Resampling

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
naveedbuttar
Posts: 8
Joined: Wed Apr 18, 2018 11:07 am America/New_York
Answers: 0

Resampling

by naveedbuttar » Fri Oct 23, 2020 1:34 am America/New_York

Hello,

I want to use SNPP VIIRS and HICO datasets in my research work, I have a few questions below, I really appreciate if you answer me.

VIIRS L2 750m and HICO L2 90m have different resolution, how can resample if I have HICO is my reference dataset?
After resampling the dataset, will it affect pixel data?
After resampling can I compare data HICO L2 vs VIIRS L2 or should I follow other protocols or standards?
What is the use of collocate in the SeaDAS? Can I resample my data using this function?
Best Regards

Naveed

Tags:

OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 269
Joined: Mon Apr 07, 2008 4:40 pm America/New_York
Answers: 0

Resampling

by OB SeaDAS - knowles » Mon Oct 26, 2020 10:17 am America/New_York

You can use the Collocate tool with HICO as the reference/master and VIIRS as the dependent file.  This will generate a new file containing all the raster data bands from both input files.  The pixels of the generated file will geographically match those of the L2 HICO pixel.  Since all data will then be in a single file, comparison tools can then be used.

Alternatively you can use the Reproject Tool or l3mapgen to individually map each file to a consistent matching map resolution. 

Note: sensors (such as MODIS) which have bowtie pixel mapping concerns can be problematic if being directly used as the reference file in the Collocate tool.

Danny

OB.DAAC - SeanBailey
User Services
User Services
Posts: 1464
Joined: Wed Sep 18, 2019 6:15 pm America/New_York
Answers: 1
Been thanked: 5 times

Resampling

by OB.DAAC - SeanBailey » Mon Oct 26, 2020 10:32 am America/New_York

Naveed,

Any resampling will impact the data.  The collocation tool in SeaDAS takes one file as the reference and identifies the data from the dependent source that match the geolocation of the reference.  It can do this via a nearest neighbor, bilinear or cubic spline approach.  It works best with more regular gridded data than with L2 swaths, but can be applied.  A better approach would be to bin the data to the same resolution (using l2bin) and compare either at the L3bin level or a mapped projection of those L3b data (using l3mpagen). 

Sean

naveedbuttar
Posts: 8
Joined: Wed Apr 18, 2018 11:07 am America/New_York
Answers: 0

Resampling

by naveedbuttar » Wed Oct 28, 2020 9:15 pm America/New_York

Thank you Mr.Sean and knowles for your reply, it's very valuable knowledge from you guys, I will continue to bother you guys with my this kind of questions

Post Reply