The island species-area relationship (ISAR) describes how species richness increases with increasing area of a given island or island-like habitat, such as freshwater lakes. While the ISAR is one of the most common phenomena observed in ecology, there is variation in both the form of the relationship and its underlying mechanisms.

We compiled a global dataset of benthic macroinvertebrates from 524 shallow freshwater lakes, ranging from 1 to 293300 ha in area. We used individual-based rarefaction to determine the degree to which ISAR was influenced by mechanisms other than passive sampling (larger islands passively sample more individuals from the regional pool and, therefore, have more species than smaller islands), which would bias results away from expected relationships between rarefied species richness (and other measures that capture relative abundances) and lake area. We also examined how climate may alter the shape of the ISARs.

We found that both rarefied species richness (the number of species standardized by area or number of individuals) and a measure of evenness emphasizing common species exhibit non-significant relationships with lake area, suggesting that the expected ISARs in these lakes most likely result from passive sampling. While there was considerable variation among ISARs across the investigated lakes, we found an overall positive rarefied ISAR for lakes in warm (i.e., tropical/subtropical) regions (n = 195), and in contrast, an overall negative rarefied ISAR in cool (i.e., north temperate) lakes (n = 329). This suggested that mechanisms beyond passive sampling (e.g., colonization-extinction dynamics and/or heterogeneity) were more likely to operate in warm lakes. One possible reason for this difference is that the area-dependent intensity of fish predation, which can lead to flatter ISARs, is weaker in warmer relative to cooler lakes.

Our study illustrates the importance of understanding both the pattern and potential processes underlying the ISARs of freshwater lakes in different climatic regions. Further, it provides a baseline for understanding how further changes to the ecosystem (i.e., in lake area or climate) might influence biodiversity patterns.

We compiled data from studies that used a similar sampling methodology to collect macroinvertebrates from lakes that varied in area. Our dataset consisted of 524 freshwater lakes from across the globe sampled at least once from 1974–2021. Data from 505 lakes was contributed by co-authors of this study and included lakes in Denmark (n = 198), China (n =131), Finland (n = 96), USA (n = 48), New Zealand (n = 15), Greece (n =11), Mexico (n = 4) and Russia (n = 2). Most of these data had been published previously. We also included data from an additional 19 lakes identified in a Google Scholar literature review using the search terms "shallow lake" AND "zoobenthos" OR "benthic macroinvertebrate". From this, we identified sources that reported taxonomic-level abundance and nutrient concentrations and met the sampling criteria applied in the other studies (mean depth < 10 m and sampling conducted in offshore zones and warm seasons). These data come from six publications. See details in paper.