Climate variability shapes dengue transmission intensity differently across urban contexts in Selangor, Malaysia

Abstract

Dengue transmission is strongly influenced by climatic conditions that regulate mosquito population dynamics and feeding behaviour. This study examines how climate variability influences mosquito biting rates and dengue transmission intensity across two districts with distinct urbanization profiles in Selangor, Malaysia, namely the highly urbanized Klang district and the more environmentally heterogeneous Sepang district. For this purpose, weekly mosquito biting rates from 2019 to 2024 were inferred from reported dengue incidence using a Susceptible-Infected mosquito vectors, Susceptible-Infected-Recovered human hosts (SI-SIR) model, yielding estimates ranging from approximately 0.4 to 1.8 per week. These inferred rates were subsequently related to lagged climatic variables (average temperature, relative humidity, and accumulated rainfall) using generalized additive models (GAMs), with an additional seasonal smooth term to capture intra-annual variability. Optimal lag structures indicated delayed climatic effects operating on 1–4 week timescales, with relative humidity consistently exerting a strong and nonlinear influence (effective degrees of freedom up to 4.6) in both districts. Average temperature showed a weak but robust negative association with biting activity, while rainfall displayed a stable positive effect in highly urbanized Klang but a less robust influence in Sepang. Seasonal forcing emerged as a dominant driver of short-term variability, particularly in Klang. The optimal GAMs explained 36.5% and 27.6% of the deviance in biting rates for Klang and Sepang, respectively, reflecting differences in the relative importance of climatic versus non-climatic drivers across urban contexts. Sensitivity analyses confirmed that inferred biting-rate dynamics were highly stable, with ± 20% perturbations in key epidemiological parameters altering mean biting rates by less than 12% and preserving temporal patterns (correlation > 0.995). Overall, the results demonstrate that delayed and nonlinear climatic effects, particularly relative humidity, play a key role in shaping mosquito biting rates and dengue transmission intensity, with their strength and lag structure differing across urban contexts in Selangor.