We determined the minimum survival temperature of eggs for two strains of Ae. albopictus and for one strain of Ae. aegypti and showed that the hatching success after the cold treatment was significantly increased in European eggs which have undergone a diapause compared to non-diapausing European eggs after exposure to cold temperatures (0°C to -15°C), for different durations. Overall, we focused on relatively short exposure times (1h, 4h, 8h, 12h and 24h), because these are the time spans on which minimum temperatures are available from climatology and meteorology. Consequently, our data can be used to make evidence-based decisions on the temporal resolution of temperature data needed for modeling approaches. Temperature was the main controlling factor, whereas the duration of the cold treatment only influenced the hatching response significantly at the thermal limits of survival for each strain/species.
To the best of our knowledge there are no comparable studies on the cold tolerance in Ae. albopictus eggs which have undergone a diapause. Previous studies on non-diapausing, non cold-acclimated eggs of a North American strain of Ae. albopictus have shown that no hatching occurs after a 24h cold period of -10°C . Our results confirmed this threshold for at which no eggs survive European non-diapausing Ae. albopictus eggs. However, the survival of the European strain in this study is reduced more strongly with decreasing temperature than the survival of the North American: for the European strain we found 50% egg mortality at -5°C, whereas for the North American strain 50% egg mortality at -8°C was observed . The minimum survival temperature of Ae. aegypti eggs in the field (Houston, Texas, USA) was found to be 7°C for a 24h cold period . But in nature short term temperature fluctuations occur. Additionally, at one sampling site small scale differences across short distances can be substantial. Ecological complexity contributes to statistical noise. This is why controlled experiments are needed to identify physiologically relevant thresholds.
In the laboratory -3°C for 24h was identified as a threshold for Ae. aegypti eggs, respectively . The present results show only a slight increase in the minimum survival temperature for Ae. aegypti (-2°C for 24h). Using a very long-lasting laboratory colony may yield deviating results compared to natural populations. However, such colonies are more appropriate to serve for experimental proof of life cycle thresholds.
It was suggested that the duration of exposure to temperatures below a certain value is an important factor influencing the hatching rate of mosquito eggs . Cold acclimation tends to increase cold hardiness in Ae. albopictus eggs [29, 33]. In this study, previous cold acclimation was avoided to distinguish between the direct impact of cold temperatures and the duration of exposure (see also ). The role of the duration of low temperatures is confirmed at the thermal limits of egg survival by the present study.
Up to now, the risk of Ae. albopictus to establish in Europe was assumed to be relevant only for regions with cold-month mean temperature of 0°C or higher [35, 36]. The -5°C coldest-month isotherm was suggested to characterize the maximum northward expansion for continental Asia and also for North America . The results presented in this study emphasize the ecological importance of absolute minimum temperatures. Up to now, thermal minima are neither adequately considered in environmental niche models nor in epidemiological models. Yet, vector niche modeling is mainly based on long-term average conditions such as annual mean temperature and annual precipitation (e.g. ). Winter conditions in terms of the mean minimum temperature of the coldest month and days with ground frost per month are considered in the niche model of Ae. albopictus. Although the incorporation of the absolute minimum temperatures would considerably improve vector risk maps, this would be difficult to implement in large scale projections due to the limited availability of meteorological data in an hourly resolution. Recent epidemiological models consider temperature and season dependent population dynamics of vectors . However, those models start each annual cycle with the same initial number of mosquito individuals. Knowing the minimum survival temperature and survival success after frost events has therefore the potential to improve epidemiological models in temperate zones substantially.
The present study has certain specifics that need to be taken into account: First, the long colonization history of the European Ae. albopictus strain raises the question to what extent this strain actually now represents the characteristics of its founding population. Second, the successful production of diapausing eggs in the laboratory in all females under the recorded circumstances is unlikely. Hatching tests during diapause still showed about 10% spontaneous hatching.
Nonetheless, winter conditions, or more specifically absolute minimum temperatures, play a decisive role for the distribution limits of a species . Spatial quantification of absolute minimum temperatures, however, is non-trivial. Microclimate may vary up to 10°C at small spatial scales . The specific thermal conditions of microsites were found to matter also for the occurrence of aedine species . For these species, indoor breeding sites  are adding anthropogenic habitats with specific conditions that do not refer to landscape traits. Thus, it would be naive to concentrate on natural and semi-natural site conditions alone.
Future studies should pay attention to winter conditions by incorporating realistic freeze-thaw cycles to identify how far these temperature fluctuations are affecting egg survival. Furthermore, the knowledge on winter survival of viruses such as dengue and Chikungunya within the eggs of vector insects is of utmost importance and should be intensively addressed in future research.