Improper scaling of ABI L1b/L2 products

Just copying the issue here in the ticketing system

File - changeo:/data/users/wstraka/DSI_glance/OR_ABI-L2-DSIF-M4_G16_s20170151527081_e20170151531484_c2017015153

This is using the LI product in the above file.

For the soundings file from G-16, at pixel x,y = 612, 83 (highlighted in both attached images. Note: HDFview is 0 based, hence why it is off by one). The (signed) value is -8835. scale factor = 7.6301E-4, add offset = -10.0.

So, following the steps to unscale the data, per Harris's writeup

1). _Unsigned = True, so convert to unscaled value:x = (-8835) + 65535 = 56700 2). Apply scale factor/add offset (mx + b) = 33.2627.

However, as can be seen in the plot dump in glance, that point is being dumped out as 23.26343. I wonder if this has to deal with how the _Unsigned flag is being applied within Glance

Looks like was done in the alphas resolved the issue for LI. Will check other files that have negative unscaled values as well.

I know there was a related issue in the L1b data (negative radiances, I believe). Does this fix that as well??

That fix should (I hope) have corrected the unsigned related issues for all the data scaling.

closed