Factors Controlling CO2 Exchange at Harvard Forest on Hourly to Annual Time Scales

Carbon dioxide eddy flux measurements at Harvard Forest have revealed annual carbon sequestration ranging from 1.1 to 2.4 MgC/ha (mean ~ 2 MgChayr) from 1991 to 2000. The observed interannual variations reflect short-term (hourly to monthly) response of the ecosystem to environmental forcing (temperature, sunlight, soil moisture) as well as delayed impact of climatic variation on factors such as decay of prior year litter, or mortality or morbidity due to drought stress. Long-term ecological trends (e.g. succession or accumulation of CWD) also likely affect C uptake at Harvard Forest. The role of these processes in controlling carbon exchange at Harvard Forest has been explored through a series of complementary modeling studies. A simple, ecophysiological based modeling study has been conducted to: (1) quantify the instantaneous ecosystem response to controlling climatic variables (temperature and sunlight), averaged over the decade and (2) identify and quantify the mechanisms responsible for inter-annual deviations of ecosystem carbon exchange from the decadal mean climatic response. A first order empirical model was developed to quantify mean ecosystem response to environmental forcing. An empirical model employing a simple respirationphotosynthesis function for ecosystem response to insolation and temperature accounts for nearly 90% of hourly variance of NEE, but very little of the monthly and inter-annual variances. To understand the role of soil moisture in controlling NEE variation on the monthly to inter-annual time scale, a two layer bucket type soil hydrology model was developed. Deviations of the ecosystem behavior from the mean short-term ecosystem response, defined as residuals from the ecophysiological based empirical model, were compared with patterns in the simulated surface layer and deep layer soil moisture. Surprisingly, little correlation was found between the empirical model respiration residuals and soil moisture, but deep layer soil moisture and model GEE residuals in the late summer were significantly correlated. When supplemented with a soil moisture correction term, the simple empirical model explains ca. 13 of the observed interannual variation in mean GEE for August and September. The empirical model explains only a few percent of the late summer variation in GEE when soil moisture is excluded. The influence of respiration from short-lived carbon pools (litter-pool and fine debris) on monthly to inter-annual scale variations in NEE was also investigated employing a monthly time step respiration - mass balance litter-pool model. The results of this investigation indicate the decay of short-lived carbon pools contribute to inter-annual variations of NEE. However, the decay of litter and fine - debris was not conclusively established as the driving mechanism behind inter-annual variations in CO2 sequestration at Harvard Forest.