干旱区生态环境知识资源中心

To improve the simulation of the surface radiation budget and related thermal processes in arid regions, three sophisticated surface albedo schemes designed for such regions were incorporated into the Biosphere-Atmosphere Transfer Scheme (BATS). Two of these schemes are functions of the solar zenith angle (SZA), where the first one has one adjustable parameter defined as SZA1 scheme, and the second one has two empirical parameters defined as SZA2 scheme. The third albedo scheme is a function of solar angle and soil water that were developed based on arid-region observations from the Dunhuang field experiment (DHEX) (defined as DH scheme). We evaluated the performance of the original and newly-incorporated albedo schemes within BATS using the in-situ data from the Oasis System Energy and Water Cycle Field Experiment that was carried out in JinTa, Gansu arid area (JTEX). The results indicate that a control run by the original version of the BATS generates a constant albedo, while the SZA1 and SZA2 schemes basically can reproduce the observed diurnal cycle of surface albedo, although these two schemes still underestimate the albedo when SZA is high in the early morning and late afternoon, and overestimate it when SZA is low during noontime. The SZA2 scheme has a better overall performance than the SZA1 scheme. In addition, BATS with the DH scheme slightly improves the albedo simulation in magnitude as compared to that from the control run, but a diurnal cycle of albedo is not produced by this scheme. The SZA1 and SZA2 schemes significantly increase the surface absorbed solar radiation by nearly 70 W m(-2), which further raises the ground temperature by 6 K and the sensible heat flux by 35 W m(-2). The increased solar radiation, heat flux, and temperature are more consistent with the observations that those from the control run. However, a significant improvement in these three variables is not found in BATS with the DH scheme due to the neglect of the diurnal cycle of albedo. Further analysis indicates that during cloudy days the solar radiation simulations of BATS with these three schemes are not in a good agreement with the observations, which implies that a more realistic partitioning of diffuse and direct radiation is needed in future land surface process simulations.