Earthdata Forum

I was recommended to repost a question I asked on the podaac github page (https://github.com/podaac/tutorials/issues/229#issuecomment-2204063649).

This is my first time working with SWOT data, and therefore possibly a naive question, however, I have the following data processing problem:

The Level 2 SWOT data from Earthdata, as well as the Level 3 SWOT data from AVISO comes with the xarray dimensions [num_lines, num_pixels, num_sides] with the coordinates latitude/longitude (num_lines, num_pixels).

I would like to be able to regrid the available data to a regular time/lat/lon grid such that it can be queried or merged with other available gridded products like the ones available on Copernicus, that come with dimensions (time, latitude, longitude)

Screenshots, of the xarray Datasets can be seen in the above linked github issue. I would welcome any input and suggestions regarding tackling this data processing issue or possibly pointing me to xarray routines or scripts that could achieve this. Thanks in advance.

Hello @n@nilsleh
My apologies that I do not really have a solution for you, but I would like to comment that the development of a (time, latitude, longitude) gridded product, and the merging of that product with others can become a heavy undertaking depending on how 'good' you want it to be and what usage it is for. For example, what grid size are you looking for, is nearest neighbor interpolation good enough for your use, and what kind of data analysis will you perform?
Perhaps the routines/scripts of researchers who take satellite L2 swaths and interpolate them for comparison with in-situ data or other regional project analyses will be useful to you.

Thank you for the reply. I am aware that a product like the https://cds.climate.copernicus.eu/cdsapp#!/dataset/10.24381/cds.4c328c78?tab=overview Copernicus Sea Level product is a substantial effort merging and interpolating different altimeters into a single aggregated product.

I am not necessarily interested in a perfect gridded product, just having a gridded product would make analysis easier I suppose. Ultimately I am interested in something similar to what you mention in the last sentence (comparing L2 observations with other data sources). For example take a certain lat/lon extent of the above mentioned Copernicus Sea Level product at a specific timestamp, then for this "patch" I would be interested in all available SWOT Level 2 observations covering that same spatial extent at roughly the same time (something like nearest neighbor would suffice). Having some gridded version of the L2 Product would just make data analysis by indexing through xarrays more straightforward. However, I would assume that other people are interested in this sort of data comparison of given spatial extent and time and therefore, I am wondering how they approach this given the L2 Product that is available for downloading. I would look for the highest possible grid size, so ideally 2km x 2km.