Remote Sensing Systems Science Grade SMAP V5 L2C Processing Algorithm

This document was written to illustrate how to run the Remote Sensing Systems Soil Moisture Active Passive (SMAP) Version 5 (V5) Level 2C (L2C) science-grade processing algorithm for sea-surface salinity (SSS) retrievals. The science-grade processing algorithm is meant to be a user-friendly simplification of the full RSS SMAP V5 L2C processing code.

The science-grade code contains both the forward model (i.e., the calculation of 'expected' brightness and antenna temperatures) and the 40-km salinity retrieval algorithm, which is, effectively, and inversion of the forward model. The smoothing to 70-km and the SMAP SSS uncertainty estimation are not contained in this package (although these are both available as part of the release of the full L2C processing code). For more information on the full RSS SMAP SSS V5.0 retrieval algorithm and forward model, please see the NASA/RSS SMAP Salinity Version 5.0 Validated Release ATBD.

This code is written in Fortran and is set up to be complied using the gfortran compiler. It inputs L2B ASCII files located in the sample_data subfolder and will output ASCII files containing relevant variables from the L2C processing. Information on the contents of the input and output ASCII files can be found in l2b_ascii_file_variable_list.md and l2c_ascii_variable_list.md, respectively.

Also contained in the sample_data subfolder are example output L2C ASCII files (these end in output_rss.txt). Users should first check to see that they can replicate these ASCII output files by following the instructions below. Once done, users may edit and adapt the science-grade L2C processing code to their own needs. Note that the data output in these L2C ASCII files are just a small subset of the variables than can be output by the L2C code. For a full list of output L2C variables see section 12.1.4 of the ATBD given above.

Compiling the code

All source files are in the L2C folder, where MAKE_SMAP_L2C_V50_ascii.f90 is the main program and it depends on the remaining .f90 files included.

Meson is used to build the code. Once meson has been installed, the code may be compiled using a build directory named build with the following commands:

meson setup build .
meson compile -C build

The output executable is build/MAKE_SMAP_L2C_V50_ascii.

Setting the directories

At present, there are two folders containing external files that the routines and subroutines of the science-grade L2C code require in order to run. These are the tables_L2C and sample_data subfolders. If users decide to keep the directory structure the same as it is on GitHub, then they may skip this step and move on to running the code. However, if the data located in tables_L2C and sample_data are placed in different locations, these locations must be set as environment variables before running the code. This is done via the following terminal commands:

export SMAP_TABLE_DIR="TABLE_DIRNAME" SMAP_DATA_DIR="DATA_DIRNAME"

Where TABLE_DIRNAME points to the location of the contents in the tables_L2C folder and DATA_DIRNAME points to the location of the contents of the sample_data folder. Note that, if the directory/folder structure is changed, all data from the tables_L2C folder must remain in a folder together and all data from the sample_data folder must remain in a folder together.

Running the code

Once the code has been compiled and the directories have been defined, the code may be run with the following command in the terminal (assuming the code is being run one directory up from the build folder):

build/MAKE_SMAP_L2C_V50_ascii 38362 38362 ##

In this command, 38362 represents the start and end SMAP orbit. This should remain unchanged for the purposes of attempting to replicate the output ASCII files located in the sample_data folder. ## represents the pixel number that the user wishes to run the code for. There are 10 different 0.25x0.25 degree Earth grid pixels that can be used in this demo, each assigned a value between 1 and 10. Therefore, ## should be an integer between 1 and 10. Information on these 10 pixels is given in pixel_descriptions.md.

After running this code, the output L2C ASCII files will be written to the sample_data folder and have the file name sample_l2c_38362_##_output.txt where ## is, again, the pixel number. The output file created by the user should be compared to the .txt file located in the sample_data folder with the name sample_l2c_38362_##_output_rss.txt.