Looking at Historical Hurricane Tracks
With the traditional North Atlantic hurricane season about to begin on June 1, interest in previous tropical storms has begun to rise. I am occasionally asked if one could use the Historical Hurricanes Tracks web site to look at past hurricane tracks to find a storm similar to some current one and then use that previous storm track to predict where some current storm is going. The Historical Hurricane Tracks (HHT) site shows the data from the International Best Track Archive for Climate Stewardship (IBTrACS), a compilation of tropical cyclone tracks from meteorological observing centers around the globe. Figure 1 shows the tracks of all tropical storms in the IBTrACS archive, 1848 – 2013.
From a casual examination of those tracks, it does appear that, within each basin, there might be a general pattern, somewhat similar to the Nike Swoosh, or a large letter C: storms move westward and, sooner or later turn poleward, then head back to the east and poleward. Hurricanes Bill, 2009 and Igor, 2010 (Figure 2.) illustrate this “pattern.”
However, does this mean that we can use tracks that start similarly to predict where a particular storm will go? Not really. Check out Hurricane Andrew, 1992,and see how similar it was to Bill and Igor over about half of its track but then it “turned left” when they “went right” (Figure 3). You can also examine these storms directly on the HHT web site.
Similar season, similar storms?
Of course, storm dynamics can vary throughout the year, so perhaps the above storms were not a good sample. What about storms that passed through the same area at similar times? Let’s take a look at all the storms that went within 100 miles of a location during one month. For this example, I’ll pick September and a location at 11° N., 55° W. The Historical Hurricane Tracks site shows 15 storms met these criteria, but I’m going to omit two storms that remained tropical depressions and three storms that lasted less than a week, so didn’t have much of a track to examine. That leave 10 strong, longer-lived storms to examine, Figure 4.
Of these ten storms, four followed relatively similar C-shaped tracks (Figure 5), three had somewhat more straight tracks before curving, if at all (Figure 6) and the remaining three were clearly Frank Sinatra fans, since they did things “their way” (Figure 7).
From just this small sample, I think we can conclude that, although hurricanes can follow the tracks of previous storms, it isn’t really something to be relied upon. They can just as often go in completely different directions. That is why it is important to rely upon the predictions from NOAA’s National Hurricane Center for each storm and to pay attention to watches and warnings when they are issued.
Some crazy storm tracks
On a somewhat related note, I thought I’d share a few interesting storm tracks I’ve stumbled across while using the HHT site. These are just tracks I thought were quite odd and made me question my assumptions about how hurricanes “generally” behave. Take a look at the pictures or click on the HHT link to explore them more thoroughly on the Historical Hurricane Tracks site.
I assume tropical storms will generally move south-to-north over their lifespan in the northern hemisphere, since they are part of a global system to move heat from the equator to the poles. Here are a couple that broke that assumption. The unnamed 1934 hurricane actually started fairly far south in the middle of the ocean, went north to whack Bermuda, then looped around and went south and east and ended over the Dominican Republic, further south than it originated. Perhaps even more oddly, Hurricane Lili, 1984, started over in the mid-Atlantic at the latitude of North Carolina, then went on a looping route south and east, again ending up at the Dominican Republic, about 15° latitude farther south than where it began.
Sometimes storms can also alternate between big swings east and west. In Figure 9, you can see Hurricane Ginger, 1971, which went about 20° longitude eastward before turning around and heading back westward. It did that with relatively little variation in latitude. Also shown is Tropical Storm Gordon, 1994, which seemingly slalomed between Caribbean islands as it moved north to strike Florida a couple of times.
Lastly, consider the pair of storms in Figure 10. Hurricane Kyle, 2002, which must have been a forecaster’s nightmare, since it went north and south, east and west, slowed down and sped up. In contrast, Tropical storm Bret, 1993, chugged along from east to west, varying less than 2° in latitude over its week-long life. It, too, might have been a forecasting challenge, since it never turned. Very odd, at least to me.
Try it yourself!
I enjoy looking at tracks like those above because they make me ponder the complexities of atmospheric dynamics and how even big, powerful systems like tropical cyclones are really just eddies in an ocean of air and can get pushed around by other air masses, bump into things on the “bottom” and get torn up by currents. It seems much like eddies in stream, but on a much much larger scale. I think that is cool!
The September example above was created by picking a random location and looking at all the storms in a particular time frame. You can try this for yourself at the HHT site. I encourage you to pick a spot and use the filtering tools to narrow things down by time if you want. I think you will be surprised by the variety of storm tracks that appear at any given location. It really drives home to me the complexity of trying to predict hurricane tracks: there are a lot of different forcing variables and each case is different.
While you were poking around on the HHT site, did you find some interesting storm tracks? How about sharing them with others in the comments below. You can use the “share” link on the HHT site to give a (large, ugly) URL that you can copy into the comments, or you can just post the name of the storm and the year. I look forward to seeing what you find!
[EDIT: Hyperlinks have been updated for the new viewer. Also please note that some storm track details may change through time as the tracks get revised when additional information is received. DLE, 7/17/2015]