Sea level rise acceleration (or not). Part V: detection & attribution

by Judith Curry

In looking for causes, I have applied the ‘Sherlock Holmes procedure’ of eliminating one suspect after another. The procedure has left us without any good suspect. Thermal expansion was the candidate of choice at the time of the first IPCC review. The computed steric rise is too little, too late, and too linear. – Walter Munk


To make credible predictions of future sea
 level rise, we must first understand the mechanisms of past and current sea level change. The issues of detecting recent acceleration in sea level rise and arguments attributing sea level rise to human caused global warming deserve critical evaluation, in context of identifying any possible indicators of a high sea level projection pathway in the 21st century.

The outstanding issue in the scientific debate surrounding sea level rise is whether the elevated rates during recent decades represent acceleration in the long-term rate of change that can be attributed to human-caused warming, or a temporary increase due to natural climate variability.

Detection and attribution of anthropogenic signals in sea level change is a new and rapidly developing field. In the terminology of the IPCC, detection seeks to demonstrate a statistical change in the climate system, while attribution evaluates the contribution of multiple causes to such a change.

There are three main challenges to identifying a sea level rise signal from human-caused global warming:

  • The timescales in the ocean are very long, and there is substantial lag time between external forcing and the realization in sea level change
  • High amplitude natural internal variability in the ocean basins on time scales from the interannual to the millennial
  • Strong regional variations in the signal.

Statements from the IPCC AR5

In the IPCC AR5 Chapter 10 on Detection and Attribution, there was no formal attribution analysis of total sea level change. However, the AR5 made the following statements regarding sea level rise, glaciers and ice sheets.  Here are the key summary statements (from the SPM):.

It is very likely that there is a substantial anthropogenic contribution to the global mean sea level rise since the 1970s. This is based on the high confidence in an anthropogenic influence on the two largest contributions to sea level rise, that is thermal expansion and glacier mass loss.

Anthropogenic influences likely contributed to the retreat of glaciers since the 1960s and to the increased surface mass loss of the Greenland ice sheet since 1993. Due to a low level of scientific understanding there is low confidence in attributing the causes of the observed loss of mass from the Antarctic ice sheet over the past two decades.

Lets look at the details on glaciers and ice sheets from Chapters 4 and 10, to see what was going on before 1993 and 1960.:

Glaciers and Greenland ice sheet

Over Greenland, temperature has risen significantly since the early 1990s, reaching values similar to those in the 1930s.” [Chapter 4, p 353]

“Along Greenland’s west coast temperatures in 2010 and 2011 were the warmest since record keeping began in 1873 resulting in the highest observed melt rates in this region since 1958.” [Section 10.5.2]

The arithmetic-mean estimate of Leclercq et al. (2011) indicates continuous mass loss from glaciers after about 1850 (Figure 4.12a, top). Most notable is the rapid loss from Greenland glaciers in the Marzeion et al. simulations during the 1930s. Other studies support rapid Greenland mass loss around this time.” [Chapter 4, p. 343-344]

Figure  4.12  Global cumulative (top) and annual (lower) glacier mass change.

Overall, there is very high confidence that globally, the mass loss from glaciers has increased since the 1960s, and this is evident in regional-scale estimates” [Chapter 4, p343-344]

There is very high confidence that the Greenland ice sheet has lost ice and contributed to sea level rise over the last two decades.” [Chapter 4, p 349]

JC COMMENT: Figure 4.12 indicates the recent onset of glacier melting occurred around 1850, with maximum rates of melting in 1900-1960.1960 was a low point, with much higher mass loss from glaciers prior to 1960.  The IPCC’s main conclusions refer to ‘since the 1960s’ and ‘since the 1970s’, conveniently neglecting what was going on in the early half of the 20th century.


“West Antarctic ice sheet and the Antarctic Peninsula are losing mass at an increasing rate, but that East Antarctica gained an average of 21 ± 43 Gt yr–1 between 1992 and 2011.” [Chapter 4, p 352]

In West Antarctica, the warming since the 1950s, the magnitude and seasonality of which are still debated, has not manifested itself in enhanced surface melting nor in increased snowfall.” [Chapter 4, p 354]

“Due to a low level of scientific understanding there is low confidence in attributing the causes of the observed loss of mass from the Antarctic ice sheet since 1993.” [Section 10.5.2]

“Anthropogenic forcing have likely made a substantial contribution to surface temperature increases since the mid 20th century over every continental region except Antartica.”  [SYR SPM 1.2]

JC comment:  with all the alarms about possible collapse of the West Antarctic Ice Sheet and the attendant sea level rise, there is little to no evidence in the recent observations to support this.

Causes of 20th century sea level rise 

The overall increase of sea level since the mid 19th century and the rising concentrations of atmospheric CO2 naively suggests a causal relation.

The two figures below illustrate the challenge of attributing long-term sea level rise to human caused global warming associated with CO2 emissions. These figures show that global sea levels were rising steadily long before fossil fuels emissions became substantial. As Munk stated in his ‘enigma’ paper: “the historic rise started too early, has too linear a trend’


Since CO2 emissions from fossil fuels didn’t become substantially large until after 1950, understanding the causes of recent sea level rise is clearly more complex than a naïve attribution to increasing levels of atmospheric CO2.

Analyzing the 20th century sea level rise budget, Gregory et al. (2013) found that the largest contribution to global mean sea level rise was from glaciers, and that its rate was the same in both the first and second half of the century, despite the secular warming trend over the course of the century. In the latter half of the nineteenth century, sea level was recovering from large volcanic eruptions and reduced solar irradiance that occurred earlier in the 19th century, which can explain the nineteenth-century onset of glacier mass loss. In the early twentieth century, warming in northern high latitudes probably increased the rate of global mean sea level rise because of mass loss by glaciers and/or the Greenland ice sheet. These natural upward fluctuations of sea level happened to lead into the start of pronounced anthropogenic forcing, and the relative constancy of the rate for most of the century was partly due to greater negative volcanic forcing since the 1960s.

Let’s take a look at the volcanic forcing since the 18th century, from Chapter 8 of the AR4. Large volcanic eruptions are seen in the first half of the 18th century, a lull in volcanic activity during 1920-1960, with an increase in volcanic activity in the last 4 decades of the 20th century.

I continue to be intrigued by the ‘bump’ in the rate of sea level rise mid-20th century. The IPCC AR5 acknowledged this ‘bump’:

“It is very likely that the mean rate of global averaged sea level rise was 1.7 [1.5 to 1.9] mm yr–1 between 1901 and 2010, 2.0 [1.7 to 2.3] mm yr–1 between 1971 and 2010 and 3.2 [2.8 to 3.6] mm yr–1 between 1993 and 2010. It is likely that similarly high rates occurred between 1920 and 1950.”

The rate of global mean sea level change as portrayed in AR5 is shown belowFigure caption: 18-year trends of global mean sea level rise estimated at 1-year intervals. The time is the start date of the 18-year period, and the shading represents the 90% confidence. The estimate from satellite altimetry is also given, with the 90% confidence given as an error bar. [AR5 WGI Figure 3.14]

I just spotted a more recent version of this diagram in the CLIVAR 2018 Winter Edition Special issue on Sea Level Rise, which gives a 20th century trend of 1.3 ± 0.2 mm/yr.

Global average sea surface temperatures from CRU since 1900 are shown below:

Figure caption: Global average sea surface temperature anomalies from the HadSST (to 2012)

 We see that global mean sea surface temperatures in the 1940’s were about 0.4oC lower than 21st century temperatures. It doesn’t seem that thermal expansion can explain the mid-century high rates of sea level rise.

So what explains the ‘bump’ in the rate sea level rise mid 20th century? I didn’t find this in any of the IPCC assessment reports.


Insights into the causes of sea level rise are provided by efforts to understand the budget of sea level rise.

This Table (from my 2014 congressional testimony) summarizes contributions to sea level rise from different sources (mm per year), from both the IPCC AR4 and AR5 Reports:Thermal expansion is one third smaller in AR5 and land water storage with a substantial amount is completely new in AR5, while the sum of these sources remained constant. The growing realization of the importance of land water storage to sea level rise is diminishing the percentage of sea level rise that is associated with warming.

Jevrejeva et al. (2016) provide a review of recent progress and challenges in balancing the 20th century sea level budget. Much recent progress has been made in clarifying the budget for the satellite era (since 1993).

Chen et al. (2017) provide an updated analysis of the budget of sea level rise during the satellite era. The key findings:

Here we show that the rise, from the sum of all observed contributions to GMSL, increases from 2.2 ± 0.3mmyr−1 in 1993 to 3.3 ± 0.3mmyr−1 in 2014. The mass contributions to GMSL increase from about 50% in 1993 to 70% in 2014 with the largest, and statistically significant, increase coming from the contribution from the Greenland ice sheet, which is less than 5% of the GMSL rate during 1993 but more than 25% during 2014.”

Figure 4 from Chen et al. shows the time series of the budget components. The steric component of the rate of sea level rise (related to ocean temperature increase) shows a lower value in 2014 than in 1993 (apparently associated with the lingering effects from the Pinatubo eruption). The increase in the rate of sea level rise is largely attributed to the Greenland ice sheet.


The recent acceleration in the mass loss from Greenland motivates a focused look at the mass balance of the Greenland ice sheet.

An article by Fettweis et al. (2008) estimates the Greenland mass balance for the 20th century.   A key conclusion:

These estimates show that the high surface mass loss rates of recent years are not unprecedented in the [Greenland ice sheet] history of the last hundred years. The minimum [surface mass balance] rate seems to have occurred earlier in the 1930s

It is seen that the largest negative values of the surface mass balance occur during the period ~1925 – 1940, and then since 1995.

When I first saw the above figure, it immediately clicked in my mind that we were seeing the affects of the Atlantic Multidecadal Oscillation (AMO). The figure below lines up the AMO index with the Fettweis et al. diagram of Greenland ice sheet mass balance:

The implications of this relationship for sea level rise attribution is rather profound – the negative early 20th century mass balance for Greenland is coincident with the ‘bump’ in rate of sea level rise noted in the section on Causes of the 19th and 20th century sea level rise, including Figure 3.14 from the IPCC AR5 WG1 Report.

Reeves and Zeng (2017) have reviewed the various surface air temperature datasets for Greenland.

There are substantial discrepancies in these datasets, although all of them show some sort of maxima in the early 20th century. I am inclined to go with Box, which is compiled by a scientist that has devoted his career to studying this. In terms of anomalies (not absolute values), Berkeley Earth agrees reasonably well with Box.

Greenland was anomalously warm in the 1930’s and 1940’s, with temperatures comparable to 21st century temperatures.

The acceleration of ice loss during the early part of the 20th century was not limited to Greenland. Marzeion et al. (2012) describe the past mass balance of glaciers (excluding Antarctica), and includes an excellent analysis of uncertainties. Excerpt:

Rates of mass loss during the 20th century were characterized by generally faster mass loss of approximately 1.5 mm SLE yr−1 during the first half of the century, caused by Greenland in the 1930s, Arctic Canada in the 1950s to early 1960s, and the Russian Arctic in the late 1950s and 1960s. Rates then dropped to a low of around 0.5 mm SLE yr−1 during the 1970s, and since then have been gaining speed again to currently approximately 1.0 mm SLE yr−1.

Marzeion et al.’s analysis includes glaciers from low latitudes and the southern hemisphere, but in terms of volume and surface area they are dominated by those from the high latitudes of the Northern Hemisphere.

Marzeion et al. published an update in 2015, reconciling some of the differences in different estimates of glacier mass change:Updates to the Leclerq and Marzeion studies show significant changes; increasing the values of the Leclerq analysis and decreasing the values of the Marzeion analysis, bringing them into closer agreement. It is not clear from the Marzeion et al. (2015) paper whether this changes the relative amplitude during the 1930’s.

LeClerq et al. have developed a new data set of worldwide glacier length fluctuations.  The paper’s key conclusion:

“The available glacier length series show relatively small fluctuations until the mid-19th century, followed by a global retreat. The retreat was strongest in the first half of the 20th century, although large variability in the length change of the different glaciers is observed.”

Well, I am not the first person to make the connection between Greenland ice sheet melt and the AMO:

A recent paper by Hover et al. provides insights into how the AMO and NAO influence the Greenland mass balance:

The Greenland Ice Sheet (GrIS) has been losing mass at an accelerating rate since the mid-1990s. We show, using satellite data and climate model output, that the abrupt reduction in surface mass balance since about 1995 can be attributed largely to a coincident trend of decreasing summer cloud cover enhancing the melt-albedo feedback. The observed reduction in cloud cover is strongly correlated with a state shift in the North Atlantic Oscillation promoting anticyclonic conditions in summer and suggests that the enhanced surface mass loss from the GrIS is driven by synoptic-scale changes in Arctic-wide atmospheric circulation.

JC note: the anticyclonic conditions would also reduce precipitation.

Detection of recent sea level rise acceleration

Identification of a statistical change in the sea level record is confounded by the existence of long-term memory and the large magnitude of internal variations in the ocean. When the magnitude of a trend is greater than some threshold that scales the influence from long-term memory and internal variations, the sign of the estimated trend can be interpreted as the underlying long-term change.

Recent research in sea level rise detection is increasingly based on more sophisticated analyses of long-term memory and persistence. Sea level variability has a strong long-term memory that implies an enhancement of the uncertainty in the observed trends. Knowing the persistence of natural sea level rise enables a statistical assessment of the probability that the recent rise is outside the range of natural variability. This persistence, which is described by the Hurst exponent, leads naturally to multi-decadal excursions from the mean state and can therefore produce relatively strong multi- decadal to centennial trends unrelated to any external forcing.

At the regional scale, detection studies on sea level are highly challenging. The internal climate variability introduces strong changes in regional sea level on timescales from years to decades (and even longer), making it very difficult to detect an external signal above the unforced internal variability. For example, the internal sea level variability related to ENSO and to the Pacific Decadal Oscillation (PDO) is of the order of ±10–20 cm, which masks the sea level changes due to any externally forced signal.

Marcos et al. (2016) summarizes recent research on sea level rise detection. Several ‘Time of Emergence’ (ToE) studies have been performed on regional sea level variability. ToE is defined as the time when the ratio of the climate change signal to the noise of natural variability exceeds a particular threshold and emerges from the natural climate variability at regional scale. Time of emergence seems to be essentially synonymous with ‘detection.’

JC query:  I’m not sure how to interpret ToE in terms of attribution to AGW, seems at least 50%, but possibly more stringent.

Lyu et al. (2014) found that the externally forced trend would be detectable in both steric and dynamic sea level by early to mid-2040s in 50% of all the oceans, under representative concentration pathways (RCP) 4.5 and RCP8.5 scenarios. Richter and Marzeion (2014) argue that the externally forced signal should be detectable in the early 2030s relative to 1990 in the South Atlantic Ocean, Arctic Ocean, eastern Pacific Ocean, and most parts of the Indian Ocean. Jorda` (2014) has shown that, on average, it would require a minimum period of 40 years to identify the externally forced signal at the regional scale. However, in regions with strong decadal and inter-annual sea level variability, the emergence time increases up to 60–80 years. Bilbao et al. (2015) argues that the local sea level change will emerge first in the northern latitudes of the Southern Ocean and in the Tropical Atlantic, where the unforced internal variability is smaller. In contrast, at southern latitudes of the Southern Ocean, it may not emerge until after 2100. Overall, these papers find that in regions of high internal variability, the trend due to externally forced signal is masked during longer time spans than in regions of low internal variability.

The Time of Emergence analyses argue that we have not yet detected unusual sea level rise in any of the individual ocean basins. But as we will see in the following section, this hasn’t impeded attribution analyses that assign cause to AGW.

Attribution of recent sea level rise

Two early papers (circa AR5) came to opposite conclusions regarding the attribution of recent sea level rise. Jevrejeva et al. (2009) used a statistical model that showed that anthropogenic forcing was responsible for up to 70% of sea level rise since 1900. Gregory et al. (2013) stated:

“The largest contribution to GMSLR during the twentieth century was from glaciers, and its rate was no greater in the second half than in the first half of the century, despite the climatic warming during the century. Of the contributions to our budget of GMSLR, only thermal expansion shows a tendency for increasing rate as the magnitude of anthropogenic global climate change increases, and this tendency has been weakened by natural volcanic forcing. Greenland ice sheet contribution relates more to regional climate variability than to global climate change; and the residual, attributed to the Antarctic ice sheet, has no significant time dependence. The implication of our closure of the budget is that a relationship between global climate change and the rate of GMSLR is weak or absent in the twentieth century. The lack of a strong relationship is consistent with the evidence from the tide gauge datasets, whose authors find acceleration of GMSLR during the twentieth century to be either insignificant or small.”

Recent attribution analyses are summarized by a review paper authored by Marcos et al. (2016):

  • Becker et al. (2014) estimated the fraction of the observed linear trends at tide gauges and for global sea level that cannot be explained by the long-term memory of sea level variations and that is thus linked to an external forcing. They concluded that the anthropogenic contribution to sea level rise during the twentieth century is more than 50% of the observed global sea level rise rate, with a 99% confidence level.
  • Dangendorf et al. (2015) concluded that it is virtually certain that at least 45% of the observed global sea level rise during the twentieth century is of anthropogenic origin, and extremely likely that it is at least 61% and very likely at least 68%.
  • Kopp et al. (2016) estimated the anthropogenic fraction as about 50%.
  • Slangen et al. (2016) found the anthropogenic forcing explains only 15 ± 55% of the observations before 1950, but increases to become the dominant contribution after 1970 (69 ± 31%), reaching 72 ± 39% in 2000 (37 ± 38% over the period 1900-2005).

With regards to the thermosteric (temperature) component of the sea level rise, Cheng et al. (2017) found that changes in OHC are relatively small before about 1980; since then, OHC has increased fairly steadily and, since 1990, has increasingly involved deeper layers of the ocean.

Marcos et al. (2016) provides the following summary of attribution of the thermosteric component of the sea level rise:

Based on the comparison between the forced signals in historical simulations and in natural forcing-only simulations using climate models, Marcos and Amores (2014) concluded that, since 1970, 87% (72–100%) of the observed warming-related sea level rise in the 0–700 m of the global ocean is of anthropogenic origin. Slangen et al. (2014), for the period 1957–2005, found that the sea level response to anthropogenic GHG forcing explains most of the magnitude of the observed differences, while the natural forcing is required to explain temporal variability.

Geothermal wild card

In the past few years, a number of papers have been published, suggesting significant geothermal heat sources for the Greenland and Antarctic ice sheets, with implications for their mass balance and contributions to sea level rise.

We identified 138 volcanoes, 91 of which have not previously been identified, and which are widely distributed throughout the deep basins of West Antarctica, but are especially concentrated and orientated along the >3000 km central axis of the West Antarctic Rift System.

“The experiments show that mantle plumes have an important local impact on the ice sheet, with basal melting rates reaching several centimeters per year directly above the hotspot.” 

Of the basic parameters that shape and control ice flow, the most poorly known is geothermal heat flux. We present a high-resolution heat flux map and associated uncertainty derived from spectral analysis of the most advanced continental compilation of airborne magnetic data.”

We show that the Thwaites Glacier catchment has a minimum average geothermal flux of ∼114 ± 10 mW/m2 with areas of high flux exceeding 200 mW/m2 consistent with hypothesized rift-associated magmatic migration and volcanism.

Ice-penetrating radar and ice core drilling have shown that large parts of the north-central Greenland ice sheet are melting from below. It has been argued that basal ice melt is due to the anomalously high geothermal flux that has also influenced the development of the longest ice stream in Greenland. Here we estimate the geothermal flux beneath the Greenland ice sheet and identify a 1,200-km-long and 400-km-wide geothermal anomaly beneath the thick ice cover.

The implications for attributing any historical or future sea level rise to ice sheet geothermal activity is unknown at this point, but this seems to be an important knowledge frontier.


Marcos et al. (2016) concluded that “In the last few years, different works have used statistical methodologies to conclude that the natural variability in sea level, i.e. decadal to multi-decadal changes associated with climate variations, cannot explain the observed linear global sea level trend during the past century. There exists therefore an irrefutable fingerprint of an anthropogenic forcing in this key climate change indicator.”

There has been substantial progress since the AR5 on detection and attribution of sea level rise, including some very nice research.  However, I am not convinced that all of this evidence is being integrated in the appropriate way in context of seeking to identify the fingerprint of CO2 emissions on sea level rise.

There has been a secular trend of increasing sea levels since the mid 19th century. This is loosely associated with ‘coming out of the Little Ice Age’, presumably caused by a combination of changes in external forcing from volcanoes and the sun and long term ocean circulations. The slow emergence of fossil fuel emissions prior to 1950 is unlikely to have contributed significantly to 19th and early 20th century sea level rise.

Before attempting to pin the recent acceleration in global mean sea level rise on fossil fuel emissions, I would argue that the sea level rise data set does not pass the ‘detection’ test. From the IPCC AR4 :

‘Detection’ is the process of demonstrating that climate has changed in some defined statistical sense, without providing a reason for that change. An identified change is ‘detected’ in observations if its likelihood of occurrence by chance due to internal variability alone is determined to be small.

So what exactly is the ‘detection’ metric we should be looking at? Two possibilities:

  • Sea level since 1950
  • Rate of sea level change since 1950

Sea level is now higher than it was in 1950. This does not seem to be a very useful metric, since here has been nearly linear sea level rise since ~1860. The rates of sea level rise since 1993 are comparable to those observed in the first half of the 20th century (given the uncertainties), i.e., no detection of change since 1950. Separating the rates of sea level rise into steric and mass components doesn’t help much either – the main acceleration signal is seen in the mass component from Greenland and glaciers, but the recent signal is smaller than was seen circa the 1930’s.

Failure to pass a ‘detection’ test is borne out by the numerous ‘time of emergence’ analyses that don’t expect the anthropogenic sea level rise signal to rise above natural variability until the mid 21st century (or later)..

Failure to pass a ‘detection’ test should not preclude attribution analyses (and it hasn’t), but any attribution analysis should recognize the reasons for the lack of detection – multi-decadal internal variability.

With regard to the detection of the impact on seal level rise of fossil fuel emissions from 1900-2005, the conclusion from Slangen et al. (2016) seems reasonable: 37 ± 38%. With regards to more recent sea level rise, estimates range from 45 to 69%.   While I have low confidence in these numbers, I find the sea level rise attribution analyses conducted by oceanographers to be eminently more sound than those conducted by atmospheric scientists regarding attribution of warming since 1950 (where their best estimate is ~100%). This is because the oceanographers consider the following:

  • Natural internal variability on multidecadal and longer timescales
  • Analyses of long-term memory and persistence

These two central factors are completely ignored by atmospheric scientists in their attribution analyses of the causes of temperature (and extreme event) increases since 1950.

Why do I have low confidence in the analyses attributing of more than half of recent sea level rise to fossil fuel emissions? The attribution of the thermosteric component relies on climate model simulations, comparing simulations with only natural forcing to those with both natural and anthropogenic forcing. Any contribution from multi-decadal internal variability is essentially zero, since the climate models do not simulate the correct phasing of this variability and averaging of multiple models leaves only the forced changes.

The bottom line for attribution is that if a metric fails the detection text (relative to internal variability), then you can’t then proceed with an attribution analysis that neglects multi-decadal internal variability — this is circular reasoning.  Which is what the climate model based attribution analyses do.

The recent acceleration in sea level rise comes from the mass balance component from Greenland, which appears to have been larger during the 1930’s. While some of this reduction in the mass balance could be caused by reduced volcanic activity in the early 20th century, a large component of this appears to be associated with  the AMO. The onset of the recent increase from Greenland (circa 1995) coincides with the 1995 shift of the AMO to its warm phase, with changing atmospheric circulation and cloudiness patterns that resulted in more solar warming of Greenland (Hover et al. 2017).

So in spite of all the new high tech data sets and complex climate models and fancy statistical methods, we still don’t know how to separate human caused variations from natural internal variability. Which to me means that we need to rely more heavily on detection and time of emergence arguments than on climate-model based attribution arguments.

I have not seen the early 20th century mass balance of Greenland and the AMO explicitly factored into the recent attribution analyses of sea level rise, I would appreciate any references that you can find on this.

And finally, in context of the ‘blame game’, Slanged et al’s. ~37% of 8 inches  since 1900 (Slangen et al.)  is enough to cover your toes but not much else if you are standing in a puddle that deep.


134 responses to “Sea level rise acceleration (or not). Part V: detection & attribution

  1. The bottom line of climate science is that we should cut emissions and the so called impacts exist to enforce compliance. Thus the idea that reducing emissions will moderate sea level rise. For that to be true there has to be a correlation between these two time series at some finite time scale that makes sense. But no such correlation exists. Please see

  2. Very interesting, thanks for sharing. I had not been aware of the mid-20th bump. AMO so often seems to be unfairly neglected.

    “The bottom line for attribution is that if a metric fails the detection text (relative to internal variability), then you can’t then proceed with an attribution analysis that neglects multi-decadal internal variability — this is circular reasoning. Which is what the climate model based attribution analyses do.”

    It’s unfortunate this is not more widely recognized as obvious.

  3. In looking for causes, I have applied the ‘Sherlock Holmes procedure’ of eliminating one suspect after another. The procedure has left us without any good suspect. Thermal expansion was the candidate of choice at the time of the first IPCC review. The computed steric rise is too little, too late, and too linear. – Walter Munk

    The above no longer entirely the case. Unlike Feynman, Munk still speaks:

    “To resolve this enigma required a decade of work and progress in three very different parts of Earth science,” Mitrovica said.

    Mitrovica said Munk was right to point out these problems when he did.

    “Munk is no climate change denier,” he said. Instead, Munk was raising issues that forced scientists to refine their models, ultimately giving them a more precise understanding of the effects of climate change.

    “If there’s a problem in Earth rotation you better solve it,” Mitrovica said. “Because that produces a really fundamental inconsistency in your argument that you’re melting glaciers.”

    Munk said he’s glad to see scientists build on his work and prove that days are in fact getting measurably longer due to climate change.

    “It’s very satisfying that work like this, when you find things that did not add up, inspired people in subsequent years to do some additional work, to try to understand it,” Munk said. “And in some instances, to succeed. I mean, that’s the way science should work.”

  4. Terrific piece.

    It’s unfortunate there are not about 50 Judiths who would look at this issue with an open, scientific and inquisitive mind. The attribution question is certainly complex. But so are the dynamics involved in contributions to SLR from Greenland and Antarctica glaciers. The current thinking doesn’t do the complexity justice. I’m not talking about the measurement challenges but just understanding the natural forces at play, and the questionable chain of causality between AGW and the meltwater from those glaciers.

    If the common orientation is toward AGW, then all the other possibilities don’t receive the appropriate attention. I’ve just read a number of interesting papers on glaciers in Greenland and Antarctica. Those papers opened my eyes on variables and uncertainties and mechanisms that I haven’t seen mentioned previously.

    I’ll go back and do some organizing to share them here.

  5. Attribution needs a baseline of comparison, and none exists. It’s not like when I ding your car and you demand that I restore the status quo ante. There are three possibilities: that without humans the world would have been cooler, or it would have been hotter, or it would have been much the same as it is now. But we just don’t know.

    On a comical note, choosing any one of the three options has odds of one in three. Like the Monty Hall problem perhaps we should all, warmists and skeptics alike, change our bets.

  6. Some interesting analyses of a complex issue by Judith but I do have to take issue with one of her central guiding assumptions – throughout this article she makes, repeatedly, the assumption that the lion’s share of the increase in CO2 emissions occurred circa 1950 and beyond. This is historically inaccurate and as such a potentially serious error in all of her subsequent calculations and conclusions – the lion’s share of the increase in CO2 emissions coincided with the substantial increase in industrialisation which was a key feature of the Industrial Revolution – from certainly 1900 onwards and possibly stretching back to the 1880’s. Hence some of the early graphs and studies highlighted at the beginning of the article are quite correct in my opinion and conflct with her negative evaluation – it is simply inaccurate to suggest that the bulk of rises to CO2 emissions occurred in 1950 and beyond since it has been historically well documented that this rise occurred some considerable time before this date. This has the effect of casting some doubt on some of Judith’s subsequent conclusions.

    • In terms of cumulative emissions, a small amount was emitted from fossil fuels prior to 1950. Is the EPA figure incorrect?

      • Of course, uptake continues to increase along with increased concentration.

        Still, the majority of accumulation, forcing, and temperature increase have all occurred after WWII, at odds with the earlier twentieth century SLR peak.

    • Maybe a better metric to use would be the rate of accumulation of CO2 in the atmosphere. If law dome ice core is to be believed, it was only about .3ppm/year from 1900 to 1950. In the MLO era, it has ranged from the initial 1 ppm/year on average circa 1960 to 2 ppm/year on average since the turn of the millennium. That would still be 3-7 times the rate of warming in the later half of the twentieth century than it was in the earlier half. (somewhat less than that when you consider the logarithmic warming of CO2) Still hard to make a case for anthropogenic warming and subsequent sea level rise before 1950…

  7. Very nice review. Many thanks. Permabookmarked.
    In a nutshell, the warmunists have attributed to AGW that which has not yet been detected. More ‘settled science’.

    OT but important. Denizens should look up and save the excellent amicus brief by Happer, Koonin, and Lindzen to Judge Alsop in the People of California v. BP lawsuit seeking climate damages from oil companies. Part of the ‘tutorial’ hearing taking place today.

  8. Global sea level rise :
    Sources: Jevrejeva, Church, and recent satellite data.
    How to explain the strong rise around 1800 that lasted 30-35 years (AMO or PDO are not known at that time)?
    Better global satellite measurements don’t imply that it is wise to overinterpret such data over a short (40 years) period of time.
    As other human beings, climatologists are too impatient.

  9. Thanks very much. A useful addition to observation based discussion of the most important potential impact of CAGW. Can’t wait to see Tonyb comments but your contributions to sea level rise discussion is magnificent.
    Thanks so much.

  10. An excellent summary. I might note that one other possible significant contribution to sea level rise may have been from ground water extraction: Wada et. al. (GEOPHYSICAL RESEARCH LETTERS, VOL. 37, L20402, doi:10.1029/2010GL044571, 2010) estimate that global ground water extraction in 2000 raised sea level by about 0.8 mm/year. That’s roughly 25% of the current total rise rate. Reservoir construction and storage offset that some earlier in the 20th century, but that seems to have approximately stabilized at present (Zhou et. al., The Contribution of Reservoirs to Global Land Surface Water Storage Variations; DOI: 10.1175/JHM-D-15-0002.1. American Meteorological Society 2016). If true, extrapolating to 2018 implies that recent contributions from ice melt and ocean thermosteric effects has been relatively small compared to 20th century averages, significantly limiting increases directly attributable to climate change.
    Keith Jackson

  11. Dr Curry, thank you for the essay.

    Identification of a statistical change in the sea level record is confounded by the existence of long-term memory and the large magnitude of internal variations in the ocean. When the magnitude of a trend is greater than some threshold that scales the influence from long-term memory and internal variations, the sign of the estimated trend can be interpreted as the underlying long-term change.

    That and the rest of the section are well written. Going from the general principles to specific applications always entails assumptions or estimates of the long-term and short-term quasiperiodicities, which at present are not known with great accuracy or reliability; and the (shorter) time spans chosen to estimate local variances. In the context of fitting many models and then reporting the results of only one or a few, the choice of significance level (e.g. p=0.1 vs p=0.01) for calling a “detection” takes on added, ahem, significance (that’s the “some threshold”).

  12. As the earth warms…
    Kip Hansen’s post over at WUWT contains a real shocker – and one with an unmistakable smoking gun of anthropogenic causation – with its figure with SLR rates at different US coastal locations. This graphic shows a correlation between rate of sea level rise and the alphabetical position of a location. Those living at locations starting with letters further up the alphabet – at no fault of their own – are imperilled by faster sea level rise and higher risk of flooding, drowning and accidental fish-swallowing.

    The (we must act now) list of human activities that we now know to be linked to grievous environmental dangers and impacts just keeps on growing: now symbolic language and a written alphabet join that list of shame. Donald Trump’s persistence in sending social media tweets irresponsibly encourages people to continue and increase their use of textual communication with the Latin alphabet, which we now know to be a real risk factor for coastal flooding.

    O hang on – I’m writing text in Latin alphabet now – better stop

  13. Why doesn’t the argument end here: “The bottom line for attribution is that if a metric fails the detection text (relative to internal variability), then you can’t then proceed with an attribution analysis…”? Most of your piece is fussing about this or that attribution. I say, if there is no signal, then there is nothing else to talk about until you get a signal. Natural variability or rebound from the LIA or whatever you want to call this unknown collection of “causes” of the very unremarkable increase in sea level, this cumulative data line of level increases represents the true null hypothesis of “climate change”. If there is no statistical deviation from that line, then there is no signal of anything new, whether it be volcano action, sun spots, expansion of the human population, heat islands, or CO2. BTW, the same applies to temperature increase. Or we might say with some degree of assurance, no hockey stick? no evidence of any AGW.
    Ronald Havelock, PhD

  14. “Anthropogenic influences likely contributed to the retreat of glaciers since the 1960s and to the increased surface mass loss of the Greenland ice sheet since 1993.”

    The anthropogenic component is the wrong sign to drive the Greenland warming, rising CO2 forcing doesn’t increase negative NAO/AO. It’s not internal variability either, declining solar plasma strength since the mid 1990’s has increased negative NAO/AO conditions, driving a warm AMO phase. If anything, rising CO2 forcing could help to explain how the AMO warming from 1995 was marginally slower than from 1925.

  15. If you compare the last 30 years to the 20th century average, the rise rates of sea level, CO2 and temperature have all doubled. Coincidence? I think not. Acceleration it is, and it is time to get used to it.

    • Somehow a little too neat. Keep re-analysing till it’s fit for purpose. Climate data a-la-carte.

      • I point things out that you don’t see in skeptic op-eds or on their blogs. Some of them are very obvious facts where they mislead by omission.

      • Who are you accusing of misleading by omission, yimmy? It’s the old serial dis-informer libel, again. I wonder why she puts up with it.

      • Squeezing acceleration out of decelerating sea level rise has required particularly inquisatorial torture of the data.

  16. AGW in the 90’s was at most 0.1K. Cloud feedback is at most 1W.^-2.k^-1 – thus 0.1W/m2 feedback. But the increase in cloud radiative forcing was some 1.5W/m2 – with oceans warming at the same rate.

    Low cloud cover changes with change in ocean and atmosphere circulation. It results in large “large natural variability in the Earth’s radiation budget…”

    Until there is a far more detailed understanding of the Earth’s radiation budget – over a longer period and with better instrumentation – it’s all a bit of a guess. I’d guess – based on the data – that it is mostly not AGW.

    Her’s the shortwave component over the 21st century thus far. It is reflected shortwave mostly from loud. So more cloud early in the century – not much change in the middle of the record – and less cloud in recent years. The power flux changes from cloud change are substantial.

    It is associated with changes in sea surface temperature in the eastern Pacific. The physics include open and closed cell cloud formation in Rayleigh–Bénard convection in a fluid heated from below.

    This is an excellent and exhaustive post from the bottom up showing that there are terms missing in sea level rise. But the resident AGW activists have taken this back over the same ground yet again with the unstated assumption that it is all AGW. From the top down – it is most evidently not.

  17. One has to consider that Greenland climate responds opposite to NH climate to changes in solar forcing as we have been able to see these past years when Greenland winters have been warmer, and with more snow, and summers have been cooler. The main explanation for this phenomenon, that probably contributes to interdecadal variability in sea level rise is provided by Kobashi et al., 2015 that I already referenced a few posts ago:
    Kobashi, T., Box, J. E., Vinther, B. M., Goto‐Azuma, K., Blunier, T., White, J. W. C., … & Andresen, C. S. (2015). Modern solar maximum forced late twentieth century Greenland cooling. Geophysical Research Letters, 42(14), 5992-5999.

    Their figure 5 shows strong Greenland warming in the 1920-1940, probably leading to an increased contribution towards the SLR acceleration that peaked ~ 1960. The cryosphere contribution to SLR appears to act with a delay to the forcing. Greenland cooling in the 1980-2000 period might therefore have contributed to lack of acceleration in the satellite era.

    • Look up “Long‐term Caspian Sea level change” or similar papers. When I worked in Azerbaijan we were quite concerned by large sea level oscillations, which changed the way we designed ports and platforms.

      • But the Caspian is an enclosed sea very sensitive to the amount brought in by rivers. Irrigation has a major effect on Caspian sea levels.

      • wasn’t the Caspian rivers used extensively and inappropriately for large scale cotton growing and this changed the landscape and water levels?


  18. It seems rather odd to use CO2 emissions as the input for this analysis.

    Total forcing would seem more logical.

    The picture from that is rather different.


    • in context of the ‘blame game’, CO2 emissions is the relevant metric.

      • Emissions?


        You would choose not to take into account either
        -the logarithmic relationship of CO2 concentration to forcing?
        – the fact that CO2 is neither the sole GHG nor that GHGs are not the sole contributor to forcing?
        – emissions relationship to concentration

        That seems an unusual approach.

  19. The plot introduced by the phrase “Global average sea surface temperatures from CRU since 1900 are shown below:” does not show sea-surface temperatures. It shows land surface air temperatures.

  20. SLR from 1900 to 1990 averaged 1.2 mm/yr.
    SLR at 1993 was ~1.8 mm/yr.
    SLR over the satellite era is 3.3 mm/yr.
    SLR over the last 20 years is 3.8 mm/yr, which includes the slowdown in SLR during the warming hiatus.
    SLR over the last 15 years is 3.6mm/yr.
    SLR over the last 10 years is 4.29 mm/yr.
    SLR over the last 5 years is 4.72 mm/yr.

  21. Pingback: Sea level rise acceleration (or not)–Part V: Detection & Attribution | Principia Scientific International

  22. If the mass component was important, the mass should be migrating from the arctic to the wider circumference at the equator. Given this, I’d expect to see the Length of Day (LOD) becoming longer (e.g. figure skaters arms being extended outwards). Instead it has been trending toward shorter days the last 50 years or so.

    • People ignore this. Earth spin rate has increased over this time. If Sea Levels did rise, spin rate should have decreased. Warm oceans and thawed polar regions start putting more snowfall on old ice than is being lost at the edges. Oceans most likely have dropped and not risen. The unknowns are larger than the knowns in measuring sea level. The spin rate of the earth is well known to a small fraction of a second.

    • NASA Study Solves Two Mysteries About Wobbling Earth

      The researchers found the answer in Eurasia. “The bulk of the answer is a deficit of water in Eurasia: the Indian subcontinent and the Caspian Sea area,” Adhikari said.

      The finding was a surprise. This region has lost water mass due to depletion of aquifers and drought, but the loss is nowhere near as great as the change in the ice sheets.

      So why did the smaller loss have such a strong effect? The researchers say it’s because the spin axis is very sensitive to changes occurring around 45 degrees latitude, both north and south. “This is well explained in the theory of rotating objects,” Adhikari explained. “That’s why changes in the Indian subcontinent, for example, are so important.”

  23. I have also looked at 21st-century sea level projections. Since anthropogenic effect is important since ~ 1950, but it is not clearly discernible in the SLR data, two possibilities are the most probable:
    1. Its effect on SLR is small and below detection after 70 years.
    2. Its effect is not small and is detectable but the result is what we are seeing, i.e. it is already baked in the cake.
    In both cases, a conservative projection of SLR for the 21st-century should extrapolate for the next 80 years what has been happening in the past 70 years. This is further supported by a reduction in the rate of increase of our emissions for the past 6 years to levels seen 20 years ago, and by the peak in coal production that took place in 2013. What we are contemplating is a future with similar emissions, not significantly higher emissions.

    So back to SLR I projected both the long term-trend with its small acceleration, and the multidecadal oscillation. Then I compared my result with central estimates from scenarios that project similar emissions pathways, like IPCC B1 (AR4), or IPCC RCP4.5 (AR5), and with an equivalent from the Horton et al. survey and with NOAA’s intermediate high projection that considers both steric and melting contributions.

    I was surprised to see that IPCC AR4 B1 produced exactly the same projection, about +0.3 m over 2000 levels. But then it got bumped up to >0.5 m by AR5 because of severe criticism. And the Horton survey is essentially another strong criticism because AR5 SLR projections are considered still too low by experts.

    It has been proposed that experts are not very good at predicting in situations that are very complex and highly uncertain. The golden rule of forecasting is to be conservative and apparently experts tend not to be conservative when uncertainty increases. It is very unlikely that we will be seeing a SLR of more than 0.5 m by 2100. I even think that my conservative scenario is probably an upper limit.

    • Javier: In both cases, a conservative projection of SLR for the 21st-century should extrapolate for the next 80 years what has been happening in the past 70 years

      That’s more like it, but based on JCH’s figure B, I’d recommend we wait 20 more years before trying an 80 year projection. 20 years was the approximate time for the drop from the 1940s peak rates to the 1960s trough in rates, and we need to learn whether the 1940s to 1960s drop in rates (approximately) repeats. That is an obvious known unknown.

      • It’s already repeated: 1998 to 2104. It slowed the rate of rise down and then it died. After WW2 there was a huge amount of impoundment added. It deepened the dip in the rate of rise. That’s not happening again. NV has done it’s deed: pop fizzle over.

      • 2014.

      • JCH: It’s already repeated: 1998 to 2104. It slowed the rate of rise down and then it died.


        Are you abandoning the processes recorded in Figure B before waiting to see how the measurements plot out?

      • I’m not abandoning it. What is being abandoned/ignored is the positive phase of the PDO like the one that started around 1924. The rate of SLR took off. That is what is repeating now. Look at the great depression and WW2:

        The repeat of that process you want to wait for has already happened. It came and went and nobody seems to have noticed because they have this preconceived notion it’s going to look like the prior dips, when ACO2 was a lot lower. There is no physical basis for expecting its repeat to look like a dip; it can merely look like a slowdown. The event called the pause; warming hiatus; no warming; 1998 to 2014. There is no physical basis for expecting the PDO to have a consistent phase length. A phase could last less than 20 years. Perhaps less than a decade.

        We are in a positive phase of the PDO, so up next: an upward climb, just like in the Great depression and WW2.

        Or, you can be like all theater skeptics and ignore natural variability.

      • Part of the post WW2 dip was caused by the global flurry in dam construction. The best sites are taken. Well, there is the Grand Canyon.

      • “There is no physical basis for expecting the PDO to have a consistent phase length. A phase could last less than 20 years. Perhaps less than a decade.”

        Or you could just make it up. The system is in principle a Lorenzian forcing of a resonant system. The physical linkages between polar annular modes, sub-polar gyre strenght and upwelling in both hemispheres are consistent. But what modulates the polar modes? Pure resonance would imply more consistent periodicity – and different periods in the north and south Pacific. Changing periodicities suggest a forcing with a somewhat irregular beat.

        “Models and data suggest that the interplay of major climate modes may result in climate shifts. More specifically it has been shown that when the network of North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO), El Nino/Southern Oscillation (ENSO) and North Pacific Index (NPI) synchronizes, an increase in the coupling between these oscillations destroys the synchronous state and leads the climate system to a new
        state. These shifts are associated with significant changes in global temperature trend and in ENSO variability. Here we probe the details of this network’s dynamics to investigate if a certain oscillation is the culprit in these shifts. From a total of 12 synchronization events observed in three climate simulations and in observations we find that the instigator of
        these shifts is NAO. Without exception only when NAO’s coupling with the Pacific increases a shift will occur. Our results suggest a dynamical sequence of events in the evolution of climate shifts which is consistent with recent independent empirical and modeling studies.”

        So what is the Lorenzian forcing for these approximately 20 to 30 year periodicities?

        “Periodicities of about 27 days that are due to the synodic rotation of the Sun are observed in several parameters of solar activity, such as sunspot numbers, or solar UV or radio fluxes, and in the solar wind (SW). The variation in galactic cosmic ray (GCR) intensity with solar rotation was first studied by Forbush (1938) using a network of ionization chambers (for reviews, see Kudela et al. 2010; Richardson 2004; Vernova et al. 2003, and references therein). Richardson et al. (1999) showed based on
        observations by space probes and neutron monitors in 1954–1998 that the amplitude of the 27-day variation of GCRs depicts the 22-yr Hale cycle. The Hale cycle consists of two successive 11-yr solar activity cycles (also called Schwabe cycles) of opposite magnetic field polarities (Hale et al. 1919).”

        The Hale cycle is a promising candidate.

      • JCH: Or, you can be like all theater skeptics and ignore natural variability.

        There is not enough information to be confident that the rate of sea level rise will not decline as it did twice in the 20th century. Wait 20 years, and we’ll see then if you have made an accurate prediction.

        Can you support the claim that the water in the dams accounts for the slowdown in sea level rise between the 40s and 60s?

      • The graphic is from Wada 2012

        The net contribution to sea level goes negative during the heyday of global dam construction.

      • And nobody is going to be waiting 20 years. You guys are just too funny.

      • JCH: And nobody is going to be waiting 20 years.

        Waiting or not, a lot of people will be analyzing the data 20 years from now.

      • JCH: And nobody is going to be waiting 20 years.

        I should clarify: I meant wait 20 years before making an 80 year prediction. I certainly do not advocate waiting 20 years before upgrading flood control and irrigation facilities. (That is what I expect Californians to do while fighting climate change, but that is a different topic, perhaps.)

      • The run up global temperature from 1910 to WW2 coincides with ramp up of the PDO. Funny that.

        The nosedive of global temperature during WW2 begins right after the nosedive in the PDO. Funny that.

        The stasis in GMST from 1952 to 1975 coincides with the stasis in the PDO. Funny that.

        The ramp up in the gmST from 1975 to 1998 coincides with the ramp in the PDO. Funny that.

        The warming hiatus, which occurred at very high levels of ACO2, just happens to coincide with the negative phase of the PDO starting in 1998 and ending in 2913. Funny that.

        The blast off of the GMST starting in 2012 just happens to coincide with the shift to the positive phase of the PDO. Funny that.

        Now let’s go through the rate of sea level rise.

        The run up PDO from 1921 to WW2 coincides with ramp up of the rate SLR. Funny that.

        The nosedive in the rate of SLR just after WW2 begins soon after the nosedive in the PDO. Funny that.

        The stasis in the rate of SLR from 1952 to 1975 coincides with the stasis in the PDO. Funny that.

        The ramp up in the rate of SLR from 1975 to 1998 coincides with the ramp in the PDO. Funny that.

        The relative satellite era SLR rate hiatus, which occurred at very high levels of ACO2, just happens to coincide with the negative phase of the PDO starting in 1998 and ending in 2013. Funny that.

        The blast off in the rate of SLR starting in 2012 just happens to coincide with the shift to the positive phase of the PDO. Funny that.

        It’s just hilarious how many big global things just happen to agree with shifts in the direction of the PDO index.

        Here comes the cut and paste attack. The warming hiatus is going to last 1 or 2 more decades. Just wait, and, well: pray really hard. And,most important of all, at all costs, ignore and deny the warming half of natural variability.

      • jch

        interesting data on the pdo.

        Do you have data on the pdo dates any further back than 1909 or was that never calculated?


      • Thanks Robert


      • Dashed lines are phase shifts:

      • Fig S3 from Wills supplementary information:

      • Above figures are from:

        Disentangling Global Warming, Multidecadal Variability, and El Niño in Pacific Temperatures

        A key challenge in climate science is to separate observed temperature changes into components due to internal variability and responses to external forcing. Extended integrations of forced and unforced climate models are often used for this purpose. Here we demonstrate a novel method to separate modes of internal variability from global warming based on differences in time scale and spatial pattern, without relying on climate models. We identify uncorrelated components of Pacific sea surface temperature variability due to global warming, the Pacific Decadal Oscillation (PDO), and the El Niño–Southern Oscillation (ENSO). Our results give statistical representations of PDO and ENSO that are consistent with their being separate processes, operating on different time scales, but are otherwise consistent with canonical definitions. We isolate the multidecadal variability of the PDO and find that it is confined to midlatitudes; tropical sea surface temperatures and their teleconnections mix in higher‐frequency variability. This implies that midlatitude PDO anomalies are more persistent than previously thought.

      • JCH: Disentangling Global Warming, Multidecadal Variability, and El Niño in Pacific Temperatures

        Thank you for the link. I could download the supporting online information, but not the main article.

      • I’ve run into the same problem in the last week on all Geophysical Research Letters papers. Previously they were not paywalled and the entire article was easily accessed. I have played around trying to get the entire paper but to no avail. I have no idea how to unlock it. There are many excellent papers out there from GRP.

      • matthew/others –

        1. Open Google Scholar.
        2. type, Wills PDO into the search box
        3. on the left click on Since 2018 (should be first paper, a PDF link from caltech is off to the right)
        4. click on All 5 Versions link. There is a PDF available, University of Washington.

      • JCH: matthew/others –

        thank you. I got it.

      • JCH: It’s just hilarious how many big global things just happen to agree with shifts in the direction of the PDO index.

        Sure: there is an explanation struggling to be born. In the mean time, we have hypotheses about processes observed for at most two oscillations from peak-to-peak-to-peak. Before being considered reliable for any practical purposes, the hypotheses have to be tested against out-of-sample data, at minimum the observations of the next 20 years for the oscillation in estimated rates of sea level rise.

      • Cloud science will be the end of it. It won’t take 20 years. The PDO is a proxy for the eastern Pacific.

    • My guess is 1-2ft. For policy purposes, I’d allow for 2ft. Might be good for 100yrs. Maybe 200. Maybe forever.

  24. Why is the contribution of geothermal heating of the deep ocean completely omitted from this discussion? We might expect such tectonic heating to occur randomly in time and space as it does on land and to be cumulatively 3 or 4 times as great as on land given the greater area of the oceans. For some reason it is always assumed to be steady state and to be negligibly small. Heating of the deep ocean by HTVs is the sole source of 3He so that this isotope is an excellent tracer of HTV heat in the deep ocean. Its distribution at 2500m in the Pacific can be seen here: The heat associated with this continental-sized 3He plume is sufficient to raise the local temperature by ~1 degC and so contribute about half a meter to global mean sea level. The standard deviation of this contribution is unknown but it is likely to be similar in magnitude.

    • And then there is also the issue of whether or not rising seas actually matter. Here in the states we’ve seen massive demographic shifts in urban areas since 1950. These fast rising demographic tides have left wholesale destruction in their wake, the city of Detroit being the poster child of that destruction. And these demo tides don’t stop rising! Here in New Orleans the white demographic largely left fifty years ago moving west of the city. In the last quarter century that population west of the city has begun its exodus to across lake ponchartrain. (so the music never stopped) If society can accept these massive and relatively quick changes, then it should have no problem with the slow moving encroachment of the seas…

      • johnvonderlin

        Are you saying that because I accepted my first wife running off with the Fuller Brush man, that I should accept my second wife putting small amounts of ant poison in every meal she serves me?

      • No, i’m saying that demographic shifts have already proven to be just as bad if not worse than projected sea level rise. (your marital problems ain’t none of my beeswax)…

  25. P. J. Watson (2011) Is There Evidence Yet of Acceleration in Mean Sea Level Rise around Mainland Australia?. Journal of Coastal Research: Volume 27, Issue 2: pp. 368 – 377.


    As an island nation with some 85% of the population residing within 50 km of the coast, Australia faces significant threats into the future from sea level rise. Further, with over 710,000 addresses within 3 km of the coast and below 6-m elevation, the implication of a projected global rise in mean sea level of up to 100 cm over the 21st century will have profound economic, social, environmental, and planning consequences. In this context, it is becoming increasingly important to monitor trends emerging from local (regional) records to augment global average measurements and future projections. The Australasian region has four very long, continuous tide gauge records, at Fremantle (1897), Auckland (1903), Fort Denison (1914), and Newcastle (1925), which are invaluable for considering whether there is evidence that the rise in mean sea level is accelerating over the longer term at these locations in line with various global average sea level time-series reconstructions. These long records have been converted to relative 20-year moving average water level time series and fitted to second-order polynomial functions to consider trends of acceleration in mean sea level over time. The analysis reveals a consistent trend of weak deceleration at each of these gauge sites throughout Australasia over the period from 1940 to 2000. Short period trends of acceleration in mean sea level after 1990 are evident at each site, although these are not abnormal or higher than other short-term rates measured throughout the historical record.

  26. Arthur Parker (2013)
    Oscillations of sea level rise along the Atlantic coast of North America north of Cape Hatteras
    Natural Hazards
    January 2013, Volume 65, Issue 1, pp 991–997 | Cite as

    Parker, A. Nat Hazards (2013) 65: 991.

    It is shown in the short comment that the sea levels are oscillating about a longer-term trend and that the sea level rise (SLR) computed with time windows of 20, 30 or 60 years also oscillates, with the amplitude of these latter oscillations reducing as the time window increases. The use of only two values of the SLR distribution is misleading to infer conclusions about the accelerating behaviour. In particular, the comparison of the 30-year SLR 1950–1979 with the 30-year SLR 1980–2009 for the tide gauges along the Atlantic coast of North America north of Cape Hatteras to infer an accelerating behaviour is particularly wrong because the 30-year time window is a too short interval to appreciate the longer-term sea level trend cleared of the multi-decadal oscillations, and the two values from the SLR distribution are computed, respectively, at the times of a valley and a peak for the 60-year Atlantic Ocean multi-decadal oscillation. By using a 60-year time window or all the data since opening when more than 60 years of recording are available and by analysing the SLR time history, the only conclusion that can be inferred from the analysis of the tide gauges along the North American Atlantic coast is that the sea levels are oscillating without too much of a positive acceleration along their longer-term trend.

  27. Sea Level Changes past Records and Future Expectations
    Nils-Axel Mörner

    Energy and Environment
    Volume: 24 issue: 3-4, page(s): 509-536
    Article first published online: June 1, 2013; Issue published: June 1, 2013

    The history and development of our understanding of sea level changes is reviewed. Sea level research is multi-fascetted and calls for integrated studies of a large number of parameters. Well established records indicate a post-LIA (1850–1950) sea level rise of 11 cm. During the same period of time, the Earth’s rate of rotation experienced a slowing down (deceleration) equivalent to a sea level rise of about 10 cm. Sea level changes during the last 40–50 years are subjected to major controversies. The methodology applied and the views claimed by the IPCC are challenged. For the last 40–50 years strong observational facts indicate virtually stable sea level conditions. The Earth’s rate of rotation records a mean acceleration from 1972 to 2012, contradicting all claims of a rapid global sea level rise, and instead suggests stable, to slightly falling, sea levels. Best estimates for future sea level changes up to the year 2100 are in the range of +5 cm ±15 cm.

  28. Mid-Pacific microatolls record sea-level stability over the past 5000 yr
    Colin D. Woodroffe Helen V. McGregor Kurt Lambeck Scott G. Smithers David Fink
    Geology (2012) 40 (10): 951-954.
    Published: October 01, 2012

    There has been geographical variation in sea level since rapid postglacial melting of polar ice ceased ~6 k.y. ago, reflecting isostatic adjustments of Earth and ocean surfaces to past (and ongoing) redistribution of ice and water loads. A new data set of over 100 fossil microatolls from Christmas (Kiritimati) Island provides a Holocene sea-level record of unparalleled continuity. Living reef-flat corals grow up to a low-tide level. Adjacent fossil microatolls, long-lived Porites corals up to several meters in diameter, occur at similar elevations (±0.1 m), and extensive fossil microatolls in the island interior are at consistent elevations within each population. Collectively, they comprise an almost continuous sequence spanning the past 5 k.y., indicating that, locally, sea level has been within 0.25 m of its present position, and precluding global sea-level oscillations of one or more meters inferred from less stable locations, or using other sea-level indicators. This mid-Pacific atoll is tectonically stable and far from former ice sheets. The precisely surveyed and radiometrically dated microatolls indicate that sea level has not experienced significant oscillations, in accordance with geophysical modeling, which implies that the eustatic contribution from past ice melt and the isostatic adjustment of the ocean floor to loading largely cancel each other at this site.

  29. Short term comparison of climate model predictions and satellite altimeter measurements of sea levels
    Author Alberto A.Boretti

    Coastal Engineering
    Volume 60, February 2012, Pages 319-322 rights and content
    Climate models (, 2011;, 2011; Rahmstorf, 2007, 2010) calculate that temperatures are increasing globally and sea level rises are increasing due to anthropogenic carbon dioxide emissions. More recent predictions (, 2011; Rahmstorf, 2007, 2010) have forecasted that sea level rises by 2100 will be higher than the 2007 projections by the Intergovernmental Panel on Climate Change (, 2011), with projected sea level rises increasing from 18–59 cm to 100 cm. In this brief communication, the predictions of Rahmstorf (2007) are validated against the experimental evidence over a 20 year period. The University of Colorado Sea Level satellite monitoring shows that the rate of rise of the sea level is not only well below the values computed in (2011) and Rahmstorf (2007, 2010), but actually reducing rather than increasing (, 2011b; 10,11). These results suggest that sea level predictions based solely on the presumed temperature evolution may fail to accurately predict the long term sea levels at the end of the century.

  30. Dear Judy – last week I had attended the local police station (London UK)after making a formal complaint about the police failing to deal with my concern about endemic fraud in relation to research which I had put forward. That it was constantly confronted with a carpet of fiction and denial, in a bid to dismiss its findings. The situation became even more absurd when your AEW density tracker was put forward along with the most basic evidence that we have changed this system – evidence which even the Keystone Cops could understand……. The police inspector then decided to argue the proof that the AEW system even existed. Your smoking gun is that that we altered at least one of the precursors to this AEW system at the start of the last century, then again in the mid 60s. The problem is the near blanket denial that this has happened and that it has affected all other systems. My preferred area of research is clinical psychology, why children from dysfunctional families end up populating the criminal justice system, rehab clinics and bookie shops. At least these kids have some kind of mitigation and excuse for their failings – the BS science community does not.
    watch this space – this pantomime has many more scenes before the curtain falls.

  31. Reblogged this on Climate Collections.

  32. Great article, thank you. Minor grammatical error here: “…we were seeing the affects [sic] of the Atlantic Multidecadal Oscillation (AMO).”

  33. Pingback: Weekly Climate and Energy News Roundup #309 | Watts Up With That?

  34. Pingback: Weekly Climate and Energy News Roundup #309 |

  35. Pingback: Sea level rise acceleration (or not): Projections for the 21st century | Climate Etc.

  36. Pingback: A look at the workings of Propaganda Inc. - Fabius Maximus website