Representing Convective Organization in Global Atmospheric Models
The central objective of the Year of Tropical Convection (YoTC)is to quantify the effects of convective organization on the large-scale atmospheric circulation. Tropical convection evolves upscale to generate coherent motion systems (organized convection) up to thousands of times the size of cumulus clouds. Such multi-scale organization involves fundamental aspects such as convection-wave interaction and the upscale cascades of mass, energy and momentum associated with the nonlinear properties of precipitating convection. However, convective organization is not represented by contemporary convective parameterizations, as demonstrated by field-campaigns over three decades ago., and this generates uncertainties in global predictions The representation of multiscale convective organization, such as in the Madden-Julian Oscillation (MJO) and convectively-coupled equatorial waves, is the subject of this talk from the perspective of parameterized global models, explicit cloud-system resolving models, and dynamical models.
A Simple Dynamical Model with Features of Convective Momentum Transport
Convective momentum transport (CMT) plays a central role in interactions across multiple space and time scales. However, due to the multiscale nature of CMT, quantifying and parameterizing its effects is often a challenge. Here a simple dynamic model with features of CMT is systematically derived and studied. The model includes interactions between a large scale zonal mean flow and convectively coupled gravity waves, and convection is parameterized using a multicloud model. The moist convective wave--mean flow interactions shown here have several interesting features that distinguish them from other classical wave--mean flow settings. First an intraseasonal oscillation of the mean flow and convectively coupled waves (CCWs) is described. The mean flow oscillates due to both upscale and downscale CMT, and the CCWs weaken, change their propagation direction, and strengthen as the mean flow oscillates. The basic mechanisms of this oscillation are corroborated by linear stability theory with different mean flow background states. Another case is set up to imitate the westerly wind burst phase of the Madden--Julian Oscillation (MJO) in the simplified dynamic model. In this case, CMT first accelerates the zonal jet with strongest westerly wind aloft, and then there is deceleration of the winds due to CMT; this occurs on an intraseasonal time scale and is in qualitative agreement with actual observations of the MJO. Also, in this case, a multiscale envelope of convection propagates westward with smaller scale convection propagating eastward within the envelope. The simplified dynamic model is able to produce this variety of behavior even though it has only a single horizontal direction and no Coriolis effect.
Cascade: Investigating Multiscale Tropical Convective Organization Using High-Resolution Numerical Simulations Over Large Domains.
The Cascade project, a NERC-funded collaborative effort among several universities and research organizations in the UK, seeks to understand the organization and interaction of tropical convection at many scales, from individual convective plumes to mesoscale clusters, easterly waves, and the Madden-Julian Oscillation (MJO). High-resolution (up to 1-km grid length) simulations of the non-hydrostatic UK Met Office Unified Model (UM) are used over domains several thousand km across. One area of focus is African easterly waves in combination with the African Monsoon Multidisciplinary Analyses (AMMA) project. The Indo-Pacific warm pool is another main focus, and equatorial channel experiments over this domain will allow for detailed investigation of possible mechanisms leading to the MJO and other large-scale convective organization. Preliminary results for a test case of an active MJO event in the warm pool are shown, with the hope that modeling and analysis methods developed for this case can be applied to a similar event studied as part of the Year of Tropical Convection (YOTC).
Large Domain Cloud Resolving Simulations of Equatorial Tropical Convection: Influence of Sea Surface Temperature on Convective Organization
It has long been recognized that organized deep convective systems play a key role in regulating the large scale circulations and thermal structure of the atmosphere in tropical oceanic regions. The interaction between convection and radiation in the tropics has been described in studies of the atmosphere at radiative convective equilibrium. Cloud system resolving models have served as a useful tool for the study of tropical convection as they are capable of realistically representing mesoscale convective organization. When a large horizontal domain is employed, CRMs can be used to examine the interaction between convection, radiation, and the large-scale overturning circulation. In this paper, we examine the tropical environment at radiative convective equilibrium using the Colorado State University Regional Atmospheric Modeling System run on a large three dimensional domain. Three separate simulations are conducted, in which sea surface temperature is set to progressively higher (constant) values. At equilibrium, each simulation exhibits coherent regions of organized convection separated by broad areas of clear air and descent. Increases in sea surface temperature are associated with notable changes in the model's thermodynamic state, as well as in the mode of organization of convection within regions of ascent. Specifically, with an increase in sea surface temperature, the mean precipitation rate and total amount increase, as does the fraction of the domain with precipitation rates greater than 100 mm/day. The magnitude of both the upward and downward mean vertical velocity increase, as does the fraction of the domain containing upward vertical velocities greater than 5 m/s. Application of a convective/stratiform partitioning scheme reveals increasing convective precipitation fraction with increasing SST. Changes to vertical velocity and precipitation are consistent with an increase in saturation vapor pressure and increase in mean convective available potential energy. In addition to changes in total vertical mass flux and precipitation, there are surprising changes in the mode of convective organization in the simulations that indicate the potential for an abrupt transition in the character of organized convection with increasing SSTs. The details of the nonlinear interaction between changes to the thermodynamic state and convective dynamics will be examined, and we will suggest implications for the response of convective organization to a warming climate.
On the importance of atmospheric and oceanic initial conditions for forecasting the MJO
Successful forecast of the MJO and its effects is paramount to subseasonal prediction. In contrast to weather and seasonal forecast which rely mainly on the quality of atmospheric and oceanic initial conditions respectively, subseasonal forecast needs good initial conditions for both the ocean and atmospheric components. In this paper we quantify the relative importance of these initial conditions by conducting a series of retrospective forecast with the NCEP Climate Forecasting System (CFS). Two different sets of atmospheric and oceanic initial conditions are used: Reanalysis-2 versus the operational NCEP analysis (GDAS) for the atmosphere; the operational NCEP ocean analysis (GODAS) versus an experimental ocean analysis in which intraseasonal modes of variability are injected. We will show that the most significant impact comes from atmospheric initial conditions and we will discuss reasons for this result.
The Madden Julian Oscillation and convectively coupled waves in NOGAPS and COAMPS®
NOGAPS ensemble runs for May – November 2007 are analyzed to evaluate the prediction of the Madden Julian Oscillation (MJO) and convectively coupled Kelvin waves. The dependence of predictability on the phase of the MJO, atmospheric circulation and diurnal SST cycle is evaluated. The results indicate reasonable forecasts of MJO propagation over the Indian Ocean, but damping of the signal in the vicinity of the Maritime continent. A mesoscale model (COAMPS) is used to examine causes for the reduced MJO signal over the Maritime Continent for the intense, coupled Kelvin wave propagating through the region in June 2006. COAMPS exhibits a similar deficiency as NOGAPS in representing the rainfall anomaly associated with the wave in the vicinity of the island of Sumatra. Results show a considerable sensitivity of this rainfall to the convective parameterization used. Tests also suggest that both NOGAPS and COAMPS may be hindered in this area by initialization/model spin-up problems in the region between Sumatra and Borneo.
Interaction of Deep and Shallow Convection in GCM Simulation of Madden-Julian Oscillation
Observations indicate that shallow convection is an important part of the Madden-Julian Oscillation (MJO). It appears ahead of the deep convection, providing moisture for the lower troposphere in preparation for deep convection. How well is this feature simulated in global climate models? The NCAR CAM3 has very weak MJO, with little shallow convection ahead of the deep convection. On the other hand, when a modified convection scheme is used, the MJO is simulated reasonably well in the CAM3. Furthermore, abundant shallow convection develops ahead of the deep convection in the MJO cycle. How important is this shallow convection and what roles does it play in the simulated MJO evolution? In this study, we will examine the interaction between deep and shallow convection using the CAM3 and the revised Zhang-McFarlane convection scheme. Sensitivity tests will be carried out to examine various aspects of shallow convection and its interaction with deep convection in the MJO. Details will be presented at the meeting.
Diagnostic evaluation of a Global Cloud-Resolving Model simulation of a Madden-Julian Oscillation event
A quantitative estimate of reproducibility of a Madden-Julian Oscillation (MJO) event that occurred in December 2006 and January 2007 is examined using the result of a Global Cloud-Resolving Model (NICAM) simulation by Miura et al. (2007) The main aim here is to advocate and demonstrate a forecast skill and reproducibility of tropical atmospheric circulation in boreal winter based on the result of simulation by directly calculating deep convection and meso-scale circulations without cumulus parameterization. Three types of horizontal mesh size are used in our experiments. The 3.5-km grid run covered 1 week, whereas the 7-km and 14-km grid runs covered 30 days. The initial atmospheric conditions were generated by linear interpolation from the National Centers for Environmental Prediction (NCEP) Global Tropospheric Analyses. In the numerical integration, mean sea surface temperature is prescribed by an weekly Reynolds- SST. The RMSEs of basic quantities in the background of a MJO event; temperature, height, and velocity field at 250-, 500-, and 850-hPa, are small in equatorial region for all mesh size. Its errors of temperature and height are smaller at upper layer while velocity error is smaller at lower layer. It is also found that the phase and amplitude of MJO are good-reproducible by using EOF modes of velocity potential at 200-hPa. The day of its maximum amplitude is predicted within 1-day error for 7-km grid run. An ensemble prediction of MJO is a next way in our GCRM.
Simulation and Prediction of Tropical Intraseasonal Variability with Contemporary General Circulation Model
Tropical Intra-Seasonal Variability (TISV) is a fundamental mode of tropical climate. The associated intraseasonal wet and dry spells strongly modulate the weather systems (e.g., TC), thus the socio-economic activities (e.g., agriculture, water management et al.) around the globe. To develop a capability in forecasting TISV with lead time beyond two weeks is extremely desirable. Unfortunately, many state-of-the-art general circulation models (GCMs) still have various problems to reasonably simulate TISV. Under real forecast context (e.g., Seo et al., 2005), the predictability of TISV is only about a week by simply extending conventional weather forecast with longer integration. This study aims to address two relevant questions: 1) what are the critical pieces of model physics for the realistic simulation of TISV that have been missed or misrepresented in many contemporary GCMs? 2) In what degree is the TISV predictability affected by different settings of initial and boundary conditions? To address the first question, a suite of sensitivity experiments has been carried out under a weather forecast mode and with three 20-year free integrations with ECHAM-4 and a coupled version. It was found that a robust TISV can be sustained in the model only when the model produces a significant proportion („d 30%) of stratiform rainfall for both the forecast experiments and long-term free integrations. When the stratiform rainfall proportion becomes small, the tropical rainfall in the model is dominated by high-frequency disturbances with neither eastward propagating nor northward-propagating TISV being sustained. This result suggests that the representation of stratiform rainfall and its connections with convective component in contemporary GCMs is probably a critical issue needed to be seriously reconsidered, in order to have overall success in the simulation and prediction of TISV. To address the second question, a series of TISV forecast experiments has been conducted under different initial and boundary conditions. Their impacts on the TISV forecast skill is under examination and will be reported in the coming meeting.
MJO Structure in the Superparameterized CAM
The Multi-scale Modeling Framework (MMF) has recently emerged as a unique approach toward bridging the gap between conventional global climate models (GCMs) and high-resolution global cloud-resolving models. The MMF replaces conventional cloud parameterizations with a small-domain cloud-resolving model (CRM; often called in this context a "super-parameterization"), embedded within each GCM grid column, to explicitly represent clouds and their effects. Recently, the Colorado State University MMF was used to conduct a 19- year Atmospheric Model Intercomparison Project (AMIP)-style simulation using observed SSTs and sea ice extent as the lower boundary conditions and the NCAR Community Atmosphere Model (CAM) as the host GCM. Highlights from a detailed examination of the MMF-simulated structure of intraseasonal convectively- coupled disturbances, i.e., the Madden-Julian Oscillation (MJO), will be presented. In the MMF simulation, the composite structure of MJO-like disturbances compares very favorably with the observed MJO based on reanalyses and satellite-derived measurements, although the magnitude of intraseasonal variability is slightly too strong in the model.