Similar issues have come up recently with rice and maize. It seems something might be wrong with the model data we have distributed. See the instructions here to download the other (possibly more correct) versions of the rice and maize yield data. And please let us know if that gives you results more like what you expect.