Hello,

My question is related to use and processing of multiple MODIS products such as MOD17A2HG (GPP and PSN), MOD16A2HG (ET and PET) and MOD11A2 (LST). Each MODIS product has its own QC (quality control) layer. Also, MOD17A2HG and MOD16A2HG are gap filled datasets.

In a multi-dataset study involving above mentioned MODIS products along with other satellite (non-MODIS) derived gridded datasets, should MODIS QC layers be used or can a custom masking approach be adopted? For instance, among all the variables, MODIS ET is known to experience more discrepancies due to atmospheric noise. Therefore, how about if a mask is made using ET data by excluding the outlier pixels and the same spatial mask is applied on all other datasets to remove same pixels (belonging to x,y) from each dataset?

I wonder using QC layer specific for each MODIS data would result in huge loss of data.

Also, if there is any reference in support or against, please share it.

FYI, the study belongs to ecology domain and doesn't involve the aspect of clouds.

Thank you,

Regards,

Akanksha