Causal Transport Inference for Computer Agriculture Irrigation Advisories Under Cross-Region Covariate Shift: An Artificial Intelligence Sensitivity-Bound Evaluation

Irrigation advisories trained on one agro-ecological region and deployed in another inherit a causal estimand that may not survive the journey. the present analysis treats cross-region transport of an irrigation-advisory effect as a sensitivity problem rather than an estimation problem, and asks how much unmeasured effect modification a before-after-control-impact design can tolerate before its transported recommendation reverses. The design's operating characteristics are characterised under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64, across cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points, with 2000 replications per design cell. Power curves, minimum detectable effect curves, coefficient intervals, a scatter fit, a kernel density of estimated power and a residual diagnostic are reported. Fifty-two clusters are required for 80 per cent power at a 5 percentage point effect. The contribution is a transport-sensitivity bound that converts a design's detectable effect into a maximum admissible covariate shift, so that an advisory's cross-region deployment can be licensed or refused on stated assumptions rather than on point estimates alone.