The current logic for updating default experiment file JSON using CDL data breaks when a field is fallow for any given year. This is temporarily fixed by leaving the crop from the default experiment file as is, but the actual fix should be on the logic where this gets updated.