Impact of 3D Var GSI-ENKF hybrid data assimilation system

The hybrid two-way coupled 3DEnsVar assimilation system was tested with the NCMRWF global data assimilation forecasting system. At present, this system consists of T574L64 deterministic model and the grid-point statistical interpolation analysis scheme. In this experiment, the analysis system is modified with a two-way coupling with an 80 member Ensemble Kalman Filter of T254L64 resolution and runs are carried out in parallel to the operational system for the Indian summer monsoon season (June–September) for the year 2015 to study its impact. Both the assimilation systems are based on NCEP GFS system. It is found that hybrid assimilation marginally improved the quality of the forecasts of all variables over the deterministic 3D Var system, in terms of statistical skill scores and also in terms of circulation features. The impact of the hybrid system in prediction of extreme rainfall and cyclone track is discussed.     

Prediction of fog/visibility over India using NWP Model

Frequent occurrence of fog in different parts of northern India is common during the winter months of December and January. Low visibility conditions due to fog disrupt normal public life. Visibility conditions heavily affect both surface and air transport. A number of flights are either diverted or cancelled every year during the winter season due to low visibility conditions, experienced at different airports of north India. Thus, fog and visibility forecasts over plains of north India become very important during winter months. This study aims to understand the ability of a NWP model (NCMRWF, Unified Model, NCUM) with a diagnostic visibility scheme to forecast visibility over plains of north India. The present study verifies visibility forecasts obtained from NCUM against the INSAT-3D fog images and visibility observations from the METAR reports of different stations in the plains of north India. The study shows that the visibility forecast obtained from NCUM can provide reasonably good indication of the spatial extent of fog in advance of one day. The fog intensity is also predicted fairly well. The study also verifies the simple diagnostic model for fog which is driven by NWP model forecast of surface relative humidity and wind speed. The performance of NWP model forecast of visibility is found comparable to that from simple fog model driven by NWP forecast of relative humidity and wind speed.     

Improvements in medium range weather forecasting system of India

Medium range weather forecasts are being generated in real time using Global Data Assimilation Forecasting System (GDAFS) at NCMRWF since 1994. The system has been continuously upgraded in terms of data usage, assimilation and forecasting system. Recently this system was upgraded to a horizontal resolution of T574 (about 22 km) with 64 levels in vertical. The assimilation scheme of this upgraded system is based on the latest Grid Statistical Interpolation (GSI) scheme and it has the provision to use most of available meteorological and oceanographic satellite datasets besides conventional meteorological observations. The new system has an improved procedure for relocating tropical cyclone to its observed position with the correct intensity. All these modifications have resulted in improvement of skill of medium range forecasts by about 1 day.     

Forecasting of cyclone Viyaru and Phailin by NWP-based cyclone prediction system (CPS) of IMD – an evaluation

An objective NWP-based cyclone prediction system (CPS) was implemented for the operational cyclone forecasting work over the Indian seas. The method comprises of five forecast components, namely (a) Cyclone Genesis Potential Parameter (GPP), (b) Multi-Model Ensemble (MME) technique for cyclone track prediction, (c) cyclone intensity prediction, (d) rapid intensification, and (e) predicting decaying intensity after the landfall. GPP is derived based on dynamical and thermodynamical parameters from the model output of IMD operational Global Forecast System. The MME technique for the cyclone track prediction is based on multiple linear regression technique. The predictor selected for the MME are forecast latitude and longitude positions of cyclone at 12-hr intervals up to 120 hours forecasts from five NWP models namely, IMD-GFS, IMD-WRF, NCEP-GFS, UKMO, and JMA. A statistical cyclone intensity prediction (SCIP) model for predicting 12 hourly cyclone intensity (up to 72 hours) is developed applying multiple linear regression technique. Various dynamical and thermodynamical parameters as predictors are derived from the model outputs of IMD operational Global Forecast System and these parameters are also used for the prediction of rapid intensification. For forecast of inland wind after the landfall of a cyclone, an empirical technique is developed. This paper briefly describes the forecast system CPS and evaluates the performance skill for two recent cyclones Viyaru (non-intensifying) and Phailin (rapid intensifying), converse in nature in terms of track and intensity formed over Bay of Bengal in 2013. The evaluation of performance shows that the GPP analysis at early stages of development of a low pressure system indicated the potential of the system for further intensification. The 12-hourly track forecast by MME, intensity forecast by SCIP model and rapid intensification forecasts are found to be consistent and very useful to the operational forecasters. The error statistics of the decay model shows that the model was able to predict the decaying intensity after landfall with reasonable accuracy. The performance statistics demonstrates the potential of the system for improving operational cyclone forecast service over the Indian seas.     

Impact of assimilation of ATOVS temperature and humidity and SSM/I total precipitable water on the simulation of a monsoon depression

The present study explored the effect of assimilation of Advanced TIROS Vertical Sounder (ATOVS) temperature and humidity profiles and Spectral sensor microwave imager (SSM/I) total precipitable water (TPW) on the simulation of a monsoon depression which formed over the Arabian Sea during September 2005 using the Weather Research and Forecast model. The three-dimensional variational (3DVAR) data assimilation technique has been employed for the purpose of assimilation of satellite observations. Statistical scores like “equitable threat score,” “bias score,” “forecast impact,” and “improvement parameter” have been used to examine the impact of the above-mentioned satellite observations on the numerical simulation of a monsoon depression. The diagnostics of this study include verification of the vertical structure of depression, in terms of temperature anomaly profiles and relative vorticity profiles with observations/analysis. Additional diagnostics of the study include the analysis of the heat budget and moisture budget. Such budget studies have been performed to provide information on the role of cumulus convection associated with the depression. The results of this study show direct and good evidence of the impact of the assimilation of the satellite observations using 3DVAR on the dynamical and thermodynamical features of a monsoon depression along with the effect of inclusion of satellite observation on the spatial pattern of the simulated precipitation associated with the depression. The “forecast impact” parameter calculated for the wind speed provides good evidence of the positive impact of the assimilation of ATOVS temperature and humidity profiles and SSM/I TPW on the model simulation, with the assimilation of the ATOVS profiles showing better impact in terms of a more positive value of the “forecast impact” parameter. The results of the study also indicate the improvement of the forecast skill in terms of “equitable threat score” and “bias score” due to the assimilation of satellite observation.     

Impact of additional surface observation network on short range weather forecast during summer monsoon 2008 over Indian subcontinent

The three dimensional variational data assimilation scheme (3D-Var) is employed in the recently developed Weather Research and Forecasting (WRF) model. Assimilation experiments have been conducted to assess the impact of Indian Space Research Organisation’s (ISRO) Automatic Weather Stations (AWS) surface observations (temperature and moisture) on the short range forecast over the Indian region. In this study, two experiments, CNT (without AWS observations) and EXP (with AWS observations) were made for 24-h forecast starting daily at 0000 UTC during July 2008. The impact of assimilation of AWS surface observations were assessed in comparison to the CNT experiment. The spatial distribution of the improvement parameter for temperature, relative humidity and wind speed from one month assimilation experiments demonstrated that for 24-h forecast, AWS observations provide valuable information. Assimilation of AWS observed temperature and relative humidity improved the analysis as well as 24-h forecast. The rainfall prediction has been improved due to the assimilation of AWS data, with the largest improvement seen over the Western Ghat and eastern India.     

利用WRF3D-Var同化多普勒雷达反演风场试验研究

为了将C波段雷达风场资料更好地应用于数值预报模式中,利用两步变分法反演多普勒雷达风场资料,并处理成标准的常规探空资料,以WRF模式及其三维变分同化系统为平台,针对2013年6月19日发生在天水的一次强暴雨过程进行同化雷达反演风的试验研究.试验结果表明:同化雷达反演风场后,对降水预报的改进能维持12h,尤其同化雷达反演风场后3~9h效果非常显著;0~3h作用不是很明显;9~12h预报具有一定的正作用.另外,循环同化比同化一次效果好,但并不是同化次数越多越好.因此,同化C波段雷达反演风场后,对降水预报具有一定的正作用.     

考虑降水预报的三峡入库洪水集合概率预报方法比较

为了考虑预见期内降水预报的不确定性对洪水预报的影响,采用中国气象局、美国环境预测中心和欧洲中期天气预报中心的TIGGE(THORPEX Interactive Grand Global Ensemble)降水预报数据驱动GR4J水文模型,开展三峡入库洪水集合概率预报,分析比较BMA、Copula-BMA、EMOS、M-BMA 4种统计后处理方法的有效性。结果表明:4种统计后处理方法均能提供一个合理可靠的预报置信区间;其期望值预报精度相较于确定性预报有所提高,尤其是水量误差显著减小;M-BMA方法概率预报效果最佳,它能够考虑预报分布的异方差性,不需要进行正态变换,结构简单,应用灵活。     

Skills of different mesoscale models over Indian region during monsoon season: Forecast errors

Performance of four mesoscale models namely, the MM5, ETA, RSM and WRF, run at NCMRWF for short range weather forecasting has been examined during monsoon-2006. Evaluation is carried out based upon comparisons between observations and day-1 and day-3 forecasts of wind, temperature, specific humidity, geopotential height, rainfall, systematic errors, root mean square errors and specific events like the monsoon depressions.It is very difficult to address the question of which model performs best over the Indian region? An honest answer is ‘none’. Perhaps an ensemble approach would be the best. However, if we must make a final verdict, it can be stated that in general, (i) the WRF is able to produce best All India rainfall prediction compared to observations in the day-1 forecast and, the MM5 is able to produce best All India rainfall forecasts in day-3, but ETA and RSM are able to depict the best distribution of rainfall maxima along the west coast of India, (ii) the MM5 is able to produce least RMSE of wind and geopotential fields at most of the time, and (iii) the RSM is able to produce least errors in the day-1 forecasts of the tracks, while the ETA model produces least errors in the day-3 forecasts.     

Experimental real-time multi-model ensemble (MME) prediction of rainfall during monsoon 2008: Large-scale medium-range aspects

Realistic simulation/prediction of the Asian summer monsoon rainfall on various space–time scales is a challenging scientific task. Compared to mid-latitudes, a proportional skill improvement in the prediction of monsoon rainfall in the medium range has not happened in recent years. Global models and data assimilation techniques are being improved for monsoon/tropics. However, multi-model ensemble (MME) forecasting is gaining popularity, as it has the potential to provide more information for practical forecasting in terms of making a consensus forecast and handling model uncertainties. As major centers are exchanging model output in near real-time, MME is a viable inexpensive way of enhancing the forecasting skill and information content. During monsoon 2008, on an experimental basis, an MME forecasting of large-scale monsoon precipitation in the medium range was carried out in real-time at National Centre for Medium Range Weather Forecasting (NCMRWF), India. Simple ensemble mean (EMN) giving equal weight to member models, bias-corrected ensemble mean (BCEMn) and MME forecast, where different weights are given to member models, are the products of the algorithm tested here. In general, the aforementioned products from the multi-model ensemble forecast system have a higher skill than individual model forecasts. The skill score for the Indian domain and other sub-regions indicates that the BCEMn produces the best result, compared to EMN and MME. Giving weights to different models to obtain an MME product helps to improve individual member models only marginally. It is noted that for higher rainfall values, the skill of the global model rainfall forecast decreases rapidly beyond day-3, and hence for day-4 and day-5, the MME products could not bring much improvement over member models. However, up to day-3, the MME products were always better than individual member models.     

The Indian summer monsoon as revealed by the NCMRWF system

In this study, we present the mean seasonal features of the Indian summer monsoon circulation in the National Centre for Medium Range Weather Forecasting (NCMRWF) global data assimilation and forecast system. The large-scale budgets of heat and moisture are examined in the analyzed and model atmosphere. The daily operational analyses and forecasts (day 1 through day 5) produced for the summer seasons comprising June, July and August of 1995 and 1993 have been considered for the purpose. The principal aim of the study is two-fold. Primarily, to comprehend the influence of the systematic errors over the Indian summer monsoon, secondarily, to analyze the performance of the model in capturing the interseasonal variability. The heat and moisture balances show reduction in the influx of heat and moisture in the model forecasts compared to the analyzed atmosphere over the monsoon domain. Consequently, the diabatic heating also indicates reducing trend with increase in the forecast period. In effect, the strength of Indian summer monsoon, which essentially depends on these parameters, weakens considerably in the model forecasts. Despite producing feeble monsoon circulation, the model captures interseasonal variability realistically. Although, 1995 and 1993 are fairly normal monsoon seasons, the former received more rainfall compared to the latter in certain pockets of the monsoon domain. This is clearly indicated by the analyzed and model atmosphere in terms of energetics.     

Indian Ocean surface winds from NCMRWF analysis as compared to QuikSCAT and moored buoy winds

The quality of the surface wind analysis at the National Centre for Medium Range Weather Forecasts (NCMRWF), New Delhi over the tropical Indian Ocean and its improvement in 2001 are examined by comparing it within situ buoy measurements and satellite derived surface winds from NASA QuikSCAT satellite (QSCT) during 1999, 2000 and 2001. The NCMRWF surface winds suffered from easterly bias of 1.0–1.5 ms^-1 in the equatorial Indian Ocean (IO) and northerly bias of 2.0–3.0 ms^-1 in the south equatorial IO during 1999 and 2000 compared to QSCT winds. The amplitude of daily variability was also underestimated compared to that in QSCT. In particular, the amplitude of daily variability of NCMRWF winds in the eastern equatorial IO was only about 60% of that of QSCT during 1999 and 2000. The NCMRWF surface winds during 2001 have significantly improved with the bias of the mean analyzed winds considerably reduced everywhere bringing it to within 0.5 ms^-1 of QSCT winds in the equatorial IO. The amplitude and phase of daily and intraseasonal variability are very close to that in QSCT almost everywhere during 2001. It is shown that the weakness in the surface wind analysis during 1999 and 2000 and its improvement in 2001 are related to the weakness in simulation of precipitation by the forecast model in the equatorial IO and its improvement in 2001.     

High resolution numerical weather prediction over the Indian subcontinent

In this study, the Florida State University Global Spectral Model (FSUGSM), in association with a high-resolution nested regional spectral model (FSUNRSM), is used for short-range weather forecasts over the Indian domain. Three-day forecasts for each day of August 1998 were performed using different versions of the FSUGSM and FSUNRSM and were compared with the observed fields (analysis) obtained from the European Center for Medium Range Weather Forecasts (ECMWF). The impact of physical initialization (a procedure that assimilates observed rain rates into the model atmosphere through a set of reverse algorithms) on rainfall forecasts was examined in detail. A very high nowcasting skill for precipitation is obtained through the use of high-resolution physical initialization applied at the regional model level. Higher skills in wind and precipitation forecasts over the Indian summer monsoon region are achieved using this version of the regional model with physical initialization. A relatively new concept, called the ‘multimodel/multianalysis superensemble’ is described in this paper and is applied for the wind and precipitation forecasts over the Indian subcontinent. Large improvement in forecast skills of wind at 850 hPa level over the Indian subcontinent is shown possible through the use of the multimodel superensemble. The multianalysis superensemble approach that uses the latest satellite data from the Tropical Rainfall Measuring Mission (TRMM) and the Defense Meteorological Satellite Program (DMSP) has shown significant improvement in the skills of precipitation forecasts over the Indian monsoon region.     

Assimilation of IRS-P4 (MSMR) meteorological data in the NCMRWF global data assimilation system

Oceansat-1 was successfully launched by India in 1999, with two payloads, namely Multi-frequency Scanning Microwave Radiometer (MSMR) and Ocean Color Monitor (OCM) to study the biological and physical parameters of the ocean. The MSMR sensor is configured as an eight-channel radiometer using four frequencies with dual polarization. The MSMR data at 75 km resolution from the Oceansat-I have been assimilated in the National Centre for Medium Range Weather Forecasting (NCMRWF) data assimilation forecast system. The operational analysis and forecast system at NCMRWF is based on a T80L18 global spectral model and Spectral Statistical Interpolation (SSI) scheme for data analysis. The impact of the MSMR data is seen globally, however it is significant over the oceanic region where conventional data are rare. The dry-nature of the control analyses have been removed by utilizing the MSMR data. Therefore, the total precipitable water data from MSMR has been identified as a very crucial parameter in this study. The impact of surface wind speed from MSMR is to increase easterlies over the tropical Indian Ocean. Shifting of the positions of westerly troughs and ridges in the south Indian Ocean has contributed to reduction of temperature to around 30‡S.     

Simulation of very severe cyclone Mala over Bay of Bengal with HWRF modeling system

Tropical cyclone is one of the most devastating weather phenomena all over the world. The Environmental Modeling Center (EMC) of the National Center for Environmental Prediction (NCEP) has developed a sophisticated mesoscale model known as Hurricane Weather Research and Forecasting (HWRF) system for tropical cyclone studies. The state-of-the-art HWRF model (atmospheric component) has been used in simulating most of the features our present study of a very severe tropical cyclone ??Mala??, which developed on April 26 over the Bay of Bengal and crossed the Arakan coast of Myanmar on April 29, 2006. The initial and lateral boundary conditions are obtained from Global Forecast System (GFS) analysis and forecast fields of the NCEP, respectively. The performance of the model is evaluated with simulation of cyclone Mala with six different initial conditions at an interval of 12?h each from 00 UTC 25 April 2006 to 12 UTC 27 April 2006. The best result in terms of track and intensity forecast as obtained from different initial conditions is further investigated for large-scale fields and structure of the cyclone. For this purpose, a number of important predicted fields?? viz. central pressure/pressure drop, winds, precipitation, etc. are verified against observations/verification analysis. Also, some of the simulated diagnostic fields such as relative vorticity, pressure vertical velocity, heat fluxes, precipitation rate, and moisture convergences are investigated for understanding of the characteristics of the cyclone in more detail. The vector displacement errors in track forecasts are calculated with the estimated best track provided by the India Meteorological Department (IMD). The results indicate that the model is able to capture most of the features of cyclone Mala with reasonable accuracy.     

Impact of variational assimilation technique on simulation of a heavy rainfall event over Pune,India

Prediction of heavy rainfall events due to severe convective storms in terms of their spatial and temporal scales is a challenging task for an operational forecaster. The present study is about a record-breaking heavy rainfall event observed in Pune (18°31′N, 73°55′E) on October 4, 2010. The day witnessed highest 24-h accumulated precipitation of 181.3 mm and caused flash floods in the city. The WRF model-based real-time weather system, operating daily at Centre for Development of Advanced Computing using PARAM Yuva supercomputer showed the signature of this convective event 4-h before, but failed to capture the actual peak rainfall and its location with reference to the city’s observational network. To investigate further, five numerical experiments were conducted to check the impact of assimilation of observations in the WRF model forecast. First, a control experiment was conducted with initialization using National Centre for Environmental Prediction (NCEP)’s Global Forecast System 0.5° data, while surface observational data from NCEP Prepbufr system were assimilated in the second experiment (VARSFC). In the third experiment (VARAMV), NCEP Prepbufr atmospheric motion vectors were assimilated. Fourth experiment (VARPRO) was assimilated with conventional soundings data, and all the available NCEP Prepbufr observations were assimilated in the fifth experiment (VARALL). Model runs were compared with observations from automated weather stations (AWS), synoptic charts of Indian Meteorological Department (IMD). Comparison of 24-h accumulated rainfall with IMD AWS 24-h gridded data showed that the fifth experiment (VARALL) produced better picture of heavy rainfall, maximum up to 251 mm/day toward the southern side, 31 km away from Pune’s IMD observatory. It was noticed that the effect of soundings observations experiment (VARPRO) caused heavy precipitation of 210 mm toward the southern side 49 km away from Pune. The wind analysis at 850 and 200 hPa indicated that the surface and atmospheric motion vector observations (VARAMV) helped in shifting its peak rainfall toward Pune, IMD observatory by 18 km, though VARALL over-predicted rainfall by 60 mm than the observed.     

Assimilation of Indian Doppler Weather Radar observations for simulation of mesoscale features of a land-falling cyclone

In this paper, impact of Indian Doppler Weather Radar (DWR) data, i.e., reflectivity (Z), radial velocity (Vr) data individually and in combination has been examined for simulation of mesoscale features of a land-falling cyclone with Advance Regional Prediction System (ARPS) Model at 9-km horizontal resolution. The radial velocity and reflectivity observations from DWR station, Chennai (lat. 13.0°N and long. 80.0°E), are assimilated using the ARPS Data Assimilation System (ADAS) and cloud analysis scheme of the model. The case selected for this study is the Bay of Bengal tropical cyclone NISHA of 27–28 November 2008. The study shows that the ARPS model with the assimilation of radial wind and reflectivity observations of DWR, Chennai, could simulate mesoscale characteristics, such as number of cells, spiral rain band structure, location of the center and strengthening of the lower tropospheric winds associated with the land-falling cyclone NISHA. The evolution of 850 hPa wind field super-imposed vorticity reveals that the forecast is improved in terms of the magnitude and direction of lower tropospheric wind, time, and location of cyclone in the experiment when both radial wind and reflectivity observations are used. With the assimilation of both radial wind and reflectivity observations, model could reproduce the rainfall pattern in a more realistic way. The results of this study are found to be very promising toward improving the short-range mesoscale forecasts.     

Impact of initial and boundary conditions on simulations of western disturbances and associated precipitation

Western Himalayas (WH) is characterized by variable topography and heterogeneous land use. During winter, it receives enormous amount of precipitation due to eastward moving extratropical cyclones, called western disturbances (WDs), in Indian parlance. This variable altitude and orientation of orographic barriers has a complex interplay with WDs in defining precipitation over the WH. To understand such complexities, three WDs are considered to study interaction with the Himalayan orography using the advanced regional prediction system. Two simulation strategies are performed and presented??first to illustrate the impact of different initial and boundary conditions and second to illustrate the impact of different horizontal model resolution with same model configuration. In the first strategy, three different initial and boundary conditions??the National Center for Environmental Prediction?CGlobal Forecast System, USA (NCEP?CGFS) (1) analysis (2) 0000UTC forecast and the National Center for Medium Range Weather Forecast, India?CT80 spectral model (NCMRWF?CT80) (3) 0000UTC forecast??are provided to the same model configuration. In the second strategy, outputs from model simulated with NCMRWF??T80 spectral model forecast at coarser horizontal model resolution of 30?km (hereafter called Experiment I) are used as input initial and boundary conditions for simulation at finer horizontal model resolution of 10?km (hereafter called Experiment II). Though there are many other dynamical factors, but in the present study, it is shown that model-simulated precipitation is sensitive to the initial and boundary conditions. Simulations at coarse resolution could capture the weather system, but detailed spatial distribution along the orography is better illustrated at finer resolution model simulation. Also, Experiment II could simulate precipitation over different ranges of the western Himalayas depicting orographic forcings.     

All-sky radiance simulation of Megha-Tropiques SAPHIR microwave sensor using multiple scattering radiative transfer model for data assimilation applications

Incorporation of cloud- and precipitation-affected radiances from microwave satellite sensors in data assimilation system has a great potential in improving the accuracy of numerical model forecasts over the regions of high impact weather. By employing the multiple scattering radiative transfer model RTTOV-SCATT, all-sky radiance (clear sky and cloudy sky) simulation has been performed for six channel microwave SAPHIR (Sounder for Atmospheric Profiling of Humidity in the Inter-tropics by Radiometry) sensors of Megha-Tropiques (MT) satellite. To investigate the importance of cloud-affected radiance data in severe weather conditions, all-sky radiance simulation is carried out for the severe cyclonic storm ‘Hudhud’ formed over Bay of Bengal. Hydrometeors from NCMRWF unified model (NCUM) forecasts are used as input to the RTTOV model to simulate cloud-affected SAPHIR radiances. Horizontal and vertical distribution of all-sky simulated radiances agrees reasonably well with the SAPHIR observed radiances over cloudy regions during different stages of cyclone development. Simulated brightness temperatures of six SAPHIR channels indicate that the three dimensional humidity structure of tropical cyclone is well represented in all-sky computations. Improved correlation and reduced bias and root mean square error against SAPHIR observations are apparent. Probability distribution functions reveal that all-sky simulations are able to produce the cloud-affected lower brightness temperatures associated with cloudy regions. The density scatter plots infer that all-sky radiances are more consistent with observed radiances. Correlation between different types of hydrometeors and simulated brightness temperatures at respective atmospheric levels highlights the significance of inclusion of scattering effects from different hydrometeors in simulating the cloud-affected radiances in all-sky simulations. The results are promising and suggest that the inclusion of multiple scattering radiative transfer models into data assimilation system can simulate the cloud-affected microwave radiance data which provide detailed information on three dimensional humidity structure of the atmosphere in the presence of cloud hydrometeors.     

订正与集成多模式的中国季度降水预测

针对两个最新换代的季度集合预测系统对中国季度降水预测中存在的系统缺陷,应用改进的贝叶斯联合概率模型（BJP）加以订正。对订正后的单一模式概率预测应用一种混合模型贝叶斯模型平均（BMA）方法加以集成,以综合各模式的优势来提高中国季度降水预测技巧。结果表明:BJP模型可有效地消除集合模式预测的系统偏差,同时大幅提高了概率预测的可靠性。经过订正的欧洲中尺度天气预报中心的 System4预测在许多季度在中国的很大区域范围内都显示出了一定的预测技巧;而澳洲气象局的POAMA2.4预测只在个别季度局部范围内具有技巧。使用BMA对订正后的单一模式预测进行集成可显著提高对中国季度降水预测的精度,相比单一模式预测,技巧得分为正值的网格百分率分别提高了13.3%和20.0%。