Cyclone activities as inferred from the Twentieth Century Reanalysis version 3 (20CRv3) for 1836-2015

Preliminary results obtained from tracking cyclones in the ensemble-average and individual members of the NOAA-CIRES-DOE Twentieth Century Reanalysis version 3 (20CRv3) ensemble for the period 1836-2015 will be presented. Comparison with tracking in the 20CRv2c ensemble-average series will also be shown.

The results indicate that the 20CRv3 is an improvement in representing cyclone climate and variability compared to previous versions (20CRv2c or 20CRv2). However, as in previous versions, the 20CRv3 ensemble-average fields are too smooth to use for tracking cyclones and studying cyclone climate, especially for the period before 1960 for the NH and for the entire reanalysis period for the SH, and that there are still temporal inhomogeneity issues in the 20CRv3, especially in the SH and in the early period for the NH, due to the increases over time of observations available for assimilation. The improvements arise from the use of a higher model resolution and the assimilation of more observations. They include that the 20CRv3 ensemble shows cyclones of higher intensities and a higher number of deep cyclones (center pressure ≤ 960 hPa) in the Northern Hemisphere than the 20CRv2c counterpart. Historical trends of cyclone activity and their uncertainties will be discussed based on the results of tracking the individual members of the 20CRv3 ensemble, with the temporal inhomogeneity issues being taken into account.