PROTOTYPE 1990 CONTERMINOUS U.S. LAND COVER CHARACTERISTICS DATA SET CD-ROM TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background. . . . . . . . . . . . . . . . . . . . . . . . . . . Classification Summary. . . . . . . . . . . . . . . . . . . . . Suggestions for Data Base Users . . . . . . . . . . . . . . . . CD-ROM DISC ORGANIZATION . . . . . . . . . . . . . . . . . . . . . . Directories and File Types. . . . . . . . . . . . . . . . . . . List of Files . . . . . . . . . . . . . . . . . . . . . . . . . Geometric Characteristics of Raster Images. . . . . . . . . . . CD-ROM FILE DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . Land Characteristics Data (LCCLASS Directory) . . . . . . . . . Source Raster Files (SOURCE Directory). . . . . . . . . . . . . Derived Raster Files (DERIVE Directory) . . . . . . . . . . . . Descriptive and Statistical Attributes (ATTRIB Directory) . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PROTOTYPE 1990 CONTERMINOUS U.S. LAND COVER CHARACTERISTICS DATA SET CD-ROM EROS Data Center U.S. Geological Survey, National Mapping Division Sioux Falls, South Dakota and Center for Advanced Land Management Information Technologies Conservation and Survey Division University of Nebraska, Lincoln INTRODUCTION Background The U.S. Geological Survey (USGS), EROS Data Center (EDC), as part of the U.S. Global Change Research Program, is developing prototype 1-km spatial data sets for global environmental research (see Loveland and Scholz, 1993, for a description of USGS global data set development activities). A significant element of the global data set development initiative involves the generation of data sets from National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer (AVHRR) satellite images. This Compact Disc-Read Only Memory (CD-ROM) contains a prototype land cover characteristics data base for the conterminous United States (lower 48 states) derived from the classification of 1990 AVHRR time-series data, with post-classification refinement based on other digital earth science data sets, including topography, climate, soils, and ecoregions. The Center for Advanced Land Management Information Technologies of the University of Nebraska at Lincoln is a major collaborator in this effort. The data on this CD-ROM consist of (1) the source data, (2) the initial and the final interpretation of land cover regions, (3) descriptive and quantitative attributes for each of the land cover regions, and (4) derived thematic data sets. It contains the complete land cover characteristics data base and is intended to provide users with the flexibility to tailor data to the unique requirements of a variety of applications. The data base provides researchers with the capability to compute new parameters, derive new classifications and aggregations of the data to suit specific needs, and to develop custom products. The data on this CD-ROM also complements the 1989-92 AVHRR CD-ROM's and the AVHRR Companion Disc produced by the USGS (see Eidenshink, 1992, for a detailed discussion of the AVHRR time-series data set program). Classification Summary The classification of land cover regions followed several steps: (1) image classification, (2) class labeling and description, and (3) post- classification refinement. Class attributes and derivative thematic maps were produced following the completion of these tasks. While a brief description of each step follows, those interested in a thorough explanation of analytic procedures should read Loveland and others, 1991, and Brown and others, 1993. Image Classification: Eight 4-week AVHRR maximum normalized difference vegetation index (NDVI) composites for March to October 1990 were initially stratified into vegetated and non-vegetated regions using a maximum NDVI image representing this 8-month period. Following this, an unsupervised classification using a clustering algorithm (ISOCLASS) and minimum-distance-to-mean classifier were used to define 70 spectral-temporal (seasonally distinct) classes within the vegetated stratum. Class Labeling and Description: Initial evaluation, labeling, and characterization of the 70 classes were based on a combination of graphic, statistical, and visual tools and techniques. For example, graphs portraying the variation of mean NDVI over the 8-month analysis period yielded a phenological profile of each class. Maps, atlases, agricultural statistics, and Landsat image maps were also used to help label the 70 classes. In addition, graphic summaries of elevation and frost-free period statistics for each class enabled association of the spatial distribution of each class with site characteristics. Tables to associate the 70 classes to ecoregions, major land resource areas (MLRA), and USGS land use and land cover (LULC) data were also created. Post-classification Refinement: In most classes, classification confusion was observed (i.e., each of the 70 classes were not uniquely associated with a single cover type). Post-classification refinement criteria were developed and used to divide about 75 percent of the original 70 preliminary vegetation greenness classes into 159 seasonally distinct land cover regions. The development of the descriptive and quantitative attributes of each region completed the final characterization of the 159 classes. Suggestions for Data Base Users The conterminous U.S. land cover characteristics data base is a prototype intended to offer flexibility in tailoring data to specific requirements for regional land cover information. Data base users may, for example, wish to aggregate the 159 classes into fewer units, grouping classes according to specific attributes. The following suggestions are provided to assist in using the data base: 1. The descriptions of vegetation types (see VEG.ASC or VEG.TXT) were derived by consulting a variety of reference materials. Because extensive site-specific ground truth was not available, it is recommended that these descriptions be used with caution. Note that the descriptors are intended to characterize, in a general fashion, the vegetation in each seasonally distinct region. All the vegetation types listed for a specific class may not necessarily occur within any given 1-km pixel. In addition, users should be aware that the order in which the vegetation types are listed does not necessarily denote their frequency of occurrence. 2. Mixtures or mosaics of land cover are often represented within each 1-km AVHRR pixel. Data base users are cautioned to consider carefully the aggregation of such classes into non-mosaic land cover types. 3. Users are urged to carefully examine the tables linking the land cover regions to descriptors abstracted from MLRA's (see MLRA.ASC or MLRA.TXT), USGS LULC data (see LULC.ASC or LULC.TXT), and ecoregions (see ECOREG.ASC or ECOREG.TXT). These are provided to augment the vegetation type descriptions and will help data base users understand the character of the regions. 4. The U.S. land cover characteristics data base is intended to complement, not replace, land cover classifications derived from source data such as Landsat. Some users may, however, wish to compare the AVHRR derived regions with other independently derived land classifications, and such a comparison can be informative. However, the following should be considered when undertaking comparative studies: a. The U.S. land cover data base was derived for the entire United States using data from 1 year and one consistent classification strategy. Other classifications may have been developed from many years of data (often much older than the 1990 AVHRR data), one seasonal observation, and/or may use quite different data analysis strategies. The concept of the "seasonally distinct" land cover region is fundamentally different than that used in most spectral classifications of remotely sensed data. b. Other classifications may differ from the U.S. data base in number of classes, class definitions, spatial resolution, occurrence of mixture and mosaic classes, and minimum mapping unit. The U.S. land cover characteristics data base is a prototype, and, therefore, a subject of ongoing research. Periodic updates of class descriptions and news on new developments may be issued as warranted. Purchasers of this CD-ROM will be notified when updates are available and will receive instructions on how to access the new information. CD-ROM DISC ORGANIZATION Directories and File Types Data on this CD-ROM are arranged so that similar data types are under the following directory names as: |--ATTRIB | |--DELIMIT | |--REPORTS |--DERIVED | |--IMAGES | |--LABELS |--LCCLASS | |--IMAGES | |--LABELS |--SOURCE | |--IMAGES | |--LABELS A README.TXT file is found at the root of each major directory that includes a brief description of the files found at that level. In addition, files with the extension .ATT contain specific file legend information for the raster files found at the root of each raster data directory (LCCLASS, SOURCE, and DERIVED). Raster image data files (with name extension .IMG) are in the IMAGES subdirectories. The binary image files are in flat headerless format with 2,889 lines (rows) and 4,587 samples (columns). The file data type (8-bit or 16-bit) varies by file. The label files (with name extension .LBL) for each raster image are included in the LABELS subdirectory using the same file name as the image file 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 NASA's Jet Propulsion Laboratory in Pasadena, California. IMDISP users can access the images on this CD-ROM by selecting the label file with the name in the LABELS subdirectory. The label file will automatically access all the header information the software requires and retrieve the image data for display on a personal computer screen. Also found with the .LBL files are selected palette (.PAL) files that can be used with IMDISP to color code categorical data. Attribute files in the ATTRIB\DELIMIT subdirectory with the extension .ASC are in an ASCII delimited format. File records are variable length and include fields separated by a comma. Character fields are enclosed in double quotes. Attribute files in the ATTRIB\REPORTS subdirectory with the extension .TXT are also ASCII files, but are formatted to be viewed on a computer screen or to be printed. List of Files Directory Data type Values Description \ README.1ST ASCII CD-ROM DOCUMENTATION \ATTRIB README.TXT ASCII ATTRIBUTE DOCUMENTATION \ATTRIB\DELIMIT VEG.ASC ASCII VEGETATION AND LAND COVER DESCRIPTIONS NDVI.ASC ASCII 1990 NDVI STATISTICS CH1.ASC ASCII AVHRR CHANNEL 1 STATISTICS CH2.ASC ASCII AVHRR CHANNEL 2 STATISTICS CH3.ASC ASCII AVHRR CHANNEL 3 STATISTICS CH4.ASC ASCII AVHRR CHANNEL 4 STATISTICS CH5.ASC ASCII AVHRR CHANNEL 5 STATISTICS ELEV.ASC ASCII ELEVATION STATISTICS LULC.ASC ASCII USGS LAND USE AND LAND COVER STATISTICS MLRA.ASC ASCII MAJOR LAND RESOURCE AREAS STATISTICS ECOREG.ASC ASCII ECOREGION STATISTICS SEASON.ASC ASCII SEASONAL PARAMETERS (ONSET, PEAK, AND LENGTH OF GREEN PERIOD) FROST.ASC ASCII FROST-FREE PERIOD STATISTICS PRECIP.ASC ASCII 30-YEAR MONTHLY AVERAGE PRECIPITATION STATISTICS TMIN.ASC ASCII 30-YEAR MONTHLY MINIMUM TEMPERATURE STATISTICS TMAX.ASC ASCII 30-YEAR MONTHLY MAXIMUM TEMPERATURE STATISTICS \ATTRIB\REPORTS VEG.TXT ASCII VEGETATION AND LAND COVER DESCRIPTIONS NDVI.TXT ASCII 1990 NDVI STATISTICS CH1.TXT ASCII AVHRR CHANNEL 1 STATISTICS CH2.TXT ASCII AVHRR CHANNEL 2 STATISTICS CH3.TXT ASCII AVHRR CHANNEL 3 STATISTICS CH4.TXT ASCII AVHRR CHANNEL 4 STATISTICS CH5.TXT ASCII AVHRR CHANNEL 5 STATISTICS ELEV.TXT ASCII ELEVATION STATISTICS LULC.TXT ASCII USGS LAND USE AND LAND COVER STATISTICS MLRA.TXT ASCII MAJOR LAND RESOURCE AREAS STATISTICS ECOREG.TXT ASCII ECOREGION STATISTICS SEASON.TXT ASCII SEASONAL PARAMETERS (ONSET, PEAK, AND LENGTH OF GREEN PERIOD) FROST.TXT ASCII FROST-FREE PERIOD STATISTICS PRECIP.TXT ASCII 30-YEAR MONTHLY AVERAGE PRECIPITATION STATISTICS TMIN.TXT ASCII 30-YEAR MONTHLY MINIMUM TEMPERATURE STATISTICS TMAX.TXT ASCII 30-YEAR MONTHLY MAXIMUM TEMPERATURE STATISTICS \DERIVED README.TXT ASCII DERIVED IMAGES DOCUMENTATION LCCUSGS.ATT ASCII LEGEND FOR LCCUSGS.IMG LCCSIB.ATT ASCII LEGEND FOR LCCSIB.IMG LCCBATS.ATT ASCII LEGEND FOR LCCBATS.IMG SEASON.ATT ASCII LEGEND FOR ONSET.IMG AND PEAK.IMG LENGTH.ATT ASCII LEGEND FOR LENGTH.IMG \DERIVED\IMAGES LCCUSGS.IMG 8-BIT 0-26 MODIFIED USGS LEVEL 2 LAND USE AND LAND COVER CLASSIFICATION LCCSIB.IMG 8-BIT 0-27 MODIFIED SIMPLE BIOSPHERE MODEL CLASSIFICATION LCCBATS.IMG 8-BIT 0-28 MODIFIED BIOSPHERE ATMOSPHERE TRANSFER SCHEME CLASSIFICATION ONSET.IMG 8-BIT 0-24 ONSET OF GREENNESS CLASSIFICATION PEAK.IMG 8-BIT 0-24 PEAK OF GREENNESS CLASSIFICATION LENGTH.IMG 8-BIT 0-255 DURATION OF GREENNESS CLASSIFICATION \DERIVED\LABELS LCCUSGS.LBL ASCII MODIFIED USGS LEVEL 2 LAND USE AND LAND COVER CLASSIFICATION IMDISP LABEL LCCUSGS.PAL ASCII MODIFIED USGS LEVEL 2 LAND USE AND LAND COVER CLASSIFICATION IMDISP PALETTE LCCSIB.LBL ASCII MODIFIED SIMPLE BIOSPHERE MODEL CLASSIFICATION IMDISP LABEL LCCSIB.PAL ASCII MODIFIED SIMPLE BIOSPHERE MODEL CLASSIFICATION IMDISP PALETTE LCCBATS.LBL ASCII MODIFIED BIOSPHERE ATMOSPHERE TRANSFER SCHEME CLASSIFICATION IMDISP LABEL LCCBATS.PAL ASCII MODIFIED BIOSPHERE ATMOSPHERE TRANSFER SCHEME CLASSIFICATION IMDISP PALETTE ONSET.LBL ASCII ONSET OF GREENNESS CLASSIFICATION IMDISP LABEL ONSET.PAL ASCII ONSET OF GREENNESS CLASSIFICATION IMDISP PALETTE PEAK.LBL ASCII PEAK OF GREENNESS CLASSIFICATION IMDISP LABEL PEAK.PAL ASCII PEAK OF GREENNESS CLASSIFICATION IMDISP PALETTE LENGTH.LBL ASCII DURATION OF GREENNESS CLASSIFICATION IMDISP LABEL \LCCLASS README.TXT ASCII LAND COVER CLASSIFICATION DOCUMENTATION LCC159.ATT ASCII LEGEND FOR LCC159.IMG \LCCLASS\IMAGES LCC71.IMG 8-BIT 0-71 PRELIMINARY LAND COVER REGIONS CLASSIFICATION LCC159.IMG 8-BIT 0-159 FINAL LAND COVER REGIONS CLASSIFICATION \LCCLASS\LABELS LCC71.LBL ASCII PRELIMINARY LAND COVER REGIONS CLASSIFICATION IMDISP LABELS LCC71.PAL ASCII PRELIMINARY LAND COVER REGIONS CLASSIFICATION IMDISP PALETTE LCC159.LBL ASCII FINAL LAND COVER REGIONS CLASSIFICATION IMDISP LABELS LCC159.PAL ASCII FINAL LAND COVER REGIONS CLASSIFICATION IMDISP PALETTE \SOURCE README.TXT ASCII SOURCE IMAGES DOCUMENTATION COUNTY.ATT ASCII STATE AND COUNTY FIPS CODES FOR STPOLY.IMG AND CTYPOLY.IMG MLRA.ATT ASCII LEGEND FOR MLRA.IMG ECOREG.ATT ASCII LEGEND FOR ECOREG.IMG LULC.ATT ASCII LEGEND FOR LULC.IMG \SOURCE\IMAGES NDVI90.IMG 8-BIT 0-200 8-CHANNEL SET OF 1990 NDVI 28-DAY COMPOSITE IMAGES NDVIMAX.IMG 8-BIT 0-200 1990 MAXIMUM GREENNESS (NDVI) COMPOSITE IMAGE MLRA.IMG 8-BIT 0-178 MAJOR LAND RESOURCE AREAS IMAGE ECOREG.IMG 8-BIT 0-76 EPA ECOREGIONS IMAGE DEM.IMG 16-BIT 0-14018 DIGITAL ELEVATION MODEL IMAGE WATER.IMG 8-BIT 0-1 WATER BODIES IMAGES LULC.IMG 8-BIT 0-91 USGS LAND USE AND LAND COVER SAMPLE IMAGE FROST.IMG 16-BIT 0-348 FROST-FREE PERIOD IMAGE STPOLY.IMG 8-BIT 0-255 STATE FIPS POLYGONS IMAGE STLINE.IMG 8-BIT 0-255 STATE BOUNDARY LINES IMAGE CTYPOLY.IMG 16-BIT 0-3112 COUNTY POLYGONS IMAGE CTYLINE.IMG 8-BIT 0-255 COUNTY BOUNDARY LINES IMAGE \SOURCE\LABELS NDVI90.LBL ASCII 8-CHANNEL SET OF 1990 NDVI 28-DAY COMPOSITE IMAGES IMDISP LABEL NDVIMAX.LBL ASCII 1990 MAXIMUM GREENNESS (NDVI) COMPOSITE IMAGE IMDISP LABEL MLRA.LBL ASCII MAJOR LAND RESOURCE AREAS IMAGE IMDISP LABELS ECOREG.LBL ASCII EPA ECOREGIONS IMAGE IMDISP LABELS DEM.LBL ASCII DIGITAL ELEVATION MODEL IMAGE IMDISP LABELS WATER.LBL ASCII WATER BODIES IMAGES IMDISP LABELS LULC.LBL ASCII USGS LAND USE AND LAND COVER SAMPLE IMAGE IMDISP LABELS LULC.PAL ASCII USGS LAND USE AND LAND COVER SAMPLE IMAGE IMDISP PALETTE FROST.LBL ASCII FROST-FREE PERIOD IMAGE IMDISP LABELS STPOLY.LBL ASCII STATE FIPS POLYGONS IMAGE IMDISP LABELS STLINE.LBL ASCII STATE BOUNDARY LINES IMAGE IMDISP LABELS CTYPOLY.LBL ASCII COUNTY POLYGONS IMAGE IMDISP LABELS CTYLINE.LBL ASCII COUNTY BOUNDARY LINES IMAGE IMDISP LABELS Geometric Characteristics of Raster Images The data dimensions of each raster file are 2,889 lines and 4,587 samples (approximately 13 megabytes for the 8-bit byte images and over 26 megabytes for integer (16 bit) images). The following is a summary of the map projection parameters used for all raster files: Projection Type: Lambert Azimuthal Equal Area (LAZEA) 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 1,000 meters For the conterminous U.S. data set: X Y Center of pixel (1,1) ( -2050000, 752000 ) Number of lines 2,889 Number of samples 4,587 LAZEA minimum bounding rectangle: In projection meters: Lower left ( -2050500, -2136500 ) Upper left ( -2050500, 752500 ) Upper right ( 2536500, 752500 ) Lower right ( 2536500, -2136500 ) In decimal degrees of longitude and latitude: Lower left ( -119.9722899, 23.5837576 ) Upper left ( -128.5300591 48.4030555 ) Upper right ( -65.3946489 46.7048989 ) Lower right ( -75.4163527 22.4793919 ) In degrees, minutes, and seconds of longitude and latitude: 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 ) CD-ROM FILE DESCRIPTIONS Land Characteristics Data (LCCLASS Directory) Preliminary Classification (LCC71.IMG, LCC71.LBL, LCC71.PAL): The classification image was derived from eight 28-day NDVI composites. The class values are in a byte range from 1 to 71, with water bodies represented by class 71 and nonvegetated land represented by class 1. Although vegetation descriptions and attributes are unavailable for this file, classes 1-70 generally are ranked from low to high total NDVI (March-November). Final Classification (LCC159.IMG, LCC159.LBL, LCC159.PAL): This is the final land cover classification and represents 159 seasonally distinct land cover regions. It was produced through the post-classification refinement of the preliminary classification using elevation, ecoregions, and frost-free period. The descriptive characteristics of these land cover regions constitute the land characteristics data base and are listed in LCC159.ATT. Source Raster Files (SOURCE Directory) AVHRR NDVI Data (NDVI90.IMG, NDVI.LBL): This file contains the eight 28-day maximum NDVI composites spanning March-October 1990 that were used to generate the preliminary land cover classification. The date range for each NDVI channel is: Channel No. Calendar Date 1 Mar 16 - Apr 12 2 Apr 13 - May 10 3 May 11 - Jun 07 4 Jun 08 - Jul 05 5 Jul 06 - Aug 02 6 Aug 03 - Aug 30 7 Aug 31 - Sep 27 8 Sep 28 - Oct 25 This data set was generated from 17 1990 biweekly maximum NDVI composites produced as part of the USGS conterminous U.S. AVHRR time-series data set development initiative (the complete 1990 biweekly data set is available from EDC on CD-ROM). This process involved the creation of a composite image in which the maximum NDVI value at each pixel location for each composite period was retained. Although the NDVI is normally in the range of -1.0 to 1.0, where negative values generally represent clouds, snow, water, and other non-vegetated surfaces, and positive values represent vegetated surfaces, NDVI values were scaled to an 8-bit 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 generally represent vegetative surfaces. 1990 Maximum NDVI (NDVIMAX.IMG, NDVIMAX.LBL): The vegetated/non- vegetation stratification used in the spectral classification was determined from this maximum NDVI image. It was developed from the March 16 through October 25, 1990, NDVI composites. In this process, the maximum NDVI value for each pixel within this 8-month period was selected. See the previous paragraph for an explanation of the NDVI scaling. Major Land Resource Areas (MLRA.IMG, MLRA.LBL): These data were digitized from a 1:7,500,000-scale map compiled by the U.S. Department of Agriculture Soil Conservation Service. Each polygon within this image represents a region of similar soil, terrain, climate, potential natural vegetation, and land use characteristics (USDA SCS, 1981). MLRA region attributes include soils, terrain, climate, potential natural vegetation, and land use. The region names for the MLRA image are found in MLRA.ATT. U.S. Environmental Protection Agency Ecoregions (ECOREG.IMG, ECOREG.LBL): This file of ecoregion data represents a unique landscape based on land surface form, geology, major soils, land use, and potential natural vegetation (Omernik, 1987; Omernik and Gallant, 1990). Ecoregion names are found in ECOREG.ATT. Digital Elevation Data (DEM.IMG, DEM.LBL): This file contains the digital elevation model (DEM) 30-arc second data used in this project. These data represent elevation above sea level (in feet). The DEM data were derived from the 30-arc second data set distributed by the National Geophysical Data Center (NGDC). Edits have been made for errors in the Great Lakes and along neatlines. The NGDC 30-arc second data were generated from Defense Mapping Agency Topographic Command 3-arc second data; sampling was done by selecting every 10th line and sample (1-percent sample), and elevations were rounded to the nearest 20 feet. The only important difference between the NGDC product and the file on this disc is that the EDC has reprojected the file using bilinear resampling to the Lambert projection with 1-km cells. Water Mask (WATER.IMG, WATER.LBL): This image is a binary mask with data values of zero for the land masses and one for the water bodies. Surface water bodies were identified based on a thresholding technique using channel 2 data from approximately 50 cloud-free AVHRR scenes. USGS Land Use and Land Cover Data (LULC.IMG, LULC.LBL, LULC.PAL): This file was assembled from 51 USGS Digital LULC files (USGS, 1986). These data, representing Anderson Level II Land Use and Land Cover, were developed by the USGS in the 1970s from visual analyses of aerial photographs (Anderson and others, 1976). The data are keyed to USGS 1:250,000-scale 1 degree by 2 degree quadrangles. Fifty-one quadrangle-based data sets were converted to a 1-km grid. The quadrangles, selected to sample major ecosystems, cover approximately 12 percent of the conterminous U.S. The file LULC.ATT contains the Level II LULC legend. Frost-Free Period (FROST.IMG, FROST.LBL): Data defining the number of frost-free days were digitized from a national climate atlas (NOAA, 1979). The information within the atlas was based on long-term records of temperature and precipitation. Digitized isoline data were subsequently interpolated to a raster image. Because of the generalized nature of the source maps, these data relate to continental climate conditions and do not represent local or microclimate conditions. Data values represent the number of frost-free days. Political Boundaries (see list below): The political boundary data sets are derived from USGS 1:2,000,000-scale digital line graph political boundaries (Domoratz, 1983). Regional digital line graph boundary files were paneled and joined with coastline and shoreline data by the USGS's Water Resources Division. These data were further processed at the EDC to correct topology and attribute coding to ensure that county representations concur with the most recent information (nominally 1983). Attribute codes conform to the Federal Information Processing Standard (FIPS) State and county numerical codes. Files included with this data set are: File Format Description STPOLY.IMG 8-Bit State FIPS polygons STLINE.IMG 8-Bit State boundary lines CTYPOLY.IMG 16-Bit County polygons CTYLINE.IMG 8-Bit County boundary lines COUNTY.ATT ASCII County five-digit codes Within CTYLINE.IMG, the county boundaries coincident with coasts and international borders are value 253, county boundaries that are defined by State borders are value 254, and all other county boundaries are value 255. Attributes for the county polygons are in COUNTY.ATT and the data fields are: cntyid the unique five-digit polygon identification where the first two-digit number is the State code and the last three-digit number is the county code npixels the number of pixels in each county FIPS the FIPS State and county code cname county name sname State name Derived Raster Files (DERIVED Directory) Six images were derived from the U.S. land cover characteristics data base by remapping the 159 data base classes into other themes. USGS Land Use and Land Cover (LCCUSGS.IMG, LCCUSGS.LBL, LCCUSGS.PAL): Twenty-six categories based on the USGS LULC system were defined to illustrate national patterns of land cover (Anderson and others, 1976). The classes include Level 2 modifications that reflect regional vegetation types and mosaics of land cover that occur at 1-km resolution. See LCCUSGS.ATT for the file legend. Simple Biosphere Model (LCCSIB.IMG, LCCSIB.LBL, LCCSIB.PAL): The Simple Biosphere Model (SiB), developed primarily by Piers Sellers at the University of Maryland, models the interactions between the atmosphere and the terrestrial biosphere, specifically, fluxes of radiation, momentum, sensible heat, and latent heat (Dorman and Sellers, 1989). The major SiB biomes were derived mainly from potential natural vegetation (Kuchler, 1983) and cultivation data (Matthews, 1983). Additional classes (21-27) were added to the original SiB scheme (1-20) to represent mosaics that occur within 1-km resolution data. See LCCSIB.ATT for the file legend. Biosphere-Atmosphere Transfer Scheme (LCCBATS.IMG, LCCBATS.LBL, LCCBATS.PAL): The Biosphere-Atmosphere Transfer Scheme (BATS) is a comprehensive model of land-surface processes suitable for use with various global circulation models. It was developed at the National Center for Atmospheric Research starting in the late 1970's (see Dickinson and others, 1986). This model includes a calculation of surface temperature in response to surface heating, heat capacity, and thermal conductivity, and the surface albedo in terms of soil moisture and vegetative cover, fractional vegetation cover, and various other heat and moisture flux parameterizations. The BATS vegetation/land-cover scheme originally was derived from vegetation and cultivation (Matthews, 1983, 1984) and land use and soils data (Wilson and Henderson-Sellers, 1985). As with SiB, additional classes (20-28) added to BATS to represent mosaics that occur within 1-km resolution data. See LCCBATS.ATT for the file legend. Onset of Greenness (ONSET.IMG, ONSET.LBL, ONSET.PAL): This image represents the 1990 biweekly period in which significant development of standing green biomass started. The onset, or start of the growing season, was estimated from 1990 biweekly NDVI composites. Onset is defined as the period in which there is a significant increase of greenness or NDVI values from a seasonal low level. Onset means different things in different landscapes. In deciduous forests, onset is when leaves develop and there is a major increase in NDVI values. In Rocky Mountain coniferous forests, onset occurs when snow cover retreats and the forest understory develops. The definition of onset is somewhat subjective, and users should determine whether this classification is suitable for their application. The legend for onset is found in SEASON.ATT. Peak of Greenness (PEAK.IMG, PEAK.LBL, PEAK.PAL): This image displays the biweekly period in which the highest level of greenness (NDVI) was recorded in 1990. It was developed by identifying the time of the maximum NDVI mean value for each of the 159 classes. This file uses a similar classification scheme as was used for the onset of greenness (see SEASON.ATT). However, class 22, green all year, is not relevant and is not found in PEAK.IMG. Duration of Greenness (LENGTH.IMG, LENGTH.LBL): This image represents the duration, or length, of greenness for each class in 1990. It was produced by determining the biweekly period in which NDVI dropped to a seasonal low and subtracting this period, or date, from the date of the period of onset of greenness. The image values represent duration of greenness in days (see LENGTH.ATT). The determination of the period in which greenness ends is as subjective as the definition of the period of onset. Again, users should determine whether this classification of 1990 duration of greenness is suitable for their application. Descriptive and Statistical Attributes (ATTRIB Directory) The attribute files contain either descriptive statistics, frequency summaries, or descriptions for each of the 159 classes found in the final land cover characteristics image file (LCC159.IMG). In the attribute files, the class numbers relate to a category in LCC159.IMG, and the number of pixels represent the sample size for the corresponding statistics or frequencies. There are two types of attribute files found on the CD-ROM. The files with a .ASC are in a comma-delimited ASCII format. These are provided to facilitate importing the data into data base management systems or spreadsheet formats (for example, DBASE IV, Paradox, Lotus 1-2-3, or Quattro Pro). The records in each file are in a standard ASCII delimited format with character fields enclosed in double quotes (") and column entries separated with commas (,). The second type of file, with a .TXT extension, are ASCII text reports that contain organized summaries of each attribute set. These can be printed or listed to see the attribute values of specific classes. It is suggested that .TXT files be viewed on a computer monitor prior to printing to determine whether they should be printed in portrait or landscape mode. For each of the following summaries, field names, value types, and suggested field widths for each variable in each attribute file are provided. Major Vegetation and Land Cover Types (VEG.ASC, VEG.TXT): The descriptions of the major vegetation and land cover types for each of the 159 classes were developed through an analysis of a wide range of sources. The descriptions represent major or dominant vegetation and land cover components of each class and are not meant to be comprehensive of all species found in a region. In addition, the order of vegetation types listed does not imply an order of dominance. The fields, value types, and suggested widths are: Class number Integer 3 General cover type Character 30 Modified Anderson Level 2 code Integer 3 Major vegetation types Character 70 The modified Anderson Level 2 code and general cover type descriptions are based on Anderson and others, 1976. The modifications were developed by the project team to represent the nature of national patterns of land cover that can be consistently identified with AVHRR data. Normalized Difference Vegetation Index - 1990 (NDVI.ASC, NDVI.TXT): These statistics were calculated from 21 biweekly AVHRR composites (January 5, 1990 - December 20, 1990). See Eidenshink, 1992, for a description of the original 1990 AVHRR data set. The statistics represent the NDVI characteristics per class. The units of measure are scaled NDVI values with values of 0, 100, and 200 translating to original NDVI values of -1, 0, and +1 (note that negative NDVI corresponds to non-vegetated surfaces and positive values relate to vegetated surfaces). Each digital count corresponds to 0.01 NDVI. The statistics indicate the temporal variations of vegetation conditions for each class. The mean, mode, median, and standard deviation values are more representative of typical conditions than the other statistics. The minimum and maximum values are influenced by anomalous outlier values and should be used with caution. The fields, value types, and suggested widths are: Class number Integer 3 Composite period Integer 2 Start date of composite Date 8 End date of composite Date 8 Number of pixels in class Integer 8 Mean Real 8 Standard deviation Real 8 Maximum value Integer 3 Minimum value Integer 3 Mode Integer 3 Median Integer 3 AVHRR Channels 1-5 - 1990 (see text for file names): The spectral (channels 1 and 2) and thermal infrared (channels 3-5) statistics found in the files named CH1.ASC, CH1.TXT; CH2.ASC, CH2.TXT; CH3.ASC, CH3TXT; CH4.ASC, CH4TXT; and CH5.ASC, CH5.TXT were calculated from 21 January 5-December 20, 1990, biweekly AVHRR composites. The statistics represent the per class AVHRR channel characteristics. The units of measure are digital counts with a potential range of 0-255. The statistics indicate the temporal variations of spectral/thermal observations of each class. The mean, median, mode, and standard deviation values are more representative of typical conditions than the other statistics. The minimum and maximum values often represent anomalous outlier values and should be used with caution. Also note that the January-June thermal data were calibrated using coefficients different than the July-December data, so direct comparisons of the yearly channels 4 and 5 values may be difficult. The fields, value types, and suggested widths are: Class number Integer 3 Composite period Integer 2 Start date of composite Date 8 End date of composite Date 8 Number of pixels in class Integer 8 Mean Real 8 Standard deviation Real 8 Maximum value Integer 3 Minimum value Integer 3 Mode Integer 3 Median Integer 3 Elevation (ELEV.ASC, ELEV.TXT): These statistics were calculated from a DEM for the conterminous U.S. (see previous discussion of DEM data in source data section). The statistics represent the per class elevation characteristics and indicate the variations of elevation conditions for each class. The mean, median, and standard deviation values are more representative of typical conditions than the other statistics. The mode, minimum, and maximum values often represent anomalous outlier values and should be used with caution. The fields, value types, and suggested widths are: Class number Integer 3 Number of pixels in class Integer 8 Mean Real 10 Standard deviation Real 10 Maximum value Integer 5 Minimum value Integer 5 Mode Integer 5 Median Integer 5 Land Use and Land Cover (LULC.ASC, LULC.TXT): This file contains the frequency counts and percentages that describe the general USGS LULC types that compose each class. The frequencies were calculated from a sample of 51 USGS 1:250,000-scale LULC quadrangle-based digital files (USGS, 1986). The frequencies are for illustrative purposes and are not based on a statistically valid sample. The frequencies provide an approximate example of the general LULC types in each class. It suggests the make-up of each class mosaic. For a complete description of the USGS LULC Classification Scheme, see Anderson and others, 1976. The fields, value types, and suggested widths are: Class number Integer 3 Number of pixels in class Integer 8 Number of pixels in sample Integer 8 Percent of class Real 8 Anderson Level 2 code Integer 2 Anderson Level 1 category Character 50 Anderson Level 2 category Character 50 Major Land Resource Areas (MLRA.ASC, MLRA.TXT): This file contains frequency counts and percentages for each MLRA in each land cover characteristics class. In addition, landscape variable descriptions of each MLRA provide information about the site characteristics of each class. For a complete description of the MLRA's, see USDA SCS, 1981. The fields, value types, and suggested widths are: Class number Integer 3 Number of pixels in class Integer 8 Number of pixels in sample Integer 8 Percent of class Real 8 MLRA map code Integer 3 MLRA name Character 60 Land use practices Character 60 Elevation range Character 60 Topography and landforms Character 60 Average annual precipitation Character 60 Average annual temperature Character 60 Average annual frost-free period Character 60 Soil characteristics Character 60 Potential natural vegetation Character 60 Ecoregions (ECOREG.ASC, ECOREG.TXT): This file contains frequency counts and percentages for each U.S. Environmental Protection Agency ecoregion in each land cover characteristics class. In addition, landscape variable descriptions of each ecoregion provide additional information about the composition of each class. For a complete description of the U.S. Environmental Protection Agency ecoregions, see Omernik, 1987. The fields, value types, and suggested widths are: Class number Integer 3 Number of pixels in class Integer 8 Number of pixels in sample Integer 8 Percent of class Real 8 Ecoregion map code Integer 3 Ecoregion name Character 50 Landform Character 50 Potential natural vegetation Character 50 Land use Character 50 Soils Character 50 Seasonal Characteristics (SEASON.ASC, SEASON.TXT): This file identifies the onset and end of greenness, the period of peak greenness, and the duration of greenness (in days). In addition, for classes with a second green season, the onset, peak, and end of the minor green period are also given. See the derived data section for a discussion of the characteristics of the onset, peak, and duration files. The fields, value types, and suggested widths are: Class number Integer 3 Period of onset Integer 2 Period of peak greenness Integer 2 Period greenness ends Integer 2 Date of onset Date (8 characters) Date of peak greenness Date (8 characters) Date greenness ends Date (8 characters) Length of green season Integer 3 Period of minor (2nd) onset Integer 2 Period of minor (2nd) peak Integer 2 Period of minor (2nd) ends Integer 2 Date of minor (2nd) onset Date (8 characters) Date of minor (2nd) peak Date (8 characters) Date of minor (2nd) end Date (8 characters) Frost-Free Period (FROST.ASC, FROST.TXT): This file represents the frost-free period conditions for the 159 classes. See the source data section for a description of the original raster data from which the statistical summary is based. The data values are days. The fields, value types, and suggested widths are: Class number Integer 3 Month Integer 2 Number of pixels in class Integer 8 Mean Real 8 Standard deviation Real 8 Maximum value Integer 3 Minimum value Integer 3 Mode Integer 3 Median Integer 3 Climate (see following text for file names): There are three attribute files that summarize per class climate (precipitation and temperature) conditions. PRECIP.ASC and PRECIP.TXT lists class monthly precipitation statistics; TMIN.ASC and TMIN.TXT contains class monthly minimum temperatures; and TMAX.ASC and TMAX.TXT contains class monthly maximum temperatures. Precipitation values are expressed in inches, and temperatures are in degrees Fahrenheit. All are based on 30-year averages. The fields, value types, and suggested widths are: Class number Integer 3 Month Integer 2 Number of pixels in class Integer 8 Mean Real 8 Standard deviation Real 8 Maximum value Integer 3 Minimum value Integer 3 Mode Integer 3 Median Integer 3 The climatic summaries for the land cover classes were derived from Earth-Info's Climate Data Summary of the Day CD-ROM series (EarthInfo, Inc., 1990). The latitude and longitude of stations and the following parameters were extracted from the CD-ROM: monthly average minimum temperature, monthly average maximum temperature, and average daily precipitation for each month. The average daily precipitation parameter was converted to average monthly precipitation. The station data were converted from x, y, and z point data to a raster image format and georeferenced to the Lambert Azimuthal Equal Area map projection to be spatially compatible with the other elements of the land cover characteristics data base. Each station is represented by one cell (1-km resolution) containing a data value corresponding to the climatic parameter. The parameter value for each weather station was assigned to all pixels within a 10-km radius of the station, and statistical summaries were calculated for each cover type. In instances where adjacent 10-km radii overlapped, the greater value was assigned to the pixel. The 10-km radius around each station, in general, is an appropriate "sphere of influence" to describe climatic conditions. Although the 10-km radius size presents some potential problems, it permits gathering a large sample of climatic data for each of the 159 land cover classes. The mean, mode, and standard deviation values are more representative of typical conditions than the other statistics. The minimum and maximum values often represent anomalous outlier values and should be used with caution. There are several factors to consider when interpreting these data, each of which could result in misleading climatic parameters assigned to some of the land cover classes: 1. The climatic summaries may be inaccurate in areas of high relief because a weather station may be significantly different in elevation than pixels that are 10-km distant. 2. Many weather stations are located in urban areas and may experience urban heat island effects. Also, the land cover descriptions tend to be less accurate in urban areas. 3. In areas where convective activity is the primary source of precipitation, weather events may be significantly different within a 10-km radius, even over a 30-year period. 4. The sample of weather stations in this data set does not include all areas of the country - especially higher elevations and remote regions. Therefore, the data must be used with caution in cases where there are large disparities between the spatial distributions of the locations of climatic and land cover data. REFERENCES Anderson, J.R., Hardy, E.E., Roach, J.T., and Witmer, R.E., 1976, A land use and land cover classification system for use with remote sensor data: U.S. Geological Survey Professional Paper 964, 28 p. Brown, J.F., Loveland, T.R., Merchant, J.W., Reed, B.C., and Ohlen, D.O., 1993, Using multisource data in global land cover characterization - concepts, requirements, and methods: Photogrammetric Engineering and Remote Sensing, v. 59, no. 6, p. 977-987. Dickinson, R.E., Henderson-Sellers, A., Kennedy, P.J., and Wilson, M.F., 1986, Biosphere-Atmosphere Transfer Scheme (BATS) for the NCAR community climate model: NCAR Technical Note, NCAR/TN-275+STR, National Center for Atmospheric Research, Boulder, Colorado, 69 p. Domoratz, M.A., 1983, USGS digital cartographic data standards - digital line graphs from 1:2,000,000-scale maps: U.S. Geological Survey Circular 895-D. Dorman, J.L. and Sellers, P.J., 1989, A global climatology of albedo, roughness length and stomatal resistance for atmospheric general circulation models as represented by the Simple Biosphere Model (SiB): Journal of Applied Meteorology, v. 28, p. 833-855. EarthInfo, Inc., 1990, Climate data user's manual: TD 3200 Summary of the Day Cooperative Observer Network, Boulder, Colorado. Eidenshink, J.E., 1992, The conterminous United States AVHRR data set: Photogrammetric Engineering and Remote Sensing, v. 58, no. 6, p. 809-813. Kuchler, A.W., 1983, World map of natural vegetation: Goode's World Atlas (16th edition), Rand McNally, p. 16-17. Loveland, T.R., Merchant, J., Ohlen, D.O., and Brown, J., 1991, Development of a land cover characteristics data base for the conterminous United States: Photogrammetric Engineering and Remote Sensing, v. 57, no. 11, p. 1453-1463. Loveland, T.R., and Scholz, D.K., 1993, Global data set development and data distribution activities at the U.S. Geological Survey's EROS Data Center: American Congress on Surveying and Mapping, American Society for Photogrammetry and Remote Sensing Annual Convention and Exposition Technical Papers, New Orleans, Louisiana, February 1993, Proceedings, Bethesda, Maryland, v. 2, p. 204-211. Matthews, E., 1983, Global vegetation and land use - new high resolution data bases for climate studies: Journal of Climate and Applied Meteorology, v. 22, p. 474-487. Matthews, E., 1984, Prescription of land-surface boundary conditions in GISS GCMII and vegetation, land-use and seasonal albedo data sets - documentation of archived data tape: NASA Technical Memoranda 86096 and 86107, National Aeronautics and Space Administration, Goddard Institute for Space Studies, 20 p. and 9 p. National Oceanic and Atmospheric Service (NOAA), 1979, Climate atlas of the United States: Environmental Data Services, Asheville, North Carolina, 80 p. Omernik, J.M., 1987, Ecoregions of the conterminous United States: Annals of the Association of American Geographers, v. 77, no. 1, p. 118-125. Omernik, J.M. and Gallant, A.L., 1990, Defining regions for evaluating environmental resources: Global Natural Resources Monitoring and Assessment Symposium, Bethesda, Maryland, Proceedings, p. 936-947. U.S. Department of Agriculture, Soil Conservation Service (USDA SCS), 1981, Land resource regions and major land resource areas of the United States: Agriculture Handbook 296. U.S. Geological Survey (USGS), 1986, Land use and land cover data from 1:250,000- and 1:100,000-scale maps: U.S. Geological Survey Data Users Guide No. 4. Wilson, M.F., and Henderson-Sellers, A., 1985, A global archive of land cover and soils data for use in general circulation climate models: Journal of Climatology, v. 5, p. 119-143.