"Climate Change and Southern Hemisphere Tropical Cyclones" International Initiative – Progress since the First International Conference on Indian Ocean Tropical Cyclones and Climate Change - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 20

with the beginning of geostationary satellite coverage in 1998, as a result of
the launch of the MeteoSat-7 satellite (Kossin et al., 2007).
In identifying TC positions and estimating intensities, operational
forecasters from Fiji and New Zealand (responsible for the eastern SPO tracks)
faced similar challenges, through gradually improved observational data and
understanding of TC development (S. Ready, pers. comm.). Within the satellite
era, low-resolution geostationary satellite imagery for the South Pacific (west
of about 155°W) became available to those forecasters from 1980. However,
only from the early 1990s, with high-resolution imagery from the GOES-West
satellites now available to the RSMC Nadi (established in 1993), did the Dvorak
technique become increasingly used for estimating storm intensity. Throughout
the 1980s and in the first half of the 1990s, there still was a reluctance to assign
intensities beyond 80 knots (central pressures below about 955 hPa). From the
mid-1990s, as forecasters in Fiji became more proficient at the Dvorak
technique, there was a greater range of intensities assigned in TC warnings. It
appears that for the SPO east of 160°E the most reliable estimates of TC intensity
start in the early 1990s, with reliable estimates of central pressure of the most
intense TCs from the mid-1990s. It is likely that prior to this time the number
of TCs with reported intensities below 950 or 945 hPa in the eastern SPO is
underestimated.
This variation between different regions, in the time period for which high
resolution satellite imagery is available, influences the accuracy of cyclone
intensity estimations and imposes limitations on the analysis of historical data.
For example, it restricts the accuracy of long-term trend analysis of changes in
cyclone intensity - something required to address the important question of
how TC activity is changing over decades and possible relationships to global
climate change. Consequently, reanalysis of the historical TCs in order to obtain
globally homogeneous records is essential (Kuleshov et al . , 2010a).
3. Southern Hemisphere Tropical Cyclone Trends
Trends in TC occurrences and intensity, and possible physical mechanisms for
change, have been discussed widely in recent years. Webster et al . (2005)
reported that the global number of very intense TCs (Saffir-Simpson categories
4 and 5) had almost doubled over recent decades. Using the TC potential
dissipation index as a measure of TC activity, Emanuel (2005) arrived at similar
conclusions for two of the major TC basins: the North Atlantic and the North
West Pacific. Other authors have rejected these findings, mainly based on the
argument that changes have been so great in observation technologies and
analysis techniques that the reported changes are artificial, and not due to any
actual trends (Landsea et al., 2006; Chan, 2006; Kossin et al., 2007). For the
Australian region (AR) and the SH, trends have been reported in the number
of TCs and in the proportion of very intense TCs by various authors, including
Nicholls et al. (1998) and Harper et al. (2008).