Wednesday, August 6, 2014

Nuisance Flooding and sea level rise

Nuisance flooding in Port Orchard, WA on January 6, 2010. Photo by Ray Garrido, courtesy of Washington Department of Ecology.

NOAA just released a brilliant analysis of patterns of "nuisance flooding" along the US coast. The report analyzes hourly water level records from 45 tide gauges around the country that:

1) have at least 30 years of more-or-less continuous water level data

2) have a "minor flood" elevation threshold defined by the National Weather Service.

The report also defines "nuisance" flooding as coastal water elevations that exceed that "minor flood" elevation as it is set by the National Weather Service, and are typically associated with "minor coastal flooding impacts". So we are talking about stuff like is shown in the photo above from Port Orchard...not just the catastrophic flooding associated with events like Superstorm Sandy or Hurricane Katrina. It turns out that this sort of nuisance flooding, while it doesn't get much media attention, happens all the time:

Average number of days per year with nuisance flooding for 2007-2009 for 45 US tide gauges. Figure 2 from Sweet et al 2014.

In the report they analyzed a variety of different components of flooding, including trends over time (expressed both as cumulative hours of flooding during the year, as well as individual days that flooding occurred), seasonal patterns (i.e. when during the year flooding tends to occur), and the contribution of various factors (astronomic tides, changes in mean sea level and non-tidal residuals) to driving total water levels that result in nuisance flooding. Here is an example for Charleston, SC which, like many east coast sites, has experienced fairly dramatic changes in patterns of nuisance flooding over the last century:

Figure from Appendix 2 of Sweet et al 2014 showing (top left)nuisance flood events (cumulative hours and impacted days) per meteorological year plotted with annual mean sea level, (bottom left) return level interval curves, (top right) maximum water levels (black) per calendar day over the 1980‐2009 period decomposed into a low‐frequency MSL cycle (blue dot), predicted tide (blue; no Sa and Ssa harmonic fits) and nontidal residual (NTR; green), and (bottom right) average nuisance flood events (shared y-axis) by calendar month over 1980‐2009.

I was particularly interested in the fundamental analysis in the upper left, and in particular the cumulative hours of the meteorological year (May of a year to April of the next, so winter is included) during which a minor flood threshold is exceeded. In their analysis only two of Washington's tide gauges were included. Seattle and Toke Point in the mouth of Willapa Bay. Here is Seattle:

A few things jump out - in most cases the rising occurrence of nuisance flooding is related to rising relative sea level, as is seen in the Charleston SC example above, as well as in Seattle. This isn't surprising - rising relative sea level means that the average level of the water is rising relative to the land, so flooding should occur. I was able to replicate their analysis process using data downloaded from NOAA, and also added a 10-year running average of the number of hours per year for which water level exceeded the nuisance flood threshold:

and ran it for Toke Point:

here, over the period of measurement (starting in 1979) there has been little to no systematic trend in relative sea level. Perhaps as a result, there is not clear pattern related to changing nuisance flood frequency.

I found only one other station that included enough hourly data to make this analysis worthwhile - Neah Bay - and applied Seattle's nuisance flood elevation threshold:

here, FALLING relative sea level has apparently resulted in a pattern of falling nuisance flood frequency. You can also start to see an apparent relationship between the average flood frequency (the black line), and annual mean sea level. Here is what it looks like for Seattle when they are plotted together:

This is a potentially useful tool, since it gives us a way to assess how many extra hours of flooding to expect for a given amount of relative sea level rise. I haven't done any formal analysis on this, but loosely it looks like relative sea level rise of about 10 cm leads to maybe about an extra hour of flooding over the course of the year, on average, in Seattle.

Sadly, most other stations in Washington don't have enough available hourly data to make this analysis worthwhile. Here is Port Angeles, for example:

There are a few interesting things here - the high water levels associated with the 1998 El Nino really stand out, but that year didn't see water levels exceeding the nuisance flood threshold. The big year for that was 2005, even though average sea level that year was almost 10 cm lower than in 1998. This is probably because of the importance of other factorss, including chance, in driving flooding in our area. Our large tidal ranges and complex tidal signal mean that a variety of factors have to line up to drive water level up into the nuisance flood realm - something along the lines of a big storm with low pressure, aligning with a spring tide period and a high tide. Rising relative sea level in the background really helps to exceed the nuisance flood threshold more frequently. Overall, in this case though, the record really isn't long enough to detect trends given all the variability.

Its also worth nothing that the "nuisance flood" threshold used here is somewhat arbitrary. In the NOAA report its an elevation that the National Weather Service has identified as being associated with "minor coastal flooding impacts". But if your home, or some important piece of infrastructure, is below that threshold, then you would see flooding more frequently then is suggested by this analysis. But the overall patterns should still hold up...

1 comment:

Dan McShane said...

Nice analyses and things to ponder. Thanks