Retrieving microphysical properties and air motion of cirrus clouds based on the doppler moments method using cloud radar

Radar parameters including radar reflectivity,Doppler velocity,and Doppler spectrum width were obtained from Doppler spectrum moments.The Doppler spectrum moment is the convolution of both the particle spectrum and the mean air vertical motion.Unlike strong precipitation,the motion of particles in cirrus clouds is quite close to the air motion around them.In this study,a method of Doppler moments was developed and used to retrieve cirrus cloud microphysical properties such as the mean air vertical velocity,mass-weighted diameter,effective particle size,and ice content.Ice content values were retrieved using both the Doppler spectrum method and classic Z-IWC(radar reflectivity-ice water content) relationships;however,the former is a more reasonable method.     

Precipitation Structures of a Thermal Convective System Happened in the Central Western Subtropical Pacific Anticyclone

In this paper, characteristics of precipitating clouds in a thermal convective system (TCS) occurred in the southeastern mainland of China at 15:00 BT (Beijing time) on August 2, 2003 in the central western subtropical Pacific anticyclone (WSPA) is studied by using TRMM tropical rainfallmeasuring mission PR (precipitution radar) and IR Infrared radiation measurements. The precipitating cloud structures in both horizontal and vertical, relationship among storm top, cloud top, and surface rain rate are particularly analyzed. Results show that a strong ascending air at 500 hPa and a strong convergence of moisture flux at 850 hPa in the central WSPA supply necessary conditions both in dynamics and moisture for the happening of the TCS precipitation. The TRMM PR observation shows that the horizontal scale of the most TCS precipitating clouds is about 30-40 km, their averaged vertical scale is above 10 km, and the maximum reaches 17.5 km. The maximum rain rate near surface of those TCS clouds is beyond 50 mm h-1. The mean rain profile of the TCS clouds shows that its maximum rain rate at 5 km altitude is 1 km lower than the estimated freezing level of the environment. Compared with the mesoscale convective system (MCS) of "98.7.20", both systems have the same altitude of the maximum rain rate displayed from both mean rain profiles, but the TCS is much deeper than the MCS. From the altitude of the maximum rain rate to near surface, profiles show that rain rate reducing in the TCS is faster than that in the MCS, which implies a strong droplet evaporation process occurring in the TCS. Relationship among cloud top, storm top, and surface rain rate analysis indicates a large variation of cloud top when storm top is lower. On the contrary, the higher the storm top, the more consistent both cloud top and storm top. And, the larger the surface rain rate, the higher and more consistent for both cloud top and storm top. At the end, results expose that area fractions of non-precipitating clouds and clear sky are 86% and 2%, respectively. The area fraction of precipitating clouds is only about 1/8 that of non-precipitating clouds.     

Radar data assimilation of the GRAPES model and experimental results in a typhoon case

Constructing β-mesoscale weather systems in initial fields remains a challenging problem in a mesoscale numerical weather prediction (NWP) model. Without vertical velocity matching the β-mesoscale weather system, convection activities would be suppressed by downdraft and cooling caused by precipitating hydrometeors. In this study, a method, basing on the three-dimensional variational (3DVAR) assimilation technique, was developed to obtain reasonable structures of β-mesoscale weather systems by assimilating radar data in a next-generation NWP system named GRAPES (the Global and Regional Assimilation and Prediction System) of China. Single-point testing indicated that assimilating radial wind significantly improved the horizontal wind but had little effect on the vertical velocity, while assimilating the retrieved vertical velocity (taking Richardson's equation as the observational operator) can greatly improve the vertical motion. Experiments on a typhoon show that assimilation of the radial wind data can greatly improve the prediction of the typhoon track, and can ameliorate precipitation to some extent. Assimilating the retrieved vertical velocity and rainwater mixing ratio, and adjusting water vapor and cloud water mixing ratio in the initial fields simultaneously, can significantly improve the tropical cyclone rainfall forecast but has little effect on typhoon path. Joint assimilating these three kinds of radar data gets the best results. Taking into account the scale of different weather systems and representation of observational data, data quality control, error setting of background field and observation data are still requiring further in-depth study.     

A 35-GHz Polarimetric Doppler Radar and Its Application for Observing Clouds Associated with Typhoon Nuri

Millimeter-wavelength radar has proved to be an effective instrument for cloud observation and research. In this study, 8-mm-wavelength cloud radar (MMCR) with Doppler and polarization capabilities was used to investigate cloud dynamics in China for the first time. Its design, system specifications, calibration, and application in measuring clouds associated with typhoon are discussed in this article. The cloud radar measurements of radar reflectivity (Z), Doppler velocity (Vr), velocity spectrum width (Sw) and the depolar-ization ratio (LDR) at vertical incidence were used to analyze the microphysical and dynamic processes of the cloud system and precipitation associated with Typhoon Nuri, which occurred in southern China in August 2008. The results show the reflectivity observed using MMCR to be consistent with the echo height and the melting-layer location data obtained by the nearby China S-band new-generation weather radar (SA), but the Ka-band MMCR provided more detailed structural information about clouds and weak precipitation data than did the SA radar. The variation of radar reflectivity and LDR in vertical structure reveals the transformation of particle phase from ice to water. The vertical velocity and velocity spectrum width of MMCR observations indicate an updraft and strong turbulence in the stratiform cloud layer. MMCR provides a valuable new technology for meteorological research in China.     

The First Observed Cloud Echoes and Microphysical Parameter Retrievals by China's 94-GHz Cloud Radar

By using the cloud echoes first successfully observed by China's indigenous 94-GHz SKY cloud radar, the macrostructure and microphysical properties of drizzling stratocumulus clouds in Anhui Province on 8 June2013 are analyzed, and the detection capability of this cloud radar is discussed. The results are as follows.(1) The cloud radar is able to observe the time-varying macroscopic and microphysical parameters of clouds,and it can reveal the microscopic structure and small-scale changes of clouds.(2) The velocity spectral width of cloud droplets is small, but the spectral width of the cloud containing both cloud droplets and drizzle is large. When the spectral width is more than 0.4 m s-1, the radar reflectivity factor is larger(over –10 dBZ).(3) The radar's sensitivity is comparatively higher because the minimum radar reflectivity factor is about–35 dBZ in this experiment, which exceeds the threshold for detecting the linear depolarized ratio(LDR) of stratocumulus(commonly –11 to –14 dBZ; decreases with increasing turbulence).(4) After distinguishing of cloud droplets from drizzle, cloud liquid water content and particle effective radius are retrieved. The liquid water content of drizzle is lower than that of cloud droplets at the same radar reflectivity factor.     

The First Observed Cloud Echoes and Microphysical Parameter Retrievals by China's 94-GHz Cloud Radar

By using the cloud echoes first successfully observed by China's indigenous 94-GHz SKY cloud radar, the macrostructure and microphysical properties of drizzling stratocumulus clouds in Anhui Province on 8 June2013 are analyzed, and the detection capability of this cloud radar is discussed. The results are as follows.(1) The cloud radar is able to observe the time-varying macroscopic and microphysical parameters of clouds,and it can reveal the microscopic structure and small-scale changes of clouds.(2) The velocity spectral width of cloud droplets is small, but the spectral width of the cloud containing both cloud droplets and drizzle is large. When the spectral width is more than 0.4 m s-1, the radar reflectivity factor is larger(over –10 dBZ).(3) The radar's sensitivity is comparatively higher because the minimum radar reflectivity factor is about–35 dBZ in this experiment, which exceeds the threshold for detecting the linear depolarized ratio(LDR) of stratocumulus(commonly –11 to –14 dBZ; decreases with increasing turbulence).(4) After distinguishing of cloud droplets from drizzle, cloud liquid water content and particle effective radius are retrieved. The liquid water content of drizzle is lower than that of cloud droplets at the same radar reflectivity factor.     

QUALITY CONTROL OF SINGLE DOPPLER RADAR DATA AND RETRIEVAL OF HORIZONTAL WIND FOR A LANDING TYPHOON

The removal of noise and velocity ambiguity and retrieval and verification of horizontal wind field is a prerequisite to make the best and fullest use of Doppler radar measurements. This approach was applied to the Doppler radar data collected during August 2005 for a landing typhoon Matsa (0509) in Yantai, Shangdong Province, and the verified result shows that the quality control for this dataset was successful. The horizontal wind field was retrieved and then verified by studying the characteristics of the radar radial velocity and large-scale wind field and the vertical cross section of the radial velocity determined with the typhoon center as the circle center and comparing it with satellite imagery. The results show that the meso- and small-scale systems in Matsa and its horizontal and vertical structure could be clearly retrieved using the dataset collected by single Doppler radar, and a shear or a convergence was corresponding with a band of severe storm around Matsa. At the same time, the retrieved wind field from single Doppler radar is proved to be a reliable and high-resolution dataset in analyzing the inner meso-scale structure of Matsa. It is also proved that the method for removing the velocity ambiguity could be an effective approach for preliminary quality control of the Doppler radar data, and the VAP method could also be a reasonable solution for the analysis of mesoscale wind field.     

Application of Radar Reflectivity Factor in Initializing Cloud-Resolving Mesoscale Model. Part Ⅱ： Numerical Simulation Experiments

Microphysics elements and vertical velocity retrieved were incorporated using the nudging method into the initial data assimilation of GRAPES （Global/Regional Assimilation and Prediction System） model. Simulation experiments indicated that nudging technique was effective in forcing the model forecast gradually consistent to the observations, yielding the thermodynamically and dynamically balanced analysis field. As viewed from the simulation results, water vapor is vital to precipitation, and it is a governing factor for the amount and duration of precipitation. The initial cloud water, rain water, and vertical velocity determine the strength distribution of convection and precipitation at the beginning time of forecast; the horizontal wind field steers the motion of the mesoscale weather system embedded in and impacts the position of precipitation zone to a large extent. The simulation experiments show that the influence of the initial retrieval data on prediction weakens with the increase of forecast time, and within the first hour of forecast, the retrieval data have an important impact on the evolution of the weather system, but its influence becomes trivial after the first three hours. Changing the nudging coefficient and the integral time-spacing of numerical model will bring some influences to the results. Herein only one radar reflectivity was used, the radar observations did not cover the whole model domain, and some empirical parameters were used in the retrieval method, therefore some differences still lie between simulation and observation to a certain extent, and further studies on several aspects are expected.     

Analysis of the Microphysical Properties of a Stratiform Rain Event Using an L-Band Profiler Radar

This paper investigates spatial and temporal distributions of the microphysical properties of precipitating stratiform clouds based on Doppler spectra of rain particles observed by an L-band profiler radar.The retrieval of raindrop size distributions（RSDs） is accomplished through eliminating vertical air motion and isolating the terminal fall velocity of raindrops in the observed Doppler velocity spectrum.The microphysical properties of raindrops in a broad stratiform region with weak convective cells are studied using data collected from a 1320-MHz wind profiler radar in Huayin,Shaanxi Province on 14 May 2009.RSDs and gamma function parameters are retrieved at altitudes between 700 and 3000 m above the surface,below a melting layer.It is found that the altitude of the maximum number of raindrops was closely related to the surface rain rate.The maximum number of large drops was observed at lower altitudes earlier in the precipitation event but at higher altitudes in later periods,suggesting decreases in the numbers of large and medium size raindrops.These decreases may have been caused by the breakup of larger drops and evaporation of smaller drops as they fell.The number of medium size drops decreased with increasing altitude.The relationship between reflectivity and liquid water content during this precipitation event was Z = 1.69×10~4M~（1.5）,and the relationship between reflectivity and rain intensity was Z = 256I~（1.4）.     

Three-Dimensional Wind Field Retrieved from Dual-Doppler Radar Based on a Variational Method: Refinement of Vertical Velocity Estimates

In this paper,a scheme of dual-Doppler radar wind analysis based on a three-dimensional variational method is proposed and performed in two steps.First,the horizontal wind field is simultaneously recovered through minimizing a cost function defined as a radial observation term with the standard conjugate gradient method,avoiding a weighting parameter specification step.Compared with conventional dual-Doppler wind synthesis approaches,this variational method minimizes errors caused by interpolation from radar observation to analysis grid in the iterative solution process,which is one of the main sources of errors.Then,through the accelerated Liebmann method,the vertical velocity is further reestimated as an extra step by solving the Poisson equation with impermeable conditions imposed at the ground and near the tropopause.The Poisson equation defined by the second derivative of the vertical velocity is derived from the mass continuity equation.Compared with the method proposed by O’Brien,this method is less sensitive to the uncertainty of the boundary conditions and has better stability and reliability.Furthermore,the method proposed in this paper is applied to Doppler radar observation of a squall line process.It is shown that the retrieved vertical wind profile agrees well with the vertical profile obtained with the velocity–azimuth display(VAD)method,and the retrieved radial velocity as well as the analyzed positive and negative velocity centers and horizontal wind shear of the squall line are in accord with radar observations.There is a good correspondence between the divergence field of the derived wind field and the vertical velocity.And,the horizontal and vertical circulations within and around the squall line,as well as strong updrafts,the associated downdrafts,and associated rear inflow of the bow echo,are analyzed well.It is worth mentioning that the variational method in this paper can be applied to simultaneously synthesize the three-dimensional wind field from multiple-Doppler radar observations.     

Characteristics of Gaseous Pollutants at a Regional Background Station in Southern China

Measurements of gaseous pollutants （03, NOx, SO2, and CO） were conducted at Dinghushan background station in southern China from January to December 2013. The levels and variations of O3, NOx, SO2, and CO were analyzed and their possible causes discussed. The annual average concentrations of 03, NOx, SO2, and CO were 24.6 ± 23.9, 12.8 ± 10.2, 4.0 ± 4.8, and 348 ± 185 ppbv, respectively. The observed levels of the gaseous pollutants are comparable to those at other background sites in China. The most obvious diurnal variation of 03 was observed in autumn, with minima in the early morning and maxima in the afternoon. The diurnal variations of SO2 showed high values during the day. The diurnal cycles of NOx showed higher values in the morning and lower values during the night. Higher CO concentrations were observed in spring followed by winter, autumn, and summer. Biomass burning, in combination with the transport of regional pollution, is an important source of CO, SO2, and NOx in spring and winter. Backward trajectories were calculated and analyzed together with corresponding pollutant concentrations. The results indicate that air masses passing over polluted areas are responsible for the high concentrations of gaseous pollutants at the Dinghushan background station.     

Integration of Multi-source Measurements to Monitor Dust over North China：A Case Study

The aerosol optical depth （AOD） data from the Moderate Resolution Imaging Spectroradiometer （MODIS） aboard Satellite Aqua, along with the altitude-resolved aerosol subtypes product from the Cloud-Aerosol LIdar with Orthogonal Polarization （CALIOP）, as well as surface PM 10 measurements, were utilized to investigate the dust activities common in springtime of northern China. Specifically, a dust storm episode that occurred over the North China Plain （NCP） during 17-21 March 2010 was identified. The PM 10 concentration at Beijing （39.8 °N, 116.47 °E） reached the peak value of 283 μgm -3 on 20 March 2010 from the background value of 15 μg m-3 measured on 17 March 2010, then dropped to 176 μgm-3 on 21 March 2010. Analysis of the CALIOP aerosol subtypes product showed that numerous large dust plumes floated over northern China, downwind of main desert source regions, and were lifted to altitudes as high as 3.5 km during this time period. The MODIS AOD data provided spatial distributions of dust load, broadly consistent with ground-level PM 10 , especially in cloud free areas. However, inconsistency between the MODIS AOD and surface PM 10 measurements under cloudy conditions did exist, further highlighting the unique capability of the CALIOP lidar. CALIOP can penetrate the cloud layer to give unambiguous and altitude-resolved dust measurements, albeit a relatively long revisit period （16 days） and narrower swath （90 m）. A back trajectory simulation using the Hybrid Single-Particle Lagrangian Integrated Trajectory （HYSPLIT） model was performed, and it was found that the sand-dust storm originated from the Gobi Desert on 18 March 2010 travelled approxi-mately 1200-1500 km day-1 eastward and passed over the NCP on 19 March 2010, in good agreement with previous findings. In addition, the multi-sensor measurements integrated with the HYSPLIT model output formed a three-dimensional view of the transport pathway for this dust episode, indicating that this episode was largely associated with the desert source regions to the northwest of the NCP. The results imply the importance of integration of multi-sensor measurements for clarifying the overall structure of dust events over northern China.     

Analysis of Nonstationary Wind Fluctuations Using the Hilbert-Huang Transform

Climatological patterns in wind fluctuations on time scales of 1-10 h are analyzed at a meteorological mast at the Yangmeishan wind farm, Yunnan Province, China, using a 2-yr time series of 10-min wind speed ob- servations. For analyzing the spectral properties of non- stationary wind fluctuations in mountain terrain, the Hil- bert-Huang transform （HHT） is applied to investigate climatological patterns between wind variability and sev- eral variables including time of year, time of day, wind direction, and pressure tendency. Compared with that for offshore sites, the wind variability at Yangmeishan wind farm has a more distinct diurnal cycle, but the seasonal discrepancies and the differences according to directions are not distinct, and the synoptic influences on wind vari- ability are weaker. There is enhanced variability in spring and winter compared with summer and autumn. For flow from the main direction sector, the maximum wind vari- ability is observed in spring. And the severe wind fluctua- tions are more common when the pressure tendency is rising.     

Results of a CLM4 Land Surface Simulation over China Using a Multisource Integrated Land Cover Dataset

In this study, the high-accuracy multisource integrated Chinese land cover （MICLCover） dataset was used in version 4 of the Community Land Model （CLM4） to assess how the new land cover information affected land surface simulation over China. Compared to the default land cover dataset in CLM4, the MICL data indicated lower values for bare soil （14.6% reduction）, nee- dleleaf tree （3.6%）, and broadleaf tree （1.9%）; higher values for shrub cover （1.8% increase）, grassland （9.9%）, cropland （5.0%）, glaciers （0.5%）, lakes （1.6%）, and wetland （1.1%）; and unchanged for urban areas. Two comparative CLM4 simulations were conducted for the 33-yr period from 1972 to 2004, one using the MICL dataset and the other using the default dataset. The results revealed that the MICL data produced a 0.3% lower mean annual surface albedo over China than the original data. The largest contributor to the reduced value was semiarid regions （2.1% reduction）. The MICL-data albedo value agreed more closely with observations （MODIS broad- band black-sky albedo products） over arid and semiarid regions than for the original data to some extent. The simulated average sensible heat flux over China increased by only 0.1 W m 2 owing to the reduced values in arid and semiarid regions, as opposed to increases in humid and semihumid regions, while an increased latent heat flux of I W m-2 was reflected in almost identical changes over the whole region. In addition, the mean annual runoff simulated by CLM4 using MICL data decreased by 6.8 mm yr-1, primarily due to large simulated decreases in humid regions.     

Marine-Atmospheric Boundary Layer Characteristics over the South China Sea During the Passage of Strong Typhoon Hagupit

The structures and characteristics of the marine-atmospheric boundary layer over the South China Sea during the passage of strong Typhoon Hagupit are analyzed in detail in this paper. The typhoon was generated in the western Pacific Ocean, and it passed across the South China Sea, finally landfalling in the west of Guangdong Province. The shortest distance between the typhoon center and the observation station on Zhizi Island(10 m in height) is 8.5 km. The observation data capture the whole of processes that occurred in the regions of the typhoon eye, two squall regions of the eye wall, and weak wind regions,before and after the typhoon's passage. The results show that:(a) during the strong wind(average velocityˉu 10 m s-1) period, in the atmospheric boundary layer below 110 m, ˉu is almost independent of height,and vertical velocity ˉw is greater than 0, increasing with ˉu and reaching 2–4 m s-1in the squall regions;(b) the turbulent fluctuations(frequency 1/60 Hz) and gusty disturbances(frequency between 1/600 and1/60 Hz) are both strong and anisotropic, but the anisotropy of the turbulent fluctuations is less strong;(c) ˉu can be used as the basic parameter to parameterize all the characteristics of fluctuations; and(d) the vertical flux of horizontal momentum contributed by the average flow(ˉu ·ˉw) is one order of magnitude larger than those contributed by fluctuation fluxes(u w and v w), implying that strong wind may have seriously disturbed the sea surface through drag force and downward transport of eddy momentum and generated large breaking waves, leading to formation of a strongly coupled marine-atmospheric boundary layer. This results in ˉw 0 in the atmosphere, and some portion of the momentum in the sea may be fed back again to the atmosphere due to ˉu ·ˉw 0.     

Seasonal Prediction of the Global Precipitation Annual Modes with the Grid-Point Atmospheric Model of IAP LASG

A right annual cycle is of critical importance for a model to improve its seasonal prediction skill. This work assesses the performance of the Grid-point Atmospheric Model of IAP LASG （GAMIL） in retrospective prediction of the global precipitation annual modes for the 1980 2004 period. The annual modes are gauged by a three-parameter metrics： the long-term annual mean and two major modes of annual cycle （AC）, namely, a solstitial mode and an equinoctial asymmetric mode. The results demonstrate that the GAMIL one-month lead prediction is basically able to capture the major patterns of the long-term annual mean as well as the first AC mode （the solstitial monsoon mode）. The GAMIL has deficiencies in reproducing the second AC mode （the equinoctial asymmetric mode）. The magnitude of the GAMIL prediction tends to be greater than the observed precipitation, especially in the sea areas including the Arabian Sea, the Bay of Bengal （BOB）, and the western North Pacific （WNP）. These biases may be due to underestimation of the convective activity predicted in the tropics, especially over the western Pacific warm pool （WPWP） and its neighboring areas. It is suggested that a more accurate parameterization of convection in the tropics, especially in the Maritime Continent, the WPWP and its neighboring areas, may be critical for reproducing the more realistic annual modes, since the enhancement of convective activity over the WPWP and its vicinity can induce suppressed convection over the WNP, the BOB, and the South Indian Ocean where the GAMIL produces falsely vigorous convections. More efforts are needed to improve the simulation not only in monsoon seasons but also in transitional seasons when the second AC mode takes place. Selection of the one-tier or coupled atmosphere-ocean system may also reduce the systematic error of the GAMIL prediction. These results offer some references for improvement of the GAMIL seasonal prediction skill.     

Characterization of Stratospheric Aerosol Distributions during the Volcanically Quiescent Period of 1998–2004

The Stratospheric Aerosol and Gas Experiment （SAGE） II aerosol extinction profiles at 1020 nm were used to study the distribution characteristics of stratospheric aerosols during the volcanically quiescent period of 1998-2004. The stratospheric aerosol distributions exhibited hemispheric asymmetry between the Northern Hemisphere （NH） and the Southern Hemisphere （SH）. In the lower stratosphere below 20 km, the zonal averaged aerosol optical depths in the NH were higher than those of the corresponding SH; whereas at higher altitudes above 20 km, the optical depths in the SH-- except the equatorial region--were higher than those of the NH. At 0-10°N and 10-20°N, the stratospheric aerosol optical depth （SAOD） exhibited larger values in boreal winter and lower values in the spring and summer; at 0-10°S and 10-20°S, the SAOD presented small seasonal variations. At 30-40°N, the SAOD presented larger values in the boreal fall and winter and lower values in the spring and summer; while at 30-40°S, the SAOD exhibited larger values in the austral winter and early spring and lower values in the summer and fall. These characteristics can mainly be attributed to the seasonal cycle of the dynamic transport, and the effects of the buildup and breakdown of the polar vortex. At 50-60°S, the SAOD exhibited extremely high values during austral winter associated with the Antarctic polar vortex boundary; at 50-60°N, the SAOD also exhibited larger values during the boreal winter, but it was much less obvious than that of its southern counterpart.     

一次冰雹过程的卫星反演分析

A case of hailstorm process occurring on 24 June 2006 in northwestern China was studied using satellite retrieval methodology. The particle effective radius （re） in the cloud tops was calculated by the reflectance in the 3.7μm channel, and cloud-top microphysical properties were vividly represented using the RGB visual multispectral classification scheme. The microphysical zones of clouds and the processes of hail formation and development are inferred using the relations of cloud-top temperature （T） versus re for the tops of convective clouds. The results show that particle effective radius was smaller near the cloud base of hailstorm. There was a deep zone of diffusional droplet growth at the low level where the particles grew slowly with height, and there existed an evident area of small ice particles in the cloud top, suggesting the existence of a strong updraft in the clouds. The low glaciated temperature indicated a great depth from the cloud base to the glaciation height, which provided a deep layer of supercooled water for hail growth.     

The Combined Effect of Initial Error and Model Error on ENSO Prediction Uncertainty Generated by the Zebiak-Cane Model

Initial errors and model errors are the source of prediction errors. In this study, the authors compute the conditional nonlinear optimal perturbation （CNOP）-type initial errors and nonlinear forcing singular vector （NFSV）- type tendency errors of the Zebiak-Cane model with respect to El Nifio events and analyze their combined effect on the prediction errors for E1 Nino events. The CNOP- type initial error （NFSV-type tendency error） represents the initial errors （model errors） that have the largest effect on prediction uncertainties for E1 Nifio events under the perfect model （perfect initial conditions） scenario. How- ever, when the CNOP-type initial errors and the NFSV- type tendency errors are simultaneously considered in the model, the prediction errors caused by them are not am- plified as the authors expected. Specifically, the predic- tion errors caused by the combined mode of CNOP-type initial errors and NFSV-type tendency errors are a little larger than those caused by the NFSV-type tendency er- rors. This fact emphasizes a need to investigate the opti- mal combined mode of initial errors and tendency errors that cause the largest prediction error for E1 Nifio events.     

The Second Decadal Leading Mode of East Asian Summer Monsoon Rainfall

The first decadal leading mode of East Asian summer rainfall （EASR） is characterized by rainfall anomalies along the East Asian subtropical rain belt. This study focuses on the second decadal leading mode （2DLM）, accounting for 17.3% of rainfall decadal vari- ance, as distinct from the other two neighboring modes of EAMR, based on the state-of-the-art in-situ rainfall data. This mode is characterized by a South-China-wet-Huaihe- River-dry pattern, and is dominated by a quasi-30-yr pe- riod. Further analysis reveals the 2DLM corresponds to an enhanced lower-level monsoon jet, an eastward extension of the western North Pacific subtropical high, and a weakened East Asian upper-level westerly jet flow. The Tibetan Plateau surface temperature and Pacific Decadal Oscillation （PDO） are closely linked with the 2DLM. The regressed SST pattern indicates the PDO-like pattern of sea surface temperature anomalies may have a telecon- nection relationship with the 2DLM of EASR.