12,270 Publications · 71,396 Citations

Kossin et al. (2014) found that tropical cyclones reached lifetime-maximum intensity (LMI) farther from the equator in 1982-2012. We used their rules (all best-track systems, agency with the greatest LMI, first LMI time) on IBTrACS v04r01 and reproduced the trends: Northern Hemisphere (NH) +54 +/- 43, Southern Hemisphere (SH) +59 +/- 49 and global (NH minus SH mean latitude) +114 +/- 70 km per decade. We then forecast the 13 new seasons 2013-2025 with the extended 1982-2012 trend and with the 1982-2012 mean (no change), and wrote the plan before the test. Global and NH: the 2013-2025 means were 148 km [16, 280] and 87 km [10, 164] poleward of the old means and inside the trend interval; the trend forecast had 41-42% lower mean squared error. SH: the result was between the two forecasts (skill -0.07). About 30% of the 1982-2012 global trend came from changes in basin storm shares; this part did not continue. The shift within basins continued at about the old rate. Weak systems (LMI below 34 kt) reach LMI nearer the equator in most basins and their share fell; with basin and intensity mix removed, the 1982-2012 global trend was +56 +/- 69 km per decade, but 2013-2025 values were still 111 km [-5, 228] above the old mean. Storms of 64 kt or more gave no clear signal, and WMO-agency winds made the trend forecast worse than no change. Most of the NH rise came from the western North Pacific during a very negative PDO; with the PDO as a regressor, the trend term there went to about zero, and a PDO-only model forecast 2013-2025 as well. Limitations: inhomogeneous best-track intensities, no ADT-HURSAT reanalysis, only 13 test seasons, and post hoc intensity and PDO analyses.