Abstract
Predicting invasive ectotherm responses to climate change requires understanding how population-level variation and diurnal temperature fluctuations shape thermal performance, and whether constant-temperature measurements accurately capture performance under fluctuating conditions. Yet many models, including those used to predict mosquito distribution and disease transmission, still rely on constant-temperature measurements. We addressed both questions in the invasive mosquito species Aedes albopictus by rearing first filial offspring (F1) from three European populations spanning a 1,200 km latitudinal gradient (Frankfurt–Rhine-Main, Germany; Trento and Palermo, Italy). We tested F1 performance under five mean temperatures (13–33°C) exposing them to constant and fluctuating (±5°C) thermal regimes, and quantifying larval survival, larva-to-adult development time, wing length, and adult longevity. The thermal regime dominated all immature traits: fluctuating temperatures reduced survival, depressed the survival optimum from ~30°C to ~24°C, and extended development by ~71% compared to equivalent constant means. A competing-risks discrete-time hazard analysis demonstrated that thermal variability simultaneously increased pre-adult mortality and decreased adult emergence. Population origin effect was significant in only a subset of traits, with divergence concentrated at the adult life-stage and thermal extremes rather than across the whole temperature range. Constant-temperature thermal performance curves overestimated fluctuating-temperature development rates by 42–60% at 23–28°C. When mapped across Europe, this inflated both predicted development rates and the geographic extent of viable development. Incorporating realistic thermal variability into experimental parameterisation is therefore progressively essential, as reliance on constant-temperature measurements can substantially bias predictions of Ae. albopictus development and spread, with potentially substantial consequences for downstream epidemiological modelling.