Rare species support vulnerable functions in high-diversity ecosystems

Abstract

Around the world, the human-induced collapses of populations and species have triggered a sixth mass extinction crisis, with rare species often being the first to disappear. Although the role of species diversity in the maintenance of ecosystem processes has been widely investigated, the role of rare species remains controversial. A critical issue is whether common species insure against the loss of functions supported by rare species. This issue is even more critical in species-rich ecosystems where high functional redundancy among species is likely and where it is thus often assumed that ecosystem functioning is buffered against species loss. Here, using extensive datasets of species occurrences and functional traits from three highly diverse ecosystems (846 coral reef fishes, 2,979 alpine plants, and 662 tropical trees), we demonstrate that the most distinct combinations of traits are supported predominantly by rare species both in terms of local abundance and regional occupancy. Moreover, species that have low functional redundancy and are likely to support the most vulnerable functions, with no other species carrying similar combinations of traits, are rarer than expected by chance in all three ecosystems. For instance, 63% and 98% of fish species that are likely to support highly vulnerable functions in coral reef ecosystems are locally and regionally rare, respectively. For alpine plants, 32% and 89% of such species are locally and regionally rare, respectively. Remarkably, 47% of fish species and 55% of tropical tree species that are likely to support highly vulnerable functions have only one individual per sample on average. Our results emphasize the importance of rare species conservation, even in highly diverse ecosystems, which are thought to exhibit high functional redundancy. Rare species offer more than aesthetic, cultural, or taxonomic diversity value; they disproportionately increase the potential breadth of functions provided by ecosystems across spatial scales. As such, they are likely to insure against future uncertainty arising from climate change and the ever-increasing anthropogenic pressures on ecosystems. Our results call for a more detailed understanding of the role of rarity and functional vulnerability in ecosystem functioning.

Figure 1. Functional distinctiveness as a function of commonness.

Species commonness is measured at the local scale as the mean abundance over all the samples where the species is present and expressed as a percentage of the maximum observed value, and at the regional scale, it is measured as the number of occurrences over all the samples and expressed as a percentage of the maximum observed value. Functional distinctiveness, expressed as a proportion of the maximum observed value, quantifies the uniqueness of species biological traits from the rest of the pool in the ecosystem. Solid lines represent ordinary least square regressions, whereas dashed and dotted lines represent 95^th and 99^th quantile regressions, respectively. Shaded areas indicate the 95% confidence intervals for each relationship. All variables are standardized to allow comparisons among ecosystems and spatial scales. ^ns p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001.

Figure 2. Functional vulnerability as a function of commonness.

Species commonness is measured as in Figure 1. Functional vulnerability, scaled between 0 and 1, quantifies the lack of functional insurance provided by the rest of the pool to the focal species in terms of functional traits and regional occupancy. Solid lines represent ordinary least square regressions, whereas dashed and dotted lines represent 95^th and 99^th quantile regressions, respectively. Shaded areas indicate the 95% confidence intervals for each relationship. ^ns p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001.

Figure 3. Percentage of rare and rarest species for different levels of functional vulnerability.

The functional vulnerability index is scaled (0–1) and was divided into four categories from high to least. Locally rare species are those with a mean abundance value over the samples where present less than 5% of the maximum local abundance value and rarest species are those with only one individual by sample where present on average. Regionally rare species are those with less than 5% of the maximum regional occupancy value and rarest species are those with only one occurrence over all the samples. For each level of vulnerability, we obtained the confidence interval at 90% by randomization and we tested whether the observed percentage of rare and rarest species is higher (red star) or lower (blue star) than expected by chance, otherwise black circle. The vertical grey line is the median obtained at random. Sample sizes are provided in Table S1.

Figure 4. Species supporting some of the most vulnerable functions on coral reefs, in alpine meadows, and in rain forests.

(A) The giant moray eel (Gymnothorax javanicus (Muraenidae)), the largest of the moray eels, hunts by night within the labyrinth of the coral reef (B). (C) Saxifraga cotyledon (Saxifragaceae) is a low-growing, rare evergreen perennial plant, with long flowering stems that make it an important resource for pollinators on species-poor siliceous alpine cliffs (D). (E) Pouteria maxima (Sapotaceae), a rain forest tree with thick, coriaceous leaves and a wide buttressed trunk with thick bark, which may buffer the impacts of drought and fire predicted to occur more frequently for tropical forests (F). Photo credits: (A) M.J. Kramer, (B) J.P. Krajewski, (C) J.P. Dalmas, (D) W. Thuiller, (E, F) C.E.T. Paine.