Comparison of MODIS/VIIRS R2018 OCX

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
pgarnesson
Posts: 13
Joined: Mon Dec 14, 2009 11:04 am America/New_York
Answers: 0

Comparison of MODIS/VIIRS R2018 OCX

by pgarnesson » Tue Jan 09, 2018 9:03 am America/New_York

We have done a first R2018 assessment to compare VIIRS & MODIS OCI chlorophyll for the last days.
We observe a bias with VIIRS < MODIS.
At global level, 75% of MODIS pixels exceed VIIRS of more than 10% (relative difference threshold),
(see attachment showing the relative difference map for the last nday).
For previous years this trend exists but lower.
In the meantime, the 5th of January, some MODIS LUTs have been switched to new LUTs (Vxx.xxx.xx.xx), so maybe this first assessment is obsolete?
By the way, is there any change about coefficient of OCX/OCI in R2018?
Thank you in advance for your support.
Philippe

Tags:

OB WebDev - norman
Subject Matter Expert
Subject Matter Expert
Posts: 142
Joined: Tue Feb 09, 2021 8:19 am America/New_York
Answers: 0

Comparison of MODIS/VIIRS R2018 OCX

by OB WebDev - norman » Tue Jan 09, 2018 11:13 am America/New_York

Hi Philippe,

I leave it to others in our group to speculate on causes, but I can confirm the
bias you are seeing.  The following are plots of the median global chlor_a values
taken from the daily level-3 bin files in our archive.

Aqua/MODIS:
https://oceancolor.gsfc.nasa.gov/cgi/l3bts?mssn=MODISA&prod=chlor_a&stat=MA&filt=60

Suomi-NPP/VIIRS:
https://oceancolor.gsfc.nasa.gov/cgi/l3bts?mssn=VIIRS&prod=chlor_a&stat=MA&filt=60

Regards,
Norman

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

Comparison of MODIS/VIIRS R2018 OCX

by OB.DAAC - SeanBailey » Tue Jan 09, 2018 11:15 am America/New_York

Philippe,
First I will say that the MODIS LUT updates will be coming more regularly.  The MODIS-Aqua instrument degradation continues and is less predictable than we'd like.
The R2018 data are only considered "refined" through mid Oct 2017, as it takes time to evaluate the degradation and implement LUT updates to account for the changes seen.
We will be making more regular LUT changes and slowly advancing the "refined" data in time, but there will be a 60-90 day (or more) lag. 

With that in mind, yes, your analysis may be premature.  However, we're seeing the same thing:
https://oceancolor.gsfc.nasa.gov/analysis/global/vt66_at155/plots/vt66_at155_olig_chl_ratio.png (VT66 is effectively R2018 for VIIRS-SNPP, AT155 is for MODIS-Aqua - but does NOT include the latest LUT)

Welcome to the world of aging instruments - Aqua will have her 16th birthday this year and VIIRS hit 6 years late last year.  We will continue to do what we can to track and correct for instrument degradation, but the problem is becoming increasingly difficult to tackle.

Oh, and, no, there were no algorithmic changes with this reprocessing, so the OC coefficients are not different.

Regards
Sean

pgarnesson
Posts: 13
Joined: Mon Dec 14, 2009 11:04 am America/New_York
Answers: 0

Comparison of MODIS/VIIRS R2018 OCX

by pgarnesson » Thu Jan 11, 2018 4:05 am America/New_York

Thank you Sean for these clarifications.
About Refined MODIS product it is said in a recent annoucement:
"The new approach means the "Quicklook" status may be retained  for up to 90 days from collection.".
- It means Refined after the 15 October should be updated in the coming days?
- I suppose Refined will be no more delivered on a regularly daily basis but extended in one step for a given period?
- About VIIRS Refined, no plan to change the today approach?
Regards, Philippe

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

Comparison of MODIS/VIIRS R2018 OCX

by OB.DAAC - SeanBailey » Thu Jan 11, 2018 8:32 am America/New_York

> - It means Refined after the 15 October should be updated in the coming days?


Yes.  The evaluation of the next LUT update is actively underway.  Once finalized, we
will process the data from 15 Oct through mid Nov (I believe the expectation is 11 Nov) as refined.

> - I suppose Refined will be no more delivered on a regularly daily basis but extended in one step for a given period?


It will be regular, but a bit staccato - with approximately monthly slugs of data with each update.   There will not be a
(nominally) fixed lag like there was previously

> - About VIIRS Refined, no plan to change the today approach?


...not yet, but I wouldn't be surprised if it eventually requires the same level of calibration "attention" as MODIS.

Sean

pgarnesson
Posts: 13
Joined: Mon Dec 14, 2009 11:04 am America/New_York
Answers: 0

Comparison of MODIS/VIIRS R2018 OCX

by pgarnesson » Tue Jan 16, 2018 8:28 am America/New_York

Dear Sean,
about L2 MODIS,
some refined products are now available until the 21 Dec.
This is fine, but in the same time we have observed that there are still  QL recently generated for October.
ncdump -h A2017288235500.L2_LAC_OC.nc (uploaded using https://oceandata.sci.gsfc.nasa.gov/cgi/getfile)
      :pversion = "2018.0QL" ;
      :date_created = "2018-01-12T00:58:17.000Z" ;
We have also observed QLP products. Please can you clarify what will be the strategy/timeliness for QLP & QL.
About L3 MODIS, same question for refined, QL, QLP
Thank you in advance.
Philippe

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

Comparison of MODIS/VIIRS R2018 OCX

by OB.DAAC - SeanBailey » Tue Jan 16, 2018 10:31 am America/New_York

Philippe,

Any refined data post 15 Oct 2017 were run prior to the implementation of the new scheme.  I'll ask the production folks to rerun those to correct the designation.

The files with QLP in the version are the preliminary quicklook files - these are made with whatever ancillary data is available at the time the data are ingested, which may even be climatology.  Once the best-available ancillary data are acquired, the files are reprocessed as "QL".  Once the refined calibration is available, the files will be reprocessed and will no longer be designated as quicklook.  This applies for L2 and L3 products.  In the past, the longer period L3 products (8-day and above) were only made once the refined L2s were available.  As the refined L2s will now take longer to be generated, we will produce "quicklook" L3s through the monthly products.

Regards,
Sean

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

Comparison of MODIS/VIIRS R2018 OCX

by OB.DAAC - SeanBailey » Wed Jan 17, 2018 10:59 am America/New_York

> Any refined data post 15 Oct 2017 were run prior to the implementation of the new scheme.  I'll ask the production folks to rerun those to correct the designation.


It had already been done :grin:
Let us know if you see any files post 01 Nov 2017 (as of now, the new refined date) that are showing up as refined.

Regards,
Sean

Post Reply