Screening Potential Control-Ring Interference in Remote-Sensing Assessments of Cascade Reservoir Microclimates: A Cautionary Case from Yunnan, China

Remote-sensing assessments of reservoir microclimates often compare areas near reservoirs with surrounding control zones, which should be both unexposed and environmentally comparable. We evaluated these requirements in a clustered cascade system using geometric screening scenarios and an illustrative outcome analysis. For 62 reservoirs in Yunnan, China, 96.8% of primary 20–50 km control rings intersected at least one neighboring reservoir’s 4× surface-water-area (SWA)-scaled proxy buffer; the rate was 88.7–98.4% across 2×–5× radius scenarios and 100% for two wider rings. Because these buffers are unvalidated geometries, overlap is interpreted only as a warning, not as evidence of physical spillover. In a 34-reservoir case study (2000–2024), the original-ring raw maximum-temperature (TMAX) estimate was −0.048 °C (95% CI [−0.073, −0.023], p = 0.00042). It changed little after proxy-overlap pixels were excluded (−0.044 °C) but attenuated after elevation–slope balancing (−0.025 °C, 95% CI [−0.053, 0.002], p = 0.065) and after combined exclusion and balancing (−0.023 °C, 95% CI [−0.049, 0.004], p = 0.092). Detrended estimates were non-significant under all four controls, and the other six outcomes were not specification-stable. Date-precision scenarios preserved the negative raw direction but did not resolve trend sensitivity. The case therefore demonstrates control-design fragility rather than a definitive reservoir-cooling effect. Geometric exposure screens should be paired with terrain diagnostics, alternative controls, and explicit uncertainty reporting before causal interpretation.