Friday, March 3, 2017

Anatomy of a Coastal Storm: March 10, 2016

Nuisance flooding in Freeland, Washington on Whidbey Island after the March 10 storm.  Unknown photographer, photo courtesy of Island County Department of Natural Resources.  
We are coming up on the one year anniversary of the 10 March, 2016 storm that really walloped many communities in northern Puget Sound (see this link as well) and on the Washington Coast.  It was an interesting storm for a variety of reasons and in particular it seemed to impact coastal areas in northern Puget Sound in ways that those communities were not accustomed to. Perhaps as a result I've heard people refer to the March 10th storm as a "100 year storm" from the standpoint of its coastal impacts.  As a result I wanted to use this blog to look in a bit more detail at some of the aspects of this event in in relationship to the coastal impacts associated with the event.

Overwash of sand, gravel and wood on to a road near Oak Harbor, Washington.  Photo courtesy of Lori Clark, Island County Department of Natural Resources
The photos above, and indeed many of the most significant impacts associated with this storm were focused in Island, Skagit and Whatcom Counties, but there is a distinct lack of tide gauge data up in those parts.  So I'm going to focus first on Port Townsend, where there is a tide gauge.  To be clear the storm was exciting there too.  The photo below shows a distinct debris line and nuisance flooding adjacent to the Jefferson Title Company building in Port Townsend on the morning of March 10:

photo courtesy of David Wilkinson
Here are the water level data collected between 5 March and 15 March, 2016 at the P.T. tide gauge, located just east of where this photo was taken:


When interpreting these data I am going to distinguish between three different water levels; first the "astronomical tidal water level", which I will also refer to as the predicted tide (shown in blue in these plots); next the non-tidal residual, which I will refer to as "storm surge", which is the difference between the predicted (or astronomical) tide and the measured water level (shown in purple in these plots); and finally the actual measured water level, or "still water level", which is the sum of the two (shown in green in these plots).

So in P.T. the total measured water level measured during this event WAS high - the peak measured on the morning of March 10 was 3.41 m relative to MLLW...but this is nowhere near a "100 year coastal flood water level" for P.T.  In fact since the current tide gauge was installed in 1972 that water level has been exceeded numerous times, with a maximum measured water level of 3.57 m relative to MLLW reached on 12/10/1993.  NOAA places the March 10, 2016 measured water level as having somewhere around a 1-in-5 year recurrence interval, or a roughly 20% chance of occurring in any given year.  Here is another way of looking at it:

These are histograms of hourly water level data dating back to 1972 in Port Townsend, with the metrics associated with March 10 in red.  At the top is the predicted or astronomical water level...and here the predicted tide on this day wasn't particularly high, but the NTR (the panel at the bottom) is pretty high, but again nowhere near the highest on record (that record was set on 1 Jan 1997 in Port Townsend when the non-tidal residual was roughly half a foot higher than the surge on March 10). So indeed, the "still water level", as measured at the tide gauge WAS indeed high, primarily because of some pretty good "storm surge", the peak of which more or less coincided with high tide...but nowhere near a 1-in-100 year event.

So what compelled us to look at March 10th as extreme?  Well, clearly wind, and wind-driven waves, played a major role in driving flooding during this event.  Here is another photo taken the morning of March 10, 2016, looking south from the south-facing shoreline of downtown Port Townsend:

Photo courtesy of David Wilkinson
There are no wind data available for that event from the P.T. tide gauge, so I turned instead to average wind data (these are averages over an 8 minute period) dating back to 2004 available from NOAA's Hein Bank buoy in the Strait of Juan de Fuca, and in particular wanted to look at winds blowing from the south, so trimmed those data to winds that were recorded as coming from between 120 and 210 degrees.  The histogram of those data suggest that extreme winds here would be those in the 18-20 kt range...


and on March 20 that buoy in the strait recorded peak average winds from the south on the order of 18 kts:


so pretty strong winds.  Again, not horribly strong from the standpoint of the available record dating back to 2004, but pretty strong.  So this leaves us with the notion that what MIGHT have been unique about this storm wasn't so much the relative magnitude of any of the individual processes at play (i.e. astronomical tides, storm surge, and wind) but rather the co-occurrence of these factors...i.e. what is the chance that you get a relatively high astronomical tide, coupled with a strong storm surge, coupled with pretty strong south winds that kick up relatively large wind waves (large by Puget Sound standards).  Taken together, maybe those things do amount to a pretty darn rare event, possibly even a 1-in-100 year event...but calculating that joint probability distribution will have to wait for a later date.