考虑有色噪声干扰的Hamm erste in非线性系统的辨识,通过梯度搜索原理推导了增广投影算法,简化增广投影算法和增广随机梯度辨识算法。基本思想是将增广信息向量中的未知噪声项用其估计残差代替。增广投影算法对噪声非常敏感,增广随机梯...考虑有色噪声干扰的Hamm erste in非线性系统的辨识,通过梯度搜索原理推导了增广投影算法,简化增广投影算法和增广随机梯度辨识算法。基本思想是将增广信息向量中的未知噪声项用其估计残差代替。增广投影算法对噪声非常敏感,增广随机梯度算法的收敛速度慢,为了解决这些不足,在增广随机梯度算法中引入遗忘因子,来改善参数估计精度,进一步通过仿真来比较算法的估计误差以及收敛速度。展开更多
The projected temperature and precipitation- change under different emissions scenarios using Coupled Model Intercomparison Project Phase 5 models over the northwestern arid regions of China （NWAC） were ana- lyzed u...The projected temperature and precipitation- change under different emissions scenarios using Coupled Model Intercomparison Project Phase 5 models over the northwestern arid regions of China （NWAC） were ana- lyzed using the ensemble of three high-resolution dy- namical downscaling simulations： the simulation of the Regional Climate Model version 4.0 （RegCM4） forced by the Beijing Climate Center Climate System Model version 1.1 （BCC_CSMI.1）; the Hadley Centre Global En- vironmental Model version 3 regional climate model （HadGEM3-RA） forced by the Atmosphere-Ocean cou- pled HadGEM version 2 （HadGEM2-AO）; and the Weather Research and Forecasting （WRF） model forced by the Norwegian community Earth System Model （NorESM1-M）. Model validation indicated that the mul- timodel simulations reproduce the spatial and temporal distribution of temperature and precipitation well. The temperature is projected to increase over NWAC under both the 4.5 and 8.5 Representative Concentration Path- ways scenarios （RCP4.5 and RCP8.5, respectively） in the middle of the 21 st century, but the warming trend is larger under the RCP8.5 scenario, Precipitation shows a signifi- cant increasing trend in spring and winter under both RCP4.5 and RCPS.5; but in summer, precipitation is pro- jected to decrease in the Tarim Basin and Junggar Basin. The regional averaged temperature and precipitation show increasing trends in the future over NWAC; meanwhile, the large variability of the winter mean temperature and precipitation may induce more extreme cold events and intense snowfall events in these regions in the future.展开更多
文摘考虑有色噪声干扰的Hamm erste in非线性系统的辨识,通过梯度搜索原理推导了增广投影算法,简化增广投影算法和增广随机梯度辨识算法。基本思想是将增广信息向量中的未知噪声项用其估计残差代替。增广投影算法对噪声非常敏感,增广随机梯度算法的收敛速度慢,为了解决这些不足,在增广随机梯度算法中引入遗忘因子,来改善参数估计精度,进一步通过仿真来比较算法的估计误差以及收敛速度。
基金supported by the National Basic Research Program of China (Grant No. 2012CB955401)the Special Fund for Public Welfare Industry (Grant No. GYHY201306026)the Key Laboratory of Oasis Ecology (KLOE) Open Fund (Grant No. XJDX02012012-04)
文摘The projected temperature and precipitation- change under different emissions scenarios using Coupled Model Intercomparison Project Phase 5 models over the northwestern arid regions of China （NWAC） were ana- lyzed using the ensemble of three high-resolution dy- namical downscaling simulations： the simulation of the Regional Climate Model version 4.0 （RegCM4） forced by the Beijing Climate Center Climate System Model version 1.1 （BCC_CSMI.1）; the Hadley Centre Global En- vironmental Model version 3 regional climate model （HadGEM3-RA） forced by the Atmosphere-Ocean cou- pled HadGEM version 2 （HadGEM2-AO）; and the Weather Research and Forecasting （WRF） model forced by the Norwegian community Earth System Model （NorESM1-M）. Model validation indicated that the mul- timodel simulations reproduce the spatial and temporal distribution of temperature and precipitation well. The temperature is projected to increase over NWAC under both the 4.5 and 8.5 Representative Concentration Path- ways scenarios （RCP4.5 and RCP8.5, respectively） in the middle of the 21 st century, but the warming trend is larger under the RCP8.5 scenario, Precipitation shows a signifi- cant increasing trend in spring and winter under both RCP4.5 and RCPS.5; but in summer, precipitation is pro- jected to decrease in the Tarim Basin and Junggar Basin. The regional averaged temperature and precipitation show increasing trends in the future over NWAC; meanwhile, the large variability of the winter mean temperature and precipitation may induce more extreme cold events and intense snowfall events in these regions in the future.
