Reexamination of Tropical Cyclone Wind-Pressure Relationships

Table 2. Statistics associated with the Eq. 7 using the observed environmental pressure(P_env), Eq. 16 using the climatological environmental pressure (P_clim) from the sample, the Atlantic Dvorak, Koba et al (1990), A&H, Love and Murphy (1985) and Crane WPRs. Bias and error statistics that are statistically different than those produced by Equation 7 are shown as italicized and gray and italicized and boldface for the 95%, and 99% levels, respectively.

Table 3. Statistics (R², bias, RMSE and MAE) associated with the Eq. 8 using the observed environmental pressure(P_env), Eq. 8 using the climatological environmental pressure (P_clim) from each regional sub sample along with the appropriate Landsea et al. (2004 ) regional WPRs utilizing a reference pressure equal to 1013 and of Penv . Bias and error statistics that are statistically different than those produced by Equation 8 are shown as italicized and gray and italicized and boldface for the 95%, and 99% levels, respectively.

Table 4. Independent comparison of results obtained from Eq. 7 and Eq. 8 vs. the operational Dvorak Tables. Data includes 491 fixes from 12 Atlantic tropical cyclones and 1 East Pacific tropical cyclone during the 2005 season. Bias and error statistics that are statistically different are shown as italicized and gray and italicized and boldface for the 95%, and 99% levels, respectively.

Table A1. Biases and MAE associated with the various fits to the raw Atkinson and Holliday (1975) dataset. Listed here are the published A&H WPR (1), the cyclostrophic form fit to the binned A&H data (2), the gradient fit for the binned A&H data (3) and for comparison the gradient fit to the Koba et al. (1990) WPR (4).

Table B3. Dvorak CI vs P tables for storms occurring poleward of 30^o latitude.