Extreme Weather #2 – Trends in Frequency and Intensity of Tropical Cyclones out over the Ocean
In Extreme Weather #1 we looked at trends in landfalling tropical cyclones (TCs) where data goes back over 100 years. Thankfully, way more TCs form over the ocean and don’t hit land. Trends on these would be informative – are they getting worse?
There isn’t much quality data before satellites started going up around 1980, so we have good data for over 40 years. More coverage was added around 1990 so we have even better data over the last 30 years.
What does the latest IPCC report say? Chapter 11 of AR6 covers extreme weather.
Here’s the simple version:
There are significant positive global trends in TC intensity.
The actual text, from p. 1585, is in the Notes at the end of this article.
This seems like bad news but it’s actually good news.
The Executive Summary for the chapter includes the “bad news”, p. 1519:
It is likely that the global proportion of Category 3–5 tropical cyclone instances has increased over the past four decades.. The global frequency of TC rapid intensification events has likely increased over the past four decades. None of these changes can be explained by natural variability alone (medium confidence).
I was confused when I read this section of the report and the paper referenced – Kossin et al., “Global increase in major tropical cyclone exceedance probability over the past four decades”, 2020. I’ve read a number of papers on TCs in the satellite era and “getting worse” didn’t seem correct. I found a paper from Klotzbach et al 2022 in my files and reread it. Both Kossin and Klozbach are heavily cited in this field, including by this IPCC report and the previous report (AR5).
Here’s Klotzbach 2022:
This study investigates 1990–2021 global tropical cyclone (TC) activity trends, a period characterized by consistent satellite observing platforms. We find that fewer hurricanes are occurring globally and that the tropics are producing less Accumulated Cyclone Energy—a metric accounting for hurricane frequency, intensity, and duration.
Here’s Kossin 2020:
Here we address and reduce these heterogeneities and identify significant global trends in TC intensity over the past four decades. The results should serve to increase confidence in projections of increased TC intensity under continued warming.
I emailed Phil Klotzbach asking for clarification – different dataset? different time period? looking at a different metric? and he very kindly replied within 24 hours explaining. (I’ve emailed a number of climate scientists during the years of writing this blog and have found them to be exceptionally responsive, courteous and helpful).
Now it’s clear. And I should have figured it out myself. Here is my plain English version:
The number of category 4-5 TCs (the most extreme) hasn’t changed. The number of category 1-3 TCs has reduced.
So this seems like good news. We can express it as “the percentage of the most extreme TCs has increased” but that’s just a confusing way of saying the same thing.
For people still confused, like a couple of friends I explained this to.. suppose the number of murders is flat but the number of other violent offences has reduced. We could say “violent crime is down”, or we could say “extreme violence has increased (as a percentage of overall violent crime)”. The first one is the plain English version.
Now, we’re looking at a short duration – 30-40 years. Is the trend due to climate variables like La Nina? Will the trend continue? Reverse? All good questions, perhaps to be considered in a future article.
This aim of this article is about the simpler question of what has been observed about trends in tropical cyclones out over the oceans. We’ll let Phil have the last word:
We find that fewer hurricanes are occurring globally and that the tropics are producing less Accumulated Cyclone Energy—a metric accounting for hurricane frequency, intensity, and duration
Notes
Text of AR6 on TC trends in the satellite era, from p. 1585:
There are previous and ongoing efforts to homogenize the best-track data (Elsner et al., 2008; Kossin et al., 2013, 2020; Choy et al., 2015; Landsea, 2015; Emanuel et al., 2018) and there is substantial literature that finds positive trends in intensity-related metrics in the best-track during the ‘satellite period’, which is generally limited to around the past 40 years (Kang and Elsner, 2012; Kishtawal et al., 2012; Kossin et al., 2013, 2020; Mei and Xie, 2016; Zhao et al., 2018; Tauvale and Tsuboki, 2019).
When best-track trends are tested using homogenized data, the intensity trends generally remain positive, but are smaller in amplitude(Kossin et al., 2013; Holland and Bruyère, 2014).
Kossin et al. (2020) extended the homogenized TC intensity record to the period 1979–2017 and identified significant global increases in major TC exceedance probability of about 6% per decade.
In addition to trends in TC intensity, there is evidence that TC intensification rates and the frequency of rapid intensification events have increased within the satellite era (Kishtawal et al., 2012; Balaguru et al., 2018; Bhatia et al., 2018). The increase in intensification rates is found in the best-track and the homogenized intensity data.
References
Seneviratne et al, 2021: Weather and Climate Extreme Events in a Changing Climate. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change
Global increase in major tropical cyclone exceedance probability over the past four decades, Kossin et al, PNAS (2020)
Trends in Global Tropical Cyclone Activity: 1990–2021, Philip J. Klotzbach et al, GRL (2022)
SOD: When AR6 appeared, I re-analyzed the data in Table 1 of Kossin (2020), AR6's primary source. Suspiciously, the data was not organized in a way that made it easy to subtract the change in the North Atlantic – where I expected the trend analyzed by Kossin to be biased by the AMO – from the global data. In the global homogenized data (ADT-HURSAT). After simply rearranging the data, I calculated:
1) There was a 5% increase in the total number of global TCs between Kossin’s early (1979-1997) and late (1998-2017) periods and a 5% DECREASE in global TCs outside the North Atlantic. The increase in the North Atlantic was 102%! (The next biggest change was a 24% DECREASE in the South Pacific.)
2) There was a 20% increase in the total number of major global TCs and only a 4% increase in major TCs outside the North Atlantic. The increase in major TCs in the North Atlantic was 289%! (The next biggest change was a 38% increase in the Southern Indian Ocean.)
3) There was a 4% increase in the percentage of major TC’s globally and only a 2% increase outside the North Atlantic. This increase in major TCs in the North Atlantic was 16% (18% to 34%).
Based on the data in Kossin (2020) and contrary to its conclusions, THERE HAS BEEN NO GLOBAL TREND IN TCs. There is merely one basin, the North Atlantic, which – during the period of this analysis – has behaved radically differently from the other six basins for well-understood reasons. The trend outside the NA appears negligible.
At this point in my research, I got bogged down in trying to understand what is known about long-term changes in TCs in the North Atlantic. In Part 1, SOD shows the lack of a long-term trend in the number of landfalling hurricanes in the NA. For all NA hurricanes, the simplest information I located was from Klotzbach (2015), which shows three active periods in the NA (1880-1900, 1926-1969, and 1995-2012) and two inactive periods (1901-1925 and 1970-1994) with 61% correlation between ACE and the AMO Index in any year. The active periods had ACE (accumulated cyclone energy) 2-3 fold higher (my estimate) than the inactive periods. This is crudely comparable to the 102% increase in NA TCs and 289% increase in NA major TCs in Kossin (2020). Klotzbach summarized the data differently: The 25 years with the highest AMO index mostly from Kossin’s late period) had 3.2-fold more major hurricane days than the 25 years with the lowest AMO (mostly from the early period). There is no reason to believe that the enormous change in the NA between Kossin’s early and late periods has anything to do with global warming.
http://www.novaweather.net/Uploads/ngeo2529.pdf
Why have Koonin and Klotzbach reached contradictory conclusions about the change in hurricanes?
1) The technology for assessing hurricanes has improved over time. Satellite images from space are processed to extract information about wind speed. If I understand correctly Klotzbach is using "best tracks" data and Koonin has reduced the resolution of satellite images of recent satellite images to the lowest resolution of match the resolution of the oldest images before processing them. In theory, the homogenized ADT HURSAT data should be superior to the apparently non-homogenized data Klotzbach is using. This leads to a big decrease in the total number of hurricanes being counted: 26% in the early period, 36% in the late period. For major hurricanes, an 8% decrease early major hurricanes and at 31% decrease in late major hurricanes. One might expect that homogenization to have caused Koonin to report a smaller increase than Klotzbach, but other factors must be more important.
2) Klotzbach studied the period 1990-2021, which isn't highly perturbed by the active phase of the AMO from about 1997-2012, while Koonin used 1979-2017. In an earlier paper Koonin started in 1982, but this new paper creates missing data for one ocean basin that wasn't monitored by satellite form 1979-1982. Koonin's early period is mostly from when the AMO Index was negative and his late period is mostly from when the AMO is positive. As I noted early, there is no global trend in Koonin's data outside the North Atlantic.
3) If you check out Koonin's Supplementary Figure 3, Best Tracks and ADT HURSAT differ substantially (often 20 kt in wind speed? more than enough to convert category 2 to 1 or 3). Quantifying hurricanes "activity" by the peak wind velocity in one 6-hour period of several dozen periods is not a very reliable process. I intend to pay more attention in the future to accumulated cyclone energy, which takes into account several dozen processed images per hurricane. It varies with v^2, as does the drag force in physics.
2) Koonin has fi