A synthesis of floodplain aquatic ecosystem metabolism and carbon flux using causal criteria analysis

The transformation of solar energy into organic matter by autotrophs (gross primary production [GPP]) and the use of that energy by autotrophs and heterotrophs (ecosystem respiration [ER]) describe the total energy available to support food webs. Rates of GPP and ER vary with temperature, light, hydrology, nutrients, and organic matter supply and quality yet despite their obvious importance, spatiotemporal variation of metabolic patterns among floodplain habitats, and their relationship to inundation dynamics remain unclear. We set out to review the peer-reviewed literature surrounding the influence of the magnitude, frequency, and duration of floodplain inundation on aquatic ecosystem metabolism and carbon flux by rigorously testing a suite of cause-effect hypotheses using a causal criteria analysis. Causal criteria analysis is a literature synthesis approach developed to address a lack of experimental data and subsequent weak inference of causal relationships. We found support for 3 of the 14 hypotheses we tested relating to putative causal relationships: (1) large floods transfer more carbon from floodplains to the river channel than small floods via the increase in inundation area leading to more overall leaching of floodplain litter, (2) in high turbidity floodplain habitats rates of GPP are reduced by restrictions to photic depth, and (3) a positive correlation between nutrients and GPP-generally GPP in floodplain wetlands increases with nutrient levels. We obtained inconsistent evidence for a causal relationship between macrophytes and aquatic GPP, with studies reporting both a negative influence from decreased light caused by macrophyte shading and a positive influence from structural support provided by macrophytes for periphyton growth. For the remaining 10 hypotheses, there was insufficient evidence to support causal relationships, including for any hypotheses relating to frequency or duration of floodplain inundation. Our results emphasize that despite an apparent wealth of metabolic studies in riverine ecosystems, floodplain metabolic dynamics remain poorly studied, likely due to less investment and increased difficulty compared to lotic waters. The review also highlighted aspects of floodplain aquatic ecosystem metabolism for which there are significant knowledge gaps in the literature, in particular metabolic responses to inundation frequency and duration. Our results call attention to the importance of site specificity and temporal changes when predicting putative cause-effect relationships between floodplain inundation and metabolic patterns.