What are the differences between SPAM 2005 v2.0 and SPAM 2005 v3.2?
  1. ALL regions: An improved cropland surface (From IIASA/IFPRI) was used
  2. ALL regions: Input statistics and results are scaled to FAOSTAT (average 2004-2006) released on Jan-8-2016
  3. ALL regions: Except for China, the existing distribution surface was not used
  4. Europe: There is no subsistence share in EU countries for any crop
  5. LAC: Oil palm in Brazil is now assigned to the correct state (previously to Rio Grande do Sul, now Para)
  6. LAC: Ecuador had wrong labeling of level-2 unit, corrected (Zamora was part of Morona Santiago, is now part of Zamora Chinchipe)
  7. ASIA: In India added all GAUL polygons identified as Arunachal Pradesh (country in conflict) to India’s state Arunachal Pradesh. This results in more pixels for the old Arunachal Pradesh (and thus India).
  8. ASIA: In India add some of the GAUL polygons identified as Jammu&Kashmir (country in conflict) to India and rename them to state Jammu&Kashmir. Level 2 (district) boundaries taken from GADM. This results in more pixels for India.
  9. SSA: In Ghana, statistics was fixed to include one missing district
  10. In V2.0, the optimization routine allowed “slacks” for cropland, irrigated areas, and sub-national statistics. In some instances, this resulted in allocated areas which exceeded the pixel size. Now only “slacks” for sub-national statistics.
  11. For the following countries, some crops had level-1 entries with -999, which yields no allocation: Czech Republic, Italy, Poland, Sweden, Yemen and New Zealand. This has been fixed.
What are the differences between SPAM 2005 v1.0 and SPAM 2005 v2.0?
  1. Better sub-national statistics for some countries
  2. We revised again the irrigated share of crops to accommodate comments made by the water team at IFPRI.
  3. Better control of model solution when initially it does not solve. One of these controls is to switch off suitability constraints. In the earlier release of SPAM2005 we had to turn off ALL suitability constraints to attempt a solution, now we can do it selectively. This selective turn off of suitability constrains leads to a crop distribution which matches better the agro-ecological conditions, and not just the information on agricultural land and irrigated areas.
  4. Still on the methodology side, when a solution through cross entropy is not found, another attempt is made with a different method (quadratic approximation) to hopefully receive some results, thus avoiding the excessive adjustment of other parameters.

SPAM 2005 evolved over the years. There were a few versions available before this latest one, the best known being SPAM 2000, which can also be downloaded from this website.

Both versions do not differ in the basic underlying methodology, but there are rather big differences in the data.


SPAM2000 allocates 20 crops, of which 15 are individual crops and the rest are aggregates of 2 or more crops. In SPAM2005 we expanded the list to 42 crops, which include 32 individual crops, 2 sub-crops and 8 aggregates.


In SPAM2000 the sub-national and national crop statistics were centered around the year 2000 and scaled with FAO’s average (1999-2001) figures. Now, in SPAM2005, we centered the sub-national and national crop statistics to the year 2005 and scaled with FAO’s average (2004-2006) figures. Along the same line we assembled, where available irrigation shares and cropping intensities which reflected the condition around the year 2005.


The previous version used the M3-Cropland described in the publication Ramankutty et al. (2008), (“Farming the planet: 1. Geographic distribution of global agricultural lands in the year 2000”, Global Biogeochemical Cycles, Vol. 22, GB1003, doi:10.1029/2007GB002952. SPAM2005 uses cropland compiled by Fritz et al at (2015)(“Mapping global cropland and field size”, Global Change Biology,doi: 10.1111/gcb.12838). Areas equipped for irrigation were previously take from The Global Map of Irrigation Areas (GMIA) V1 and are now based on V5 (http://www.fao.org/nr/water/aquastat/irrigationmap/index10.stm).


The data files generated by GAEZ v 2.0, FAO/IIASA, 2002 (http://webarchive.iiasa.ac.at/Research/LUC/SAEZ/index.html) were the basis for our suitability surfaces for SPAM2000. For the current version we relied on the files generated by GAEZ V3.0 2009 (http://webarchive.iiasa.ac.at/Research/LUC/GAEZv3.0/). Not only is GAEZ 2009 an updated version of GAEZ 2000, but the list of crops for which there are suitability calculations is more extensive, and thus a better match for SPAM2005.