What Provinces do Tropical Cyclones Visit?

One interesting question arose when I was looking at our previous Tropical Cyclone heatmaps. The Tropical Cyclone densities were mostly above the ocean. But where do these Tropical Cyclones go in the Philippines?

What provinces do these storms visit?

From the Philippine Administrative Boundaries by NAMRIA and PSA, we begin with a simple intersection between the Philippine Provinces and Tropical Cyclone tracks to see where they go.

Now, something is amiss here. Why does it seem like besides CAR, Cagayan and Quezon, Tropical Cyclones seem do visit Palawan the most? CAR, Cagayan and Quezon we can understand as they’re directly in the path of the storms, but why Palawan?

I think it’s simply a case of larger targets have a tendency to be hit more often than the others. As Palawan’s landmass is longer with respect to the North-South direction, there are higher chances of Tropical Cyclones intersecting this area. Comparing that with NCR that has a teensy-weensy area, storms are more likely to hit or “make landfall” along Palawan.

David vs Goliath

Now, how can we make a more meaningful map? We now take into consideration the size of a Tropical Cyclone.

The sizes of Tropical Cyclones vary greatly. Their radius can be as small as around 50km, and can span up to hundreds of kilometers (see the paper “Observed Tropical Cyclone Size Revisited” by Chavas et al, 2016 for a more in depth discussion).

This means that we shouldn’t just be looking strictly at the tracks of the storm, but the areas around the track as the effects of a storm is not focused on its track

Now, as a quick exercise (not really quick because it took the whole night to run in my laptop), I generated a 100km buffer around the tracks to see how many times a province has been within a Tropical Cyclone’s 100km radius.

Now that’s a more believable map as it now looks closer to the heatmap that we made before.


There are some caveats with the previous maps, however. As mentioned earlier, provinces differ in sizes and may not be the best represent the tracks of Tropical Cyclones. The radius 100km is totally arbitrary on my part, and better analyses could be done if we took the individual radii of each Tropical Cyclone. Unfortunately this data is not readily available.

Second the strength of Tropical Cyclones weren’t considered in this story. While certain provinces experienced a number of Tropical Cyclones we haven’t really looked at how strong these Tropical Cyclones were. It is for consideration for a future story.

But in the end how you make your maps totally depends on what you really want to see.

Say if you are a scientist and you want to look at the general trends of Tropical Cyclone tracks, heatmaps provide a perspective with a finer resolution.

But for example if you are a policy maker and would want to see which provinces experience more Tropical Cyclones, the purple map above might be a better visualization.


As a post-script, I would just like to thank David Garcia for continuing to serve as my inspiration to continue making these maps (as I continue to procrastinate for my thesis).

He makes very, very, very beautiful maps using data that are open and free.

For an example related to this story:

He also used the same IBTrACS dataset for this map

For more beautiful maps, do check out his website Mapmaker: Maps by David Garcia (https://www.mapmakerdavid.com/)


References:

  1. Chavas, D. R., Lin, N., Dong, W., & Lin, Y. (2016). Observed Tropical Cyclone Size Revisited. Journal of Climate, 29(8), 2923–2939. https://doi.org/10.1175/JCLI-D-15-0731.1