5.19.26

[scottyhq]

Attendees: Eric, Scott

Time & Location: 10am eScience Seminar Room

Agenda:

~5 min: Voluntary updates and/or introductions
~10 min: Scott to share figures and learnings from NISAR proposal
~40 min: Co-working
~5 min: Wrap-up

Notes:

• reviewed figures of GUNW, GCOV, GSLC
• discussed coherence in NISAR vs Sentinel-1. NISAR tends to be lower than comparable 12-day C-band Sentinel-1 products from HYP3 or OPERA CLSCs. Maybe due to HH vs VV polarization, different incident angles, weather conditions a day a part. Maybe some histogram matching and differencing would be interesting... not sure of scientific use case though. Also should wait until calibrated products are released.
• GCOV shows 5dB scatter over Mt Rainier snow fields... but uncalibrated and not during a time of much snow, so hard to interpret. Also need to wait for more data and calibrated data
• Discussed visualization options for Zarr/Icechunk. Eric - Icechunk seems a bit slower if anything using https://github.com/carbonplan/zarr-layer, which is neat for geozarr with overviews but definitely a bit buggy! Also keeping an eye on https://github.com/developmentseed/deck.gl-raster which is adding Zarr support in addition to COG.
• Discussed idea for creating a remote cache of virtualized datasets for NISAR - see https://earthaccess.readthedocs.io/en/stable/api/#earthaccess.virtual.virtualize

people are welcome to linger in the eScience studio and continue to work!