Simulation of land surface heat fluxes in permafrost regions on the Qinghai-Tibetan Plateau using CMIP5 models

The variations in land surface heat fluxes affect the ecological environment, hydrological processes and the stability of surface engineering structures in permafrost regions of the Qinghai-Tibetan Plateau (QTP). Based on observation data from a meteorological station in the Tanggula site in 2005, which is located in a permafrost region on the QTP, the performances of seventeen selected the phase 5 of the Coupled Model Intercomparison Project (CMIP5) were evaluated. The results showed that these simulations did not perform well using sensible heat flux, downward shortwave radiation or upward shortwave radiation, and differences exist among the models. The average multimodel ensemble results were similar to the observed land surface heat fluxes. The results revealed that the monthly average latent heat flux and the net radiation were small in December and January and large in May, June and July. The fluctuation in the soil heat flux was well correlated with the net radiation, and the sensible heat flux was negative in January and December in northwest of the Plateau. The latent heat flux was the strongest over the southeastern QTP from May to August, and it decreased over the northwestern QTP. In contrast, the sensible heat flux was the weakest over the southeastern QTP, and it gradually increased and became dominant over the northwestern QTP. The results also indicated that there was a good correlation between the surface heating field intensity and the net radiation, with a correlation coefficient of 0.99; this indicates stronger heating over the eastern QTP than over the western QTP and stronger heating over the southern QTP than over the northern QTP. Furthermore, the Bowen ratio was higher during the freezing and thawing stages than that during the completely thawed stage. This ratio was larger over the central and northeastern QTP and smaller along the northwest edge of the QTP, which was lower (range from -0.81 to 4.86) due to the overestimation of precipitation, a smaller difference between the simulated monthly average surface temperature and the observed air temperature, and a decrease in wind speed when using the CMIP5 models in the permafrost region of the QTP. This research provides a foundation for understanding land surface heat flux characteristics in the permafrost regions on the QTP under climate change.