Topic 1.6: TC Structure Analysis and Forecasting Techniques

Rapporteur: Edwin S.T. Lai
Hong Kong Observatory
134A Nathan Road
Hong Kong

E-mail: stlai@hko.gov.hk
Fax: 852 2375 2645

Working Group: K. Brueske, K. Caesar, J. Callaghan, R. Edson, B. Hanstrum,
W.K. Leong, D. Li, M. Nagata, T. Peng, A. Waqaicelua.


Statistical dynamical methods remain the mainstay in structural analysis and hence intensity assessment at various operational centres, with reported gains in performance after using more realistic SST (sea-surface temperature) analyses and OHC (ocean heat content) as predictors. Techniques using neural networks are also being explored in some countries with a view to possible operational implementation. More new data types from increasingly sophisticated satellite sensors have also begun to make an impact for nowcasting and NWP (Numerical Weather Prediction) applications.

Inner-core structural dynamics and maximum potential intensity (MPI) have attracted the most research attention in recent years. Core vortex dynamical adjustments suggest that a feedback loop exists with the convection and the larger scale environment. Asymmetric cloud distribution is attributed to be the reason for vortex structure failing to attain its MPI. While numerical experiments have shown that intensity and structural details are sensitive to initial conditions, model resolution as well as SST, enhanced observations within the tropical cyclone (TC) have also contributed positively to NWP intensity forecasts.

Further works need to be done in translating the knowledge gained into conceptual models or operational guidelines for forecasters’ assessment of TC intensity changes based on real-time information. Careful considerations should be given to the formulation of a concerted global or regional effort in the definition, collection, analysis and archival of multi-faceted intensity parameters for a systematic study of TC structure.

1.6.1: Overview

In the IWTC-IV held in April 1998 at Haikou, Hainan Island, China, the following recommendations were made in connection with this sub-topic:

(a) For WMO: Operationally, it was considered desirable to have peak gust, minimum pressure and time of eye passage included in the SYNOP code. For validaton purposes, there were also suggestions on standardization of wind-averaging times and compilation of error statistics of Dvorak technique and pressure-wind relationships. To establish a quality database, WMO was asked to draw up guidelines to encourage TC archiving centres to incorporate size (e.g., gale radius) and radius of maximum winds in their best track information.

(b) For Operational Centres: They were encouraged to make better use of statistical dynamical models for short-term intensity forecasts, to introduce thermodynamically derived potential intensity guidance for future intensity or peak intensity assessment, and to incorporate real-time ocean data for general intensity forecast improvement.

(c) For Research Community: On intensity analysis, alternative diagnostic tools to Dvorak technique should be explored. On significant intensity changes, knowledge-based methodologies for trend forecast should be developed. On structural dynamics, better understanding on various issues such as inner-core mechanisms, vortex-induced waves, the role of convection and spiral rainbands evolution, and formation of polygonal and outer eyewalls should be acquired. For facilitation of concerted research initiatives, analysis schemes should be designed to accommodate meaningfully various data types onto a common framework for numerical modelling efforts. There were also suggestions to attempt a re-analysis of archived TC data and to make available on the public domain a fully tested parametric wind model as a standard for comparison of results.

This report covers as far as possible the notable progress that has been made four years down the line, and the outstanding issues that remain unresolved. The review is presented in 1.6.2 – 1.6.4 organized in a way that generally follows the above sub-headings, with a summary in 1.6.5. While structural analysis and intensity forecasting are intricately linked with many other aspects in TC studies, more specialized areas such as environmental effects and interface, boundary and convective processes will be covered in detail by other sub-topics under Topic 1. They will nonetheless be mentioned in passing for ease of discussion or cross reference, and also if members of this working group have made specific reference to such issues in the context of this sub-topic.

1.6.2: WMO and Coordination at International Level

Scanning through recently published WMO technical documents (TD), direct reference to the current topic is few and far between. TD-No.966 on ‘TC-related NWP products and their guidance’ and TD-No.975 on ‘Estimating the amount of rainfall associated with TCs using satellite techniques’, both published in 1999, may be considered to be of some indirect relevance, particularly in view of the fact that NWP and satellites will no doubt play increasingly important roles in TC structure analysis and forecasting. It is noted that many of the issues left over from the last IWTC will be on the agenda of the forthcoming ‘4th TC RSMCs Technical Coordination Meeting’ to be held in Nadi, Fiji ahead of IWTC-V. Presumably, outcome of the Fijian discussion will be reported in the Cairns gathering.

One related issue of potential relevance concerns the planned migration towards the more flexible BUFR (Binary Universal Form for the Representation of meteorological data)and CREX (Character form for the Representation and Exchange of data)codes in the coming years, which renders the proposed inclusion of additional TC intensity and structure details in international information exchanges a more attainable target in the foreseeable future.

One recurring issue in many WMO meetings (e.g., WMO 2002) relates to storm surges and their disastrous impact in certain TC-infested basins. Although this can be a self-contained topic, improved TC structural analyses (also see Rapporteur Report on Topic 2.5 ‘Storm Surge’ by Professor S. Dube) and forecasting based on knowledge gained and techniques developed in various aspects under Topic 1 will no doubt contribute positively to the mitigation of such catastrophes.

1.6.3 Operational Development in Warning Centres

As suggested in the last IWTC, statistical dynamical methods remain the mainstay in structural analysis and hence intensity assessment. One major reported improvement is the successful implementation and enhanced performance of SHIPS (Statistical Hurricane Intensity Prediction Scheme) in the Atlantic and eastern Pacific basins as the scheme evolves towards a statistical dynamical version with inclusion of weekly SST analyses (DeMaria and Kaplan 1999). It is reported in Elsberry (2002) that recent effort to include satellite-derived monthly ocean heat content (OHC) distribution as an additional predictor has led to an improvement in SHIPS forecasts by 3-5%. Meanwhile, Petty and Hobgood (2000) also reports favourable results from the EPIC (Eastern Pacific Intensity Change) scheme that is based on time tendencies of wind shear and thermal variables.

In Australia, intensity change forecasts are closely linked with synoptic assessment of cyclone interaction with upper-air troughs. Similar considerations are adopted in Fiji through the use of three progressive checklists as the cyclone evolves. It is noted in Fiji that terrain modulation by seemingly benign islands is also an important consideration in cyclone intensity changes.

Other advances with operational implications relate mainly to the increasing use of satellite information. Apart from the development towards automated and enhanced Dvorak techniques (Velden et al. 1998), new data types such as SSM/I are also being incorporated into the Dvorak intensity estimates (Cocks et al. 1999) at JTWC (Joint Typhoon Warning Center). In general, as more satellite-based information becomes widely available, efforts to incorporate data such as scatterometer-derived sea-surface winds and AMSU (Advanced Microwave Sounding Unit) radiance into operational analyses (e.g., Brueske and Velden 2003) continue to gather momentum. It forces the issue on whether the conventional Dvorak approach can be adapted to adequately make use of the diversity of new data and emerging techniques.

At JMA, the usefulness of QuikSCAT winds in defining the radius of 30-kt winds has already found its way into the initial NWP analysis through a more realistic construction of the bogus vortex. In Hong Kong, near real-time analyses of 3D wind fields incorporating QuikSCAT, radar TREC (Tracking of Radar Echoes by Correlation) and Doppler winds, AMDAR (Aircraft Meteorological Data Relay), wind profilers and observations from regional mesoscale networks are also being explored for landfalling TCs with a view to operational implementation for improving very short-range NWP forecasts as well as nowcasting of TC-related weather.

In response to a need for more reliable estimates of the storm-force wind radius, JMA has adopted a TC analysis software for operational use. The introduction of the software enables forecasters to compute an idealized pressure profile of a target TC that is a best fit to the sea-level pressure observations in the least-square sense, and hence to provide an estimate of central pressure. From the resultant profile of gradient wind speeds with respect to the distance from cyclone centre, an upper bound to the maximum sustained wind can be obtained. Apart from wind-related assessments, forecasting problems in China are highlighted in respect of rapid intensity decay and contrasting rain scenarios (even given similar TC intensity) associated with landfalling TCs, e.g., during TCs Sinlaku, Kammuri, and Vongfong in 2002.

1.6.4 Progress in Research Development Works

a) Theory and Diagnostics

In many ways, inner-core structure and dynamics have attracted the most attention in recent years through observational studies and numerical simulations, as well as in the development of conceptual models. Inner-core vortex dynamical adjustments suggest that a feedback loop exists with the convection and the larger scale environment. It is hypothesized that the inner-core wind structure (intensity) evolution may be described as a balancing act between generation and breakdown in eyewall mixing events. A similar feedback loop is postulated with the onset of environmental wind shear. The shift in the preferred regions of ascent and descent leads to asymmetric convection and precipitation. The associated thermodynamical changes lead to adjustments in the dynamics of the inner-core vortex, which then feedbacks to the environment and convection via vortex Rossby waves.

Emanuel and Holland have developed separate relationships between the maximum potential intensity (MPI) and environmental conditions such as SST, static stability, upper tropospheric conditions and relative humidity. The relative merits between the Emanuel model and the Holland model are discussed in depth in Tonkin (2000) and Camp (2001). Emanuel (2000) analyzed the cumulative distribution of TC wind speeds from NHC and JTWC data for the Atlantic and northwestern Pacific respectively, and showed that the average cyclone reaches a sharp intensity peak followed by a decline at a rate of roughly two-thirds of its prior intensification rate. This average intensity evolution is distinctly different from that simulated by axisymmetric numerical models.

Recent observational and numerical modelling studies have demonstrated that following the onset of environmental vertical wind shear, the favoured ascent region is shifted to the downshear left quadrant of the TC. Clearing of the deep convection and formation of a dry slot occurs on the upshear side in response to forced subsidence. Whereas a symmetric vortex would in principle spin up to its MPI given sufficient time in quiescient favourable conditions, the asymmetric cloud distribution leads to a vortex structure that fails to attain its MPI.

b) Observations and Related Techniques

For accurate specifications of the initial vortex structure, observations on the scale of 5 km must be sought from manned aircraft Doppler radar, unmanned aircraft in situ observations, high resolution satellite soundings (microwave or future GIFTS (Geosynchronous Imaging Fourier Transform Spectrometer)). Highest priority is for winds in the TC and its environment, while moisture measurements are also critical. Research studies including dropwindsondes data into GFDL model has led to improved forecasts in terms of both intensity and motion.

Applying neural networks for TC intensity prediction is also being explored (Baik and Paek (2000)) and its potential for operational implementation evaluated. Results show that vertical wind shear is consistently the most important predictor over the entire forecast period and the errors from the neural networks model are generally 7% to 16% smaller than those from the multiple linear regression model using the same predictors.

c) NWP

A WMO Commission on Atmospheric Science (CAS) mesoscale NWP model inter-comparison case study on Typhoon Flo reported by JMA shows that TC track and intensity are very sensitive to the initial fields, either as a result of different data assimilation schemes or changes in the bogus vortex. Horizontal resolution, in particular, has a large impact on intensity prediction, presumably due to a better representation of the inner structure.

Another analysis on Typhoon Flo by Titley and Elsberry (2000) using TCM-90 enhanced field data has found that rapid intensification appears to be an internal (within 300 km) adjustment process. Favourable development and rapid decay are linked to eddy flux convergence of angular momentum.

Case studies using a high resolution coupled-ocean model confirm that intensity changes are closely related to SST changes. More realistic SST information inevitably leads to better intensity forecasts.

1.6.5 Summary

Despite a steady gain in knowledge and an accelerating increase in observations from remote sensors, operational meteorologists are still scratching at the surface of the problems, and making the best use of whatever information that is available. Better understanding of the internal dynamics that impact on TC structural evolution, and hence intensity changes, is yet to be translated into well-presented conceptual models that benefit operational applications. This prevents the forecasters from reacting intelligently to real-time observational clues in their continual evaluation of NWP reliability in TC intensity prognoses. Some knowledge-based techniques derived from satellite/radar, synoptic and NWP traits, like those developed for track forecast, may be just as useful in the foreseeable future.

On NWP guidance, the general feeling is that improvement in intensity forecast is less evident than for the track forecast. In fact, performance of global operational models from major centres continues to exhibit significant fluctuations in intensity forecast on a day-to-day basis as well as on an interannual basis. The uncertainty is most noticeable in monsoon gyre situations and for multiple TCs when the relative strength of the various vortices involved, including some short-lived mesoscale convective systems or secondary circulation centres, are often poorly represented. While ensemble approach is fast becoming the way to go in many forecasting applications, the relatively poor skill in the models’ ability in correctly depicting TC structure and evolution would probably make it a more difficult proposition for operational implementation in terms of reliable TC intensity prediction.

Lack of observations and inefficient data assimilation techniques are still considered to be the perennial culprits that are stifling significant operational advances. Insufficient communication bandwidth for transmitting potentially useful information and inadequate computer resources to run sophisticated TC models are also regarded as the limiting factors.

In many ways, the study and forecasting of TC structure and intensity evolution are still very much handicapped by the syndrome of ‘single intensity descriptor’, either in the form of ‘minimum central pressure’ or ‘maximum sustained winds near the centre.’ These two oft quoted intensity parameters are not even the easiest to observe or deal with, and yet are still in vogue probably by default or simply for the lack of credible alternatives! Given the complexity of the structural characteristics associated with a TC, intensity definition based on one generalized parameter is always going to be inadequate. Quantified intensity descriptors or derived indices for other aspects are equally important: e.g., TC with no eye (multiple centres, sheared-off, depth of inner-core?), TC with eye (size, shape, multiple, concentric?), surface winds (radius of maximum winds, extent of gale/storm/hurricane winds, asymmetry?), upper winds (outflow pattern, vorticity changes?), convection (extent of spiral bands, rain intensity, MCS, radiance analysis?), environmental factors (monsoon gyre, upper troughs, dry/cold air intrusion, SST changes?).

With enhanced observations, particularly from satellites, and increasingly reliable NWP-assisted analyses, we now stand a better chance of extracting such information from the atmosphere in the coming years. Careful considerations should therefore be given to the formulation of a concerted global or regional effort in the definition, collection, analysis and archival of multi-faceted intensity parameters for a systematic study of TC structure.


Aberson, S. D., and J. L. Franklin, 1999: Impact on hurricane track and intensity forecasts of GPS dropwindsonde observations from the first-season flights of the NOAA Gulfstream-IV jet aircraft. Bull. Amer. Met. Soc., 80, 421-428.

Atlas, R., R. N. Hoffman, S. M. Leidner, J. Sienkiewicz, T.-W. Yu, S. C. Bloom, E. Brin, J. Ardizzone, J. Terry, D. Bungato, and J. C. Jusem, 2001: The effects of marine winds from scatterometer data on weather analysis and forecasting. Bull. Amer. Meteor. Soc., 82, 1965–1990.

Baik, J.-J., and J.-S. Paek, 2000: A neural network model for predicting typhoon intensity. J. Meteor. Soc. Japan, 78, 857-869.

Bankert, R. L., and P. M. Tag, 2002: An automated method to estimate tropical cyclone intensity using SSM/I imagery. J. of App. Met., 41, 461-472.

Bender, M. A., and I. Ginis, 2000: Real-case simulations of hurricane-ocean interaction using a high-resolution coupled model: Effects on hurricane intensity. Mon. Wea. Rev., 128, 917-946.

Braun, S. A., 2002: A cloud-resolving simulation of Hurricane Bob (1991): Storm structure and eyewall buoyancy. Mon. Wea. Rev., 130, 1573-1592.

Camp, J. P., and M. T. Montgomery, 2001: Hurricane maximum intensity: Past and present. Mon. Wea. Rev., 129, 1704-1717.

Croxford, M., and G. M. Barnes, 2002: Inner core strength of Atlantic tropical cyclones. Mon. Wea. Rev., 130, 127-139.

DeMaria, M., and J. Kaplan, 1999: An updated Statistical Hurricane Intensity Prediction Scheme (SHIPS) for the Atlantic and eastern North Pacific basins. Wea. Forecasting, 14, 326-337.

Edson, R. T., and J. D. Hawkins, 2000: A comparison of scatterometer-derived wind data over tropical cyclones as determined from ERS-2 and QuikSCAT data. 24th Conference on Hurricanes and Tropical Meteorology, American Meteorological Society, Boston, MA, 195–196.

Elsberry, R. L., 2002: Numerical modeling for tropical cyclone intensity and precipitation – a workshop report (draft). USWRP Workshop on Numerical Modeling for Tropical Cyclone Intensity and Precipitation Prediction, San Diego, CA, USA, 3-4 May 2002.

Emanuel, K., 2000: A statistical analysis of tropical cyclone intensity. Mon. Wea. Rev., 128, 1139-1152.

Hawkins, J. D., and M. J. Helveston, 1998: ERS-NSCAT capabilities/limitations in mapping tropical cyclone wind fields. 9th Conference on the Interaction of the Sea and Atmosphere, American Meteorological Society, Boston, MA, 45–48.

Katsaros, K. B., E. B. Forde, P. Chang, and W. T. Liu, 2001: QuikSCAT’s SeaWinds facilitates early identification of tropical depressions in 1999 hurricane season. Geophys. Res. Lett., 28, 1043–1046.

Kossin, J. P., and W. H. Schubert, 2001: Mesovortices, polygonal flow patterns, and rapid pressure falls in hurricane-like vortices. J. of Atmos. Sci., 58, 2196-2209
Liu, K. S., J. C. L. Chan, 1999: Size of tropical cyclones as inferred from ERS-1 and ERS-2 Data. Mon. Wea. Rev. 127, 2992–3001.
Liu, W. T., W. Tang, and R. S. Dunbar, 1997: Scatterometer observes extratropical transition of Pacific typhoons. Trans. Amer. Geophys. Union, 237-240.

Nagata, M., et al., 2001: A mesoscale model intercomparison: A case of explosive development of a tropical cyclone (COMPARE III). J. Meteor. Soc. Japan, 79, 999-1033.

Persing, J., M. T. Montgomery, and R. E. Tuleya, 2002: Environmental interactions in the GFDL Hurricane Model for Hurricane Opal. Mon. Wea. Rev., 130, 298-317.

Petty, K. R., and J. S. Hobgood, 2000: Improving tropical cyclone intensity guidance in the eastern North Pacific. Wea. Forecasting, 15, 233-244.

Ritchie, E., J. Simpson, W. T. Liu, C. S. Velden, K. Brueske, and J. Halverson, 2002: A closer look at hurricane formation and intensification using new technology. Coping with Hurricanes, R. Simpson, M. Garstang, and R. Anthes, eds., American Geophysical Society, Washington, DC (in press).

Sakai, R., H. Mino, and M. Nagata, 2002: Verifications of tropical cyclone predictions of the new numerical models at JMA. RSMC Tokyo – Typhoon Center Technical Review, No.5, Japan Meteorological Agency, 1-18.

Sharp, R. J., M. A. Bourassa, and J. J. O’Brien, 2002: Early detection of tropical cyclones using SeaWinds-derived vorticity. Bull. Amer. Meteor. Soc., 83, 879–889.

Titley, D. W., and R. L. Elsberry, 2000: Large intensity changes in tropical cyclones: A case study of Supertyphoon Flo during TCM-90. Mon. Wea. Rev., 128, 3556-3573.

Tonkin, H., G. J. Holland, N. Holbrook, and A. Henderson-Sellers, 2000:
An evaluation of thermodynamic estimates of climatological maximum potential tropical cyclone intensity. Mon. Wea. Rev., 128, 746-762.

Velden, C. S., T. L. Olander, and R. M. Zehr, 1998: Development of an objective scheme to estimate tropical cyclone intensity from digital geostationary satellite infrared imagery. Wea. Forecasting, 13, 172-186.

Wang, Y., 2002: Vortex Rossby waves in a numerically simulated tropical cyclone. Part I: Overall structure, potential vorticity, and kinetic energy budgets. J. of Atmos. Sci., 59, 1213-1238.

World Meteorological Organization, 2002a: TC-related NWP products and their guidance (updated version). WMO/TD-No.966, Geneva, 68 pp.

World Meteorological Organization, 2002b: The science and forecasting of tropical cyclones, 2002. WMO/TD-No.1129, Geneva, 88 pp.

World Meteorological Organization, 2002c: Twenty-eighth Status Report on the Implementation of the WMO Tropical Cyclone Programme. WMO, Geneva.

Zhang, D.-L., Y. Liu, and M. K. Yau, 2000: A multiscale numerical study of Hurricane Andrew (1992). Part III: Dynamically induced vertical motion. Mon. Wea. Rev., 128, 3772-3788.