Continent-wide parallel urban evolution of increased heat tolerance in a common moth

Abstract

Urbanization and its urban-heat-island effect (UHI) have expanding footprints worldwide. The UHI means that urban habitats experience a higher mean and more frequent extreme high temperatures than rural habitats, impacting the ontogeny and resilience of urban biodiversity. However, many organisms occupy different microhabitats during different life stages and thus may experience the UHI differently across their development. While evolutionary changes in heat tolerance in line with the UHI have been demonstrated, it is unknown whether such evolutionary responses can vary across development. Here, using common-garden-reared Chiasmia clathrata moths from urban and rural populations from three European countries, we tested for urban evolution of heat shock tolerance in two life stages: larvae and adults. Our results indicate widespread urban evolution of increased heat tolerance in the adult stage only, suggesting that the UHI may be a stronger selective agent in adults. We also found that the difference in heat tolerance between urban and rural populations was similar to the difference between Mid- and North-European regions, suggesting similarity between adaptation to the UHI and natural, latitudinal temperature variation. Our observations incentivize further research to quantify the impact of these UHI adaptations on fitness during urbanization and climate change, and to check whether life-stage-specific adaptations in heat tolerance are typical of other ectothermic species that manage to survive in urbanized settings.

Keywords: heat knock‐down time (HKDT); heat tolerance; latitudinal variation; lepidoptera; urban evolution; urban‐heat‐island effect (UHI).

FIGURE 1

Map of Europe showing the locations of three urban (red triangles), three rural (orange triangles) sampled populations, as well as one semi‐urban (red triangle; site 6) sampled population. For each sampling site 3 km radius maps of land cover classes (Malinowski et al., 2020) are visualized. Note: map lines delineate study areas and do not necessarily depict accepted national boundaries.

FIGURE 2

Larval heat knock‐down time in relation to mass at measurement in both urban (blue symbols and lines) and rural (yellow symbols and lines) populations from Mid‐Europe (circles) and North‐Europe (triangles). The shaded areas around the regression lines show 95% bootstrap confidence intervals.

FIGURE 3

Scatterplot of family means showing the weak and only marginally significant correlation between larval and adult heat knock‐down time (HKDT).

FIGURE 4

Heat knock‐down time (HKDT) (s) in C. clathrata adult females (left column) and males (right column) from North‐European (top row) and Mid‐European (bottom row) populations (N = 606) in relation to age (days) since eclosion. Each panel shows the original measurements for individuals deriving from rural (orange symbols) and urban (blue symbols) populations. Note that the age variable has only integer values but that data points are slightly jittered along the horizontal axis for visualization. The regression lines (orange and blue lines for rural and urban populations, respectively) and their 95% percentile confidence intervals (shaded areas around the lines) are derived from 5000 model‐averaged estimates based on bootstrap resampling. The gray vertical dashed lines indicate the age before which the fitted regression line for the urban population falls outside of the confidence interval of the regression for the rural population; that is, there is statistical support for a longer HKDT in the urban than in the rural population below that age (note that the confidence intervals are narrow at young ages as most data are concentrated in the youngest age classes) (see also Figure 5).

FIGURE 5

Raw data (open circles) on heat knock‐down time (s) in young C. clathrata adults (Mid‐European males: ≤4 days since eclosion; all other groups: ≤3 days since eclosion; i.e., all data left to the vertical dashed lines in Figure 4). Data are shown for females (left) and males (right) from urban (blue) and rural (orange) populations from North‐European (left panel) and Mid‐European (right panel) populations (N _total = 546). Black circles show group‐specific means and black whiskers around them the 95% confidence intervals of the means. Group‐specific sample sizes are indicated at the bottom of the figure.