•The DTC conducted extensive tests of the Community HWRF model in order to designate its second HWRF Reference Configuration (dubbed H3RC). Over 1200 HWRF runs for the Eastern North Pacific and North Atlantic basins for the 2010 and 2011 seasons were conducted, demonstrating that a robust testing environment is functional at the DTC.

    •Results show that mean track errors increase linearly with time from near zero at initialization to approximately 300 nm and 250 nm in the North Atlantic and Eastern North Pacific basins, respectively.

    •Mean cross-track errors are positive for both basins at all lead times, indicating that the forecast storm is usually to the right of the observed one. Along-track errors indicate that the storm is too slow at all lead times in the Atlantic basin. In the Eastern North Pacific basin, along-track errors are only statistically significant in the last two days of the forecast, and also show that the storm moves too slow when compared to observations.

    •While the average track errors are smaller in the Eastern North Pacific basin, both the along- and cross- track errors are larger in this basin than in the North Atlantic one. This indicates that the Eastern North Pacific track errors have more bias (too slow and too much to the right), while the Atlantic track errors are less systematic.

    •The mean of the absolute intensity errors displays a sharp growth in the first 12 h of forecast, from near 3 to 10 kt. After that the errors grow more slowly and reach 20 kt in the North Atlantic and 30 kt in the Eastern North Pacific basin.

    •The mean of the intensity errors (bias) indicates that the model underpredicts intensity at the initial time and throughout the first half of the forecast period. In the second half of the forecast period, the bias remains negative in the Eastern North Pacific but is not statistically significant, while it becomes positive (and statistically significant) in the North Atlantic. This suggests that, in the Atlantic, the forecast intensity tends to increase with forecast lead time, that is, that the model is overly cyclogenetic.

    •The forecast storm size is consistently larger than the observed one, with the errors being larger in the Atlantic than in the Eastern North Pacific. For most radii and quadrants, the errors decrease in the first few hours of the forecast, indicating that the initial vortex is too large. In the North Atlantic basin, after the initial adjustment period, errors grow for the 64-kt radii, suggesting that the core of the storm expands in the forecast. On the other hand, errors for the 34-kt radii tnd to decrease throughout the forecast period, suggesting that the overall storm size improves after initialization.

    •Model output files have been archived and are available to the community for future studies. Forecast maps and verification graphics, along with this report and additional information are available in the DTC website.