The availability of free CO₂ to submersed macrophytes can vary five-fold or more among freshwater ecosystems, in part because many systems are not in equilibrium with the atmosphere. In greenhouse growth experiments, some macrophyte species show a greater response to CO₂ enrichment than others, much as C3 and C4 plants do in the terrestrial realm, suggesting that CO₂ availability may influence community structure in the field. This study tested the hypothesis that lakes high in [CO₂] favor more [CO₂]-responsive species, just as increasing atmospheric [CO₂] may tend to favor C3 plants on land over time. We coupled a [CO₂]-sensitivity index (CSI, the ratio of relative growth rate at high to that at low [CO₂]) determined experimentally in the greenhouse with quantitative sampling of submersed vegetation along depth contours within Adirondack Mountain lakes to derive a physiologically-based index of community composition. This index, the Community CSI (“CCSI”), is the sum of the products of individual species CSI values and their relative frequencies within each community sampled. We also performed non-metric multidimensional scaling (NMS) ordination, and used first axis scores as an index of community composition based on vegetation data alone.
Results/Conclusions

Over a range of acidic lakes, CCSI showed a significant positive correlation (P<0.05) with [CO₂], in line with our central hypothesis. Over a broader pH range, including circumneutral lakes, the CCSI correlation with [CO₂] was lost. For this latter group of lakes, however, significant positive correlations emerged between CCSI and pH, [Ca], [HCO₃^-], and [total dissolved inorganic carbon] (P<0.01 for pH and P<0.0001 for the others), but not with light availability (Secchi depth) or sediment traits (bulk density and organic content, likely surrogates for mineral nutrient availability). Axis 1 scores from NMS ordination were highly correlated to CCSI values (r² = 0.554, P<0.0001), indicating a strong parallel between our physiologically-based index of community composition and a synthetic index of community composition revealed by the vegetation itself. Our findings link physiological traits and community composition, and strongly suggest that existing spatial variation in both free CO₂ and bicarbonate availability provide important structuring influences on community composition of submersed freshwater vegetation.