The IPCC recently released the policy-maker’s summary (SREX-SPM) on extreme weather and climate events. The background for this report is a larger report that is due to be published in the near future, and one gets a taste of this in the ‘wordle‘ figure below. By the way, the phrase ‘ET’ in this context does not refer ‘extra-terrestrial’, and ‘AL’ is not a person, but these refer to the way of citing many scholars: ‘et al.‘
The fact that the summary is released before the main report is bound to cause some confusion, and has lead to a number of false allegations in the past, such as the main report being written to suit the conclusions of the summary. This is not the case, but I personally think that the IPCC handles the release of these reports in a strange way.
The main report has already been written, but there are some fine details that need to be approved by the member states before it is finalized. My understanding is that the whole process will be open and transparent, and that the previous drafts and review comments will be available in time. Those who already have read the main report are not supposed to cite it before it’s out.
I must also confess that one of the aspects that I’m most curious about concerns tropical cyclones (TCs). Hence, these phenomena was one of the things I looked at first. Here are some quotes:
Average tropical cyclone maximum wind speed is likely to increase, although increases may not occur in all ocean basins. It is likely that the global frequency of tropical cyclones will either decrease or remain essentially unchanged.
The message from the summary of policy-makers is therefore that it is likely [66-100% probability] that there will be fewer or same number but more intense tropical cyclones (including tropical storms, hurricanes, and typhoons) in the future. This conclusion is not new, however, as it was also the concusion of the AR4, as well as the most recent WMO consensus statement on tropical storms.
A combination of stronger tropical cyclone maximum winds but fewer tropical cyclones is nevertheless quite interesting. My feeling is that this statement is still a bit premature, as it surely is based on projections made with global climate models (GCMs). The tropical cyclones are represented differently in the GCMs compared to real world measurements, where the wind speed changes continuously in space.
The message from the SREX-SPM is similar to that of a 2010 study from Nature Geoscience, for which the abstract reads (my outline):
Whether the characteristics of tropical cyclones have changed or will change in a warming climate — and if so, how — has been the subject of considerable investigation, often with conflicting results. Large amplitude fluctuations in the frequency and intensity of tropical cyclones greatly complicate both the detection of long-term trends and their attribution to rising levels of atmospheric greenhouse gases. Trend detection is further impeded by substantial limitations in the availability and quality of global historical records of tropical cyclones. Therefore, it remains uncertain whether past changes in tropical cyclone activity have exceeded the variability expected from natural causes. However, future projections based on theory and high-resolution dynamical models consistently indicate that greenhouse warming will cause the globally averaged intensity of tropical cyclones to shift towards stronger storms, with intensity increases of 2–11% by 2100. Existing modelling studies also consistently project decreases in the globally averaged frequency of tropical cyclones, by 6–34%. Balanced against this, higher resolution modelling studies typically project substantial increases in the frequency of the most intense cyclones, and increases of the order of 20% in the precipitation rate within 100 km of the storm centre. For all cyclone parameters, projected changes for individual basins show large variations between different modelling studies.
In GCMs, these phenomena appear as vortex-like features in the discrete representation of the flow represented by neighboring grid boxes. It’s quite remarkable that these phenomena are present at all in these models (sometimes they are not, though), even though they may have too weak or exaggerated features. In the real world, the definition of a tropical storm is a synoptic scale low-pressure system with maximum sustained surface wind speed greater than 17 m/s, and in hurricanes greater than 33 m/s.
If we look at wind speed measurements at a given location, we see that there are relatively few days with zero wind, more often there are moderate wind speeds, and it is typically rare when the wind speed exceed the threshold defining a tropical storm or a hurricane. In statistical terms, the wind speed may be described by a distribution function – e.g. a Weibull distribution (e.g. here and here). The situation is illustrated below showing wind speed statistics, where the curve is the probability distribution function (pdf) for the wind speed and where the x-axis represents the wind speed and the y-axis the likelihood (frequency) of occurrence. The threshold marking tropical storms is shown as the first vertical line (the others mark typical TC categories), and the area under the curve to the left of this treshold (denoted “a” in the diagram) is proportional to number of observations (e.g. days) with no tropical storms. The area above (“A”) is proportional to the frequency of tropical cyclone occurrence.
Let’s consider the implication of fewer tropical cyclones but an increase in their intensity. In terms of wind speed statistics, this suggests a shift in the pdf (grey gurve in Fig. 3), with an increase in the area under the curve with wind speeds lower than 17 m/s (“a”). This also implies a decrease in the area under the curve for which wind speeds exceed 17 m/s (“A”), as the area under the total curve of a pdf must be constant (unity by definition). But if the tropical cyclones are getting more intense (increased mean TC maximum wind speed), there must be a second threshold, e.g. 33 m/s for which the area under the curve for the new pdf is greater than for the old curve.
It is certainly possible that the requirement in these changes in the wind speed statistics can occur, but the question is whether it is likely and whether we are able to detect this. If the shape of the wind speed is constrained to being a Weibull type, then it is easy to simulate the probability that the area under the curve is greater both for the portion of the curve with wind speeds lower than say 17 m/s and greater than 33 m/s (Monte-Carlo simulations – R-script). The fraction of Weibull shapes satisfying this, accoring to a simple Monte-Carlo simulation, is 1.9% (i.e. not very likely). Another issue is the required size of a statistical sample to be able to detect such changes, and the GCMs’ ability to provide such details (there are not that many simulations with high-resolution GCMs, the number of TCs is sensitive to a number of factors, such as ENSO, AMO, MJO, and the annual cycle – I must admit that I don’t know if the GCMs capture these dependencies well).
An analysis of empirical data provides strong indications that the number of TCs, N, varies with the surface area of warm surface (warmer than 26C). However, N also depends on the number of triggering events (organized convection e.g. African easterly waves in the North Atlantic), the wind shear, and the convective available potential energy (CAPE). The question about future trends in number and intensity of TCs depend on these aspects, in addition to the magnitude of the sea surface temperature.
It is well-known that tropical cyclones are influenced by a number of factors, such as El Nino Southern Oscillation (ENSO), the Madden Julian Oscillation (MJO), wind shear, organized convection, and sea surface temperatures. The GCMs, however, provide different accounts as to which direction the ENSO takes, struggle with reproducing the MJO, and may have biases with respect to the sea surface temperatures and wind shear. There are also some problems, as they produce a spurious “double” inter-tropical convergence zone (ITCZ), as well as biases in sea surface temperatures and wind shear.
Furthermore, small-scale processes may still not be sufficiently resolved by the GCMs used for projecting the future climates. Having said that, high-resolution global atmosphere models provide realistic-looking pictures of tropical cyclones, and the question is not whether the models in principle are capable to capture these events, but rather whether the current set-up of GCM experiments is sufficient for providing reliable information about how these will evolve in the future. The main report may shed more light on this, and we should keep in mind that the models must be evaluated against the past and reproduce known dependencies, in addition to reproducing the wind speed distributions.
There is still a debate about the past trends (see previous discussions here, here, here, and here). Has the tropical cyclone activity or the number of cyclones increased? Note, the trends may not not necessarily linear, and if one tries to fit a straight line in time, it may not provide the best picture of the situation. As long as we have no reliable records on tropical cyclones for the past, I’d argue that we don’t know how well our models are able to capture long-term changes in tropical cyclones. However, this is only one small issue (see Fig. 1), and the SREX-SPM covers many other topics on which I have little expertise.