Font Size: a A A

Thermal And Dynamical Parameterization Over A Typical Temperate Desert Steppe

Posted on:2015-03-05Degree:DoctorType:Dissertation
Country:ChinaCandidate:G ZhangFull Text:PDF
GTID:1260330428457590Subject:Science of meteorology
Abstract/Summary:PDF Full Text Request
The land-atmosphere interactions in coupled weather and climate models are describednumerically by land surface parameterizations. The improvements on land surfaceparameterizations are at the price of increasing number of parameters specified and moreaccurate parameterization functions used. Over the arid and semi-arid region, it is still adaunting challenge for land surface models (LSMs) to correctly represent surface heatexchange for water-limited desert steppe ecosystems. This study aims at improving theability of the Noah LSM to simulate surface heat fluxes through addressing uncertainties inprecipitation forcing conditions, rapidly evolving vegetation properties, soil hydraulicproperties (SHPs), and key parameterization schemes.3-year (2008-2010) observed surfaceheat fluxes and soil temperature over a desert steppe site in Inner Mongolia of China are usedto verify model simulations. Through the sensitivity analysis algorithm and optimizationmethod, we analyze the influences of parameter uncertainties and selected sensitiveparameters on flux simulations. The major conclusions are as below:(1) Parameters have different contributions to different flux simulations. We havedone sensitivity analysis of27parameters in Noah LSM for net radiation (Rn), sensible heatflux(H), and latent heat flux (LE) respectively, by employing the distributed evaluation oflocal sensitivity analysis (DELSA) and the shuffled complex evolution algorithm, Universityof Arizona (SCE-UA). The study indicates that different parameters have substantiallydifferent influence on each flux and the same parameter also has different contributions todifferent fluxes. For Rn and H, there are3~5most sensitive parameters. But for LE, there aremore sensitive parameters and they are more dispersed. The parameter czil has substantialinfluences on all the three fluxes.(2) Distinct seasonal and interannual variabilities in canopy resistance wereidentified in2008-2010and the new exponential function of water stress for the widelyused Jarvis scheme can improve the LE simulation over the desert steppe. Based onprincipal component regression method, the analysis of the relative contribution of five majorenvironmental factors [soil-water content (SWC), leaf-area index (LAI), photosyntheticallyactive radiation (PAR), VPD, and air temperature] to canopy resistance shows thatcanopy-resistance variation is most responsive to SWC (with>35%contribution), followedby LAI, especially for water-stressed soil conditions (>20%influence), and VPD(consistently with an influence of approximately20%). Canopy-resistance variations do notrespond to PAR due to the small interannual variability in PAR during the three years. Theseanalyses are used to develop a new exponential function of water stress for the widely usedJarvis scheme, which substantially improves the calculation of canopy resistance and latent heat fluxes, especially for moist and wet soils, and effectively reduces the high bias inevaporation estimated by the original Jarvis scheme. This study highlights the importantcontrol of canopy resistance on plant transpiration and growth for the investigated desertsteppe site with a relatively low LAI.(3) The changing precipitation, vegetation parameters, surface coupling strengthand soil hydraulic properties impact the simulations of surface heat flux over the desertsteppe site of Inner Mongolia. The proper seasonal distribution of precipitation along withmore realistic vegetation parameters can improve the simulation of H and the seasonalvariability of LE. Changes in precipitation seasonal distribution due to climate change wouldcause significant consequences in energy flux simulations when the precipitation is relativelyample over this desert steppe. But when the site suffers from a drought (for instance,2010),the difference in precipitation distribution has less influence on flux simulations.Relating the Czil coefficient in the Noah surface exchange coefficient calculation with thecanopy height improves the calculation of surface coupling strength, the simulated H and thediurnal range of soil temperature over this site compared to using the default constant Czil.The exponential water stress formulation we propose for the Jarvis scheme improves thepartitioning between soil evaporation and transpiration, and the simulation of LE to somedegree. The uncertainties in observed soil moisture and the inconsistency between observedand simulated soil moisture could contribute to the bias between the development ofexponential soil-water-stress function and its application in Noah LSM. It is found that thesurface energy fluxes are very sensitive to SHPs.(4) Modified parameterizations and optimal parameter set reduce the uncertaintiesof the model, and the modified parameterizations physically improve the fluxsimulations. Parameter calibration can efficiently improve the model performance inlong-term and dynamical properties of fluxes. Parameters have different influences ondifferent fluxes, and the interaction between parameters substantially impacts the surfacefluxes. The modified parameterizations obviously reduce the systematic error and each termof mean squared error decomposition in flux simulations..
Keywords/Search Tags:Typical temperate desert steppe, Thermal parameterization, Dynamicalparameterization, Land-atmosphere interaction, Impact, Sensitivity analysis
PDF Full Text Request
Related items