This repository is the R code base for the Short-term Drought Forecaster (STDF). The STDF is a tool that predicts weather and soil moisture for user-generated, specific coordinates (lat, long) for the upcoming year.
The application works by evaluating user definition of site-specific attributes (location, soil, vegetation). Location information is used to gather gridded historical weather data and weather predictions for the upcoming year. Soils and vegetation information are needed to simulate soil moisture conditions, alongside weather information.
The application currently returns predictions of weather and soil moisture in both graphical and tabular formats, for both the upcoming year (12 months from current), as well as the historical period (1980 - current).
Weather forecasts for the next 13 months are generated by meteorologists at the National Weather Service for the United States at a regional scale. This tool takes this forecast information and downscales it to the local (i.e. lat/long coordinates) level and runs simulations of the future that represent a range of potential conditions. More information about this process and the logic is here.
To generate soil moisture predictions, this tool runs iterations of the mechanistic water-balance model, SOILWAT2. The source code for this model can be found here. SOILWAT2 is a point based (lat, long) model that requires users inputs of climate, soil texture, and vegetation structure in order to simulate water balance and soil moisture conditions for any time periods in which it is provided climate data.
Application source code was written exclusively in R.
The STDF is currently deployed as an API using the plumber package.
Functionality for soil-water modeling relies on the R package: rSOILWAT2, a R plugin for the C based code of SOILWAT2.
This app is containerized in docker-compose and hosted on an AWS EC2 instance.
Details about deployment are contained in the repository's cloud formation script.