A12D-01 INVITED
High Resolution Modeling of Hurricanes in a Climate Context
Modeling of tropical cyclone activity in a climate context initially focused on simulation of relatively weak tropical storm-like disturbances as resolved by coarse grid (200 km) global models. As computing power has increased, multi-year simulations with global models of grid spacing 20-30 km have become feasible. Increased resolution also allowed for simulation storms of increasing intensity, and some global models generate storms of hurricane strength, depending on their resolution and other factors, although detailed hurricane structure is not simulated realistically. Results from some recent high resolution global model studies are reviewed. An alternative for hurricane simulation is regional downscaling. An early approach was to embed an operational (GFDL) hurricane prediction model within a global model solution, either for 5-day case studies of particular model storm cases, or for "idealized experiments" where an initial vortex is inserted into an idealized environments derived from global model statistics. Using this approach, hurricanes up to category five intensity can be simulated, owing to the model's relatively high resolution (9 km grid) and refined physics. Variants on this approach have been used to provide modeling support for theoretical predictions that greenhouse warming will increase the maximum intensities of hurricanes. These modeling studies also simulate increased hurricane rainfall rates in a warmer climate. The studies do not address hurricane frequency issues, and vertical shear is neglected in the idealized studies. A recent development is the use of regional model dynamical downscaling for extended (e.g., season-length) integrations of hurricane activity. In a study for the Atlantic basin, a non-hydrostatic model with grid spacing of 18km is run without convective parameterization, but with internal spectral nudging toward observed large-scale (basin wavenumbers 0-2) atmospheric conditions from reanalyses. Using this approach, our model reproduces the observed increase in Atlantic hurricane activity (numbers, Accumulated Cyclone Energy (ACE), Power Dissipation Index (PDI), etc.) over the period 1980-2006 fairly realistically, and also simulates ENSO-related interannual variations in hurricane counts. Annual simulated hurricane counts from a two-member ensemble correlate with observed counts at r=0.86. However, the model does not simulate hurricanes as intense as those observed, with minimum central pressures of 937 hPa (category 4) and maximum surface winds of 47 m/s (category 2) being the most intense simulated so far in these experiments. To explore possible impacts of future climate warming on Atlantic hurricane activity, we are re-running the 1980- 2006 seasons, keeping the interannual to multidecadal variations unchanged, but altering the August-October mean climate according to changes simulated by an 18-member ensemble of AR4 climate models (years 2080- 2099, A1B emission scenario). The warmer climate state features higher Atlantic SSTs, and also increased vertical wind shear across the Caribbean (Vecchi and Soden, GRL 2007). A key assumption of this approach is that the 18-model ensemble-mean climate change is the best available projection of future climate change in the Atlantic. Some of the 18 global models show little increase in wind shear, or even a decrease, and thus there will be considerable uncertainty associated with the hurricane frequency results, which will require further exploration. Results from our simulations will be presented at the meeting.
A12D-02 INVITED
Modeling the Upper Ocean Response to a Hurricane.
The upper ocean response to a hurricane is a three-dimensional and time-dependent phenomenon that can be analyzed usefully from several different perspectives. From a (numerical) modeling perspective, a significant issue is whether a given phenomenon is resolved by an ocean model (and model hurricane) or has to be parameterized. Ocean models suitable for coupled air-sea forecasting applications will likely have a horizontal and vertical resolution of about 5 km and 10 m. Resolved phenomena will then include a hurricane-scale rightward bias of the surface current amplitude, and the downward and outward propagation of near inertial- frequency energy. Parameterized phenomena include diapycnal mixing within the upper ocean. This mixing determines the effective depth of the ocean surface layer and is the main contributor to the cooling of SST that is thought to be an important aspect of air-sea interaction. Evidence from the CBLAST deployment of EM-Apex floats is that this mixing is a response to stratified shear flow instability. The rightward bias of SST cooling is then a direct result of the rightward bias of the current amplitude. Satellite infrared images show a very clear signature of a rightward bias in SST cooling. They also show a significant along-track variability of SST cooling on scales of tens of kilometers whose origin is unclear. Certainly this kind of variability is not reproduced by ocean models forced by smoothly varying model hurricanes and started with homogeneous ocean initial conditions.
A12D-03
Comparison of three parameterizations of subgrid-scale feedback physics for sea spray contributions to near-surface heat fluxes in hurricanes
Air-sea interaction is unquestionably a critical aspect of hurricane structure and evolution. Recent experiments with high resolution (km-scale) nested models have shown strong sensitivity to the representation of air sea fluxes - both the direct (interfacial) transfers and contributions of sea spray. With the advent of fully coupled (air-wave-ocean) models, it becomes possible to represent the surface fluxes in more physical detail (as opposed to simple wind-speed driven bulk flux algorithms). Recent theoretical work has led to a new representation of sea spray droplet flux that is driven by the fundamental processes associated with blowing large droplets off the tops of breaking waves. In today's high-resolution numerical models sea spray effects are not represented explicitly but are input at modifications to bulk flux relationships applied to the lowest atmospheric grid level. These bulk flux relationships assume a subgrid-scale structure for the scalar profiles (essentially Monin-Obukhov similarity). The actual contribution of the droplet spectrum to the sensible and latent heat fluxes is computed assuming the droplets fall through this specified temperature/humidity environment. However, this is complicated by coupling of the droplet heat fluxes to the mean temperature and humidity profiles near the surface. Thus, the heat and moisture transferred by the droplets modifies the local environment that drives these transfers - a process referred to a feedback. Feedback has little or no effect on the total enthalpy flux carried by the sea spray but it affects the partition of sensible versus latent heat (Bowen ratio). In this paper we will describe and compare three representations (one old and two new) of this subgrid-scale feedback process.
A12D-04
Calibration Experiments for Nested HYCOM Coupled to Hurricane-WRF
The next generation Ocean-Atmosphere coupled models for Hurricane Prediction at NCEP include a triple nested atmospheric model (GFS-fix, HWRF coarse-moving and HWRF fine) coupled to a double nested ocean model (RTOFS-HYCOM). A sensitivity study of (numerical and physical) model parameters for storms derived from the historical record, available observations, and other modeling results are used to guide the calibration of the nested ocean model component.
A12D-05
Hurricane Modeling and Supercomputing: Can a global mesoscale model be useful in improving forecasts of tropical cyclogenesis?
Hurricane modeling, along with guidance from observations, has been used to help construct hurricane theories since the 1960s. CISK (conditional instability of the second kind, Charney and Eliassen 1964; Ooyama 1964,1969) and WISHE (wind-induced surface heat exchange, Emanuel 1986) are among the well-known theories being used to understand hurricane intensification. For hurricane genesis, observations have indicated the importance of large-scale flows (e.g., the Madden-Julian Oscillation or MJO, Maloney and Hartmann, 2000) on the modulation of hurricane activity. Recent modeling studies have focused on the role of the MJO and Rossby waves (e.g., Ferreira and Schubert, 1996; Aivyer and Molinari, 2003) and/or the interaction of small-scale vortices (e.g., Holland 1995; Simpson et al. 1997; Hendrick et al. 2004), of which determinism could be also built by large-scale flows. The aforementioned studies suggest a unified view on hurricane formation, consisting of multiscale processes such as scale transition (e.g., from the MJO to Equatorial Rossby Waves and from waves to vortices), and scale interactions among vortices, convection, and surface heat and moisture fluxes. To depict the processes in the unified view, a high-resolution global model is needed. During the past several years, supercomputers have enabled the deployment of ultra-high resolution global models, obtaining remarkable forecasts of hurricane track and intensity (Atlas et al. 2005; Shen et al. 2006). In this work, hurricane genesis is investigated with the aid of a global mesoscale model on the NASA Columbia supercomputer by conducting numerical experiments on the genesis of six consecutive tropical cyclones (TCs) in May 2002. These TCs include two pairs of twin TCs in the Indian Ocean, Supertyphoon Hagibis in the West Pacific Ocean and Hurricane Alma in the East Pacific Ocean. It is found that the model is capable of predicting the genesis of five of these TCs about two to three days in advance. Our real-data simulations are the first of this type of global model experiment to support the following hypothesis: the MJO may play an essential role in the formation of this kind of TC, and therefore its occurrence could represent a crucial precursor to TC genesis.
A12D-06
Representing Hurricanes with a Nested Global Forecast Model
A global forecast model is essential for predicting hurricane tracks beyond a period of ~2 days since global processes that may influence the longer-term storm tracks can be represented explicitly and there are no errors from the lateral boundary conditions that can propagate into the model domain and diminish the accuracy of the track forecasts. However, global models usually do not have enough horizontal and vertical resolution to produce meaningful hurricane intensity forecasts. Most current operational global forecast models represent the atmosphere horizontally using spherical harmonic basis functions with an equivalent resolution of ~40-50 km. The NOAA Science Advisory Board Hurricane Intensity Research Working Group recommends approximately 1-km-resolution hurricane forecasts in order to represent the important physical processes in the core region of hurricanes that are important to accurately predict hurricane intensity. Even with state-of-the-art computers, it will be many years before global forecasts with 1-km horizontal resolution are practical. To predict both hurricane tracks and intensity well, a nested global model is necessary. Large-scale processes are represented on a coarser, computationally-efficient grid while features such as hurricanes are represented on a high-resolution nest. The global model used in this study is the Ocean-Land-Atmosphere Model (OLAM) being developed at Duke University. OLAM is the global successor to the Regional Atmospheric Modeling System (RAMS), which originated at Colorado State University in 1986. OLAM uses the same physics parameterizations as RAMS, but it solves the governing equations by discretizing the atmosphere on an unstructured triangular finite-volume grid. The triangular grid uses the Arakawa-C staggering and is fully mass conservative. Since the triangular mesh is unstructured, the mesh can be refined to produce much higher horizontal resolution in areas of interest such as near hurricanes. Here, we examine hurricane track and intensity forecasting in a global nested model using a real-data case. Using a high-resolution nest in the vicinity of a hurricane, we examine how well the inner core hurricane structure can be resolved in order to produce meaningful intensity forecasts. We also determine if the better representation of hurricanes also leads to better longer-term hurricane track predictions.
A12D-07
Representation of Hurricanes in the new ECMWF Nature Run
One crucial component of any Observing System Simulation Experiment (OSSE) is the Nature Run (NR) which has the purpose of representing a virtual atmosphere from which observations can be simulated so that the impact of future instruments can be assessed. A community-based, Joint OSSE Nature Run (NR) was designed in an international collaborative effort and was produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) to contribute building a next-generation OSSE capability. In this work this new 13-month T511 Nature Run is being analyzed with emphasis on tropical development over the western African monsoon region and the tropical Atlantic. In particular, the NR representation of the African Easterly Jet (AEJ) and the characteristics of African Easterly Waves (AEWs) including their propagation and development in tropical cyclones are investigated. It is shown that, given the resolution limitations, the NR provides a very realistic representation of different processes such as genesis, development, extra-tropical transitions, track variability and multiple vortex interaction.