|
|
||
| |
Climate change, climate variability, and extreme weatherWhereas in the past, meteorology and climatology were separate fields, be it only because of disparate time (and length scales as well), it appears today that the two fields are strongly coupled, not only as the climate gives the boundaries for investigating the weather, but also because localized events can influence the larger climatological scales. The specific items on which ESSL scientists focused in this year are related to the role of aerosols in climate and weather, to the coupling of eco-systems, biochemistry and climate, to climate change, climate variability and extreme weather such as hurricanes, to interactions of the water cycle with climate and weather, to the impacts of climate and weather on society and ecosystems and finally to megacities and the effects of urbanization; the latter priority is a highlight for NCAR and concerns the international multi-agency field campaign that took place in Mexico city and combined interactive modeling as well. The laboratory highlights are related to the role of aerosols, to the regional carbon cycle, to a numerical simulation of turbulence, to landfall of hurricanes, to the global and regional water cycles and to polar climate. Landfall of hurricanes [LAR Highlight] - MMM
Landfall of hurricanes
The 2005 Atlantic hurricane season is a vivid reminder of the economic and societal consequences of landfalling tropical cyclones. Improved forecasts of hurricane intensity change, and time-extensions of skillful track prediction are vital for evacuation strategies. Furthermore, accurate assessment of the uncertainty in hurricane forecasts is critical in a variety of economic sectors. Progress requires solving difficult problems such as the inner-core hurricane dynamics and how it affects intensity, quantifying the net enthalpy flux from the ocean in high-wind-speed conditions, and incorporation of a variety of remotely sensed data into model initial conditions. The purpose of ESSL/MMM research in hurricane simulations is to create the next generation hurricane-prediction system, and a community hurricane-prediction model that can be used for process and predictability studies. During the past year, analysis of real-time WRF-ARW forecasts of landfalling Atlantic hurricanes during 2004 and 2005 were examined. Notable improvements of track and intensity forecasts over operational models were found at lead times from 48-96 h for two different model configurations, one whose finest grid increment was 12-km and another with a moving 4-km mesh. Shortcomings in these real-time forecasts have led to exploration of three research areas: (1) finer resolution in the inner core; (2) advanced data assimilation using 3DVAR and the Ensemble Kalman Filter; and (3) coupling with an ocean model to simulate the mixing-induced cooling of the upper ocean and negative feedback on storm intensity. Studies with hurricane Katrina showed that only with a second concentric moving nest could any of the rapid intensification of the storm be captured. With the EnKF method to initialize the model, a nearly correct forecast of landfall location was achieved about 1 day earlier than using the global analysis alone for initial conditions. Significant sensitivity to surface exchange coefficients was found as well, in accord with theory, and improved parameterizations have been implemented. Finally, a simple, mixed-layer ocean model was found able to produce much of the observed ocean cooling in Katrina, and, moreover, produced changes in the ocean mixed layer in agreement with more sophisticated ocean models. This mixed-layer model is now part of WRF. Real-time forecasts during the 2006 season have so far employed the EnKF initialization method in parallel with a conventional initialization, both using a 12-km single-domain configuration of WRF. The addition of a second nest, now with a 1.33-km spacing - about 1/6 of the grid increment of current operational models – captures the inner core dynamics. In the coming year, the performance of all the new physics and nesting for hurricane simulations will be critically examined with an emphasis on inner core Rossby waves, intensity change due to ocean coupling, improvements in track with the EnKF and the inclusion of airborne radar data using the EnKF and 3DVAR initialization schemes for selected cases. Movie: Animation of model-generated composite radar reflectivity for hurricane Wilma. This real-time forecast shows the initial stall of the storm over the Yucatan Peninsula, its later landfall in Florida (within a few hours of the correct time) and ultimate transition to a more frontal structure off the East Coast of the U. S. Real-time model simulations of hurricanes are important to researchers and the development of the next generation hurricane-prediction system. Convection organization: Observational analysis and resolved simulations
The propensity for deep, moist convection to organize and project onto larger spatial and temporal scales requires numerical simulations spanning convection-resolving scales to continental scales. Furthermore, simulation studies must be closely constrained by observational analysis of the organizing properties of convection. Prediction of tropical and warm-season higher-latitude convection, and the response of the synoptic-scale and planetary-scale flow, is vital for increasing our ability to anticipate significant weather events more than a day in advance. It is also vital for credible representations by models of regional climate. Much of the work in the previous year has been related to the USWRP or the Water Cycle Across Scales efforts. As an important application to global rainfall climatologies, satellite-derived estimates of rainfall-propagation speed were found to exceed systematically those derived from radar by 15% over the U.S. Moreover, roughly 70% of warm season rainfall was attributed to propagating convection. Based on WRF simulations and composite radar and gridded analyses, the occurrence of multi-day “corridors” of propagation, and the origin of strong, elevated nocturnal convection was crucially dependent on frontogenesis induced by the low-level jet impinging on a nearly stationary surface. Explicit simulations of episodes indicate more sensitivity to land surface and boundary layer formulations than to cloud microphysics, but model resolution studies have revealed a lack of numerical convergence until the grid spacing reaches only a few hundred meters. Therefore, sensitivity and process studies using models with grid increments greater than this value should be interpreted cautiously. Convection in Central America was found to exhibit its most diurnally synchronized propagation away from the mountains, rather than along the mountains. No clear relation to easterly waves was found, rather, deeper easterly flow and stronger easterly shear separated propagating and non-propagating regimes (Figure). Over North Africa, easterly waves modulate convection, but the clearest signal is that of terrain forcing and westward propagation at nearly the speed of the African easterly jet. In the coming year, more attention will be focused on the role of mountain ranges upstream from convection episodes on many continents, starting with North America, and looking for lower- and mid-tropospheric mesoscale disturbances shed by terrain that can focus convection downstream. Work will commence on simulations of Asian convective episodes with emphasis on the role of the Tibetan Plateau. Further observational analysis of convection characteristics is planned for southern Africa, with WRF simulations begun for active and inactive convection regimes in Sahelian North Africa. Idealized simulations of squall lines will be used to investigate appropriate turbulence parameterizations for grid spacings of 1-4 km that will be important for future process studies with cloud resolving models. Super parameterization
Representation of convection and cloud processes is the most uncertain aspect of numerical modeling of weather and climate. Horizontal grid spacing around 1 km is required to resolve gross features of deep convection and even higher resolution is needed for shallow clouds. Only recently cloud-resolving (or convection-permitting) resolutions have become feasible for the continental-scale NWP using nonhydrostatic models, e.g., the Weather Research and Forecasting (WRF) Model (see http://www.wrf-model.org). For climate models, cloud-resolving modeling of atmospheric general circulation is in its infancy, with the first global cloud-resolving simulations just recently completed in Japan (see http://www.ccsr.u-tokyo.ac.jp/~satoh/nicam/index.html). As far as traditional climate models are concerned, a different approach was proposed by ESSL/MMM. The cornerstone is to apply a two-dimensional (2D) cloud-resolving model in each column of a large-scale model to represent small-scale and mesoscale processes, and the two-way coupling between convection and larger scales of motion. This approach was termed the Cloud-Resolving Convection Parameterization (CRCP) and has since been often referred to as the "superparameterization" (SP). The superparameterization approach has been applied to investigate the large-scale organization of tropical convection in idealized simulations, focusing on physical processes underlying the tropical intraseasonal oscillations and Madden-Julian Oscillation (MJO). The most recent investigation contributes to the debate concerning the role of the oceanic processes in MJO. Traditional climate models suggest that interactive sea surface temperature (SST) improves the model’s climate variability on intraseasonal timescales. In contrast, results obtained using SP approach suggest that interactive SST impedes the development of large-scale organization (see figure) and has an insignificant impact on the dynamics of mature MJO-like systems. Work has also begun to improve microphysical parameterizations used in the SP model. The latter will provide a basis for collaborative research within the recently-created NSF Science and Technology Center for Multi-Scale Modeling of Atmospheric Processes (CMMAP), based at the Colorado State University. More recent results suggest that SP provides a useful approach to convective parameterization in a mesoscale model (i.e., a model with gridlength of 10-50 km). For such resolutions, the mesoscale model handles mesoscale convective organization, whereas SP models handle small-scale convective dynamics only. The mesoscale model featuring SP will be applied in the near future to the problem of summertime organized convection over continental U.S. Support for this work includes NSF, NSF/ROCEW and NOAA. NAMMA
The NASA-sponsored African Monsoon Multidisciplinary Analyses (NAMMA) campaign was a field research investigation conducted during August-September 2006 based in the Cape Verde Islands, 350 miles off the coast of Senegal in west Africa. The goals of this project include an examination of the formation and evolution of tropical cyclones in the eastern and central Atlantic, a characterization of the composition and structure of the Saharan Dust Layer, and a determination of whether aerosols affect cloud precipitation and influence cyclone development. As part of this effort, ESSL/MMM scientists participated in the field campaign and provided microphysical probes on the NASA DC-8 aircraft, with the goal of characterizing the conversion of liquid water to ice in the updrafts of intense convection in the NAMMA domain. Conversion of liquid water to ice, especially in the updrafts where additional buoyancy can be produced by ice latent heating, may be a central factor in mesoscale vortex spinup. Saharan dust may significantly affect this process because it provides copious ice nuclei that could lower the height of ice formation in the updrafts. From NCAR's suite of instruments, scientists will be able to characterize the chemical composition and physical sizes of the Saharan dust nuclei that were active in cloud droplet and ice particle formation and determine their role in cloud droplet and ice formation processes. Researchers have an excellent data set to address the following questions:
Future analyses efforts will focus on these four questions. Support for this work includes NSF and NASA. |
|
