X hits on this document

63 views

0 shares

0 downloads

0 comments

29 / 30

1772

MONTHLY WEATHER REVIEW

VOLUME 133

12

24

36

48

72

96

120

2003 official error (n mi) 2003 CLIPER error (n mi) No. of cases 19932002 official error (n mi) No. of cases 2003 official error relative to 19932002 official error (%) 2003 official error relative to 2003 CLIPER error (%)

37 49 362 45 2983 18% 24%

64 103 327 81 2724 21% 37%

93 164 291 116

123 221 256 150

161 334 197 224 1818 28% 52%

191 467 156

223 617 125

2480 20% 43%

2229 18% 44%

59%

64%

TABLE 7. A homogeneous set of official and CLIPER average track forecast errors in the Atlantic basin for the 2003 season for all tropical cyclones, including depressions. Ten-year averages for the period 19932002 are also given for the 12- through 72-h forecasts.

Forecast period (h)

The depression dissipated on 10 September over the western Cape Verde Islands.

the difference between a forecast wind speed for a par- ticular time and the best-track wind speed for the same time.

4. Forecast verification

The NHC issues an official forecast, every 6 h, of the position (latitude and longitude of the circulation cen- ter) and the intensity (maximum 1-min wind speed at 10 m above the surface) for all tropical cyclones in the Atlantic and eastern North Pacific basins. These fore- casts are valid at 12, 24, 36, 48, 72, 96, and 120 h from the initial synoptic time of the forecast (0000, 0600, 1200, or 1800 UTC). It is noted that 2003 is the first year that 96- and 120-h forecasts were issued. Forecasts are verified by comparison with the best trackdescribed in section 1 above. Forecasts are verified for all tropical and subtropical cyclones, including depressions. A track forecast error is defined as the great circle dis- tance between a forecast position valid at a particular time and the best-track position for the same time. An intensity forecast error is defined as the magnitude of

Table 7 lists the average Atlantic basin official track forecast errors for 2003. For comparison, the average official track errors for the previous 10 yr are also listed. The 2003 average track errors range from 37 n mi at 12 h to 161 n mi at 72 h and are from 18% to 28% smaller than the previous 10-yr official average errors. There are no previous 10-yr averages to compare to this sea- sons 96- and 120-h average forecast errors. Table 7 also lists the average errors for a climatology and persis- tence (CLIPER) track model (Aberson 1998). The CLIPER model represents a no skilllevel of forecast accuracy and the 2003 average official errors are from 24% to 64% smaller than their corresponding CLIPER errors.

Average official track forecast errors have decreased over the years (Franklin et al. 2003b) in consonance with advances in numerical weather prediction. Figure 7 is a plot of official yearly average official track fore-

FIG. 7. Official yearly average 24-, 48-, and 72-h track forecast errors and linear trend lines. Tropical depressions are not included.

Document info
Document views63
Page views63
Page last viewedSun Dec 04 12:32:30 UTC 2016
Pages30
Paragraphs1627
Words16885

Comments