Abstract

A model of the tangential wind speed in the outer regions of tropical cyclones is proposed based on approximate conservation of angular momentum. The purpose is to derive an operationally useful model of the beta-effect propagation (BEP), which barotropic numerical models have shown to be primarily related to the outer wind structure. The functional dependence of the predicted BEP speeds over a range of latitudes and the radii R _o at which the cyclonic winds are reduced to zero is determined from dimensional analysis. Given the empirical nature of the profile and imprecise estimates of R _o , only four tropical cyclone sizes or outer wind structures are defined based on the magnitude of the BEP speeds and their relative contributions to the total motion. A second aspect of BEP from the barotropic model integrations is the development of a trailing anticyclone to the southeast of the tropical cyclone as a result of Rossby wave dispersion. The four tropical cyclone size categories or outer wind structures are also characterized in terms of the potential for the trailing anticyclone to be part of a change in the tropical cyclone environment structure. Whereas small tropical cyclones have a small BEP speed and do not change their environment, large tropical cyclones have BEP speeds exceeding 2 m s⁻¹ and have a large amplitude peripheral anticyclone that may introduce significant changes in the environment structure, which can indirectly contribute to a track change.

Abstract

Objective criteria for distinguishing among three modes of binary tropical cyclone (TC) interactions (TCI) are defined and tested with an eight-year sample of western North Pacific TCs. These objective criteria lead to an approximate 4:1 ratio of hits to misses of the subjectively assigned TCIs. False alarm rates of approximately 20% are also found, which indicate that similar track deflections may occur for other reasons, and the objective identifications must still be carefully validated by a forecaster.

Abstract

The relationships between consensus spread of five dynamical model tracks and the consensus mean error is explored for a western North Pacific tropical cyclone database of 381 cases. Whereas a small spread of the five tracks is often indicative of a small consensus track error, some cases with large errors also are found even though the consensus spread is small. Some of the success of the dynamical model consensus approach arises because a substantial number (21%) of the cases with a large consensus spread have reduced errors after the consensus averaging. In nearly all the cases in this sample, the best of the five models has a 72-h track error of less than 300 n mi, but no tools are available to allow the forecaster to always select this best model. It is demonstrated that the forecaster can also add value by forming a selective consensus after first eliminating one or more likely erroneous track(s) and averaging the remaining tracks. Conceptual models and symptoms in the predicted fields to assist the forecaster in this error detection have been separately described by the authors, and their successful application would result in more accurate selective consensus forecasts than nonselective consensus forecasts.

Abstract

Sudden poleward track changes of tropical cyclones embedded in monsoon gyres in the western North Pacific are documented. During these track changes, which are generally not well forecast, the cyclones are often accompanied by a separate comma-shaped area of gale-force winds and deep convection along the eastern periphery. This monsoon surge is distinct from the tropical cyclone. Synoptic analyses often reveal a building anticyclone to the east or southeast of the monsoon gyre. The hypothesis that the sudden track change is initiated by a binary interaction of the tropical cyclone and monsoon gyre is tested with a nondivergent barotropic model. Tropical cyclone-scale vortices with initial positions within the eastern semicircle of a larger monsoon gyre-scale vortex initially coalesce with the monsoon gyre and then exhibit sudden poleward track changes that are similar to the observations. During the coalescence phase, the large and relatively weak monsoon gyre undergoes a β-induced dispersion in which nonlinear vorticity advection also plays an important role. This dispersion process produces strong ridging to the east and southeast of the coalesced tropical cyclone and monsoon gyre. An intermediate region of high winds that resembles the observed monsoon surge develops between the monsoon gyre and the peripheral ridging. A southerly steering develops across the coalescing tropical cyclone and monsoon gyre and causes the poleward acceleration. Key features of the simulated streamfunction and isotach patterns associated with the sudden track changes are substantiated with synoptic analyses of observed casts with similar track changes. Thus, it is concluded that the tropical cyclone-monsoon gyre interactions are a likely explanation for monsoon surge track changes and that the observed phenomena may be explained to first order by conservation of absolute vorticity on a β plane.

Abstract

The flow field around a tropical cyclone can be decomposed into the symmetric vortex, an environmental steering flow, and an asymmetric circulation that is the residual after the first two components are removed. In a barotropic model with a quiescent environment, the asymmetric circulation is associated with a propagation vector that is defined to be the difference between the storm-motion vector and the steering flow. An analytical specification derived by Carr for the asymmetric circulation and the corresponding propagation vector is proposed for use in the initial conditions of dynamical track-prediction models.

One widely used method for estimating the environmental steering flow has been to average the wind components in radial bands centered on the storm position. It is shown that the portion of the asymmetric circulation that is included within these radial band averages results in a systematic rotation of the propagation vectors defined relative to the steering flow, which is quite similar to that calculated by compositing observations. Thus, some method other than the radial band-average estimate of the steering will be required in the wind-field decomposition to separate the lame-scale environment from the influence of the tropical cyclone.

It is demonstrated that inclusion of the analytically generated asymmetric circulation in the initial conditions of a barotropic track-prediction model leads to a propagation vector with the correct speed and direction. It is proposed that inclusion of the asymmetric circulation would eliminate much of the initial slow bias in dynamical track-prediction models that include only the symmetric vortex circulation and thus improve the tropical cyclone track forecasts.

Abstract

All highly erroneous (>300 n mi or 555 km at 72 h) Navy Operational Global Atmospheric Prediction System (NOGAPS) and U.S. Navy version of the Geophysical Fluid Dynamics Laboratory model (GFDN) tropical cyclone track forecasts in the western North Pacific during 1997 are examined. Responsible error mechanisms are described by conceptual models that are all related to known tropical cyclone motion processes that are being misrepresented in the dynamical models. Error mechanisms that predominantly occur while the tropical cyclone is still in the Tropics are described in this paper, and those errors that are more related to midlatitude circulations are addressed in a companion paper. Of the 69 NOGAPS large-error cases, 39 were attributed to excessive direct cyclone interaction (E-DCI), 12 cases of excessive ridge modification by the tropical cyclone (E-RMT), and 10 cases of excessive reverse trough formation (E-RTF). Of the 50 GFDN large-error cases, 31 were E-DCI, and only two E-RMT and two E-RTF cases were found, but 9 cases involving a single cyclone were attributed to excessive tropical cyclone size (E-TCS). Characteristics and symptoms in the forecast tracks and model fields that accompany these frequently occurring error mechanisms are documented and illustrative case studies are presented. When a sudden deviation from previous track guidance or a track outlier from the other dynamical model guidance appears, the forecaster should diagnose whether this is an error, or is indicative of a real track change. If the conceptual models of large-error mechanisms proposed from this retrospective study can be applied in real time, track forecasting will be improved.

Abstract

All highly erroneous (>300 n mi or 555 km at 72 h) Navy Operational Global Atmospheric Prediction System (NOGAPS) and U.S. Navy version of the Geophysical Fluid Dynamics Laboratory model (GFDN) tropical cyclone track forecasts in the western North Pacific during 1997 are examined. Error mechanisms that are more related to midlatitude circulations are described in this paper and those errors that predominantly occur while the tropical cyclone (TC) is still in the Tropics are addressed in a companion paper. Responsible error mechanisms are described by conceptual models that are all related to known tropical cyclone motion processes that are being misrepresented in the dynamical models. As in the companion paper, characteristics and symptoms in the forecast tracks and model fields that accompany these frequently recurring error mechanisms are documented and illustrative case studies are presented. Whereas 21 GFDN forecasts were degraded by an improper prediction of a midlatitude system evolution, only seven NOGAPS track forecasts had 72-h errors exceeding 300 n mi. The NOGAPS model is more prone to excessive responses to vertical shear, and a useful indicator is that the 500-mb circulation of the TC becomes noticeably displaced downshear of the low-level center in the 48–72-h forecast fields. Both the NOGAPS and GFDN forecasts are degraded by the baroclinic cyclone interaction, and none of the five dynamical models provided consistently accurate or time-consistent TC track forecasts in these situations. These midlatitude cases are more difficult than the tropical errors to diagnose as multiple error sources may be present.

When a sudden deviation from previous track guidance or a track outlier from the other dynamical model guidance appears, the forecaster should diagnose whether this is an error, or is indicative of a real track change. If the conceptual models of large-error mechanisms proposed from this retrospective study can be applied in real time, track forecasting will be improved.

Abstract

The authors have developed error mechanism conceptual models with characteristic track departures and anomalous wind or sea level pressure patterns for dynamical tropical cyclone track predictions primarily occurring in tropical regions or those associated with midlatitude circulation patterns. These conceptual models were based on a retrospective study in which it was known that the 72-h track error exceeded 300 n mi (555 km). A knowledge-based expert system module named the Systematic Approach Forecast Aid (SAFA) has been developed to assist the forecaster in the information management, visualization, and proactive investigation of the frequently occurring error mechanisms.

A beta test of the SAFA module was carried out for all available track forecasts for the western North Pacific cyclones 19W–30W during 1999. The objective was to determine if the SAFA module could guide the team to apply the conceptual models in a real-time scenario to detect dynamical model tracks likely to have 72-h errors greater than 300 n mi. The metric was a selective consensus from the remaining model tracks that had smaller errors than the nonselective consensus track of all dynamical model tracks.

This beta test demonstrated that the prototype SAFA module with the large-error mechanism conceptual models could be effectively applied in real-time conditions. The beta-test team recognized 14 cases in which elimination of one or more dynamical model forecasts before calculating the consensus track resulted in a 10% improvement over the nonselective consensus track. A number of lessons learned from the beta test are described, including that rejecting a specific model track is not normally successful if the tracks are tightly clustered, that the availability of the model-predicted fields is critical to the error detection, and that at least three of the five model tracks need to be available.

Abstract

An observational study of western North Pacific tropical cyclones (TC) revealed many cases of two TCs whose tracks were altered by processes that were quite different from the mutual advection (Fujiwhara-type) processes. Thus, four conceptual models are proposed to describe these track alterations. A conceptual model called direct interaction is proposed that is a modification of one by Lander and Holland and has three modes: 1) a one-way influence in which the track of a smaller TC that is embedded in the circulation of a larger TC has a cyclonic orbiting motion, but no significant track alteration of the larger TC is apparent; 2) a similar case in which a mutual advection occurs with the tracks of both the smaller and larger TCs being altered; and 3) a subset of 2) in which the mutual advection includes an attraction component such that the two similarly sized TC circulations eventually merge into a larger circulation with a single center. During the 7-yr period (1989–95), the one-way influence, mutual interaction, and merger modes were detected seven, three, and two times.

A semidirect interaction conceptual model is proposed in which the two TCs have a relative cyclonic rotation as in the Lander and Holland model, but the TCs are separated by 10°–20° longitude so that a direct (advective-type) interaction is excluded. Rather, the track alteration is attributed to an environmental flow established by the juxtaposition of a TC on one side and a subtropical anticyclone cell on the opposite side. In an east–west orientation of the two TCs and a subtropical anticyclone cell to the east (west), the height gradient between the western (eastern) TC and the eastern (western) subtropical anticyclone establishes a poleward (equatorward) environmental steering flow across the eastern (western) TC. In the 1989–95 sample, a semidirect interaction that altered the tracks of the eastern or the western TC occurred 18 and 14 times, respectively.

An indirect interaction conceptual model is proposed in which the distinguishing feature is the Rossby wave dispersion-induced anticyclone to the east and equatorward of the western TC. This anticyclone imposes an equatorward (poleward) steering flow across the eastern (western) TC. Several variations of the indirect interaction are possible depending on the separation distance, sizes of each TC, and their relative orientation. During the 7-yr period, an indirect interaction affecting the western TC or the eastern TC occurred 36 and 22 times, respectively.

A fourth conceptual model of track alterations involving two TCs is proposed in conjunction with a reverse-oriented monsoon trough formation. The distinguishing feature of this conceptual model is the combination of the peripheral anticyclones of both TCs as the eastern TC moves into an east–west orientation and has a separation of 10°–20° longitude. In the 1989–95 sample, a reverse-oriented trough formation involving two TCs occurred seven times.

The frequency of track alterations whenever two TCs are present emphasizes that forecasters must give special attention to such situations. The four conceptual models proposed here emphasize that the physical mechanisms are complex and in the vast majority of cases cannot be attributed to the mutual advection (Fujiwhara-type) process implied in the Lander and Holland model.

Abstract

A simple statistical-synoptic technique for tropical cyclone (TC) track forecasting to 72 h in the western North Pacific is derived. This technique applies to the standard (S) pattern/dominant ridge region (S/DR) and poleward/poleward-oriented (P/PO) combinations, which are the two most common and represent about 73% of all situations. Only eight predictors that involve present and past 12-h and 24-h positions, intensity, and date are used. The track predictions are simple to calculate and understand; are available in near–real time each 6 h; apply at all intensities, as compared to the complex global or regional dynamical model predictions that are only available each 12 h at about 3–4 h after synoptic time; are not calculated for weak TCs; and tend to have accurate predictions only for tropical storm stage and above. The statistical-synoptic technique for S/DR cases has an improvement (skill) relative to the operational climatology and persistence (WPCLPR) technique of 12% after 12 h and 24% after 72 h if the TC remains in the S/DR pattern/region for the entire 72 h. The statistical-synoptic technique for P/PO cases have an improvement relative to WPCLPR of 11% after 12 h and about 13% for 72-h forecasts if the TC remains in P/PO for the entire 72 h.

Assuming a perfect knowledge of the S/DR to P/PO and P/PO to S/DR transitions, a simple blending of a composite post-transition track with the statistical-synoptic technique is tested. For the 72-h forecasts initiated 12 h before the S/DR to P/PO transition, the statistical-synoptic track error is about 290 n mi (537 km) compared to 410 n mi (759 km) for WPCLPR. For corresponding P/PO to S/DR transition, the statistical-synoptic technique 72-h error is 215 n mi (398 km) compared to about 485 n mi (898 km) for WPCLPR.