Smooth support vector machine （SSVM） changs the normal support vector machine （SVM） into the unconstrained op- timization by using the smooth sigmoid function. The method can be solved under the Broyden-Fletcher-G...Smooth support vector machine （SSVM） changs the normal support vector machine （SVM） into the unconstrained op- timization by using the smooth sigmoid function. The method can be solved under the Broyden-Fletcher-Goldfarb-Shanno （BFGS） algorithm and the Newdon-Armijio （NA） algorithm easily, however the accuracy of sigmoid function is not as good as that of polyno- mial smooth function. Furthermore, the method cannot reduce the influence of outliers or noise in dataset. A fuzzy smooth support vector machine （FSSVM） with fuzzy membership and polynomial smooth functions is introduced into the SVM. The fuzzy member- ship considers the contribution rate of each sample to the optimal separating hyperplane and makes the optimization problem more accurate at the inflection point. Those changes play a positive role on trials. The results of the experiments show that those FSSVMs can obtain a better accuracy and consume the shorter time than SSVM and lagrange support vector machine （LSVM）.展开更多
General circulation models （GCMs） are often used in assessing the impact of climate change at global and continental scales. However, the climatic factors simulated by GCMs are inconsistent at comparatively smaller ...General circulation models （GCMs） are often used in assessing the impact of climate change at global and continental scales. However, the climatic factors simulated by GCMs are inconsistent at comparatively smaller scales, such as individual river basins. In this study, a statistical downscaling approach based on the Smooth Support Vector Machine （SSVM） method was constructed to predict daily precipitation of the changed climate in the Hanjiang Basin. NCEP/NCAR reanalysis data were used to establish the statistical relationship between the larger scale climate predictors and observed precipitation. The relationship obtained was used to project future precipitation from two GCMs （CGCM2 and HadCM3） for the A2 emission scenario. The results obtained using SSVM were compared with those from an artificial neural network （ANN）. The comparisons showed that SSVM is suitable for conducting climate impact studies as a statistical downscaling tool in this region. The temporal trends projected by SSVM based on the A2 emission scenario for CGCM2 and HadCM3 were for rainfall to decrease during the period 2011–2040 in the upper basin and to increase after 2071 in the whole of Hanjiang Basin.展开更多
The climate impact studies in hydrology often rely on climate change information at fine spatial resolution. However, the general circulation model (GCM), which is widely used to simulate future climate scenario, oper...The climate impact studies in hydrology often rely on climate change information at fine spatial resolution. However, the general circulation model (GCM), which is widely used to simulate future climate scenario, operates on a coarse scale and does not provide reliable data on local or regional scale for hydrological modeling. Therefore the outputs from GCM have to be downscaled to obtain the information fit for hydrologic studies. The variable infiltration capacity (VIC) distributed hydrological model with 9×9 km2 grid resolution was applied and calibrated in the Hanjiang Basin. Validation results show that SSVM can approximate observed precipitation and temperature data reasonably well, and that the VIC model can simulate runoff hydrograph with high model efficiency and low relative error. By applying the SSVM model, the trends of precipitation and temperature (including daily mean temperature, daily maximum temperature and daily minimum temperature) projected from CGCM2 under A2 and B2 scenarios will decrease in the 2020s (2011―2040), and increase in the 2080s (2071―2100). However, in the 2050s (2041―2070), the precipitation will be decreased under A2 scenario and no significant changes under B2 scenario, but the temperature will be not obviously changed under both climate change scenarios. Under both climate change scenarios, the impact analysis of runoff, made with the downscaled precipitation and temperature time series as input of the VIC distributed model, has resulted in a decreasing trend for the 2020s and 2050s, and an overall increasing trend for the 2080s.展开更多
基金supported by the National Natural Science Foundation of China (60974082)
文摘Smooth support vector machine （SSVM） changs the normal support vector machine （SVM） into the unconstrained op- timization by using the smooth sigmoid function. The method can be solved under the Broyden-Fletcher-Goldfarb-Shanno （BFGS） algorithm and the Newdon-Armijio （NA） algorithm easily, however the accuracy of sigmoid function is not as good as that of polyno- mial smooth function. Furthermore, the method cannot reduce the influence of outliers or noise in dataset. A fuzzy smooth support vector machine （FSSVM） with fuzzy membership and polynomial smooth functions is introduced into the SVM. The fuzzy member- ship considers the contribution rate of each sample to the optimal separating hyperplane and makes the optimization problem more accurate at the inflection point. Those changes play a positive role on trials. The results of the experiments show that those FSSVMs can obtain a better accuracy and consume the shorter time than SSVM and lagrange support vector machine （LSVM）.
基金supported financially by the National Natural Science Foundation of China (Grant Nos.50679063 and 50809049)the International CooperationResearch Fund of China (2005DFA20520)the Re-search Fund for the Doctoral Program of Higher Education(200804861062)
文摘General circulation models （GCMs） are often used in assessing the impact of climate change at global and continental scales. However, the climatic factors simulated by GCMs are inconsistent at comparatively smaller scales, such as individual river basins. In this study, a statistical downscaling approach based on the Smooth Support Vector Machine （SSVM） method was constructed to predict daily precipitation of the changed climate in the Hanjiang Basin. NCEP/NCAR reanalysis data were used to establish the statistical relationship between the larger scale climate predictors and observed precipitation. The relationship obtained was used to project future precipitation from two GCMs （CGCM2 and HadCM3） for the A2 emission scenario. The results obtained using SSVM were compared with those from an artificial neural network （ANN）. The comparisons showed that SSVM is suitable for conducting climate impact studies as a statistical downscaling tool in this region. The temporal trends projected by SSVM based on the A2 emission scenario for CGCM2 and HadCM3 were for rainfall to decrease during the period 2011–2040 in the upper basin and to increase after 2071 in the whole of Hanjiang Basin.
基金Supported by the National Natural Science Foundation of China (Grant Nos. 50679063, 50809049)the International Cooperation Research Fund of China (Grant No. 2005DFA20520)the Research Fund for the Doctoral Program of Higher Education (Grant No. 200804861062)
文摘The climate impact studies in hydrology often rely on climate change information at fine spatial resolution. However, the general circulation model (GCM), which is widely used to simulate future climate scenario, operates on a coarse scale and does not provide reliable data on local or regional scale for hydrological modeling. Therefore the outputs from GCM have to be downscaled to obtain the information fit for hydrologic studies. The variable infiltration capacity (VIC) distributed hydrological model with 9×9 km2 grid resolution was applied and calibrated in the Hanjiang Basin. Validation results show that SSVM can approximate observed precipitation and temperature data reasonably well, and that the VIC model can simulate runoff hydrograph with high model efficiency and low relative error. By applying the SSVM model, the trends of precipitation and temperature (including daily mean temperature, daily maximum temperature and daily minimum temperature) projected from CGCM2 under A2 and B2 scenarios will decrease in the 2020s (2011―2040), and increase in the 2080s (2071―2100). However, in the 2050s (2041―2070), the precipitation will be decreased under A2 scenario and no significant changes under B2 scenario, but the temperature will be not obviously changed under both climate change scenarios. Under both climate change scenarios, the impact analysis of runoff, made with the downscaled precipitation and temperature time series as input of the VIC distributed model, has resulted in a decreasing trend for the 2020s and 2050s, and an overall increasing trend for the 2080s.
