Application Of Four Methods Of Calculating Surface Turbulent Heat Fluxes In Western China And Error Analysis

Accurate estimates of the surface turbulent heat fluxes have always been the central and practical concern in atmospheric science and hydrometeorology.With the improvement of numerical weather prediction and finer climate model,it is important to improve the accuracy of surface turbulent heat flux and further provides accurate lower boundary to the model.Numerous methods in calculating surface turbulent heat flux have been developed with the increased observation data from different land surface,but most of them exists different uncertainties.When these methods are applied over western China with complex land use and topographic heterogeneities that are widely spread,the results would be affected by different error sources.Therefore,it is necessary to evaluate the performance of different methods estimating surface turbulent heat flux in western China from the perspective of the methodology.These methods are eddy covariance(EC)method,variational method,Noah MP land surface model,and the Ensemble Kalman filter(EnKF)assimilation method,respectively.This thesis focuses on analyzing the possible error source in these methods,and partially would be solved by proposing the improvement scheme,thus the accuracy of the surface turbulent heat flux would be improved significantly.The main results are summarized as follows:(1)For large uncertainties of the surface turbulent heat flux measured by EC method over one specific type of oasis terrain in western China,the error sources are quantitively analyzed in this paper.Firstly,the results of the EC method over heterogeneous oasis show that one possible error source may be from large vertical advection and downward vertical velocity over oasis,especially before and after midday.The daily root mean square(rms)value of sensible(latent)vertical advection flux contributes 16.5(7.4)% to the surface sensible(latent)heat flux.Secondly,to supply the representative of the conclusion obtained from single EC measurement in oasis,the representative error of single EC method at different locations over oasis-desert heterogeneity are further studied using large eddy simulation in WRF-LES,and the result indicates that the EC fluxes at different locations in oasis have different representative of the area-averaged flux,whichis influenced by turbulent organized structure(TOS)and secondary circulation(SC)induced over strip-like heterogeneous oasis and desert.The sum of sensible and latent heat flux caused by TOS and SC accounted for area-averaged flux for both heat and water vapor flux is high in the middle and high layer in convective boundary layer,causing that the EC flux measured by single point cannot represent the area-averaged flux completely.Therefore,the flux measured by single EC method underestimates the area-averaged flux over some locations.(2)For the problem that the accuracy of surface turbulent heat flux estimated by variational method would be affected over some type of land surface with large energy storage,it is solved by proposing two approaches that introduce the surface energy storage into the cost function in variational method for improving the estimates of surface turbulent heat fluxes.In the first approach,each of the energy storage terms is directly calculated based on available observations,and in the second approach,the total energy storage is fitted by the piecewise linear regression function.The heat flux estimates are validated with the EC measurements at two carefully selected stations with different land covers and weather conditions in Northwestern China over Daman(canopy energy storage and soil heat storage caused by irrigation)and east of the Tibetan Plateau(soil freeze-thaw process at near-surface soil layer with large energy absorbed/frozen),respectively.In contrast to the variational method without considering the energy storage in the cost function,two new approaches have improved the heat flux estimates,with the first approach being slightly better,especially around midday and/or under strong unstable conditions.If there are no sufficient observations in calculation each energy storage term,it is also reasonable that the calculated/fitted energy storage with the measurements could be transferred for the heat flux estimates in the other time period.(3)Using the observations at Daman oasis station located at Zhangye City,Gansu Province,the Noah-MP land surface model is used to calculate the surface turbulent heat flux,and the error sources are analyzed carefully.Firstly,the results by default schemes show that net radiation,sensible heat flux and latent heat flux are well simulated in sunny days.The deficiencies of the default scheme are that the latent heat flux and net radiation is less than the observations during midday,the latent heat flux of plant transpiration is underestimated,and the fraction of soil evaporation is overestimated in sunny days.Secondly,the error sources are analyzed from physical process,and the default parameterization schemes are replaced by other parameterization schemes.Both the Cai11 surface albedo scheme and/or Jarvis stomatal resistance scheme could partially solve the underestimation of net radiation and latent heat flux,improving the accuracy of simulated net radiation and latent heat flux,reducing the fraction of soil evaporation and increase the fraction of plant transpiration effectively.The Alpha soil resistance scheme can reduce thefraction of soil evaporation.(4)The En KF assimilation method has been used to improve the estimates of surface turbulent heat flux using the observations at SACOL station over the native vegetation terrain which located at Yuzhong County,Gansu Province in Loess Plateau.The results show that:comparing with the results by Noah-MP without assimilating any observation data,assimilating the observed near-surface soil water content could improve the underestimated sensible heat flux and the overestimated latent heat flux effectively,especially during 9:00-12:00 LT.Finally,the sensitivity tests have been carried out to investigate the influence of observation error and background field error on the results of EnKF,and the results shows that: the analysis value becomes accurate with the decrease of observation error.The flux estimated by EnKF increases with the inflation coefficient.Therefore,both the high quality observation data and the accurate sample that could represent the true background error covariance are key factors to improve the estimation of surface turbulent heat flux by EnKF.In a word,four kinds of estimation methods have its advantage and disadvantage,and their performances are all feasible for estimating the surface turbulent fluxes in western China.However,it is hard to compare the advantage and disadvantage of them each other for lack of different observations.Researchers can choose any one method to use with considering different observations and underlying surface.