The findings of the invasive mosquito species Ae. albopictus in used tire companies in the Netherlands demonstrate the long-distance human-aided dispersal capabilities of this species [37]. Given its potential as a vector of infectious diseases, the question that arose was whether this non-native mosquito species could survive under the climate conditions of the Netherlands, and whether it would be able to establish permanent populations in this country. In this study, we predicted the potential areas that may be suitable for Ae. albopictus populations in the Netherlands using two modelling approaches. Results show differences in the predicted suitability values between the models. The MaxEnt model is the more restrictive for the species in the Netherlands. The winter temperature is the variable that most contributes to the performance of the model, thereby reducing the chance of establishment of the species in northern parts of Europe. However, we have found the species outdoors at used tire companies every year since 2010 at different locations across the country. These findings can be explained by the continuous and frequent introduction of used tires at the facilities containing Ae. albopictus diapausing eggs from areas where the species has established populations (e.g. northern Italy or southern France). Conversely, the LST model shows that coastal areas are suitable for overwintering of eggs and the entire country is suitable for successful completion of the life cycle if the species is introduced after the winter months.
The MaxEnt results show lower habitat suitability than the LST model probably due to the earlier time-frame temperature period used (MaxEnt: 1970-2000; LST (2009-2015). According to the Royal Netherlands Meteorological Institute (KNMI) climate scenarios [38], the area of the Netherlands experienced a change of monthly temperature increase that has favoured the invasive mosquito since 2000. Additionally, in the MaxEnt model, the time period of the climatic data used for training (1970-2000) does not widely match the time period of the used occurrence data. First records of populations of the species in Europe (Italy) date back to the early 1990’s, but for example records in Spain and Germany were reported after the year 2000 [39, 40]. Therefore, we assume that if the species would have been introduced before the year 2000 in these areas, it would have successfully become established due to the proven invasiveness of Ae. albopictus. Unlike other similar models [27, 32], our MaxEnt model has only been fed with known current established populations or locations where overwintering of the species has been observed and reported. These studies were performed using also occurrence data from interceptions without confirmation of establishment or overwintering (e.g. interceptions in Lucky bamboo greenhouses in the Netherlands). In our opinion, results of these models could lead to an overestimation of the northern limits of the species in Europe.
As observed in previous species distribution models [23, 27, 35], an expansion of climatically suitable habitats in western and central Europe is expected in the near future for Ae. albopictus. In our study, climate change scenarios have not been taken into account. As known from the climate scenarios developed by the KNMI, even under conservative and optimistic scenarios, the temperature in the Netherlands is expected to continue to rise, with an increase of values between 1 and 2.3 ºC in 2050, and 1.3 and 3.7 ºC in 2085 [38]. That means that climate change will undoubtedly increase the winter temperatures, leading to the increase of the probability for overwintering of eggs, and consequently increasing the risk of establishment of Ae. albopictus in some parts of the Netherlands within 30 years.
The LST model shows very low suitability values in forested areas and greenhouse farming houses, and the very high values in urban areas (e.g. built-up areas, residential, industrial). These results are in accordance with Guo et al. [41] who found that built-up areas with paved roads, residential, and factory buildings have a higher LST, and vegetated land covers have the lowest LST. A possible explanation for a low LST in vegetated areas is that the LST is collected from the forest canopy, which is influenced by the cooling effect of evaporation. The low suitability accounted in the greenhouse buildings situated in the western part of the country, is directly related to the highly reflective roofs of the greenhouses, an artefact which suggests that these locations are cool places.
The results of our study provide a new perspective on the previous study of Takumi et al. [28]. The latter study was based on climatological information from weather stations (only from 2006), and concluded that winter conditions in the Netherlands were permissive for the establishment of temperate strains of Ae. albopictus. In our study, two parameters that act against the establishment of Ae. albopictus in areas where diapausing eggs may overwinter, were also taken into account: the annual mean temperature (11ºC), and the day of year when 1,350 Growing Degree Days (GDDs) are reached. These parameters are considered important because they affect the time needed for Ae. albopictus larvae to mature into adult mosquitoes during the period of seasonal activity. It should be mentioned that in comparison with the study of Takumi et al. [28], we did not use precipitation data in the LST model. For optimal development of Ae. albopictus, the species requires at least 500 mm of annual rainfall [42]. The mean annual rainfall in the Netherlands ranges from about 700 to 900 mm [43], and is not considered a limiting factor for the species to be taken into account in our LST model.
In our LST model study, the results depend on the selected thresholds for the January temperature (1ºC), annual temperature (11ºC), and day of year when 1,350 GDDs are reached. These threshold values were based on [44, 45] and [23]. Modifying these temperature thresholds with a few degrees could result in a substantial change in the suitable area for the species. For example, lowering the January and annual temperature thresholds a few degrees would increase the area regarded as suitable for Ae. albopictus. In the same way, moving by several weeks before the threshold value when 1,350 GDDs are reached may result in a decrease in the suitable area. We based our LST model study on temperature thresholds studied in other areas of the world. For the temperate strain of Ae. albopictus spreading in Europe, the threshold of 11ºC has been accepted as lower threshold for larval growth in the USA [46] and in Japan [44]. Furthermore, in the risk maps produced by the European Centre for Disease Prevention and Control (ECDC) in 2009 [22], it was concluded that the criterion of an annual mean temperature of 11ºC seemed to fit well with the overall observed distribution of the temperate strain of Ae. albopictus in Europe. Compared with results from studies in Italy [47] and Switzerland [48], our GDDs model results do not reveal new suitability hotspots for the species in the Netherlands. As shown in the results, the species could develop from larvae to adult every year in the whole study area for the period 2009-2015.
As stressed by Thomas et al. [49] and Tippelt et al. [50], single events of extreme temperatures in short periods may have a strong impact on overwintering of Ae. albopictus because they can cause irreversible damage to the eggs. Laboratory experiments of Thomas et al. [49] have shown that eggs of European Ae. albopictus could endure temperatures of -10ºC for at least 24 hours, and even show low rates of hatching after exposure to temperatures as low as -12ºC for short periods. At the location Weert, average daily temperatures below -10ºC registered at the beginning of 2012 could have affected the overwintering of eggs, and the finding of the species during surveillance could be related to the introduction of eggs with new tire imports after the winter months. In general, only a few locations could be affected by short periods of low temperatures in the Netherlands suggesting that at these locations, egg hatching rates after the winter can produce sufficient individuals to build up a population during the next season.
In the Netherlands, after the implementation of a risk-based surveillance strategy on invasive mosquitoes, Ae. albopictus has been intercepted yearly at used tire companies or its surroundings [30]. However, the species has not successfully established in these interception areas. Because surveillance and control of invasive mosquitoes started in 2010, and used tire companies had already imported tires before the start of the surveillance, Ae. albopictus may have been introduced many years before 2010. The question is whether in the absence of surveillance and control measures, what factors could have worked against the successful establishment in these areas. The MaxEnt results indicate low suitability for the species in the Netherlands, in comparison with areas in southern Europe, and only one location (Moerdijk) in the LST model results showed moderate suitability for overwintering of eggs (value < 0.5). The length of the reproductive season in the Netherlands could also contribute to lowering the probability of suitable conditions. In fact, this length is considerably shorter in the Netherlands than at southern latitudes in Europe, due mainly to differences in temperature. That means that the temperature necessary to stimulate the hatching of diapausing eggs will occur later in the season in the Netherlands than in southern European regions, and the temperature causing adult mortality will arrive early in the season. However, Ae. albopictus is a species with a high ecological plasticity that can cope with a wide range of climatic conditions, has competitive ability, and has demonstrated to have the potential to adapt fast to the environment during the invasion process [51]. All these facts might allow Ae. albopictus to adapt to colder temperatures in the new invaded locations, facilitating establishment in colder regions. Considering the applied methodological approach and the used data, the physiological plasticity of the species [51] can be considered as the most limiting factor affecting the prediction in both models. In MaxEnt, if in the future the species will become established in drier environments (i.e. in southern Spain), or in colder northern regions (i.e. Belgium or The Netherlands), the contribution of these occurrence locations used in MaxEnt will result in a different output than our study. In the LST model, the adaptation to colder temperatures will make the use of survival temperature thresholds not useful. Possible adaptation to colder climates will increase the areas suitable in northern regions.
We conducted our study using climatic variables exclusively as the main factors effecting the distribution of Ae. albopictus in the Netherlands. However, other factors such as land cover/land use [52], habitat availability and microclimate [53] can significantly influence the success of the establishment in an area. In the case of the Netherlands, the winter and annual temperatures are the limiting factors in our models. In this case, hibernation of diapausing eggs indoors or in warm and protected areas in cities, may protect the mosquito from cold events and may be responsible for local establishments under these special conditions. Especially the effect on the microclimate created by the tires could not be assessed within the framework of this study but is potentially influential. We believe that the tires can provide a suitable site for egg overwintering, larval development and adult shelter. Several factors can influence the microclimate in the tire locations, such as storage method (piles, pyramids, etc.), amount and type of tires, or exposure to wind. As also shown in our results, higher temperatures can be expected around cities (heat island effect) [52]. In cities, potential breeding sites such as urban catch basins can provide favourable microclimatic conditions for overwintering of diapausing eggs compared to more cold-exposed sites [53]. Unfortunately, these are factors that could not be considered in this study and are recommended for future investigations. Because the species is currently invading areas northwards in Europe [54], regular updates of the modelling using updated occurrence and climate data are recommended.
In summary, the results obtained lead to a better understanding of the species’ potential distribution, and identified areas with a risk for the establishment of the species in the Netherlands. However, the results of the two modelling approaches were different, and for the interpretation of these results, one needs to be aware of the limitations of both modelling approaches. For the MaxEnt model, regular updates of the model using the most recent occurrence and climate data are recommended. Specially, new occurrence data on established populations in northern regions in Europe coupled with climate data corresponding to the occurrence period (1990-2018) will be of interest to accurately predict the geographical limits of the species. For the LST model, yearly updates of the model after the winter months, and using the recent data will provide the egg overwintering probability in the study area, including also the used tire locations where the species was found before the winter months. This information will be relevant for the authorities in charge of mosquito control operations, especially at the finding locations where high probability of overwintering is predicted. At these sites, eggs would be expected to survive the winter, larvae will be expected to emerge during the mosquito season in the next year, and as a consequence mosquito control will be advised to prevent establishment of the species. If an introduction of Ae. albopictus is found, special attention should also be taken in urban and peri-urban areas, where the species may inhabit artificial containers and catch basins as breeding sites. The variability of microclimate among sites in complex urban conditions could provide more favourable conditions for the species, and thus those areas should be placed under intensive surveillance. Experiences from the invasion in Europe tell us that once this species has colonized an area, eradication might be difficult or impossible to achieve [51, 55].