MapSPAM.info

Observed as: Incompatible Crop/Production System Maps containing non zero pixels for production but zero area for the same pixels (for example)

Cause: [High Certainty] SPAM results are shown to 1 decimal place in the following units (Area harvested and physical area: ha, Yield: kg/ha, Production: mt). Calculations within SPAM have the accuracy of the software used, while SPAM results posted on the web show only 1 decimal place. As a consequence, values less than 0.05 are shown as 0 (e.g., for production) but may have associated values which are larger than 0.05 (e.g., for area). These inconsistencies arise mostly in cases where the area and/or yield values are very small.

Resolution: The next release of SPAM will resolve these inconsistencies, and should soon be on the website. We'll keep you posted, but our target date is by the end of May 2010.

Interim User Options: In the interim users could mask out (set as zero) yield and production pixels for crop * system layers (e.g. subsistence maize) whose area pixel values are zero. This will provide consistent databases but may result in aggregate production values at the national level being less consistent with equivalent FAO reported national production values (used in the initial scaling of SPAM sub-national values)

Note: Here is a list of feedback items that we've received since the release of SPAM 2000 V3 R1 on November 9th, 2009, and this list will be keep updated. We are committed to investigate the issues and revise/improve SPAM in upcoming releases. Thanks!

SHIFT IN SPAM V3R1 DATA BY APPROXIMATELY ONE HALF A CELL

The lower left coordinates (xll & yll) provided in the header portion of the ASCII data download files for SPAM version 3 release 1 are incorrect resulting in a shift in the data of approximately one half a cell (0.04167dd or approximately 5km) north and west. This problem has been corrected and the revised files are available for download as version 3 release 2.The content of the files has NOT changed so if you have already downloaded the data and would prefer to just correct the coordinates, this can be done in one of two ways. 1) Before converting the ASCII file to a raster, change the xll corner value to -159.8333 and the yll corner to -55.5000. (The ASCII files can be manually edited using Notebook and then saved as .txt files). 2) To correct the SPAM data that have already been imported: in ArcToolbox, use Data Management Tools, Projections and Transformations, Raster, Shift with the shift values of +0.041667 for the x value and -0.041667 for the y value. This will shift the surface to its correct location.

Banana in Portugal

Portugal shows banana-plantain production, and it seems a bit odd. After checking some more, we learned that Portugal's banana production stems from Madeira, which is Portuguese territory. In that sense it is OK, but it should not show up in Portugal's mainland. The next revision of SPAM will take account of this.

Maize in Southern DRC

There is a missing maize growing area. We'll look into the modeling process for the country and see if there was any error occurred while processing different levels of sub-national production statistics.

Cassava in Western Kenya and Northern Tanzania

SPAM shows low production of cassava, which is not consistent with the data being collected by the GLCI project (reported by Chris Legg).