2013 Hurricane season off to a slow start

The 2013 North Atlantic tropical cyclone count stands at 6/0/0 meaning 6 named tropical storms so far, and no hurricanes or obviously major hurricanes. This is a similar start to 2011 which stood at 8/0/0 until Irene formed and quickly became a hurricane near Haiti on 22 August. Going back a bit further, this is the longest wait for a hurricane since 2002 when the count was also 6/0/0 at the end of August and Gustav (not the one in 2008 that hit Louisiana, but a visitor to Nova Scotia, Canada) became a hurricane on 11 September.

Well back into the past archives of storms, we can find seasons without hurricanes at all. This could be partly due to the less ubiquitous observations available today from satellite imagery, but it’s unlikely that in 1907, all North Atlantic shipping would have avoided detecting a hurricane (weak tropical storms probably got missed quite routinely before 1950). That year, the count reached 5/0/0 for the entire season. But the futility record was soon shattered when 1914 produced just one tropical storm.

The modern (1993-2012) average frequency of named storms is 14 with 8 becoming hurricanes and 4 reaching major (category 3 or higher) intensity. The average had been more like 11-12 with 7 hurricanes and 3 majors for decades before 1993. There has been both an upturn in activity similar to earlier peaks in the 1950s and 1930s, as well as a slight inflationary factor from the ability to discover even the weakest and short-lived tropical storms. It’s doubtful that hurricane frequency since around 1900 has been much influenced by our ability to detect storms, but perhaps in the statistics from the 19th century, it would be necessary to adjust those numbers upwards by 10-20 per cent to bring them into line with modern reporting and observation. I could see where a 3-day hurricane between the Azores and Bermuda might go undetected by Atlantic shipping in 1850, but there were plenty of observers to catch these storms closer to land and certainly on land.

So, what’s causing this year to be so weak in production so far? Many who study tropical cyclone formation are blaming the influx of dry Saharan air out into the tropical Atlantic. Vast outflows of dry, dusty air have been noted in July and August and this seems to have been starving the cyclones that do get started from enough moisture at mid-levels to find the energy to intensify. Thus they have all shown a short burst of activity and died out. Another reason often quoted is the lack of supportive environments given by weak upper level lows in the subtropics, but of course that just shifts the question to what happened to them, and we don’t really have a very well-developed understanding of what causes such disturbances. They would be more likely to develop if the subtropical highs were perhaps a bit stronger and further north. Much of August has been wasted away in a low-energy upper-level pattern that has been associated with below-normal temperatures over much of the southeast U.S. (Atlanta had a four-day interval in mid-August with temperatures 10-15 deg below normal!).

I expect we will see a gradual awakening of the hurricane-creating machine in the next week or two, with a peak to this season likely to come late in September and perhaps into October. I have my eye on two conducive periods (from my own research), namely 18-27 Sept and 16-24 October for possible major storms. The east coast of the U.S. seems more likely to see them than the Gulf or Caribbean this year. Out of the activity I am expecting, there is something like a 30-40 per cent chance of one named storm remnant affecting the weather around Ireland or Britain and a 10-20 per cent chance of two such events. By that, I mean heavy rainfalls and/or stronger wind gusts than 40 knots. The chances that any named storm would ever reach any part of Ireland or Britain intact as a functioning tropical storm is always low, and the closest calls such as Debbie (Sept 16, 1961) which produced major wind damage in parts of Ireland, are still debatable as to whether they were extra-tropical before reaching nearby regions of the North Atlantic.

Debbie produced very similar wind patterns to transitioning storms that hit eastern Canada on a more regular basis. Just about every tropical cyclone that hits Newfoundland is about mid-way through extra-tropical transition when its winds reach the Avalon Peninsula (the most vulnerable area) and we frequently see small and disintegrating remnants of hurricane or tropical storm “eyes” on St John’s radar. But the 10-14 C water temperatures around southeast Newfoundland rapidly destroy any tropical characteristics these storms have left. Nova Scotia, however, can see intact hurricane landfalls such as Juan 2003 which did enormous damage around Halifax with cat-2 wind gusts, and so on rare occasions can New Brunswick (Edna in 1954 hit St. John, NB with very strong wind gusts). Southern Ontario seems to be more prone to heavy rainfalls from already-transitioned tropical cyclones coming inland from the Gulf of Mexico or the Atlantic. Hazel in 1954 produced massive flooding in Toronto that resulted in about 80 deaths. More recently, I observed nearly 100 mm of rain from Tropical Storm Candy (June, 1968) and from remnants of Hurricane Hugo (Sept 1989) which rapidly moved in from South Carolina. Wind gusts with Hugo (and Hazel) were around 120 km/hr so these tropical systems were not quite fully extra-tropical. There was even a very small eye feature left with Hugo and it went pretty much over my location at midnight, leading to a very brief drop in wind speeds and a complete turn from SE to NW winds there, within an hour.

The west coast of Canada saw one actual hurricane landfall. Freda (1962) was a member of the central Pacific family of storms, that form near or west of the Hawaiian Islands. These aren’t very numerous and some seasons come and go without this family showing up at all. Further west, there are often 20-30 typhoons in the western Pacific and here on the west coast of Canada we get hit with the remnants of those quite often. But Freda managed to accelerate past Hawaii and reached the Oregon and Washington coasts (on October 12, 1962, hence it’s known in the U.S. as the Columbus Day storm) with wind gusts to 120 mph. The centre came inland over Victoria B.C. late on that day and winds gusted to very strong although unrecorded speeds over Vancouver’s Stanley Park, causing a massive blow-down of mature trees. That was also the situation around Puget Sound in western Washington. The same location in Stanley Park got hit by a tornadic wind streak in December 2006 and many trees were lost then too. This 1962 hurricane could be compared to the 1987 windstorm in southern England although Freda was probably more tropical in structure. Both are likely very long-period return type storms, in the case of England we know that it might be on the order of 100-200 years but for western Canada it’s hard to say since records only extend back to the late 19th century here.

The record high number of named storms was 28 in 2005, when we had to resort to the Greek alphabet after exhausting ours (although we don’t use Q or U,X,Y,Z so that 21 storms were from the list of names, one “named” storm was not named but added to the list in reanalysis, and then six Greek letters were employed. The second most active season on record was 1933 with 21 storms, can’t say named because 1950 was the first year that names were actually used. For a few years they used Morse code type names, such as Able, Baker, and King. Then in 1953 came the switch to female names, for a few years they used the same ones each year, then had lists and began the practice of dropping severe storm names. The switch to alternating girl and boy hurricanes came in 1979. There are also three named sub-tropical storms in the 1970s, One, Two and Three. This practice must have died out soon after it started because 1979 has an un-named one. Before names were routinely given, major hurricanes became known by names that referred to their history, such as the Long Island Express of 1938 or the Great Galveston Hurricane of 1900. That one resulted in about eight to twelve thousand deaths and massive destruction in a city unprotected by high sea-walls as it is today. Two hurricane disasters in southeast Florida in 1926 and 1928 also had higher death tolls than Katrina.

My final thought on hurricanes is that you often hear they will increase in frequency and severity with climate change. This is yet to be proven conclusively. The historical record tends to favour the concept of constant natural cycles of hurricane activity, but in any case, the human impacts tend to be very much a factor of the hurricane tracks. Some very powerful storms like Gilbert (1988) have done relatively moderate damage because they hit sparsely populated coastlines or reached their maximum intensity over the oceans. Other less severe storms like Isobel (2003) did considerable damage because they had just the right tracjectory and interactions with high tides to max out their impacts. As with the theory of increasing tornadoes, this all has to be viewed through a lens of increasing population and vulnerability, but also, better warnings, awareness and improving building codes. That makes comparing data from period to period a very tricky business. My gut feeling is that neither type of severe storm is really that much on the increase, but that the past had just as many over long periods of time, and in fact, some of the impacts even with today’s warning systems would still be catastrophic if these past monsters reappeared today.

Peter O’Donnell

One thought on “2013 Hurricane season off to a slow start

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s