Catchment properties and the photosynthetic trait composition of freshwater plant communities

Change in plants as bicarbonate rises Freshwater plants can be broadly divided into two major categories according to their photosynthetic traits: Some use carbon dioxide as their carbon source, whereas others use bicarbonate. Iversen et al. found that the relative concentrations of these two inorganic carbon forms in water determine the functional composition of plant communities across freshwater ecosystems (see the Perspective by Marcé and Obrador). They created global maps revealing that community composition is structured by catchment geology and not climate (in contrast to the terrestrial realm, where the trait composition is structured by temperature and rainfall). Anthropogenic influences from land-use change are causing large-scale increases in bicarbonate concentrations in freshwater catchments and are thus leading to wholesale changes in the composition of their aquatic plant communities. Science, this issue p. 878; see also p. 805 The geographical distribution of bicarbonate use in freshwater plants is controlled by catchment characteristics. Unlike in land plants, photosynthesis in many aquatic plants relies on bicarbonate in addition to carbon dioxide (CO2) to compensate for the low diffusivity and potential depletion of CO2 in water. Concentrations of bicarbonate and CO2 vary greatly with catchment geology. In this study, we investigate whether there is a link between these concentrations and the frequency of freshwater plants possessing the bicarbonate use trait. We show, globally, that the frequency of plant species with this trait increases with bicarbonate concentration. Regionally, however, the frequency of bicarbonate use is reduced at sites where the CO2 concentration is substantially above the air equilibrium, consistent with this trait being an adaptation to carbon limitation. Future anthropogenic changes of bicarbonate and CO2 concentrations may alter the species compositions of freshwater plant communities.

T he biogeography of terrestrial plants is influenced by climatic factors-primarily air temperature and precipitation (1). Furthermore, the distribution of biochemical traits, such as the two terrestrial CO 2concentrating mechanisms, C 4 photosynthesis and crassulacean acid metabolism, are linked to temperature and water availability (2). Although freshwater angiosperms evolved from terrestrial ancestors (3), their growth is controlled by light, nutrients, and inorganic carbon (4) rather than water, and therefore the factors influencing their biogeography are likely to be different. Inorganic carbon potentially limits photosynthesis in aquatic systems, because the diffusion of CO 2 is 10 4 -fold lower in water than in air. Consequently, the CO 2 concentration needed to saturate photosynthesis is up to 12 times the air equilibrium concentration (5). Moreover, rapid photosynthesis can reduce CO 2 in water substantially below air saturation (4).
In response to carbon limitation, a few aquatic angiosperms evolved the same CO 2concentrating mechanisms found in their terrestrial ancestors, but the most frequent mechanism, found in about half of studied submerged freshwater plants, is the exploitation of bicarbonate (HCO 3 − ) (4, 6), which is derived from mineral weathering of soils and rocks in the catchment. Bicarbonate is the dominant form of inorganic carbon in fresh waters when the pH is between~6.3 and~10.2, and its concentration often exceeds that of CO 2 by a factor of 10 to 100 (6). The ability to use bicarbonate is present in most taxonomic groups and appears to have evolved independently in cyanobacteria, eukaryotic algae, and vascular aquatic plants (7). This shows the fundamental importance of bicarbonate use to plant fitness (6); increase of photosynthesis, growth, and primary productivity at higher bicarbonate concentrations has been documented (8)(9)(10). However, bicarbonate use is not ubiquitous, because it involves costs as well as benefits. Costs include energy, as it is an active process (11) and rates of photosynthesis at limiting concentrations of inorganic carbon are greater in CO 2 users than in bicarbonate users (5,12). Thus, where CO 2 concentrations are substantially above air saturation, as is often the case in streams, the benefit of bicarbonate use will be reduced (13). Furthermore, obligate CO 2 users can exploit alternative CO 2 sources in the air, lake sediment, or the water overlying the sediment (14), allowing continued photosynthesis without the need to invest in mechanisms required for bicarbonate use.
We hypothesized that because limitation of photosynthesis by inorganic carbon supply is widespread in freshwater plants, the relative concentrations of bicarbonate and CO 2 at a particular site should determine the proportion of plants that are obligate CO 2 users versus bicarbonate users. Because geochemical catchment characteristics determine bicarbonate concentration, there should be broad biogeographical patterns in the proportion of freshwater plants able to use bicarbonate, whereas at a smaller scale both the CO 2 and bicarbonate concentrations in lakes and streams might structure the functional group composition.
To test these hypotheses, we generated a database of freshwater angiosperms and their ability to use bicarbonate as an inorganic carbon source, based on data found in the literature. These were complemented with new data we gathered on 35 species from mainly tropical regions where few prior data existed (table S1) (15). The resulting 131 species represent~10% of known species with a submerged life stage (16), and of these, 58 (44%) can use bicarbonate. To quantify the distribution of bicarbonate users versus CO 2 users, we used: (i)~1 million geo-referenced plant records, (ii) global plant ecoregion species lists, and (iii) 963 sitespecific plant compositions from Northern Hemisphere lakes and streams (fig. S1). In each of the investigated 963 sites, plant composition was related to measured concentrations of CO 2 and bicarbonate. The geo-referenced plant records and ecoregion species lists were linked to local bicarbonate concentrations derived from a constructed global map of bicarbonate concentrations ( fig. S2) (15).
In the analyzed lake and stream sites, concentrations of both bicarbonate and CO 2 affected the occurrence of obligate CO 2 users versus bicarbonate users, albeit differently within and between lakes and streams ( Fig. 1 and fig.  S3). The chance of observing a bicarbonate user in lakes and streams correlated directly with concentrations of bicarbonate and  (Fig. 1C). The present study shows that the concentration of bicarbonate has a different effect on the proportion of bicarbonate users in lakes versus streams. Unlike in lakes, no relationship between bicarbonate availability and bicarbonate users was found in streams. This upholds our hypothesis that where concentrations of CO 2 are high, the competitive advantage of using bicarbonate as a carbon source for photosynthesis will be reduced even if bicarbonate is available.
Across global plant regions (17), the shifting proportions of bicarbonate users versus obligate CO 2 users showed distinct spatial patterns ( Fig. 2A). Compared to the overall mean, a higher proportion of bicarbonate users  was observed in Africa, temperate Asia, and the northern part of North America ( Fig. 2A). Globally, species using bicarbonate were found in areas with higher bicarbonate concentrations [bicarbonate users − CO 2 users = 0.16 mM (0.02, 0.30)] ( Fig. 2C; see Fig. 3 for a local example). The proportion of bicarbonate-using species increased with bicarbonate concentration within ecoregions [b = 0.14 (0.05, 0.24)] (Fig. 2B). Because catchment geology and geological history shape the distributions of lakes and rivers, as well as the bicarbonate concentrations in freshwater ecosystems (18,19), they are the chief determinants of plant distributions in fresh waters. CO 2 concentrations are largely regulated by local CO 2 supersaturated inflow (20) and ecosystem metabolism, making modeling difficult at large spatial scales (19,21). Thus, future models of freshwater CO 2 concentrations may improve the prediction of plant distributions even further. Although global lake and river data exist to some extent as annual means (22), given the temporal varia-bility in CO 2 concentration, the appropriate concentration would be that during the growing season at the specific site (20).
Anthropogenic changes as a consequence of deforestation, cultivation of land, application of nitrate fertilizers, and reduced atmospheric acid deposition (23) are causing large-scale increases in bicarbonate concentrations (24,25). The observed increasing bicarbonate concentrations are expected to cause a severe impact on bicarbonate-poor lakes, because higher bicarbonate concentrations will markedly change species composition (26) by allowing tall, fastgrowing bicarbonate users to colonize and suppress smaller species adapted to the use of CO 2 alone in or near the sediment (27). There is evidence for reestablishment of species that are able to use bicarbonate, after the bicarbonate has increased because of liming (28) or as a result of reduction in acid deposition (29). Moreover, systematic changes in species composition caused by changes in CO 2 concentration have also been demonstrated in a river system where the proportion of CO 2 users declined as CO 2 decreased downstream (13). In contrast, increasing atmospheric CO 2 concentrations, even if they influence dissolved CO 2 , will have little effect on the abundance of bicarbonate users, because increases in CO 2 will be small relative to bicarbonate concentrations and will have little effect on plant photosynthesis rate (30).
Our study shows that bicarbonate use by aquatic angiosperms is widespread in fresh waters around the globe and that the proportion of obligate CO 2 users to bicarbonate users is significantly related to the bicarbonate concentration. Among terrestrial plants, the evolution of leaf traits and different photosynthetic pathways that enable rapid carbon assimilation and improved water economy (31) has resulted in global biogeographical patterns that are linked to variations in climate (32,33). In contrast, for freshwater plants, we show that biogeographical patterns of bicarbonate use exist and that these are caused by catchment properties that determine the concentrations of bicarbonate and CO 2 . This insight will help evaluate the repercussions of future changes in concentrations of bicarbonate and CO 2 on the biodiversity and ecosystem functions for fresh waters.  . S2), and species data were extracted from the geo-referenced plant occurrences (15). of the United Nations Environment Programme. The authors alone are responsible for the views expressed in the publication, and they do not necessarily represent opinions, decisions, or policies of the United Nations Environment Programme. The authors declare no other competing interests. Data and materials availability: All R scripts and cleaned datasets used for this analysis are available at the Dryad Digital Repository (34).