Prediction of all India summer monsoon rainfall using error-back-propagation neural networks

Summary In this paper, multilayered feedforward neural networks trained with the error-back-propagation (EBP) algorithm have been employed for predicting the seasonal monsoon rainfall over India. Three network models that use, respectively, 2, 3 and 10 input parameters which are known to significantly influence the Indian summer monsoon rainfall (ISMR) have been constructed and optimized. The results obtained thereby are rigorously compared with those from the statistical models. The predictions of network models indicate that they can serve as a potent tool for ISMR prediction.     

Prediction of summer monsoon rainfall over India using the NCEP climate forecast system

The performance of a dynamical seasonal forecast system is evaluated for the prediction of summer monsoon rainfall over the Indian region during June to September (JJAS). The evaluation is based on the National Centre for Environmental Prediction’s (NCEP) climate forecast system (CFS) initialized during March, April and May and integrated for a period of 9 months with a 15 ensemble members for 25 years period from 1981 to 2005. The CFS’s hindcast climatology during JJAS of March (lag-3), April (lag-2) and May (lag-1) initial conditions show mostly an identical pattern of rainfall similar to that of verification climatology with the rainfall maxima (one over the west-coast of India and the other over the head Bay of Bengal region) well simulated. The pattern correlation between verification and forecast climatology over the global tropics and Indian monsoon region (IMR) bounded by 50°E–110°E and 10°S–35°N shows significant correlation coefficient (CCs). The skill of simulation of broad scale monsoon circulation index (Webster and Yang; WY index) is quite good in the CFS with highly significant CC between the observed and predicted by the CFS from the March, April and May forecasts. High skill in forecasting El Nino event is also noted for the CFS March, April and May initial conditions, whereas, the skill of the simulation of Indian Ocean Dipole is poor and is basically due to the poor skill of prediction of sea surface temperature (SST) anomalies over the eastern equatorial Indian Ocean. Over the IMR the skill of monsoon rainfall forecast during JJAS as measured by the spatial Anomaly CC between forecast rainfall anomaly and the observed rainfall anomaly during 1991, 1994, 1997 and 1998 is high (almost of the order of 0.6), whereas, during the year 1982, 1984, 1985, 1987 and 1989 the ACC is only around 0.3. By using lower and upper tropospheric forecast winds during JJAS over the regions of significant CCs as predictors for the All India Summer Monsoon Rainfall (AISMR; only the land stations of India during JJAS), the predicted mean AISMR with March, April and May initial conditions is found to be well correlated with actual AISMR and is found to provide skillful prediction. Thus, the calibrated CFS forecast could be used as a better tool for the real time prediction of AISMR.     

The role of different links of the river net in forming winter river runoff

Results of investigations of the winter river runoff formation in the Northern Dvina River basin (without the Vychegda River basin) are under consideration. The peculiarities of the winter runoff formation are revealed from the analysis of the conditions of the runoff reduction in different parts of the river basin. A regulatory effect of the upper links of the river net on the lower links is established.     

Chemical characteristics of haze during summer and winter in Guangzhou

Airborne particles were collected with a 10-stage MOUDI and a PM₁₀ sampler in Guangzhou, China, during both haze and normal days in the summer of 2002 and 2003, and winter 2002. The characteristics of PAHs, organic carbon, elemental carbon and water-soluble inorganic ions were studied under four periods (summer normal, summer haze, winter normal and winter haze). In this study, secondary pollutants (OC, SO₄²⁻, NO₃⁻ and NH₄⁺) were the major chemical components and appeared to show a remarkably rapid increase from normal to haze days. The particle mass size distributions were bimodal and dominated by fine particles in haze days. A significantly higher OC/EC ratio was found in haze days (3.2–4.7) compared to normal days (1.8–2.8), indicating secondary organic aerosol formation might be significant during haze days. Correlation analysis between visibility and chemical species showed that the major scattering species were TC (total carbon) and sulfate in normal days and nitrate and TC in haze days, respectively. Simultaneously, correlation analysis between visibility and meteorological factors demonstrated that visibility increased with both temperature and wind speed, while it decreased with relative humidity. Furthermore, the relatively higher value of IcdP/(BghiP + IcdP) and the low value of Cmax, CPI, and BghiP/BeP in winter haze could be due to the growth of motor vehicle usage and energy consumption in winter.     

新疆夏季降水和冬季气温预测方法及效果评估

本文利用4个国内外先进的气候模式（国家气候中心、ECMWF、NCEP和JMA）业务预测数据,采用2种多模式集合方法（等权平均和超级集合）、3种降尺度方法（BP-CCA、EOF迭代、高相关回归集成）和3种统计方法（CCA、最优气候值、高相关回归集成）以及降尺度集成和降尺度-统计方法集成,分析了目前季节模式、多模式集合、降尺度、统计方法、降尺度-统计集合等目前常用气候预测技术对新疆夏季降水和冬季气温的业务预测能力。 研究表明,以上技术方法对新疆夏季降水和冬季气温的预测预测能力有较大差别。目前先进的气候业务模式的预测技巧普遍很低,多模式超级集合和降尺度方法的技巧常高于单个模式,并且最佳的降尺度方法通常技巧高于最佳多模式集合方法。同时,统计方法和降尺度方法的预测技巧通常较为接近,而对二者进行超级集合可以具有相对很高的预测技巧。此外,现有常用气候预测技术方法对新疆夏季降水和冬季气温的趋势有一定的预测能力,但对气候异常的空间分布基本无预测能力。建议新疆气候预测技术围绕统计和降尺度方法集合发展。     

The Prediction of winter surface air temperature

The author investigates the prediction of Northeast China’s winter surface air temperature (SAT),and first forecast the year to year increment in the predic-tand and then predict the predictand.Thus,in the first step,we determined the predictors for an increment in winter SAT by analyzing the atmospheric variability associated with an increment in winter SAT.Then,multi-linear re-gression was applied to establish a prediction model for an increment in winter SAT in Northeast China.The pre-diction model shows a high correlation coefficient (0.73) between the simulated and observed annual increments in winter SAT in Northeast China throughout the period 1965-2002,with a relative root mean square error of -7.9%.The prediction model makes a reasonable hindcast for 2003-08,with an average relative root mean square error of -7.2%.The prediction model can capture the in-creasing trend of winter SAT in Northeast China from 1965-2008.The results suggest that this approach to forecasting an annual increment in winter SAT in North-east China would be relevant in operational seasonal forecasts.     

The impact of weather variability on British outbound flows

Climate can be understood both as a resource and a motivation for tourism. This study focuses on the second issue trying to establish the sensitivity to weather anomalies of the outbound flows from United Kingdom, the third biggest international tourist spender country. Using transfer function models it is possible to analyze the significance of the short-term weather conditions in the determination of outbound British flows and simulate the effects of different climate change scenarios. Results show how mean temperature, heat waves, air frost and sunshine days are the weather variables that can be significantly related to the dynamics of the outbound British flows time series.     

Variations of the summer Somali and Australia cross-equatorial flows and the implications for the Asian summer monsoon

The temporal variations during 1948-2010 and vertical structures of the summer Somali and Australia cross-equatorial flows(CEFs) and the implications for the Asian summer monsoon were explored in this study.The strongest southerly and northerly CEFs exist at 925 hPa and 150 hPa level,respectively.The low-level Somali(LLS) CEFs were significantly connected with the rainfall in most regions of India(especially the monsoon regions),except in a small area in southwest India.In comparison to the climatology,the lowlevel Australia(LLA) CEFs exhibited stronger variations at interannual time scale and are more closely connected to the East Asian summer monsoon circulation than to the LLS CEFs.The East Asian summer monsoon circulation anomalies related to stronger LLA CEFs were associated with less water vapor content and less rainfall in the region between the middle Yellow River and Yangtze River and with more water vapor and more rainfall in southern China.The sea-surface temperature anomalies east of Australia related to summer LLA CEFs emerge in spring and persist into summer,with implications for the seasonal prediction of summer rainfall in East Asia.The connection between the LLA CEFs and East Asian summer monsoon rainfall may be partly due to its linkage with El Nino-Southern Oscillation.In addition,both the LLA and LLS CEFs exhibited interdecadal shifts in the late 1970s and the late 1990s,consistent with the phase shifts of Pacific Decadal Oscillation(PDO).     

Climate change effects on winter river runoff in Transbaikalia

Considered are interannual variations of the river runoff in Transbaikalia during the freeze-up and their dependence on the climate change. The air temperature in winter has increased since the middle of the 20th century that results in the ice thickness decrease on the majority of rivers on the territory under study. The runoff volume in winter depends significantly on the preliminary moistening of catchments. Warming effect is manifested in the increase in the fraction of the winter runoff in the annual runoff volume during the long-term period.     

Characteristics of carbonaceous particles in Beijing during winter and summer 2003

Campaigns were conducted to measure Organic Carbon (OC) and Elemental Carbon (EC) in PM2.5 during winter and summer 2003 in Beijing. Modest differences of PM2.5 and PM10 mean concentrations were observed between the winter and summer campaigns. The mean PM2.5/PM10 ratio in both seasons was around 60%, indicating PM2.5 contributed significantly to PM10. The mean concentrations of OC and EC in PM2.5 were 11.2±7.5 and 6.0±5.0μg m-3 for the winter campaign, and 9.4±2.1 and 4.3±3.0 μg m-3 for the summer campaign, respectively. Diurnal concentrations of OC and EC in PM2.5 were found high at night and low during the daytime in winter, and characterized by an obvious minimum in the summer afternoon. The mean OC/EC ratio was 1.87±0.09 for winter and Z39±0.49 for summer. The higher OC/EC ratio in summer indicates some formation of Secondary Organic Carbon (SOC). The estimated SOC was 2.8 μg m-3 for winter and 4.2μg m-3 for summer.     

Potential evaporation in the peace river region of British Columbia

Abstract

Micrometeorological data collected over pasture in the Peace River area of British Columbia during the wet summer of 1977 were used to test the Priestley and Taylor (1972) model for potential evaporation. The model performed very well. RMSE was less than 10% of the mean evaporation rate on a daytime basis using an alpha value of 1.26. Since the model is mainly dependent on net radiation, which is rarely measured in such remote areas, this quantity was also estimated from more readily available meteorological data and used to calculate evaporation. Results were encouraging; calculated values were generally within 20 and 10% of energy balance estimates on daily and 5‐day mean bases.     

Spatiotemporal distributions of thermal peaks in summer and winter over South Korea

This study investigated the spatial distributions and long-term trends of the annual highest and lowest temperatures (summer peak temperature, SPT; winter peak temperature, WPT) and their timings (summer peak day, SPD; winter peak day, WPD) in South Korea and analyzed their relationship with the general circulation patterns. The two peaks were determined by selecting the highest and the lowest points after extracting temperature variations longer than the seasonal scale (91 days) in the time series of daily mean temperatures. For the long-term trend, we examined data for 100 years (1909–2008) for five stations and data for 35 years (1974–2008) for 61 stations. The SPD in South Korea is August 4 on average. It is earliest (July 31) in the central inland region, the central hilly sections, and the southern inland region and latest (August 10) in the southern coastal region and on Jeju Island (Seogwipo). The WPD in South Korea is January 16 on average. It is earliest (January 13) in the central inland region and southern inland region and latest (January 24) on Jeju Island (Jeju) and in Ulleungdo. The SPT and WPT are highest on Jeju Island (Seogwipo; 27.3 and 6.4 °C, respectively) and lowest in the central hilly sections (Daegwallyeong; 20.2 and ?7.9 °C, respectively). The interannual variations in the WPD and WPT are greater than those in SPD and SPT. A significant increasing trend in the WPT was observed for all of the analyzed stations only for the second half of the 100-year period (1959–2008). In the case of the 35-year period, the SPD did not show any clear changes at all stations, but the WPD tended to occur earlier at three stations in the east coastal area. The WPT showed an increasing trend at 55 stations for 35 years, but the SPT showed an increasing trend only in Seogwipo (0.041 °C/year) and even showed a decreasing trend in Mungyeong (?0.049 °C/year). General circulation patterns were indexed and their correlations with the seasonal peaks were investigated. No correlations were found with the SPD. However, the WPD showed a significant positive correlation with the day of the highest Siberian High Intensity (SHI) and the day of the lowest Arctic Oscillation Index (AOI). Furthermore, the SPT showed negative correlations with the intensities of the Okhotsk High and North Pacific High, whereas the WPT showed a negative correlation with SHI and a positive correlation with AOI and with the intensity of the northerly wind that flows into the Korean Peninsula.     

Causes of mid-Pliocene strengthened summer and weakened winter monsoons over East Asia

The mid-Pliocene warm period was the most recent geological period in Earth's history that featured long-term warming.Both geological evidence and model results indicate that East Asian summer winds(EASWs) strengthened in monsoonal China, and that East Asian winter winds(EAWWs) weakened in northern monsoonal China during this period, as compared to the pre-industrial period. However, the corresponding mechanisms are still unclear. In this paper, the results of a set of numerical simulations are reported to analyze the effects of changed boundary conditions on the mid-Pliocene East Asian monsoon climate, based on PRISM3(Pliocene Research Interpretation and Synoptic Mapping) palaeoenvironmental reconstruction. The model results showed that the combined changes of sea surface temperatures, atmospheric CO2 concentration,and ice sheet extent were necessary to generate an overall warm climate on a large scale, and that these factors exerted the greatest effects on the strengthening of EASWs in monsoonal China. The orographic change produced significant local warming and had the greatest effect on the weakening of EAWWs in northern monsoonal China in the mid-Pliocene. Thus,these two factors both had important but different effects on the monsoon change. In comparison, the effects of vegetational change on the strengthened EASWs and weakened EAWWs were relatively weak. The changed monsoon winds can be explained by a reorganization of the meridional temperature gradient and zonal thermal contrast. Moreover, the effect of orbital parameters cannot be ignored. Results showed that changes in orbital parameters could have markedly affected the EASWs and EAWWs, and caused significant short-term oscillations in the mid-Pliocene monsoon climate in East Asia.     

辽宁地区夏季高温极值预测模型

利用1957-2006年辽宁地区夏季23站极端最高气温资料和国家气候中心气候监测室的74项环流特征量资料,应用EOF方法对高温极值样本进行分解,研究辽宁极端高温的时空分布规律。结果表明：第一特征向量表现为区域整体一致的特征,中心区位于辽西北、辽北,第二、三特征向量空间分布表现为东西部反位相和南北反位相的特征。普查了前3个时间系数与前期环流指数的相关关系,认为前3个时间系数的显著影响因子是不同的。采用CSC准则确定最优预测因子,分别建立各时间系数的回归统计模型,并对高温极值历史拟合序列进行回报检验和预测检验。回报结果表明,各站的历史拟合率都保持在一定水平,但拟合率在辽西地区较差。各年的历史拟合率极不均衡,多数年份较为稳定,但个别年份拟合率较低。未来3 a试验性预测效果逐年下降,模型对未来1 a预测能力较好,可以作为业务预测的参考。     

机器学习方法在湖南夏季降水预测中的应用

利用湖南97个国家站的逐月降水资料、国家气候中心130项气候指数集以及国家气候中心和美国国家环境预报中心两套季节预测模式的降水预测资料,采用递归特征消除法确定预测因子并使用多层前馈神经网络、支持向量回归和自然梯度提升三种算法建立了两种湖南夏季降水统计预测方案的模型,检验了预测效果.结果表明:基于机器学习的预测模型对湖南...