When we live in the West Indies, or in other exposed areas, the risk of cyclones is an integral part of our life 6 months a year. It influences our decisions, our choices, our vacations, companies organize themselves in front of it and politicians try to apprehend it as best as possible (there, there is progress to be made).
It is also a source of fantasies, fake news and a lot of irrationality. The aim of this article is to make it more palpable and to give it its proper importance, without generating unnecessary stress and without passing over it in silence. All accompanied by examples and statistics. Let's go !
Introduction to the concepts of cyclonic risk and warning
WARNING: whatever the statistics and the associated risk, the reality is that in the hurricane season and in the short term, the risk is proven and almost identical for all regions. Once that said, let's move on to the notions of risk.
It is necessary to separate 2 concepts of risk: the long-term statistical risk and the “real” risk at the end of a current season. As I indicated above, the seasonal risk over a period of a few months is difficult to quantify at the scale of a territory because it is very cyclical. We must therefore take it with the same approach regardless of the territory in which we are. It is the end of June and overall the risk for Martinique, Guadeloupe or the northern islands is identical at 5 months.
In this article I will therefore focus on trying to quantify the long-term risk (units in decades) for the different zones and territories. This quantification will be based on cyclone statistics since 1950, ie 858 cyclones listed. This is a sufficient sample to draw some trends.
Zone size scale
To assimilate the statistical risk, it is necessary to understand the size scales and the risk exposure of the areas we are going to talk about:
- The Atlantic basin: 858 cyclones since 1950
- The West Indies: 103 cyclones crossed it over the same period (12% of the total)
- The most impacted island: 15 cyclones have passed directly over Guadeloupe since 1950 (1,7% of the total and 14,5% of the arc)
As we can see, at the scale of a territory, even that most affected historically in the West Indies, the statistical risk becomes extremely low compared to the activity of the Atlantic basin. And even the West Indies as a whole, which is nevertheless in the front row, represents “only” 11% of the impacts of cyclones, while at the same time the US coasts are around 25%. So the statistical risk must be understood on its own scale.
Long-term risk factors
If we take the statistics since 1950 by territory, we can determine 3 risk factors: the latitude of the territory, its linear distance of exposure and finally an inevitable part of uncertainty. Even with a statistical sampling approaching 1.000 samples, it is absolutely not possible to ignore the part of uncertainty.
The latitude of the territory
This is the factor that is at the base of the biggest fantasies. “The further north we go in the arc, the more cyclones there are”… it's the vox populi who says it!
In order to determine the reality of this statement, I divided the West Indies into 4 zones: The south, from Grenada to Saint-Vincent, the south-central from Saint Lucia to Martinique, the central north which goes from Dominica in Guadeloupe and the north for the rest. From this breakdown, I analyzed the trajectories of the cyclones that crossed the arc to distribute them in each zone and determine to what extent there is a significant difference and the zones which are above or below the average which is therefore 25% since there are 4 roughly equal zones.
The numbers are quite interesting:
(The total of the 4 zones is greater than that of the arc (97) since quite a few systems have passed through several zones.)
We see that there is indeed a preponderance of the passage of cyclones between the north and the south of the arc. But it is very little marked between the north of the arc and the center-north. Only 2% difference over 70 years is not significant and finally the northern half of the arc has a roughly equivalent risk from Dominica to Angilla.
Going down in the southern half, the difference becomes more and more marked with 6% less for the center-south and another 6% of months for the south.
We therefore actually have 3 statistical zones: The northern half which represents 60% of cyclone passages, the south-central and the south with a difference of around 6 or 7% between each. It is very marked between the northern half and the south.
We can conclude that the vox populi was generally right and that the northern half of the arc is more exposed than the south. But the differences are less important than what the imagination wanted to say. Guadeloupe is almost as exposed as the Northern Islands for example, which many people in the Northern Islands could not really imagine. And it still passed 17 cyclones in the south which remains significant.
I would do these comparisons again later, but breaking them down by intensity of the systems.
The linear exposure distance of each territory
This is an element that no one takes into account but which is nevertheless essential and which largely explains the differences in exposure between 2 nearby islands. To determine this distance, we will separate the arc into 2 parts: from Martinique to the north, 90% of cyclones cross on a road between the west and the north-west. And in the south of Martinique, cyclones crossed 90% on a road to the west. So from Martinique to the north, the exposure area of each island is the south coast and the east coast while in the south of Martinique, the exposure area of an island is essentially its east coast. I will take 3 examples for the demonstration: Saint Lucia, Martinique and Guadeloupe.
We can see that according to the latitude (and therefore the axes of circulation of the cyclones) and according to the size of an island, the exposure is really different. Guadeloupe has a linear exposure three times greater than that of Saint Lucia in this example.
An island with a much higher linear exposure, at roughly equivalent latitude, will statistically be much more likely to be directly affected by the passage of a cyclone.
Uncertainty and the cyclical factor
Beyond all the statistical explanations, there remains uncertainty and the economic situation. No statistic had seen Lenny cross the Caribbean Sea from west to east and torment the Northern Islands for very long hours after being appointed off the coast of Colombia! Or even Ivan who managed to strengthen himself in Cat. 4 while running along the 10N, which is quite surprising in terms of physics. Like what, even in Granada, the most southerly of the Arc, we are not immune to bad karma. So yes, the uncertainty and the global meteorological situation (various oscillations) have their say on the intensity of a season but also on the trajectories of the systems, which sometimes lead them to brave statistical trends.
Conclusion
You have to take into account that these stats and calculations have biases. They are not at all applicable to very little exposed territories because of their size and / or their latitude. If we take the example of Carriacou, it will be difficult to measure the risk with these methods because of its very limited exposure. It is market conditions and luck that will determine the risk overall, as in the short term. If we take that of Saba, yet in the very active area of the arc, the risk of the center of a system passing directly over it is extremely low due to its tiny size. But for areas such as the French departments, Dominica or Saint Lucia, we see that the assessed risk is in line with the past.
These figures show the predominance of certain risk factors over the long term, over several decades. And it's a safe bet that in 30 years they will still be verified even if, in the meantime, over a season, they can be completely called into question. And don't forget that even Guadeloupe, yet the most affected island of the West Indies, is only directly impacted on average by 1,7% of the cyclones that affect the Atlantic basin (and a little more than 2% pass within 30 km).
I hope this article will allow you to better understand the risk for your territory. Of course, it cannot be a generalization and even less a basis for prevention. I just tried to show you with statistics why this island is more affected than another. You now know why Guadeloupe is much more affected than the other islands overall!
Finally, I will repeat something that I often say: “As a human, by nature selfish, I would always prefer an extremely active season that does not affect me rather than the least active of all time with a single cyclone but passed over me!“. Statistics and forecasts can say what they want and be very relevant, the reality of the moment in a given place is absolutely IM-PRE-VI-SI-BLE!
