Causal attribution of satellite-observed vegetation browning to grazing pressure versus rainfall deficit: a climate-analytics instrumental-variable estimation for rangeland early warning

Satellite records of vegetation greenness across rangelands confound two drivers that early-warning systems must separate: rainfall deficit and grazing pressure. The framework treats rainfall as endogenous to vegetation state and instruments it with ocean-atmosphere teleconnection indices, while grazing pressure is instrumented with livestock-price and herd-cycle variables. Formal conditions for point identification are stated, and a bound on the bias of the naive ordinary-least-squares attribution is derived. The Lagos series yields a mean temperature of 26.414 °C and a warming trend of 0.1553 °C per decade, which the framework uses to calibrate the rainfall-deficit channel and to demonstrate the magnitude of confounding that uncorrected attribution would incur. Attribution without explicit instrumentation systematically overstates grazing pressure in semi-arid rangelands, and that early-warning systems which report a single browning signal without decomposing it will misdirect pastoralist and veterinary interventions.