谷口 博

By comparison with satellite and field observations, the comprehensive performance and potential utility of near real-time forecasts using Nonhydrostatic Icosahedral Atmospheric Model (NICAM) are demonstrated by exploiting the Cooperative Indian Ocean Experiment on Intraseasonal Variability in the Year 2011 (CINDY2011)/Dynamics of the Madden Julian Oscillation (DYNAMO) campaign. A week-long forecast was run each day using a regionally stretched version of NICAM, with the finest mesh size of 14 km over the tropical Indian Ocean (10), throughout the intensive observation period (IOP). The simulated precipitation time series fairly represented the evolution and propagation of the observed Madden-Julian Oscillation (MJO) events, although a 30 % overprediction of precipitation over the IO domain (60-90 degrees E, 10 degrees S-10 degrees N) was found on average. Frequencies of strong (> 40 mm day(-1)) precipitation were over predicted, while those of weak precipitation were underpredicted against satellite observations. Compared with the field observations at Can island, the biases in precipitation frequency were less obvious, whereas the growth of lower to middle tropospheric dry (similar to 1 g kg(-1)) and warm (similar to 1 K) biases were found. Despite these mean biases, temporal variations of the moisture and zonal wind profiles including the MJO events were reasonably simulated. Using the forecast data the moisture and energy budgets during the IOP were investigated. The diagnosis using the 7-day-mean fields captured the observed features of the MJO events. Meanwhile, significant upward transport of moisture by the grid-resolved high-frequency variability was detected throughout the IOP. The relationship between these high-frequency effects and the simulated WO or mean biases is also discussed.

The present study assesses the forecast skill of the Madden-Julian Oscillation (MJO) observed during the period of DYNAMO (Dynamics of the MJO)/CINDY (Cooperative Indian Ocean Experiment on Intraseasonal Variability in Year 2011) field campaign in the GFS (NCEP Global Forecast System), CFSv2 (NCEP Climate Forecast System version 2) and UH (University of Hawaii) models, and revealed their strength and weakness in forecasting initiation and propagation of the MJO. Overall, the models forecast better the successive MJO which follows the preceding event than that with no preceding event (primary MJO). The common modeling problems include too slow eastward propagation, the Maritime Continent barrier and weak intensity. The forecasting skills of MJO major modes reach 13, 25 and 28 days, respectively, in the GFS atmosphere-only model, the CFSv2 and UH coupled models. An equal-weighted multi-model ensemble with the CFSv2 and UH models reaches 36 days. Air-sea coupling plays an important role for initiation and propagation of the MJO and largely accounts for the skill difference between the GFS and CFSv2. A series of forecasting experiments by forcing UH model with persistent, forecasted and observed daily SST further demonstrate that: (1) air-sea coupling extends MJO skill by about 1 week; (2) atmosphere-only forecasts driven by forecasted daily SST have a similar skill as the coupled forecasts, which suggests that if the high-resolution GFS is forced with CFSv2 forecasted daily SST, its forecast skill can be much higher than its current level as forced with persistent SST; (3) atmosphere-only forecasts driven by observed daily SST reaches beyond 40 days. It is also found that the MJO-TC (Tropical Cyclone) interactions have been much better represented in the UH and CFSv2 models than that in the GFS model. Both the CFSv2 and UH coupled models reasonably well capture the development of westerly wind bursts associated with November 2011 MJO and the cyclogenesis of TC05A in the Indian Ocean with a lead time of 2 weeks. However, the high-resolution GFS atmosphere-only model fails to reproduce the November MJO and the genesis of TC05A at 2 weeks' lead. This result highlights the necessity to get MJO right in order to ensure skillful extended-range TC forecasting.

REVOLUTIONIZING CLIMATE MODELING WITH PROJECT ATHENA: A MULTI-INSTITUTIONAL, INTERNATIONAL COLLABORATION The importance of using dedicated high-end computing resources to enable high spatial resolution in global climate models and advance knowledge of the climate system has been evaluated in an international collaboration called Project Athena. Inspired by the World Modeling Summit of 2008 and made possible by the availability of dedicated high-end computing resources provided by the National Science Foundation from October 2009 through March 2010, Project Athena demonstrated the sensitivity of climate simulations to spatial resolution and to the representation of subgrid-scale processes with horizontal resolutions up to 10 times higher than contemporary climate models. While many aspects of the mean climate were found to be reassuringly similar, beyond a suggested minimum resolution, the magnitudes and structure of regional effects can differ substantially. Project Athena served as a pilot project to demonstrate, that an effective international collaboration can be formed to efficiently exploit dedicated supercomputing resources. The outcomes to date suggest that, in addition to substantial and dedicated computing resources, future climate modeling and prediction require a substantial research effort to efficiently explore the fidelity of climate models when explicitly resolving important atmospheric and oceanic processes. (Page 231)

Project Athena is an international collaboration testing the efficacy of high-resolution global climate models. We compare results from 7-km mesh experiments of the Nonhydrostatic Icosahedral Atmospheric Model (NICAM) and 10-km mesh experiments of the Integrated Forecast System (IFS), focusing on the Intra-Seasonal Oscillation (ISO) and its relationship with tropical cyclones (TC) among the boreal summer period (21 May-31 Aug) of 8 years (2001-2002, 2004-2009). In the first month of simulation, both models capture the intra-seasonal oscillatory behavior of the Indian monsoon similar to the observed boreal summer ISO in approximately half of the 8-year samples. The IFS simulates the NW-SE-oriented rainband and the westerly location better, while NICAM marginally reproduces mesoscale organized convective systems and better simulates the northward migration of the westerly peak and precipitation, particularly in 2006. The reproducibility of the evolution of MJO depends on the given year; IFS simulates the MJO signal well for 2002, while NICAM simulates it well for 2006. An empirical orthogonal function analysis shows that both models statistically reproduce MJO signals similar to observations, with slightly better phase speed reproduced by NICAM. Stronger TCs are simulated in NICAM than in IFS, and NICAM shows a wind-pressure relation for TCs closer to observations. TC cyclogenesis is active during MJO phases 3 and 4 in NICAM as in observations. The results show the potential of high-resolution global atmospheric models in reproducing some aspects of the relationship between MJO and TCs and the statistical behavior of TCs.

The environmental field of tropical cyclogenesis over the Bay of Bengal is analyzed for the extended summer monsoon season (approximately May-November) using best-track and reanalysis data. Genesis potential index (GPI) is used to assess four possible environmental factors responsible for tropical cyclogenesis: lower-tropospheric absolute vorticity, vertical shear, potential intensity, and midtropospheric relative humidity. The climatological cyclogenesis is active within high GPI in the premonsoon (similar to May) and postmonsoon seasons (approximately October-November), which is attributed to weak vertical shear. The genesis of intense tropical cyclone is suppressed within the low GPI in the mature monsoon (approximately June-September), which is due to the strong vertical shear. In addition to the climatological seasonal transition, the authors' composite analysis based on tropical cyclogenesis identified a high GPI signal moving northward with a periodicity of approximately 30-40 days, which is associated with boreal summer intraseasonal oscillation (BSISO). In a composite analysis based on the BSISO phase, the active cyclogenesis occurs in the high GPI phase of BSISO. It is revealed that the high GPI of BSISO is attributed to high relative humidity and large absolute vorticity. Furthermore, in the mature monsoon season, when the vertical shear is climatologically strong, tropical cyclogenesis particularly favors the phase of BSISO that reduces vertical shear effectively. Thus, the combination of seasonal and intraseasonal effects is important for the tropical cyclogenesis, rather than the independent effects.

Tropical cyclone Nargis, generated over the Bay of Bengal in late April 2008, caused catastrophic destruction after making landfall in Myanmar. Here, the large-scale environment of cyclogenesis was investigated using re-analysis datasets and a cloud-system resolving model. The reanalysis datasets showed that a westerly wind axis over the Bay of Bengal shifted northward from mid-April to early May. This shift is attributed to a seasonal transition of the Asian summer monsoon and a boreal summer intraseasonal oscillation. The timing of this environmental modulation was consistent with the climatologically early tropical cyclogenesis over the Bay of Bengal. This period was also characterized by high genesis potential, which is an empirical index of the environmental field favorable for tropical cyclogenesis. An analysis of genesis potential showed that the genesis of Nargis was associated with reduced vertical shear and increased lower-tropospheric vorticity.A cloud-system-resolving model successfully reproduced the high probability of tropical cyclogenesis during the observed period of cyclogenesis in late April. The model also simulated the large-scale environment including the northward shift of the westerly wind axis, although the precise location of tropical cyclogenesis was sensitive to initial conditions in the model. The anomaly of sea surface temperature in 2008 had little influence on the simulated probability of tropical cyclogenesis. Therefore, a cloud-system-resolving atmospheric model even without ocean feedback is a promising tool for predicting the probability of tropical cyclogenesis over the Bay of Bengal around the onset of the Asian summer monsoon, which is a favorable environment for tropical cyclogenesis.

An ensemble simulation of cyclone Nargis was performed using the Non-hydrostatic ICosahedral Atmospheric Model (NICAM) at 14-km mesh size in order to examine the effect on cyclogenesis of disturbances associated with intraseasonal oscillations. An analysis of observational data reveals that cyclone Nargis formed during the northward propagation of low-level zonal wind, associated with active cloud areas and precipitation from the equator to 20 degrees N in the Bay of Bengal, when the active convective region associated with the Madden-Julian Oscillation (MJO) passed through the bay and then resided over the Maritime continent. The northward migration of low-level zonal wind, outgoing longwave radiation (OLR), and precipitation are successfully simulated in the ensemble results. Each simulated tropical cyclone (TC) genesis also occurs with northward migration and with a timing such that the active convective region associated with the MJO resides over the east side of the Maritime continent. The incipient disturbances that contributed to the initiation of cyclone Nargis formed during the period when the westerly wind burst passed through the Bay of Bengal after the monsoon onset and developed to TCs in the ensemble simulation. However, for an ensemble member for which northward migration as a monsoon onset is not simulated, no TC is formed in the Bay of Bengal. It is also found that the effect of the easterlies across the northern part of the Malay Peninsula is important for TC formation in our simulation.

We have been challenging the simulation of tropical cyclone (TC) geneses using a global/regional cloud-system resolving model (GCSRM), non-hydrostatic ICosahedral-grid atmospheric model (NICAM) (Satoh et al. 2008). A GCSRM has the advantages that it can deal with the organization of meso-scale cloud systems into a tropical cyclone during the cyclogenesis process, and that it can cover the long-distance movement of TC sources such as tropical waves. Using a GCSRM, our goal is to reveal the relation between predictabilities and mechanisms of TC geneses. In the previous studies, two important perspectives have been suggested related to TC geneses: environments and sources. The large-scale environments explain probability distribution of TC geneses including climatological seasonal change, interannual variability like ENSO, and intraseasonal oscillation such as Madden-Julian Oscillation. On the other hand, the precise timings of TC geneses seem to be triggered by synoptic-scale sources such as tropical waves, extratropical disturbances, and energy dispersions from neighboring TCs. Recently we have demonstrated that the 14-km-grid GCSRM can predict the timing of Typhoon 21st in 2006 with the lead time of more than 3 days, which was controlled by the westward propagating wave over the North Pacific. In 2008, Cyclone Nargis caused terrible disaster in Myanmar we examined the predictability of Nargis genesis using the NICAM model. © 2010 Springer Science+Business Media B.V.

This study discloses detailed Madden-Julian oscillation (MJO) characteristics in the two 30-day integrations of the global cloud-system-resolving Nonhydrostatic Icosahedral Atmospheric Model (NICAM) using the all-season real-time multivariate MJO index of Wheeler and Hendon. The model anomaly is derived by excluding the observed climatology because the simulation is sufficiently realistic. Results show that the MJO has a realistic evolution in amplitude pattern, geographical locations, eastward propagation, and baroclinic- and westward-tilted structures. In the central Indian Ocean, convection develops with the low-level easterly wind anomaly then matures where the low-level easterly and westerly anomalies meet. Anomalous moisture tilts slightly with height. In contrast, over the western Pacific, the convection grows with a low-level westerly anomaly. Moisture fluctuations, leading convection in eastward propagation, tilt clearly westward with height. The frictional moisture convergence mechanism operates to maintain the MJO. Such success can be attributed to the explicit representation of the interactions between convection and large-scale circulations. The simulated event, however, grows faster in phases 2 and 3, and peaks with 30% higher amplitude than that observed, although the 7-km version shows slight improvement. The fast-growth phases are induced by the fast-growing low-level convergence in the Indian Ocean and the strongly biased ITCZ in the west Pacific when the model undergoes a spinup. The simulated OLR has a substantial bias in the tropics. Possible solutions to the deficiencies are discussed.

Unstable modes of a linear shear flow in shallow water on an equatorial β-plane are obtained over a wide range of a non-dimensional parameter corresponding to Lamb parameter, and are interpreted physically by the use of the concept of resonance between neutral waves. The results indicate that there exist two resonating types for most unstable modes: (i) the resonance of equatorial Kelvin modes and continuous modes, (ii) the resonance of equatorial Kelvin modes and westward mixed Rossby-gravity modes. Examinations of dispersion curves suggests that unstable modes of type (ii) mode with zonally asymmetric structure is the same kind of instability as so-called zonally symmetric inertial unstable modes.

We investigate the linear stability of linear shear flows in order to consider the selection mechanism of the zonal wavenumber of disturbances. The linearized non-dimensional primitive equations on the equatorial beta-plane are solved. Values of Lamb's parameter (E) are 0.0031622 ∿ 31622777. The results indicate that unstable modes appear with E greater than 0.1,and that symmetric modes are most unstable for E greater than 16.