MUR-L4 SST: Persistent Striping Artifacts near the Brazil Current

[yuzuru]

I observe strong horizontal striping artifacts in MUR_L4 SST and SST gradient fields near the Brazil Current (approx. 12°S–28°S, 27°W–42°W). The striping appears almost daily. And these areas are flagged as quality level = 1 (good data).

I would like to know why these striping artifacts occur. Is this a known issue in the MUR_L4 product, and can it be technically mitigated or corrected in future versions?

Attached is an example (2010-09-23).

https://worldview.earthdata.nasa.gov/?v=-78.64821285045814,-42.40743885885213,-12.516134871357451,-9.858056728513507&l=Coastlines_15m,GHRSST_L4_MUR_Sea_Surface_Temperature&lg=false&t=2024-02-03-T08%3A06%3A25Z

[yuzuru]

Correct: 1 represents the open-sea, instead of meaning good data

[miawang]

Hi all, I can confirm a similar issue in MUR_L4 SST: the gradient maps show oblique, along-track banded/striping artifacts that appear non-physical. What’s notable in my case is that the collocated MODIS swath SST for the same day does not show these stripes, suggesting the pattern is unlikely to be a real oceanic front.

I have observed this on multiple dates/regions as well, and I’m attaching:
the MUR_L4 SST image showing the stripes, and the corresponding MODIS SST for the same date/time window, where the stripes are absent.

[PODAAC - edward.m.armstrong]

Hello
Thanks for the info on the issues with MUR. I'm the lead POC for this dataset and help maintain the production system and software.

I do agree that there is striping in many of the MUR images I have reviewed both recent and in the Sep 2010 time period from the original poster. But it appears to wax and wane. When I do see strips in the Brazil Current region I dont necessarily see them in other parts of the world's oceans. It could be an artifact in the high resolution SST data from MODIS Aqua or Terra, or even the AVHRR data used. I'll have to dig a little deeper into the data sources for specific days.

[PODAAC - edward.m.armstrong]

Empty

[PODAAC - edward.m.armstrong]

@miawang do you have the exact dates for the MODIS L3 SST and MUR SST images you posted ?

[miawang]

PODAAC - edward.m.armstrong wrote:
> @miawang do you have the exact dates for the MODIS L3 SST and MUR SST images you posted ?

Hi Edward, the exact date for both the MODIS L3 SST and MUR SST images I posted is January 13, 2022 (20220113). Hope this helps with the investigation.

[PODAAC - edward.m.armstrong]

Hi all

Using this Worldview URL I was able to research the different SST layers:

https://worldview.earthdata.nasa.gov/?v=-76.23076163791865,-44.67172589734214,-37.46569352384574,-24.964707453190115&l=Reference_Labels_15m(hidden),Reference_Features_15m(hidden),Coastlines_15m(hidden),OrbitTracks_Aqua_Descending,GHRSST_L4_MUR_Sea_Surface_Temperature(hidden),MODIS_Aqua_L2_Sea_Surface_Temp_Night(hidden),MODIS_Aqua_L3_SST_MidIR_4km_Night_Daily,MODIS_Terra_Brightness_Temp_Band31_Night,OCI_PACE_True_Color(hidden),VIIRS_NOAA21_CorrectedReflectance_TrueColor(hidden),VIIRS_NOAA20_CorrectedReflectance_TrueColor(hidden),VIIRS_SNPP_CorrectedReflectance_TrueColor(hidden),MODIS_Aqua_CorrectedReflectance_TrueColor(hidden),MODIS_Terra_CorrectedReflectance_TrueColor(hidden)&lg=true&t=2022-01-13-T22%3A15%3A27Z

By toggling between the MUR L4, MODIS Aqua L2, and MODIS Aqua L3 layers you can see the L2 layer has the strips. The L2 is used in the L4 fusion algorithm so they appear in MUR. The L3 is a 4 km binned dataset so it does not exhibit much "stripping". But L3 is not used in L4 MUR.

I'm glad we got to the bottom of this. Something to note and try to correct once MUR has a reprocessing campaign.

Thanks for you contributions and pointing out the issue with MUR.