The seasonal prediction of both the surface air temperature and the first-flowering date (FFD) over South Korea are produced using dynamical downscaling. Dynamical downscaling is performed using Weather Research and Forecast (WRF) v3.0 with the lateral forcing from hourly outputs of Pusan National University (PNU) coupled general circulation model (CGCM) v1.1. Gridded surface air temperature data with high spatial (3km) and temporal (daily) resolution are obtained using the physically-based dynamical models. To reduce systematic bias, simple statistical correction method is then applied to the model output. The FFDs of cherry, peach and pear in South Korea are predicted for the decade of 1999-2008 by applying the corrected daily temperature predictions to the phenological thermal-time model. The WRF v3.0 results reflect the detailed topographical effect, despite having cold and warm biases for warm and cold seasons, respectively. After applying the correction, the mean temperature for early spring (February to April) well represents the general pattern of observation, while preserving the advantages of dynamical downscaling. The FFD predictabilities for the three species of trees are evaluated in terms of qualitative, quantitative and categorical estimations. Although FFDs derived from the corrected WRF results well predict the spatial distribution and the variation of observation, the prediction performance has no statistical significance or appropriate predictability. The approach used in the study may be helpful in obtaining detailed and useful information about FFD and regional temperature by accounting for physically-based atmospheric dynamics, although the seasonal predictability of flowering phenology is not high enough.

Acknowledgements

This work was carried out with the support of the Rural Development Administration Cooperative Research Program for Agriculture Science and Technology Development under Grant Project No. PJ009953 and the Korea Meteorological Administration Research and Development Program under Grant CATER 2012-3100, Republic of Korea.

Reference

Hur J, Ahn J-B. 2015 Seasonal Prediction of Regional Surface Air Temperature and First-flowering Date over South Korea, Int. J. Climatol., under revision (conditionally accepted)

The summer monsoon considerably affects water resource and natural hazards including flood and drought in East Asia, one of the world’s most densely populated area. In this study, we investigate future changes in summer precipitation over East Asia induced by global warming through dynamical downscaling with the Weather Research and Forecast model. We have selected a global model from the Coupled Model Intercomparison Project Phase 5 based on an objective evaluation for East Asian summer monsoon and applied its climate change under Representative Concentration Pathway 4.5 scenario to a pseudo global warming method. Unlike the previous studies that focused on a qualitative description of projected precipitation changes over East Asia, this study tried to identify the physical causes of the precipitation changes by analyzing a local moisture budget. Projected changes in precipitation over the eastern foothills area of Tibetan Plateau including Sichuan Basin and Yangtze River displayed a contrasting pattern: a decrease in its northern area and an increase in its southern area. A local moisture budget analysis indicated the precipitation increase over the southern area can be mainly attributed to an increase in horizontal wind convergence and surface evaporation. On the other hand, the precipitation decrease over the northern area can be largely explained by horizontal advection of dry air from the northern continent and by divergent wind flow. Regional changes in future precipitation in East Asia are likely to be attributed to different mechanisms which can be better resolved by regional dynamical downscaling.

The Philippines is one of the most vulnerable countries to the impacts of climate change, as evidenced by the damage to communities and economies caused by recent climate-related disasters. This study aims to provide high-resolution precipitation and temperature data for four Philippine cities. The cities were selected based on their geographical location, economic importance and recent history of climate-related disasters. Output from four global climate models (GCMs) (CanESM2, HadGEM2-ES, GFDL, and CSIRO-Mk3.6.0) was downscaled for the Philippines at 25-km resolution using the regional climate model  (RegCM4.3.5.6) for the baseline period (1971-2000) and future periods (2025 and 2050). Comparison of model output with observed data from the Philippine Atmospheric Geophysical and Astronomical Services Administration (PAGASA) showed that the monthly and seasonal distribution of precipitation and temperature are simulated well by RegCM4. However, bias correction techniques have also been done to address differences between model and observed. The projection runs showed that the amount of precipitation will not significantly shift, although rainfall patterns will become more extreme relative to current trends. For the same period, temperature across all cities is projected to increase by 1.2-2.9°C relative to the baseline mean. This study reflects similar results from previous work, which can be used to improve disaster risk reduction management in the study cities and build communities that are more resilient to the impacts of future climate extremes.

In this study, we evaluate the performance of the regional climate model version 4 (RegCM4), which includes the Biosphere–Atmosphere Transfer Scheme (BATS) and Community Land Model (CLM3) land-surface schemes, in simulating the summer precipitation over East Asia. The precipitation characteristics are investigated in terms of mean amount, frequency and intensity of daily precipitation. The results show that the simulation of the summer precipitation is significantly sensitive to the choices of the land-surface schemes. Despite several deficiencies, the simulation of daily precipitation with CLM3 exhibits superior performance to that with BATS. The BATS simulation tends to systematically overestimate both precipitation frequency and intensity, and hence total precipitation across the whole domain. On the other hand, the CLM3 simulation substantially reduces the wet biases produced in the BATS simulation. The difference in performance between the two simulations mainly results from convective precipitation rather than large-scale precipitation. Since excessive convective precipitation tends to suppress large-scale precipitation, the BATS simulation also exhibits a limitation in properly simulating the ratio of convective and large-scale precipitation. Such behaviour can be explained by the influence of soil moisture on convective precipitation. Persistently wetter soil moisture in the BATS land-surface scheme can modulate the partitioning of surface heat fluxes inadequately, leading to overestimation of latent heat lux and underestimation of sensible heat lux over South China, in particular. Consequently, it affects the thermodynamic structure, which in turn affects the atmospheric stability to determine the convective activity. The CLM3 simulation generates a more realistic representation of equivalent potential temperature, convective available potential energy and convective inhibition, and thus improves the characteristics of daily precipitation.

Most of areas in the Southwestern United States (SW US) receive a significant amount of precipitation (over half their annual precipitation) during the North American Monsoon (NAM) season.  In this arid to semi-arid region, the overall agricultural system is substantially affected by climate factors. Therefore, accurate climatic data (i.e. precipitation, temperature, and radiation) are essential to assess agricultural productivity.  However, 21 years of WRF Regional Climate Model (RCM) results over SW US show a strong dry bias during the warm season, especially in Arizona.  Consequently, crop model results shows very low crop yields compared to observed yields in this region.  We suspect that the coarse resolution of reanalysis 2 data could not resolve the relatively warm Sea Surface Temperature (SST) in the Gulf of California, causing the SST to be up to 10 degrees lower than the climatology data.  During the monsoon season, the low-level moisture is advected to the SW US via the Gulf of California, which is known to be the dominant moisture source. Thus, high-resolution SST data in the Gulf of California are required for RCM simulations to accurately represent a reasonable amount of precipitation in the region, allowing reliable evaluation of the impacts on regional eco- and agricultural systems. 

To evaluate the influence of SST on agriculture in the SW US, two sets of numerical simulations were employed in this study. As a control run, we used unresolved SST for the Gulf of California, 0.5 by 0.5 degree AMIP II SST data, used widely in RCM studies (e.g. NARCCAP). And another set employed daily updated high resolution (9.26km) SST data from the MODIS satellite sensor. First, we analyzed the differences of meteorological drivers from each of the 6 year RCM runs. And then the meteorological data were provided as input to the Agricultural Production Systems sIMulator (APSIM) crop model to determine resultant crop yields. Analyses of the simulated crop production, and the interannual variation of the meteorological drivers, demonstrate the influence of SST on crop yields in the SW US.

Recent efforts at the International Pacific Research Center (IPRC) of the University of Hawaii at Manoa have being made to conduct long-term sub-kolimeter dynamical downscaling for regional climate and its change over the Pacific Islands with steep orography under the financial support of the USGS Pacific Island Climate Science Center (PICSC). In this presentation both the theoretical basis and the practical feasibility will be briefly discussed first. Some preliminary results from two funded projects, one for the Hawaiian Island and one for Guam and American Samoa, will be introduced as a demonstration of the merits with the use of sub-kolimeter model resolution to resolve the island-scale microclimate zones in thses two regions. In particualr the dynamically downscaled tropical cyclone activity over the western North Pacific for the present cliamte and projected its change by the later 21st century will be highlighted. Potential applicaitons of the approach that we have developed to other Pacific Islands will be discussed.

Extreme hot spells can have significant impacts on human society and ecosystems, and therefore it is important to assess how these extreme events will evolve in a changing climate. In this study, the impact of climate change on hot days, hot spells, and heat waves, over 10 climatic regions covering Canada, based on 11 Regional Climate Model (RCM) simulations from the North American Regional Climate Change Assessment Program (NARCCAP) for the June to August summer period, will be presented. These simulations were produced with six RCMs driven by four Atmosphere-Ocean General Circulation Models (AOGCM), for the A2 emission scenario, for the current 1970–1999 and future 2040–2069 periods. Two types of hot days, namely HD-1 and HD-2, defined respectively as days with only daily maximum temperature (Tmax) and both Tmax and daily minimum temperature (Tmin) exceeding their respective thresholds (i.e., period-of-record 90^th percentile of Tmax and Tmin values), are considered in the study. Analogous to these hot days, two types of hot spells, namely HS-1 and HS-2, are identified as spells of consecutive HD-1 and HD-2 type hot days. In the study, heat waves are defined as periods of three or more consecutive days, with Tmax above 32°C threshold. Results suggest future increases in the number of both types of hot days and hot spell events for the 10 climatic regions considered. However, the projected changes show high spatial variability and are highly dependent on the RCM and driving AOGCM combination. Extreme hot spell events such as HS-2 type hot spells of longer duration are expected to experience relatively larger increases compared to hot spells of moderate duration, implying considerable heat related environmental and health risks. Regionally, the Great Lakes, West Coast, Northern Plains, and Maritimes regions are found to be more affected due to increases in the frequency and severity of hot spells and/or heat wave characteristics, requiring more in depth studies for these regions to facilitate appropriate adaptation measures.

Climate change is expected to exacerbate the extremes in the climate variables, mainly, rainfall and temperature. Precipitation extremes are expected to increase in intensity and frequency over many regions in the world due to global warming. The general circulation models (GCMs), which are primary physically based tools that give information on current and future climates, provide reasonable climate simulations (when compared to observations) only in a global or a continental scale, but at a regional/sub-regional/city scale, such climate simulations are rather not good enough due to their coarse spatial resolution. Thus, there is a need to ‘downscale’ the GCM information to regional/sub-regional/city (urban) scales for applications in impact studies. One of the widely used methods to downscale is the use of high resolution regional climate models (aka RCMs) and the technique called ‘dynamical downscaling’. This presentation focus on the applications of a regional climate model, Weather Research and Forecasting (WRF) for climate impacts studies, especially over regions that lack robust observed data. The usefulness of the RCM as ‘proxies’ in such a context is evident and showcases the potential of RCMs to be used for varied climate impacts studies.

The retrospective forecasts of SINTEX-F2 CGCM with one month lead are downscaled over three different study regions viz Southern Africa, Australia and India using the Weather Research and Forecasting (WRF) regional model. During boreal summer the Indian subcontinent receives most of its annual precipitation and also the southwest of West Australia receives rainfall during the same period. Southern Africa and the eastern parts of Australia receive rainfall during the boreal winter season. Hence, the downscaling was for the boreal summer (winter) from June to Sep (Dec to Feb) for former (latter) study domain(s). Improving the seasonal forecasts over these regions would contribute to the economics of the regions.

Various experiments are carried out to improve the value addition of the WRF model generated forecasts to the SINTEX-F2 forecasts. One of the experiments, the WRF model was coupled to a simple mixed later ocean model. The other technique was to bias correct the SINTEX-F2 forecasts before they are downscaled. Calculation of the anomaly correlation, bias adjusted equitable threat score, Relative Operating Characteristic (ROC) and Relative Operating Level (ROL) scores shows that the WRF downscaling adds value to the SINTEX-F2 forecasts when a mean bias correction is employed prior to the downscaling. The downscaling improved the spatial and temporal distribution of the anomalies. It is found that the WRF model driven by the bias corrected SINTEX-F2 forecasts results in larger value addition to the SINTEX-F2 forecasts compared to the WRF model driven by the uncorrected SINTEX-F2 forecasts.

An assessment of the Regional Climate Model version 4.3 (RegCM4.3) using the Community Land Model version 3.5 (CLM), a recently implemented coupled land surface model, over the Loess Plateau in China is performed against the older Biosphere-Atmosphere Transfer Scheme (BATS). The Loess Plateau is located in a semi-arid transition zone characterized by a pronounced heterogeneity in topography and vegetation cover. Experiments with three convective parameterizations (Grell with the Fritsch-Chappell closure, Grell with the Arakawa-Schubert closure, and the Emanuel scheme) spanning a period from 1990 to 2009 with a 50-km horizontal resolution are conducted for CLM and BATS respectively.

There are two significant differences between the corresponding CLM and BATS simulations: 1) the CLM tends to reduce the amount of precipitation during both the winter and summer, and 2) the CLM increases the temperature during the winter but decreases it during the summer. These differences persist, to a greater or lesser degree for each corresponding convective scheme and are mostly related to differences in the water and energy budgets between the two land surface models. For example, the reduced precipitation simulated by the CLM is primarily caused by a significantly lower evapotranspiration rate and excessive runoff, although the drier conditions are mitigated slightly by the positive moisture advections resulted from stronger southerly monsoon flows. The temperature differences between the two land surface models may be explained by the differences in cloud coverage: the excessively high cloud coverage, which is likely triggered by stronger simulated southerlies, over the central band of the Loess Plateau during the summer significantly corresponds to the lower surface temperature present in the CLM. Given the improvements brought by the CLM, however, the limitation in representing the realistic land uses over the Loess Plateau and potential impacts to projecting regional climate is noted.

The Weather@Home project, originally developed at the University of Oxford, utilises volunteers' computing power to generate very large ensembles of simulations representing regional climate. The Weather@Home Australia-New Zealand (ANZ) project has been set up over the CORDEX ANZ domain to examine the anthropogenic influence on climate over the region. Tens of thousands of model realisations are generated representing the world as it is and a world without human influences. These can then be used to make quantifiable estimates of anthropogenic influences on extreme events. Here, I will outline the model set-up and some results highlighting the benefits that this project offers over global climate modelling projects such as CMIP5. For example, an analysis of the record-breaking heat over Australia in early 2013 on different spatial and temporal scales shows significant and substantial increases in the likelihood of extreme heat due to anthropogenic influences. The data from Weather@Home is freely available and we encourage other scientists to make use of it.

In this study, the Weather Research and Forecasting (WRF) model was employed to verify the performance of the model in downscaling climate variability over Southeast Asia at 27-km horizontal resolution. The model was initialized at 00 UTC on 1 January 1990 and integrated for 23 years (1990-2012). The ECMWF interim reanalysis data were used as initial and boundary conditions for the model. The observed sea surface temperature is used as a bottom boundary condition and updated every six hours. Model performance is evaluated by verifying mean climatology and interannual variability in terms of circulation features and rainfall over Southeast Asia for the different seasons of the year. The model output is compared with the available observation and reanalysis data. The capability of WRF model in downscaling the large scale climate mode such as El Niño Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) is verified over southeast Asia - for the continental (consisting Myanmar, Cambodia, Laos, Thailand, Vietnam and Peninsular Malaysia) and maritime (consisting Brunei, East Malaysia, East Timor, Indonesia, Papua New Guinea, the Philippines, and Singapore) regions. Results indicate that the model could reproduce the observed mean climatology and intraseasonal to interannual rainfall variability reasonably well compared to the ERA-Interim data. The analysis also suggests that the model could simulate the rainfall variation associated with ENSO and IOD.  The better simulation of the rainfall in the model seems to be due to the realistic simulation of large scale circulation and moisture fluxes over this region. The results obtained in this study will be useful for the forecasting experiments with the boundary conditions from the coupled general circulation model.

As part of the NSW/ACT Regional Climate Modelling (NARCliM) project three model configurations of the Weather Research and Forecasting (WRF) model have been chosen based on performance and independence. The independence measure is based on the covariance of model errors and ensures that error characteristics differ between the configurations. Each configuration has been run for 60-years over the CORDEX-AustralAsia domain driven by the NCEP/NCAR reanalysis. This talk examines the performance of these WRF configurations over Australia. Results demonstrate the independence of the model errors. Further examination of the teleconnections with large scale climate modes (e.g. El Nino) and trends in extremes generally demonstrates that the regional models perform better than the driving reanalysis. These improvement vary between configurations.

Here we introduce a downscaling program named “Development of basic technology for risk information on climate change”, which is a theme of SOUSEI program (FY2012 – FY2016) supported by MEXT, Japan. The program produces hydro-meteorological hazard information for risk assessment of climate change.

In the program, changes in extreme events like heavy precipitation or typhoons are estimated. For this purpose, various statistical and dynamical downscaling methods have been applied. As the 1st stage of the program has already completed, we show here the results of the 1st stage of the program, and application example of some impact study researches, such as climate-analogue, echo-system, global drought distribution, and typhoon disaster change caused by climate change.

This research is supported by SOUSEI program supported by MEXT, Japan.

To increase understanding of the regional and local impacts of climate change and to support national resilience planning, the Centre for Climate Research Singapore (CCRS) embarked on the Second National Climate Change Study in collaboration with the UK Met Office (UKMO) to develop climate change scenarios for the island and the surrounding region. Owing to its small size and the convective nature of its weather patterns, generating future climate change scenarios at high spatial and temporal resolution poses fairly unique challenges to Singapore. In this talk, an overview of the project components (user engagement, GCM sub-selection and dynamical downscaling) will be covered. In the sub-selection of GCMs, a subset of models was chosen based on two criteria. Firstly, models that have poor representation of the baseline climate in the simulation of the regional monsoon systems for example were excluded. Secondly, models were chosen to span the range of future projected changes in all of the CMIP5 models as fully as possible for key variables such as temperature and rainfall. Regional downscaling was performed at 12 km in an attempt to better capture higher temporal and spatial resolutions patterns of key atmospheric variables. The realism in the simulations was evaluated against the observed distribution of the corresponding variables. Further post-processing was then carried out to correct, in part, the biases found in the results and bring them closer to the observed climate. The methodology and outcomes for sub-selection of Global Climate Models (GCMs) and the use of selected GCMs for subsequent dynamical downscaling work will be discussed in more detail.

The global climate models have difficulties in capturing the observed decrease of the Indian summer monsoon precipitation, thus limiting our understanding of the regional land surface response to monsoonal changes. This issue is investigated by performing two long-term simulation experiments, with and without anthropogenic forcing, using a variable grid global climate model having high-resolution zooming over the South Asian region. The present results indicate that anthropogenic effects have considerably influenced the recent weakening of the monsoon circulation and decline of precipitation. It is seen that the simulated increase of surface temperature over the Indian region during the post-1950s is accompanied by a significant decrease of monsoon precipitation and soil moisture. Our analysis further reveals that the land surface response to decrease of soil moisture is associated with significant reduction in evapotranspiration over the Indian land region.  A future projection, based on the representative concentration pathway 4.5 (RCP4.5) scenario of the IPCC, using the same high-resolution model indicates the possibility for detecting the summer-time soil drying signal over the Indian region during the 21^st century, in response to climate change. While these monsoon hydrological changes have profound socio-economic implications, the robustness of the high-resolution simulations provides deeper insights and enhances our understanding of the regional land surface response to the changing South Asian monsoon.

In this study, we analyzed the simulation skills of four regional climate models (RCM: RegCM4, SNU-RCM, WRF, YSU-RSM) for the temperature and precipitation over South Korea for the 30-year period(1981-2010) and projected 21st century(2021-2100) climate change, based on RCP(Representative Concentration Pathway) scenarios(4.5, 8.5). The HadGEM2-AO data provided by the National Institute of Meteorological Research were used as the lateral boundary conditions for the four RCMs with 12.5 km of horizontal resolution. In general, the four RCMs well simulated the temporal and spatial variations of temperature and precipitation compared with CRU (Climate Research Unit) data. However, the RCM’s simulation skills are clearly dependent on the climate element, seasons and geographic locations. The annual mean temperature over South Korea in the late 21st century is expected to increase by about +2.71ᵒC and +4.53ᵒC, respectively, compared with the present climate under the RCP 4.5 and 8.5 scenarios. Annual precipitation is projected to increase by about +0.40 mm/day under RCP8.5. Interannual variability (IAV) of temperature was projected to increase and decrease significantly under RCP8.5 and RCP4.5, respectively. The significant increase in IAV of temperature under RCP8.5 indicates the possibility that abnormal temperatures may increase in the late 21st century. The portion of heavy rainfalls (> 50 mm/day) was also expected to increase irrespective of scenarios, which indicates the possibility that extreme droughts and heavy rainfalls may be increased.The detailed simulation and projection results will be presented in this work.

Lightning network data, as a useful supplement to radar observations, can record lightning information (including location, intensity, and frequency) and is limited in temporal and spatial coverage. To improve the accuracy of convective system, lightning data assimilation is conducted on the basis of converting lightning data to water vapor mixing ratio via a simple smooth continues function between using 9 km gridded resolution total flash rate and simulated graupel mixing ratio as input variables. Then the water vapor mixing ratio is switched into relative humidity, which is written in a form of sounding data. Finally, the relative humidity is assimilated into the background field utilizing the three dimensional variational method in WRFDA (the weather and research forecasting model data assimilation system). The benefits of assimilating lightning data are demonstrated by a series of experiments using data from a strong convection event that occurred in Beijing, China, on 31 July 2007. For this case, the results show that there were significant improvements both in the prediction of reflectivity and precipitation (3 h and 6 h), and the improvements could be maintained for several hours. The spatial location and intensity of precipitation were pretty close to the observation.

For predictions of a heavy rainfall in a limited area, detail information of hydrometeors (i.e., cloud water, cloud ice, snow, graupel, water vapor) in a numerical model is important. Radiosonde observations can give three-dimensional structures of water vapor, but no direct observation for other hydrometeor. At the same time, the radiosonde observation is too coarse to obtain details in a small area. In this study, the brightness temperature data observed by a spaceborne microwave radiometer, or Advanced Microwave Scanning Radiometer for EOS (AMSR-E) onboard Aqua, is assimilated by using Ensemble Kalman Filter (EnKF). Making the ensemble members, cloud top heights estimated by satellite observation in an infrared channel are used to give a reasonable horizontal distribution of cloud in initial states. This procedure is also effective to reduce the range of perturbations given for the ensemble members. Experiments by different number of ensemble members are made. The experiments by 45 and 15 ensemble members show similar results in precipitation predictions. This indicates that initialization by satellite-based cloud top heights can give a good estimation in mean state among the ensemble members and reduce the number of ensemble. On the other hand, the results of assimilation is quite sensitive for the selection of members in different ranges of perturbation. In this study, the brightness temperature in 36.5 GHz is assimilated. This frequency has high sensitivity for cloud water, but useful for estimation of water vapor. An experiment without assimilation of water vapor shows overestimation of precipitation. However, results of application of EnKF assimilation for water vapor showed reducing overestimation of precipitation over the ocean.

A tropical convective-scale Numerical Weather Prediction (NWP) model, SINGV, based on the Unified Model, is being developed at CCRS in collaboration with the UK Met Office. The objective is to better address convective-scale weather forecasting up to 36 hours ahead over Singapore and its surrounding region. The model adopts a 1.5 km horizontal resolution and uses 80 levels in the vertical, with finest resolution near the ground such that there are 20 levels in the lowest 2 km, and with the model top at 38.5km. A series of downscaling experiments have been conducted for a pair of one month trial periods, which are in the South-West Monsoon and North-East monsoon seasons respectively. The simulation experiments include 1) two nesting domains to downscale from the global model analysis and forecast, 2) a single large domain to downscale the global model outputs with resolutions of 1.5km, 2.2km and 4.5km. The preliminary results indicate the single large domain with 1.5km resolution outperforms the two coarser resolution ones in terms of the Fractions Skill Score (FSS) for rainfall across various spatial scales, especially for large threshold rainfall events (e.g. >32 mm/d), which is exactly of our interest. In general the SINGV model is able to capture the diurnal cycle over this maritime continent area, and is able to predict severe afternoon thunderstorms, as well as the squalls which develop over Sumatra Island and propagate eastwards across Singapore.

In the Strategic Programs for Innovative Research (SPIRE) Field 3 state-of-the-art data assimilation methods were implemented in the K Computer, among them 4DVAR, Hybrid-4DVAR, and Hybrid-4DEnVAR. These methods are expected to represent the atmospheric state more accurately, thus improving the forecast quality, especially for severe weather phenomena like local heavy rainfall. All the above methods belong to the variational data assimilation technique which estimates the mode of the posterior distribution through minimization of a cost function. While 4DVAR and Hybrid-4DVAR use the tangent linear and adjoint models to propagate the uncertainty in time, Hybrid-4DEnVAR retrieves this information from the nonlinear forecasts of ensemble members. Both Hybrid-4DVAR and Hybrid-4DEnVAR take “errors of the day” in consideration when using the background covariance from an ensemble forecast. To provide this ensemble background covariance for the two hybrid systems, a 4D-LETKF system was run in parallel to the 4DVAR module. However, this is not a one-way interaction between 4DVAR and 4D-LETKF. The 4D-LETKF analysis is replaced by the one of 4DVAR, combining this with the ensemble analyses estimated by 4D-LETKF to propagate the system state and its uncertainty in time. Performance of three methods is shown in experiments with real observations. Other aspects like computational cost, complexity, and balance of the analyzed field will be discussed.

The background error (BE) covariance of the Numerical Weather Prediction (NWP) model is important for the variational data assimilation in short term weather forecasts. The characteristics of BE is strongly associated with local weather features. According to the experiences of local meteorologists of Singapore, the wind information is critical for the development/ turning of the weather in this region. However, unlike in mid- or high-latitudes, there is a lack of geostrophic balance between wind and mass fields in the tropics. With the aid of the gen-be utility, a tool comes with the WRF release, the BE characteristics of SINGV model (a high resolution NWP model based on the Unified Model) is investigated in the tropical area. The results of the balance among the analysis variables, as well as the data assimilation behaviour with the BE over the complex topography in tropical region will be presented.

Large scale patterns usually cannot be represented very well by high resolution forecasts generated from a limited domain. A scale-dependent blending scheme has been developed in WRF (e.g., Hsiao et al., 2014; Wang et al., 2014) for reintroducing large scale information into the WRFDA analysis. The Unified Model (UM) has been implemented in Singapore since 2013, and a similar blending scheme is firstly implemented in the Unified Model (UM) within the project of SingV.

In our experiment, smaller scale features from the 1.5 km UM forecasts are extracted and merged with the UM/ECMWF analysis (at the resolution of 17 km approximately) through the Raymond filter scheme. The blended analysis is applied in two ways:

(1) This scheme can be beneficial to the model run/cycle by using it to initialize the model. The purpose of doing this is that, with the model forward integrating, due to limited area, large scale information may be gradually degraded and consequently affect the model performance. The constraints from large scale information may be a way to partially address this problem.

(2) We also use the blending scheme to improve the desktop forecasts which are presented to end users. It could work well for island countries such as Singapore, especially for improving the representativeness of surface fields. For example, due to the lifting effects of orography, we may expect better forecasts of winds from high resolution model, but over South China Sea, lateral boundary for the limited area could block the transportations of large scale waves and thus affect the winds estimates. Therefore, in the area with complex surface, it is essential to merge the small scale features from high resolution model and large scale information from regional/global model (with bigger domain) within one chart to better represent the system evolutions.

At this conference, the preliminary results of using the UM based blending scheme over the tropical region will be presented.

Snow albedo plays a critical role in calculating the energy budget, but snow-covered albedo are still very different among the land surface models, especially over forest region. The structure and density of forest affect snow albedo, because it is hard to completely cover whole forest with snow thus canopy has the shading effect. In the Noah-MP, all vegetation types have the minimum values of leaf area index (LAI) and stem area index (SAI) in most of winter, which are too low and do not consider the vegetation type. LAI and SAI are dealt with as photosynthetic activeness; therefore, in winter, there is little vegetation effect on surface albedo with only these parameters. We found that albedo calculation without proper consideration to the vegetation effect is mainly responsible for the large positive bias in winter. Therefore, we develop a method to introduce new parameters for quantifying all leaves and stems that have an effect of vegetation structure on the winter albedo and optimize the parameters using the micro-genetic algorithm (GA). This study also shows the variation of radiation flux with the improvement of snow-covered albedo.

Lightning activity in the Eastern Mediterranean and the vicinity of Israel is prevalent in winter, mostly between November and March, in conjunction with the passage of Cyprus Lows over the warm Mediterranean Sea or with the intrusion of a tropical Red Sea trough from the south. Thunderstorm cells are found in cold-fronts and within the ensuing cloud streets, and reach 6-8 km. Most lightning activity occurs over the sea, with seasonal changes according to the dominant synoptic condition.

The multiplicity, the percentage of single stroke flashes and the geographical distribution of single vs. multiple-stroke flashes for thunderstorms in the region were derived from the Israeli Lightning Detection Network data. Results show that the percentage of single stroke flashes in Israel was 37% and the average negative multiplicity was 1.7. The fraction of +CG is found to be positively correlated with the wind shear in the layer between 0ºC- -25ºC, increasing from the ~1% in early autumn to 17±7.5% in late winter. Optical observations conducted during the ILAN sprite campaigns (Imaging of Lightning and Nocturnal Flashes) confirm that +CGs in winter thunderstorms are sometimes accompanied by Transient Luminous Events. The optical campaigns were augmented by ELF and VLF data and lightning location systems. Results of a statistical analysis of 436 sprites observed in 7 winter campaigns from 2006/7-2012/13 show a clear peak in the frequency of sprite detections, with maximum values (< 40% of events) between 00:30-02:15 LST (22:30-00:15 UT; LST=UT+2). The detection times of sprites are well-correlated with a relative increase in the fraction of +CG strokes, which exhibit maxima between 00:00-02:00 LST. This TLE climatology shows that the Eastern Mediterranean is a major sprite producer in Northern Hemisphere winter, and so Israel offers an important ground-based coverage for future space missions.

It’s difficult to understand the inside structure and developing process of thunderstorm only with conventional meteorological instruments or their networks since the horizontal extent of the storm cell is sometimes smaller than an order of 10 km while the densest ground network consists of sites located every an order of 10 km and the typical resolution of meteorological radar is 1-2 km in general. Even X-band radar, the latest system with highest spatial resolution of 25 m, is not sufficient to calculate the convergence of horizontal wind, namely, the updraft. Here we suggest a thunderstorm monitoring system consisting of the network of sferics receivers and the super dense meteorological observation system with simple and low cost plate-type sensors that can be used for measurement both of raindrop and vertical electric field change caused by cloud-to-ground lightning discharge. Horizontal location, height and charge amount of each lightning discharge are estimated successfully based on the information of quasi electric field changes measured at several observing sites. Moreover, it was found that the thunderstorm has a structure well smaller than 300 m that cannot be measured by any other ways, counting the positive and negative pulses caused by attachment of raindrop to the sensor plate. We plan to construct a new super dense observation network in the north Kanto region as a test site, Japan, where the lightning activity is most prominent in summer Japan and surrounded by our VLF receiving systems developed for detecting radio waves emitted from lightning discharge, distributing more than several tens of sensors at every 4 km or shorter, such as an order of 100 m at minimum. This kind of new type network will reveal the unknown fine structures of thunderstorms and open the door for constructing real time alert system of torrential rainfall and lightning stroke. 

Transient Luminous Events (TLEs) are triggered by inter-cloud lightning (IC) or cloud to ground (CG) lightning. The polarity of lightning and TLEs have been studied for decades. For example, more than 99% of sprites are initiated by +CG lightning, but the polarity of elve-associated lightning is still unclear now. We can explore the connection between TLEs occurrence and lightning polarity by investigating the spatial distribution of the discharge polarity. Since there is no single-source data providing all necessary information for this study, multiple sources including the magnetic field measurements at extremely low frequency (ELF) band-pass, the database of the World Wide Lightning Location Network (WWLLN) and the ISUAL TLEs list are used to explore the discharge polarities of lightning and TLEs.

The polarities of the lightning and TLE events are investigated by an efficient algorithm developed in this study. The results show that the successful detection of lightning polarity can be up to 70% within 3000 km from Taiwan. In this region, -CG lightning accounts for approximately 90% of the total polarity-resolved events. Moreover, the polarity detection rate decays along the distance, as well as the ratio of –CG lightning also drops to about 50%. This result implies that +CG discharge is averagely more energetic than –CG events.

The in-depth analysis exhibits that the observed sprite in this period are dominantly triggered by +CG events (95%), only one -CG sprite out of 48 is confirmed and this result is comparable with previous findings. It is surprisingly founded that elves are majorly initiated by –CG lightning (95%), and this result is new and confident because the accuracy of ISUAL onboard trigger time is improved by other ELF study, and the uncertainty of ISUAL event trigger time shrinks from 25 milliseconds to 3 milliseconds. In addition, it is revealed that the density of TLE is proportional to lightning occurrence in the seasonal distribution of lightning and TLEs, but the polarity of lightning seems no significant seasonal variation is observed. This result also suggests that the distribution of TLEs is mainly governed by lightning density and energy, and less correlated with polarity.

Elves in the mesosphere produced by intense lightning return strokes in the troposphere are the dominant type of transient luminous event (TLE) and predominate near 85-87 km of altitude.  The OH* Meinel band airglow is a conspicuous optical phenomenon in the same height range. By analyzing the limb-viewed images from the Imager of Sprite and Upper Atmosphere Lightning (ISUAL) onboard Formosat-2 in a wavelength range 623-754 nm, it is found that ~70% of elves are collocated in altitude with the OH* nightglow peak brightness within +/- 1 pixel at the mean altitude 85.6 +/- 4km. This collocation has not been discussed extensively and is frequently dismissed as coincidence, since the physical mechanisms for the formation of elves and the OH* Meinel band are macroscopically quite different.  Elve radiance requires the presence of free electrons, though not so many as to exclude the lightning-produced electric field by high conductivity.  These requirements make the characteristic ledge in nighttime electron density a favorable location for elves. However, the sharp increase in monatomic oxygen density in the same height range as this ledge and the OH* concentration is a hint that the OH* nightglow and the nighttime ledge in electron density are physically/chemically linked. In this study, we will discuss correlations between the key chemical species and the electron density with a focus on their diurnal and annual variations, and also the wavenumber-4 structure in OH* airglow intensity at local midnight. These comparisons of marked changes in key quantities over a narrow range of special altitudes lend support to O being the main player in linking electrical and optical behavior in elve and nightglow emission.

JEM-GLIMS mission conducts the nadir observations of lightning discharges and TLEs from the International Space Station (ISS).  The main goal of this mission is to identify the global occurrence distributions and rates of lightning discharges and TLEs, to characterize the horizontal distribution of sprites, to clarify the conditions determining the horizontal distribution of sprites. JEM-GLIMS was launched on July 21, 2012 and started the continuous observations of lightning and TLEs since November 20, 2012. For the period between November 2012 and December 2014, a total of 5,254 lightning events were detected.  In 588 events of these lightning events, optical emissions originated in TLEs were also measured by CMOS cameras and spectrophotometers. Using these JEM-GLIMS optical data, we have estimated the global occurrence distributions and rates of lightning and TLEs. It is found that most of the lightning events were detected over continental regions in the local summer hemisphere. On the other hand, TLEs including both sprites and elves occurred over not only continental regions but also oceanic regions, which is similar to the results derived from the ISUAL measurements. JEM-GLIMS succeeded in detecting sprite emissions over the active thundercloud regions. The detected sprite emissions showed clear horizontal displacement from the peak location of the parent lightning emissions. Using electric field waveform data obtained by the VHF interferometer onboard JEM-GLIMS, source locations of VHF pulses excited by the parent lightning discharges are estimated.  It is found that the source locations were located within the area of the parent lightning emissions and near the sprite emissions. At the presentation, we will show the global occurrence distributions and rates of lightning and TLEs in detail. Also, we will show detailed characteristics of the sprite distributions and their relation to the parent lightning discharges and VHF sources.

The initiation and propagation of lightning is a process of charge separation along positive and negative leaders extending in opposite directions as driven by the thunderstorm generator. The propagation of positive leader is difficult to observe due to its weak optical emission and very-high frequency (VHF) radiation. It is the critical problem of lightning physics as how to effectively observe the propagation of positive leader inside clouds and the micro-scale features of its propagation.

During the artificially triggered lightning experiment in summer of 2013, the lightning research group of Institute of Atmospheric Physics (IAP), Chinese Academy of Sciences deployed the low-frequency magnetic sensor typically used for the remote sensing of lightning and recorded the electromagnetic (EM) pulse train of more than 20 milliseconds accompanying the initial continuous current in triggered lightning. The timescale of these EM pulses is about several microseconds, and the inter-pulse interval is generally about 25 µs. The comparison with the two-dimensional (2D) observation result of high-speed camera and the short-baseline VHF lightning imaging system indicates that the EM pulse train was related to the stepping of positive leader. The analyses of these pulses suggest that the propagation of positive leader bears the similar stepping feature to the negative counterpart during its propagation, whereas the interval between individual stepping is relatively long. This implies some difference in the physical mechanism of formation and propagation for the positive and negative leaders.

The electromagnetic pulse train has been commonly observed in the triggered lightning experiment during summer of 2014, indicating that the sensitive magnetic sensor will contribute to further understand the propagation features of positive leader. At present, the lightning research group of IAP is looking for the electromagnetic pulse trains possibly associated with the stepping feature of positive leader in natural lightning.

We present results of recent measurements of daytime tweek atmospherics, which we have recorded in a favorable electromagnetic environment on the summit of La Grande Montagne (1028 m, 43.9410N, 5.4836E), Plateau d'Albion, France in 2014 and 2015. Waveforms of the vertical electric field and two horizontal components of the magnetic field have been measured by a ground-based version of the VLF/ELF analyzer with a sampling interval of 20 ms which we are preparing for the Luna-resource and Resonance spacecraft missions. The analyzer is coupled with a 10-cm spherical electric antenna located 2 m above ground and two 12-turn magnetic loop antennas of 4 m² with preamplifiers (E-W and S-N components).

The observed daytime tweek atmospherics had a clear frequency dispersion which we were able to analyze above the first ionospheric cutoff, while tweeks up to the second-order mode were identified in our data. We have done a search of model parameters for the obtained frequency dispersion as a function of time. We were able to estimate the reflection height, attenuation, and propagation distances of these unusual daytime tweek atmospherics.

We develop a numerical code to study the plasma streamer observed in the sprite events. The plasma streamer of sprites at an altitude of 40-90 km is the ionization wave propagating through the atmosphere. Just after the lightning occurs, the associated electric field could drive the development of plasma streamers in the sprite altitudes. A possible explanation is described as follows. For sources of ionizing by cosmic rays, the electrons and ions will be accelerated in the opposite directions in a background electric field. That causes the positively/negatively charged shell in both the opposite edges of the ionization source. In the positively/negatively charged shell, the local electron field plus background electric field can exceed the breakdown electric field for ionizing gases. The high electric field will ionize more electrons and ions ahead the positively/negatively charged edges. The positively/negatively charged head with a peak electric field will propagate through the atmosphere, like an ionization wave. In this presentation, we will study the chemically active particles behind the ionization wave and estimate the degree of ionization after the passage of streamers. The sprite occurring rate and detailed chemical reactions are poorly known. With the FORMOSAT-2 ISUAL data and our modeling results of the sprite streamers, we try to resolve the problem of chemical and ionization effect by sprites in the middle atmosphere.

Analysis of interannual variability of surface air temperature during boreal winter in the East Asian (EA) region from 1960 to 2009 reveals that positive temperature anomaly occurs frequently after the mid-1980s, implying the weakening of the East Asian winter monsoon (EAWM). The robust warming over the EA region in the lower and middle troposphere as well as at the surface is primarily caused by changes in atmospheric circulations over the North Pacific and Eurasian continent. It is evident that upper-level jet in its core region and 500-hPa trough over the EA region significantly weaken after the mid-1980s. The enhanced westerly and weakened northerly in the Siberian high region and to the north of the EA region, which is related to the positive phase of arctic oscillation, interfere cold advection toward the EA region. The anomalous southeasterlies over the East China Sea due to the enhanced North Pacific oscillation (NPO)-like SLP pattern lead to the anomalous warm advection over the EA region. This finding suggests that these anomalous circulations provide a more favorable condition for the weakening of EAWM. Moreover, the advection of mean temperature by anomalous wind and advection of anomalous temperature by mean wind mainly contribute to the anomalous warm advection in the EA region after the mid-1980s.

Southeast Asia often suffers from severe floods and droughts, and these disasters have caused huge socio-economical impact to the region. For example, Thailand alone lost 50 million USD due droughts in 2010 and lost another 45.7 billion USD due to floods in the following year. Thailand flood in 2011 alone has caused an acute shortage of HDD in the global market, which shows the extent of the climate influence on the interconnected global economy in the modern world. Though several studies in the past have tried to link the region’s rainfall with the dominant modes of tropical climate variations, there still exist a lot of uncertainties. 

In a recent study, it is found that El Nino/Southern Oscillation (ENSO) and the recently found ENSO Modoki influence the southern Myanmar and Thailand rainfall but only during March-May. Interestingly, from the correlation and composite analyses, it is found that the variation in the large-scale monsoon influences the local monsoon wind and thereby the rainfall anomalies over those two regions through anomalous transports of moisture in boreal summer and fall seasons in addition to the boreal spring season. Furthermore, the regional monsoon winds that influence the local rainfall is in fact connected to the basin-scale Indian Monsoon and the Western North Pacific Monsoon through the large-scale processes. 

The variations of the regional monsoons are captured by a set of newly defined indices. Moving away from the Indo-China monsoon region, the Sabah, Mindanao, and Cebu rainfall is influenced by the winter monsoon in Southeast Asia. Partial correlation analysis shows that the region is influenced by the variations of Siberian High and Maritime Continent Low, particularly during boreal winter months. Hence, both summer and winter monsoons play important roles in the rainfall variations of the Southeast Asia.

The Indian Ocean Dipole (IOD), an air-sea coupled climate mode, characterized by an aperiodic oscillation of sea surface temperature (SST) in the Indian Ocean, has considerable influence on the climates of many regions around the global, such as the floods in East African and droufghts in Indonesia and parts of Australia [Ashok et al., 2003; Saji et al., 1999; Webster et al., 1999]. IOD is measured by an index called the Dipole Mode Index (DMI), describing the difference in SST anomaly between the tropical western Indian Ocean (50°E-70°E,10°S-10°N) and the tropical south-eastern Indian Ocean (90°E-110°E,10°S-Equator). As the dominant modes in, respectively, the Pacific and Indian Oceans, the relationship between El Niño and IOD is of enormous interest. In general, it is believed that El Niño may induce IOD [Wang and Wang, 2014]. Ashok et al. [2003] found that the IOD has significant impact on the winter rainfall of western and southern Australia. Using an AGCM, they found the mechanism is that an anomalously anticyclonic circulation, due to the cold SSTA to the west of the Indonesian archipelago in a positive IOD event, is introduced at lower levels over the eastern tropical and subtropical Indian Ocean as well as much of the Australian continent.

Luo et al. [2010] demonstrated that extreme IOD has important influence of El Niño and its predictability, and vice versa. The interbasin coupling via the atmospheric Walker circulation is crucial to both El Niño and IOD evolutions and predictions.

The purpose of the present study is to examine the possible impacts of IOD events on rainfall in Singapore, and its adjacent area.

Synoptic-scale dryline formed in the pre-monsoon season (March-May) along the eastern boundary of India due to interaction between a hot deep dry air advected from the arid region of the west to the India and warm shallow moist air advected from the Bay of Bengal. The NCEP-CFSR 20-years reanalysis data of horizontal resolution 0.5°x0.5° are used for analyzing the characteristics of the dryline and its propagation. The dryline oriented from the northeast to southwest direction has the horizontal length of ~2000 m at surface. The highest dewpoint gradient is found of approximately 2⁰C/10 km and equivalent potential temperature gradient of 3⁰C/10 km across the dryline. During March the low-level moist southwesterly is initially wedged below to the northwest dry air and it gradually increases with height by the progress of the season, which in turn, helps to withdraw the seasonal dryline in June. The humidity averaged over 1000 hPa to 600 hPa increases 2.5 times of the initial values (4.5 g/kg) at the end of the pre-monsoon season.  

The diurnal variation of dryline motion is observed for the recent three years (2012 -2014) which oscillates from the eastern-most position (near coast) at noon to 300-500 km far  west at night. The dryline is retrograde back to the original position for the next noon which provides an approximate sinusoidal diurnal variation over the season. The average dryline motion is found faster (~2 m/s) during evening to midnight otherwise the motion is slow (>1 m/s).

The extreme dry spells in Taiwan show clear decadal-scale variations. The overall dry spell extremity became more sever after mid-1950s. The enhanced extremity showed clear seasonal dependence in different decades. In 1960s the enhanced extremity occurred during the earlier half of the summer months from May to July.  During the decade from mid-1960s to mid-1970s the enhanced extremity occurred during the late summer and early autumn months from August to October, while it occurred in the mid-summer months of July and August during the 1980s. After 1990 the enhanced dry extremity was seen during the winter half year from October to March, the dry season in Taiwan. The enhanced dry extremity during the winter months increases the general concern that the dry season will become drier and the wet season will become wetter under the global warming trend, although the decadal-scale variations of the extremity during the wet season did not show concurrent relationship with the dry extremity. This paper shows that the summer dry spell extremes can be explained by more intense and westward extended western Pacific subtropical high (WPSH). The winter dry spell extremes can be explained by weaker East Asian winter monsoon. It is worth noting that the most recent and still going dry situation in Taiwan, which is the worst in 10 years, occurred concurrently with more westward extended WPSH during the 2014 summer and autumn and weaker than normal meridional mode of the East Asian winter monsoon during the 2014/15 winter. Understanding how extreme dry spells can be influenced by the East Asian monsoon variations is of critical importance to advance the climate change adaptation research in Taiwan.

This study finds that the winter (December - February) decadal variability of northerly wind in East Asian Winter Monsoon (EAWM) over the northern part of East Asia coast was influenced by the forcing from the middle latitudes during 1950s to 2000s related to the Pacific Decadal Oscillation (PDO). It is proposed that this fluctuation of northerly wind in winter is effected by the position of the Aleutian low, which has been known to be associated with the phase change of the PDO. The mechanism accounts for the change of the westward sea-level pressure (SLP) gradient along the Northeast Asia coast, it’s effected by the movement of Aleutian low. Moreover, the location of the Aleutian low is associated with the phases of the PDO. As the Aleutian low is influenced by the negative phase of the PDO and moves westward, the SLP gradient between the Siberian high and the Aleutian low and the northerly wind at 850hPa will increase. This can be associated with the increased EAWM near the Sakhalin Island to Hokkaido.

The two dominant modes of boreal summer intraseasonal oscillation (BSISO) that influence onset and retreat of the Asian summer monsoon system and strongly modulate extreme rainfall events over Asia have been identified. One is the canonical northward propagating mode (BSISO1) with quasi-oscillating periods of 30-60 days and strong variance throughout the May to October period often in conjunction with the eastward Madden-Julian Oscillation (MJO). The other is the northwestward propagating mode (BSISO2) with periods of 10-30 days during primarily the pre-monsoon and monsoon-onset season. This study investigates how the two modes are different from each other dynamically and thermodynamically. The BSISO1 circulation cells are more Rossby wave like with a northwest to southeast slope, whereas the circulation associated with the BSISO2 is more elongated and front-like with a southwest to northeast slope. The associated rainfall anomalies over the Indian monsoon and Western North Pacific region are out-of-phase in the BSISO1 but rather in-phase in the BSISO2. For both modes, moisture convergence plays more important role in development and propagation of BSISO than moisture advection but with different energetics. In the BSISO1, convective instability characterized by the low-level warm advection of moisture-laden air and upper-level cold advection of dry air is dominant but the baroclinic instability with warm-air rising and cold-air sinking tends to control the BSISO2. Phase-latitude composite analysis over the western North Pacific-East Asian region further indicates differences in intensification and propagation mechanism between the two modes.

We have investigated long-term variation of Japanese summer climate using observation records at 60 stations nationwide since 1901. Observational data show that surface air temperature has increased at a lower rate for northeastern Japan in July and August compared to those in other regions of Japan and in other seasons. Precipitation shows positive trends on the Japan Sea side of Japan in July and August, which is different from the Pacific side of Japan, where the trends are unclear or negative in this season. A detailed analysis for Baiu rainfall, a part of early summer rainfall over East Asia, reveals decreased trends in the early stage (1 June to 20 June), contrasting to significantly increased trends in the late stage (11 July to 31 July) particularly on the Japan Sea side of Japan, indicating a delay of seasonal progress of the Baiu season. Thus, the spatial distribution of the observed trends is strongly affected by topography, especially for precipitation. The observed trends in July may result from southward contraction and westward expansion of the Pacific high and also from intensification of the Okhotsk high, although the latter is unclear in observation records in Japan. AMIP-type simulations using a 60-km mesh AGCM for the period of 1901-2005 can reproduce the observed contrast in the precipitation trend between the Japan Sea side and the Pacific side, which seems to result from enhanced westerly moisture flow and its interaction with topography.

Recent energetics analysis of individual tropical convective events, suggested that a large portion of the energy available to drive weather global weather events resides in the tropics in the form of geopotential energy cached in the Upper Troposphere-Lower Stratosphere of the tropics. We have termed this energy “Jet Available Potential Energy”, because it represents the excess potential energy placed in the UTLS that is available to be converted into kinetic energy and so form a subtropical jet.  Therefore we term this UTLS energy cache, associated with the elevated tropical tropopause, the  “tropical JAPE bubble”.  It resides near or above the level of zero net radiation and at isentropic layers high enough to have no direct contact with the surface. 

Our analysis suggests that the trapped high entropy and high angular momentum of the tropical JAPE bubble must eventually move into the extratropics where radiative processes can complete the energy cycle by bleeding off the energy to space and surface friction.  Analysis of global prediction models, suggests that the energy is bled off of the tropical bubble through occasional poleward surges, usually in the winter hemisphere, resulting in poleward directed  “tropical plumes” , of JAPE, reflected often as atmospheric rivers at the surface.  Such plumes are typically associated with the poleward arching of the subtropical jet and its connection into the Rossby wave stream where energy of the tropical bubble is injected into the extratropics. In the absence of sufficiently frequent tropical plumes of this type leads to infrequent extreme surges.  They appear to be set off by the build up of tropical energy held in the bubble that weaknesses in the inertial walls containing the bubble can be breached.  Such events seem to be a mechanism causing at least some of the unusual weather events being experienced globally.

The South Asian summer monsoon contributes to more than ¾th of the annual rainfall, and dictates the socio-economic livelihood in the subcontinent. However, there are large uncertainties looming over the status and fate of the monsoon, with several studies debating on whether the summer monsoon is weakening or strengthening in a changing climate. The observations and climate models have suggested that anthropogenic warming in the past century have increased the moisture availability, as well as the land-sea thermal contrast in the tropics, favoring an increase in tropical rainfall. However in the current study, we observe that the summer monsoon rainfall during 1901-2012 shows a significant weakening trend over the South Asian subcontinent, extending from Pakistan through central India to Bangladesh. We notice that the subcontinent experienced a relatively subdued warming during this period. In contrast, the tropical Indian Ocean experienced a nearly monotonic warming, at a rate faster than the other tropical oceans. Using multiple observed datasets and model simulations, we demonstrate that the subdued warming of the subcontinent along with the enhanced Indian Ocean warming results in a reduced land-sea thermal contrast, which extends from the surface to the upper troposphere, weakening the summer monsoon circulation and rainfall.

The present study shows the Madden and Julian Oscillation (MJO) appearing a general circulation model (GCM) with a full representation of cloud microphysics at 50 km horizontal resolution and the MJOs are compared with those of GCMs with conventional convective parameterizations. The coarse-resolution GCM requires modifications of several parameters of cloud microphysics and an additional vertical mixing process in the lower troposphere. The newly developed GCM, which includes explicit cloud microphysics and additional vertical mixing, produces more heavy precipitation and less light precipitation than conventional GCMs, thus simulating a precipitation frequency that is closer to the observation, and the model simulates the MJO reasonably well. The dynamical structures and mechanisms associated with the MJO simulated by the present GCM will be presented in the talk.

This work provides a new perspective on the major factors controlling the East Asian summer monsoon (EASM) in July, and a promising physical–statistical forecasting of the EASM ahead of summer. Dominant modes of the EASM are revealed from the variability of large-scale air masses discerned by equivalent potential temperature, and are found to be dynamically connected with the anomalous sea surface temperatures (SSTs) over the three major oceans of the world and their counterparts of prevailing atmospheric oscillation or teleconnection patterns. Precipitation over Northeast Asia (NEA) during July is enhanced by the tropical central Indian Ocean warming and central Pacific El Niño-related SST warming, the northwestern Pacific cooling off the coast of NEA, and the North Atlantic Ocean warming. Using these factors and data from the preceding spring seasons, we build a multiple linear regression model for seasonal forecasting. The cross-validated correlation skill predicted for the period 1994 to 2012 is up to 0.84, which far exceeds the skill level of contemporary climate models.

Seasonal prediction skills of the intensity of the Northeast Asian summer monsoon are relatively low using state-of-the-art general circulation models. It is introduced that seasonal-mean precipitation in Northeast Asian region are strongly correlated to western North Pacific (WNP) subtropical High variability in summertime. The relationship between Northeast Asian and WNP summer monsoons keeps with a high correlation coefficient over the entire period. Many general circulation models fail to predict the Northeast Asian summer precipitation but can well capture and predict interannual variability of the western North Pacific subtropical High, which dominates climate anomalies in the western North Pacific-East Asian region. Besides, interannual variability of the Northeast Asian summer monsoon is highly correlated to the WNP subtropical High variability. Based on this relationship, we suggest a seasonal prediction model using several coupled general circulation models and canonical correlation analysis for Northeast Asian summer precipitation anomalies in this study. This methodology provides with considerable benefit of seasonal prediction skills for Northeast Asian summer rainfall anomalies.

Climatological seasonal changes of rainfall and lower tropospheric circulation in the Philippines were analyzed by utilizing 5-day mean TRMM 3B-42 and station rainfall data provided by PAGASA, and ERA-Interim wind data for the period 1998-2013. In particular, climatological onset and withdrawal processes of the southwest monsoon were investigated.

It was found that the onset of southwest monsoon occurs abruptly in mid-May. Interestingly, it starts from the north in the Philippines both in rainfall and wind, which is a peculiar monsoon seasonal change feature in this region. After the onset, anti-cyclonic flow from the Pacific high is predominant, and it changes into cyclonic flow in mid-June. Easterlies still remain in the south until early July, afterwards SW monsoon covers the whole country and enhances from late July.

Southwest monsoon begins to retreat from the north in mid-September, and fully retreated from the southern tip of the Philippines in late October. 

These seasonal change processes are closely related with a stepwise seasonal change processes of the Western North Pacific Monsoon (WNPM), and even in phase with the East Asian monsoon seasonal changes.

Acknowledgment: Part of this study was supported by The JSPS KAKENHI, The GRENE program of the MEXT, and the Asian Human Resource Fund of the Tokyo Metropolitan Government.

A recent renewed emphasis has been placed on understanding the role of the mesoscale meridionally-elongated mountain ranges in South and Southeast Asia in the monsoon scale. By simulations using a global climate model with and without the Arakan and Annamese Mountains in the Indochina Peninsula, individually and also combined, we found that the presence of the Arakan/Annamese Mountains substantially influences the monsoon onset in South Asia/East Asia-Western North Pacific (EA-WNP); nevertheless, the Arakan Mountains affect the large-scale monsoon opposite to that of the Annamese Mountains.

In the developing phase of overall summer Asian monsoon around late May, the presence of the two mountain ranges in the model markedly improves the simulation of the upper-tropospheric high over the Indochina Peninsula, the midlatitude jet stream, and precipitation. Compared with the control simulation without the two mountain ranges, the atmospheric response to the mountains shows forced midlatitude wavy pattern in the upper troposphere, which contributes to the better simulated Meiyu-Baiu precipitation.

The presence of the two mountains also better simulate the EA-WNP monsoon transition in late July, which can be characterized by the weakened (also northward shift) upper-level jet stream and the deepened lower-level monsoon trough. The enhanced monsoon activity over the subtropical WNP, as a result of the presence of the mountains, results in higher sea surface temperature over the region. The potential mechanistic connection of the EA-WNP monsoonal subseasonality to the narrow mountain ranges is being investigated.