Agroclimatic risk of Andean crops under CMIP6 scenarios in Imbabura, Ecuador: A reproducible Random Forest-Bayesian Network pipeline

This dataset contains the complete computational pipeline (22 Python scripts, 10 phases) and all outputs for assessing agroclimatic risk of four Andean crops (potato, maize, bean, quinoa) across 42 parishes of Imbabura Province, Ecuador, under CMIP6 climate scenarios (SSP1-2.6, SSP3-7.0, SSP5-8.5) for 2021–2040, 2041–2060, and 2061–2080. Methodology integrates: (1) Random Forest Species Distribution Models on BASD-CMIP6-PE climate projections, (2) 7-node Bayesian Network integrating Hazard × Exposure × Vulnerability (IPCC AR6 framework), generating 1,512 Risk Index values, 58 cartographic maps (300 DPI), and 42 parish technical sheets. Key results: García Moreno IR=0.689 (highest risk, SSP5-8.5, 2061–2080); Imbaya IR=0.350 (lowest); potato most vulnerable (IR≈0.596); quinoa largest relative increase (+46.6–46.8%); bean most climate-stable (ΔIR≤+0.5%); AUC 0.804–0.871; total agricultural exposure 10,706.1 ha. GitHub repository: https://github.com/VICTORSIG1985/agroclimatic-risk-imbabura