EloteSat Tamaulipas
Map grain maize and sorghum in Tamaulipas (Mexico) from their phenology: SIAP crop calendars, Sentinel-2 and Landsat downloads, NDVI-curve variables, Random Forest and municipal SIAP maps.
A tool to tell maize from sorghum and other crops in Tamaulipas. Searches Sentinel-2 L2A (Earth Search or Microsoft Planetary Computer) and Landsat 8/9 Collection 2 Level-2 (Planetary Computer, or Earth Search with the requester-pays USGS bucket) through STAC. Checks cover and cloud inside the study-area polygon with SCL / QA_PIXEL before downloading, reads only the window over the study area without resampling, and files each scene in a stage folder (cycle/stage/sensor) from an editable crop calendar. Then: variables in windows relative to each pixel's own NDVI curve (green-up, peak, post-peak, senescence) plus curve-shape metrics on a common grid, calendar profiles per municipality and water regime built from SIAP (Avance de Siembras y Cosechas) with an in-plugin updater, profile assignment from the study area, crop reference windows for interpretation, mapped vs SIAP sown area check, k-means groups without field data, Jeffries-Matusita separability and threshold rules, and hierarchical/flat Random Forest (scikit-learn if present, numpy otherwise) with field-grouped cross-validation. A SIAP maps tab makes styled municipal maps of any Avance variable (sown, harvested, lost area, production, yield, sowing and harvest dates) with an A4 print layout, and a Help tab ships the user manual (PDF). Works only in Tamaulipas. Interface in Spanish and English. Formerly FenoSat (up to 0.4.0). No mandatory extra Python dependencies.