Agricultural Biodiversity Weblog

More today on where bananas grow. Or where people think they may grow, anyway. First, let’s tie up some loose ends. Yesterday I mentioned Chad Monfreda‘s dataset from the Land Use and Global Environmental Change project, but said that I hadn’t been able to get an actual map. Thanks to Julian and Andy, I now have. To the left is the result for banana.

And to the right for plantain.

Andy also pointed me to another dataset, called MIRCA2000: “Global data set of monthly irrigated and rainfed crop areas around the year 2000.” Here’s Andy’s pragmatic take on this proliferation of datasets:

You do have information on quality and uncertainties and there is enough info — if you are interested — to work out which to use. I don’t believe that there is a right or wrong answer here, they all try to make the best of what is often very cruddy and incomplete data.

In short – you’ll have to sort it out yourself, along with the rest of us. I just take the average of these when I need to make a rice area map – though the average of a bunch of wrong datasets is still a wrong dataset.

It’s a choice between modelled crop distributions or statistical data or a combination of both. Neither is wholly accurate and like most global data sets, it is best not to look too closely at the results.

4 data sets are better than none right?

Glenn also chimed in with his work for the Generation Challenge Programme, which uses the SPAM dataset I mentioned yesterday. It’s still in beta, so you can’t play around with it yet, but, again, this is what banana in Africa looks like:

Julian sums it up very well in his very comprehensive comment on yesterday’s post, and I leave the final word to him:

This is what you’d normally expect when there’s no coordination or data sharing among institutions: you end up with several institutions, each one re-inventing the wheel. The public then does not know which source to use.
…
So, you end up with two sources of data: (1) FAO country level statistics, and (2) expert knowledge; two methods: Monfreda et al.’s and HarvestChoice’s; and one visualization tool. Anyone interested in comparing or merging?

Where does crop X grow? Important question. And pretty simple too, no? No! Because just a little looking around yields about half a dozen different answers, and no clear idea of which to trust, or how they relate to each other, or how they were arrived at, or even if there are more.

Here’s what I came up with in only about half an hour of searching. The following are all data for banana/plantains. First, there’s MapSpaM (that would be Spatial Production Allocation Model), from HarvestChoice:

Then there’s FAO’s AgroMAPS, which has some really weird data in it. Try looking at the distribution of cassava in Africa, for example. Anyway, here’s banana, which actually looks pretty weird itself:

And then there’s CIAT’s Crop Atlas of the World. ² That says it is based on the FAO data, but doesn’t really seem like it to me, at least not on this evidence:

In its time, CIAT has also used the dataset from the Land Use and Global Environmental Change project called “Harvested Area and Yields of 175 crops (M3-Crops Data),” but I haven’t been able to get a map of that.

And then there’s IITA’s banana mapping effort, which admittedly is very much still a work in progress:

Well, I suppose I could sort out some of the questions I have about all this if I spent a little more time at it. But really, should a poor boy have to? Shouldn’t FAO, or the CGIAR, have all this sorted out by now? Anyone out there want to guide me through this?