National 50th percentile mosaics have missing cells or missing scenes

[NormanGA]

I'm reviewing the outputs of the synthetic national mosaics being created by Biswajit. The outputs are mosaics of pixels created from the 50th percentiles of NBAR drawn from 2-year seasonal periods. As an example the 1989-1990 dataset is created from the 50th percentile of values observed between April and September in 1989 and 1990.

Datasets are located in v10/std_prod_epoch_2, named ARG25_Date1_Date2_APR_TO_SEP_BAND_PERCENTILE_50.tif

There is one file per band, so a complete mosaic set is 6 files.

We need to investigate the cause of errors that turn up in these outputs so we can determine if they are caused by underlying NBAR, PQ or the processing code and fix where appropriate.

Issues apparent in the ARG25_19890401_19900930_APR_TO_SEP_ ... dataset:

Location and issue
129_-22.5 to 131_-19.3 : Two scene area with mostly null results. It looks like nearly the entire area has been masked for cloud.

121.1_-19.6 to 122.1_-18.1 : One scene all null.

Cell 138_-035 : no data

Cell 138_-034 : Data missing from SWIR-1 and NIR (at least)

Cell 144_-035 : no data

Cell 144_-034 : no data

Cell 144_-033 : Data missing from SWIR-1 and NIR (at least)

Cell 145_-029 : Data missing from SWIR-1 (at least)

145.5_-42.5 to 146.6_-40.9 : Two scene are, all null.