Skip to content

Uncertainties in Tropical-Cyclone Translation Speed

June 6th, 2019

Key Findings

  • The apparent slowing of tropical cyclones occurs abruptly, rather than gradually, and primarily in the earlier part of the period of record, rather than the later part when the effects of climate change have been accelerating. These factors argue against a dominant role of climate change.
  • While some of the slowing is associated with natural climate variability much of it may be artificial, due to the introduction of satellite remote sensing in the 1960s.
  • The conclusions of this study are tentative and will require further research to firm them up. Such research could involve scrutiny of both observations and the output from climate models. In particular, a better understanding of the effects of the introduction of satellite remote sensing in the 1960s on the record of cyclone movement is needed.

John R. Lanzante. Nature. DOI: 10.1038/s41586-019-1223-2

A recent study found a downward trend from 1949-2016 in the speed at which tropical cyclones move. If this could be attributed to climate change the implications would be enormous. Slower moving storms, as exemplified by Hurricane Harvey in 2017, have the potential to produce much more rainfall than faster ones.

This study finds that the bulk of the decrease in speed is related to abrupt changes that occur in the earlier part of the period of study. Both the abruptness along with the lack of change during more recent times argues against a dominant role for climate change. The results suggest that the changes are likely due to a combination of natural climate variability and changes over time in the manner in which tropical cyclones were tracked. In particular the introduction of satellite remote sensing in the 1960s may have distorted the record by yielding more observations in areas which had previously been uncharted. It appears that such areas are ones where storms naturally move more slowly.

Figure 1 Latitudinal profiles of climatological average tropical cyclone (TC) translational speed (km/hr) (black) over 1949-2016, with labeling along the bottom, and relative change in number of TC observations (%) (solid red) from the pre-satellite to satellite eras, with labeling along the top. Latitude is labeled on the left and the dashed red line represents zero change between eras. Each panel corresponds to a different ocean basin: (a) Northern Atlantic, (b) Eastern North Pacific, (c) Western North Pacific, (d) Northern Indian, (e) Southern Pacific, and (f) Southern Indian. Of relevance is the tendency for positive (negative) values of change to correspond to lower (higher) values of TC speed. This is most evident in (a), (c), (e) and (f) indicating that the introduction of satellite remote sensing produced an artificial decrease in TC speed because satellites are able to detect slower-moving TC’s that would have previously gone undetected.


Figure 2 Each bar represents the linear trend (km/hr per year) in tropical cyclone (TC) translational speed, with labeling on the left, corresponding to one of nine scenarios, as labeled at the top. Stippled bars indicate the trend is statistically significant at the 5% level. In each panel the left-most bar (black) represents the trend based on the raw data. Moving from left to right, as the scenario number increases, we systematically alter the data to remove more of the effects of the extraneous factors of satellite changes and/or natural internal climate variability. It is noteworthy that the trends, which for the raw data are predominantly negative and significant, tend to become positive or less negative, as well as non-significant, as we alter the data to remove the extraneous effects. This indicates that the “true” trends are not likely to be indicative of actual climate change in which TC speed decreases over time. Note that because some scenarios are redundant for a given basin, some bars are missing as we plot only the first instance. The regions for (d)-(i) are as in Figure 1 with additional aggregate regions: (a) Global, (b) Northern Hemisphere, and (c) Southern Hemisphere.