Research

Global warming increasingly threatens biodiversity through local extinctions, yet empirical validations of predictive species distribution models remain rare. We validated a historical extinction model using site-level "disappearances"—defined as the current absence of a species at sites where it was recorded 20–60 years ago. Focusing on a cold-adapted salmonid, the southern Asian Dolly Varden charr (Salvelinus curilus) in Hokkaido, Japan, we enhanced risk models by incorporating seasonal thermal regimes, groundwater effects, and biotic interactions.

We electrofished 138 historical sites (2018–2024) to assess species presence, and analyzed long-term air and modeled groundwater temperatures to estimate warming trends. Generalized linear mixed models were developed to evaluate disappearance risks against summer temperature trends and competition with white-spotted charr, masu salmon, and rainbow trout.

Local disappearance occurred at 12 sites (8.70%), doubling the prediction of the previous annual-temperature-based model (6 sites, 4.35%). Disappearances did not correlate with latitude or elevation. Models incorporating summer warming and interspecific competition yielded the highest predictive performance, whereas annual mean or subsurface temperature models performed poorly.

Our findings demonstrate that seasonal thermal dynamics and biotic interactions—rather than annual means—drive local disappearances. Limited dispersal and habitat specificity likely exacerbate these climate impacts. Integrating these ecological complexities significantly improves the accuracy of risk forecasting, providing crucial guidance for international freshwater conservation strategies.

How and why ecologically similar species can coexist has been a central theme in ecology. A classic model system for exploring coexistence mechanisms is the congeneric charr with parapatric distributions along water temperature gradients: Dolly Varden charr (DV, a cold-adapted species) and white-spotted charr (WSC, a warm-adapted species). However, past experiments have shown that WSC always outcompetes DV regardless of temperature, which does not explain why DV dominates in cold waters in the wild.

To address this paradox, we conducted a field survey at 28 stream habitat mosaics. Species composition shifted from DV to WSC with increasing water temperature. Interestingly, we also discovered a new pattern: WSC was abundant even in cold habitats when DV was absent. Based on this finding, we proposed a novel coexistence mechanism in which the distribution is determined by the dominant species’ cost–benefit balance, independent of the subordinate species.

To test this hypothesis, we carried out enclosure experiments in four cold streams using a BACI design. We excluded DV and then released 50 WSC into two “impact” streams and 50 DV into two “control” streams. After one month, WSC persisted in cold streams at similar rates (26–46%) as DV. This result supports our hypothesis that costs imposed by DV, rather than water temperature, reduce WSC’s use of cold-water habitats.

Together with previous studies, our results suggest the following distributional process: DV first colonized cold habitats but could not expand into warmer habitats due to competitive disadvantage against WSC. Conversely, WSC avoid cold habitats when DV is present, owing to costs such as reduced food availability through density-dependent competition. Thus, strong asymmetric competition and resulting spatial segregation may drive the regional coexistence of these two ecologically similar species.

Artificial transplantation of species can lead to the extinction of native species not only through cascading effects in predator–prey interactions but also through unexpected hybridization, known as invasive hybridization. In western Japan, the distributions of two bitterling fish (subfamily Acheilognathinae), Tanakia lanceolata and T. limbata, overlap. Bitterlings deposit their eggs in the gills of freshwater bivalves, where the early juvenile stages develop. However, populations of freshwater bivalves are declining worldwide, reducing available spawning substrates for bitterlings. T. limbata has been artificially introduced into some rivers around Matsuyama, Japan, where it now coexists with native T. lanceolata, and hybrids between the two species have been observed.

We collected individuals of both species from multiple sites in western Japan and analyzed their genetic population structure using multilocus microsatellites and mitochondrial cytochrome b sequences. Population structure analysis identified three genetically distinct groups: T. lanceolata, T. limbata “West Kyushu,” and T. limbata “Setouchi.” These two clades of T. limbata were also supported by molecular phylogenetic analyses based on cytochrome b. In Matsuyama, most hybrids originated from crosses between male T. lanceolata and female T. limbata “West Kyushu,” comprising more than 10% of the collected specimens. This suggests that hybrids commonly arise between females of colonizing species and males of native species. In contrast, interspecific hybrids were detected at high frequencies in some rivers on Kyushu Island, where the two bitterlings naturally occur in sympatry. Furthermore, we detected a few T. limbata “Setouchi” individuals in the Midori and Kase Rivers, likely introduced from other regions, coexisting with native T. limbata “West Kyushu.” This cryptic invasion may have triggered interspecific hybridization.

These results suggest that the artificial introduction of fish, coupled with the decline of unionid mussels and habitat degradation, has driven extensive hybridization among bitterlings in western Japan.