Chironomid-climate continentality conundrum

PLOS One | https://doi.org/10.1371/journal.pone.0327780 August 5, 2025 11 / 18 lake-water pH, lake-water depth, April and January mean air temperature, and KOI. The independence of KOI can be explained by its formula: it includes October, January and April mean air temperatures. Further, KOI explains the highest amount of variation in the chironomid assemblages (18.4%; Table 2). Thus, KOI appears to be a key driver of change in the chironomid assemblages in the present dataset (Fig 4; Table 2). However, the chironomid assemblages from the dataset could be influenced by unmeasured environmental variables (e.g., water trophic state, conductivity, and oxygen content; catchment vegetation cover/type). In our study, GDD5 has higher explanatory power (17.2%) than in previous studies: GDD5 explains 9% of the variation in Swiss Alps chironomid assemblages [70] and 9.7% in New Zealand Alps assemblages [71]. This could be due to the wide spatial spread of our sampling sites combined with the long GDD5 gradient in our dataset (Fig 2A) and suggests the importance of growing-season length for chironomid assemblages. The high performance of July, relatively high of October, and significant April mean air temperatures in the RDA, which are all related to growing season duration, also support this interpretation. Furthermore, April mean air temperature, while having the lowest explanatory power (4%) among the tested climatic variables, is aligned with the second RDA axis, together

with January mean air temperatures suggesting the distinctive impact of the growing-season start on the chirono- mid assemblages. While, the relationship of chironomid assemblages and October air temperatures has not been studied before, its relatively high explanatory power suggests that the time of the autumn water column mixing may also influence chironomid assemblages. January mean air temperature has been found to impact chironomid assem- blages and explained 5.2% in Swedish chironomid training set [45]. The higher influence of January mean air tem- peratures on the studied dataset (8%) is explained by a high range of this climate variable: from −14 °C in Urals to −1−0 °C in coastal Baltic area. Ice cover is known to cause depletion in the dissolved oxygen content of lake water [16,72]. Thus, the importance of ice cover can be explained by affecting dissolved oxygen changes, to which chironomids are known to be sensitive [35,36,73,74]. Warmer air temperature combined with shorter ice-cover duration have been shown to increase lake water pH [75] over a longer period of time. Our dataset covers a long gradient of ice-cover duration (35−196 days) and ATR, Fig 6. (A) Diagnostic plot of cross-validated estimates of the dataset compared with observed Kerner Oceanity Index (KOI) values and (B) residuals plot from a weighted-average partial least squares (WA-PLS) model based on two components. Continental sites (orange) correspond to −10–0 KOI, transitional sites (green) to 0–10 KOI, and oceanic sites (blue) to 10–20 KOI. https://doi.org/10.1371/journal.pone.0327780.g006