| Land surface is an important part of weather,climate and earth system.The complex interaction between the land surface and the atmosphere is achieved through the turbulent exchange of surface momentum,heat and moisture.Therefore,accurate calculation of surface flux is of great significance for understanding surface heat source distribution,atmospheric boundary layer structure,atmospheric circulation and climate change research,and improving numerical simulation capabilities.The classical Bulk Transfer Model is based on the Monin-Obukhov similarity theory,which is not suitable for non-uniform underlying surface and morning-evening transition period.The calculation of latent heat flux in this model is an empirical formula,and the model usually uses surface wind speed and surface roughness to parameterize the turbulent transport coefficient,so the model has a large uncertainty.In addition,the surface fluxes given by this model do not necessarily satisfy the energy balance because there is no energy balance condition as a constraint.The Maximum Entropy Production(MEP)model takes net radiation,surface temperature and surface specific humidity as inputs,and surface turbulent fluxes are derived from the distribution of effective surface radiation fluxes,always satisfying the energy balance.The model does not depend on vertical gradient variables and does not require surface wind speed and surface roughness as input variables,which greatly reduces the uncertainty of surface heat flux estimation,and provides an effective way to improve the observation and simulation of existing surface heat flux processes.Previous studies have shown that the MEP model can reasonably estimate the surface heat flux under uniform underlying surface conditions such as ocean,bare soil,forest and Qinghai-Tibet Plateau.However,the applicability of MEP model in complex underlying surfaces and the effect of MEP model in numerical model needs to be explored.Accordingly,this paper conducts relevant research in three aspects:Firstly,the surface sensible and latent heat fluxes at three sites at complex underlying surface(complex terrain and urban)are estimated using observation data and the MEP model,and then verified and evaluated using Eddy Covariance(EC).In addition,the results of MEP are compared with the daily average surface sensible heat fluxes and latent heat fluxes from another simple non-gradient calculation method(M20)and seven reanalyses(including: NCEP/NCAR,NCEP-DOE,MERRA2,ERA-5,ERAInterim,JRA55,CRA),to compare and analyze their performance at complex underlying surface.The results show that the surface latent heat fluxes and sensible heat fluxes calculated by the MEP model are close to observation,and have similar daily,monthly and seasonal variation characteristics.Compared with the M20 method and reanalysis,the calculated results of the MEP model are in better agreement with the observations,with smaller root mean square errors and mean bias errors.Therefore,MEP model has high accuracy and reliability in calculating surface heat fluxes at complex underlying surface,and has good applicability.Secondly,based on the regional rapid-refresh multi-scale analysis and prediction system Short term(CMA-BJ)developed by Institute of Urban Meteorology,a control experiment was conducted using the original surface heat flux scheme(BTM)and MEP,to evaluate the impact of the MEP model on the simulation of summer precipitation in China using average precipitation,heavy rainfall frequency and statistical scores.The results show that the difference in the comprehennsive scores of simulated precipitation between the BTM and MEP schemes increases with increasing rainfall,and the MEP scheme improves the comprehennsive scores of heavy rainfall forecasts in northern and central China.Overall,the MEP scheme produces more precipitation in the model,which improves the simulation of heavy precipitation events.The daily variation curves of MEP precipitation have higher correlation coefficients with observations,and the descriptions of the daily variation characteristics of precipitation are in better agreement with observation.Finally,the effects of the two schemes on the surface energy balance simulations in China were compared using tresidual of energy balance and energy closure rates.And the effects of the two surface heat fluxes schemes on the 2m temperature,2m specific humidity and 10 m wind speed simulations were analyzed using root mean square errors.The results show that the MEP scheme has an advantage in energy balance,and the BTM scheme may underestimate the surface sensible heat fluxes or surface latent heat fluxes,resulting in energy non-closure.From the full-field average,the two surface heat flux schemes of BTM and MEP have the same ability to simulate the near-surface meteorology.In terms of spatial distribution,the MEP scheme simulates 2m air temperature and specific humidity better in Ningxia,Guangxi and Guangdong regions with smaller root mean square errors. |
