Sevilleta National Wildlife Refuge, near Albuquerque, New Mexico

\log

12/08/97 - Date this file created. G. Shore.
10/30/01 - Added date/time fields to APPENDIX II section. G. Shore.


\doc

##############################################################
DATA SET CODE AND TITLE

SEV031 AVHRR Biweekly Composites (1990)


##############################################################
ABSTRACT
This dataset contains 19 separate 14-day composited AVHRR
images for 1990 clipped approximatedly to the New Mexico State
boundaries (plus 50 Km buffer around boundary). These were
obtained from the U.S. Geological Survey's EROS Data Center (EDC),
National Mapping Division, from the "CONTERMINOUS U.S. AVHRR BIWEEKLY
COMPOSITES" CD product series. NOTE: two early 1990 14-day composite
periods were contained on the 1989 CD's as periods 22 and 23, and are
described in the 1989 SIMS IAF documentation file.


##############################################################
KEYWORDS
SEV031 AVHRR Advanced_Very_High_Resolution_Radiometer NOAA Satellite_Imagery Remote_Sensing USGS_EROS_Data_Center 1990


##############################################################
TABLE OF CONTENTS
I. WHY THE DATA WERE COLLECTED
II. WHEN THE DATA WERE COLLECTED
III. WHO IS INVOLVED WITH THE DATA
IV. WHERE TH DATA WERE COLLECTED
V. HOW THE DATA WERE COLLECTED AND PROCESSED BY THE USGS EDC
VI. HOW THE DATA WERE PROCESSED BY THE SEVILLETA IMS (SIMS)
VII. APPENDIX I - USGS EDC Metadata
VIII. APPENDIX II - Pixel Date Attribute Table


##############################################################
I. WHY THE DATA WERE COLLECTED

See the U.S. Geological Survey's EROS Data Center (EDC)
documentation in APPENDIX I below.


##############################################################
II. WHEN THE DATA WERE COLLECTED

Nineteen 14-day composites for 1990. See the U.S. Geological
Survey's EROS Data Center (EDC) documentation in APPENDIX I below
for specific composite periods.


##############################################################
III. WHO IS INVOLVED WITH THE DATA

SOURCE AGENCY CONTACTS:
See the U.S. Geological Survey's EROS Data Center (EDC) documentation
in APPENDIX I below for USGS EDC contacts.

LOCAL SEVILLETA LTER CONTACTS:
Primary contact:
Greg Shore, Sevilleta LTER (gshore@sevilleta.unm.edu).

Principal investigators:
Bruce Milne, Sevilleta LTER (bmilne@sevilleta.unm.edu)

GIS/GPS specialist:
Greg Shore, Sevilleta LTER (gshore@sevilleta.unm.edu).

Data Management
Greg Shore, Sevilleta LTER (gshore@sevilleta.unm.edu)


##############################################################
IV. WHERE THE DATA WERE COLLECTED

The original biweekly composited images covered the
Conterminous United States. However, the images were clipped
to the approximate New Mexico State boundaries (plus 50 Km buffer
around boundary) for online access purposes, while the full U.S.
scenes are stored offline on tape. The clipping coordinates were
selected to perform exact clipping, so no resampling was required.
The approximate online (NM clipped) boundaries are:

Latitude/Longitude, decimal degrees (Clarke 1866 spheroid)
XMIN: -109.515170 YMIN 30.759247 XMAX -102.454601 YMAX 37.817184

Lambert Azimuthal Equal-area (see EDC document for projection info)
XMIN: -914000 YMIN -1529000 XMAX -216000 YMAX -795000

The source and the clipped images are in the following map projection:
Lambert Azimuthal Equal Area projection
Parameters:
Radius of sphere 6,370,997.0 meters
Longitude of central meridian 100 00 00 West
Latitude of origin 45 00 00 North
False easting 0
False northing 0
Units of measure meters
Pixel size 1,000 meters

Each clipped image has 735 rows and 699 columns, and has a cell size
of 1000 x 1000 m.

See the U.S. Geological Survey's EROS Data Center (EDC) documentation
in APPENDIX I below for spatial extent, projection information, etc.,
related to the full Conterminous US scenes.


##############################################################
V. HOW THE DATA WERE COLLECTED AND PROCESSED BY THE USGS EDC

See the U.S. Geological Survey's EROS Data Center (EDC)
documentation in APPENDIX I below.


##############################################################
VI. HOW THE DATA WERE PROCESSED BY THE SEVILLETA IMS (SIMS)

The general SIMS processing steps for each biweekly composite
period were to read the 10 image bands off the source USGS-EDC CD,
concatenate/import them into a 10-band ERDAS Imagine format image
file, georegister to the USGS-EDC specifications, clip the scene
to the approximate New Mexico State boundaries, move the clipped
scene to the online SIMS archive, and write the full scene to
offline tape.

The band order in the archived scenes is:

1- NOAA CHANNEL 1 6- NDVI
2- NOAA CHANNEL 2 7- SATELLITE ZENITH
3- NOAA CHANNEL 3 8- SOLAR ZENITH
4- NOAA CHANNEL 4 9- RELATIVE AZIMUTH
5- NOAA CHANNEL 5 10- DATE

The bands are described in detail in the U.S. Geological Survey's
EROS Data Center (EDC) documentation in APPENDIX I below. The
attribute data for the "DATE" band is found in APPENDIX II below.

A more precise description of the processing steps is as follows:

1. Mount AVHRR CD

2. Read the 10 bands of information off the CD for each biweekly composited
scene, concatenate 21 NULL bytes onto the last line of each band, concatenate
the 10 bands into a single file, then import as an ERDAS Imagine 10-band
image file (of size 2889 rows x 4587 cols). This process was automated
with the C-shell script:

/db/local/imagery/bincom/avhrr_import.csh 90

NOTE: this generates 10-band ERDAS Imagine format files, that are of
image size 2889 rows x 4587 cols, with filenames as follows:

avhrr90pPP.img

where: PP = bi-weekly growth period number

and bands in following order::
1- NOAA CHANNEL 1 6- NDVI
2- NOAA CHANNEL 2 7- SATELLITE ZENITH
3- NOAA CHANNEL 3 8- SOLAR ZENITH
4- NOAA CHANNEL 4 9- RELATIVE AZIMUTH
5- NOAA CHANNEL 5 10- DATE

3. Georegister the scene as follows:
Bring up ERDAS Imagine GUI, then Tools->ImageInfo tool and do:
1. File->Open->avhrr90pPP.img
2. Edit->Change Map Model:
a. Upper Left X: -2050000
b. Upper Left Y: 752000
c. Pixel Size X: 1000
d. Pixel Size Y: 1000
e. Units: meters
f. Projection: Lambert Azimuthal Equal-area
NOTE: click OK, then answer "Yes" to changing Map Model in all layers.
3. Edit->Add/Change Projection:
a. Spheroid Name: Sphere of Radius 6370997m
b. Datum Name: Undefined
c. Longitude of center of projection: 100:00:00 W
d. Latitude of center of projection: 45:00:00 N
e. False easting: 0.0 meters
f. False northing: 0.0 meters
NOTE: click OK, then answer "Yes" to changing Map Model in all layers.
4. Edit->Change Layer Name:
a. Change bands 1 - 5 to Channel_1, Channel_2, ..., Channel_5,respectively
b. Change band 6 to NDVI
c. Change bands 7 - 10 to SATELLITE_ZENITH, SOLAR ZENITH,
RELATIVE_AZIMUTH, and DATE, respectively

4. Clip full-scene to New Mexico minimum bounding box (+50 Km buffer) with
coordinates ULx = -914000, ULy = -795000, LRx = -216000, LRy = -1529000,
and dimensions 735 rows x 699 columns. This process was automated with
the C-shell scripts:

/db/local/imagery/bincom/batch_avhrr_clip2nm.csh, which calls:
/db/local/imagery/bincom/avhrr_clip2nm.csh avhrr90pPP.img avhrr90pPPnm.img

NOTE: this generates 10-band ERDAS Imagine format files, that are of
image size 735 rows x 699 cols, with filenames as follows:

avhrr90pPPnm.img

5. Unix compress the NM clipped image, make an archive directory, and move it
to the archive destination (/db/archive/imagery/avhrr/avhrr90pPP/). This
process was automated with the C-shell scripts:

/db/local/imagery/bincom/avhrr_archive.csh

NOTE: this generates Unix compressed files with filenames as follows:

avhrr90pPPnm.img.Z

6. Compress (gzip) and archive the full scene image to 4mm DAT tape, then
remove from online disk:

gzip avhrr90pPP.img
mt -f /dev/rmt/0cn fsf <#>
tar cvf /dev/rmt/0cn avhrr90pPP.img.gz
rm avhrr90pPP.img.gz

7. Copy the most current yearly "USGS-EDC-AVHRR Dataset README" file off CD for
inclusion in the online SIMS IAF metadata file (avhrr90.dbf) for the images
(note, must also convert from DOS to Unix file):

dos2unix /cdrom/cdrom0/readme.1st avhrr90readme.1st

8. Copy the "Date of Acquisition by Pixel" attribute file off each CD, then
split by biweekly period for inclusion in GIS/RS Metadata Abstract file
(avhrr90pPPnm.mda), and merge by year for inclusion in SIMS IAF metadata
file (avhrr90.dbf):

dos2unix /cdrom/cdrom0/geom/date.att avhrr90pPP-PPdate.att

/db/local/imagery/bincom/batch_avhrr_split_dateatt.csh, which calls:
/db/local/imagery/bincom/avhrr_split_dateatt.csh

/db/local/imagery/bincom/avhrr_concat_dates.csh 90 > avhrr90date.att

9. Generate (this) SIMS IAF metadata file (i.e., avhrr90.dbf) using template
file (avhrr_dbf.tmpl), appending readme file (avhrr90readme.1st), and date
file (avhrr90date.att).

10. Generate GIS/RS Metadata Abstract file for each image, and put with
image in appropriate directory (/db/archive/imagery/avhrr/avhrr90pPP/):

/db/local/imagery/bincom/get_avhrr_dates.csh 90 avhrr90.dbf 97 \
> avhrr90periods.txt

/db/local/imagery/bincom/batch_make_avhrr_abstract.csh avhrr90periods.txt
which calls:
/db/local/imagery/bincom/make_avhrr_abstract.csh


##############################################################
VII. APPENDIX I - USGS EDC Metadata

THE 1990 CONTERMINOUS U.S. AVHRR DATA SET


INTRODUCTION

In 1987, the U.S. Geological Survey, EROS Data Center (EDC) began
real-time reception of Advanced Very High Resolution Radiometer
(AVHRR) data from operational NOAA polar orbiting satellites.
EDC receives AVHRR High Resolution Picture Transmission (HRPT)
data for the entire conterminous United States, southern Canada,
and northern Mexico. The AVHRR Data Acquisition and Processing
System (ADAPS) includes a tracking antenna, a data receiving
subsystem, and a minicomputer with associated peripherals for
data processing. The central location of the EROS Data Center
within the U.S. enables direct reception of all satellite
overpasses of the conterminous U.S., as well as much of Canada
and Mexico.

Early in the 1990 growing season EDC began to use the NOAA-11
AVHRR (1km resolution) daily observations to produce weekly and
biweekly maximum normalized difference vegetation index (NDVI)
composites of the conterminous United States. This work follows
experiments conducted in 1988 over a five state area in the
northern Great Plains and the western half of the U.S. in 1989.

The objective of the program was to compile a comprehensive time
series data set of calibrated, georegistered daily observations
and biweekly maximum NDVI composites for the 1990 growing season.
This data set, and similar data sets planned for the future, have
applications for environmental monitoring and assessing impacts
of global climate change.

The diversity of the area provides opportunities for studying the
uses of AVHRR data and NDVI for monitoring vegetation condition
in several different ecosystems including forests, agricultural
crops, and grasslands. The data set provides a comprehensive
growing season profile of these ecosystems and is extremely
useful for assessing seasonal variations in vegetation condition
and provides a foundation for efforts to study long-term changes
resulting from human interactions or natural factors.


DATA SET CHARACTERISTICS

The data set is comprised of nineteen biweekly maximum NDVI
composites which were created from nearly four hundred processed
NOAA-11 daily observations. The first seventeen composite
periods represent a continuous period from March 2, 1990 to
October 22, 1990. The last two composites (periods 18 and 19)
represent a two week period in November and December since a
single biweekly composite in each month was deemed sufficient to
document any changes during the winter period.


Each daily observation includes nine bands of information which
are: AVHRR channels 1-5, NDVI, satellite zenith, solar zenith,
and relative azimuth. The daily observations have been
calibrated to reflectance, scaled to byte data, and geometrically
registered to the Lambert Azimuthal Equal Area map projection.

Each biweekly composite includes ten bands of information which
include the nine described for a daily observation and a tenth
band which is a pointer to identify the date of the source daily
observation scene. The data for each pixel in the composite is
extracted from the daily observation scene based on the maximum
NDVI compositing process.

The following is a list of the composite period dates:

Calendar date Julian Period
range
-------------------------------------------------------
Mar 2 - Mar 15 03/02 - 03/15/1990 061/074 1
Mar 16 - Mar 29 03/16 - 03/29/1990 075/088 2
Mar 30 - Apr 12 03/30 - 04/12/1990 089/102 3
Apr 13 - Apr 26 04/13 - 04/26/1990 103/116 4
Apr 27 - May 10 04/27 - 05/10/1990 117/130 5
May 11 - May 24 05/11 - 05/24/1990 131/144 6
May 25 - Jun 7 05/25 - 06/07/1990 145/158 7
Jun 8 - Jun 21 06/08 - 06/21/1990 159/172 8
Jun 22 - Jul 5 06/22 - 07/05/1990 173/186 9
Jul 6 - Jul 19 07/06 - 07/19/1990 187/200 10
Jul 20 - Aug 2 07/20 - 08/02/1990 201/214 11
Aug 3 - Aug 16 08/03 - 08/16/1990 215/228 12
Aug 17 - Aug 30 08/17 - 08/30/1990 229/242 13
Aug 31 - Sep 13 08/31 - 09/13/1990 243/256 14
Sep 14 - Sep 27 09/14 - 09/27/1990 257/270 15
Sep 28 - Oct 11 09/28 - 10/11/1990 271/284 16
Oct 12 - Oct 25 10/12 - 10/25/1990 285/298 17
Nov 9 - Nov 22 11/9 - 11/22/1990 313/326 18
Dec 7 - Dec 20 12/7 - 12/20/1990 341/354 19

The actual data dimensions of each band is 2889 lines and 4587
samples (13.2 megabytes). However, in order to accommodate use of
the disc on the network which can be accessed by the Land
Analysis System (LAS) software at EDC and other facilities it was
necessary to create images for which the number of samples is a
multiple of 512. As a result the images are stored as 2889 lines
and 4608 samples. The last 21 samples of each image are blank
(zero).

PROCEDURES

The following text will describe the data processing flow that
was used at EDC to create a composite data set. The flow will be
covered under the topics of scene selection, calibration,
computation of satellite and solar viewing geometry, geometric
registration, computation of NDVI, compositing, and archiving
(including tape formats). All image processing was conducted
using software in the Land Analysis System (LAS) (Ailts et al,
1990).


SCENE SELECTION

Cloud free AVHRR observations of the land surface are necessary
for monitoring the vegetation conditions. The likelihood of a
single AVHRR overpass being completely cloud free is small.
Holben (1986) showed that compositing AVHRR data acquired over
several days produces spatially continuous cloud-free imagery
over large areas with sufficient temporal resolution to study
green-vegetation dynamics. The duration of consecutive daily
observations is referred to as the compositing period. On a
daily basis during a composite period each observation of NOAA-11
data over the conterminous U.S. was evaluated for cloud cover.
Generally there are two observations per day, an eastern and a
western pass. Every image which provided a clear observation of
a large ground surface area at reasonable nadir viewing angles
was included in the composite. On an average, 18 daily
observations per biweekly period were included in the composite.


RADIOMETRIC CALIBRATION

Radiometric calibration is an important consideration with all
remote sensing data. Calibration of AVHRR data is especially
troublesome since there is no onboard calibration capability for
the visible and near-infrared channels. The calibration
coefficients provided with the data are based on prelaunch
measurements of the sensors. Field calibration studies of
several AVHRR sensors have shown apparent sensor degradation over
time. However, such studies regarding NOAA-11 are still
inconclusive. There has been a recent update of prelaunch
coefficients provided by NOAA (September 1990), but in order to
maintain consistency within the 1990 data set EDC has chosen to
utilize prelaunch coefficients published by NOAA in April 1989.

The calibration and a correction for solar illumination has been
performed using the complete 10-bit range. The visible and near-
infrared channels are converted to reflectance using the
following formula:

R=(d*d/z)*(a+b*c)

=(d*d*a)/z + (d*d*b*c)/z

where:
R is reflectance

d is the mean earth-sun distance (A.U.)

z is the cosine of the solar zenith angle

a is the intercept

b is the gain coefficient

c is the digital count


When the visible and near infrared values are corrected for
illumination variability we update (d/z) every five scan lines,
assuming (d/z) does not vary across a scan line.

The resulting data range is 0 - 1000 where each bin represents
0.1% reflectance. The calibrated channel 1 and 2 data are
converted to byte range by first scaling the 0 - 1000 range to
0 - 400 so that each bin represents 0.25% reflectance. Then the
0 - 400 range is scaled to byte such that values in the range 0 -
254 remain the same and all values equal to or greater than 255
are scaled to 255. The range 0 - 254 represents 0 - 63.5%
reflectance (254 * 0.25%) and the value 255 is a grouping of
reflectance values greater than 63.5%. Generally, any surface
with greater than 63% reflectance is a cloud, snow, or other
bright non-vegetated surface.

The AVHRR channel 3,4, and 5 data have been converted to
radiance using in-flight coefficients in the following
formula:

R=a+bc

where:

R is radiance

a is the intercept

b is the gain coefficient

c is the digital count

Radiance is then converted to brightness temperature using the
inverse of Planck's radiation function. The brightness
temperatures are represented in Kelvin units. Two different
scaling factors were used to convert to byte data. For data
processed up through June 21, 1990, 190 is subtracted from the
brightness temperature values and the difference is multiplied by
two in order to scale the brightness temperature values down
to byte range and to maintain one half percent accuracy (i.e., a
brightness temperature value of 250.58 becomes 121). For data
processed after June 21, 1990, 202.5 is subtracted from the
brightness temperature values and the difference is multiplied by
two in order to scale the brightness temperature values down to
byte range and to maintain one half percent accuracy (i.e., a
brightness temperature value of 250.58 becomes 96). The early
scaling factor tends to lump high brightness temperatures at value
255, where as, the later scaling factor provides more sensitivity
at high brightness temperatures.


SATELLITE AND SOLAR VIEWING GEOMETRY

The availability of the viewing geometry information enables
studies on the effects of off-nadir viewing and the investigations
of potential data correction techniques.

The computation of the solar/satellite geometry is a process which
derives the satellite zenith, solar zenith and relative azimuth
angle for each image pixel. The relative azimuth is the absolute
difference between the satellite and solar azimuth angles. A seperate
single band image file is created for each of these three angle
computations.

The computed angles do not exceed 180 degrees. The satellite
zenith angle is computed in degrees in which nadir is represented
as 90 degrees. Therefore values less than 90 represent view
angles in the negative (westerly) direction and values greater
than 90 represent positive (easterly) view angles. Note that the
effective field of view of the satellite is approximately 55
degrees each side of nadir, but computed satellite zenith angles
can exceed 55 degrees because of the curvature of the earth.

The relative azimuth angle is computed as the absolute difference
between the solar azimuth and the satellite azimuth angles. The
computed values are in the range 0 - 180. The relative azimuth
angle is computed instead of the separate azimuth angles for two
reasons. First, the relative azimuth angle along with satellite
and solar zenith angles are required for atmospheric correction
algorithms and secondly, the relative azimuth angle requires only
one band in a daily observation and composite image instead of
two.


DATE OF ACQUISITION

The date of acquisition images are provided to allow a user to
identify the specific daily observation used for each pixel. The
date images for each composite identify each daily input image as
a unique value. The unique value is linked to an inventory of
the daily observations. A complete list of daily observations
used in each composite period is on this disc under the \GEOM
directory as file DATE.ATT.


GEOMETRIC REGISTRATION

The compositing process requires each daily overpass to be
registered to a common map projection with precise registration
to ensure that from day to day each 1 km pixel represents the
exact same ground location.

An evaluation of image-to-image registration using automated
correlation techniques showed an improvement in throughput and
geometric accuracy (RMSE less than 1.0 pixel) over image-to-map
procedures. To provide a base image of the conterminous U.S. for
correlation, the U.S. Geological Survey (USGS) 1:2,000,000
Digital Line Graph data set was transformed to the Lambert
Azimuthal Equal Area map projection. This map projection was
chosen because of it is appropriate for the North American
continent and because the equal area characteristic enables easy
measurement of area throughout the data. Approximately 20 near
nadir cloud-free segments of NOAA-11 channel 2 daily observations
from the 1989 growing season were manually registered to the DLG.
Each segment was verified for accuracy (RMSE less than 1.0
pixel). The segments were digitally mosaicked to produce a
single base image of the conterminous U.S. for use in registering
the 1990 growing season data. The accuracy of this base image was
verified with an root mean square error less than 1.0 pixel.
Table 1 provides details on projection parameters.


Table 1. LAMBERT AZIMUTHAL EQUAL AREA projection parameters:
-----------------------------------------------------------------
Longitude of central meridian 100 00 00 W
Latitude of origin 45 00 00 N
False Easting 0
False Northing 0
Units of measure Meters

Pixel size 1000 meters



For the Conterminous U.S. (1990)

Center of pixel (1,1) ( -2050000, 752000 )
Number of lines 2889
Number of samples 4587
LAZEA minimum bounding rectangle:
Lower Left ( -2050500, -2136500 )
Upper Left ( -2050500, 752500 )
Upper Right ( 2536500, 752500 )
Lower Right ( 2536500, -2136500 )

Lower Left ( -119.9722899, 23.5837576 )
Upper Left ( -128.5300591 48.4030555 )
Upper Right ( -65.3946489 46.7048989 )
Lower Right ( -75.4163527 22.4793919 )

Lower Left ( -119 58 20 23 35 02 )
Upper Left ( -128 31 48 48 24 11 )
Upper Right ( -65 23 41 46 42 18 )
Lower Right ( -75 24 59 22 28 46 )

-----------------------------------------------------------------

Each daily observation for the 1990 growing season was
registered, using image-to-image correlation, to the base image,
using the following procedures. First, the channel 2 data for
each daily observation are roughly transformed using the
satellite transformation information in the ephemeris data.
Next, correlation is performed using a set of 255 selected ground
control points. If all the ground control points are cloud
covered in the daily observation no correlation is defined and
the image is rejected. Otherwise, the correlation is determined
and the satellite transformation coefficients are revised. Then
the raw data (channels 1 - 5) and satellite geometry data are
transformed using the revised coefficients and nearest neighbor
resampling.


NORMALIZED DIFFERENCE VEGETATION INDEX (NDVI)

NDVI is calculated from calibrated data which has been scaled to
byte range and geometrically registered. NDVI is the difference
of near-infrared (AVHRR Channel 2) and visible (AVHRR Channel 1)
reflectance values divided by total reflectance as follows:

IR(Band 2) - Visible(Band 1)
NDVI = ----------------------------------------
IR(Band 2) + Visible(Band 1)

The equation produces NDVI values in the range of -1.0 to 1.0,
where negative values generally represent clouds, snow, water,
and other non-vegetated surfaces while positive values represent
vegetated surfaces.

In order to scale the computed NDVI results to byte data range the
NDVI data range of -1.0 to 1.0 is scaled to the range of 0 to
200, where computed -1.0 equals 0, computed 0 equals 100, and
computed 1.0 equals 200. As a result, NDVI values less than 100
now represent clouds, snow, water, and other non-vegetative
surfaces and values greater than 100 represent vegetative
surfaces.


COMPOSITING

The method for determining the portion of each overpass to be
included in the composite image is to retain pixels having the
highest NDVI values. NDVI is examined pixel-by-pixel for each
overpass within the biweekly compositing period to determine the
maximum value.

The retention of the highest NDVI reduces the number of cloud
contaminated pixels because values for clouds and cloud shadows
are generally less than 100 (in the byte scaled data) while clear
day observations of vegetated surfaces are greater than 100 (in the
byte scaled data). The result is a near cloud free image which
depicts the maximum vegetative greenness for the compositing period.

There is one inherent problem with this process over water bodies.
The NDVI value of water is much lower than it is for a cloud and as
a result a cloudy observation is chosen instead of a clear observation
over water. In an attempt to retain a clear observation over water,
the negative NDVI values (-1.0 to -0.01), which are the values 0 to 99
in the scaled to byte data, were flipped in the daily scenes so that
clear observations of water would have a higher NDVI value than a
cloud. For example, the NDVI value of 75 (water in the scaled to
byte data) and 99 (cloud in the scaled to byte data) would be 25
and 1 respectively after the data were flipped and the water would
be selected as the maximum value in the compositing process.
Unfortunately, clouds are often slightly greater than 100 in the
scaled to byte data and cloudy observations routinely are selected
over water bodies.


The output of the compositing process is a ten band image which
includes the maximum NDVI value for each pixel during the
composite period, the channel 1-5 and satellite viewing geometry
data from the chosen daily observations, and a pointer value
which identifies the scene id of the observation. Table 2 lists
the data included in each of the ten bands.

Table 2.
-----------------------------------------------------------------

BAND DESCRIPTION OF COMPOSITE IMAGES
_________________________________________________________________

1- NOAA CHANNEL 1 6- NDVI
2- NOAA CHANNEL 2 7- SATELLITE ZENITH
3- NOAA CHANNEL 3 8- SOLAR ZENITH
4- NOAA CHANNEL 4 9- RELATIVE AZIMUTH
5- NOAA CHANNEL 5 10- DATE


-----------------------------------------------------------------

The date of acquisition pointer is provided to allow a user to
identify the specific daily observation used for each pixel. To
determine the date and scene id of the daily images you must
identify the date pointer value for a specific pixel within a
specific period and use the reference table in file DATE.ATT to
determine the scene id.

MISCELLANEOUS DATA

Often when displaying data covering large areas with AVHRR data
it is beneficial to include as an overlay or mask of familiar
linework such as county boundaries as a location aid. Several
images have been included which provide location information.
All of the linework images represent lines in raster format as 1
km cells just as the AVHRR data. These data sets include
climatic division boundaries (CDLINES), major land resource areas
boundaries (LRALINES) and county boundaries (CTYLINES). The
climatic division lines have been digitized from NOAA base maps.
The county lines are a modified version of the USGS 1:2,000,000
Digital Line Graph (DLG) data. The major land resource area
boundaries have been digitized from Soil Conservation Service
maps.

The linework in the CDLINES and LRALINES images is coded at the
byte value 255. In the CTYLINES image, the county boundaries
identified by the coasts and international borders are at value
253, the county borders that are coincident with state are at
value 254, and other county boundaries are at value 255. This
variable coding provides the capability to display coastal,
state, or county boundaries from the same image.

Also included are three raster polygon images that can be used,
for example, in an overlay process where histograms or
descriptive statistics could be computed for the NDVI values
within a polygon. These images include climatic divisions
(CDPOLY), counties (CTYPOLY), major land resource areas
(LRAPOLY).

Each polygon is in raster format and has a unique numeric
identifier. Images which include more than 256 unique polygons
are stored in I*2 integer (16 bit) format.

The attribute information which identifies or characterizes each
polygon have been included on this disc under the \MISC directory.
The attributes for the county polygons are in CTYPOLY.ATT. The
fields in the file are:

cntyid -- the unique polygon id number
npixels -- the number of pixels in each county
FIPS -- the FIPS state and county code for each county
cname -- the county name
sname -- the state name

The attributes for the major land resource area polygons are in
LRAPOLY.ATT. The fields in the file are:

polyid -- the unique polygon id number
mlra -- the MLRA identification code used by SCS
lratext -- text description of the MLRA used by SCS

The attributes for the climatic division polygons are limited.
In fact, the unique polygon id can be parsed into the state and
climatic division number within the state. For example, climatic
district one in Arizona is polygon number 401, 4 is the FIPS
state identification number and 01 is the number one division;
climatic district one in Oklahoma is polygon number 4001, where
40 is the FIPS state code and 01 is the number one division.


Table 3 provides a list of miscellaneous image file
characteristics.

Table 3.
-----------------------------------------------------------------
NAME TYPE BANDS LINES SAMPLES
-----------------------------------------------------------------
LRAPOLY I*2 1 2889 4608
LRALINES BYTE 1 2889 4608
CDPOLY I*2 1 2889 4608
CDLINES BYTE 1 2889 4608
CTYPOLY I*2 1 2889 4608
CTYLINES BYTE 1 2889 4608
-----------------------------------------------------------------




DISK ORGANIZATION

A large volume of data was generated during the construction of
this data base. In fact, the data stored on this one CD-ROM
required 10 6250 bpi magnetic tapes. The data on this disc have
been organized in a directory structure which logically separates
the data components. This structure is shown below:

README.1ST

\AVHRR
README \LABELS \IMAGES

\NDVI
README \LABELS \IMAGES

\GEOM
README \LABELS \IMAGES

\MISC
README \LABELS \IMAGES


Each directory on the disc contains data which is similar in
type. Each directory also contains an ASCII text file (README)
that details the contents of the directory.

The data files and LAS header files (.DDR) are in the \IMAGES
subdirectories, the image label files in the \LABELS
subdirectories. The binary image files have been put on the disc
with 512 byte header record. This header record is utilized by the
LAS image processing system.

To help get a quick start looking at the image files, label files
for each image have been included in the \LABELS directory using
identical file names as the image file it describes in the
\IMAGES subdirectory. These label files were designed for use by
the public domain MS-DOS personal computer IMDISP image display
software developed by Mike Martin at NASA's Jet Propulsion
Laboratory in Pasadena, CA. IMDISP users can access the images
on this CD-ROM by selecting the image name in the \LABELS
subdirectory which will automatically access all header information
the software requires and quickly retrieve the image data itself.

Remember that the actual data dimensions of each band is 2889
lines and 4587 samples (13.2 megabytes), but the images are
stored as 2889 lines and 4608 samples to accommodate LAS software.
The last 21 samples of each image are blank (zero).

The \AVHRR directory contains the five channels of AVHRR data
associated with the four biweekly composites on this disc. Each
band of each biweekly composite file is uniquely named using the
convention:

P01CH1.IMG
-- -
| |___Channel #
Biweekly
Period #

and stored in the \IMAGES subdirectory.


The \NDVI directory contains the single band computed normalized
difference vegetation index for the biweekly AVHRR composite data
sets and name using the convention:

P01NDVI.IMG
--
|
Biweekly
Period #

The \GEOM directory contains the satellite and solar zenith
and relative azimuth information for each pixel in the AVHRR
composite images. This directory also contains the date images
for each of the composites as well as the DATE.ATT attribute
file.

The \MISC directory contains the political (CTYLINES, CTYPOLY),
climatic divisions (CDLINES, CDPOLY), and land resource area
(LRALINES, LRAPOLY) raster line and polygon images which are
useful for display or in digital analysis procedures. Attribute
files related to these are included as files CTYPOLY.ATT and
LRAPOLY.ATT.

For more detailed technical information please contact Customer
Services, EROS Data Center, Sioux Falls, SD 57198, (605)594-6151,
FAX (605)594-6589.


REFERENCES

Ailts, B., Akkerman, D., Quirk, B., and Steinwand, D., 1990. LAS
5.0 -- An image processing system for research and production
environments, Proceedings of the American Society of
Photogrammetry and Remote Sensing-American Congress of Surveying
and Mapping Annual Convention, Denver, Colorado, Vol. 4, p. 1-12.

Holben, B.N., 1986. Characteristics of maximum-value composite
images from temporal AVHRR data, The International Journal of
Remote Sensing,Vol. 7,No.11,p. 1417.


##############################################################
VIII. APPENDIX II - Pixel Date Attribute Table

NOTE: Periods 18 and 19 were not found on the distribution CD.


PERIOD INDEX SCENEID Date GMT
------ ----- ---------------- ------- --------
1 1 av119006318215 90-063 18:21:5
2 AV119006621120 90-066 21:12:0
3 av119006617511 90-066 17:51:1
4 av119006521230 90-065 21:23:0
5 av119006320021 90-063 20:02:1
6 av119006120240 90-061 20:24:0
7 av119006220130 90-062 20:13:0
8 av119006918573 90-069 18:57:3
9 av119007118361 90-071 18:36:1
10 av119007121595 90-071 21:59:5
11 av119007120170 90-071 20:17:0
12 av119007421265 90-074 21:26:5
2 1 av119007619224 90-076 19:22:4
1 av119007619224 90-076 19:22:4
1 av119007720534 90-077 20:53:4
2 av119007719115 90-077 19:11:5
3 av119007619224 90-076 19:22:4
4 av119007820433 90-078 20:43:3
5 av119007920324 90-079 20:32:4
6 av119007918520 90-079 18:52:0
7 av119008018412 90-080 18:41:2
8 av119008221414 90-082 21:41:4
9 av119008521085 90-085 21:08:5
10 av119008519275 90-085 19:27:5
11 av119008619170 90-086 19:17:0
12 av119008820361 90-088 20:36:1
3 1 av119008920251 90-089 20:25:1
2 av119009121451 90-091 21:45:1
3 av119009221341 90-092 21:34:1
4 av119009219524 90-092 19:52:4
5 av119009321232 90-093 21:23:2
6 av119009419312 90-094 19:31:2
7 av119009319420 90-093 19:42:0
8 av119009521012 90-095 21:01:2
9 av119009519202 90-095 19:20:2
10 av119009620503 90-096 20:50:3
11 av119009921593 90-099 21:59:3
12 av119009918372 90-099 18:37:2
13 av119010119561 90-101 19:56:1
14 av119010221264 90-102 21:26:4
4 1 av119010419234 90-104 19:23:4
2 av119010321154 90-103 21:15:4
3 av119010319344 90-103 19:34:4
4 av119010317545 90-103 17:54:5
5 av119010520535 90-105 20:53:5
6 av119010720320 90-107 20:32:0
7 av119010818404 90-108 18:40:4
8 av119010820211 90-108 20:21:1
9 av119010918300 90-109 18:30:0
10 av119011119483 90-111 19:48:3
11 av119011219375 90-112 19:37:5
12 av119011319265 90-113 19:26:5
13 av119011217580 90-112 17:58:0
14 av119011420565 90-114 20:56:5
15 av119011419160 90-114 19:16:0
16 av119011519051 90-115 19:05:1
17 av119011620350 90-116 20:35:0
5 1 av119011718434 90-117 18:43:4
2 av119011720240 90-117 20:24:0
3 av119011820131 90-118 20:13:1
4 av119012021325 90-120 21:32:5
5 av119011920021 90-119 20:02:1
6 av119012119403 90-121 19:40:3
7 av119012121215 90-121 21:21:5
8 av119012217501 90-122 17:50:1
9 av119012219294 90-122 19:29:4
10 av119012320595 90-123 20:59:5
11 av119012420485 90-124 20:48:5
12 av119012520375 90-125 20:37:5
13 av119012720160 90-127 20:16:0
14 av119012718354 90-127 18:35:4
15 av119012818250 90-128 18:25:0
16 av119012921355 90-129 21:35:5
17 av119012620265 90-126 20:26:5
18 av119013019433 90-130 19:43:3
6 1 av119013119323 90-131 19:32:3
2 av119013319110 90-133 19:11:0
3 av119013221024 90-132 21:02:4
4 av119013320514 90-133 20:51:4
5 av119013420405 90-134 20:40:5
6 av119013622004 90-136 22:00:4
7 av119013620190 90-136 20:19:0
8 av119013720080 90-137 20:08:0
9 av119013819570 90-138 19:57:0
10 av119014021164 90-140 21:16:4
11 av119013821384 90-138 21:38:4
12 av119013918062 90-139 18:06:2
13 av119014121053 90-141 21:05:3
14 av119014220543 90-142 20:54:3
15 av119014320433 90-143 20:43:3
16 av119014418520 90-144 18:52:0
7 1 av119014520214 90-145 20:21:4
2 av119014522033 90-145 22:03:3
3 av119014620103 90-146 20:10:3
5 AV119014818085 90-148 18:08:5
6 av119014819484 90-148 19:48:4
7 av119014919380 90-149 19:38:0
8 av119014518411 90-145 18:41:1
9 av119014719594 90-147 19:59:4
10 av119014921190 90-149 21:19:0
11 av119015019195 90-150 19:19:5
12 av119015119090 90-151 19:09:0
13 av119015120505 90-151 20:50:5
14 AV119015218585 90-152 18:58:5
15 AV119015420171 90-154 20:17:1
16 AV119015520053 90-155 20:05:3
17 av119015521502 90-155 21:50:2
18 av119015621384 90-156 21:38:4
19 av119015618162 90-156 18:16:2
20 av119015619545 90-156 19:54:5
21 av119015821153 90-158 21:15:3
8 1 AH119015921041 90-159 21:04:1
2 AH119016020525 90-160 20:52:5
3 AH119015919222 90-159 19:22:2
4 ah119016019113 90-160 19:11:3
5 ah119016220302 90-162 20:30:2
6 ah119016218502 90-162 18:50:2
7 ah119016320192 90-163 20:19:2
8 ah119016318395 90-163 18:39:5
9 ah119016420081 90-164 20:08:1
10 ah119016421530 90-164 21:53:0
11 AH119016521410 90-165 21:41:0
12 ah119016819245 90-168 19:24:5
13 ah119016621293 90-166 21:29:3
14 ah119016919135 90-169 19:13:5
15 ah119017019032 90-170 19:03:2
16 ah119017120325 90-171 20:32:5
17 ah119016920552 90-169 20:55:2
18 AH119017020441 90-170 20:44:1
9 1 ah119017320104 90-173 20:10:4
2 ah119017321554 90-173 21:55:4
3 ah119017519484 90-175 19:48:4
6 ah119017419594 90-174 19:59:4
7 ah119017621203 90-176 21:20:3
8 ah119017421435 90-174 21:43:5
9 ah119017521315 90-175 21:31:5
10 ah119017721090 90-177 21:09:0
11 ah119017619375 90-176 19:37:5
12 ah119017717501 90-177 17:50:1
13 ah119017719271 90-177 19:27:1
14 AH119017819162 90-178 19:16:2
15 ah119017919054 90-179 19:05:4
15 ah119017919054 90-179 19:05:4
16 ah119018020351 90-180 20:35:1
17 ah119018220130 90-182 20:13:0
18 ah119018120241 90-181 20:24:1
19 ah119018221582 90-182 21:58:2
20 ah119018318231 90-183 18:23:1
21 ah119018519402 90-185 19:40:2
22 ah119018518024 90-185 18:02:4
20 ah119018318231 90-183 18:23:1
23 ah119018118441 90-181 18:44:1
24 ah119018421344 90-184 21:34:4
10 1 ah119018918572 90-189 18:57:2
2 ah119018721001 90-187 21:00:1
3 ah119018920374 90-189 20:37:4
4 ah119018719184 90-187 19:18:4
5 ah119019018464 90-190 18:46:4
6 ah119019020264 90-190 20:26:4
7 ah119019120152 90-191 20:15:2
8 ah119018820485 90-188 20:48:5
9 ah119019218253 90-192 18:25:3
10 ah119019319532 90-193 19:53:2
11 ah119019421252 90-194 21:25:2
12 ah119019419423 90-194 19:42:3
13 ah119019418044 90-194 18:04:4
14 ah119019720510 90-197 20:51:0
15 ah119019517542 90-195 17:54:2
16 ah119019719100 90-197 19:10:0
17 ah119019820401 90-198 20:40:1
18 ah119019818593 90-198 18:59:3
19 ah119019918485 90-199 18:48:5
11 1 ah119020121510 90-201 21:51:0
2 ah119020421161 90-204 21:16:1
3 ah119020221393 90-202 21:39:3
4 ah119020118274 90-201 18:27:4
5 ah119020419335 90-204 19:33:5
6 ah119020519231 90-205 19:23:1
7 ah119020619122 90-206 19:12:2
8 ah119020620532 90-206 20:53:2
9 ah119020719014 90-207 19:01:4
10 ah119020720422 90-207 20:42:2
11 ah119020820311 90-208 20:31:1
12 ah119021018295 90-210 18:29:5
13 ah119020818510 90-208 18:51:0
14 ah119021118192 90-211 18:19:2
15 ah119021121414 90-211 21:41:4
16 ah119021221301 90-212 21:30:1
17 ah119021219470 90-212 19:47:0
18 ah119021319360 90-213 19:36:0
19 ah119021421071 90-214 21:07:1
12 1 ah119021520555 90-215 20:55:5
2 ah119021519143 90-215 19:14:3
3 ah119021619035 90-216 19:03:5
4 ah119021820222 90-218 20:22:2
5 ah119021921554 90-219 21:55:4
6 ah119021920110 90-219 20:11:0
7 ah119022020000 90-220 20:00:0
8 ah119022021435 90-220 21:43:5
9 ah119022119490 90-221 19:49:0
10 ah119022118110 90-221 18:11:0
11 ah119022219395 90-222 19:39:5
12 ah119022519055 90-225 19:05:5
13 ah119022319273 90-223 19:27:3
14 ah119022419163 90-224 19:16:3
15 ah119022618551 90-226 18:55:1
16 ah119022718444 90-227 18:44:4
17 ah119022620352 90-226 20:35:2
18 ah119022820130 90-228 20:13:0
13 1 ah119022918233 90-229 18:23:3
2 ah119022921461 90-229 21:46:1
3 ah119022920020 90-229 20:02:0
4 ah119023221111 90-232 21:11:1
5 ah119023019510 90-230 19:51:0
6 ah119023321000 90-233 21:00:0
7 ah119023317421 90-233 17:42:1
8 ah119023420484 90-234 20:48:4
9 AH119023419075 90-234 19:07:5
10 AH119023620262 90-236 20:26:2
11 AH119023921362 90-239 21:36:2
12 ah119023919530 90-239 19:53:0
13 ah119023818252 90-238 18:25:2
14 ah119024021245 90-240 21:24:5
15 ah119024019420 90-240 19:42:0
16 ah119024121131 90-241 21:13:1
17 ah119024119312 90-241 19:31:2
18 ah119024221015 90-242 21:01:5
19 ah119024219203 90-242 19:20:3
14 1 ah119024320503 90-243 20:50:3
2 ah119024319094 90-243 19:09:4
3 ah119024418585 90-244 18:58:5
4 ah119024720055 90-247 20:05:5
5 ah119024721500 90-247 21:50:0
6 ah119024618373 90-246 18:37:3
7 ah11090590195456 09-05-90 19:54:56
8 ah11090590213824 09-05-90 21:38:24
9 ah11090690194357 09-06-90 19:43:57
10 ah11090690212645 09-06-90 21:26:45
11 ah11090990205223 09-09-90 20:52:23
12 ah11090790193314 09-07-90 19:33:14
13 ah11091090204118 09-10-90 20:41:18
14 ah11090790211507 09-07-90 21:15:07
15 ah11091090190040 09-10-90 19:00:40
16 ah11091190203001 09-11-90 20:30:01
17 ah11091190185002 09-11-90 18:50:02
18 ah11091290201858 09-12-90 20:18:58
19 ah11091390200741 09-13-90 20:07:41
20 ah11091390215205 09-13-90 21:52:05
15 1 ah11091490214009 09-14-90 21:40:09
2 ah11091490195640 09-14-90 19:56:40
3 ah11091590194540 09-15-90 19:45:40
4 ah11091890205405 09-18-90 20:54:05
5 ah11091890191315 09-18-90 19:13:15
6 ah11091990204258 09-19-90 20:42:58
7 ah11091790192359 09-17-90 19:23:59
8 ah11092290200915 09-22-90 20:09:15
9 ah11092290215340 09-22-90 21:53:40
10 ah11092490194711 09-24-90 19:47:11
11 ah11092490213001 09-24-90 21:30:01
12 ah11092590193628 09-25-90 19:36:28
13 ah11092590211837 09-25-90 21:18:37
14 ah11092590175904 09-25-90 17:59:04
15 ah11092690192528 09-26-90 19:25:28
16 ah11092690210658 09-26-90 21:06:58
17 ah11092790191444 09-27-90 19:14:44
16 1 ah11092890204426 09-28-90 20:44:26
2 ah11093090202157 09-30-90 20:21:57
3 ah11093090220718 09-30-90 22:07:18
4 ah11092990203303 09-29-90 20:33:03
5 ah11100190201036 10-01-90 20:10:36
6 ah11100190183142 10-01-90 18:31:42
7 ah11100290214315 10-02-90 21:43:15
8 ah11100390213133 10-03-90 21:31:33
9 ah11100390181053 10-03-90 18:10:53
10 ah11100490193757 10-04-90 19:37:57
11 ah11100490212008 10-04-90 21:20:08
12 ah11100590192658 10-05-90 19:26:58
13 ah11100590210843 10-05-90 21:08:43
14 AH11100690191559 10-06-90 19:15:59
15 ah11100990202330 10-09-90 20:23:30
16 ah11100690205705 10-06-90 20:57:05
17 ah11101090201214 10-10-90 20:12:14
18 ah11101090215640 10-10-90 21:56:40
19 ah11101190200110 10-11-90 20:01:10
20 ah11101190214457 10-11-90 21:44:57
17 1 ah11101290195008 10-12-90 19:50:08
2 ah11101290213315 10-12-90 21:33:15
3 ah11101590191631 10-15-90 19:16:31
4 ah11101590205846 10-15-90 20:58:46
5 ah11101690190620 10-16-90 19:06:20
6 ah11101690204718 10-16-90 20:47:18
7 ah11101890202449 10-18-90 20:24:49
8 AH11102290212243 10-22-90 21:22:43
9 ah11102190213431 10-21-90 21:34:31
10 AH11102390211120 10-23-90 21:11:20
11 ah11101990201328 10-19-90 20:13:28
12 ah11102090182006 10-20-90 18:20:06
13 ah11101990183010 10-19-90 18:30:10
14 ah11102490205954 10-24-90 20:59:54
15 ah11102090214610 10-20-90 21:46:10
16 ah11102190180822 10-21-90 18:08:22
17 ah11102590204829 10-25-90 20:48:29
18 ah11102590190643 10-25-90 19:06:43


#################### END DOC SECTION #########################

\header

\data