Difference of rainfall distribution for tropical cyclones over land and ocean and rainfall potential derived from satellite observations and its implication on hurricane landfall flooding prediction
Conference
Jiang, H, Halverson, JB, Simpson, J. (2006). Difference of rainfall distribution for tropical cyclones over land and ocean and rainfall potential derived from satellite observations and its implication on hurricane landfall flooding prediction
.
Jiang, H, Halverson, JB, Simpson, J. (2006). Difference of rainfall distribution for tropical cyclones over land and ocean and rainfall potential derived from satellite observations and its implication on hurricane landfall flooding prediction
.
Rainfall distributions over land and ocean in Atlantic TCs have been studied using observations from MPA. Between 1 January 1998 and 31 December 2004, 2680 3-hourly instantaneous MPA measurements were collected in 37 Atlantic landfall TCs with intensity ranging from tropical depression to category 5 hurricane. PDFs and azimuthal averages are constructed as a function of TC intensity and location. Correlations between rain rate and maximum wind intensity for over-land and over-ocean observations are compared. (Table presented) The PDF and azimuthal average analysis is used for comparison of MPA TC rainfall distributions with previous studies and for comparison of over-land and over-ocean differences. It is found that over-ocean distributions derived from the MPA product have the similar characteristics to those results in previous studies (Rodger et al. 1994, Rodger and Pierce 1995, Lonfat et al. 2004). However, peak value and location of peak for over-land distributions are significantly different with those for over-ocean distributions, indicating stronger rains for over-land observations. Many previous studies (Rodger et al. 1995, Rodger and Pierce 1995, Rodger et al. 2001) have shown that the inner-core mean rain is well correlated with TC intensity. However, this study found that this statement might be only true for over-ocean observations. For TCs over land, the correlation is much weaker. Using the same dataset, the relationships between rainfall potential and TC landfall rain parameters have been examined. High correlations are found between rainfall potential for 1 day before TC landfall and maximum storm total rain over land. A prediction index is proposed based this finding. Hurricane Katrina (2005) case is applied to verify this prediction index. By comparing with surface rain gauge observations, the error of the forecasted TC landfall rain is within 10%. Future work is to verify this index using more 2005 hurricane cases.
