The past decade has been marked by significant advancements in numerical weather prediction of hurricanes, which have greatly contributed to the steady decline in forecast track error. Since its operational implementation by the U.S. National Weather Service (NWS) in 1995, the best-track model performer has been NOAA’s regional hurricane model developed at the Geophysical Fluid Dynamics Laboratory (GFDL). The purpose of this paper is to summarize the major upgrades to the GFDL hurricane forecast system since 1998. These include coupling the atmospheric component with the Princeton Ocean Model, which became operational in 2001, major physics upgrades implemented in 2003 and 2006, and increases in both the vertical resolution in 2003 and the horizontal resolution in 2002 and 2005. The paper will also report on the GFDL model performance for both track and intensity, focusing particularly on the 2003 through 2006 hurricane seasons. During this period, the GFDL track errors were the lowest of all the dynamical model guidance available to the NWS Tropical Prediction Center in both the Atlantic and eastern Pacific basins. It will also be shown that the GFDL model has exhibited a steady reduction in its intensity errors during the past 5 yr, and can now provide skillful intensity forecasts. Tests of 153 forecasts from the 2004 and 2005 Atlantic hurricane seasons and 75 forecasts from the 2005 eastern Pacific season have demonstrated a positive impact on both track and intensity prediction in the 2006 GFDL model upgrade, through introduction of a cloud microphysics package and an improved air–sea momentum flux parameterization. In addition, the large positive intensity bias in sheared environments observed in previous versions of the model is significantly reduced. This led to the significant improvement in the model’s reliability and skill for forecasting intensity that occurred in 2006.

This study documents a new parametric hurricane rainfall prediction scheme, based on the rainfall climatology and persistence model (R-CLIPER) used operationally in the Atlantic Ocean basin to forecast rainfall accumulations. Although R-CLIPER has shown skill at estimating the mean amplitude of rainfall across the storm track, one underlying limitation is that it assumes that hurricanes produce rain fields that are azimuthally symmetric. The new implementations described here take into account the effect of shear and topography on the rainfall distribution through the use of parametric representations of these processes. Shear affects the hurricane rainfall by introducing spatial asymmetries, which can be reasonably well modeled to first order using a Fourier decomposition. The effect of topography is modeled by evaluating changes in elevation of flow parcels within the storm circulation between time steps and correcting the rainfall field in proportion to those changes. Effects modeled in R-CLIPER and those from shear and topography are combined in a new model called the Parametric Hurricane Rainfall Model (PHRaM). Comparisons of rainfall accumulations predicted from the operational R-CLIPER model, PHRaM, and radar-derived observations show some improvement in the spatial distribution and amplitude of rainfall when shear is accounted for and significant improvements when both shear and topography are modeled.

[1] Rainfall and flooding associated with landfalling tropical cyclones are examined through empirical analyses of three hurricanes (Frances, Ivan, and Jeanne) that affected large portions of the eastern U.S. during September 2004. Three rainfall products are considered for the analyses: NLDAS, Stage IV, and TMPA. Each of these products has strengths and weaknesses related to their spatio-temporal resolution and accuracy in estimating rainfall. Based on our analyses, we recommend using the Stage IV product when studying rainfall distribution in landfalling tropical cyclones due to its fine spatial and temporal resolutions (about 4-km and hourly) and accuracy, and the capability of estimating rainfall up to 150 km from the coast. Lagrangian analyses of rainfall distribution relative to the track of the storm are developed to represent evolution of the temporal and spatial structure of rainfall. Analyses highlight the profound changes in rainfall distribution near landfall, the changing contributions to the rainfall field from eyewall convection, inner rain bands and outer rain bands, and the key role of orographic amplification of rainfall. We also present new methods for examining spatial extreme of flooding from tropical cyclones and illustrate the links between evolving rainfall structure and spatial extent of flooding.

A scheme for validating quantitative precipitation forecasts (QPFs) for landfalling tropical cyclones is developed and presented here. This scheme takes advantage of the unique characteristics of tropical cyclone rainfall by evaluating the skill of rainfall forecasts in three attributes: the ability to match observed rainfall patterns, the ability to match the mean value and volume of observed rainfall, and the ability to produce the extreme amounts often observed in tropical cyclones. For some of these characteristics, track-relative analyses are employed that help to reduce the impact of model track forecast error on QPF skill. These characteristics are evaluated for storm-total rainfall forecasts of all U.S. landfalling tropical cyclones from 1998 to 2004 by the NCEP operational models, that is, the Global Forecast System (GFS), the Geophysical Fluid Dynamics Laboratory (GFDL) hurricane model, and the North American Mesoscale (NAM) model, as well as the benchmark Rainfall Climatology and Persistence (R-CLIPER) model. Compared to R-CLIPER, all of the numerical models showed comparable or greater skill for all of the attributes. The GFS performed the best of all of the models for each of the categories. The GFDL had a bias of predicting too much heavy rain, especially in the core of the tropical cyclones, while the NAM predicted too little of the heavy rain. The R-CLIPER performed well near the track of the core, but it predicted much too little rain at large distances from the track. Whereas a primary determinant of tropical cyclone QPF errors is track forecast error, possible physical causes of track-relative differences lie with the physical parameterizations and initialization schemes for each of the models. This validation scheme can be used to identify model limitations and biases and guide future efforts toward model development and improvement.