saifeldinkhedir-coder/sudan-irrigation-monitor

Multi-sensor Google Earth Engine engine for Sudan's irrigation schemes: canal water, head-to-tail equity, crop condition and water requirement — with scientific-integrity rules enforced in code, not promised in docs. # Sudan Irrigation & Agriculture Monitor A multi-sensor Google Earth Engine analysis engine for irrigated agriculture in Sudan, answering two questions usually asked separately: - **FIELD** — how is this field doing, and is it short of water? - **NETWORK** — did the water actually arrive, and was it shared fairly from source to farm? The network layer — and specifically the **head-to-tail equity** figure — is the uncommon part, and it is where a gravity-fed scheme's real problem lives. See `docs/STRATEGY.md` for the build plan, the validation design, the feasibility split, and the stage-by-stage status table (§8). ## Layout ``` src/ decision_logic.py pure, ee-free decisions (thresholds, Otsu, the head-tail slope fit + CI, the nitrogen gate, agreement). The tested heart of the integrity rules. engine.py network + field layers, EE plumbing, and the assembly (analyse) that wires everything together. nutrition_climate_ground.py red-edge nutrition ladder, calibration store, climate (server-side dry spells / season history), ground-observation store, drone roadmap. attribution.py Stage 3: stratified control for soil/crop/planting date, adjusted gap + CI, placebo test, green-up, persistence, negative controls. agri_engine.py THE AGRICULTURE ENGINE - fields in, farm report out. Imports nothing from engine.py. farm_cli.py agriculture entry point (--fields only). network.py continuity (where the water stopped), siltation candidates, water-use efficiency + its refusal, radar resolvability qualifier. agronomy.py FAO-56 ET0, Kcb from NDVI, crop water requirement, irrigation requirement, GFS outlook, yield gate. rangeland.py rangeland productivity + timing, water points, corridors — with the neutrality guard enforced. farm_records.py REPORTED-side store (operations, costs, sales, margins) and the rule-based advisory. farmer_channel.py …