Changes in Typhoons Over the Last Decades As Given in Observations and Climate Model Simulations

Abstract

Typhoons pose a threat for costal populations, offshore and onshore industries, shipping, agriculture, or forestry. Future predictions of those phenomena are important, and depend to a great extent on the evaluation of changes in typhoon statistics of the past. In this study an analysis of typhoon frequencies and intensities over the last more than six decades was conducted. ‗Best Track Data‘ sets were assessed, these are provided by several meteorological institutes, which compile reanalyzed in-situ- and satellite-based measurements of tropical cyclones (TC). But, the intensity estimations provided by individual data sets differ (Barcikowska et al., 2012), which influences the derived climatological statistics. A source for these discrepancies are the varying approaches to estimate TC intensities used by the individual weather services, like the definition of TC intensity or measurement products. As a result, TC activity trends are ambiguous and depend on the analyzed data set. An alternative approach to obtain typhoon statistics is to dynamically downscale reanalysis data with a regional climate model (Feser and von Storch, 2008a, b). Ensemble simulations are useful to evaluate the effect of internal model variability on typhoon simulations (Feser and Barcikowska, 2012). A spectral nudging technique (von Storch et al., 2000) is beneficial for constraining the RCM‘s internal variability, as it keeps the large-scale weather state inside of the model domain close to the forcing reanalysis. A long-term climatology was computed with the regional climate model CCLM and either the NCEP/NCAR or the ERA40 reanalyses as external forcing. The variability of intense TC numbers shows large similarity to the observed one, especially for the last three decades (when satellite data were available). But, long-term statistics derived for the whole simulation period (six decades) differ. Both RCM simulations indicate mostly runs presumably reflect inhomogeneities of the forcing reanalyses. An upward shift in TC intensities and in the magnitude of its variability in 1978 was detected in both regional simulations. At this time satellite measurements were introduced to the reanalyses. This seems to be responsible for the detected shift, and may contribute to the increase in TC activity measures.