Characteristics Of Sensible Heat Flux Measured By Large Aperture Scintillometer(LAS) And Its Improvement Of Energy Flux On Grid Scale From Numerical Model

The common covariance technique Eddy Covariance (EC) has a smaller observation on flux measurement, which results a large deviation compared with the flux of numerical model grid scale, and the observation doesn’t apply to the improvement of the land surface model parameterization scheme. Looking for new ways to improve energy flux observation instead of EC is a key aspect of improving the land surface model parameterization schemes. The emergence of Large Aperture Scintillometer (LAS) solved the traditional observation EC well. With the extensive application of the LAS, previous workers did a lot of meaningful research on LAS. However, there is still a lot of research space on the systematic study of sensible heat flux characteristics observated by LAS overall, LAS’s improving observation of energy flux of numerical model grid scale compared to EC is also rarely involved. For this reason, On the one hand, based on the data of synchronous observations measured during January, June and September,2010in Ding-xi station and January and June,2010and June and July,2012in Qing-yang station over the Loess Plateau, characteristic quantity T*of sensible heat flux measured by LAS is initially analyzed first. Secondly, physical factor of sensible heat flux by LAS is analyzed. Thirdly, from the start of an important reason underlying surface inhomogeneity which causing difference of sensible heat flux measured by LAS and EC, the affect of the underlying surface inhomogeneity is analyzed. Finally, With reducing EC’s energy balance unclosure as a starting point, the affect of difference of sensible heat flux measured by LAS and EC caused by a physical process mesoscale vertical advection is studied. On the other hand, combined with the simulated data from typical land surface model CLM and NOAH, with comparing the fitting between observations of two instruments and results of model simulations, LAS’s improving observation of energy flux of numerical model grid scale compared to EC is analyzed. The main conclusions are as follows: (1) Through analyzing the relationship between the characteristic scale of temperature T*and its difference in the process of sensible heat flux measured by LAS and EC and the meteorological elements of surface layer such as wind and potential temperature gradient, we find that:There is a very good correlation between T*LAS measured by LAS and T*LAS measured by EC on the underlying surface over Loess Plateau with the correlation coefficient up to0.955and the fitting linear trend coefficient is1.482. The prevailing wind direction is NNE and SE after we make a wind statistics for September, the results show that the correlation coefficient is0.960and the fitting linear trend coefficient is1.349in NNE wind direction while in SE wind direction the correlation coefficient is0.968and the fitting linear trend coefficient is1.619, which indicates that wind direction has a significant influence on T*. There is a good linear correlation between potential temperature gradient and T*, but the correlation between T*LAS and potential temperature gradient is better than that between T*LAS and potential temperature gradient.(2) Through analyzing the spatial and temporal difference of sensible heat flux measured by LAS and EC over different underlying surface and the relationship between the difference in the process of sensible heat flux measured by LAS and EC and the physical factors of underlying surface such as net radiation, wind direction, wind speed and stability, we find that:The difference of sensible heat flux by LAS and EC is related to the heterogeneity of underlying surface with the difference more obvious over the more complicated underlying surface. Net radiation Rn is the main driving factor to sensible heat flux. The value of HLAS-HEC in the growing season is greater than the value in the non-growing season and positively correlated with Rn. Wind direction has a significant effect on sensible heat flux by LAS and EC. The presence of low-frequency volution in convective boundary layer issues in lower flux statistics to EC, which results in that:HLAS/HEC decreases with the increasing of wind speed and increases with the increasing of z/L.(3) On the basis of quantifying the degree of inhomogeneity of underlying surface, not only the relationship between underlying surface inhomogeneity and surface temperature variability but also the impact of underlying surface inhomogeneity causing difference between LAS and EC is analyzed. The results show that:The degree of inhomogeneity of underlying surface η in Qingyang station has a good correlation with surface temperature variability τ, with the correlation coefficient up to more than0.566. The degree of inhomogeneity of underlying surface η has a good consistency with the difference of sensible heat flux by LAS and EC HLAS-HEC, with the correlation coefficient up to0.634. The more uneven the underlying surface is, the lager the difference by LAS and EC is. And when the underlying surface is completely uniform, the difference tends to zero. Which indicates that:the inhomogeneity of underlying surface has a significant influence on difference of sensible heat flux by LAS and EC. Analyzing the two prevailing wind, the correlation coefficient of η and HLAS-HEC is0.430and the fitting linear trend coefficient of HLAS and HEC is1.279in E-SE wind direction, while in SW-W wind direction the correlation coefficient of η and HLAS-HEC is0.680and the fitting linear trend coefficient of HLAS and HEC is1.297. Which indicates that:the greater the inhomogeneity of underlying surface affects, the larger the difference of sensible heat flux by LAS and EC is.(4) With an important reason EC’s energy balance unclosure which causing difference of sensible heat measured by Large Aperture Scintillometer (LAS) and Eddy Covariance (EC) as a starting point, from the start of a physical process mesoscale vertical sensible heat advection, the contribution of vertical sensible heat advection is corrected to sensible heat flux directly measured by EC as a part of effective surface energy, and the difference size of sensible heat flux by LAS and EC before and after correction is compared. The results show that:The movement near the ground in Ding-xi station is mainly upward with a peak value of0.055m/s, which provides the necessary dynamic conditions for vertical advection. In particular, the region stands in the semi-arid, the surface is heated significantly by solar radiation, leading to a large temperature gradient with a peak value of0.31K/m, which provides energy basis to produce vertical advection. After adding the contribution of vertical advection to sensible heat flux directly measured by EC, the daily average of EC’s imbalance error of land surface energy reduces from77W/m2of direct observation to65W/m2in daytime, with the daily average of surface energy unclosure reducing from0.28to0.23. As a result, the fitting linear trend coefficient between sensible heat flux measured by LAS and EC is reduced from1.330of direct observation to1.132, which means that the difference between them can be reduced after the correction.(5) Through analyzing the comparison of sensible heat flux observed by LAS and EC and simulated by CLM and NOAH, the influence of vegetation over the observation and simulation of sensible heat flux and the improvement of surface energy balance from LAS, we find that:The result of surface energy observed by LAS improves the balanced error which by EC. The fit coefficients and related coefficient between the observed value from LAS and the simulated value from model is bigger than that between EC and model, which reduced the fault bit between simulation and observation. The difference between observation and simulation of sensible heat flux has a significant response to vegetational cover, because the improved status of LAS to EC in growing season is better than which in non-growing season, which tells us that the heat flux observed by LAS is more superior than what by EC over inhomogeneous land surface. There is a great difference between both sensible heat flux and latent heat flux observed by EC and which simulated by model, while the observation of LAS reduced the fault bit, which improved the Surface energy balance.