Flow and turbulence near the edge of a forest canopy: Wind tunnel and numerical experiment

Most towers used to measure CO2 flux are set up in complex terrain in Asia (e.g., AsiaFlux homepage: http://asiaflux.yonsei.ac.kr/). The insufficient fetch with regard to such sites is frequently a problem. Numerical and wind tunnel experiments have been conducted on the flow and turbulence near the edge of forests. However, the results from such experiments are not always similar. We conducted numerical experiments, both RANS and LES, and a wind tunnel experiment to investigate the flow and turbulence near the edge of a forest canopy. The wind tunnel experiment was conducted in the AIST wind tunnel, which has a length, width, and height of 20 m, 3 m, and 2 m, respectively. The forest canopy in the wind tunnel was modeled with wire mesh equally spaced along the flow in an area of 5 m X 2 m X 0.1 m on the floor. All the results showed that the area was divided into three regions near the edge of the forest canopy, that is, a deceleration zone of wind velocity near the ground, an acceleration zone of wind velocity, and a constant velocity zone in the canopy interior. The wind velocity even had a negative value under some conditions. For this zoning, turbulence generated near the top of the canopy played an important role. The acceleration zone was generated by the momentum transport with the turbulence generated near the top of the canopy, which carried a high wind velocity above the canopy into the canopy. The numerical results suggest that the usual RANS model could not accurately reproduce the features in these regions. Since the turbulence carried CO2 only in the upper foliage near the edge of forest, the NEE measured near the edge is likely to be underestimated. The fact that too much NEP was estimated with the eddy covariance method in SKR (Thailand) and TKY (Japan) in the AsiaFlux sites could be attributed to this anomaly; however, further investigation will be necessary.