UPDATE (July 3, 2012): I’ve provided this update on both of my Sallenger et al (2012) posts.
Tamino (Grant Foster) has once again written a misleading rebuttal, this time to my two posts about Sallenger et al (2012). Refer to his post Sea Level Rises … Tisdale Falls. In the past, I’ve taken the time to write detailed responses to his critiques. See here, here, here and here. I’ve also reposted a comment by blogger JJ that’s really worth reading here. Every time I do make the effort, however, many of those commenting on the cross posts at WattsUpWithThat suggest that there’s no reason for anyone to respond to Tamino’s nonsense, because few people find Tamino’s posts to be credible, that the methods and tactics Tamino employs have been found to be questionable in a number of instances by a number of other parties. See here and here, for example. So I won’t bother with a new post that provides a detailed response to Tamino this time. For those interested, I did respond to Tamino’s recent post in two comments at one of the WattsUpWithThat cross posts, where one of Tamino’s disciples introduced me to another of Tamino’s classic examples of misdirection. See my July 1, 2012 at 7:33 pm and July 2, 2012 at 3:46 am comments.
John Droz Jr. asked me to comment on the Sallenger et al (2012) paper Hotspot of Accelerated Sea Level Rise on the Atlantic Coast of North America. As you’ll recall, John Droz Jr. has reported on pending North Carolina legislation about rising sea level. Refer to the WattsUpWithThat posts here and here. Apparently, North Carolina lawmakers are concerned about the use of the word “accelerated” in Sallenger et al (2012). So I took an initial look at the sea level data for the Atlantic Coast “Hotspot”.
The shift in trends around 1990 noted in Sallenger et al (2012) appears to be a function of the natural variability of the North Atlantic.
In fact, Sallenger et al (2012) in their closing comments state that this is a possibility, but it was impossible for them to tell due to the limited availability of data (my boldface):
In regard to the role of cycles, the single ~ 60-yr pattern in Fig. 4a does not seem associated with 10–30 yr sea-level variations discussed in ref. 27. With our limited series length, the presence of cycles, for example associated with natural ocean variability and/or AMO, is indeterminate. In the Holocene geologic record of an NEH marsh, the authors of ref. 28found evidence of several rapid SLR increases separated by 900 yr or more that they associated with gyre changes. Regardless, our correlations suggest that should temperatures rise in the twenty-first century as projected, the NEH SLRD will continue to increase. If future sea-level variability is forced by aerosols and/or is part of a cycle, SLR in the NEH may also alternately fall below and rise above projections of IPCC scenarios alone.
Our analyses support a recent acceleration of SLR on ~ 1,000 km of the east coast of North America north of Cape Hatteras. This hotspot is consistent with SLR associated with a slowdown of AMOC.
In other words, Sallenger et al (2012) note there is an accelerated rate at which sea level is rising for a portion of the Atlantic coast of the United States, starting around 1990, but they don’t know the cause. In the opening sentence of the above quote, they comment about the observed 60-year cycles in the Atlantic Multidecadal Oscillation and how they do not agree with the results of their reference 27, which is Frankcombe and Dijkstra (2009) Coherent multidecadal variability in North Atlantic sea level.
The abstract of Frankcombe and Dijkstra (2009) reads:
Tide gauge records from around the North Atlantic are examined and found to show variability on 20–30 year time scales. Sea surface height variability along the western boundary of the North Atlantic shows a particularly strong and coherent signal. Similar variability is also found in an ensemble of runs using a state-of-the-art climate model (GFDL CM2.1). This sea surface height variability is linked to variability of temperatures in the upper layers of the ocean and thence to the Atlantic Multidecadal Oscillation. The variability is consistent with the excited internal ocean mode mechanism of multidecadal variability derived from idealised models and the timescale is consistent with that derived from observations of sub-surface temperature variability.
The last sentence Frankcombe and Dijkstra (2009) reads:
The long tide gauge records analysed here certainly point to the importance of the 20–30 year period over the 50–70 year period more commonly associated with the AMO.
Immediately, one wants to examine a time-series graph of the “Hotspot” sea level data in the Sallenger et al (2012) paper to get an idea of what the multidecadal variability in the data looks like. If there’s a 20-30 cycle in the data, and if the last cycle started in 1990, then odds are the cycle has already peaked and has started back down again. Also, picking the last break point in a dataset with known multidecadal variability, adding before and after trends to the data, and then stating the data during the last period has accelerated has a familiar ring to it. It sounds very similar to the nonsensical tactic the IPCC tried in FAQ 3.1 Figure 1 from their 4thAssessment Report to show that global surface temperatures were warming at an accelerated rate.
Curiously, there were no time-series graphs of the “hotspot” sea level anomalies in Sallenger et al (2012). That raises the question: what didn’t they want to show?
To determine the multidecadal signal in the East Coast “Hotspot” sea level data, I downloaded Permanent Service for Mean Sea Level (PSMSL) tidal gage-based sea level data, same dataset used in Sallenger et al (2012), from the KNMI Climate Explorer. It can be found there on their Monthly station datawebpage. As you many of you are aware, the tidal gage-based sea level data is sparse, rarely continuous, and few have more than a few decades of data. In other words, it’s troublesome to work with, which is one of the reasons I rarely look at sea level data. So I selected the stations with the longest data, 70 years or more, along the Atlantic coast that were included in the “hot-spot” group of Sallenger et al (2012), from North Carolina to Massachusetts, that were also relatively continuous; that is, they did not have a lot of missing data. They included New York (Battery), Philadelphia (Pier9N), Atlantic City, Boston, Newport, Hampton Roads, and Sandy Hook. Since John Droz was interested in North Carolina, I added Wilmington, NC to the mix. We’ll call the combination of those datasets “Hotspot+1”. The Wilmington data had the shortest term, so it set the start year of 1935. Not too surprisingly, it had the lowest trend from 1935 to 2008. Refer to Figure 1. However, since we’re going to detrend the data, its lower trend would have little influence on this preliminary look at the East Coast sea level data. Note: The PSMSL data at the KNMI climate Explorer ends in 2008. Sallenger et al (2012) ended their analyses in 2009, so we’ve only lost a year.
Figure 2 presents the average sea level anomalies of those “Hotspot+1” stations along with the linear trend. Assuming the average sea level anomalies of the 8 stations used in this presentation provide a curve that’s similar to the stations Sallenger et al (2012) used for their “Hotspot”, this may illustrate the reason why Sallenger et al (2012) excluded a time series graph for their average “Hotspot” sea level data. It appears sea level anomalies for the “Hotspot+1” stations have been relatively flat since about 1996.
Since one of the points being discussed is decadal to multidecadal variability, we’ll detrend the “Hotspot+1” average, to help show if it exists. Figure 3 presents the detrended average of the “Hotspot+1” sea level anomalies.
First things first: Is there a break point around 1990 when the trends shift? There is definitely a shift in trends before and after 1989/90 in the detrended “Hotspot+1” sea level data, Figure 4. There are also comparable shifts in trend around 1976/77 and about 1981/82, Figure 5 and 6. But Sallenger et al (2012) went through a detailed analysis and elected to discuss the shift around 1990.
I’ve filtered the detrended “Hotspot+1” average sea level anomalies with a 13-month running-average filter (centered on the 7thmonth) in Figure 7 to reduce the month-to-month variability. The spikes in 1972/73 and 1997/98 may lead one to believe that El Niño-Southern Oscillation (ENSO) events have a major influence on it.
Whatever is causing the sea level anomalies to vary, there is a multidecadal signal visible in the detrended “Hotspot+1” sea level anomalies. See Figure 8. In it, I’ve smoothed the data with a 121-month running-average filter, which is commonly used when illustrating decadal and multidecadal variability like the Atlantic Multidecadal Oscillation. Now we can understand why Sallenger et al (2012) were discussing a multidecadal signal. One very plainly exists.
Looking at the trailing decadal (120-month) trends, Figure 9, we can see very clearly that the trend of the most recent decade (through December 2008) is not as high as it has been in the past. Also note that the peaks in the decadal trends occurred about the time of the three largest El Niño events that weren’t counteracted by volcanic aerosols. They were the El Niño events of 1972/73, 1986/87/88 and 1997/98. The eruption of El Chichon occurred at the same time as the 1982/83 El Niño and likely offset the effects of that El Niño-Southern Oscillation (ENSO) event on the sea level anomaly trends. Oddly, ENSO isn’t mentioned in Sallenger et al (2012). Volcanic aerosols are mentioned in the closing of Sallenger et al (2012), however. Are the volcanic eruptions of El Chichon in 1982 and of Mount Pinatubo in 1991 the reasons for the low, actually negative, decadal sea level trends around the same time? If so, why did sea level trends drop in the early 2000s?
What about the Wilmington, NC sea level anomalies?
The sea level anomalies of Wilmington have a relatively low linear trend compared to the long-term stations used by Sallenger et al (2012). Refer back to Figure 1. Looking at the detrended Wilmington sea level data when smoothed with a 121-month filter, they have a different multidecadal signal than the “Hotspot” average. See Figure 10. The latest upswing started about 1976, then peaked near 1996, and the downswing was still in progress though the end of 2008.
The running decadal trends for Wilmington, Figure 11, reveal they have been negative since 2002. That is, on a decadal basis, sea level anomalies in Wilmington have been falling, not rising, since 2002. Have they risen since 2008? It makes one wonder why Sallenger et al (2012) chose to exclude the last few years of data from their analyses!
The source of the data presented in this post is the KNMI Climate Explorer, specifically their Monthly station data.
North Carolina is considering new legislation about the rise is sea level. It would be a shame if the incomplete analysis and data presentation by Sallenger et al (2012) and their use of the word “accelerated” influenced it.
SHAMELESS BOOK PLUG
My book If the IPCC was Selling Manmade Global Warming as a Product, Would the FTC Stop their deceptive Ads? is available in pdf and Kindle editions. An overview of my book is provided in the above-linked post. Amazon also provides a Kindle preview that runs from the introduction through a good portion of Section 2. That’s about the first 15% of it. Refer also to the introduction, table of contents, and closing in pdf form here.
I continue to make progress on my upcoming book about El Niño-Southern Oscillation and hope to publish pdf and Kindle editions by late July, early August 2012.