These are chat archives for EcoForecast/GPP

25th
Mar 2016
Alexey Shiklomanov
@ashiklom
Mar 25 2016 00:16
@bcow If MODIS GPP defeats you, grab EVI. We can do something like build a model of GPP based on combined MODIS and OCO SIF data, calibrated to flux towers.
It'll be janky, but no more janky than this project will be as is.
And we can have a cool figure showing how GPP precision improves as soon as OCO2 kicks in.
Michael Dietze
@mdietze
Mar 25 2016 00:24
If modis GPP defeats you I'd recommend fAPAR over EVI since GPP ~ LUE * fAPAR
Alexey Shiklomanov
@ashiklom
Mar 25 2016 00:32
:+1:
Alexey Shiklomanov
@ashiklom
Mar 25 2016 20:49
@bcow If you're going to grab MODIS stuff manually anyway, have a look at product 20 -- Chlorophyll fluorescence. If it's only available for the ocean (which is what it's intended for), can you grab the surface reflectance (product 1B) for bands 13, 14, and 15 (from which I can manually calculate something resembling fluorescence following the general methods described here). I think it may be possible, without a ton of work, to cross-calibrate MODIS and OCO fluorescence observations, which would dramatically extend our record. And since these are traditionally ocean bands, I'm pretty sure they don't factor into the GPP algorithm and therefore can be considered independent observations.
Let me know if you need help. I'm assuming you can basically use the same script to grab all of these, but if not, I can write the scripts for the stuff I just described.