Codes for the manuscript of prescreening-based subset selection for improving predictions of Earth system models for regional environmental management of red tide

Subset selection

This repo contains the codes for the manuscript of "prescreening-based subset selection for improving predictions of Earth system models for regional environmental management of red tide." The folder Subset_selection contains two codes for processing the data and figures in the manuscript. Check the Jupyter notebook DataVisualization_zos for Figs. 1 and 2, and Jupyter notebook SubsetSelection.ipynb for Figs. 4-7. Interactive versions of these two notebooks are c1.ipynb and c2.ipynb, respectively, which can be found under folder i. Using colab, the interactive notebooks can be executed by anyone, who has a web browser and Google. Thus, these interactive versions make the codes immediately reproducible and thus can be executed with different opitions along with reading the manuscript.

The folder Karenia_brevis_data_analysis contains the codes for processing the Karenia brevis data.

The folder zos_data_extraction contains the codes for preprocessing zos data (sea surface height above geoid). These are mainly data crunching tasks such as hyber-slabing, and stacking using the netCDF Operators (NCO, http://nco.sourceforge.net/). This is followed by data extraction along 300 m Bathymetry (B300) in the study area. This is mainly to reduce the size of the Earth system models (ESMs) data and reanalysis data from 80 GB to about 10 MB for the ESMs data zos_data_B300_ESMs43210.npy and less than 1 MB for the reanalysis data zos_data_B300_Reanalysis10_phy001_030_r1.csv. These two files are used as input files for SubsetSelection.ipynb and its interactive version c2.ipynb.