SST

[sigridurvala]

I am trying to obtain satellite sea surface temperature data around Iceland and I have found the data on the ocean color map website. The problem I am having is that I need to download the data from each day from the past 22 years (since 2002) and I can only seem to download a file from one day at a time. Is it possible to obtain all the data in one file? (All the temperature data available from 2002 to 2024 within certain coordinates)

[PODAAC - wenhaoli]

Hi,

Thank you for your inquiry. Could you provide us more detail information about the dataset? It would be helpful to know the dataset short or long name, plus the web links. There are many satellite sea surface temperature products archived at the PO.DAAC, see https://podaac.jpl.nasa.gov/cloud-datasets?search=ghrsst.
We recommend to download the data with podaac-data-download script (https://github.com/podaac/data-subscriber). Please let us know if you have more questions.

Best Regards,

[sigridurvala]

The dataset short name is MODIS_A-JPL-L2P-v2019.0. Is it possible to download the data using R, or is it only possible with Python? And if it is only possible with Python, do you know if I can get the data through Python and then save it as a file that I can use in R?

[PODAAC - wenhaoli]

Hi,

We recommend to download the MODIS_A-JPL-L2P-v2019.0 data with the podaac-data-downloader tool, see an example of the command.
podaac-data-downloader -c MODIS_A-JPL-L2P-v2019.0 -d ./data --start-date 2002-07-04T00:00:00Z --end-date 2002-07-11T00:00:00Z -e ""
The SST data is in netCDF format. PO.DAAC provides a generic reader tool in R to read the data. Here is the link to the tool: https://github.com/podaac/data-readers/tree/main/r.

[mdsumner]

R can lazy load from netcdf via URLs in the usual ways, here the {terra} package leverages GDAL to do the work (I'm not using your dataset exactly but happy to rejig if you want to share specifics):


```
library(terra)
url <- "https://archive.podaac.earthdata.nasa.gov/podaac-ops-cumulus-protected/MUR-JPL-L4-GLOB-v4.1/20240318090000-JPL-L4_GHRSST-SSTfnd-MUR-GLOB-v02.0-fv04.1.nc"

## make a GDAL dsn for an online file
dsn <- sprintf("/vsicurl/%s", url)

## we need our earthdata token set (how you input the token is an open topic, maybe use an env var with Sys.getenv)
Sys.setenv("GDAL_HTTP_HEADERS" = sprintf("Authorization: Bearer %s", mytoken))

## open it and select analysed_sst
sst <- rast(dsn, "analysed_sst")
```

https://gist.github.com/mdsumner/63274b2069bbdfbc297b350f5d53dd2f

Note the GDAL version requirement, so you might not easily have access to that version (and OS support can differ).

Once loaded by terra you can use `crop()` or `project()` methods to obtain subsets or resample windows or `extract()` to interrogate by point, line, polygon objects.

But, this is a massive dataset and it's challenging to hit it efficiently (I wouldn't want to download entire NetCDF files to interrogate a few pixel values for example). Terra can potentially load multiple urls in one step as above (make a vector of dsns), but whether that's efficient downstream depends on what happens next exactly.

Happy to follow up with discussion and alternatives, we're still exploring this problem.

[sigridurvala]

Thank you for the information. When I download the data with the podaac-data-downloader tool I get many files where each file contains data for a particular time. Is there any way to download the data where I get everything in one file?

[PODAAC - wenhaoli]

As far as I know. Currently, there is no such tool to merge all granules into a single file.

[sigridurvala]

Thank you! I'm new here and I just saw the reply from mdsumner, and I could use some further assistance. What I am trying to obtain is just one specific pixel value, but all of the values in that pixel from every day for the past ~20 years. Since you wouldn't suggest using this method, what would you suggest instead?

Another problem I've been having is that I can only download one timing in one file, meaning I would have to download thousands of files to obtain my required dataset.

I don't know if you can answer my next question but just in case, what is the difference between the data with processing level 2 and level 4? When I download data with processing level 4 it shows the same temperature over a very big area which doesn't make sense, and when I download data with processing level 2, the values are always temperatures that are not possible for the ocean.

Thank you very much!

[GES DISC - jimacker]

Hello,

What spatial resolution do you require? The Giovanni system can provide a time-series of MODIS-Aqua SST data for the length of the mission (starting in mid-2022) using the MODIS-Aqua Level 3 data with 4 km resolution.

Please reply if this would be acceptable and I can explain further.

[GES DISC - jimacker]

Note, Level 4 data refers to data that is a model output of some kind. You would have to specify what dataset you were visualizing to provide a better answer.