AN APPLICATION OF THE VIC-3L LAND SURFACE MODEL IN SIMULATING STREAMFLOW FOR CONTINENTAL-SCALE RIVER BASINS IN CHINA

Transcription

1 AN APPLICATION OF THE VIC-3L LAND SURFACE MODEL IN SIMULATING STREAMFLOW FOR CONTINENTAL-SCALE RIVER BASINS IN CHINA XIE Zhenghui 1, SU Fengge 1, LIANG Xu 2, LIU Qian 1 1 Institute of Atmospheric Physics, Chinese Academy of Sciences Beijing , China ( zxie@lasg.iap.ac.cn) 2 Department of Civil and Environmental Engineering University of California, Berkeley, CA ,USA. ABSTRACT The hydrologically based three-layer variable infiltration capacity (VIC-3L) land surface model coupled with a routing scheme is applied to simulate streamflow for the entire region of China and for a semi-arid river basin with a drainage area over 10,000 km 2. The routing scheme is represented by a linear reservoir method for overland flow and the Muskingum-Cunge method for channel flow. Soil parameters needed are derived from the soil classification information of global 5-min data provided by the NOAA hydrology office and re-gridded to km 2 resolution. The vegetation parameters are derived based on AVHRR (Advanced Very High Resolution Radiometer) and LDAS (Land Data Assimilation System) information. The forcing data are obtained through interpolation methods based on 740 meteorological stations. All of the data (i.e., soil, vegetation, and forcings) needed by VIC-3L are compiled at the km 2 resolution for the entire region of China, and the daily forcing data are available for the period of 1980 to The application of VIC-3L for a sub-basin of the Weihe River basin is conducted for the period of 1980 to The VIC-3L simulated daily runoff is routed to the outlet of Yangjiaping station and compared to the monthly observed streamflow at the station. Results show that the model can simulate the observations quite well. Keywords:VIC-3L, land surface model, routing scheme, streamflow, river basin INTRODUCTION Due to the impact of global climate change on temporal and spatial distributions of precipitation and temperature, the temporal and spatial distributions of runoff and evapotranspiration are changing over time. Therefore, it is very important to adequately simulate the runoff patterns over large areas in space and time to facilitate water resources planning and management, and regional sustainable development. Hydrological models have been developed and applied to study the effects of future climate change on water resources by a number of researchers (Kite et al, 1999; Bergstorm et al, 1998; and Arnell, 1999). In China, specific models are developed for specific river basins to study the hydrological effects. For example, models are developed for applications to the Yellow River (Wang et al, 2000), Haihe River and Yangtse River (Xiong et al, 1996), and Huaihe River (Hao et al, 2000). These models are conceptual hydrological models and are developed based on water balance framework. These hydrological models have three important limitations: (1) the effects of vegetation are not explicitly considered; (2) the energy budget is not considered, and (3) it is difficult to couple these models to climate models due to the specific structure of these models. Liang et al (1994) developed a two-layer Variable Infiltration Capacity (VIC-2L) model which considers both water and energy budgets. Also, the VIC-2L model considers explicitly the effects of multiple vegetation covers on water and energy budgets. Different from many other land 1

2 surface schemes, the VIC-2L model incorporates the representation of subgrid spatial variability of This paper presents an application of the precipitation with the representation of spatial VIC-3L model coupled with a routing scheme to variability of infiltration to simulate energy and simulate runoff for the entire region of China at a water budgets (Liang et al., 1994; 1996a). The spatial resolution of km 2. Also, an feature of considering subgrid variability within the application of the VIC-3L to a tributary of the Weihe model framework is important because it facilitates river basin to simulate the streamflow is compared applications of VIC-2L to large spatial domains. with monthly observations with encouraging results. VIC-2L includes two different time scales (i.e., fast and slow) for runoff to capture the dynamics of 1 DATA AND MODEL PARAMETERS runoff generation. The upper soil layer of the model Vegetation, soil, and forcing data needed to is designed to represent the dynamic response of the apply the VIC-3L model are prepared at soil to rainfall events, and the lower layer is used to km 2 resolution for the entire region of China. In this characterize the seasonal soil moisture behavior. The section, preparation of each of the data sets is briefly subsurface flow from the low layer is represented by described. the ARNO parameterization (Franchini et al, 1991). To better represent quick bare soil evaporation following small summer rainfall events, a thin soil layer is included in VIC-2L (Liang et al., 1996b), and VIC-2L becomes VIC-3L. Also, soil moisture diffusion processes between the three soil layers are considered in VIC-3L. Liang and Xie (2001) developed a new parameterization to represent the infiltration excess runoff mechanism in VIC-3L and combined it effectively with the original representation of saturation excess runoff mechanism. Cherkauer and Lettenmaier (1999) improved the representation of processes for cold climate within VIC. VIC-2L/VIC-3L has been tested and applied to various basins of different scales in the United States (e.g., Arkansas-Red River Basin, Columbia River Basin, Delaware River Basin, and Mississippi River Basin) with good performance (e.g., Nijssen et al., 1997; Lohmann and Raschke, 1998; Liang and Xie, 2001; Parada et al., 2002, Xie and Liang et al, 2003). Also, VIC-2L/VIC-3L has performed well under humid excess runoff mechanism. 1.1 Vegetation data set Vegetation data set at km 2 resolution is derived based on AVHRR (Advanced Very High Resolution Radiometer) and LDAS (Land Data Assimilation System) information. AVHRR provides information on global land classification at 1 km resolution (Hansen et al., 2000). For each type of vegetation, the vegetation parameters such as architectural resistance, minimum stomata resistance, leaf-area index, albedo, roughness length, zero-plane displacement, and fraction of root depth of each soil layer are derived based on the vegetation parameter information from the Land Data Assimilation System (LDAS). The vegetation parameters in LDAS are estimated based on information from IGBP, BATS, NCAR LSM, SiB, SiB2 and Mosaic. Table 1 of Su and Xie (2003) lists the vegetation parameters used in VIC-3L for different vegetation class. 1.2 Soil data set and cold conditions in the various phases of the The classification of soil texture over the entire Project for Intercomparison of Land surface region of China is based on the information of Parameterization Schemes (PILPS) (e.g., Lohmann global 5-min soil data provided by the NOAA et al., 1998; Bowling et al., 2003; Nijssen et al., hydrology office and re-gridded to km ; Liang and Xie, 2003a, Liang and Xie et al, resolution. The individual soil parameters needed by 2003b). Liang and Xie (2001) and Parada et al. VIC-3L, such as, porosity θ s (m 3 m -3 ), saturated soil (2002) showed that the performance of VIC under potential ψ(m), s saturated hydraulic conductivity K s drier conditions could be improved with the (ms -1 ), and the exponent b parameter, are then inclusion of the new representation of infiltration derived based on the work of Cosby et al (1993) and 2

3 Rawls et al (1993). Table 2 of Su and Xie (2003) shows the soil classifications and the corresponding values of soil parameters used in the VIC-3L model. 1.3 Forcing data set The forcing data needed by VIC-3L at the km 2 resolution are obtained through interpolation methods based on 740 meteorological stations, which contain 11 years of daily precipitation and air temperature data from 1980 to Such station information is mapped to the resolution of km 2 grids through interpolation methods: (1) minimum distance method, i.e., the value observed at the nearest rain gauge station is taken as the mean value of a grid; and (2) linear interpolation weighted by distance between the rain gauge and the grid cell of interest. The above interpolation methods do not considerer the impact of topography. In the further, effects of the topography will be considered in the interpolation. applied to the entire region of China, the above parameters are assigned to be: 0.3 for B; 10, 0.02 and 0.8 for D m, D s and W s ; and 0.5 m and 2.0 m for the upper and lower layer soil depths, respectively, without conducting any model parameter calibrations, while the five model parameters are calibrated in the application to a sub-basin of the Weihe river basin. 2.1 Runoff simulations over entire region of China The VIC-3L model is applied to 4355 grid cells from 1980 to 1990 without calibration of the five VIC-3L model parameters. Daily runoff (mm) series of each grid cell are generated independently. Fig.1 shows the distribution of mean annual precipitation from the generated grid-based forcing data described in section 1. Fig.2 shows the distribution of simulated mean annual runoff. It can be seen that the spatial patterns of the simulated mean annual 2 NUMERICAL SIMULATIONS The vegetation, soil, and forcing data of each grid described in Section 1 are applied to the VIC-3L model to simulate evapotranspiration, runoff, and soil moisture over the entire region of China from 1980 to 1990 and over sub-basin of the Weihe River basin for the period of 1980 to The simulated runoff at each grid is then routed to the outlet of Yangjiaping station in the Weihe River basin application through a linear reservoir method for overland flow and the Muskingum-Cunge method for channel flow. The routed daily runoff at Yangjiaping station is aggregated to monthly runoff and compared to the monthly observed streamflow at the Yangjiaping station. In general, before conducting the numerical simulations, five model parameters of VIC-3L need to be calibrated since they cannot be determined well based on the soil information currently available. These five model parameters are the depth of the upper and lower soil layers (d i ), the exponent (B) of VIC-3L curve which describes the spatial variability of the soil moisture capacity, and the three parameters in the ARNO subsurface flow parameterization (i.e., D m, D s and W s ). In this study Fig.1 Distribution of mean annual precipitation from 1980 to 1990 (unit: mm). Fig.2 Distribution of simulated mean annual runoff from 1980 to 1990 (unit: mm). runoff and mean annual precipitation are consistent with each other. The correlation coefficient is The spatial pattern shows that water resources are 3

4 basically decreased from southeast to northwest in China. Because of the absence of observed data for evaporation, soil moisture and runoff for each grid cell, the model simulation results cannot be evaluated quantitatively. Streamflow ( m 3 /s ) is arguably the most easily measured and best documented component of the regional surface water budget component, and it offers an opportunity to evaluate the performance of model simulations. 2.2 A streamflow simulation Discharge/m 3 s Observed Simulated Month Fig. 3 Observed and simulated monthly streamflow at the Yangjiaping station from 1980 to The Weihe river basin within the Yellow River Basin is selected for initial validation of the VIC-3L model runoff simulations. Most of the large tributaries of the Weihe River are located on the north side of the main river. The three large tributaries are the Huluhe River, the Jinghe River, and the Beiluohe River. The drainage area of each river basin is more than 10,000 km 2. The Weihe river basin has a semi-arid climate. Its mean annual precipitation is 400 ~ 600 mm. In this application, the five VIC-3L model parameters are calibrated. The VIC-3L simulated runoff at each grid is routed to the outlet through a linear reservoir method for overland flow and the Muskingum-Cunge method for channel flow. Fig.3 shows a comparison of monthly streamflow over the period of 1980 to 1986 between model simulations and observations at the outlet of Yangjiaping station which is located at o ' o E, 35 20' N, and controls a drainage area of km 2. Fig.3 shows that there is a good agreement between the observed and simulated streamflow in the semi-arid region and the model can generally reproduce the monthly streamflow for the study watershed. 3 CONCLUSIONS In this study, a framework of applying the VIC-3L model to conduct hydrological simulations for large river basins in China is described. Vegetation, soil, and forcing data sets at the 50 km 50 km grid resolution over the entire region of China are produced. Daily forcing data from 740 stations for the period of 1980 to 1990 are mapped to each grid through interpolation methods. The VIC-3L model is applied to each of the 4355 grids for the period of 1980 to The results show that the spatial distribution of the mean annual simulated runoff has a good agreement with the mean annual precipitation. Comparison of monthly streamflow at Yangjiaping station in the Weihe river basin shows a good agreement between model simulations and observations. This modeling framework will be applied to investigate the impact of climate change on water resources in China. ACKNOWLEDGMENT: We wish to thank Prof. Chunzhen Liu for beneficial discussions and Mr. Silong Zhang, Ms. Yan Li for providing sufficient hydrology data. This work was supported by the National Key Planning Development Project for Basic Research (Grant Nos. 2001CB and G ), the National Natural Science Foundation of China (Grant Nos ), the Hundred Talents Program of the Chinese Academy of Sciences, and the Knowledge Innovation Key Project of Chinese Academy of Sciences(Grant No. KZCX2-SW- 317). 4

5 REFERENCES 1 Arnell,N.W., 1999: A simple water balance model for the simulation of streamflow over a large geophysical domain. J.Hydro.,217, Bergstrom, S.,., L. Phil Graham,1998: On the scale problem in hydrological modeling. J.Hydrol., 211, Bowling et al., 2003: Simulation of high latitude hydrological processes in the Torne-Kalix basin: PILPS Phase 2(e) 1: Experiment description and summary intercomparisons, Global and Planetary Change, 38, Bristow, K. L., G.S. Campbell, 1984: On the relationship between incoming solar radiation and daily maximum and minimum temperature. Agricultural and Forest Meteorology, 31, Liang, X., D.P. Lettenmaier, E.F. Wood, and S.J. Burges, A simple hydrologically based model of land surface water and energy fluxes for general circulation models, J. Geophys. Res., 99(D7), 14,415-14,428, Liang, X., D.P. Lettenmaier, and E.F. Wood, A one-dimensional statistical dynamic representation of subgrid spatial variability of precipitation in the two-layer variable infiltration capacity model, J. Geophys. Res., 101(D16), 21,403-21,422, 1996a. 13 Liang, X., E.F. Wood, and D.P. Lettenmaier, Surface soil moisture parameterization of the VIC-2L model: Evaluation and modifications, Global and Planetary Change, 13, , 1996b. 14 Liang, X., and Z. Xie, A new surface runoff parameterization with subgrid-scale soil heterogeneity for land surface models, Advances in Water Resources, 24, , Cherkauer, K. A. and D. P. Lettenmaier, 1999: 15 Liang X., Z. Xie, Important factors in Hydrologic effects of frozen soils in the upper land-atmosphere interactions: surface runoff Mississippi River basin. J. Geophys. Res., 104(D16): generactions and interactions between surface and groundwater, Global Planetary Change, 38, , 6 Cosby, B. J., G. M. Hornberger, R. B. Clapp, 2003a. and T.R. Ginn, 1984: A statistical exploration of the 16 Liang X., Z. Xie, and M. Huang, A new relationships of soil moisture characteristics to the parameterization for surface and groundwater physical properties of soils. Wate Resour. Res. 20, interactions and its impact on water budgets with the variable infiltration capacity(vic) land surface model, Journal of Geophysics Research,108(D16), 7 Franchini, M., Pacciani,M., 1991: Comparative GCP 8-1 -GCP 8-17,2003b. analysis of several conceptual rainfall runoff models. J. Hydrol.,122:, Lohmann, D. and E.Raschke, 1998: Regional scale hydrology: Ⅰ.Formulation of the VIC-2L 8 Hansen, M., DeFries, R., Townshend, J. R. G. model coupled to a routing model. Hydrol. Sci.J., and Sohlberg, R., 2000:Global land cover 43(1): classification at 1km resolution using a decision tree classifier. International Journal of Remote Sensing, 18 Lohmann, D., et al., The Project for 21: Intercomparison of Land-surface Parameterization Schemes (PILPS) phase 2(c) Red-Arkansas River 9 Hao Z., and F. Su, 2000: Application of the Basin experiment: 3. Spatial and temporal analysis monthly distributed hydrological model in the Huai of water fluxes, Global and Planetary Change, 19, River basin. Advances in Water Science, , (Supplement): Nijssen, B., D.P. Lettenmaier, X. Liang, S. 10 Kite, G. W. and U.haberlandt, 1999: Wetzel, E.F. Wood, 1997: Streamflow simulation for Atmospheric model data for macroscale hydrology. continental-scale river basins, Water Resour. Res., J.Hydrol.,217,

6 33, Nijssen et al., Simulation of high latitude hydrological processes in the Torne-Kalix basin: PILPS Phase 2(e) 2: Comparison of model results with observations, Global and Planetary Change, 38,31-53, Parada, L. M., J. P. Fram, and X. Liang, Multi-resolution calibration methodology for hydrologic models: Applications to a sub-humid catchment, Editors, Q. Duan, H. Gupta, S. Sorooshian, A. Rousseau, and R. Turcotte, Advances in Calibration of Watershed Models, Water Science and Application 6, American Geophy. Union, , Rawls, W. J.,L.R. Ahuja, D.L. Brakensiek, and A. Shirmohammadi, 1993:Infiltration and soil water movement. (ed. Maidment, D. R.). Handbook of Hydrology. New York: McGraw-Hill Inc., assessing effects of climate change on runoff of China, Progress in Natural Science,13(9), Wang Guoqing, Jian Li, and Yunzhang Wang, 2000: Grid monthly hydrological model in Yellow River and its application. Advances in Water Science, 11(Supplement, in Chinese): Xie Z., F. Su, X. Liang, et al, 2003, Applications of a surface runoff model with Horton and Dunne runoff for VIC, Advances in Atmospheric Sciences. 20(2), Xiong L., S. Guo, X. Fu, et al.,1996: Two-parameter monthly water balance model and its application. Advances in Water Science, 7:80(Supplement). 23 Su F., and Z. Xie, 2003: A model for 6

7 7