SST algorithms performances at night

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
madjidh
Posts: 39
Joined: Wed Feb 27, 2019 10:05 am America/New_York
Has thanked: 2 times

SST algorithms performances at night

by madjidh » Wed Dec 15, 2021 9:22 am America/New_York

Hello,

I have read that the SST4 algorithm performs poorer during daylight compared to SST11 due to glint sensitivity around 4 microns mostly. I also read that it is the opposite during night time, with SST4 having better performances compared to SST11, which is why NASA's SST product are SST11 during day and SST4 during night. Please correct me if I am wrong.

I was analyzing the SST SeaBASS data set, for MODIS Aqua, since 2002 to present, and plotted the SST4 and SST11 algorithms during night time (22:00 to 02:00), full data set across the globe. However, when I have a look at the data, SST4 seems to show poorer performances during night compared to SST11. On the figures below, I am comparing in situ values with the value returned by one of the algorithm in attached plots (the temporal window is 30min in SeaBASS, area of 10km if I am right).

1) Was a similar observation made before? I was suspecting a mistake, but I didn't really process the data, I only removed values with negatives, and retained values from 22:00 and 02:00. Happy to share the data if needed.
2) During night, can we have access to some kind of flags despite not having access to visible light?

Best
Attachments
SST4_All_night.jpg
SST4_All_night.jpg (200.09 KiB) Not viewed yet
SST11_All_night.jpg
SST11_All_night.jpg (195.5 KiB) Not viewed yet

Tags:

madjidh
Posts: 39
Joined: Wed Feb 27, 2019 10:05 am America/New_York
Has thanked: 2 times

Re: SST algorithms performances at night

by madjidh » Tue Jan 11, 2022 7:32 am America/New_York

Hello, I am still looking for an answer for this question.

I have one more question that may solve my problem, is the time of the matchups in the SeaBASS SST dataset associated to a specific location (like UTC or Eastern US time) or to the matchup point ?

Cheers

OB SeaBASS - cwproctor
Subject Matter Expert
Subject Matter Expert
Posts: 18
Joined: Mon Feb 05, 2007 12:20 pm America/New_York

Re: SST algorithms performances at night

by OB SeaBASS - cwproctor » Tue Jan 11, 2022 12:38 pm America/New_York

Hello,

The date and time information in the SeaBASS SST dataset (and all SeaBASS data in general) is in GMT/UTC, not local time. Since the dataset is global, if you wish to organize the data by specific local times, you will need to use a method to convert to local time.

FYI in case it is helpful: Note that the SeaBASS SST datasets contain a column called <sensor platform name>_day_night flag (e.g., MODIS_Aqua_day_night_flag) containing text values of "Day", "Night", or "Mixed". Every matchup is assigned one of those, which comes from the global attributes of the relevant satellite file.

madjidh
Posts: 39
Joined: Wed Feb 27, 2019 10:05 am America/New_York
Has thanked: 2 times

Re: SST algorithms performances at night

by madjidh » Thu Jan 13, 2022 7:06 am America/New_York

Thank you for the answer, solving my problem ! I should have checked the header carefully.

Best

Post Reply