Earthdata Forum

I have been checking out the new NISAR GUNW release and have found three GUNW products that seem to have very very large ramps or random noise over the whole scene. They are from different orbits (77 and 149) but both over the same frame (24). I don't know if nearby frame have this issue as they don't have GUNWs released for this period. See attached screenshot of one of the GUNW quick look images.

GUNWs I have found:
NISAR_L2_PR_GUNW_010_077_A_024_012_4000_SH_20260114T124618_20260114T124653_20260207T124619_20260207T124654_P05023_N_F_J_001
NISAR_L2_PR_GUNW_008_077_A_024_010_4000_SH_20251221T124616_20251221T124651_20260114T124618_20260114T124653_P05023_N_F_J_001
NISAR_L2_PR_GUNW_008_149_A_024_009_4000_SH_20251226T123757_20251226T123832_20260107T123757_20260107T123832_P05023_N_F_J_001

when I plot the azimuth offset for these three GUNWs they show signals of +-100 pixels which seems unlikely given that the other one I checked was sub pixel.

might be related but looks different since it is a hard cut?
viewtopic.php?t=8118&sid=0e7cb496bbc56a9ff257e6a3b75e86f2

Thanks for reporting this issue. The problem originates from corrupted input raw data. Looking at the reference RSLC shows that the image is fully defocused and therefore the interferometry fails to produce any useful product. For example see this RSLC product:

NISAR_L1_PR_RSLC_010_077_A_024_4005_DHDH_A_20260114T124618_20260114T124653_P05023_N_F_J_002

We are working on recovering useful parts of data, and through that process, the P05023 has a known edge case where the recovered data from an observation with missing/discarded data has wrong time tags and therefore causing downstream issues in level-1 and level-2 products. We are testing a fix for the next delivery.

Thanks that is very helpful. What is corrupting the input data that means you need to discard echoes? Is there any flag (other than terrible looking phase data) for picking out when an RSLC has failed to focus I should be checking for?

Attaching RSLC image of poorly focused RSLC for folks who are interested.

thanks again

P05023 provisional products, do not include masks and metrics to help users discard such data. We are indeed adding these metrics in the next delivery. We understand the importance of such metrics for users and will update as they become available.