Verification of Forecasts of Tropical Cyclone
Issued on 16 Jan 2009
Since 2000, City University of Hong Kong has been issuing real-time predictions of the annual number of tropical cyclones (TCs) affecting the western North Pacific (WNP). Verifications of the predictions have shown that the predictions are mostly correct within the error bars.
These are all statistical predictions with predictors drawn from a large group of indices that represent the atmospheric and oceanographic conditions in the previous year up to the spring of the current year. The most prominent ones include the proxies for El Niño/Southern Oscillation (ENSO), the extent of the subtropical ridge, and the intensity of the India-Burma trough. Details can be found in Chan et al. (1998, 2001) and Liu and Chan (2003).
2. Verification of the 2008 forecasts
a. Summary of the forecasts issued
Our April forecasts (issued on 18 April 2008) suggested “slightly above-normal activity for all the categories”. A similar forecast was made on 24 June 2008. Detailed numbers are summarized in Table 1, together with the observed numbers based on the warnings from JTWC and the Tokyo Regional Specialised Meteorological Center (RSMC) (Table 2).
Disagreements occurred among the warning centres on the intensity of some of the systems. Four systems (01W, 14W, 16W and 22W) were classified by the JTWC as having reached tropical storm intensity but were not named by the RSMC. On the other hand, RSMC issued tropical storm warning for one system (Phanfone) but no warning was given by JTWC. In addition, Halong was considered by JTWC as having reached typhoon intensity but not by RSMC Tokyo
b. Verification and discussion
The TC activity in 2008 over the WNP was below-normal (Fig. 1), with 25 TCs reaching at least tropical storm intensity and 12 TCs reaching typhoon intensity (based on JTWC, Table 2). The overall TC intensity is also lower. The number of typhoons is 5 less than the normal number (the normal being 17) but the number of tropical storms and typhoons is only 2 less than the normal number (the normal being 27). Our forecasts from both April and June over-estimated the TC activity especially for the number of typhoons. This is the first time that our forecasts were incorrect, the possible reasons of which are discussed below.
The mean genesis location shifted westward and no storm formed over the tropical WNP east of 150oE, which is rather unusual if 2008 is considered to be an ENSO-neutral year. However, as discussed below, the atmospheric conditions (see Figs. 2 and 3) were very much like that of, or a continuation of, a La Niña event.
The ENSO is generally considered to be in its neutral status during the TC season of 2008. The mean Jun-Nov Nino4 index is -0.31, indicating the slightly cooler-than-normal SSTs over the equatorial Central Pacific (Fig. 2). However, a large area of negative SST anomalies is found over the subtropical Pacific east of 170oE. Correspondingly, easterly anomalies are found over the entire tropical WNP, with the maximum amplitude near 150oE (Fig. 3). As a result, the monsoon trough is weaker. Therefore, atmospheric and oceanic conditions are not favorable for TC genesis and these partly explain the below-normal TC activity in the year of 2008.
It is worth to note that the PDO is in its negative phase during the TC season of 2008. The mean PDO index averaged between June and November is the lowest in the last five decades (Fig. 4). The unfavorable atmospheric and oceanic conditions for TC genesis may be related to this strong negative PDO event. It is therefore of interest to investigate the TC activity in a year after a La Nina event and under the background of a strong negative PDO event. During the period of 1960-2008, five such TC seasons (1975, 1999, 2000 and 2008) are found (Table 3). Out of these seasons, four are associated with slightly below-normal or below-normal TC activity. For the other TC seasons without a strong negative PDO background (i.e. either a weak or positive PDO signal), six of the eight seasons are associated with above-normal TC activity, which is a main reason for our forecasts. The strong negative PDO in 2008 may be one of the factors responsible for the below-normal TC activity. Our forecast scheme does not include the PDO predictor and this may be the reason for the over-estimation of the TC activity in 2008. Further research into the effect of PDO is necessary in future forecasts.
Chan, J. C. L., J. E. Shi and C. M. Lam, 1998: Seasonal forecasting of tropical cyclone activity over the western North Pacific and the South China Sea. Weather Forecasting, 13, 997-1004. Abstract
Chan, J. C. L., J. E. Shi and K. S. Liu, 2001: Improvements in the seasonal forecasting of tropical cyclone activity over the western North Pacific. Weather Forecasting, 16, 491-498. Abstract
Liu, K. S. and J. C. L. Chan, 2003: Climatological characteristics and seasonal forecasting of tropical cyclones making landfall along the South China coast. Monthly Weather Review, 131, 1650-1662. Abstract
Table 1. Forecasts of TC activity in 2008 issued in April and June. The observed activity based on both the JTWC and RSMC-Tokyo warnings and the normal values are also shown.
Table 2. 2008 summary of tropical cyclones over the western North Pacific.
Table 3. Summary of tropical cyclone activity in the years followed the La Niña events. Blue and red shadings indicate the below-normal and above-normal TC activity respectively.
Fig. 1. Annual number of tropical storms and typhoons in the western North Pacific between 1960 and 2008. The horizontal line indicates the climatological mean. The green circle indicates the year of 2008.
Fig. 2. SST anomalies between June and November in 2008.
Fig. 3. 850-hPa wind anomalies (vector) between June and November in 2008. Shadings indicates the wind speed (interval = 1 m s-1).
Fig. 4. Time series of the July-October PDO index between 1960 and 2008. The green circle indicates the year of 2008.