Understanding the phase of cyclones gives a broader, yet insightful, perspective on the overwhelming distribution of all cyclones. Cyclones aren't simply tropical or extratropical; there is a great continuum of cyclone types, with a significant fraction of them having characteristics of both tropical and extratropical cyclones.

The phase of a cyclone (warm vs cold-core, in particular) is related to its intensity, size, forecast uncertainty, and ultimately the threat it poses to us. Generally, a warm core cyclone has more forecast uncertainty associated with it than a cold-core cyclone. For example, the development and intensification of tropical cyclones is one of the most difficult forecasts to make. The diagnostics shown here give an indiciation of whether a cyclone that is developing within the models is a pure warm-core development, or hybrid development.

Forecasting the phase conversion or transition of cyclones (e.g. tropical into extratropical and subtropical into tropical) is a difficult but important task, and the phase diagnoses provided here provide considerable insight into the essence of model cyclone development. In concert with real-time surface, ship, and satellite observations, a more accurate diagnosis and forecast of cyclone evolution is possible when the phase diagrams are used.

During the 2001 hurricane season, the phase diagnoses provided here accurately forecast the conversion of two subtropical cyclones into Hurricanes Karen and Olga. The extratropical transition of Hurricanes Gabrielle and Michelle were both well-forecast by the phase diagrams as well.

These diagrams also provide for forecast track and intensity comparison.

NCEP AVN, Eta and NGM; U.S. Navy NOGAPS; U.S. Air Force MM5; Canadian Meteorological Center Global CMC; UK Met Office UKMET. The ECMWF model data available to us is (unfortunately) far too coarse to allow reliable phase analysis.

A cyclone will not be analyzed on the phase diagrams if: 1) the cyclone is currently too close to the edge of the model domain, 2) the cyclone does not last for 24h, or 3) the cyclone has a minimum SLP > 1018mb.

The black ones are existing cyclones, the red ones are forecast cyclones that have not yet developed.

Either the cyclone has just formed, has not yet formed, or the database has a large gap in it from missing data that has prevented continuous tracking of the cyclone from one model analysis to the next.

The detection of cyclones and tracking them is a very complex and difficult process. Cyclones may form downstream from an already existing cyclone. In this case, it can become difficult to determine where the path of one cyclone ends and another begins. As a result, the track and phase evolution of a cyclone can appear to jump, if another, stronger, cyclone forms relatively close to an existing cyclone. Also, if a cyclone moves extremely quickly (> 30 ms^-1), the tracking algorithm may interpret the new cyclone position as the formation of a new cyclone. This is most likely in the NOGAPS, UKMET, and CMC analyses which are available only every 12h -- and, consequently, large changes in cyclone position can result.

Thumbnail versions of the phase diagrams for a given cyclone are compared, allowing the user to compare the current and forecast phase evolution for a single cyclone among several models. Note that not all models will necessary resolve the same cyclones, especially in the tropics. The consensus forecast (unweighted mean of all available models) for the given cyclone is also presented. The 1 standard deviation variability from this mean is given as the shading either side of the mean phase analysis.

The two parameters used to describe the cyclones (B, -V_T) are calculated over the pressure ranges of 900-600hPa. If high terrain enters this pressure range, then the field values for thickness and geopotential height are contaminated by interpolation below ground. As a result, the diagnostics are not representative of the atmosphere or the cyclone's true phase, and it will not be analyzed.

Existing computer resources do not allow us to examine the entire globe.

The phase analyses shown here are derived completely from model analyses and forecasts. They will not and cannot represent the complete or true structure of the observed cyclone. The phase diagrams are only as accurate as the model analyses and forecasts from which they are derived. As model analyses become increasingly more accurate, so will the phase diagrams.