Floods, like other natural phenomena, have strong influence on the evolution of organisms. Floods play a major role in the natural selection and development of morphological and behavioral adaptations of plants and animals living in floodplains. Roughly speaking, natural disasters may be divided into smaller, more frequent, that reshape the behavior of organisms and trigger multigenerational adaptation to living conditions through natural selection, and to larger, rarer natural disasters that cause the extinction of one or more species, thereby accelerating the development of surviving species (McEvan and Punyasena 2007). Flooded habitats are considered as ecosystems with one the highest biodiversity on Earth due to high evolutionary pressure of the challenging environment caused by regular disturbances (Naiman et al. 1993). The fauna and flora of these ecosystems was formed as a result of hydrological gradient, floods, and different types of soil, resulting in a high diversity of plants, fungi and animals. However, man has indirectly and directly drastically reshaped the flood regimes across Europe over the past 60 years, rendering these habitats as the most threatened in Europe (Joyce et al. 1998).
The construction of dams, embankments, draining meadows, deepening riverbeds, fortifying banks, building infrastructure, have major impact on floodplains, which are often no longer flooded or, on the contrary, remain flooded for prolonged periods (Leyer 2005). In addition, climate change is also causing floods that are longer and more inconsistent, to which existing local vegetation and fauna cannot adapt (Sperna Weiland et al. 2012). The changed regimes and duration of floods directly affect the composition of ecosystems (Asselman 2003), which reduces the plant and indirectly the animal diversity of the area. Terrestrial animals are most at risk for drowning during floods when they are submersed in poorly oxygenated standing water. They are also threatened by the possibility of being flushed from their habitat, and by the impact of toxic substances that form in the soil from decaying vegetation such as methane, acetic acid, and butyric acid. With the presence of sulfates (fertilizers, pesticides), the formation of sulfuric acid can also occur (Plum 2005). All these threats only increase with the prolonged duration of floods and ultimately cause local extinctions of maladapted species.
Ljubljansko Barje plain has a long history of anthropogenic interventions, in particularly drainage, aiming on prevention of floods in populated areas. In addition, due to climate change, extensive precipitation has become more frequent in the region (ARSO 2019), causing longer flooding of low-lying areas along main rivers, especially in the autumn (Fig. 1). Far-reaching potential effects of a prolonged flooding in Ljubljansko barje was first noted in 2012 when a sharp decline in abundance and distribution of the Scarce large blue (P. teleius) was documented during regular monitoring of the Habitats directive species (Verovnik et al. 2009; Zakšek et al. 2012). The decline coincides with 2010 long-lasting autumn flood, which persisted for up to 29 days in the surveyed area (ARSO 2019). No specific research was conducted at that time to identify the reasons for the sharp decline but direct mortality of larvae and pupae due to long lasting flooding, and retreat of the host ants from the flooded areas, were speculated as a potential factor.
The main host of the P. teleius in Central Europe are Myrmica scabrinodis and M. rubra ants (Pech et al. 2007). The caterpillars feed in first three instars with inflorescences of the Great Burnet (Sanguisorba officinalis), which is their only host plant (Thomas 1984, Elmes et al. 1991). The development of caterpillars on the flowers usually takes about 3 to 4 weeks (Elmes et al. 1991). After the transition to the fourth instar, they descend from the host plant and begin to emit acoustic (Sala et al. 2014) and chemical signals (Sliwinska et al. 2006) which attract and confuse host ants, so that they adopt them as their own brood (Thomas 1984, Elmes et al. 1991). The caterpillars are predatory on ant brood and their development takes from 10 to 22 months depending on the size of the ant colony (Witek et al. 2006). After pupation in early summer adults emerge from the ant nests at the end of June or beginning of July depending on season (Verovnik et al. 2012).
Although the effects of flooding on ants is not well studied, several adaptations that allow ants to survive floods –has been described (Marx et al. 2012). The Myrmica ants, in particular, are especially exposed to flooding as they build underground nests. Representatives of this very widespread genus have developed two strategies that allow them to survive several days of inundation. The first strategy is to use the trapped air bubbles in the anthill during the flooding; they also tolerate increased concentrations of carbon dioxide. The species is able to take advantage of the air bubbles that form in the anthill gallery and with their help spend several days completely underwater. If the floods do not recede for a longer time, the concentration of carbon dioxide in the bubble rises to the limit value causing ant mortality (Robson 2010). Another survival strategy is swimming and finding a refuge on dry land. Representatives of the genus Myrmica have often been observed swimming several meters to land or the nearest branch (Radchenko and Elmens 2010). Even less is known on Myrmica ants recolonization of suitable habitats, but it likely appears through budding from the existing colonies (Hicks, 2012) as nuptial flights, especially of females, are very localized (Elmes 1991).
Our aim was to study the effects of flooding on host ant M. scabrinodis distribution and recolonization. Specifically, we wanted to investigate: (i) how floods and their duration affect the presence – the probability of survival of the M. scabrinodis host ant. We assumed that using above described strategies ants survive shorter floods, but with the prolongation of flood duration, the probability of survival of ants decreases and finally drops to zero resulting in local extinctions; (ii) whether and how the recolonization of host ants to unoccupied areas is proceeding after the end of floods. Habitat recolonization in mobile species occurs in at least two manners, either by gradual marginal expansion of the population range, as is budding of ant colonies, or by long distance dispersal and the establishment of new (also remote) metapopulation nuclei, as e.g. is the case in nuptial flights in social insects. Since ants are less migratory for most of their lives (except for the nuptial flights), we expect, that the first of both listed systems of recolonization would prevail. Therefore, we predict: (i) that the probability of occurrence of ants decreases with distance from the nearest topographic refuge (elevated site enabling ant survival during floods), and (ii) this probability simultaneously increases over the time (elapsed since the last major flood) due to the (gradual) process of recolonization. These patterns of mortality and recolonization are of great conservation importance also for the P. teleius, as it fully depends on the presence of its host ant.