| Making use of the boundary layer comprehensive observation system built with the JICA Project of Sino-Japan Joint Research Center of Meteorological Disaster at Yunnan Dali National Climatic Station, and use EDDYPRO software to complete the eddy covariance turbulence observation quality control and data correction. We compare EDDYPRO turbulent flux results with JICA fortran program results. The results of Dali data processing indicate that the results of profile method and eddy covariance method are considerably different. We analyse the surface layer kinetic characteristic of the turbulence. The turbulence parameters including stability, turbulent variance, turbulent intensity, turbulent kinetic energy(TKE), buoyancy term, shear term and so on. We estimate land surface parameters, such as aerodynamic roughness length(zo) and flux bulk transfer coefficients. Analysis of atmospheric boundary layer height characteristics which include stable boundary layer(SBL) and convective boundary layer(CBL) using the GPS Soundings. Based on the diagnostic analysis surface layer turbulent variables, In the viewpoint of the different scales interaction between turbulence-convection motion, we discuss the correlation between boundary layer turbulent components and thermodynamic process in plateau, and we find that there are significant correlation between turbulent components and vertical motion.Calculation of surface layer turbulent fluxes is very important for atmosphere numerical models. How to parameterize the turbulent fluxes is one of the key research questions in current atmosphere boundary layer study. The paper uses two common schemes (MM5similarity non-iterative Scheme A and ETA similarity iterative Scheme B) in Weather Research Forecast Model(WRF) to make offline test and intercomparison of the parameterization results with PBL eddy-correlation observation. The aerodynamic roughness length(zo) obtained on Dali boundary layer is determined before calculating the turbulent fluxes. The aerodynamic roughness length(z0) by fitting four different heights wind speed from PBL tower data under neutral condition varies significantly with season due to the obvious changes in underlaying surface during the whole year,(horsebean in winter half year and paddy in summer half year) Offline test results offer basis for evaluating online simulation effect. Based on the offline test two surface layer schemes, online simulation work is progressing gradually. This article mainly includes two parts, one is the regional simulation effects of the two surface layer schemes, another is second-order moment term in the turbulent energy equation. Because of our interest in second-order moment term,1.5-order local closure model2.5level MYNN boundary layer scheme has been used(3level MYNN scheme can not be combined with two surface layer schemes).The present study has mainly drawn the conclusions as follows:(1)Overall turbulent intensity in Dali cropland is larger than Sichuan basin, slightly larger than Litang grassland. The normalized covariance of three dimensional wind speeds obeys the power law of1/3. Buoyancy term in Dali is obviously smaller plateau Dangxiong area, and the value is close to plain area. Shear term which is considerably larger than plain area in Dali is close to plateau Dangxiong area. The aerodynamic roughness length presents seasonal variation due to agricultural activities. The flux bulk transfer coefficients in eastern Tibetan Plateau is obviously larger the western area.(2)Thermal inhomogeneity of the underlying surface on the plateau results from the strong heating source or complex topography of the middle part of the plateau. Large scale vertical motion in Tibetan Plateau is significantly correlated with apparent heat source (Q1) and moisture sink(Q2). Meanwhile, surface TKE, buoyancy term and shear term also have notable correlation with NCEP800hPa vertical montion in Dali. This result shows the heating source process drived by the plateau, and the turbulent contribution to the vertical motion. The situation in Linzhi surface layer also appear the similar correlation characteristics bellow600hPa. The further applying of satellite data and other Multi-source information, and adopt the vertical motion data provided by LAPS model in Linzhi, we find there exist more significant correlation between vertical motion and TKE. The result confirms strong vertical motion occurs with strong turbulence in the southern slopes of the Tibet plateau. The correlations attenuated above600hPa in Linzhi, the coefficients of correlation reach peak between600hpa-700hpa.(3)Under cloudless sky condition, CBL height in Dali can reach about2000m. There are statistically significant correlation between CBL height and surface heat flux. The correlation between CBL height and shear term is uncertain. Shear term produces the horizontal direction turbulence, while buoyancy term produces the vertical direction turbulence. This conclusion has been confirmed by clear physical explanation.(4)Surface layer turbulent fluxes have great significance to mesoscale numerical simulation, therefore the parameterization scheme of turbulent fluxes is one of the hot spots in current atmosphere boundary layer research. The authors of this paper use two schemes(MM5similarity non-iterative method A and ETA similarity iterative method B)in common usage in Weather Research Forecast Model(WRF) to make offline test, and compare the parameterization schemes results with PBL eddy-correlation observation. We determine the momentum roughness length(zo) before calculating turbulent fluxes. The method that nonlinear fitting four different heights wind speed under neutral condition has been used, and fitting result suggest that the momentum roughness length(zo) appear obvious seasonal variation due to the underlaying surface changes obviously during the whole year(winter half year horsebean and summer half year paddy). The turbulent fluxes result show that the vegetation sparseness degree have significant effect to turbulent fluxes calculation in Dali cropland. In general, momentum flux is estimated well for both MM5and ETA scheme. In unstable condition, MM5scheme overestimates momentum flux, however, ETA scheme underestimates momentum flux. In unstable condition, ETA Scheme significantly underestimates momentum flux in bare soil underlying surface, MM5Scheme is superior to ETA Scheme. However, ETA Scheme is superior to MM5Scheme for vegetation underlying surface. In stable condition, MM5scheme appears more lower values while ETA scheme appears more larger values. ETA scheme takes account of the difference between momentum roughness length(zo)and heat roughness length(z0h), the exchange of heat is accorded with actual condition better, and the heat flux result is superior to MM5scheme especially in bare soil or sparse vegetation. Athough ETA scheme has been used to calculate heat flux, the result is still larger than the observation. In order to solve this problem, we adopt (Zeng et al.,1998) method. This method has significant effect to correct the bare soil underlaying heat flux when the radiometric surface temperature data have been used.(5) The online results show that MM5Scheme generates larger momentum flux or friction velocity values in unstable condition, and ETA Scheme generates larger values in stable condition. These results agree well with offline test. We find that the principal types of underlying surface is grassland and forestland in southeastern side of Tibetan Plateau by MODIS data, so we think ETA Scheme results seem more plausible and reliability. The online results also show that2.5level MYNN boundary layer scheme has adequate performance to simulate TKE and turbulent terms in TKE equation. Because of the larger friction velocity values in unstable condition calculated by MM5Scheme, larger TKE appeared in MM5Scheme relative to ETA Scheme. According to the conclusion from offline in unstable condition, we think TKE simulated by ETA Scheme is more reliable. |
PDF Full Text Request