Negative sprectral values of processed MODIS L0/L1A data

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
3sunanda.m
Posts: 16
Joined: Tue Dec 11, 2018 10:42 am America/New_York
Answers: 0

Negative sprectral values of processed MODIS L0/L1A data

by 3sunanda.m » Wed Apr 10, 2019 3:43 am America/New_York

Dear All

I have processed Modis L0 to L2 in Seadas 7.5.1. I had to work with Lw & nLw bands so I opted for these two in the Products tab of the l2gen dialog box while processing. I exported the final reprojected file (L2) into .hdf5 format so that I can use the same data in other platforms. But when I am checking the values of the spectral reflectances I could see negative values for all bands. Can anyone give insights about the issue.

Regards
Sunanda


Tags:

OB.DAAC - SeanBailey
User Services
User Services
Posts: 1381
Joined: Wed Sep 18, 2019 6:15 pm America/New_York
Answers: 1

Negative sprectral values of processed MODIS L0/L1A data

by OB.DAAC - SeanBailey » Thu Apr 11, 2019 8:56 am America/New_York

Without seeing your output files, we can provide no assistance other than to suggest that you check the data at each stage of the process to verify the values.

Sean

gnwiii
Posts: 691
Joined: Fri Jan 29, 2021 5:51 pm America/New_York
Answers: 2

Negative sprectral values of processed MODIS L0/L1A data

by gnwiii » Thu Apr 11, 2019 9:33 am America/New_York

There is a famous article titled "How to Report Bugs Effectively", which might better be titled "How to Report Unexpected Results Effectively".  You haven't provided enough detail to allow others to reproduce your calculation, so there is very little anyone can do to help you.  It is worth noting that it is often much easier to communicate your processing steps if you use a terminal as opposed to a GUI so you can paste text into your post (as opposed to attaching screen shots).   Doing this forces you to review your processing steps and may help you solve the problem without outside help (if so, please post your findings in case others encounter a similar issue). The NASA experts have plenty to do making SeaDAS better, so we all benefit if you avoid placing extra demands on their time.  In the long run, posting information that may help others avoid similar issues reduces demands on the NASA folks.

3sunanda.m
Posts: 16
Joined: Tue Dec 11, 2018 10:42 am America/New_York
Answers: 0

Negative sprectral values of processed MODIS L0/L1A data

by 3sunanda.m » Tue Apr 16, 2019 1:22 am America/New_York

Thank you for reverting. I have attached the screenshots of the images and spectral values (in the first post of this issue) where I was facing the issue, though I received positive X & Y pixel values in Seadas. I imported the .hdf5 files in ENVI and was checking the spectral values which were eventually processed in SEADAS. Is there any issue related to .hdf5 conversion or do we need to follow other procedures for exporting to use the images at other software platforms. Kindly check the attachments, if this would help, please let me know the reasons behind the anomaly.

Regards
Sunanda

gnwiii
Posts: 691
Joined: Fri Jan 29, 2021 5:51 pm America/New_York
Answers: 2

Negative sprectral values of processed MODIS L0/L1A data

by gnwiii » Tue Apr 16, 2019 7:21 am America/New_York

Using ENVI you seem to be viewing unscaled data.   The scaling coefficients should be in the hdf5 file, but hdf5 is a storage format so lacks metadata standards.  The netcdf4-cf format is essentially a subset of hdf5 plus metadata standards that include scaling and missing value codes.   In the good old days the scaling factors had to be published separately for each geophysical variable, but with netcdf4-cf they can differ from one file to the next and the software will use the values from each file.  Current versions of ENVI should be able to apply scaling and missing value handling when it reads a netcdf4-cf file, so you may find it easier to work with netcdf4-cf format data.   Current versions of ENVI should allow you to view the metadata in both netcdf4-cf and hdf5 files.  You could try to find the scaling factors (usually names that inlude "scale" and "offset") in the hdf5 file but you have to guess at the equation: y = scale*(x+offset) or y = scale*x + offset.

Post Reply