Evaluation of Multiple Dynamic Initialization Schemes for Tropical Cyclone Prediction

2013 ◽  
Vol 141 (11) ◽  
pp. 4028-4048 ◽  
Author(s):  
Eric A. Hendricks ◽  
Melinda S. Peng ◽  
Tim Li
Keyword(s):  
Tropical Cyclone ◽  
Three Dimensional ◽  
Equation Model ◽  
Naval Research ◽  
Lead Times ◽  
Second Stage ◽  
The North ◽  
Initialization Scheme ◽  
Prediction Systems ◽  
Tc Prediction

Abstract Three different dynamic initialization schemes for tropical cyclone (TC) prediction in numerical prediction systems are described and evaluated. The first scheme involves the removal of the analyzed vortex, followed by the insertion of a dynamically initialized vortex into the model analyses. This scheme is referred to as the tropical cyclone dynamic initialization scheme (TCDI) because the TC component is nudged to the observed surface pressure in an independent three-dimensional primitive equation model prior to insertion. The second scheme is a 12-h relaxation to the analyses' horizontal momentum before the forecast integration begins, and is called the dynamic initialization (DI) scheme. The third scheme is a combination of the previous two schemes, and is called the two-stage dynamic initialization scheme (TCDI/DI). In the first stage, TCDI is implemented in order to improve the representation of the TC vortex. In the second stage, DI is invoked in order to improve the balance between the inserted TC vortex and its environment. All three dynamic initialization schemes are compared with a control (CNTL) scheme, which creates the initial vortex using synthetic TC observations that match the observed intensity and structure in a three-dimensional variational data assimilation (3DVAR) system. The four schemes are tested on 120 cases in the North Atlantic and western North Pacific basins during 2010 and 2011 using the Naval Research Laboratory's TC prediction model: Coupled Ocean–Atmosphere Mesoscale Prediction System-Tropical Cyclones (COAMPS-TC). It is demonstrated that TCDI/DI performed the best overall with regard to intensity forecasts, reducing the average minimum central pressure error for all lead times by 24.4% compared to the CNTL scheme.

Extended Prediction of North Indian Ocean Tropical Cyclones

2012 ◽  
Vol 27 (3) ◽  
pp. 757-769 ◽  
Author(s):  
James I. Belanger ◽  
Peter J. Webster ◽  
Judith A. Curry ◽  
Mark T. Jelinek
Keyword(s):  
Tropical Cyclone ◽  
Indian Ocean ◽  
Tropical Cyclones ◽  
North Indian Ocean ◽  
Ensemble Prediction ◽  
Tropical Cyclone Genesis ◽  
Lead Times ◽  
Prediction System ◽  
Ensemble Prediction System ◽  
The North

Abstract This analysis examines the predictability of several key forecasting parameters using the ECMWF Variable Ensemble Prediction System (VarEPS) for tropical cyclones (TCs) in the North Indian Ocean (NIO) including tropical cyclone genesis, pregenesis and postgenesis track and intensity projections, and regional outlooks of tropical cyclone activity for the Arabian Sea and the Bay of Bengal. Based on the evaluation period from 2007 to 2010, the VarEPS TC genesis forecasts demonstrate low false-alarm rates and moderate to high probabilities of detection for lead times of 1–7 days. In addition, VarEPS pregenesis track forecasts on average perform better than VarEPS postgenesis forecasts through 120 h and feature a total track error growth of 41 n mi day−1. VarEPS provides superior postgenesis track forecasts for lead times greater than 12 h compared to other models, including the Met Office global model (UKMET), the Navy Operational Global Atmospheric Prediction System (NOGAPS), and the Global Forecasting System (GFS), and slightly lower track errors than the Joint Typhoon Warning Center. This paper concludes with a discussion of how VarEPS can provide much of this extended predictability within a probabilistic framework for the region.

Naval Research Laboratory Multiscale Targeting Guidance for T-PARC and TCS-08

2010 ◽  
Vol 25 (2) ◽  
pp. 526-544 ◽  
Author(s):  
Carolyn A. Reynolds ◽  
James D. Doyle ◽  
Richard M. Hodur ◽  
Hao Jin
Keyword(s):  
Tropical Cyclone ◽  
Research Laboratory ◽  
Large Scale ◽  
Adjoint System ◽  
Horizontal Resolution ◽  
Naval Research Laboratory ◽  
Naval Research ◽  
Prediction System ◽  
Initial State ◽  
The North

Abstract As part of The Observing System Research and Predictability Experiment (THORPEX) Pacific Asian Regional Campaign (T-PARC) and the Office of Naval Research’s (ONR’s) Tropical Cyclone Structure-08 (TCS-08) experiments, a variety of real-time products were produced at the Naval Research Laboratory during the field campaign that took place from August through early October 2008. In support of the targeted observing objective, large-scale targeting guidance was produced twice daily using singular vectors (SVs) from the Navy Operational Global Atmospheric Prediction System (NOGAPS). These SVs were optimized for fixed regions centered over Guam, Taiwan, Japan, and two regions over the North Pacific east of Japan. During high-interest periods, flow-dependent SVs were also produced. In addition, global ensemble forecasts were produced and were useful for examining the potential downstream impacts of extratropical transitions. For mesoscale models, TC forecasts were produced using a new version of the Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS) developed specifically for tropical cyclone prediction (COAMPS-TC). In addition to the COAMPS-TC forecasts, mesoscale targeted observing products were produced using the COAMPS forecast and adjoint system twice daily, centered on storms of interest, at a 40-km horizontal resolution. These products were produced with 24-, 36-, and 48-h lead times. The nonhydrostatic adjoint system used during T-PARC/TCS-08 contains an exact adjoint to the explicit microphysics. An adaptive response function region was used to target favorable areas for tropical cyclone formation and development. Results indicate that forecasts of tropical cyclones in the western Pacific are very sensitive to the initial state.

Tropical Cyclone Formation Guidance Using Pregenesis Dvorak Climatology. Part I: Operational Forecasting and Predictive Potential

2013 ◽  
Vol 28 (1) ◽  
pp. 100-118 ◽  
Author(s):  
Joshua H. Cossuth ◽  
Richard D. Knabb ◽  
Daniel P. Brown ◽  
Robert E. Hart
Keyword(s):  
Tropical Cyclone ◽  
Tropical Cyclogenesis ◽  
Lead Times ◽  
Eastern Region ◽  
Regional Variability ◽  
Central Pacific ◽  
The North ◽  
Predictive Skill ◽  
Out Of Sample ◽  
Frequency Maximum

Abstract While there are a variety of modes for tropical cyclone (TC) development, there have been relatively few efforts to systematically catalog both nondeveloping and developing cases. This paper introduces an operationally derived climatology of tropical disturbances that were analyzed using the Dvorak technique at the National Hurricane Center (NHC) and the Central Pacific Hurricane Center from 2001 to 2011. Using these Dvorak intensity estimates, the likelihood of genesis is calculated as a historical baseline for TC prediction. Despite the limited period of record, the climatology of Dvorak analyses of incipient tropical systems has a spatial distribution that compares well with previous climatologies. The North Atlantic basin shows substantial regional variability in Dvorak classification frequency. In contrast, tropical disturbances in the combined eastern and central North Pacific basins (which split at 125°W into an eastern region and a central region) have a single broad frequency maximum and limited meridional extent. When applied to forecasting, several important features are discovered. Dvorak fixes are sometimes unavailable for disturbances that develop into TCs, especially at longer lead times. However, when probabilities of genesis are calculated by a Dvorak current intensity (CI) number, the likelihood stratifies well by basin and intensity. Tropical disturbances that are analyzed as being stronger (a higher Dvorak CI number) achieve genesis more often. Further, all else being equal, genesis rates are highest in the eastern Pacific, followed by the Atlantic. Out-of-sample verification of predictive skill shows comparable results to that of the NHC, with potential to inform forecasts and provide the first disturbance-centric baseline for tropical cyclogenesis potential.

A 3DVAR-Based Dynamical Initialization Scheme for Tropical Cyclone Predictions*

2012 ◽  
Vol 27 (2) ◽  
pp. 473-483 ◽  
Author(s):  
Shengjun Zhang ◽  
Tim Li ◽  
Xuyang Ge ◽  
Melinda Peng ◽  
Ning Pan
Keyword(s):  
Tropical Cyclone ◽  
Data Assimilation ◽  
Global Analysis ◽  
Three Dimensional ◽  
Preliminary Test ◽  
Intensity Change ◽  
Observation System ◽  
Wind Fields ◽  
Warm Core ◽  
Initialization Scheme

Abstract A combined tropical cyclone dynamic initialization–three-dimensional variational data assimilation scheme (TCDI–3DVAR) is proposed. The specific procedure for the new initialization scheme is described as follows. First, a first-guess vortex field derived from a global analysis will be spun up in a full-physics mesoscale regional model in a quiescent environment. During the spinup period, the weak vortex is forced toward the observed central minimum sea level pressure (MSLP). The so-generated balanced TC vortex with realistic MSLP and a warm core is then merged into the environmental field and used in the subsequent 3DVAR data assimilation. The observation system simulation experiments (OSSEs) demonstrate that this new TC initialization scheme leads to much improved initial MSLP, warm core, and asymmetric temperature patterns compared to those from the conventional 3DVAR scheme. Forecasts of TC intensity with the new initialization scheme are made, and the results show that the new scheme is able to predict the “observed” TC intensity change, compared to runs with the conventional 3DVAR scheme or the TCDI-only scheme. Sensitivity experiments further show that the intensity forecasts with knowledge of the initial MSLP and wind fields appear more skillful than do the cases where the initial MSLP, temperature, and humidity fields are known. The numerical experiments above demonstrate the potential usefulness of the proposed new initialization scheme in operational applications. A preliminary test of this scheme with a navy operational model shows encouraging results.

Surface Track Discontinuity of Tropical Cyclones Crossing Taiwan: A Statistical Study

2012 ◽  
Vol 140 (1) ◽  
pp. 121-139 ◽  
Author(s):  
L. Peng ◽  
S.-T. Wang ◽  
S.-L. Shieh ◽  
M.-D. Cheng ◽  
T.-C. Yeh
Keyword(s):  
Tropical Cyclone ◽  
Tropical Cyclones ◽  
Quadratic Function ◽  
Training Sample ◽  
Equation Model ◽  
Mountain Range ◽  
Accuracy Score ◽  
The North ◽  
Error Range ◽  
Rms Error

Abstract Surface tracks of some cross-Taiwan tropical cyclones were discontinuous as a result of the blockage of the north-northeast–south-southwest-oriented Central Mountain Range (CMR). This paper tries to identify the variables that may be used to diagnose track continuity in advance. The track records of 131 westbound cross-Taiwan tropical cyclones between 1897 and 2009 are examined. It is found that the track continuity of a westbound cross-Taiwan tropical cyclone depends mostly upon the landfall location (YLF), the approaching direction (ANG), and the maximum wind (VMX) of the cyclone. According to the empirical probability of track continuity estimated from the data, the dependence on YLF, which is nonlinear and remarkably asymmetric with respect to the midpoint of the east coast, may be well approximated by a quadratic function of YLF. The nonlinearity and asymmetry can be interpreted in terms of the length scale of the CMR and the north–south antisymmetry of the cyclonic flow. The estimated dependence of track continuity on cyclone intensity and size may be approximated by a linear function of VMX. The estimated dependence of track continuity on ANG may be approximated by a single term of the modified variable DIR (=|ANG − 110|, where 110 is the direction, in degrees, perpendicular to the CMR’s long axis). Using the 64 tracks between 1944 and 1996 as the training sample, a logistic regression equation model, built in terms of YLF, YLF square, DIR, and VMX gives an overall accuracy score of 89%. As to the probability estimates of individual tracks, 49 of the 64 tracks have estimated probabilities outside the (0.5 − 0.127, 0.5 + 0.127) RMS error range and are correctly classified. A prediction test using another set of 67 tracks not included in the model-training sample, scores a success rate of 82%. As to the probability predictions for individual tracks, 49 of the 67 tracks have predicted probabilities outside the RMS error range and are correctly predicted. These results confirm the appropriateness of the model and moreover demonstrate that the three parameters, YLF, DIR, and VMX, primarily control the surface track continuity of a westbound tropical cyclone crossing Taiwan.

Performance of a Dynamic Initialization Scheme in the Coupled Ocean–Atmosphere Mesoscale Prediction System for Tropical Cyclones (COAMPS-TC)

2011 ◽  
Vol 26 (5) ◽  
pp. 650-663 ◽  
Author(s):  
Eric A. Hendricks ◽  
Melinda S. Peng ◽  
Xuyang Ge ◽  
Tim Li
Keyword(s):  
Tropical Cyclone ◽  
Tropical Cyclones ◽  
Initial Conditions ◽  
Layer Structure ◽  
North Pacific Ocean ◽  
Surface Wind ◽  
Naval Research ◽  
Prediction System ◽  
Ocean Atmosphere ◽  
Initialization Scheme

Abstract A dynamic initialization scheme for tropical cyclone structure and intensity in numerical prediction systems is described and tested. The procedure involves the removal of the analyzed vortex and, then, insertion of a new vortex that is dynamically initialized to the observed surface pressure into the numerical model initial conditions. This new vortex has the potential to be more balanced, and to have a more realistic boundary layer structure than by adding synthetic data in the data assimilation procedure to initialize the tropical cyclone in a model. The dynamic initialization scheme was tested on multiple tropical cyclones during 2008 and 2009 in the North Atlantic and western North Pacific Ocean basins using the Naval Research Laboratory’s tropical cyclone version of the Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS-TC). The use of this initialization procedure yielded significant improvements in intensity forecasts, with no degradation in track performance. Mean absolute errors in the maximum sustained surface wind were reduced by approximately 5 kt for all lead times up to 72 h.

Tropical Cyclone Energy Dispersion in a Three-Dimensional Primitive Equation Model: Upper-Tropospheric Influence*

2008 ◽  
Vol 65 (7) ◽  
pp. 2272-2289 ◽  
Author(s):  
Xuyang Ge ◽  
Tim Li ◽  
Yuqing Wang ◽  
Melinda S. Peng
Keyword(s):  
Tropical Cyclone ◽  
Rossby Wave ◽  
Wave Energy ◽  
Wave Train ◽  
Three Dimensional ◽  
Equation Model ◽  
Energy Dispersion ◽  
Primitive Equation ◽  
Primitive Equation Model ◽  
Rossby Wave Train

Abstract The three-dimensional (3D) Rossby wave energy dispersion of a tropical cyclone (TC) is studied using a baroclinic primitive equation model. The model is initialized with a symmetric vortex on a beta plane in an environment at rest. The vortex intensifies while becoming asymmetric and moving northwestward because of the beta effect. A synoptic-scale wave train forms in its wake a few days later. The energy-dispersion-induced Rossby wave train has a noticeable baroclinic structure with alternating cyclonic–anticyclonic–cyclonic (anticyclonic–cyclonic–anticyclonic) circulations in the lower (upper) troposphere. A key feature associated with the 3D wave train development is a downward propagation of the relative vorticity and kinetic energy. Because of the vertical differential inertial stability, the upper-level wave train develops faster than the lower-level counterpart. The upper anticyclonic circulation rapidly induces an intense asymmetric outflow jet in the southeast quadrant, and then further influences the lower-level Rossby wave train. On one hand, the outflow jet exerts an indirect effect on the lower-level wave train strength through changing TC intensity and structure. On the other hand, it triggers downward energy propagation that further enhances the lower-level Rossby wave train. A sudden removal of the diabatic heating may initially accelerate the energy dispersion through the increase of the radius of maximum wind and the reduction of the lower-level inflow. The latter may modulate the group velocity of the Rossby wave train through the Doppler shift effect. The 3D numerical results illustrate more complicated Rossby wave energy dispersion characteristics than 2D barotropic dynamics.

The Naval Research Laboratory’s Coupled Ocean/Atmosphere Mesoscale Prediction System – Tropical Cyclone ensemble (COAMPS-TC ensemble)

2021 ◽  
Author(s):  
William A. Komaromi ◽  
Patrick A. Reinecke ◽  
James D. Doyle ◽  
Jonathan R. Moskaitis
Keyword(s):  
Tropical Cyclone ◽  
Boundary Conditions ◽  
Skill Score ◽  
Horizontal Resolution ◽  
Naval Research ◽  
Lead Times ◽  
Prediction System ◽  
Intensity Forecast ◽  
Probabilistic Forecasts ◽  
Ocean Atmosphere

AbstractThe 11-member Coupled Ocean/Atmosphere Mesoscale Prediction System for Tropical Cyclones (COAMPS-TC) ensemble has been developed by the Naval Research Laboratory (NRL) to produce probabilistic forecasts of tropical cyclone (TC) track, intensity and structure. All members run with a storm-following inner grid at convection-permitting 4-km horizontal resolution. The COAMPS-TC ensemble is constructed via a combination of perturbations to initial and boundary conditions, the initial vortex, and model physics to account for a variety of different sources of uncertainty that affect track and intensity forecasts. Unlike global model ensembles, which do a reasonable job capturing track uncertainty but not intensity, mesoscale ensembles such as the COAMPS-TC ensemble are necessary to provide a realistic intensity forecast spectrum.The initial and boundary condition perturbations are responsible for generating the majority of track spread at all lead times, as well as the intensity spread from 36-120 h. The vortex and physics perturbations are necessary to produce meaningful spread in the intensity prediction from 0-36 h. In a large sample of forecasts from 2014-2017, the ensemble-mean track and intensity forecast is superior to the unperturbed control forecast at all lead times, demonstrating a clear advantage to running an ensemble versus a deterministic forecast. The spread-skill relationship of the ensemble is also examined, and is found to be very well calibrated for track, but is under-dispersive for intensity. Using a mixture of lateral boundary conditions derived from different global models is found to improve upon the spread-skill score for intensity, but it is hypothesized that additional physics perturbations will be necessary to achieve realistic ensemble spread.

scholarly journals Evaluation of Tropical Cyclone Structure Forecasts in a High-Resolution Version of the Multiscale GFDL fvGFS Model

2018 ◽  
Vol 33 (2) ◽  
pp. 419-442 ◽  
Author(s):  
Andrew T. Hazelton ◽  
Lucas Harris ◽  
Shian-Jiann Lin
Keyword(s):  
High Resolution ◽  
Tropical Cyclone ◽  
Doppler Radar ◽  
Three Dimensional ◽  
Research Division ◽  
The North ◽  
Separate Model ◽  
High Resolution Structure ◽  
Cubed Sphere ◽  
Structural Metrics

Abstract A nested version of the cubed-sphere finite-volume dynamical core (FV3) with GFS physics (fvGFS) is capable of tropical cyclone (TC) prediction across multiple space and time scales, from subseasonal prediction to high-resolution structure and intensity forecasting. Here, a version of fvGFS with 2-km resolution covering most of the North Atlantic is evaluated for its ability to simulate TC track, intensity, and finescale structure. TC structure is evaluated through a comparison of forecasts with three-dimensional Doppler radar from P-3 flights by NOAA’s Hurricane Research Division (HRD), and the structural metrics evaluated include the 2-km radius of maximum wind (RMW), slope of the RMW, depth of the TC vortex, and horizontal vortex decay rate. Seven TCs from the 2010–16 seasons are evaluated, including 10 separate model runs and 38 individual flights. The model had some success in producing rapid intensification (RI) forecasts for Earl, Edouard, and Matthew. The fvGFS model successfully predicts RMWs in the 25–50-km range but tends to have a small bias at very large radii and a large bias at very small radii. The wind peak also tends to be somewhat too sharp, and the vortex depth occasionally has a high bias, especially for storms that are observed to be shallow. Composite radial wind shows that the boundary layer tends to be too deep, although the outflow structure aloft is relatively consistent with observations. These results highlight the utility of the structural evaluation of TC forecasts and also show the promise of fvGFS for forecasting TCs.

Sign in / Sign up

Export Citation Format

Share Document