Delineating the effects of water demand and availability on assimilatory and respiratory carbon fluxes in managed forest ecosystems in northern Wisconsin

The diurnal and seasonal variation within ecosystem and differences between forest types in regard to assimilatory and respiratory carbon fluxes is analyzed as a function of vapor pressure deficit (VPD) and soil moisture potential (M). The work was undertaken to address the autocorrelation between environmental parameters (temperature, moisture and radiation), as they are interdependent and have similar diurnal and seasonal patterns. The method of path coefficient analysis, used to delineate moisture effects, is a more powerful tool compared to residual analysis commonly used in environmental regulation analyses of carbon fluxes. Net ecosystem exchange of carbon (NEE) was measured in five closely spaced forest ecosystems (mature mixed hardwood, mature red pine, pine barrens, young mixed pine and recent hardwood clear–cut), in a managed northern Wisconsin (USA) landscape using eddy covariance method. The dependence of diurnal and seasonal patterns and seasonal sums of estimated ecosystem respiration (R) and gross ecosystem production (GEP) on radiation, soil and air temperature, VPD and M was analyzed using regression and path coefficient analysis techniques. The analysis of the seasonal course of light response parameters of growing season NEEc clearly differentiated between coniferous and deciduous stands, whereas the effect of disturbance was only marginally significant. The respiration estimates from light response relationships of NEE showed minimal differences among the stands, whereas base respiration estimates based on night-time data exhibited a negative logarithmic relationship with mean tree age (R²=0.5) and a strong positive linear relationship with coarse woody debris (R²=0.74). No statistically significant differences in the temperature sensitivity of ecosystem respiration could be detected between sites at either seasonal or annual time scale. Our results suggest that disturbed forests in upper Midwest become carbon sinks within 10 years after stand replacing disturbance and achieve net carbon balance comparable to mature stands in 20-30 years, sooner than assumed in commonly used forest growth models.