Robustness of Climate-Risk Screening for Maize: Indicator Choice and Event-Year Sensitivity in Heilongjiang, China

Reliable climate-risk screening can inform agricultural decision-making, but the robustness of screening rankings is distinct from adaptation effectiveness or sustainability performance. We assessed the robustness of maize climate-risk screening across 13 prefecture-level units in Heilongjiang, China, during 2011–2024 (N = 182). Reported maize yield was verified against the corresponding Heilongjiang Statistical Yearbook tables and combined with cropland-weighted precipitation, climatic water deficit (CWD), and Palmer Drought Severity Index (PDSI). Fixed-effects models quantified pooled conditional associations using cluster-robust inference and a fully enumerated restricted wild-cluster bootstrap-t, while descriptive rankings were compared across indicator definitions, temporal adjustments, and synchronized year deletions. Seasonal climate–yield associations were generally imprecise; for example, a one-within-unit-standard-deviation increase in precipitation corresponded to a 0.43% yield difference (95% CR1 interval, −0.79% to 1.66%). Although CWD and PDSI rankings showed relatively high aggregate agreement, synchronized year deletion produced median rank spans of five to six positions, and recalculation of percentile-based event membership confirmed substantial local sensitivity. These findings show that apparently consistent climate-risk rankings can remain unstable when based on short administrative records. By identifying where screening results are sensitive to indicator choice and record composition, the proposed robustness-oriented framework can support more transparent climate-risk screening and help determine when local rankings should be treated as provisional inputs to decision-making. The analysis evaluates the robustness of the screening procedure rather than adaptation outcomes or sustainability performance.