Eric Blake spent from 1998-2001 as a graduate student at Colorado State University. His research efforts went into the development of an Atlantic basin August-only hurricane forecast scheme which was used for our 6 August 2003 forecast. See Blake (2002) or Blake and Gray (2003) for background information. Blake attained a high degree of hindcast skill for the period of 1950-2000. Note that Blake's forecast scheme showed a near perfect result for this August.

Information obtained through August 2003 shows that we have had 46 percent of the average season Net Tropical Cyclone (NTC)⁵ activity. In the average year, 38 percent of the seasonal average NTC of 100 occurs by the end of August.

Our new September-only forecast calls for four named storms, two hurricanes, one major hurricane and NTC activity of 55 which is slightly above the mean September-only average value of 48.

Our October-only forecast calls for three named storms, two hurricanes, one major hurricane and NTC activity of 30 which is above the mean October average (17).

We now anticipate that the 2003 Atlantic basin tropical cyclone (TC) season will be about the same as the activity we anticipated in our 6 August forecast (see table below) with 7 hurricanes (rather than 8). United States and Caribbean basin landfall probability for the remainder of the 2003 season is now estimated to be slightly above the long-term average.

We are grateful to AIG - Lexington Insurance Company (a member of the American International Group) for providing partial support for the research necessary to make these forecasts.

1. Chris Landsea, Eric Blake and John Knaff (former Gray project graduate students) deserve (by virtue of their many contributions to our project climate studies) to be co-authors on this forecast. As each of these individuals now work for NOAA (or a branch of NOAA), they have been directed to remove their names from CSU forecasts.

2. Due to technical difficulties the regular NOAA reanalysis products that we use have not been available to us through the later half of August. This has led to our estimating August conditions from less accurate alternate data sources.

Colorado State University (CSU) has issued seasonal hurricane forecasts for the last 20 years. These forecasts, which are issued in early December of the prior year, and in the early part of the months of April, June and August of the current year, have steadily improved through continuing research. CSU forecasts now include individual monthly predictions and seasonal updates of Atlantic basin activity and monthly U.S. hurricane landfall probabilities which are issued in early August, early September and early October.

The National Oceanic and Atmospheric Administration (NOAA) has also begun to issue Atlantic basin seasonal hurricane forecasts, but they do not issue monthly forecasts or landfall probability forecasts. NOAA issues two forecasts per season and gives a range of numbers. The NOAA forecasts are independent of the CSU forecasts, although they utilize prior CSU research augmented with their own insights. The NOAA and the CSU forecasts will not necessarily be in agreement.

1  Predictions of Individual Monthly Atlantic TC Activity

       A new aspect of our climate research is the development of TC activity predictions for individual months. On average, August, September and October have about 26%, 48%, and 17% or 91% of the total Atlantic basin NTC activity. Initial August-only forecasts have now been made for the last four seasons, and September-only forecasts have been made for the last two seasons. This is the first year that we have issued an October-only forecast.

There are often monthly periods within active and inactive hurricane seasons which do not conform to the overall season. To this end, we have recently developed new schemes to forecast August-only, September-only and October-only Atlantic basin hurricane activity by the beginning of each respective month. These efforts have been recently documented in CSU project reports by Eric Blake (2002) for our August-only forecast and by Phil Klotzbach (2002) for our September-only forecast - see the last page for references. Klotzbach is working on documentation for the October-only forecast.

Quite skillful August-only, September-only and October-only prediction schemes have been developed based on 51 years (1950-2000) of hindcast testing using a statistically independent jackknife approach. Predictors are derived from prior month, usually June and July (NCEP global reanalysis) data for all three (August-only, September-only and October-only) individual monthly forecasts and include August's data for the early September update of the September-only and October-only forecasts. Table 1 gives an outline and timetable of the different forecasts and verifications we issue after the end of July.

Through August, the 2003 hurricane season has had 46 percent of the NTC activity of the average hurricane season. As of 1 September, seven named storms, three hurricanes and one major (cat. 2-3-4) hurricane have developed. The average number of named storms through August is 4.2, the average number of hurricanes is 2.4, and the average number of major hurricanes is 0.7. This season has thus far been more active than the average season.

Table 2 and Fig. 1 list the nine September-only predictors with their location and sign for higher TC activity. Table 3 shows which of the nine predictors are used for each forecast parameter of NS, NSD, H, etc.

Table 4 lists the value of each September-only predictor and whether its 2003 value indicates above or below average September-only TC activity. A majority of the September predictors call for below-average activity. We have had to estimate the values of the August predictors for September activity due to a data failure at the Climate Diagnostics Center.

Our early September update of September TC activity has been increased considerably (see Table 5). This is largely due to Hurricane Fabian. Fabian has already produced two intense hurricane days and is forecast to remain intense for the next 2-3 days. It should also be noted that the NS, H and IH that are forecast for September do not include Fabian since it became an intense hurricane on August 30th.

1.1  Independent October-only Statistical Forecast

       We are issuing our first monthly forecast for October-only tropical cyclone activity this year. Through examination of the NCEP/NCAR reanalysis, we have discovered four predictors that in combination explain about 60 percent of the October cross-validated variance in Net Tropical Cyclone activity by 1 September for the hindcast period of 1950-2001. We are currently unable to find combinations of predictors that explain large amounts of variance for the individual tropical cyclone parameters (i.e., named storms, hurricane days, etc.). Therefore, our October forecast consists of predicting NTC and consequently increasing or decreasing October's values for the other parameters accordingly. For example, if October NTC was 150 percent of normal, and a typical October had two named storms, we would forecast three named storms for October.

Due to data problems at the Climate Diagnostics Center, we were unable to obtain new data for our early September update of October activity. We are therefore going to adjust our October forecast upwards due to low sea level pressure in the subtropical Atlantic SST's. In round numbers we are forecasting 3 NS, 2 H, 1 IH and a NTC of 30.

Table 6 displays a summary of this year's hurricane activity through August and our projection for the rest of the season. We expect September and October activity to be slightly above normal. We assume no activity in November.

2  U.S. Landfall Probability Forecast for September and October

       We have recently developed a methodology for calculating probability of landfall along the entire U.S. coastline for the months of September and October. Based on a 1950-2001 dataset, the probability of various intensity classes of tropical cyclones making landfall in September is given in Table 7. Table 8 displays the landfall probabilities for October. The probability of September and October landfalling tropical cyclones in any particular year is directly related to the NTC forecast. All landfall probabilites are slightly above average based on an NTC forecast of 55 for September and an NTC forecast of 20 for October.

3  Discussion

       August had about average activity. We expect the remainder of the 2003 hurricane season to be about as active as an average year. We foresee a slightly above-average probability of U.S. and Caribbean basin hurricane landfall for the rest of this year.

4  Forecast Verification

      We will issue a seasonal update and a revised October-only forecast on 2 October 2003. Our forecasts for 2003 will be verified in late November 2003. Our first seasonal hurricane forecast for the 2004 season will be issued on 5 December 2003. All forecasts and verifications are available at our web address given on the front cover (http://tropical.atmos.colostate.edu/forecasts/index.html).

5  Papers Which Document Our August-only and September-only Atlantic Basin Forecasts

1 Blake, E. S., 2002: Prediction of August Atlantic basin hurricane activity. Dept. of Atmos. Sci. Paper No. 719, Colo. State Univ., Ft. Collins, CO, 80 pp.

1 Blake, E. S., and W. M. Gray, 2003: Prediction of August Atlantic basin hurricane activity. Submitted to Wea. and Forecasting

1 Klotzbach, P. J., 2002: Forecasting September Atlantic basin tropical cyclone activity at zero and one month lead times. Dept. of Atmos. Sci. Paper No. 723, Colo. State Univ., Ft. Collins, CO, 91 pp.

1 Klotzbach, P. J., and W. M. Gray, 2003: Forecasting September Atlantic basin tropical cyclone activity. Wea. and Forecasting, 18 (in press).

Footnotes:

¹Professor of Atmospheric Science

²Research Associate

³Research Associate

⁴Research Associate

⁵NTC is a combined measure of the yearly mean of six indices (NS, NSD, H, HD, IH, IHD) of hurricane activity as a percent deviation from the 1950-2000 year average.