Code and data for manuscript submitted to JGR: Atmospheres

Comprehensive analysis of the NOAA National Water Model: A call for heterogeneous formulations and diagnostic model selection

J. Michael Johnson, Shiqi Fang, A.Sankarasubramanian, Arash Modaresi Rad, Luciana Kindl da Cunha, Keith S Jennings, Keith C. Clarke, Amir Mazrooei, Lilit Yeghiazarian

Data Creation

0. Precomputed

• nhdplus_nwis_mapping.parquet
  • This file stores the GAGE_ID/feature_id relation found with the NLDI
• data/usgs_metadata.parquet
  • This files stores the metadata about NWIS flow records found with dataRetrieval::readNWISdata()

1. Get basins: data/R/01_basin_boundaries.R

• This file downloads the gagesII shapediles from the USGS and joins it to precomputed data (0)
• The output is a 680MB geopackage called basins.gpkg

2. Get flows: 02_flow_comparisions.R

• This file uses the GAGE_ID/feature_id sets in basins.gpkg to query daily timeseries data from a NWIS and NWM2.0 NetCDF file.
• All stats are computed based on the comparison of these series
• The stats are joined to the basins and re-exported to the basins.gpkg

3. Get NLDAS data: 03_nldas_data_download.R

• This file downloads all requisite NLDAS data from the NASA GES DISC based on the file data/subset_NLDAS_FORA0125_M_20230907_224958_.txt
• Directions for generating that file are in the notes of the R file, however, it should work for everyone as is.

4. Process NLDAS data: 04_nldas_data_processing.R

• This file prepares basin summarized NLDAS data
• The stats are joined to the basins and re-exported to the basins.gpkg

5. Process NLCD data: 05_nlcd_data_processing.R

• This file prepares basin summarized NLCD data
• The stats are joined to the basins and re-exported to the basins.gpkg

6. Process NID data: 06_nid_data_processing.R

• This file prepares the basin summarized NID data
• The stats are joined to the basins and re-exported to the basins.gpkg

7. Export Table: 07_table_export.R

• This simply reads the basins.gpkg and exports the non-spatial data as a parquet file. This file is called data/v20_summary_stats.parquet and is used in all image creation (see section called Create all images)

7. Format NextGen outputs: 08_ngen_experiment.R

• This file re-formates the data for the NextGen experiment from CSV files produced by co-authors from ngen simulation.

Key Data Artifacts

The key artifacts used in the analysis of this paper are: - data/v20_summary_stats.parquet - this is the basins.gpkg stripped of geometry and saved as a parquet file

Create all images

Run this from the terminal (in this directory):

bash ./make_images.sh

All images will be generated in the ./images/png directory.