Airflow and CO2 transport over double forested hills

Accurate quantification of net ecosystem-atmosphere exchanges (NEE) of mass and energy is a fundamental and critical step in reducing the uncertainty from the potential effects of climate change on ecosystems as sources or sinks for atmospheric CO2. The eddy covariance technique has proven to be a useful approach to quantify net ecosystem C sequestration when strong turbulent mixing occurs, while measurements carry significant advection errors when flux sites are located in complex terrain (Massman and Lee,2002). Although many multiple-tower advective experiments (Feigenwinter et al.2008) have been conducted, the impact of hill geometry on advective CO2 flux remains unknown. We examined the impacts of hill geometry and canopy structure on airflow and CO2 dynamics by performing numerical simulations over double forested hills. Varying hill geometry is performed by altering the height of the hill (H) with constant length scale (L). Our results show that recirculation regions are formed behind hills and create complex patterns of air flow(Figure 1a). We found that H/L=0.8 is a threshold value of flow-pattern formation in the recirculation region: (1) below 0.8 the reversed flow is characterized by intermittent positive and negative streamwise velocity; (2) at 0.8 one vortex is formed; and (3) above 0.8 a pair of counter-rotating vortices are formed. Depth of recirculation region increases with increasing H/L. Contribution of advective fluxes to CO2 budget is topographic-dependent (Figure 1b). When H/L<0.8, horizontal advective flux Fh is positive on windward slope and negative on leeward slope. When H/L≥0.8, recirculation vortices lead to positive and negative Fh on both windward and leeward slope in the valley. Vertical advective flux Fv is opposite in sign to Fh. Fh and Fv cannot exactly offset in magnitude. Depth-integrated advective fluxes indicate significant variation in fluxes across double-hill. Domain-integrated total advective flux decreases with increasing slope as H/L≤1.0. This indicates that gentle slopes can cause large advection error, which is opposite to the traditional thought. Acknowledgement: This research was supported by NSF Grants ATM-0930015, CNS-0958379 & CNS-0855217, PSC-CUNY ENHC-42-64 & CUNY HPCC.