🌾 Agricultural Forecasting API

🌾 Agricultural Forecasting API#

FastAPI-based ASGI service for crop disease risk forecasting using multi-source weather data, developed by the University of Wisconsin–Madison Data Science Institute and integrated with Open-Lambda technology.

Overview#

The API provides geospatial agricultural intelligence for Wisconsin, combining weather data with validated agronomic models.

Key Features#

  • 🌽 Crop disease risk forecasting (corn & soybean)

  • 🌱 Winter rye biomass estimation

  • 🌦 Multi-source weather integration (IBM EIS, WiscoNet)

  • 📍 Coordinate and station-based queries

  • 🗺 GeoJSON outputs for GIS applications

  • ⚡ Async batch processing for multi-station analysis

System Architecture#

The system is structured into four main layers:

  • API Layer (FastAPI)
    Handles incoming requests and routing (/ibm, /wisconet_g, /models)

  • Data Layer

    • IBM EIS: high-resolution global weather API

    • WiscoNet: Wisconsin mesonet station network

  • Processing Layer
    Weather normalization, unit conversion, GDD calculation, rolling features

  • Model Layer
    Disease risk models and winter rye biomass model

Core Modules#

  • Weather ingestion (IBM + WiscoNet)

  • Disease risk modeling

  • Winter rye biomass estimation

  • Async pipeline orchestration

  • GeoJSON response formatting

API Endpoints#

IBM Forecasting (Coordinates)#

GET /v2/ag_models_wrappers/ibm

Returns disease risk + biomass using IBM weather data.

WiscoNet Forecasting (Stations)#

GET /v2/ag_models_wrappers/wisconet_g

Returns station-based time-series disease risk and biomass.

Model Metadata#

GET /v2/ag_models_wrappers/models

Returns available disease and biomass models.

Disease Models#

  • Tarspot (corn) – humidity and temperature-based risk

  • Gray Leaf Spot (corn) – temperature + dew point model

  • Frogeye Leaf Spot (soybean) – GDD + rainfall model

  • White Mold (soybean) – precipitation and soil moisture model

Winter Rye Biomass Model#

Predicts dry biomass (lb/acre) using:

  • Growing Degree Days (0°C base)

  • Planting date (day-of-year)

  • Fall precipitation

  • Logistic growth curve

Outputs#

  • Biomass (lb/acre)

  • Color class (gray / yellow / green)

  • Interpretation message

Data Sources#

IBM Environmental Intelligence Suite (EIS)#

  • High-resolution global weather data

  • Hourly forecasts and historical data

  • Requires authentication

WiscoNet#

  • Wisconsin mesonet (~100 stations)

  • Daily weather observations

  • Public API access

Response Format#

All outputs are returned as GeoJSON FeatureCollections, including:

  • Weather variables

  • Disease risk scores

  • Biomass predictions

  • Station metadata

Performance Features#

  • Async multi-station processing

  • Cached weather and station data (6h–7d TTL)

  • Parallel risk computation

  • Optimized data aggregation pipeline

Setup#

git clone https://github.com/UW-Madison-DSI/ag_forecasting_api.git
cd ag_forecasting_api

python -m venv .venv
source .venv/bin/activate

pip install -r requirements.txt

uvicorn app:app --reload