Title: Alternative expression for the maximum potential intensity of tropical cyclones

Abstract: Emanuel's concept of Maximum Potential Intensity (E-PI) estimates the maximum velocity of tropical cyclones from environmental parameters. At the point of maximum wind, E-PI's key equation relates proportionally the centrifugal acceleration (squared maximum velocity divided by radius) to the radial gradient of saturated moist entropy. The proportionality coefficient depends on the outflow temperature. We show that a different relationship between the same quantities derives straightforwardly from the definition of saturated moist entropy and gradient wind balance. Here the proportionality coefficient depends on the radial gradient of local air temperature. The robust alternative reveals a previously unexplored constraint: for E-PI to match the observations, the radial gradient of air temperature at the top of the boundary layer should be a function of the outflow temperature. When, as commonly assumed in various E-PI applications, the air is horizontally isothermal, E-PI's key equation underestimates the squared maximum velocity by approximately twofold. This provides a new explanation to the superintensity phenomenon (maximum wind speeds exceeding E-PI). The discrepancy diminishes (grows) when the air temperature grows (declines) towards the storm center, and with increasing (decreasing) local temperature. The latter pattern predicts less superintensity at higher temperatures, corroborating recent results from numerical simulations. An increase in air temperature towards the center is shown to mask the effect of supergradient wind (which E-PI is known to underestimate). The established theoretical relationships are illustrated with the data for Hurricane Isabel 2003 and their implications for assessing storm intensities are discussed.