Measurements of long-term exchange of CO2 between the ecosystem and the atmosphere have improved the scientific understanding of the role that terrestrial ecosystems play in the global carbon budget. Eddy covariance (EC), a micrometeorological technique, allows estimation of continuously measured net ecosystem exchange of CO2 (NEE) between the atmosphere and the ecosystem. While the EC method is designed to collect data continuously, gaps in the continuous data are inevitable because conditions frequently do not meet the assumptions of the EC method, sensors fail or are fouled by wildlife, and systems are down during maintenance. The goal of this study was development of improved model estimates for carbon dynamics of the Everglades while maintaining data structural variance when imputing (“gap-filling”) missing values. Here, we utilized EC-based carbon flux data from two Everglades freshwater marsh sites and applied a non-linear regression approaches to Michaelis-Menten and Van’t Hoff functions to estimate the functional parameters of physiological activity. We used a 31-day moving window to determine annual pattern of these parameters. Multiple regression was used to quantify the relationship between these parameters and the 31-day running mean of environmental variables such as relative humidity, wind speed, pressure, water level and vapor pressure deficit (VPD).
Preliminary results showed that in addition to photosynthetically active radiation, significant variability in CO2 assimilation was explained by wind speed, atmospheric pressure and water level. The variability in respiration rates was explained by air temperature, with significant effects of wind speed, pressure, water level, and VPD explain most of the variability of the respiration (beyond that of air temperature). Using these results, we investigated causality interaction among environmental variables that ultimately alters NEE leading to a new regime in an ecosystem, which would then affect environmental variables at varying time domain. We will expand this investigation using path analysis and structural equation modelling to conceptualize and confirm the changes in ecological functional parameters in response to variation in environmental factors. The mechanistic causal framework and models developed will advance our understanding of physiologically-driven carbon dynamics of wetland ecosystems in time and space in the context of changing climate.