The Sensitivity of WRF Simulations of Hurricane Ivan to Choice of Cumulus Parameterization

As progressively smaller grid spacing is used in numeral modeling of mesoscale weather systems, explicit simulation of convection without parameterization has become more common. However to our knowledge, the appropriateness of omitting the cumulus parameterization (CP) scheme at smaller grid spacing (4 km or less) has not been systematically determined in a tropical environment. The objective of this study is to observe the effects of explicitly resolving convection in a tropical cyclone, as opposed to using a convective parameterization.

Weather Research and Forecasting (WRF) model simulations of Hurricane Ivan (2004) are run with a one-way nested grid of three domains, the innermost of which had a grid spacing of 3 km. Five runs are performed with identical specifications and physics, with the exception of the cumulus scheme, which was turned on and off to test sensitivity. The first run is designated 000 and has no CP scheme on any of the 3 domains. The subsequent two are run with the Betts-Miller-Janjic (BMJ) scheme on and off on the inner-most grid and designated BMBMBM and BMBM0, respectively. The next two are run as before but with the Kain-Fritsch (KF) scheme and designated KFKFKF and KFKF0.

Ivan strengthens the most in the BMBM0 run, and best matches the minimum central pressure observed. The BMBMBM run simulates the second-most intense hurricane, followed by the KFKF0, KFKFKF, and 000 runs. Comparison of the precipitation rates indicates that turning off the convective scheme does cause more intense precipitation, as the BMBM0 simulation has an average maximum hourly rainfall rate more than a tenth of an inch higher than the BMBMBM run. Overall, the appearance of the precipitation differs, as also observed in the radar imagery. Run 000 simulates a more discontinuous, ragged eyewall with numerous, larger reflectivity cores in the spiral bands. Run BMBM0 shows these same traits to a lesser degree. The BMBMBM simulation has the greatest tendency to exhibits a closed-off eyewall, and the precipitation field of the spiral bands is more continuous.

Future work will focus on comparison of soundings from the different model runs. Cross-sections of other variables, such as model simulated reflectivity and vorticity, will also be used to examine the effects of the turning the cumulus scheme off. Ultimately, these model simulations will be compared to runs with smaller grid spacing to determine what resolution is appropriate for turning off the CP scheme.