Spatial Production Allocation Model 2000 Version 3 Release 2

The datasets are pixel-level estimates of crop area, production and yield. While the datasets are consistent with actual census data at aggregate levels (e.g. country, state, counties), the pixel level distributions are modeled results which may differ from the actual condition at the farm level. We recommend they be used for regional and global analysis, and use caution when used in country or sub-national studies.

Data files

The [Download/Data] section contains the following three folders: Grid Cell ID, Raster, and Raw Data Table. Grid Cell ID contains the HarvestChoice 5 minute resolution unique id surface. This can be used in correlation with the raw data tables for further analyses and mapping. The Raster folder contains zip files of the SPAM data results in ascii format. Each zipped package includes a set of raster files of 20 crops for each results variable (physical area, harvested area, production and yield). These files are in the ESRI grid ASCII format and are easily imported into various GIS software. The Raw Data Table folder contains zip files of the complete set of SPAM results in ASCII text and dbf formats including results for each results variable for 20 crops by input system (irrigated, high-input/commercial rainfed & low-input/subsistence rainfed). 

HarvestChoice grid cell ID

The HarvestChoice grid is a global 5 minutes (0.08333dd) resolution id grid containing sequential numbers starting from 0, and originating at -180,90 (upper left corner). This file contains both a unique sequential cell id (cell5m) covering global land area and an allocation id (Alloc_key) unique to SPAM. The HarvestChoice id grid originates from a series of global sequential id files developed by Robert Hijmans that includes raster files, shapefiles, country ID look-ups, and documentation which can be found at IRRI.

The allocation id (Alloc_key) is a unique id based on latitude and longitude used as part of the SPAM allocation process.To create these ids, the world is divided into 4320 (longitude) * 2160 (latitude) cells. Each cell is approximately 5 arc minutes (0.0833334 degree). The Alloc_key has eight digits. The first four digits represent row number and the last four digits represent column number. The equation to get row and column number are shown as below:

col = ((gridx + 180)/(0.0833334) ) + 1
row = ((90 - gridy)/(0.0833334) ) + 1

This means:

gridx = (col - 1) * 0.0833334 - 180
gridy = 90 - (row - 1) * 0.0833334

where gridx, gridy are longitude and latitude in radians, respectively.

SPAM results

The SPAM results data are organized in two formats:  Tables (ascii text & dbf), and Raster (ESRI ASCII). The original model results are in tables, which were converted into ascii for easy import and mapping.

The tables are divided by record type: P- Physical area(ha), H - Harvested area(ha), R-pRoduction(ton), Y - Yield (kg/ha). In the raw data table these are identified by "Rec_type" column. In the ascii files these are divided into four separate types of files: _production: production (ton, production = harvested area * yield), _harvested-area: harvested area (hectare), _physical-area: physical area (hectare, physical area *cropping intensity = harvested area), and _yield: crop yield (kg/ha).

SPAM results - raw data format

The Raw Data folder contains the SPAM results in ascii text and dbf formats. In the raw data tables the first two fields are geopolitical codes that relate to the country and sub-national units. The fourth field (“alloc_key”) is pixel id concatenated by row/column numbers (each four digits). The allocation was run based on these unique cells. The fifth field ("cell5M") is a unique mapping field that is used to link the data table to the HarvestChoice grid for mapping purposes. The following columns are the SPAM results by crop. 

In the raw data table each crop is divided into 3 sub-systems: irrigated ("I"), rainfed high-input/commercial ("H"), rainfed low-input/ subsistence ("L"). These are stored by columns and in total. So "Rice" refers to the total rice physical area (if rec_type="P"), total harvested area ("H") or total production ("R"). "Rice-H" refers to high-input rainfed physical area (if rec_type="P"), high-input rainfed harvested area ("H") or high-input rainfed production ("R") and so on for each of the 20 crops. By definition, the total column (e.g. "Rice") should be equal to the sum of the sub-systems, (e.g. "Rice-H","Rice-L","Rice-I") for each row. Currently only the total values are available in ascii format.

The 20 crops are: wheat, rice, maize, barley, millet, sorghum, potato, sweet potato, cassava and yams, plantain and banana, soybean, dry beans, other pulse, sugar cane, sugar beets, coffee, cotton, other fibres, groundnuts, and other oil crops.  Additionally, other crops are always zero values except in the physical area table where the value for other crops is estimated by subtracting the sum of the above 20 crops from total cropland area for each pixel.

SPAM results - Raster ascii format

The Raster folder contains 5 zip files. There is one file that contains all of the results data (e.g. spam2000v3r2.zip) and 4 additional files with results organized by variable: physical area, harvested area, production & yield (e.g. spam2000v3r2_harvested-area.zip); Each zipped package by variable includes 84 raster files (20 crops + other-crops by 3 input systems and total). 

All the raster files are in the ESRI Grid ASCII format, a non-proprietary format widely being used in many GIS software. Each file contains the data for the given variable and crop along with the following header information:

• ncols 4060
• nrows 1506
• xllcorner -159.83333
• yllcorner -55.50000
• cellsize 0.08333
• NODATA_value -9999

Note: Version 3 release 1 contained incorrect xllcorner and yllcorner values. These have been corrected in Version 3 release 2.

Each zip file also contains a copy of the license agreement between you and HarvestChoice which you agreed to in order to access the data.

Contact

All data files and maps were generated by the HarvestChoice project ( This e-mail address is being protected from spambots. You need JavaScript enabled to view it )