The Development And Application Of A Lake-atmosphere Coupled Model

A one-dimensional (1-D) physically based lake model was coupled to the WeatherResearch and Forecasting (WRF) model version3.2developed by the National Center forAtmospheric Research to dynamically simulate physical processes of lakes and their effectson weather and climate at local and regional scales. Our study area is focused on the GreatLakes. This coupled model realistically reproduces the lake surface temperature (LST) at abuoy station in a shallow lake (Lake Erie), while it generates strong LST biases ranging from-20to20oC at a buoy station in a deep lake (Lake Superior). Through a large number ofsensitivity tests, we found that the biases in the deep lake LST simulations result from thedrastic underestimation of heat transfer between the lower and upper parts of the lake througheddy diffusion. Additional tests were made to calibrate the eddy diffusioncoefficient inWRF-Lake. It is found that when this parameter is multiplied by a factor ranging from102to105for various lake depths deeper than15m, the LST simulations for the deep lake buoystation show good agreement with observations, and the bias range reduces to±4oC.Essentially, the enlarged eddy diffusivity strengthens heat transfer within the lake columns inthe deep lake, which is significantly underestimated in the lake model without calibration.Validation simulations were carried out to simulate winter lake surface temperature(LST) and its effects on precipitation for the the whole Great Lakes using the calibratedcoupled model.Our study period is December, January, and February over2003-2008. Theresults show that the simulated LSTs for the Great Lakes with WRF-Lake agree very wellwith observations and have better quality than those from the North American RegionalReanalysis (NARR) system that show strong warm biases. The coupled model alsoreasonably reproduces the observed LIC for the Great Lakes during the three winter months.In addition, lake-effect precipitation is more realistically simulated with WRF-Lake than withWRF driven by NARR LST, indicating that the LST plays a very important role in theprecipitation processes over the Great Lakes and their surrounding areas. However, the warmbiases of the NARR LST generate stronger surface water and heat fluxes in WRF and reduce the stability in the lower atmosphere, resulting in overestimated precipitation. The coupledWRF-Lake model adds to a strong capacity in dynamically predicting lake processes andlake-atmosphere interactions.The performance of the one-dimensional (1-D) thermal diffusion lake model for LakeTaihu has been examined using observations from August11to28,2010. The originallysimulated lake surface temperature has significant biases, and the simulated temperaturediurnal cycle range is smaller than observations. Eighteen sensitivity experiments wereconducted with inclduing the ecological environment and pollution conditions in Lake Taihu,the sensitivity tests show that three parameters that depend on the lake type should bechanged as follows: increase extinction coefficient by a factor of3, change the fraction of thenon-reflecting shortwave radiation absorbed by the lake surface from0.4to0.8; parameterizethe roughness length a function of wind and lake depth. The new lake model using the newparameters has greatly improved the temperature simulations when compared withobservations and original simulations. The vertical thermal structure in Lake Taihu is wellreproduced using the three new parameters in the lake model. The simulated sensible flux andlatent heat flux are also improved significantly.At last, the coupled WRF_Lake model withnew parameters was tested in the Lake Taihu. Results show the coupled model hassignificantly improve the model simulation ability in Lake Taihu when compared with thatsimulated by WRF without including a lake scheme. Those researches in Lake Taihu providea good guide for the application of the lake model in the same type of lake and necessarypremise for the study of lake-atmosphere interactions.