Relationships between meteorological factors and mosquito vector density in an Urban District of Shanghai, China (2018-2024).

Weather-based early warning systems for mosquito-borne diseases require quantitative understanding of the non-linear and delayed effects of meteorological factors on vector dynamics, yet these remain poorly characterised for Aedes albopictus in subtropical urban settings. We analysed weekly Mosquito Oviposition Index (MOI) data from 64,340 ovitraps deployed in Shanghai during 2018-2024, using Generalised Additive Models and Distributed Lag Non-Linear Models to quantify lag-specific temperature effects over 4 weeks. Univariate Generalised Additive Models (GAMs) identified six significant meteorological predictors (p < 0.01), but multivariate GAMs retained only precipitation as independently significant (p = 0.008), with minimum temperature marginally so (p = 0.064). DLNM confirmed a significant overall temperature effect (F(9, 18) = 4.727, p = 0.0025), revealing a unimodal pattern: positive cumulative effects at 14-20°C, peaking at ~18-20°C and negative effects above 23°C. Lag dynamics were asymmetrical: cool conditions (13°C) had immediate positive effects (lag 0-1 weeks), whereas warm conditions (25°C) produced delayed negative effects (lag 2-3 weeks). Minimum temperature exerts significant non-linear, delayed effects on Aedes albopictus oviposition in urban Shanghai, with an optimal range of ~18-20°C and distinct lag-specific patterns. These quantitative estimates support the development of temperature-based early warning indicators for vector surveillance.