Surface soil moisture has great impact on both meso-and microscale atmospheric processes,especially on severe local convection processes and on the dynamics of short-lived torrential rains.To promote the performance of the land surface model (LSM) in surface soil moisture simulations,a hybrid hydrologic runoff parameterization scheme based upon the essential modeling theories of the Xin'anjiang model and Topography based hydrological Model (TOPMODEL) was developed in preference to the simple water balance model (SWB) in the Noah LSM.Using a strategy for coupling and integrating this modified Noah LSM to the Global/Regional Assimilation and Prediction System (GRAPES) analogous to that used with the standard Noah LSM,a simulation of atmosphere-land surface interactions for a torrential event during 2007 in Shandong was attempted.The results suggested that the surface,10-cm depth soil moisture simulated by GRAPES using the modified hydrologic approach agrees well with the observations.Improvements from the simulated results were found,especially over eastern Shandong.The simulated results,compared with the products of the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) soil moisture datasets,indicated a consistent spatial pattern over all of China.The temporal variation of surface soil moisture was validated with the data at an observation station,also demonstrated that GRAPES with modified Noah LSM exhibits a more reasonable response to precipitation events,even though biases and systematic trends may still exist.
There is an increasing trend to incorporate the basin hydrological model into the traditional land surface model （LSM） to improve the description of hydrological processes in them. For incorporating with the Noah LSM, a new rainfall-runoff model named XXT （the first X stands for Xinanjiang, the second X stands for hybrid, and T stands for TOPMODEL） was developed and presented in this study, based on the soil moisture storage capacity distribution curve （SMSCC）, some essential modules of the Xinanjiang model, together with the simple model framework of the TOPMODEL （a topography based hydrological model）. The innovation of XXT is that the water table is incorporated into SMSCC and it connects the surface runoff production with base flow production. This improves the description of the dynamically varying saturated areas that produce runoff and also captures the physical underground water level. XXT was tested in a small-scale watershed Youshuijie （946 km2） and a large-scale watershed Yinglouxia （10009 km2） in China. The results show that XXT has better performance against the TOPMODEL and the Xinanjiang model for the two watersheds in both the calibration period and the validation period in terms of the Nash-Sutcliffe efficiency. Moreover, XXT captures the largest peak flow well for both the small： and large-scale watersheds during the validation period, while the TOPMODEL produces significant overestimates or underestimates, so does the Xinanjiang model.
Early and effective flood warning is essential for reducing loss of life and economic damage. Three global ensemble weather prediction systems of the China Meteorological Administration （CMA）, the European Centre for Medium-Range Weather Forecasts （ECMWF）, and the US National Centers for Environmental Prediction （NCEP） in THORPEX （The Observing System Research and Predictability Experiment） In- teractive Grand Global Ensemble （TIGGE） archive are used in this research to drive the Global/Regional Assimilation and PrEdiction System （GRAPES） to produce 6-h lead time forecasts. The output （precipita- tion, air temperature, humidity, and pressure） in turn drives a hydrological model XXT （the first X stands for Xinanjiang, the second X stands for hybrid, and T stands for TOPMODEL）, the hybrid model that combines the TOPMODEL （a topography based hydrological model） and the Xinanjiang model, for a case study of a flood event that lasted from 18 to 20 July 2007 in the Linyi watershed. The results show that rainfall forecasts by GRAPES using TIGGE data from the three forecast centers all underestimate heavy rainfall rates; the rainfall forecast by GRAPES using the data from the NCEP is the closest to the obser- vation while that from the CMA performs the worst. Moreover, the ensemble is not better than individual members for rainfall forecasts. In contrast to corresponding rainfall forecasts, runoff forecasts are much better for all three forecast centers, especially for the NCEP. The results suggest that early flood warning by the GRAPES/XXT model based on TIGGE data is feasible and this provides a new approach to raise preparedness and thus to reduce the socio-economic impact of floods.
Hydrological processes exert enormous influences on the land surface water and energy balance, and have a close relationship with human society. We have developed a new hydrological runoff parameteriza- tion called XXT to improve the performance of a coupled land surface-atmosphere modeling system. The XXT parameterization, which is based upon the Xinanjiang hydrological model and TOPMODEL, includes an optimized function of runoff calculation with a new soil moisture storage capacity distribution curve （SMSCC）. We then couple XXT with the Global/RegionM Assimilation Prediction System （GRAPES） and compare it to GRAPES coupled with a simple water balance model （SWB）. For the model evaluation and comparison, we perform 72-h online simulations using GRAPES-XXT and GRAPES-SWB during two torrential events in August 2007 and July 2008, respectively. The results show that GRAPES can reproduce the rainfall distribution and intensity fairly well in both cases. Differences in the representation of feedback processes between surface hydrology and the atmosphere result in differences in the distributions and amounts of precipitation simulated by GRAPES-XXT and GRAPES-SWB. The runoff simulations are greatly improved by the use of XXT in place of SWB, particularly with respect to the distribution and amount of runoff. The average runoff depth is nearly doubled in the rainbelt area, and unreasonable runoff distributions simulated by GRAPES-SWB are made more realistic by the introduction of XXT. Differences in surface soil moisture between GRAPES-XXT and GRAPES-SWB show that the XXT model changes infiltration and increases surface runoff. We also evaluate river flood discharge in the Yishu River basin. The peak values of flood discharge calculated from the output of GRAPES-XXT agree more closely with observations than those calculated from the output of GRAPES-SWB.
For the Z-R relationship in radar-based rainfall estimation, the distribution of corresponding R values for a given Z value (or the corresponding Z value for a given R value) may be highly skewed. However, the traditional power-law model is physically deduced and fitted under the normal-distribution presumption of radar wave echoes associated with a rain rate value, and it may not be very appropriate. Considering this problem, the authors devised several generalized linear models with different forms and distribution presumptions to represent the Z-R relationship. Radar-reflectivity scans observed by a CINRAD/SC Doppler radar and 5-minute rainfall accumulation recorded by 10 ground gauges were used to fit these models. All data used in this study were collected during some large rainfalls of the period from 2005 to 2007. The radar and all gauges were installed in the catchment of the Yishu River, a branch of the Huaihe River in China. Three models based on normal distribution and a dBZ presumption of gamma distribution were fitted using maximum-likelihood techniques, which were resolved by genetic algorithms. Comparisons of estimated maximized likelihoods based on assumptions of gamma and normal distribution showed that all generalized linear models (GLMs) of presumed gamma distribution were better fitted than GLMs based on normal distribution. In a comparison of maximum-likelihood, the differences between these three models were small. Three error statistics were used to assess the agreement between radar estimated rainfall and gauge rainfall: relative bias (B), root mean square error (RMSE), and correlation coefficient (r). The results showed that no one model was excellent in all criteria. On the whole, the GLM-based models gave smaller relative bias than the traditional power-law model. It is suggested that validations conducted in many previous works should have been made against a specific criterion but overlooked others.
LIU Yong-HeZHANG Wan-ChangSHAO Yue-HongZHANG Jing-Ying
