Lieven Claessens resolves the Great Millet Mystery:
In GYGA we use HarvestChoice’s SPAM crop distribution masks for our analysis. SPAM in turn uses FAO statistics to create spatially disaggregated maps of crop distribution. I looked in more detail to the FAOstat numbers and found out that their numbers for ‘millet’ are actually for all sorts of millets combined… So finger and pearl millet are combined, even with teff to my big surprise! Identity crisis? ;-) In GYGA we have used a generic model for both pearl and finger millet so they are combined in the analysis….
Bottom line: the GYGA results for “millet” are unusable. But it’s FAO’s fault.
I think FAOSTAT make a mistake. They say that Teff is included in the FBS category ‘Millet‘.
However, when you look at Ethiopia’s 2011 Food Balance Sheet, there are only 5.8 kg/capita/yr of ‘Millet and Products’ but 27.4 kg/capita/yr of ‘Cereals, other’. For comparison, figures for ‘Maize and products’ and ‘Wheat and products’ are 42.9 and 31.9 kg/capita/yr, respectively, and since teff is considered the staple grain of Ethiopia one would expect its consumption to be up there with the other main grains.
So, I reckon the teff production data is actually falling into the ‘Cereals, other’ category.
Cheers!
You are right re: ‘teff production data is actually falling into the ‘Cereals, other’ category’. I remember searching for teff data on FAOSTAT a while back, and, asking around, someone (was it @FAOstatistics on Twitter?) confirmed it.