RESTORE/makESTUSAL, Source code for construction of various statistical models and prediction of daily salinity in coastal regions of the Gulf of Mexico, United States

The RESTORE/makESTUSAL software repository contains R language source code that can be used for the construction of various statistical models and output time series of predicted daily salinity coastal regions of the Gulf of Mexico, United States. The source code is expansive, and the repository is organizationally deep following logical organization units. One major subsystem of the repository is the documentation of what is termed "full scale" machine learning and subsequent predictions and uncertainties of daily salinity by looping through 15 salinity groups (coastal subareas) with group-specific lists of daily salinity monitoring sites. There are a total of 91 salinity sites across the groups. Covariates used for prediction of salinity include features such as streamflow from daily climatology, daily streamflow, and daily wind components. A second major subsystem of the repository is a catalog of detailed statistical model testing and method comparison with as "companion software" supporting interpretative study of specific salinity sites supporting publications.