SIMCAST Spatial: A Python Pipeline for Gridded Climate Processing, Point Masking, and Distributed Late Blight Modeling

This dataset contains SIMCAST Spatial, a Python-only pipeline for spatial late blight modeling that processes gridded climate datasets and runs the SIMCAST model on a point grid at scale using Dask Distributed. The workflow supports: (1) resampling precipitation NetCDF rasters to a target grid, (2) masking/extracting monthly NetCDF values onto a point grid and writing monthly Parquet files, (3) running SIMCAST per point/season to produce per-region/year results and optional application (APP) rasters, (4) generating BU/FU raster summaries and an “all lands” animation, and (5) optional trend significance analysis (OLS over years) for APP/ABU/AFU. The pipeline is configured and executed via main.py and config.py, and is currently tuned to a Peru potato-zoning workflow but intended to be reusable by passing paths and resource settings. This deposit includes only software; input data (NetCDFs, point grids) are provided by the user and are not included.