Multi-scale modeling of clouds and process studies using CRM and MMF frameworks (Invited)
Tropical convection often organizes itself into large-convective systems, from squall lines to Madden-Julian Oscillation (MJO). The effects of the multi-scale organization of convection on large-scale circulation are poorly understood. The talk will demonstrate the application of a cloud-resolving model (CRM) to meso-scale organization of convection in the Inter-Tropical Convergence Zone (ITCZ) and to the problem of tropical cyclogenesis. The application of a Multi-scale Modeling Framework (MMF) to large-scale organization of convection such as the MJO is also discussed. The MMF is a global climate model in which most of conventional sub-GCM-grid-scale parameterizations in each grid-column are replaced with imbedded CRMs, called super-parameterization. The MMF based on super-parameterized version of the Community Atmosphere Model, SP-CAM, produces robust MJO signal both in Earth and so-call Aqua-planet simulations, when the continents are removed, and observed sea-surface temperature (SST) is replaced with zonally symmetric SST. Sensitivity of the MJO on Aqua-planet to cold and warm SST perturbations, to zonal homogenization of surface fluxes and radiation, and to water vapor anomalies is discussed. Preliminary results of running the MMF in weather-forecasting mode during the YOTC period are reported.
Parameterizing convection in high-resolution global atmospheric models (Invited)
Representing the effects of atmospheric convection in global models has been, and remains, one of the major obstacles facing climate researchers. As the horizontal resolution used in global climate simulations increases to scales much finer than 100 km, problems related to convection show no signs of diminishing. This talk will present results from high resolution global atmospheric models, as well as cloud resolving model results, and satellite measurements from the CloudSat and TRMM instruments. These results suggest that climate simulations at high horizontal resolutions may in fact present new challenges to convection parameterization. Tropical cyclone simulations conducted at ¼o (or ~25 km) resolution show that standard climate-style convection parameterizations may interfere with the organization and strengthening of tropical systems. A three-way comparison of cloud resolving model (CRM) results, satellite data, and global simulations, suggests that assumptions about scale-separation and statistical equilibrium between convection and the resolved flow begin to break down for spatial scales smaller than 100 km. Quantities such as convective cloud height exhibit large variance when sorted into regimes with similar background meteorology. Simply put, a one-to-one relationship between convective parameters and resolved model fields may not exist, even approximately, for scales smaller than 100km. Possible remedies, including a stochastic component for parameterized convection, based on CRM results and satellite measurements, are discussed
The MJO in an Aquaplanet General Circulation Model
An aquaplanet atmospheric general circulation model simulation with a robust intraseasonal oscillation is analyzed. The SST boundary condition resembles the observed December-April average, although with the meridional SST gradient reduced to be one-quarter of that observed poleward of 10 degrees latitude. Slow, regular eastward propagation at 5 m/s in winds and precipitation with amplitude greater than those in the observed MJO is clearly identified in unfiltered fields. The local relationship between precipitation rate column precipitable water is strongly increasing and nonlinear, as in observations. The model intraseasonal oscillation resembles a moisture mode that is destabilized by wind-evaporation feedback, and that propagates eastward through advection of anomalous humidity by the sum of perturbation winds and mean westerly flow. Moistening of the troposphere occurs to the east of (in quadrature with) enhanced precipitation, and is dominated by a column-integrated horizontal advection moistening rate of greater than 2 mm/day. Zonal and meridional moisture advection are of approximately equal amplitude, although meridional advection tends to damp tropospheric moisture anomalies, while zonal advection propagates them eastward. At the time of peak moistening in the model, the total zonal wind near 850 hPa is 5 m/s. This value is approximately the same as the phase speed of the intraseasonal disturbances; together with other diagnostics, this suggests that horizontal advection of moisture is the dominant propagation mechanism. Latent heat flux is the second largest term in the intraseasonal moisture budget and has a positive covariance with precipitation anomalies. A mechanism denial experiment in which intraseasonal latent heat flux variability is removed largely eliminates intraseasonal wind and precipitation variability, although weak, small spatial scale convective features that move slowly eastward, presumably through advection, are still present in the simulation. Aquaplanet simulations with different basic states are also conducted and discussed. An SST boundary condition that resembles the December-April average produces realistic amplitude intraseasonal wind variability, although precipitation contains more variance at low frequencies than in observations. A zonally symmetric SST basic state produces weak and unrealistic intraseasonal variability between 30 and 90 day timescales, indicating the importance of mean low-level westerly winds and hence a realistic phase relationship between precipitation and surface flux anomalies for producing realistic tropical intraseasonal variability.
Relative Roles of Convection Closure, Trigger Condition and Convective Momentum Transport in GCM Simulations of the Madden-Julian Oscillation
The uncertainty in the representation of convection is responsible for the poor simulation of the Madden-Julian Oscillation by most general circulation models (GCMs). With several modifications made to the deep convection scheme, the Iowa State University (ISU) GCM, which is based on a version of the NCAR Community Climate Model, is able to simulate many features of MJO as revealed by observations such as the amplitude, spatial distribution, eastward propagation, horizontal and vertical structures, and the coherent feature of eastward propagating convection and the precursor sign of convective center. In this paper, four 10-year (1979-88) ISUGCM simulations with observed sea surface temperatures are analyzed and compared to examine relative roles of the revised convection closure, convection trigger condition and convective momentum transport (CMT) in the MJO simulations. The revised convection closure plays a key role in the improvement of eastward propagation of MJO. The convection trigger helps produce less frequent but more vigorous moist convection and enhance the amplitude of the MJO signal. The inclusion of CMT results in more coherent structure for the MJO deep convective center and its corresponding atmospheric variances.
What can Cloud-Resolving Models Tell us About Critical Phenomena in Atmospheric Precipitation?
Recent work suggests that observations of Tropical precipitation conform to properties associated with critical phenomena of other systems (Peters and Neelin 2006). The precipitation retrievals are averages over 25-km by 25-km areas and are snapshots in time, and therefore unable to reveal the underlying, smaller-scale physical processes. We are using a 3D cloud-resolving model (CRM) to resolve these processes in space and time, and thereby allow us to investigate the underlying physics in detail. The CRM was run over a large domain (1000 km by 1000 km) for a long time (~10 days) in order to adequately sample the rare large events. In addition, we are using results from a 4-year global simulation using a climate model based on the multi-scale modeling framework (MMF). Whereas conventional parameterizations are based on statistical theories involving uncertain closure assumptions, MMFs represent cloud processes on their native scales by embedding a 2D CRM with a 4-km horizontal grid size in each climate model grid column. We have analyzed the model results following the methodology of Peters and Neelin. We used the results to produce rainfall rates conditioned on column water vapor and column temperature over the Tropical oceans. We have also analyzed additional statistical aspects of Tropical convection in the 3D CRM simulations that are related to critical behavior. We have found that: (1) CRMs are able to reproduce nearly all of the observed statistics of strong convective precipitation over tropical oceans. (2) CRMs and MMFs do not generally reproduce the observed roll-off of precipitation rate at large column water vapor values. (3) Analysis of CRM results suggests that many of the observed features are due to the tight coupling between dynamics and moist thermodynamics in convective updrafts.
Representation of intra-seasonal activity in coupled GCMs with varying atmospheric and oceanic resolutions
The Geophysical Fluid Dyanmics Laboratory (GFDL) is actively involved in the development of high resolution coupled GCMs to be used for prediction and predictability studies from intra-seasonal to decadal time scales. A number of coupled GCMs with different atmospheric and oceanic grid resolutions have been integrated from 20 to 100 years. An assessment of how well these GCMs simulate observed MJO and other intra-seasonal activity is done by applying several diagnostics developed by the MJO CliVAR working group to daily OLR and precipitation fields. Improvement in the geographical distribution of daily precipitation variance is seen with higher atmospheric resolution and there appears to be some further improvement with higher oceanic resolution. From Wheeler-Kiladis wave frequency analyses, it appears that higher atmospheric horizontal resolution improves the speed (slows down) and amplitude of the MJO. There is also an indication of a better Kelvin mode representation with higher resolution, although some other studies show that changes to the convective parameterization also have a significant impact on the Kelvin modes and the capability to capture the asymmetric mixed Rossby gravity modes. These coupled GCM experiments show an encouraging improvement in the representation of intra-seasonal activity with increased atmospheric and oceanic resolution.
Evaluating parameterized variables in the Community Atmospheric Model along the GCSS Pacific cross-section during YOTC
We use short-term forecasts along the GCSS Pacific cross-section to evaluate parameterizations in several versions of the Community Atmospheric Model (CAM) including the version used for the IPCC AR4 and the ones to be used for the IPCC AR5. Climate models are commonly validated against various statistics based on observations. However, climate models can achieve a reasonable mean state as the result of compensating errors, which are impossible to untangle after long integrations. An innovative way to evaluate parameterizations in climate models is to use the weather forecasting approach: the state of the atmosphere is initialized with realistic conditions and the model is run for short-term forecasts. This approach allows a direct comparison of the parameterized variables (e.g. clouds, precipitation, radiation) with observations. Therefore, it is possible to gain insight into the parameterization deficiencies and to diagnose the processes behind the drift away from reality. The YOTC period is well suited for this approach because of the availability of state-of-the-art analyses to initialize the model and a wealth of integrated observational datasets to evaluate the forecasts. The GCSS cross-section (that runs from the stratocumulus regions off the coast of California, across the shallow convection dominated trade winds, to the deep convection regions of the ITCZ) is particularly relevant for such a comparison because it includes several important cloud regimes, and their interactions through the large-scale circulation. In this study, forecasts are initialized from ECMWF analyses and the CAM is run for 5 days to determine the differences with satellite data along the cross-section. We focus on JJA 2008 and the forecasts are evaluated against A-train, AIRS, TRMM, SSM/I, ISCCP and CERES products. The mean forecast biases grow very quickly along the cross-section, and after 5 days, the error pattern is very similar to the mean climate error. Around the ITCZ, most of the temperature and moisture errors have developed after a single day. Over the stratocumulus region, the error grows more slowly and it takes 5 days before the mean forecast error reaches the amplitude of the mean climate error. The CAM3 (used for the IPCC AR4 runs) significantly overestimates the temperature along the Pacific cross-section. These warm biases are attributable to deep convection errors and their propagation to other cloud regimes. The temperature and moisture biases are reduced in IPCC AR5 versions of the CAM. With a modified deep convection scheme, the local tropical errors are reduced and there is a dramatic improvement of the precipitation in the ITCZ region. Also, in association with the adjusted atmospheric state, the biases are also reduced away from the region of significant deep convection (i.e. in the shallow cumulus and stratocumulus regimes). This clearly illustrates the interaction between the tropical convection and the large-scale circulation. We also show that the inclusion of a two-moment microphysics scheme further reduces upper troposphere errors and that a more realistic representation of cloud-topped boundary layers deepens the boundary layer and produces more realistic low-level clouds.
Mesoscale weather and climate modeling with the global non-hydrostatic Goddard Earth Observing System Model (GEOS-5) at cloud-permitting resolutions
The Goddard Earth Observing System Model (GEOS-5), an earth system model developed in the NASA Global Modeling and Assimilation Office (GMAO), has integrated the non-hydrostatic finite-volume dynamical core on the cubed-sphere grid. The extension to a non-hydrostatic dynamical framework and the quasi-uniform cubed-sphere geometry permits the efficient exploration of global weather and climate modeling at cloud permitting resolutions of 10- to 4-km on today's high performance computing platforms. We have explored a series of incremental increases in global resolution with GEOS-5 from it's standard 72-level 27-km resolution (~5.5 million cells covering the globe from the surface to 0.1 hPa) down to 3.5-km (~3.6 billion cells). We will present results from a series of forecast experiments exploring the impact of the non-hydrostatic dynamics at transition resolutions of 14- to 7-km, and the influence of increased horizontal/vertical resolution on convection and physical parameterizations within GEOS-5. Regional and mesoscale features of 5- to 10-day weather forecasts will be presented and compared with satellite observations. Our results will highlight the impact of resolution on the structure of cloud features including tropical convection and tropical cyclone predicability, cloud streets, von Karman vortices, and the marine stratocumulus cloud layer. We will also present experiment design and early results from climate impact experiments for global non-hydrostatic models using GEOS-5. Our climate experiments will focus on support for the Year of Tropical Convection (YOTC). We will also discuss a seasonal climate time-slice experiment design for downscaling coarse resolution century scale climate simulations to global non-hydrostatic resolutions of 14- to 7-km with GEOS-5.