The7-year Variability Of Carbon And Water Fluxes Of An Oak-dominated Forest In The USA

Carbon (C) and water fluxes of temperate deciduous forests are sensitive to changing climate. Therefore, the multiyear measurements are required for understanding the response of C and water fluxes to the environmental and biophysical variability at an ecosystem scale. The goals of this study were to quantify daily, monthly, seasonal, annual and interannual variability of C and water fluxes and environmental and biophysical factors; evaluate the effects of these factors on C and water fluxes among7years. Net ecosystem exchanges of C and water over a7-year period (2004-2010) were measured using the eddy-covariance technique in a60-yr-old mixed oak woodland in northwestern Ohio, USA. Over the7years, at different time scales, the variability of C fluxes (NEE, GEP and ER) and water flux (ET) were controlled by the environmental and biophysical factors.(1) The variability of environmental and biophysical factors:mean annual Ta was9.9℃and P was824.3mm over the7-year period, which were higher than the averages of last30years (1971to2000). Year2004had the lowest P of668mm yr-1while2006had the highest P of1019mm yT-1. Over the7years, the changing ranges of the aridity/humidity index (Ia) were from0.9in2006to1.5in2010. Year2006was the only humid year with the lowest Iα of0.9. Year2006had the most annual P (1019mm yr-1), the highest annual Ta (10.7℃), a highest VWC (18.3%) summer, a wettest autumn (P=143.3mm; VWC=19.56%) and a warmest and wettest (Ta=1.7℃; P=238.9mm) winter, Year2010was the only semiarid year, with the highest Ia of1.5. Year2010had the lowest VWC (18.5%) and relative higher Ta (10.5℃). Ta, Ts5, PAR, VPD and LAI reached peak in summer and bottom in winter. VWC presented maximum in spring and minimum in summer. Each warmer summer followed a warmer spring with soil water stress. Annual Ts5increased and VWC decreased significently.(2) The responses of C fluxes to the environmental variability:This forest acted as a net C sink in each year, with mean (STD) of NEE, GEP and ER of-339.2(34.4),1552.3(82.2) and1213.1(83.9) g C m-2yr-1, respectively. Year2006had the highest ER and the least net C uptake year (1342and289g C m-2yr-1, respectively), because of the hot and wet conditions. Year2010had the second least net C uptake year (298g Cm-2yr-1), because of the hot and dry conditons. The wet warming increased ER more than GEP and thus C uptake decreased. The dry warming suppressed GEP more than ER and thus C uptake also decreased. The moderate warming with sufficient water enhanced GEP more than ER and thus C uptake increased. Monthly spring LAI, summer PAR and winter Ta were the most important single determinants of seasonal NEE (the highest p<0.03) for each season. Nevertheless, no factors explained NEE significantly in autumn. PAR and VPD had no significant effects on C fluxes in autumn and winter. Summer, autumn and winter Ta and spring Ts5were the main determinants of seasonal ER (p<0.03). Spring Ts5, summer PAR and autumn Ta were the main determinants of seasonal GEP (p<0.01) and no factor explain GEP in winter. At the interannual scale, all factors influenced monthly C fluxes. The contributions on C balance were vegetation dynamics (LAI)> temperature (Ta&Ts5)> energy (PAR&VPD)> water (P and VWC)(p<0.0001; N=84). Importantly, seasonal and interannual decreased GEP was both correlated with decreased ER and less C uptake (p<0.001&0.02, respectively).(3) The response of water flux to the environmental variability:Although annual ET was less than annual P in all years, the7-year mean of summer ET was8%higher than that of P. Annual ET varied from578mm in2009to671mm in2010corresponding to ET/P ratios of68.3%and95.8%, respectively. Year2010had the highest ET due to the highest annual Ts5(12.1℃), the least VWC (18.5%) and highest PAR, comparing with other six years.2009had the lowest ET resulting from the second annual lowest Ta (9.4℃), intermediate VWC and PAR. Warming climate increased ET under a drought condition but a low temperature condition suppressed ET when water was not limited. Monthly spring LAI, summer PAR, autumn and winter Ta were the most important single determinants of seasonal ET for each season(p<0.02), respectively. Except for winter, water (P&VWC) had no significant effect on seasonal ET. At the interannual scale, all the factors could influence monthly ET. The contributions on ET were vegetation dynamics (LAI)> temperature (Ta&Ts5)> energy (PAR&VPD)> water (P&VWC)(p<0.0001; N=84).