mksurfdata_esmf should interpolate "abm" value 13 among the dominant values to work correctly #3188
Description
I'm preparing to open a new issue that @lifang0209 brought up. I'm starting the conversation here because it's related. @lifang0209 and everyone, please correct and/or comment on the below:
@lifang0209 points out that
- abm is "agricultural fire peak month" and should not be spatially averaged from the raw data to the fsurdat file
- Instead we should pick the dominant value when interpolating
- We are not doing this correctly in mksurfdata_esmf and this is the issue that I plan to open
- And my question: Could this be done correctly if abm were changed to a stream, which is why I'm bringing up the topic here
Also @lifang0209 has a fix in the fire model to address this until we correct mksurfdata_esmf. In an email @lifang0209 wrote:
"I checked the f1.9 dataset, and it’s not correct. I’ve added an IF statement in the fire model to assume no crop fire if crop coverage is less than 10%. It's consistent with observations as crop fires are rare in these grids and it also helps resolve the issue."
Originally posted by @slevis-lmwg in #2617