ORNL/CDIAC-119 NDP-068 Geographical Distribution of Biomass Carbon in Tropical Southeast Asian Forests: A Database Contributed by Sandra Brown (1), Louis R. Iverson (2), and Anantha Prasad (2) Department of Natural Resources and Environmental Sciences University of Illinois Urbana, Illinois and Illinois Natural History Survey Champaign, Illinois (1) Present address: Winrock International Arlington, Virginia (2) Present address: United States Forest Service Northeast Research Station Delaware, Ohio Prepared by Tammy W. Beaty, Lisa M. Olsen, Robert M. Cushman, and Antoinette L. Brenkert Environmental Sciences Division Environmental Sciences Division Publication No. 4879 Date Published: March 2001 Prepared for the Environmental Sciences Division Office of Biological and Environmental Research Budget Activity Number KP 12 04 01 0 Prepared by the Carbon Dioxide Information Analysis Center Environmental Sciences Division OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831-6335 managed by UNIVERSITY OF TENNESSEE-BATTELLE, LLC for the U.S. DEPARTMENT OF ENERGY under contract DE-AC05-00OR22725 CONTENTS LIST OF TABLES ABSTRACT 1. BACKGROUND INFORMATION 2. APPLICATIONS OF THE DATA 3. DATA LIMITATIONS AND RESTRICTIONS 4. QUALITY-ASSURANCE CHECKS AND DATA-PROCESSING ACTIVITIES PERFORMED BY CDIAC 5. REFERENCES 6. HOW TO OBTAIN THE DATA AND DOCUMENTATION 7. LISTING OF FILES PROVIDED 8. DESCRIPTION OF THE DOCUMENTATION FILE 9. DESCRIPTION, FORMAT, AND PARTIAL LISTINGS OF THE ARC/INFO GRID FILES 10. DESCRIPTION, FORMAT, AND PARTIAL LISTINGS OF THE 24 ASCII DATA FILES PRODUCED BY THE ARC/INFO GRIDASCII COMMAND 11. DESCRIPTION, FORMAT, AND PARTIAL LISTING OF THE COMPOSITE 3.75-KM AND 0.25-DEGREE ASCII DATA FILES 12. STATISTICS OF THE FILES PROVIDED IN THIS NUMERIC DATA PACKAGE LIST OF TABLES 1 Redistribution of the data as a result of the resampling process 2 GRIDASCII syntax used to produce the ASCII data files 3 Files in this numeric data package 4 Item descriptions for the ten ARC/INFO export grids 5 Format and description of variables for the composite ASCII data files in this numeric data package (se_asia.dat and se_asiax.dat) 6 Item statistics for the data files in this numeric data package ABSTRACT BROWN, S., L. R. IVERSON, AND A. PRASAD. 2001. Geographical Distribution of Biomass Carbon in Tropical Southeast Asian Forests: A Database. ORNL/CDIAC-119, NDP-068. Carbon Dioxide Information Analysis Center, U.S. Department of Energy, Oak Ridge National Laboratory, Oak Ridge, Tennessee, U.S.A. doi: 10.3334/CDIAC/lue.ndp068 A database was generated of estimates of geographically referenced carbon densities of forest vegetation in tropical Southeast Asia for 1980. A geographic information system (GIS) was used to incorporate spatial databases of climatic, edaphic, and geomorphological indices and vegetation to estimate potential (i.e., in the absence of human intervention and natural disturbance) carbon densities of forests. The resulting map was then modified to estimate actual 1980 carbon density as a function of population density and climatic zone. The database covers the following 13 countries: Bangladesh, Brunei, Cambodia (Campuchea), India, Indonesia, Laos, Malaysia, Myanmar (Burma), Nepal, the Philippines, Sri Lanka, Thailand, and Vietnam. The data sets within this database are provided in three file formats: ARC/INFO (TM) exported integer grids, ASCII (American Standard Code for Information Interchange) files formatted for raster-based GIS software packages, and generic ASCII files with x, y coordinates for use with non-GIS software packages. This database includes ten ARC/INFO exported integer grid files (five with the pixel size 3.75 km x 3.75 km and five with the pixel size 0.25 degree longitude x 0.25 degree latitude) and 27 ASCII files. The first ASCII file contains the documentation associated with this database. Twenty-four of the ASCII files were generated by means of the ARC/INFO GRIDASCII command and can be used by most raster-based GIS software packages. The 24 files can be subdivided into two groups of 12 files each. These files contain real data values representing actual carbon and potential carbon density in Mg C/ha (1 megagram = 10^6 grams) and integer- coded values for country name, Weck's Climatic Index, ecofloristic zone, elevation, forest or non- forest designation, population density, mean annual precipitation, slope, soil texture, and vegetation classification. One set of 12 files contains these data at a spatial resolution of 3.75 km, whereas the other set of 12 files has a spatial resolution of 0.25 degree. The remaining two ASCII data files combine all of the data from the 24 ASCII data files into 2 single generic data files. The first file has a spatial resolution of 3.75 km, and the second has a resolution of 0.25 degree. Both files also provide a grid-cell identification number and the longitude and latitude of the centerpoint of each grid cell. The 3.75-km data in this numeric data package yield an actual total carbon estimate of 42.1 Pg (1 petagram = 10^15 grams) and a potential carbon estimate of 73.6 Pg; whereas the 0.25-degree data produced an actual total carbon estimate of 41.8 Pg and a total potential carbon estimate of 73.9 Pg. Fortran and SAS (TM) access codes are provided to read the ASCII data files, and ARC/INFO and ARCVIEW command syntax are provided to import the ARC/INFO exported integer grid files. The data files and this documentation are available without charge on a variety of media and via the Internet from the Carbon Dioxide Information Analysis Center (CDIAC). Keywords: biomass, carbon, carbon cycle, climate, elevation, forest, land use, organic matter, population, slope, soil, Southeast Asia, tropics, vegetation 1. BACKGROUND INFORMATION Quantification of the role of changing land use in the global cycling of carbon (and, consequently, in controlling atmospheric concentrations of carbon dioxide, the single most important anthropogenic greenhouse gas) requires complete, consistent, and accurate databases of vegetation, land use, and biospheric carbon content. The Carbon Dioxide Information Analysis Center (CDIAC) has previously made available several important quality-assured and documented databases on this topic (Olson et al. 1985, Richards and Flint 1994, Houghton and Hackler 1995, and Brown et al. 1996). This database (NDP-068) expands the series by providing detailed geographically referenced information on actual and potential biomass carbon (1 g biomass = 0.5 g C) in tropical Southeast Asia and all the background information used to generate those files. A geographic information system (GIS) was used to incorporate spatial databases of climatic, edaphic, and geomorphological indices and vegetation to estimate potential (without human influence) carbon densities of forests in 1980. The resulting estimates were then modified to produce estimates of actual carbon density as a function of population density and climatic zone. Estimates of carbon in the biomass (aboveground and belowground) of tropical Southeast Asian forests for the year 1980 were generated by means of a GIS modeling approach, on the basis of the assumption that "the present distribution of forest biomass density is a function of the potential biomass the landscape can support under the prevailing climatic, edaphic and geomorphological conditions and the cumulative impact of human activities such as logging, fuel- wood collection, shifting cultivation, and other activities that reduce the biomass" (Brown et al. 1993). The database covers the following 13 countries: Bangladesh, Brunei, Cambodia (Campuchea), India, Indonesia, Laos, Malaysia [Peninsular (Malaya) and Insular (Sabah, also known as North Borneo, and Sarawak)], Myanmar (Burma), Nepal, the Philippines, Sri Lanka, Thailand, and Vietnam. A thorough description of the methods and data sources can be found in Brown et al. (1993). To calculate potential and actual aboveground biomass carbon densities, the general methodology of Risser and Iverson (1988) and Iverson et al. (1994) was followed. This consisted of a simple weighted additive model of data layers of elevation and slope, precipitation, Weck's Climatic Index, and soil texture to arrive at a score for potential biomass density for each pixel. Elevation data were derived from a U.S. National Geophysical Data Center elevation map; soil texture data and slope data were derived from the Soil Map of the World produced by the Food and Agriculture Organization (FAO) United Nations Educational, Scientific, and Cultural Organization; and annual precipitation and a modified Weck's Climatic Index (Weck 1970) were interpolated from about 600 stations in the FAO agro- meteorological database. Results were compared with independent ground-truth information and iteratively reprocessed to within certain bounds to obtain a satisfactory result. The map results were overlaid with forest/non-forest data from circa 1980, resulting in a map of potential carbon densities. The forest/non-forest data were derived from a FAO vegetation map of continental tropical Southeast Asia and a World Conservation Monitoring Center map of forested areas of insular Asia. The resulting potential biomass was compared with ecofloristic zones derived from an FAO map, confirming the reasonableness of the model-derived estimates. Ratios of forest degradation (from increasing population) were calculated from forest inventory data and the calculated potential biomass densities for 47 subnational units in Bangladesh, India, Malaysia (Peninsular and Insular), the Philippines, Sri Lanka, Thailand, and Vietnam. Linear regression of the forest degradation ratio versus population density (natural-log transformed) showed the effect of population density on the forest degradation ratio to be greatest in dry, followed by seasonal, then moist, forests. The regression equations were then used in conjunction with the potential biomass carbon density, population , and precipitation maps [used to delineate climatic zones: aseasonal moist (>2000 mm/year), seasonally moist (1500 to 2000 mm/year), and dry (<1500 mm/year)] to estimate the actual biomass carbon densities of the forests. At very high and very low population densities, default degradation ratios of 0.06 and 1.0, respectively, were used. Population density was based on data from the FAO Demographic and Statistics Department. Root:shoot ratios were calculated from previously published data of belowground biomass and stratified according to climate zones based on precipitation and elevation. Three climate zones were recognized: dry (<1200 mm/year for lowland), seasonal (1200 to 2000 mm/year for lowland and 500 to 1200 mm/year for montane), and moist (>2000 mm/year for lowland and >1200 mm/year for montane), where lowland is defined as elevation <=1000 m and montane as elevation >1000 m. Moist forests were assigned a root:shoot ratio of 0.18; seasonal forests, 0.10; and dry forests, 0.5. These ratios were used to calculate belowground biomass from the aboveground biomass estimate for each pixel. Total biomass was calculated as the sum of the below-ground and above-ground estimates. Brown et al. (1993) compared their estimates of biomass carbon density with those of other recent assessments for the same 13-country study area. They found that estimates of biomass carbon densities derived from the FAO Tropical Forest Resource Assessment 1990 Project were about 75% of their own, and that estimates of 1980 biomass carbon density of Flint and Richards (1994) for forests and woodlands were about 65% of their own. Although differences exist between the estimates of Brown et al. and the other two studies, the three sets of values are similar in order of magnitude despite differences in methodology, input data, and time of assessment. The general similarity of the estimates provides compelling evidence that forests of tropical Asian countries have generally low biomass carbon densities; these low densities are most likely due to the long history of human use in the region. 2. APPLICATIONS OF THE DATA The maps generated from the these data lend themselves to comparisons with, for example, spatial representations of land-use changes determined from satellite imagery. Consequently, uncertainties associated with carbon fluxes from tropical Southeast Asia can be reduced, and processes in the global carbon cycle (e.g., forest clearing, degradation, and regrowth) can be better quantified. 3. DATA LIMITATIONS AND RESTRICTIONS The biomass estimates are limited to trees with a diameter of at least 10 cm (5 cm in more open forests); this would result in a slight underestimate (less than 5%) of aboveground biomass in closed forests and an unknown amount of underestimate in open forests; the estimates also exclude litter (Brown et al. 1993). Brown et al. (1993) compared their model-derived biomass carbon density estimates with values from forest inventories. They report that their model tended to produce slight overestimates: <5% for carbon densities of <250 Mg/ha and <=8% for carbon densities of 250-400 Mg/ha. The estimates provided in this numeric data package also exclude soil carbon, although Brown et al. (1993) describe the estimation of soil carbon for tropical Southeast Asia. Brown et al. (1993) evaluated the errors in estimates of carbon densities from both methodology and data limitations. In general, they caution that, while general patterns would be reliable, carbon densities cannot be precisely located to the level of an individual pixel. The original vegetation and soil maps showed insufficient detail and might not always have been fully accurate. The precipitation, Weck's Climatic Index, and population maps were generated from point data, although interpolation error from these types of data was minimized by using a two-dimensional interpolation method and by comparing results with other maps. Potential error in weighting schemes was minimized by developing varying-width classes for each of the input variables. Omitting the effects of roads, shifting cultivation, and the differentiation between broadleaf and conifer species was considered acceptable, given the scale of the final maps. Correlation between population density and the calculated forest degradation index was low for some regions. New information on forest inventories can alleviate these uncertainties. It must also be noted that understory and fine and coarse litter were not included in the total carbon estimates; correction for this omission could add another 20 to 30% to the estimates of total biomass carbon density. Explicit accounting of large-scale disturbances was also not included. As new data become available, the same basic methodology for calculating carbon densities in biomass and soils can be readily applied, differences analyzed, and uncertainties further reduced (Brown et al. 1993, Iverson et al.1994). The gridded database described in this numeric data package defines tropical Southeast Asia as originating at -5879340.56205 m longitude (44.25875 degrees), -1221655.95152 m latitude (-16.52954 degrees) and extending to 3863159.43795 m longitude (149.50875 degrees), 4808344.04848 m latitude (42.97046 degrees). Data are provided for the following 13 countries: Bangladesh, Brunei, Cambodia (Campuchea), India, Indonesia, Laos, Malaysia [Peninsular (Malaya) and Insular (Sabah, also known as North Borneo, and Sarawak)], Myanmar (Burma), Nepal, the Philippines, Sri Lanka, Thailand, and Vietnam. The country boundary information was originally received from the contributors in vector format as continental and insular polygon coverages. The source polygons contained boundaries for fifteen countries in Asia, including the 13 aforementioned countries, as well as Pakistan and Papua New Guinea. For distribution purposes, the polygon data were joined together into single polygon coverage and then converted to a grid. As a result of this rasterization process, a few grid cells are defined as Pakistan or Papua New Guinea although they contain no real data. Furthermore, the gridded country boundaries within this database should not be used to define countries for other datasets, because the rasterization process produces generalized boundary lines. 4. QUALITY-ASSURANCE CHECKS AND DATA-PROCESSING ACTIVITIES PERFORMED BY CDIAC An important part of the data packaging process at CDIAC involves the quality assurance (QA) of data before distribution. To guarantee data of the highest possible quality, CDIAC performs extensive QA checks, examining the data for completeness, reasonableness, and accuracy. The data as obtained from the contributors consisted of 17 ARC/INFO-exported integer grids with a pixel size of approximately 3.75 km x 3.75 km. Actual and potential carbon densities (Mg C/ha), as well as ecofloristic zone and vegetation classification, were provided individually for continental and insular Southeast Asia. Separate country boundaries were provided for insular and continental Southeast Asia. These ten grids were transformed into an Albers Projection, but with a unique set of projection parameters for continental and insular Southeast Asia. Population density, mean annual precipitation, elevation, slope, soil texture, forest/non-forest designation, and Weck's Climatic Index data were assembled collectively for all of Southeast Asia. These seven grids were not projected (i.e., they can be referred to as being in a "geographic projection"). For distribution purposes the continental and insular data were combined into common grids. The following methodology was used: 1. Each of the 17 grids originally received from the contributors was re-projected into an Albers Projection with a cell size of 3750 m by using the following parameters: 1st standard parallel: 30 08 24.000 2nd standard parallel: -4 17 24.000 central meridian: 107 28 12.000 latitude of projection's origin: 0 0 0.000 false easting (meters): 0.00000 false northing (meters): 0.00000 2. Each of the 17 newly projected grids was assigned a missing-value indicator of -9999. 3. The value attribute tables for the continental and insular grids were reviewed for consistency and redundancy. Numeric data values were re-assigned as necessary. 4. The population grid was designated as a base grid because it included the combined spatial extent of real data contained in each of the 17 grids. 5. The continental grids for actual carbon density, potential carbon density, ecofloristic zone, vegetation code, and country name were combined with their corresponding insular grids and the base grid by using the ARC/INFO GRID command COMBINE. 6. The seven remaining grids were reformatted to match the extent of the base grid. This was accomplished by using the ARC/INFO GRID command COMBINE with the base grid. Performing these six steps resulted in 12 grids with identical parameters. The 12 grids became the core data layers used to prepare the 37 data files included in this numeric data package. The first file is merely a flat ASCII text file containing a copy of this documentation. Ten of the 37 files are exported ARC/INFO integer grids, five with 3.75-km pixel size and five with 0.25-degree pixel size. The 3.75-km exported ARC/INFO grids included within this numeric data package were generated by using the ARC/INFO GRID command COMBINE and were grouped as follows: 1. Actual and potential carbon were combined into a common grid called BIOMASS. 2. Mean annual precipitation and Weck's Climatic Index were combined into a common grid called CLIMATE. 3. Population and country were combined into a common demographic grid called DEMOG. 4. Slope, soil texture, and elevation were grouped into a common landform grid called LAND. 5. Forest designation, ecofloristic zone, and vegetation index were grouped into a common vegetation grid called VEGT. Resampling is the process of determining values for grid cells that are geometrically transformed from a source grid into a grid of a different spatial resolution. ARC/INFO GRID offers three resampling techniques: nearest neighbor assignment, bilinear interpolation, and cubic convolution.The nearest neighbor assignment process identifies the input grid cell closest to the output grid-cell center and assigns this value to the entire output grid cell. The bilinear interpolation method of resampling identifies the four nearest input cell centers surrounding the output grid-cell center, then calculates a weighted mean of those values, and assigns the mean to the output grid-cell center. Cubic convolution is a computationally intensive interpolation method that fits a cubic polynomial surface to a 4 x 4 (16-pixel) neighborhood of cells to produce a smooth resultant from a distance-weighted mean. The mean and variance of the output distribution match the input distribution; however, the range of data values may be altered as a result of this process of smoothing the data. The online documentation for ARC/INFO Version 7.2.1 offers the following guidance on resampling methods. Nearest neighbor is the preferred resampling method for categorical data because it does not alter the value of the input cells. It should be used for nominal or ordinal data where each value represents a class of data values rather than discrete data values. Bilinear interpolation is recommended for continuous surfaces because a known point or phenomenon determines the assigned value (e.g., elevation, and slope). Cubic convolution tends to smooth the data more than bilinear interpolation because of the smooth curves used as well as the larger number of points evaluated. Cubic convolution is the best method when total yields need to be determined (e.g., total CO2 emissions per country). All three techniques can be applied to continuous data, with nearest neighbor producing the most blocky output, and cubic convolution, the smoothest. However, neither bilinear interpolation nor cubic convolution should be used to resample categorical data. The 0.25-degree ARC/INFO exported integer grids were generated as follows: 1. Each of the five 3.75-km grids was unprojected (i.e., re-projected from an Albers into a geographic projection). 2. Missing data values were changed from -9999 to "NO DATA" by using the ARC/INFO GRID SELECT command for resampling purposes. 3. Nearest neighbor, bilinear interpolation, and cubic convolution algorithms were each used to resample actual and potential carbon biomass estimates in the BIOMASS grid to a 0.25-degree resolution. 4. The products of the resampled data were then projected back to Albers and summed. Based on a comparison of the following actual and potential biomass carbon estimates with values published by Brown et al. (1993), the cubic convolution method of resampling was used to produce the 0.25-degree biomass grid in this numeric data package. Resampling method Actual carbon (Pg) Potential carbon (Pg) Nearest neighbor 41.7256 73.8159 Bilinear interpolation 41.7286 73.8847 Cubic convolution 41.7583 73.9194 5. The remaining four grids were resampled by using the nearest neighbor assignment method because each grid contained only categorical data. 6. The resulting five 0.25-degree grids (i.e., BIOMASSX, CLIMATEX, DEMOGX, LANDX, and VEGTX) have attributes comparable, but not identical, to those found in the 3.75-km grids in this numeric data package. Table 1 displays the data ranges for the variables in each of the ten ARC/INFO GRIDS to illustrate the redistribution of the data after the resampling process. ********** Table 1. Redistribution of the data as a result of the resampling process Variable Number of name unique values Minimum Maximum Cell size Grid name AC 281 7 383 3.75 km BIOMASS AC 279 7 336 0.25 degree BIOMASSX PC 30 14 393 3.75 km BIOMASS PC 288 43 402 0.25 degree BIOMASSX CLIMI 20 1 20 3.75 km CLIMATE CLIMI 20 1 20 0.25 degree CLIMATEX PRECIP 13 1 13 3.75 km CLIMATE PRECIP 13 1 13 0.25 degree CLIMATEX POP 14 1 14 3.75 km DEMOG POP 14 1 14 0.25 degree DEMOGX CNTRY 16 1 16 3.75 km DEMOG CNTRY 16 1 16 0.25 degree DEMOGX SLOPE 6 1 6 3.75 km LAND SLOPE 6 1 6 0.25 degree LANDX ELEV 10 1 10 3.75 km LAND ELEV 10 1 10 0.25 degree LANDX SOILT 6 1 6 3.75 km LAND SOILT 6 1 6 0.25 degree LANDX FOREST 2 1 2 3.75 km VEGT FOREST 2 1 2 0.25 degree VEGTX EFZ 6 2 9 3.75 km VEGT EFZ 6 2 9 0.25 degree VEGTX VEG 16 1 20 3.75 km VEGT VEG 16 1 20 0.25 degree VEGTX ********** The cubic convolution method of resampling was used to transfer the data values of AC and PC in the BIOMASS (3.75-km) grid to the BIOMASSX (0.25-degree) grid. The data for AC in the 3.75-km grid range from 7 to 383, with 281 unique data values. After the resampling, the data for AC in the BIOMASSX (0.25-degree) grid ranged from 7 to 336, with 279 unique data values. The data for PC in the BIOMASS (3.75-km) grid ranged from 14 to 393, with 30 unique data values. After the resampling, they ranged from 43 to 402, with 288 unique data values in the BIOMASSX (0.25-degree) grid. The nearest neighbor method of resampling was used to transfer the data values of CLIMI, PRECIP, POP, CNTRY, SLOPE, ELEV, SOILT, FOREST, EFZ, and VEG in the remaining 3.75-km grids (CLIMATE, DEMOG, LAND, and VEGT) to the 0.25-degree grids (CLIMATEX, DEMOGX, LANDX, and VEGTX). Note that the data range and number of unique data values did not change for these variables. Twenty-four of the 26 remaining ASCII files were generated directly from the 10 ARC/INFO GRIDS (five 3.75-km grids and five 0.25-degree grids) by using the GRIDASCII command. The GRIDASCII command produces raster-based data files that can be used by most GIS software packages (and read by non-GIS software packages, as well). Each file contains R lines (where R = the number of rows in the grid + six header lines). Lines 1 through 6 contain the following values: the number of columns in the grid (line 1), the number of rows in the grid (line 2), the lower left-hand x (longitude) coordinate (line 3), the lower left-hand y (latitude) coordinate (line 4), the grid-cell size (line 5), and a definition of the grid's no-data value (line 6). The remaining lines in the file represent individual columns of data in the grid. For example, if there are 3066 columns and 1736 rows of data, there would be 1743 lines in the file. Lines 1 through 6 would contain the aforementioned header information, while lines 7 to 1743 would contain 3066 data values, each separated by a single space. Table 2 shows the arguments used with the GRIDASCII syntax to produce the 12 ASCII data files from the 3.75-km data and the 12 ASCII data files from the 0.25-degree data. ********** Table 2. GRIDASCII syntax used to produce the ASCII data files Grid name Output file name Variable name Variable description BIOMASS ac.dat AC Actual biomass carbon in Mg C/ha BIOMASS pc.dat PC Potential biomass carbon in Mg C/ha CLIMATE climi.dat CLIMI Weck's Climatic Index code CLIMATE precip.dat PRECIP Mean annual precipitation code DEMOG pop.dat POP Population density code DEMOG cntry.dat CNTRY Country code LAND slope.dat SLOPE Slope code LAND elev.dat ELEV Mean elevation code LAND soilt.dat SOILT Soil texture code VEGT forest.dat FOREST Forest or non-forest code VEGT efz.dat EFZ Ecofloristic zone code VEGT veg.dat VEG Vegetation code BIOMASSX acx.dat AC Actual biomass carbon in Mg C/ha BIOMASSX pcx.dat PC Potential biomass carbon in Mg C/ha CLIMATEX climix.dat CLIMI Weck's Climatic Index code CLIMATEX precipx.dat PRECIP Mean annual precipitation code DEMOGX popx.dat POP Population density code DEMOGX cntryx.dat CNTRY Country code LANDX slopex.dat SLOPE Slope code LANDX elevx.dat ELEV Mean elevation code LANDX soiltx.dat SOILT Soil texture code VEGTX forestx.dat FOREST Forest or non-forest code VEGTX efzx.dat EFZ Ecofloristic zone code VEGTX vegx.dat VEG Vegetation code ********** The remaining two generic ASCII files with longitude and latitude (x, y) coordinates were produced as follows: 1. A point coverage was generated from the BIOMASS grid by using the ARC/INFO GRIDPOINT command. 2. The ARC/INFO PROJECT command was used to project the meter coordinates into decimal degrees. 3. The output coverage from step 2 was ungenerated to produce an ASCII file containing a grid- cell id number, longitude, and latitude for each of the 4,177,584 grid-cell centers. 4. The 12 ASCII files produced by the GRIDASCII command for the 3.75-km data were then merged, one file at a time, with the file produced in step 3. 5. The result of steps 1 through 4 is a file called se_asia.dat with 4,177,584 records containing the following variables: grid-cell identification number, longitude in decimal degrees of the centerpoint of each grid cell, latitude in decimal degrees of the centerpoint of each grid cell, actual biomass carbon, potential biomass carbon, precipitation, population, country, slope, soil texture, forest designation, ecofloristic zone, and vegetation index. 6. A point coverage was generated from the BIOMASSX grid by using the ARC/INFO GRIDPOINT command. 7. The 12 ASCII files produced by the GRIDASCII command for the 0.25-degree data were then merged, one file at a time, with the files produced in step 6. Note that, because these data are provided in an unprojected format, there was no need to use a projection step to assemble these data for distribution. 8. Steps 6 and 7 resulted in a file called se_asiax.dat containing 100,198 records, with the same variables listed in step 5. Actual and potential biomass were each totaled for tropical Southeast Asia from the data sets included with this numeric data package, for comparison with the totals published by Brown et al. (1993). For each data set, the number of pixels with a specific carbon density was multiplied by the carbon density then multiplied by the pixel area to yield total carbon; finally, this product was summed for all carbon densities. For tropical Southeast Asia, estimated total biomass is 42.1 Pg C actual and 73.6 Pg C potential. The same totals were calculated from the 0.25-degree gridded data to be 41.8 Pg C actual and 73.9 Pg C potential. These totals agree with the corresponding values of 42 Pg C actual and 74 Pg C potential reported in Brown et al. (1993), verifying that overall the database included with this numeric data package reflects the data used by the authors in their publication. As an additional check, the actual biomass carbon density estimates for 1980 in this database can be compared with the carbon content data in Table 5 of NDP-046 (Richards and Flint 1994) for the same year. The 1980 total carbon in forest cover is estimated by Richards and Flint (1994) to be 23.95 Pg C in contrast with 42 Pg C estimated herein. This level of difference is similar to the differences in carbon densities observed in the estimates of Brown et al. (1993) and of Flint and Richards (1994). In addition to the methodological and data-source differences mentioned in Section 3, it must be noted that Richards and Flint (1994), but not Brown et al. (1993), include Singapore, whereas the converse is true for Nepal; furthermore, there are differences between the two databases in terms of the estimated area covered by forests. 5. REFERENCES Brown, S., L. R. Iverson, A. Prasad, and D. Liu. 1993. Geographical distributions of carbon in biomass and soils of tropical Asian forests. Geocarto International 4:45-59. Brown, S., G. Gaston, and R. C. Daniels. 1996. Tropical Africa: Land use, biomass, and carbon estimates for 1980. ORNL/CDIAC-92, NDP-055. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee. Flint, E. P., and J. F. Richards. 1994. Trends in carbon content of vegetation in South and Southeast Asia associated with changes in land use, pp. 201-299. In V. Dale (ed.), Effects of Land-Use Change on Atmospheric CO2 Concentrations: South and Southeast Asia as a Case Study. Springer-Verlag, New York. Houghton, R. A., and J. L. Hackler. 1995. Continental scale estimates of the biotic carbon flux from land cover change: 1850 to 1980. ORNL/CDIAC-79, NDP-050. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee. Iverson, L., S. Brown, A. Prasad, H. Mitasova, A. J. R. Gillespie, and A. E. Lugo. 1994. Use of GIS for estimating potential and actual forest biomass for continental South and Southeast Asia, pp. 67-116. In V. Dale (ed.), Effects of Land-Use Change on Atmospheric CO2 Concentrations: South and Southeast Asia as a Case Study. Springer-Verlag, New York. Olson, J. S., J. A. Watts, and L. J. Allison. 1985. Major world ecosystem complexes ranked by carbon in live vegetation: A database. NDP-017. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee. Richards, J. F., and E. P. Flint. 1994. Historic land use and carbon estimates for South and Southeast Asia. ORNL/CDIAC-61, NDP-064. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee. Risser, P. G., and L. R. Iverson. 1988. Geographic information systems and natural resource issues at the state level, pp. 231-239. In D. B. Botkin, M. E. Casswell, J. E. Estes, and A. A. Orio (eds.), Our Role in Changing the Global Environment: What Can We Do About Large Scale Environmental Issues? Academic Press, New York. Weck, J. 1970. An improved CVP-index for the delimitation of the potential productivity zones of forest lands of India. Indian Forester 96:565-572. 6. HOW TO OBTAIN THE DATA AND DOCUMENTATION This database (NDP-068) is available free of charge from CDIAC. The files are available from CDIAC's Web site (http://cdiac.ess-dive.lbl.gov) or from CDIAC's anonymous FTP (file transfer protocol) area (cdiac.esd.ornl.gov) as follows: 1. FTP to cdiac.esd.ornl.gov (128.219.24.36). 2. Enter "ftp" as the user id. 3. Enter your electronic mail address as the password (e.g., fred@zulu.org). 4. Change to the directory "pub/ndp068" (i.e., use the command "cd pub/ndp068"). 5. Set ftp to get ASCII files by using the ftp "ascii" command. 6. Retrieve the ASCII database documentation file by using the ftp "get ndp068.txt" command, and retrieve the ASCII data files by using the ftp "mget *.dat" command. 7. Set ftp to get *.e00 data files by using the ftp "binary" command. 8. Retrieve the *.e00 data files by using the ftp "mget *.e00" command. 9. Exit the system by using the ftp "quit" command. Uncompress the files on your computer, if they are obtained in compressed format. For non-Internet data acquisitions (e.g., floppy diskette or compact disk) or for additional information, contact: Carbon Dioxide Information Analysis Center Oak Ridge National Laboratory P.O. Box 2008 Oak Ridge, Tennessee 37831-6335, U.S.A. Telephone: 1-865-574-3645 Telefax: 1-865-574-2232 E-mail: cdiac@ornl.gov 7. LISTING OF FILES PROVIDED This database consists of 37 files: This documentation file (ndp068.txt, File 1), 10 exported ARC/INFO integer grid files, and 26 ASCII data files (Table 3). Five of the 10 exported ARC/INFO grid files have a pixel size of 3.75 km by 3.75 km, whereas the other five have a pixel size of 0.25 degrees by 0.25 degrees. Each core data layer in this database was also grouped into one of five thematic grids (see Sect. 4). The 3.75-km data were aggregated to a resolution of 0.25 degrees; the data at the two levels of resolution contain identical attributes. Each grid when imported into ARC/INFO is an integer grid and contains a value attribute table (vat) and a statistics table (sta). Except for the biomass carbon measures, each grid contains data classes identified by a numeric value code and defined by a character description of the class. Twenty-four of the 26 ASCII data files were generated by using the ARC/INFO GRIDASCII command. As such, these files can be used with or without ARC/INFO software and can be used by raster or vector GIS software packages as well as non-GIS software packages. These 24 files each represent one data item and, when used as a GRID in ARC/INFO, contain the same information found in the 10 ARC/INFO export grids in this numeric data package. The two remaining ASCII data files are aggregates of all the data within this database in ASCII format, one at a spatial resolution of 3.75 km and the other at 0.25 degree. Table 3 describes the files provided in this numeric data package. ********** Table 3. Files in this numeric data package File File size Projection number File name (kbytes) File description type File type 1 ndp068.txt 94 Descriptive file (i.e., this n/a ASCII text document) 2 Biomass.e00 59,468 Exported ARC/INFO gridded Albers ARC/INFO (3.75-km) estimates of actual export GRID and potential biomass carbon 3 Biomassx.e00 1,534 Exported ARC/INFO gridded Geographic ARC/INFO (0.25-degree) estimates of export GRID actual and potential biomass carbon 4 ac.dat 24,607 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75- ASCII data km) estimates of actual file biomass carbon 5 acx.dat 593 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) estimates of actual file biomass carbon 6 pc.dat 24,655 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75- ASCII data km) estimates of potential file biomass carbon 7 pcx.dat 594 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) estimates of potential file biomass carbon 8 Climate.e00 59,382 Exported ARC/INFO gridded Albers ARC/INFO (3.75-km) Weck's Climatic export GRID Index and mean annual precipitation 9 Climatex.e00 1,448 Exported ARC/INFO gridded Geographic ARC/INFO (0.25-degree) Weck's export GRID Climatic Index and mean annual precipitation 10 climi.dat 23,218 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data Weck's Climatic Index file 11 climix.dat 566 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) Weck's Climatic file Index 12 precip.dat 23,047 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data mean annual file precipitation 13 precipx.dat 562 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) mean annual file precipitation 14 Demog.e00 59,381 Exported ARC/INFO gridded Albers ARC/INFO (3.75-km) population density export GRID and country name 15 Demogx.e00 1,449 Exported ARC/INFO gridded Geographic ARC/INFO (0.25-degree) population export GRID density and country name 16 pop.dat 22,865 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data population density file 17 popx.dat 559 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) population density file 18 cntry.dat 23,056 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data country name file 19 cntryx.dat 563 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) country name file 20 Land.e00 59,401 Exported ARC/INFO gridded Albers ARC/INFO (3.75-km) slope, elevation, export GRID and soil texture 21 Landx.e00 1,461 Exported ARC/INFO gridded Geographic ARC/INFO (0.25-degree) slope, export GRID elevation, and soil texture 22 slope.dat 22,884 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data slope file 23 slopex.dat 559 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) slope file 24 elev.dat 22,873 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data elevation file 25 elevx.dat 559 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) elevation file 26 soilt.dat 22,884 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data soil texture file 27 soiltx.dat 559 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) soil texture file 28 Vegt.e00 59,392 Exported ARC/INFO gridded Albers ARC/INFO (3.75-km) forest or export GRID non-forest designation, ecofloristic zone, and vegetation type 29 Vegtx.e00 1,452 Exported ARC/INFO gridded Geographic ARC/INFO (0.25-degree) forest or export GRID non-forest designation, ecofloristic zone, and vegetation type 30 forest.dat 22,880 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data forest or non-forest file designation 31 forestx.dat 559 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) forest or non-forest file designation 32 efz.dat 22,895 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data ecofloristic zone file 33 efzx.dat 560 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) ecofloristic zone file 34 veg.dat 23,037 ASCII file of ungenerated Albers GRIDASCII ARC/INFO gridded (3.75-km) ASCII data vegetation type file 35 vegx.dat 562 ASCII file of ungenerated Geographic GRIDASCII ARC/INFO gridded (0.25- ASCII data degree) vegetation type file 36 se_asia.dat 407,372 ASCII file of gridded (3.75- n/a composite km) grid-cell identification ASCII data number, longitude and file latitude (of the centerpoint of each grid cell), estimate of actual biomass carbon, estimate of potential biomass carbon, Weck's Climatic Index, mean annual precipitation, population density, country name, slope, elevation, soil texture, forest or non-forest designation, ecofloristic zone, and vegetation type 37 se_asiax.dat 9,780 ASCII file of gridded (0.25- n/a composite degree) grid-cell ASCII data identification number, file longitude and latitude (of the centerpoint of each grid cell), estimate of actual biomass carbon, estimate of potential biomass carbon, Weck's Climatic Index, mean annual precipitation, population density, country name, slope, elevation, soil texture, forest or non-forest designation, ecofloristic zone, and vegetation type Note: GRIDASCII is an ARC/INFO (TM) command that produces an ASCII file containing data for an individual gridded data layer. ********** 8. DESCRIPTION OF THE DOCUMENTATION FILE ndp068.txt (File 1) This file is identical to this document. 9. DESCRIPTION, FORMAT, AND PARTIAL LISTINGS OF THE ARC/INFO GRID FILES Ten of the 37 files contained within this database are exported ARC/INFO grids. Each exported grid file was generated using the EXPORT command in ARC/INFO with the 'grid' and the 'no data compression' options (e.g., EXPORT GRID BIOMASS BIOMASS NONE). Five of the exported grid files contain a pixel size of 3.75 km, whereas the other five have a pixel size of 0.25 degrees. The five 3.75-km ARC/INFO export grids are named BIOMASS, CLIMATE, DEMOG, LAND, and VEGT, and the five 0.25-degree grids are called BIOMASSX, CLIMATEX, DEMOGX, LANDX, and VEGTX. Each grid has been projected into Albers with a unit base of meters by using the following parameters: 1st standard parallel: 30 08 24.000 2nd standard parallel: -4 17 24.000 central meridian: 107 28 12.000 latitude of projection's origin: 0 0 0.000 false easting (meters): 0.00000 false northing (meters): 0.00000 The five 3.75-km ARC/INFO grids originate at -5879340.56205, -1221655.95152 m and extend to 3863159.43795, 4808344.04848 m; these values are approximately equal to an origin of 44.25875 longitude, -16.52954 latitude and an extent of 149.50875 longitude, 42.97046 latitude. There are 1608 rows and 2598 columns in each grid. The five 0.25-degree grids originate at 44.25875 degrees longitude, -16.52954 degrees latitude and extend to 149.50875 degrees longitude, 42.97046 degrees latitude. There are 238 rows and 421 columns in each grid. The 3.75-km grids and the 0.25-degree grids differ only in spatial resolution. The files with an "x" suffix are associated with the aggregated data. For example, BIOMASS and BIOMASSX have exactly the same attributes, as is true for CLIMATE and CLIMATEX, DEMOG and DEMOGX, LAND and LANDX, and VEGT and VEGTX. Table 4 defines the attributes of each grid. Table 4. Item descriptions for the ten ARC/INFO export grids ********** 3.75-km grid 0.25-degree Item Input Output Variable name grid name Column name width width Item type description BIOMASS BIOMASSX 1 Value 4 10 Binary Unique value for (2,200 records (2,209 records each grid cell in .vat file) in .vat file) 5 Count 4 10 Binary Cell count associated with each unique value 9 ac 4 16 Binary Actual biomass carbon (Mg C/ha) 13 pc 4 16 Binary Potential biomass carbon (Mg C/ha) CLIMATE CLIMATEX 1 Value 4 10 Binary Unique value for (201 records (154 records each grid cell in .vat file) in .vat file) 5 Count 4 10 Binary Cell count associated with each unique value 9 Climi 4 16 Binary Weck's Climatic Index (code) 13 precip 4 16 Binary Mean annual precipitation (code) 17 climi-c 12 12 Character Weck's Climatic Index (code definition) 29 precip-c 12 12 Character Mean annual precipitation (code definition) DEMOG (166 DEMOGX 1 Value 4 10 Binary Unique value for records in .vat (147 records each grid cell file) in .vat file) 5 C 4 10 Binary Cell count associated with each unique value 9 pop 4 16 Binary Population density (code) 13 pop-c 18 18 Character Population density (code definition) 31 cntry 4 16 Binary Country (code) 35 cntry-c 24 24 Character Country (code definition) LAND (333 LANDX 1 Value 4 10 Binary Unique value for records in .vat (244 records each grid cell file) in .vat file) 5 C 4 10 Binary Cell count associated with each unique value 9 Slope 4 16 Binary Slope (code) 13 Elev 4 16 Binary Mean elevation (code) 17 Soilt 4 16 Binary Soil texture (code) 21 slope-c 18 18 Character Slope (code definition) 39 elev-c 18 18 Character Mean elevation (code definition in meters) 57 soilt-c 18 18 Character Soil texture (code definition) VEGT (258 VEGTX 1 Value 4 10 Binary Unique value for records in .vat (156 records each grid cell file) in .vat file) 5 Count 4 10 Binary Cell count associated with each unique value 9 Forest 4 16 Binary Forest (code) 13 Forest-c 12 12 Character Forest (code definition) 25 efz 4 16 Binary Ecofloristic zone (code) 29 efz-c 18 18 Character Ecofloristic zone (code definition) 47 veg 4 16 Binary Vegetation type (code) 51 veg-c 24 24 Character Vegetation type (code definition) ********** The ARC/INFO IMPORT command or the ARCVIEW IMPORT program must be used to read the ten ARC/INFO export grids. The syntax for the ARC/INFO IMPORT command is "IMPORT