Abstract:The research and application of crop growth simulation models in agrometeorolo gical services were carried out in recent years in China.The necessity of applyi ng crop growth simulation models to agrometeorological services was analyzed o n the basis of reviewing application of crop growth simulation models around the world and the state of agrometeorological services in China. The methodology of localization research of foreign models and the approach of applicationof model from single site to regional scale were discussed. It is focused on the progres s in application of crop growth simulation models to agrometeorological servic es in China, and effect assessment of meteorological condition on crops growth and dynamic forecasting of crop yield by using maize growth modal in NorthEast China, wheat growth model in North China and double cropping rice model in the middle Yangtze River valley. Finally, the existing problems and possible resolut ion in applying crop growth simulation models were proposed from mechanism impro vement, regional scale application, as well as operational implement.

Abstract:A new treatment of twotimelevel semiLagrangian scheme: Stable Extra polation TwoTimeLevel Scheme (SETTLS) has been successfully applied in the dynamic frame of T213L31 and the reduced Gaussian grid is transferred to linear Gau ssian grid, which made the number of spectral truncation increase from 213 to 63 9 and formed socalled T639L60 system. More than one year continuous forecast e xperiments have been done, and the results of objective verification on the perf ormance of 5day precipitation forecasting and 10day geopotential height fo recasting at 500hPa level are compared with the counterparts of operational system—T213L31.It shows that T639L60 system makes obvious improvement on 500hPa geo potential height forecasting over East Asia with improving one day valuable fore cast time. And it also improves the precipitation forecasting at light to heavy rain levels over all efficient times according to threat score and bias. Althoug h the threat score of T639 at storm rain level is higher than that of T213, the bias of T639 is worse than T213.

Abstract:During Beijing 2007 Olympic drilling the refined objective forecast techniques, i.e. theSupport Vector Machine (SVM) and the Half Periodic Function Fit (HPFF), gave a powerful technical support for the venues refined forecasts. Using the five weather elements including temperature, relative humidity, wind direction, wind speed, and 3hour rainfall observati ons at different Olympic Game venues, the 3hour products of SVM and H PFF refined objective forecast and forecaster's forecast for next 3 d ays at five different Olympic Game venues are verified. The results show that: (1) the continuous variables (e.g. temperature, relative humidity , wind speed) HPFF forecast skill is higher than SVM, and for the noncontinuous variables(e.g. wind direction) less. (2) The objective forecast techniques are t he basis of the forecasters producing refined forecast products. For temperature, relative humidity, wind speed, forecaster forecast skill is a little higher than the objective techniques; and for rainfall and wind direction, forecaster and t he objective techniques have not much difference on forecast skill. Although the forecasters have the limited ability of modification for the objective techniqu es, but have the comprehensive differentiation ability of the best choice for two objective products. (3) The mean absolute error from 0h to 63h valid time of the temperature that forecaster produced is about 1.8℃, the forecast accuracy of tem perature when considering the absolute difference between forecast and observati on is less and equal 1℃ is around 43%, and forecast error has no evident change with increasing valid time. the mean absolute error of relative humidity (0～24 h) is about 10%, the forecast accuracy of relative humidity when considering the absolute difference between forecast and observation is less and equal 10% is a round 60%, and forecast error has evident change with increasing valid time. The mean absolute error of wind direction (0～63h) is about 20%, the forecast accur acy has no obvious change with increasing valid time; the mean absolute error of wind speed is between 0.8m·s-1 and 1.4m·s-1.

Abstract:Herein Doppler radar reflectivity and radial velocity data assimilation experime nts are carried by means of ARPS ADAS (ARPS Data Analysis System) and 3Dvar (3 D Variation System) respectively. Roles played by ADAS and 3Dvar in improving in itial fields and forecast results are analyzed. Through investi gating one rainstorm process occurred in Anhui Province, July 2003, it indicates that radial velocity data affects initial wind field, along with integral begin ning, water vapor field is adjusted too. Reflectivity data impacts initial water vapor field, and after integration, the wind field is improved. The effect of r eflectivity is more evident than that of radial velocity. Moreover, for the 3h r ainfall simulation, reflectivity and radial velocity assimilation experiments ar e both different with observations in situ, the former one is greater than the o bservation. While the both results of assimilation experiments are better than t hose of the nonassimilation ones.

Abstract:Meiyu onset over YangtseHuaihe river valley is of obvious features of interannual variations. Based on NCEP/NCAR reanalysis dataset and NOAA OLR an d ERSST data, the previous stronger signal of anomalous Meiyu onset over Yangtse Huaihe river valley is investigated by employing correlation and composite analy ses. Preliminary analyses for the impact of the stronger signal on the Meiyu ons et are also studied. The results show that ENSO(Southern Oscillation) events are the previous stronger signal. The Meiyu begins late (early) with the warm (cool) phase of ENSO. The SSTA of Nino4 in February and spring can be regar ded as credible implications for forecasting the onset of Meiyu. The weak convective activ ities around the AsianAustralian “land bridge", Philippines, the western Paci fi c warm pool and Indian peninsula correspond with the warm phase of ENSO. The wea k convective activities around the regions abovementioned may not facilitate n orthward jump of the western Pacific subtropical high and the outbreak of Indian summer monsoon . Consequently, the seasonal transition of atmospheric circulation in the easter n Asian regions is late, which causes the late onset of Meiyu.

Abstract:Numerical forecast models are widely used as an important tool of proj ecting global climate change. However, their resolution is too coarse to provide the regional scale information. Therefore, downscaling methods for extracting r egional scale information from output of numerical forecast models are dev eloped. Statistical downscaling techniques can be used to predict regional scale climate by using statistical relationship between the largescale climate and t he regional scale climate. On the basis of introducing the theory and approach o f statistical downscaling, application of statistical downscaling to the medium term weather forecast was simply discussed. 

Abstract:By using routine observation data, the climate characteristics of the hail in fo ur areas of Guangxi are analyzed. The circulation patterns that causes hail are c lassified to be three kinds: Huabei trough, east plateau trough, south branch troug h. By searching the forecast factors with physics meaning from numerical forecas t products, the potential forecast of hail is produced by using the discriminate analytical method and index accumulating method. The results show that the inde x accumulating method is better than the discriminate analytical method. Based o n the parameter estimation of numerical model output fields, different threshold s for different parameters are set up to produce the probability forecast of the falling area for the hail happened in Guangxi. Process forecast has some effect while the falling area forecast needs to be improved.

Abstract:The TRMM data, conventional observation data, NCAR/NCEP2 reanalysis data and global data assimilation system (GDAS) data are taken to investigate the strong enhancement of tropical cyclone rainstorm when tropical cyclone accompanied with South China Sea monsoon trough lands in South China. Based on the observational facts, the event that the tropical cyclone accompanied with a monsoon trough la nds in South China is defined. The results show that strong enhancement of tropi cal cyclone rainstorm usually happened in July to September. The synopticclima tic characteristics are that subtropical high intensifies zonal spreading wester ly and dominates over middle China. Meanwhile, South China Sea monsoon is in its active period, South China Sea monsoon trough locates over north of South China Sea. The lifetime of tropical cyclone after its landing is prolonged due to the accompanying South China Sea monsoon trough. Cloud system of tropical cyclone has a process of regeneration, strengthening and spreading. Tropical cyclone rai nstorm gains strong increment on rainfall intensity and tropical cyclone lifetim e. The rainfall area stretches on the south of the monsoon trough about 1500～25 00km.

Abstract:Using the Penn StateNCAR nonhydrostatic mesoscale model (MM5), the later per iod of landfall of typhoon Matsa is simulated. A lowpass filter is used to sep arate the flow into a larger and a subsynoptic scale component. Then the sub synoptic scale features of Matsa were investigated based on movable coordinate. From the view of the effect of terrain on the interactions between two scales, t he influence of terrain on tropical cyclone is proved to be nonnegligible furt her. The results are as follows: The existence of terrain helps to maintain the intens ity of tropical cyclone and the effect intensifies with the minishing of distanc e between the corner of cyclone and terrain. The intensification is more evident at upper troposphere. At low level the subsynoptic system converts smaller kinetic energy into tropical cyclone and gets more kinetic energy from tro pical cyclone if terrain is taken out. Meanwhile, the su bsynoptic system gets more positive vertical vorticity from tropical cyclone. In a word, tropical cyclone dos not get but losses positive vertical vorticity i f without terrain.

Abstract:Selforganizing neural network method is applied to classify weather p atterns based on daily NWPs of ECMWF from Jan. 1,2003 to Dec. 31,2006. It shows that the weather pattern is similar over China from November to March. Dynamic di agnosis is applied to daily NWPs of ECMWF in November to March in year 2004-2007 to pick up predictors which have good correlation coefficients with daily maxim um wind speed at 16 coastal weather stations. An updated KNN method is used to s et up wind speed forecast models for November to March. Daily wind speed forecas t for January to March of 2007 is carried out. Results show that KNN method is of good ability in daily maximum wind forecast.

Abstract:With the realtime data minutely measured by the lightning locator, and the pre cipitation data minutely measured from the encrypted precipitation station, as w ell as the satellite cloud picture, the cloudtoground lightning character in a super rainstorm occurred in the south of Shanxi on July 29-30, 2007 was anal yzed. The results show that the cloudtoground lightning appeared in the area of 500hPa, between 5840gpm and 5880gpm, and TBB≤-43℃. The areas surrounded by 3 longitude distance, the left side of low jet flow, and 2 - 3 latitude distance, the south side of 700hPa warm shear, enclosed by the area o f TBB≤-63℃ or the south of cloud cluster, big value zone of TBB horizontal gra ds, are the high frequency and dense areas of the cloudtoground lightning. T here is a good correspondence between these areas and rainstorm falling area. T hree meso βscale convective cloud cluster and one MCC are the main convective systems, which resulted in the super rainstorm. The analysis result indicates t hat the time when a couple of meso βscale cloud cluster incorporates is the t ime when the frequency of lightning is higher. The position where a couple of me so βscale cloud cluster incorporates is the area where the frequency of light ning is higher and the precipitation is stronger. There is a good relationship b etween the minutely cumulated number of cloudtoground lightning from single station and the minute precipitation from the encrypted precipitation station. I t can be used to identify mesoγ scale convective system, remotely survey smal l scale strong precipitation, and forecast the coming of rain intensity peak val ue ahead of 3540 minutes. Only in favorable allocation of upper and lower air system configuration, there will be a perfect correspondence between the local c loudtoground lightning frequency and rain intensity change along with time.

Abstract:The spatialtemporal characteristics, the background of atmospheric ci rculations and the relevant physical elements of the haze weather are statistica lly analyzed based on the observational data and NCEP/NCAR reanalysis data durin g the period from October 2006 to February of the next year. The results show th at higher frequency of the haze occurs in October, December, January and Februar y, and a high incidence of haze mainly occurs at about 11:00-14:00. The main syn optic situations of haze are partitioned into three types, namely the blocking h igh over middle latitude pattern, southern branch trough pattern and zonal circu lation pa ttern. The physical elements exhibit variations form north to south, especially obvious in humidity. During haze occurrence at 14pm, relative humidity is less t han 80% in Huanan Region, 70% in the middlelowvalley of the Yangtze River, less than 6 0% in Huabei Region. The visibility is relatively low in parts of Guangdong Province and Huanghuai Region. Above results are valuable to understand the forming mechanism and to establish forecast system of haze.

Abstract:A largescale haze process that happened on Feb.5, 2007 over middle and eastern China was analyzed based on the air pollution index, conventional observation d ata, and NCEP reanalyzed data. The results show that both the forwardtilling t rough and the temperature inversion are favorable condition for the haze formati on. The absolute values of vorticity, divergence and vertical velocity below 800 hPa are all quite smaller, which favorites to the persistence of haze. The varia tio n of air mixing height is a good indication to the formation and disappearance o f the haze. The mixing layer height appears higher in North China than that in S outh China during the haze periods, which makes it easier to form heavy hazy wea ther in South China. In addition, it was found that the relative humidity is low er, the difference between air temperature and dew temperature is higher, the in tensity of air temperature inversion is weaker and the mixing layer height is hi gher in the haze periods than in the fog periods，which can provide some helpful clew for haze forecasting.

Abstract:Based on principal components analysis (PCA),the BP (Back Propagation) neural ne twork forecast method is introduced in air pollution prediction and the SO2 co ncentration prediction model is established. The results indicate that by applyi ng the principal component analysis in the data preprocessing and taking the principal components of primitive predictor as the input of neural network, it can reduce the dimension of data, eliminate the correlation between the samples, and largel y speed up the convergence rate. The verification of forecast model shows that t he absolute error between the forecasts and the real value is 0.0098, and the co rrelation coefficient between them reaches 0.885. The PCABP model has a fit ac curacy better than the common BP model.

Abstract:By statistical analysis of Doppler radar velocity field data of fortythre e rai nfall processes from June 2003 to October 2005 at Dalian district, the effects o f radar velocity field characters on future rainfall variation and precipitation enhancement condition were gained, the indications of several familiar echo vel ocity characters of the cloud systems on the developing changes of the coming ra infall were obtained. Several criterions were concluded for analyzing the precip itation potential region and the opportunity for precipitation enhancement opera tion while radar echo showed “s” or bow shape. By using these characters as as sistant criterions in precipitation enhancement command, the practical validatio n showed good results in the precipitation enhancement operation at Dalian distr ict in spring and summer of 2006. These criterions have become as important estimating indexes in the realtime precipitation enhancement operat ion control at Dalian district.

Abstract:In order to meet the demand of high spatialtemporal resolution operational for ecast under the situation of abundant NWP products, based on the former recogniz ing study about upper air primary system with contour analysis method, targeting the MICAPS format grid data, using the vector analysis, the algorithms of quant itative parameters for a certain point were setup such as the distance to the af fecting weather system, the intensity, the moving direction and velocity, the te mporal changing rate of the system, the relative position to that system，the lo cation of the jet and the jet intensity also. The quantitative autoanalyzing s oftware for upper air major weather systems was developed with visual BASIC language. The experiment on sandstorm i n northern part of China and cold air activity in southern China shows that the algorithm is effective with high accuracy. The algorithm provides a good basis f or developing automatic forecast system and to rise the modernization level for objective forecast.