optimized  for
 identification from the MSS  instrument.


  Land Cover Change Products

     Two techniques will be used for production of change detection
products. These include images generated from the brightness value
data   through   differencing  of   multiple  band  data   (pre-
categorization).   The second approach will  employ comparisons of
categorized and labeled scenes or a post-categorization effort.
                              383

-------
     The two techniques are to be proposed' due to some of the great
differences found in different image scenes.  The first technique
is the  easiest  to  implement  and  will  be  used  in  a number  of
locations.  The  second technique will  be useful where there is a
high level of cloud cover, or a great difference in seasonal dates
to be compared.

     One  valuable   result   of  MSS-based   NALC  data   is  the
identification of both carbon sinks (afforestation for example) as
well  as  carbon  sources (deforestation) .   NALC will be  able to
identify both and provide quantities and spatial distribution.


  Global Change Research Program Pilot

     It  is  necessary to develop further the  utility  of  NALC and
related products in Agency global change  research experiments.  To
develop the NALC products and their applications to science issues
there will be Process  Pilot  studies.  These studies will link the
NALC and related products to field  and  lab based measurements, and
allow use of results in modeling activities.

     The Process Pilot studies will examine measurements of soil
carbon and  soil moisture with field and sensor instruments, and
will  seek  to  identify  forest or  woody land  cover  type  using
combinations of sensors such as NALC MSS and satellite and airborne
radars.   The evaluation of  radar  data  would be performed under
EPA'S core  remote sensing research program as  it  is  beyond the
current  scope of the NALC project.


  Data Quality Objectives

      It was deemed important to conduct  an  accuracy assessment of
the land cover categories.  The uncertainties of estimates in land
cover, or for  that matter in process estimates, are important to
supplying error  bounds on calculations that use  NALC results.  The
accuracy assessments of individual  categories are vital in carbon
flux  calculations.  Large variations in uncertainties will greatly
weaken  estimates  and   broaden  those  estimates  to  the  point of
diminishing value.  Hence, it is important to assign accuracies or
uncertainties to address carbon science  issues.

      The data quality  objectives as presented at the Workshop were
deemed adequate  for  the proposed uses  of the  data.

      Various  science   users may  require  more  detail  than  is
available from  the  land cover  categorization  system and the  level
of accuracy supplied by NALC.  Hopefully, needs  that go beyond the
requirements  of  the  NALC   project  can be met  with additional
processing  of the pre-categorized  or categorized NALC  data  sets.
                               384

-------
                     A Report to the NALC Overview Committee
               following the April 27th, 1993 workshop in Las Vegas, NV

                                  R.A. Houghton
                                   April 29, 1993
       THE IMPORTANCE OF NALC PRODUCTS TO SCIENTIFIC ISSUES

      As part of the Landsat Pathfinder program, the North American Landscape
Characterization (NALC) Project helps prepare the EOS community for handling large
volumes of data, searches the current Landsat archive, accumulates additional current
scenes, and makes products available and useful to the broader user community.

      More specifically, the NALC Program will provide information useful for analysis
of the global carbon cycle. In particular, it has the potential to offer three products that
are essential for carbon cycle research:

      1.     Triplicate (1973, 1986, 1991) Landsat scenes for all of North and Central
America, including the Caribbean (coverage of Canada will be provided independently by
the Canadian Center for Remote Sensing). For the approximately 20-year record of
Landsat, these triplicate products will provide estimates of change in forest area. Change
in the area of forests is currently the single most important datum for determining the
flux of carbon from land use. (As satellite data are used to provide these data globally
over the next years, variation in  carbon density (biomass and soil carbon) will become the
factor contributing most uncertainty to estimates of terrestrial carbon flux. Rate of
change in forest area is most important at present, however.)

      2.  Images of change between the three dates over the 1972-1991 period. Change
in the area of forests is most important and is readily determined from Landsat MSS.
Carbon  density is not. However, carbon density may be determined indirectly from
Landsat data if change  detection is  carried out with pre-classified data. For example,
change detection may be able to identify logging and regrowth if young and old forests
are not  first classified as identical. And there may be other changes and variations
observable with Landsat MSS data  that could serve to determine carbon densities
indirectly (see Level of Classification, below). Every  effort should be made to include not
only changes in forest - nonforest, but changes in the age, structure, and condition of
ecosystems.

      3.     A base map (1991) for North and Central America against which to
compare future land covers. The value of such a digital product can be appreciated by
considering how much more we  would know now about sources and sinks of terrestrial
carbon if we had such  images of land cover, globally, over the last 100 years. (AVHRR
would provide such images for less  money, but not with a resolution that allows
                                        385

-------
monitoring of deforestation/reforestation) (see below).

      Potential uses of the NALC products (base map and 20-year record of changes)
include:

      A.  Determination of net flux of carbon from changes in land cover (Note: The
net flux of carbon in 1990 is dependent on  regrowth of forests logged before the launch
of Landsat. Such forests are currently accumulating carbon, contributing to a carbon sink.
For this reason, sinks are difficult to document with only 20 years of satellite data. The
20-year record  of Landsat will overestimate terrestrial emissions because releases of
carbon from  logging are greater per unit area than long-term sinks of carbon associated
with regrowth.  Over the period of Landsat, deforestation will have released some amount
of carbon. But  regrowth of forests on lands harvested or abandoned before 1972 will be
accumulating carbon and will not be 'seen'  by Landsat. This bias must be considered by
those who would use the NALC products to determine flux.

       B.  Determination of the areal extent of major land cover types. A knowledge of
land cover will, in turn, be useful for:

             a.     assigning carbon densities (biomass and soil carbon)  for
                    determination  of North American carbon stocks. (Note: assigning
                    carbon densities is required for A, above, for at least that part of the
                    region that shows change);

             b.     estimating the  potential for carbon sequestration. Again, carbon
                    stocks for each cover type will have to be assigned independently of
                    the NALC products. NALC will provide the detailed digital product
                    (the map);

             c.     assigning greenhouse gas fluxes of CH4, N2O, CO  to cover types;

             d.     assigning climatological parameters (e.g., albedo, roughness, rooting
                    depth, etc.) to cover types.

             e.     other?

       C,  The NALC 1991 base map will be useful in the future to determine areas that
 indicate change, either as a result of
             a.  land-use change, or
             b.  climate change.
 Monitoring programs such as NALC must  be continued in the future. Ten-year intervals
 may be adequate in temperate and boreal regions;  five-year intervals or  less may be
 required in the tropics.

                                       386

-------
Why North America?

      The contribution of North America to the global terrestrial flux of carbon is
thought to be small. Furthermore, the net flux cannot be determined with data provided
by NALC alone (independent data on historical land use and on carbon densities are
required). Nevertheless, the flux from North America is important because the global flux
is the sum of all regions. If we consider only the regions of large source strength, we will
have only part of the flux — most of it, perhaps, but only part. Accuracies in global
estimates are now ±0.5 PgC or better (combustion of fossil fuels releases 6 PgC/yr +_
10% (15%)?). Furthermore, we probably don't know areas of flux well enough to be
certain of identifying the "important" regions. AH regions'are important, or have the
potential to be so in the future. We certainly do not know the regions where a major
carbon sink exists. My suspicion is that we don't know well even the land-use flux from
temperate and boreal regions. What is the role of fire, for example? Has the frequency
of forest fires increased? Will it in the future?

      We need to know the flux of carbon from land-use change (including both sources
from deforestation and sinks from regrowth), globally. The need results from the fact that
there are two components to the terrestrial flux of carbon: a component related to
changes in land use (direct human effects), and a component related to other, more
subtle effects (climatic change, elevated CO2, increased availability of fixed nitrogen or
other nutrients or toxins). Changes in these environmental factors affect metabolism and,
hence, the storage of carbon per unit area. Such changes are very difficult to appraise
locally, let alone globally. So far, only the component resulting from direct human effects
can be determined; the other component has never been measured. It has been inferred
indirectly from analyses based on geophysical data — in other words, by difference. The
land-use component (the first one) is important to determine accurately because it helps
constrain the magnitude of the second component. One should  measure what can be
measured.

      Sources are important even in a region that is thought to be largely a sink.
Together the sources and sinks provide the net flux, and the net flux is required for the
global carbon budget. (Bank accounts are not balanced by considering only deposits.)

      We need to know the flux associated with global changes in land use  to ±0.5 PgC
or better (0.2 PgC?) in order to reduce the uncertainty to a level consistent  with other
terms in the global carbon equation (atmosphere, fossil fuels, oceans). Thus, despite the
fact that North and Central America may not be regions of high flux at present, their
contribution to atmospheric carbon  must nevertheless be determined. The determination
need  not require wall-to-wall coverage with triplicates of Landsat MSS. For
determination of land use change over the last 20 years, wall-to-wall inventories or wall-
to-wall application of change detection might be applied only to hot spots, with some
intensity of sampling used to cover areas not thought to be undergoing change. But while
sampling may be  adequate for retrospective estimates of land-use change, it will not

                                       387

-------
provide a baseline map. A map with usefulness for the future will require wall-to-wall
coverage, and I think establishment of such a map is highly justified (see Landsat MSS vs
AVHRR, below).
Level of Classification Required

       If the NALC products are to be used to calculate carbon emissions from land-use
change, a level 1 classification is adequate. Forest - nonforest is the change involving the
greatest change in carbon. Types of forest are generally less important.

       If the NALC products are to provide a baseline or inventory of cover types for
subsequent assessment of change or subsequent assigning of biomass, then the highest
level of classification possible should be used. However, even the highest levels of
classification (at great expense) would probably be inadequate. What is needed for
carbon cycling is a combination of sensors that will  give biomass (and soil carbon). The
techniques are not yet developed, but when or if they are, my guess is that existing land-
cover classifications will not be adequate  for assigning variations in carbon stocks. New
classification will be required, based on the new sensors of biomass.

       The state of the art classification of  85% accuracy 75% of the time suggests that
pre-classification change detection is better than post-classification, especially if level 1
classification is used. For example, level 1 recognizes woody vegetation, which would
include in the same class both old growth forest and young  forests. Post-classification
change detection would show no change after 5-10  years of regrowth following logging.
Pre-classification change detection, on the other hand, would show a change — and the
change is important. It allows aging of the forest and indirect estimation of structure or
biomass.

       Pre-classification change detection should be used as an exploratory tool in all
 regions, and should be explored in the pilot studies of NALC. Level 1 is probably
 appropriate for classification but pre-classification change detection should be used to
 determine the level of classification appropriate for cover types that have changed.
 Landsat MSS vs. AVHRR
       AVHRR is cheapen
       AVHRR has high temporal resolution (twice daily).
       AVHRR is good for broad categories.
       Landsat has high spatial resolution.
       Landsat is good for land-use change. AVHRR overestimates deforestation. (Could
 MSS be used to determine a correction for AVHRR? Early studies suggest not; in areas
 where forest cover is <70%, AVHRR may either over- or under-estimate it.) Landsat
 MSS or TM is  near the lower limit of resolution for determining short-term changes in

                                        388

-------
forest area.

Implications

      In a strict sense, there are at least two scientific issues here, and they might be
most economically addressed with different approaches. First, for determination of the
net flux of carbon between North American ecosystems and the atmosphere, one needs
high resolution Landsat data, but not necessarily wall-to-wall. Stratified sampling would
probably be adequate and would be more economical. On the other hand, for
establishment of a baseline map, LAC AVHRR might be adequate and might be more
economical than Landsat (significantly more?). The NALC program is in a position to
address both issues (20-year  change and base map). It is not clear from these two issues,
alone, whether NALC should attempt to address both. There may be other uses of the
products that help determine the appropriate goals for NALC. If the net flux of carbon is
important not only over the  last 20 years, but over the next decades, as well, then a good
argument exists for obtaining, now, wall-to-wall coverage with Landsat MSS.
                                      389

-------
Letter Report on NALC Pathfinder
Product Refinement Workshop
US EPA EMSL-LV April 27, 19S3

.submitted by:

William T, Lawrence
University of Maryland
Dept. of Geography
[301] 405-S809/wl33@umail. umd.edu
My immediate impression of the NALC Pathfinder Project is that it
is perceived within EPA as a very carefully designed data product
•with an as yet poorly defined science product and end-user
community.  The fault in the product's definition seems more
conceptual than actual.  I find the project very much on- track.

It is clear to me  that the proposed products of NALC will prove
to be extremely va-luable to the science community,   The mapped,
cloud-screened MSS image triplicates are in themselves an
invaluable product.  The addition of a straightforward change
classification and lineage attributes make the basic dataset even
more useful.  This community-wide value will become apparent once
the product stream is robust enough to allow general announcement
and release of the products to the ex-EPA user community.  At
present, it seems  that there is still a strong outreach effort to
be undertaken within EPA to generate some excitement about the
NALC product and its wide application.  The dual roles of this
project; fulfilling both NASA Pathfinder and EPA Lab-wide needs
makes it much more difficult to build a strong consensus.
However, the initial interactions I observed between NALC, Athens
and Corvallis labs seems to be the very promising beginning of
collaboration that will be of benefit to all involved.

Products/End Uses:

There is a high level of interest in generating products beyond
the triplicates and associated land cover/land cover change maps
within NALC .Pathfinder.  Such an effort on the part of NALC would
be an over- commitment of both time and resources, and is best
left for later collaborative, user community or EMSL-LV led
effotts that will  surely follow given the tremendous potential
applications of the basic data product .  I think it became clear
during the Product Refinement Meeting that, given appropriate
examples and some  in-depth familiarization, projects at both
Athens and Corvallis labs could make very profitable use of the
basic suite of NALC data.

Some of the immediate uses of this basic dataset were made very
apparent during the presentations and discussions .  I have
outlined a few below.

     Soil carbon contents - The brief discussion of work at the

                           Lawrence - 1
                              390

-------
Athens lab brought many potential uses of NALC data to mind.   The
large scale soil carbon modeling of this group could benefit
greatly from the utilization of a spatial dataset such as NALC.
This is not a trivial application, but a project that would
require some time to carry out.  However, the results would not
only test the linkage of statistical and spatial data, but also
help define the accuracy of the current methodologies with their
broad spatial generalization.

     Global Carbon Stocks - Both Athens and Corvallis labs are
interested in global carbon stocks.  The standard approach uses
non-spatially explicit values for models.  The use of spatial
data, such as the NALC product, for distribution of carbon
densities across landscapes present very real difficulties, so
have not been widely used.  This is .an area of research that
could profitably engage both NALC and other EPA labs in
development of state-of-the-art science products.  Other closely
related modeling efforts that could use the NALC products are the
whole suite of Earth-system modeling projects, assessment of
management practices on carbon cycling, greenhouse gas emission,
and global carbon balance studies.

     State-of-the-Art Maps - The NALC land cover products will be
the first ever available for accurate assessment of the spatial
extent and -change in important land use type.s.  No such baseline
maps exist at the scale of the NALC.  Simple spatial analyses of
the land cover maps can yield statistics on the state of
cultivated lands, deforestation, and even wetland distribution.
Such maps are critical to accurate parameterization of ecosystem
models and assessment of management planning.  Applications in
social and political science, and other as yet unidentified
fields are sure to follow the wide distribution of this dataset.

Methods:

     Digital Triplicates - The digital triplicates seem extremely
well thought out in terms of the content, mapping, and cloud-
reduction mosaic processing.  I am a bit concerned that the types
of distribution, files are proliferating as changes are made in
the data stream, ie. the method of copying mosaics.  If a data
layer is made for each of several cloud contaminated images,
rather than the earlier mosaic process, things begin to get very
complicated.  A simplification might be the distribution of two
cloud cover exclusive scenes for any one year without masking or
mosaicing.  These datasets may get so voluminous that they are
impossible to distribute once classification, lineage, land-cover
and_land cover change layers are added.  Distribution of a very
basic land cover and land cover change dataset would be useful.
Many users need only the product, not the entire digital dataset.
Additional information might include 8-bit color composites for
each year simulating visible or near-lR photoproducts.

     ^supervised Classification.  - Care should be taken not to
distribute too many classes.  Tens of classes from an initial
classification should be sufficient to describe the basic

                           Lawrence -  2

                             391

-------
variability for the collaborators that are doing the labeling and
aggregation,  I understand that this process is still in
development ,  Some image smoothing may also be useful .   We find
that our digital classification produces a data product that is
entirely too complicated, so majority filter smoothing and
removal of groups of pixels below a minimum mapping unit is used
to simplify the basic data prior to editing and analysis.

The type of change detection to be used need very careful
attention.  Whether to use pre- or post-classification detection
will probably depend on the types of change and land cover-
present, as well as your ability to normalize MSS scenes through
time.  If normalization, is successful, then I would suggest the
distribution of the normalized images/ rather than the raw
products -
                     of Land Cover Types - There was some
discussion of  the level at which to distribute the land cover
information.   I would let your collaborators classify to the
finest detail  that they can, using digital and ancillary
datasets, with later post -process ing to aggregate the data to the
lowest common  denominator, which would be based on your
Classification criteria*  In any case < level 2 [Ecologically-
oriented Land  Cover Classification System] would be unacceptable
for most ecological or Earth-science modeling uses, especially
when concerned with carbon modeling.  The finer classifications,
when successful,  could be distributed as case studies, or even as
separate data  layers.  A useful paradigm might be distribution of
the MSS digital triplicates separately from the land cover and
land cover change information.

Results of

The Product Refinement meeting was valuable for me and hopefully
our outside point of view will prove useful to the EPA NALC
Project as well.   During the meeting I learned much of the NALC
and its placement within EPA and among other EPA laboratories .
The basic triplicate product of NALC is of exceptional value, as
it makes a uniform set of image products available to EPA and
other researchers,  "Hie near-term science product, the land cover
characterization  and change maps, will prove to be even more
important to the  user community, but clearly are still in a
developmental  stage.  Much intensive research remains before a
truly operational land cover and/or change product is
forthcoming, but  this should be accomplished in the near-term.
The research planned for the Chiapas and Oregon Transect pilot
studies will probably meet these developmental needs and answer
many methodological questions that remain.  A very fruitful
collaboration  with Athens and Corvallis labs in terms of NALC
dataset development and utilization seems highly likely.  These
collaborations will not only enhance the NALC process but also
serve to further  the linkage of spatial and statistical datasets
in modeling.
                           Lawrence - 3

                            392

-------
Recommendations -

* use pilot studies to test utility of unsupervised
     classification and appropriate level of detail for land
     cover class identification
* a pilot study in agriculturally-dominated landscapes may be
     useful to test land cover classification methodologies [if
     such land cover types are not well represented in current
     pilot study sites]
*• go into production mode with basic products - triplicates, land
     cover change - let later projects develop the more
     discipline-specific science products
> build constituency with other EPA labs, using them to help
     define future products/collaborative research efforts - this
     will be somewhat of an MALC 'outreach' until others catch on
     to the utility of this multi-temporal dataset
+ formalize inter-lab NALC data based research in some way - eg,
     travel funds, some limited support
*• consider wide distribution of a simple dataset, eg. land cover
     classes for three periods,.plus change and perhaps an 8-bit
     color composite  [visible or near-IR - no MSS digital) as
     this will reduce the data volume and suit many applications
»• MSS digital triplicates might be distributed separately from
     all land cover/land cover change data, if datasets become
     unmanageably large
»- 'radiometric rectification' or some normalization of the
     digital data may be useful prior to distribution - research
     work on pre-classification change detection will clearly
     show if this is a viable option
*• maintain close ties with other Landsat Pathfinder projects,
     especially in change detection methodologies - NALC land
     cover classification could be applied to other Pathfinder
     datasets
                      ^C-MJ
William T.  Lawrence
May 10, 1993
                           Lawrence - 4

                            393

-------
         North American Landscape Characterization  (NALC)
                   Product Refinement Workshop
                      EMSL-LV April 27,  1993

Review comments by William Salas.

     A product refinement workshop for the North American
Landscape Characterization project (NALC)  was held at the
Environmental Monitoring Systems Laboratory in Las Vegas on Apri 1
27, 1993. This workshop was extremely informative and useful.

     As part of the Landsat Pathfinder program, NALC is making
important contributions in data processing, distribution, and
archiving techniques needed for development of EOS-DIS. In
particular, the use of EDC as the Land DAAC to create triplicate
MSS images that are image to image registered demonstrates how a
"science computing facility" could use a DAAC in the EOS era.
While this point may seem trivial,  it is an important test of the
DAAC-SCF infrastructure design of EOS-DIS. NALC's assistance in
completing the MSS 1992 acquisition for complete coverage of
North America is another important contribution in itself. The
now complete 20 year archive of MSS data is an extremely
important data set, in that it provides the capability to study
temporal phenomena over a 20 year period with the same
instrument.

     The processing algorithms being developed by NALC for land
cover and land cover change mapping are important but difficult
tasks. Characterization of thematic classes for large areas over
time is extremely difficult using remote sensing data, and as a
result has not been attempted until the recent implementation of
the Landsat Pathfinder projects. Success of these programs will
demonstrate the ability to handle massive, multitemporal, high
spatial resolution data sets. However, in order to succeed
efficient processing algorithms are needed. NALC use of pilot
studies to assess the utility of several candidate processing
algorithms is  imperative since the success of the algorithms may
depend on the  regional under study. NALC pre-classification
change detection strategy is exciting and should provide a better
product  for the science community.

     Overall  the NALC project has a solid start toward obtaining
their goals of producing "land cover and land cover change data
products at sub-kilometer spatial resolution" for most of North
America and the project will be a success. However, as with any
project, there are a  few areas of concern that if addressed could
possibly enhance the  success of the project.

     The project needs  to address how the NALC products may be
used throughout the science  community, not just within the EPA. I
realize  one of the goals of  this meeting was to address  this
issue, but, for example, there are many EOS investigators that
would be extremely interested in hearing about the NALC project
and could  integrate the NALC products with ongoing EOS research.

                              394

-------
For instance, there is a group at the University of New Hampshire
that would be extremely interested in providing inputs to NALC
regarding the land cover data sets for the northeast US and how
the data could be used to supplement their efforts to model net
primary production and evapotranspiration in the region. As
another example, NALC data products could be used as test sites
for the North American AVHRR 1-km project. This is an important
exercise in how can high spatial resolution be scaled up for use
in testing the lower. As a result I suggest that another pilot
project be developed to directly link the NALC data set with an
existing global change research project. This would enhance the
visibility and broaden the base of support for NALC, as well as
demonstrate how the data products can be used to address .specific
global change research needs.

     The use of a clustering algorithm for an entire scene may
present some-problems. From experience we have scene that
variations in atmospheric conditions within a scene can cause
different land cover types to have the same spectral
characteristics. This occurred often occurred in scenes along or
near coastal areas, due to differences in the atmosphere over
terrestrial and marine areas. As a result, the scene needed to be
broken up into smaller areas to accurately classify land cover
types.

     The data analysis tracking techniques are an important part
of a large processing oriented project like NALC and the current
system of creating a bitmap relating the origin of each pixel is
important and properly addressed. However, some thought as to how
to account and track sidelap (7% at equator to 60+% in Alaska) in
the processing stream is needed.

     Verification procedures and accuracy assessments are
difficult to provide for large projects like NALC. The procedures
for spatial sampling outlined during the workshop and in the
technical work plan are good. A measure of variability in
labeling due to differences in individual interpretation should
be addressed.

     The concerns provided here are minor and an emphasis on all
the good aspects of the project should be inferred. The project
has a strong understanding of its goals and has presented~a~well
thought out and realizable approach to insure that the goals are
met. It was extremely instructive for me to sit in on this
workshop and I hope my brief comments can be of use to the
project.
                             395

-------
                        Appendix   IX:
   Landsat Pathfinder Initiative and the Canada Centre for Remote
Sensing (CCRS), and Documents from the December, 1992 Meeting with
CCRS.
                              396

-------
LU
V)


a
a:
O
Q
Z
<
LU
LU


<
CO
cc
O
O
                                   O)
                                   O
                                   05
                                  •o
                                   c
                                   (O
                                  a.
                                  •»-•
                                   (0
                                   (O
                                  •a
                                   c
                                   (0
                                                  CN
                                                  O
                                                  O)
                                                                  CO
 o
JD
 E
 o
 o
 a)
Q
                                               397
                                                                                                        O)

                                                                                                         d
                                                                                                         cu
                                                                                                        Q
                                                                                                         o»
                                                                                                        .c
                                                                                                         0)
                                                                                                         

                                                                                                       z a:
                                                                                                       52
                                                                                                       cc
                                                                                                       o
                                                                                                       o

-------
                        EPA/EDC/CCRS Meeting
                        December  9  &  10,  1992

                               Agenda
December 9. 1992 in Room 301 - 09:00 - 12:00
  Introduct ion
  Welcome to CCRS
  Review Agenda
  Pathfinder Overview
  North American Land Characterization (NALC)
  NALC Data Processing Overview
  NALC Standard Data Processing Methods
  Discussion
December 9, 1992 in Room 201 - 13:00 - 16:00
Terry Fisher
Florian Guertin
Terry Fisher
Ross Lunetta
Ross Lunetta
James Sturdevant
John Lyon
• Detailed discussion on pathfinder processing methods and
algorithms
          • catalog guerying
          • radiometric processing
          • geometric prcessing
          • product definition & distribution

December 10, 1992 in Room 201 - 09:00 - 15:00

     This day will be devoted to discussion of topics of mutual
     interest including: CCRS Global Change initiatives, change
     detection, etc. and possibly a tour of CCRS facilities at Booth
     Street.

     Suggested areas:
          • CREO Optical tape
          • GEOSCOPE
          • ERS-1 Mosaics / system
          • GCNet/Video Disk
                              398

-------
                LANDSAT PATHFINDER INITIATIVES
                            AT THE
            CANADA CENTRE FOR REMOTE SENSING
      In February 1993 a project was initiated at the Canada Centre for
Remote Sensing (CCRS) to  develop and demonstrate capabilities to
manipulate and post-process LANDSAT Pathfinder data.  Of particular
interest is the ability to merge scene-based imagery into seamless mosaics
for large area monitoring purposes.  The project, which will be completed
at the end of March, 1995, will include the development of a PC-based,
prototype workstation.  As a means to demonstrate this technology and
to increase Canadian scientific community awareness of the potential of
Pathfinder imagery, an example mosaic of the Great Lakes region will be
generated for the 1988 time window.  The mosaic will include both U.S.
Pathfinder products and Canadian coverage pre-processed  on Canadian
systems. The prototype workstation will expand upon a current capability
for spaceborne SAR image mosaicking and will include (a) radiometric and
geometric normalization, (b) automated  seam delineation, (c) selectable
spatial resolution and (d) a simple DBMS to aid in planning processing
strategies and in monitoring and assessing processing status and product
quality.

      The present U.S. program does not include plans for the routine
inclusion of  Canadian  coverage.   As an added objective,  CCRS will
conduct a technical assessment of Canadian operational products with a
view to their potential as a source of Pathfinder data.
                         399

-------
                                                                                   CM
                                                                                   ffJ
                                                                               I
                                                  08
                                                  en

                                                  u

                                                  Q
                                                  t

                                                  Oi
                                                  JC
                                                  "-P
                                                  o
                                                  4)
                                                                                    O
                                                                                    T3
                                                                                    C
                  H-
                  Z


                  CC
                                                                                    a
                                                                                    a.
                                                                                   •a

                                                                                   m
CO

§
Q
O
CC
a.
UJ

LU
O
CO
a

a
2
<
LU
z
LU
O
CO

                            CO
                            CO
                            CO
                            Q
G
LU
O

LU Q
CC UJ

Ufa
UJ O
CC O
o o
LU LU
O O
CO
o
ID
LL

Q
LU
O


y Q
LU O
CC O

O O
LU LU
O CD
                                                                                  UJ 111
                                                                                  u a



                                                                                  O

                                                                                  O
                              ^ w
                              " O
                              Q ut
                              ? «
                              3 E
                              < a
                              o "


                              CO

                              CC

                              O

                              O
                                      400

-------
co
LU

U



DC
UJ

CO



flC
ui
DC

O


O
LU

CC
o

(D

OtO

-1 Q


3» UJ
02

05
UJ o
CO ll
< —i
OQ O

co"-
LLJ LU
Z X
UJ H-

£"-
^o



1°


1<
Si
UJ
_J

LU

CO
  CD

  ^

  O

  O


SATELLITE ID





SENSOR ID



F APPLICABLE)
LU
SENSOR MOD


<
APPLICABLE)
~^
u.
VIEW ANGLE



UJ
H
Q
ACQUISITION



UJ
UJ
O
CO
X
o
LU
LL.
O
LU
H
ACQUISITION




H
IMAGE QUALI




DC
CLOUD COVE



ORDINATES
O
O
GEOGRAPHIC



EARCHES
CO
SS INVENTOR
O
CC
o
ALL SATELLITE/SENSOR
DC
O
SEARCH ANY



i/l A SINGLE USER REQUEST
^
O
CC
)ATABASES F
L_J


                             401
                                                                CM
                                                                0>

                                                                O
                                                                <^

                                                                •a

                                                                O)


                                                                u
                                                                0

                                                                Q
                                                                 o»



                                                                 I
                                                                 0
                                                                1
                                                      u>
                                                      •a

                                                      5
                                                                51
                                                                UJ -4U
                                                     O
                                                     o
                                                                 55
                                                                C7)

                                                                CC

                                                                O

                                                                O

-------
                                                                                                CD

                                                                                                O
a
2
X
O
DC
O
2
CO
CO
LU
O
O
cc
Q.
^
o
o
v£
o

d
o

o

  1h
r-
CO

cc
LU
CO

^
LU
O
LU
O
LL.
O

CO


LU
o
cc
LU
LU


<
CO

CO
DC
O
O
tD to
s ®




Total
Scenes
Received
t! -fl>
£ «
co Q

c
"55




IN, s
co ;
JN

rM
i
JN
ico
CO
IN,
t—
"5

CO
CO
•+-» :
CO
W
-o
.C
CO
_J
IN
^

00
T—

o
LO
CD"
o
CM
CO
05
,'t~J
O)

H
^_,
fO
(0
TD
C
10

O :
CT>

CM

O
CM
LO"
CO
CO
CO
v"
a.
~
o
Q.
'3
i^
"*~*
1-
O
CL
CO
LO
O

CO

CO
LO
CO
CO
00
0)
r-
Q.
"•+3
CO
anchrom
CL.
^^
J-
0
Q.
CO
CO
LO
0)

^J*

•MH
CO
CO
00
CO
O)
T~
CO

cc
CO
CO
LU

T—
CO
o
^
T—
*^

CO

r—
CO
en




TOTAL




                                                                                                CD


                                                                                                'O
                                                                                                O)
                                                                                                Q
                                                                                                TD
                                                                                                c
                                                                                   ro
                                                                                   CL
                                                                                                T3

                                                                                                C
                                                                                                03
                                                                                  o
                                                                                  o
                                                                                               < CC
                                                                                               •Z K
                                                                                               < O
                                                                                               o Z.

                                                                                               CO
                                                                                               a:
                                                                                               o
                                                                                               o
                                           402

-------
CO
LU
o
DC
LU
OC
o
LU

z
or
o

o
UJ
DC
CO
UJ
O

DC
UJ
CO
>•
cc
o
H
o
LU
DC
G

LU
Z
o
o
               u. Q
               o z
               LU O
               DC P

               > O

               <3
               03 I

               l-<
               CO QC
               LU Jn
               £P
                                 -JO <
                                 SJ *- *-
                                 OQ .  - CO

               g<
               H DC
               O LU ;?
               ii i |— UJ
               es E I
               Q O h-
                         uj co
                         co >•
                         D tO
                         LU
                UJ <
                iZ O


                Z H
in "J
CO CC

ii
I UJ
                         CO
                         LU
   DC

   G
   o
   LU
   z
o z
i- O
uj O
CO >-
                         < o
                         G iZ
                            o   F;
                                 
-------
                                                                                          en
                                                                                          o
                              LU
                                                             0)

                                                             o
                                                             c
                                                             Q
V*
HI
 CC
 LU


 CC

I

 LU

 2'
'" MWK '
 Q
 2
 <
o
LU
CC
a
                             O
                             OQ
O
                              CC
                              O
O
O

CO
I- Q
CC LU
O H-
CL O
LU LU
CC -J
— LU
cc co
          CO
          LU

          LU
          O
          CO

          o
          111

          o
          LU
          —I
          UJ
          CO
          LL.
          O
          o
          H

          o
          O LU

          >2
CO
LU

LU
O
CO

Q
LU

O
LU
_l
LU
CO

CC
o
LL.


CC
LU
LU
CO


O
CC
GQ
                                 O


                                 O LU
                                 UJ CO
                                 h- < CO
                                 o m LLJ
                                 10 — 2
                                 CC < LLJ
                                                             o
                                                             0


                                                             k.
                                                             o
                                                             •a
                                                                                             S
                                                  CL
                                                  CO
                           CD

                           Q
                 Q LL
                                                           " o

                                                           II
                                                            CO
                                                            LT
                                                            O
                                                            O
                                        404

-------
ULf
Q

I
Ul
cc
O
Q  I

1


CC
O

O
LU
CC
Q
                                                                                            I
 2?
 o
 r-
 00
 «

  U
  0
 a

  o»

 •*:

  o

 •o

 3
 2z
_ o

< o
O
O
                                                                                               Z ui
                                                                                               iw Jr
                                                                                               " o

                                                                                               il
                                                                                                CO
                                                                                                CC
                                                                                                o
                                                                                                o
                                         405

-------
               APPENDIX X
LAND COVER CATEGORIZATION SYSTEM FOR
  USE IN THE NALC-PATHFINDER PROJECT
                 406

-------
INTRODUCTION

      The United States Environmental Protection Agency's (USEPA's) North
American Landscape Characterization (NALC) Pathfinder program requires a
categorization system that can be used to categorize Landsat MSS data across the
entire North American continent. The two main categorization systems in use by
Federal Agencies are Anderson et al. (1976): A Land Use and Land Cover
Classification System for Use with Remote Sensor Data, and Cowardin et al.
(1979):  Classification of Wetlands and Deep Water Habitats of the United States.

      The NALC categorization system combines features from the Anderson et
al., Cowardin et al., and other systems. It provides the capability to "crosswalk"
between the NALC system and other systems.  Crosswalk is a categorization term
referring to the  ability to integrate data which has been categorized with qne
system with data that has been categorized  using a different system. The NALC
categorization scheme is a hierarchial system with broad categories at the more
general levels and detailed land cover and land use categories at the finer levels.
The resulting system was designed to remain relatively simple at the basic level,
while still defining features most important for assessing carbon stocks  and
monitoring  changes.
      The NALC categorization system has been developed with major input from
the USEPA Environmental Monitoring Systems Laboratory-Las Vegas (EMSL-LV);
the NALC Technical Review Panel; and the NALC Technical Working Group.
Considerable consideration was given to the interagency working group
categorization activities. The Interagency Categorization System was developed
through the input from many sources including the U. S. Geological Survey
(USGS); the U. S. Fish and Wildlife Service - National Wetlands Inventory
(USFWS-NWI); the National Oceanic and Atmospheric Administration (NOAA); the
NOAA - National Marine Fisheries Service (NMFS); the University of Delaware; the
Oak Ridge National Laboratory; the Salisbury State University;  and the Florida
Department of Natural Resources.
NALC CATEGORIZATION SYSTEM

      The standard NALC categorization product for North America will be a Level
II database product.  However, the NALC categorization system (Table 1) currently
contains three levels of information. The third level will have regional and site
specific applications for global change process pilot studies and unique regional
scale applications. A level is a stratum or tier of appropriate effort and technology
required to derive a given level of land cover information.  Each of these levels
corresponds to different degrees of information processing, progressing from

                                   407

-------
NALC Pathfinder Categorization System
LEVEL 0
Land



























Wotor




Other



LEVEL 1
1.0 Barren or Developed
Land

2.0 Woody






3.0 Herbaceous



4.0 Arid










5.0 Snow/Ice


6.0 Water & Submerged
Land



7.0 Other



LEVEL II
1.1 Exposed Land
1 .2 Developed Land

2.1 Forest


2.2 Scrub/Shrub



3.1 Herbaceous



4.1 Arid Vegetation





4.2 Riparian




5.1 Snow/Ice


6.1 Ocean
6.2 Coastal
6.3 Near-Shore
6.4 Inland

7.1 Cloud
7.2 Shadow
7.3 Missing
7.4 Indeterminable
LEVEL III
'"$**& * m^r^f4yff
., ,,, ,,,, , ,,,.;,,,, , ,r/f

2.1 '.1 0p&ftF2 Mode««& Forest
3,1,$ £>e«$0 F>W«9t
£,&f <}pftff.&<;ft(h/$jw|
2,£3fj$»3l+g! tj^afts Sors^^tub,

3t .1.1 Fastara/SrdsBliand
3.1.2R«iyvf>Ci'Qj*
3*1*3 Arctic TUrtdfa

4.1,1 Alpine Tundra
4<1,2 Arid Fores*
4*14 $$ V^dcWIland w
4.1 ,4 Arid StertifcfaCtd
*(1>S Arid &^$(fto«t
4<1 <$ Arid (?«se«bfld
4.^1 Ariif Sty/amp't'Qfiesit^,
4-5 4 A^rf SwannpBOfHif '
4,2,3, Arid Marahtsn^
4,2:4 tfncf 'Strandland s

S-1 1 $rt»W% v 1 v
§4i ^/s^<^^ ; _"• ,

..'i>-, -•,-, --.-. <<^
f \$Y,*,f •*
- „„', „ &?'**•-
'~ '~f jfe. * V-WM-V o %
^,,«

' - ,»'>"•<« ' £ ;r
S ff fft ff>ts,jf"S '": •':" <"• \-~ 'f .<•
, ,,v~.
, ' ' ','>
               408

-------
automated spectral clustering of satellite and aircraft image
data, to manual collection of field measurements.

     This categorization system is intended to provide a
framework for categorizing a broad variety of land cover types.
If more specific information is needed about a particular area,
the user can modify the system by adding categories at deeper
levels.  Any of the Level III categories can be subdivided
creating Level IV categories.


DEFINITIONS

     This section contains the definitions for the categorization
system.  Areas are categorized and labelled according to the
resources present on the ground at the time of imagery
acquisition.  If a agricultural field is devoid of vegetation and
contains bare soil at the time of image acquisition, the field
would be labelled as barren or developed rather than herbaceous.

The categorization system is divided into three organizational
categories: Land Division, Water Division, and Other.  The actual
categories in the NALC database would be the Level I and II
categories.

LAND DIVISION

The Land Division includes all categories other than open water
and other.  It is divided into five Level I categories: 1.0
Barren or Developed Land; 2.0 Woody; 3.0 Herbaceous; 4.0 Arid;
and 5.0 Snow and Ice.

1.0 Barren or Developed Land

     This category is composed of cleared, burned, or otherwise
     barren areas.   This includes areas of artificial
     anthropogenic cover, with much of the land covered by
     artificial materials such as buildings and other man-made
     impervious surfaces.  This category is subdivided into two
     Level II categories: 1.1 Exposed Land; and 1.2 Developed
     Land.

     1.1 Exposed Land

          The Exposed Land category includes naturally occurring
          areas that have limited ability to support life and
          other areas disturbed by man.   In general, the surface
          of these areas are covered with soil,  sand, or rocks
          and contains less than one-third vegetation or water.
          Naturally barren areas includes rock outcroppings,
          talus slopes,  sandy beaches,  etc.  The exposed areas
          may be transitional or permanent.  These areas include
                               409

-------
          lands cleared for a variety of purposes,  i.e., building
          construction, quarries, landfills, gravel pits, or
          strip mines.  Recently clearcut forests with all the
          vegetation removed  and bare soil agricultural fields
          would be included in this category.

     1.2 Developed Land

          This category includes areas in which 70% or more of
          the land surface is covered by structures and other
          man-made impervious features.  This corresponds with
          those areas termed "Urban or Built-up Land" by Anderson
          et al. (1976) who states:

               Included in this category are cities; towns;
               villages; strip developments along highways;
               transportation, power, and communications
               facilities; and areas such as those occupied by
               mills, shopping centers, industrial and commercial
               complexes, and institutions that may, in some
               instances, be isolated from urban areas.
2.0 Woody
     The category Woody refers to areas dominated by plant
     species that have an aerial stem which persists for more
     than one season, and a cambium layer for periodic growth in
     diameter (Harlow and Harrar, 1969).   Succulents such as
     cacti would be included in the Arid category.  The woody
     vegetation category requires a crown closure of 25% or
     greater.  This category includes scrub/shrubs and trees and
     is subdivided into two Level II categories: 2.1 Forest; and
     2.2 Scrub/Shrub.

     2.1 Forest

          The forest category contains single stemmed woody
          vegetation (trees).  Harlow and Harrar (1969) defines a
          tree as a "woody plant which at maturity is 6 meters or
          more in height, with a single trunk, unbranched for at
          least several feet above the ground, and having a more
          or less definite crown".  This category is subdivided
          into three Level III categories: 2.1.1 Open Forest;
          2.1.2 Moderate Forest; and 2.1.3 Dense Forest

          2.1.1 Open Forest

               Corresponds to communities dominated by trees with
               a mean potential height usually under 15 meters a
               canopy which is usually open, and is singularly
               layered.  The crown closure is between 25 and 50

                               410

-------
               percent.  These areas typically have a grassland
               understory.   Savannas are an ecotone between woody
               and herbaceous.

          2.1.2 Moderate Forest

               Corresponds to forests having a canopy closure of
               greater than  or equal to 50%, but less than 75%.

          2.1.3 Dense Forest

               Corresponds to forests having a canopy closure of
               greater than  or equal to 75%. -


     2.2 Scrub/Shrub

          The Scrub/Shrub category contains woody vegetation
          which at maturity  are usually less than 6 meters in
          height.  The crown closure is greater than 25%.
          Scrub/Shrub are smaller than trees and unusually
          exhibit several erect, spreading, or prostrate stems
          and a general bush appearance  This category is
          subdivided into three Level III categories: 2.2.1 Open
          Scrub/Shrub; 2.2.2 Moderate Scrub/Shrub; and 2.2.3
          Dense Scrub/Shrub.

          2.2.1 Open Scrub/Shrub

               Corresponds to all Scrub/Shrub,  having a canopy
               closure of greater or equal to 25%, but less than
               50%.

          2.2.2 Moderate Scrub/Shrub

               Corresponds to all Scrub/Shrub,  having a canopy
               closure of greater than 50%, but less than 75%.

          2.2.3 Dense Scrub/Shrub

               Corresponds to all Scrub/Shrub,  having a canopy
               closure of greater than or equal to 75%.
3.0 Herbaceous
     This category includes areas dominated by herbaceous cover.
     This category encompasses natural and managed areas of
     herbaceous cover, including lawns, natural grasslands,
     cultivated (agricultural)  fields, pastures, or herbaceous
     wetlands (marshes).   Land used for food production such as
     corn,  soybeans,  wheat, etc.,  are included in this category.
                              411

-------
     Some grasslands may have been seeded to introduce or
     domesticate plant species.   Other grasslands are
     successional in an area such as areas that have been burned
     by fire or agricultural fields are left fallow.  This
     category is not divided at the Level II category.
     Herbaceous is subdivided into three Level III categories.

     3.1 Herbaceous

          This category is subdivided into:

          3.1.1 Pasture/Grassland

               Pasture/Grassland encompasses those areas that
               have both managed and natural grassland cover.
               This category includes small grains such as oats
               and wheat, hayfields, pastures, lawns, herbaceous
               road right-of-ways, herbaceous fields, or natural
               prairie areas.

          3.1.2 Row Crop

               Row crop refers to areas cultivated for
               agricultural purposes.  Examples include fields
               used to grow soybeans, corn, and potatoes.  Areas
               that are used for agricultural purposes that are
               temporarily unvegetated due to cropping or tilling
               practices would be categorized as barren or
               developed land.

          3.1.3 Arctic Tundra

               The Arctic Tundra category exists in an
               environment so cold that moisture is unavailable
               for most of the year, precluding the establishment
               of trees, and in which the maximum vegetation
               development is of herbaceous root perennials,
               lichens and mosses, with grasses poorly
               represented or at least not dominant  (Brown et
               al., 19.79).  Tundra categorized in this category
               consists of arctic and near arctic areas north of
               60 degrees latitude.
4.0 Arid
     An Arid region is characterized by dryness, variously
     defined as rainfall insufficient for plant life or for crops
     without irrigation; receives less than 25 cm of annual
     rainfall; or has a higher evaporation rate than
     precipitation rate (Dictionary of Geological Terms, 1976).
     In alpine settings above 5000 feet, the precipitation may be

                               412

-------
between 25 and 75 cm.  The areas with higher amounts of
precipitation support forest communities.  Land cover types
included in this category have a soil regime which is dry
more than half the time when not frozen, and is never moist
for more than 90 consecutive days (Buol et al., 1989).  This
category is divided into two Level II categories: 4.1 Arid
Vegetation; and 4.2 Riparian.

4.1  Arid Vegetation

     The vegetation categories within the arid represent
     upland communities which derive moisture primarily from
     available precipitation, rather than from surface
     water.  Vegetation in these categories would often be
     considered the background or regional vegetation.  The
     category is subdivided into six Level III categories:
     4.1.1 Alpine Tundra; 4.1.2 Arid Forest; 4.1.3 Arid
     Woodland; 4.1.4 Arid Scrubland; 4.1.5 Arid Grassland;
     and 4.1.6 Arid Desertland.

     4.1.1 Alpine Tundra

          Plant communities existing in an environment so
          cold that moisture is unavailable during most of
          the year, which precludes the establishment of
          trees.  The climax vegetation consists of
          herbaceous root perennials, shrubs,  lichens, and
          mosses with grasses poorly represented.  Areas
          categorized as 4.1.1 differ from tundras
          categorized as 3.1.3 by lying at high altitudes
          above treeline and in an alpine setting.

     4.1.2 Arid Forest

          Arid Forest areas are typically open forests
          dominated by coniferous trees over 6 meters in
          height.   The understory may consist of grasses or
          other mesophytic plants,  or there may be no
          understory vegetation.   These forest areas differ
          from woody forested areas in that the annual
          precipitation is less than 75 centimeters and is
          an alpine setting.   If found below 5000 feet the
          annual rainfall is less than 25 centimeters.

     4.1.3 Arid Woodland

          Arid Woodlands are those areas which typically
          contain trees under 15  meters in height which are
          widely spaced with singular canopies.


     4.1.4 Arid Scrubland

                          413

-------
          Communities dominated by either sclerophyllic or
          microphyllic shrubs which generally do not exceed
          6 meters in height.  The canopy can be closed, or
          open with a perennial understory interspersed
          between shrubs.  Areas with more than 25 percent
          of the ground area covered by shrubs fall into
          this, category.

     4.1.5 Arid Grasslands

          Arid Grasslands include those areas in which the
          dominant vegetation consists of grasses and other
          herbaceous plants.  There may be some shrubs
          scattered within this category but less than 25%.
          In some cases the differentiating factor between
          this category and the woodland is the spacing of
          the trees.  Areas with more than 75 percent of the
          ground cover containing grass fall into this
          category.

     4.1.6 Arid Desertland

          Vegetation in this category can be either annual
          or perennial.  The areas between plants usually
          consists of rock or sand with virtually no
          intervening vegetation.
4.2  Riparian
     It is understood that riparian communities exist within
     all types of landscapes.  These communities are
     wetlands associated with a stream, or river.  The
     defining factor in this category is that the average
     annual precipitation is less than 25 centimeters.  Four
     Level III categories lie within this category: 4.2.1
     Arid Swamp Forest; 4.2.2 Arid Swampscrub; 4.2.3 Arid
     Marshland; and 4.2.4 Arid Strandland.

     4.2.1 Arid Swamp Forest

          This community has an overstory of trees over 6
          meters high.  There is often a shrub and
          herbaceous understory within this category.
          Cottonwoods (Populus spp.) and sycamores (Platanus
          spp.) are typical dominants within this category.

     4.2.2 Arid Swampscrub

          The dominant vegetation within this category
          consists of low trees or shrubs below 6 meters in
          height.  Willows (Salix spp.,) and alder  (Alnus
          spp.) are common dominants within this category.

                          414
                                                                 f
                                                                ,; i

-------
           4.2.3 Arid Marshland

               Wetland  communities  in which the principle plants
               are herbaceous emergents which have  their basal
               portions annually, periodically, or  continually
               submerged.   Examples of dominant vegetation  in
               this category would  be cattails  (Typha  spp.), reed
                (Phragmites  spp.), and bulrush  (Scirpus spp.).  At
               the edge of  these communities there  is  often some
               intermingling of nearby upland species.  In
               addition to  the emergent plants, aquatic bed
               vegetation such as pondweed (Potamogeton spp.) are
               often present.  These areas are often associated
               with fairly  extensive bodies of standing water.

           4.2.4 Arid Strandland

               The dynamic  nature of these communities precludes
               the establishment of perennial vegetation,
               although annuals occasionally are found.  The
               surface  of these areas often consists of sand or
               cobbles.  Examples of these communities include
               desert washes, and beach areas within an arid
               context.  Beach and  river channel communities
               subject  to infrequent but periodic submersion,
               wind driven  waves and/or spray are also included.
               Plants are separated by significant  areas devoid
               of perennial vegetation1  (Brown et  al.,  1979).
5.0  Snow/Ice
     This category includes areas that are covered with snow or
     ice.  Anderson et al.  (1976) defines snow and ice in the
     following way:

          Certain lands are covered either seasonally or have a
          perennial cover of either snow or ice.  A combination
          of environmental factors cause some of these areas to
          survive the summer melting season, while in other areas
          the snow and ice melt to reveal another category or
          type of land cover.
     !Strand communities are situated in harsh physical
environments that produce their characteristic physiognomy.
Accordingly, strandland is' treated as the wetland equivalent of
desert land.  While occurring in the usual sense on beaches and
other seacoast habitats, freshwater  (or interior) strands also
occur in river channels, along lake margins, and below reservoir
high water lines (Brown et al., 1979).
                               415

-------
     This category is not subdivided into Level II categories,
     rather it is subdivided at Level III.

     5.1 Snow/Ice

          This category is subdivided into two Level III
          categories.

          5.1.1 Snow

               This category includes all areas which are covered
               with snow.

          5.1.1 Ice/Glacier

               This category refers to those areas that are
               permanently covered with ice as well as glaciers.
               Ice and superficial snow persists throughout the
               year and flows downhill under its own weight.
WATER

The Water Division contains all areas that are covered either
permanently or semi-permanently covered by water as the primary
substrate.  There is only one Level 1 category: 6.0 Water and
Submerged Land.

6.0 Water and Submerged Land

     This category contains areas representing standing and deep
     water, either natural or man-made.  The single feature of
     this category is that the soil or substrate is saturated
     with and covered by water.  Man-made areas of water would
     include reservoirs, impoundments, diked areas, ponds, and
     canals.  The Water and Submerged Land category is subdivided
     into three Level II categories: 6.1 Ocean; 6.2 Coastal; 6.3
     Near-Shore; and 6.4 Inland.

     6.1 Ocean

          This category contains those marine waters which extend
          seaward from the land and coastal region.  Waters with
          relatively high salinities also exist in this category.
          The category does not include bays or estuaries that
          extend inland, for these are included under Coastal.
          Examples are: the Atlantic Ocean, Pacific Ocean, Gulf
          of Mexico, and the Arctic Ocean.

     6.2 Coastal

          Coastal waters include bays, estuaries,  and lagoons

                               416

-------
          which lie between terrestrial and marine environments.
          Sheltered water bodies behind barrier beaches, and
          tidal mudflats are examples of this type of
          environment.  The water regime in these areas include
          marine and estuarine.  Examples of these areas would be
          Chesapeake Bay, Puget Sound, or San Francisco Bay.

     6.3 Near-Shore

          Includes the zone which lies between the normal high
          water mark and the extent of the emergent and floating
          vegetation into the open water.

     6.4  Inland

          This category includes ponds, lakes, reservoirs,
          streams, and rivers that are not considered coastal or
          ocean.  The water is most often fresh but may be
          brackish or highly saline (Great Salt Lake).


OTHER

The categories included in this division appear both over Land
and Water.  There is only one Level 1 category under the
division: 7.0 Other.

7.0 Other

     The Other category contains unidentified resources or
     anomalous data which can not be interpreted to fit into any
     of the land cover categories or water above.   This category
     is divided into:

     7.1 Cloud

          Clouds that  are present in the imagery that obscure
          land cover categories that exist underneath them.   This
          area also contains hazed areas where the land cover
          categories  are obscured.

     7.2  Shadow

          Shadows  obscure the land cover categories underneath
          them.  The shadows may be caused by clouds or terrain.

     7.3  Missing

          This category includes any imagery  that  may be missing
          due to data  drop out,  striping of data,  or other
          anomalies in the data set.
                               417

-------
7.4 Indeterminable

     This category includes those areas that cannot be
     identified with remote sensing sources due to
     conditions such as fire.  These categories could be
     spectral clusters that contain only a few pixels, or
     may be large clusters that contain a wide variety of
    . land cover resources.
                           418

-------
                             REFERENCES


 Anderson, J.R., Hardy, E.E., Roach, J.T.,  and  R.E.  Witman,  1976,
 A Land Use and Land Cover Classification System  for Use with
 Remote sensor Data.  U.S. Geological  Survey Professional Paper
 964,  Washington, D.C., 28 pp.
 Brown,  D.E., 1982, Biotic Communities of the American Southwest-
 United States and Mexico.  Vol. 4, Numbers  1-4, University of
-Arizona,  Superior, AZ, 342 pp.
 Boul,  S.W.,  Hole, F.D., and R.J. McCracken, 1989,  Soil  Genesis
 and Classification.  Iowa State University Press,  Ames,  IA  446
 pp.

 Cowardin,  L.M.,  Carter, V., Golet, F.C., and E.T.  LaRoe,  1979,  A
 Classification of Wetlands and Deepwater Habitats  of  the United
 States,  Office of Biological Services, Fish and Wildlife Service,
 U.S. Department of the Interior, Washington, D.C.,  103  pp.

 Dictionary of Geological Termsr  1976, Anchor Books, Anchor
 Press/Doubleday,  Garden City, N.Y., 472 pp.

 Harlow,  W.M., and E.S. Harrar, 1969, Textbook of Dendrology;
 Covering the Important Forest Trees of the United  States  and
 Canada.  McGraw-Hill Book Company, New York, N.Y.  511 p.
                               419
                                       •6-U.S. GOVERNMENT PRINTING OFFICE: 1994 - 550-001/80329

-------

-------

-------
EPA/600/R-
CD
00
in










Official Bus
Penalty for
$300
"05"
3-8
a. »
CD
(/>
CD









o o m c
5- CD 3 3
ttfl
l?ig
~_ m 3 ??
i|S w
^n O ^^
sii
00 CD O
if
3J >
CD (a
W CD
CO 3
c5 ^
rr
a
o
3.
o

-------