Does Temporal Mismatch Between Species Occurrence Records and Climate Data Affect the Reliability of Ecological Niche Models for Cereal Stemborers in Kenya?

ABSTRACT Species distribution modelling (SDM) is a central tool in ecological forecasting and biogeography studies, yet uncertainty remains regarding the optimal temporal scale of predictor variables. We investigated how predictor temporality influences SDM performance using two economically significant cereal stemborers, Busseola fusca and Chilo partellus, across different agroecosystems in Kenya. Models were developed under two scenarios: (1) long‐term climate baselines (1970–2000) and (2) climate data temporally aligned with species records (2001–2006), while using identical topographic and vegetation predictors. Contrary to our hypothesis, the model with temporally aligned predictors (scenario 2) did not outperform the one based on long‐term baseline (scenario 1). Internal cross‐validation showed no significant differences in mean area under the curve (AUC) or true skill statistic (TSS) ( p  ≥ 0.05). However, marginal differences were observed for B. fusca (0.86/0.71 vs. 0.87/0.69) and C. partellus (0.76/0.51 vs. 0.75/0.52). External validation consistently favoured scenario 1, with higher precision and F 1‐scores, particularly for C. partellus (0.88/0.93 vs. 0.83/0.88; p  = 0.024). These findings indicated that long‐term climatic baselines better capture stable species‐environment relationships. More broadly, our results highlighted the significance of predictor temporality as a source of uncertainty in SDMs and underlined the importance of combined internal and external assessments to enhance good modelling practices (GMP). Such rigour is essential for reliable ecological forecasting and for developing climate‐informed Integrated Pest Management (IPM) strategies for cereal stemborers. The model outputs from our study highlighted hotspots of B. fusca and C. partellus habitat suitability across Kenyan agroecosystems, providing a foundation for targeted and effective management strategies.