Study finds global ocean warming has decelerated 50% over the past 50 years

The currently-favored excuse du jour for no statistically-significant global warming over the past 20 years is that the oceans "ate the man-made global warming." However, a 2012 paper published in Nature Climate Change torpedoes this notion, finding the global oceans started warming at least 135 years ago just after the Little Ice Age, on or before the historic voyage of the HMS Challenger in the 1870's. More importantly, the study finds that ocean warming has decelerated 50% over the past 50 years. 

If, as claimed, man-made greenhouse gases are causing the oceans to warm, the opposite would have been expected, namely an acceleration of ocean warming over the past 60 years, beginning in the ~1950's. The fact that the oceans were warming long before CO2 levels significantly increased, and at a higher rate before 50 years ago, clearly demonstrates ocean warming is a natural recovery from the Little Ice Age, and not due to man-made CO2.

The paper is corroborated by a recent paper finding the oceans have warmed only 0.09C over the past 55 years, a rate of 0.016C per decade, and a 34% deceleration from the rate of 0.024C per decade over the past 135 years found by this study.

Further, climate alarmists claim that the "missing heat" is hiding below 1,500 meters deep, but this paper finds the oceans have instead cooled below 1,500 meters over the past 135 years [see third figure below].

In addition, if man-made CO2 was warming the oceans, there should have been an acceleration of steric sea level rise over the past 50 years due to thermal expansion, but no acceleration of sea level rise has been found over the past 203 years.

Related: An additional 60+ links that torpedo the "oceans ate my global warming" theory

Related: Ship find [of a similar ship the HMS Investigator] shows Arctic Sea Ice conditions similar to 1853

New Comparison of Ocean Temperatures Reveals Rise over the Last Century; Ocean robots used in Scripps-led study that traces ocean warming to late 19th century

A new study contrasting ocean temperature readings of the 1870s with temperatures of the modern seas reveals an upward trend of global ocean warming spanning at least 100 years.

The research led by Scripps Institution of Oceanography at UC San Diego physical oceanographer Dean Roemmich shows a .33-degree Celsius (.59-degree Fahrenheit) average increase in the upper portions of the ocean to 700 meters (2,300 feet) depth. The increase was largest at the ocean surface, .59-degree Celsius (1.1-degree Fahrenheit), decreasing to .12-degree Celsius (.22-degree Fahrenheit) at 900 meters (2,950 feet) depth.

The report is the first global comparison of temperature between the historic voyage of HMS Challenger (1872-1876) and modern data obtained by ocean-probing robots now continuously reporting temperatures via the global Argo program. Scientists have previously determined that nearly 90 percent of the [alleged, "missing"] excess heat added to Earth's climate system since the 1960s has been stored in the oceans. The new study, published in the April 1 advance online edition of Nature Climate Change and coauthored by John Gould of the United Kingdom-based National Oceanography Centre and John Gilson of Scripps Oceanography, pushes the ocean warming trend back much earlier.

"The significance of the study is not only that we see a temperature difference that indicates warming on a global scale, but that the magnitude of the temperature change since the 1870s is twice that observed over the past 50 years," said Roemmich, co-chairman of the International Argo Steering Team. "This implies that the time scale for the warming of the ocean is not just the last 50 years but at least the last 100 years."

Although the Challenger data set covers only some 300 temperature soundings (measurements from the sea surface down to the deep ocean) around the world, the information sets a baseline for temperature change in the world's oceans, which are now sampled continuously through Argo's unprecedented global coverage. Nearly 3,500 free-drifting profiling Argo floats each collect a temperature profile every 10 days.

Roemmich believes the new findings, a piece of a larger puzzle of understanding the earth's climate, help scientists to understand the longer record of sea-level rise, because the expansion of seawater due to warming is a significant contributor to rising sea level. Moreover, the 100-year timescale of ocean warming implies that the Earth's climate system as a whole has been gaining heat for at least that long.

Launched in 2000, the Argo program collects more than 100,000 temperature-salinity profiles per year across the world's oceans. To date, more than 1,000 research papers have been published using Argo's data set.

The Nature Climate Change study was supported by U.S. Argo through NOAA.

135 years of global ocean warming between the Challenger expedition and the Argo Programme

Dean Roemmich, W. John Gould & John Gilson

Nature Climate Change 2, 425–428 (2012) doi:10.1038/nclimate1461

Published online 01 April 2012

Changing temperature throughout the oceans is a key indicator of climate change. Since the 1960s about 90% of the excess heat added to the Earth’s climate system has been stored in the oceans^1, 2. The ocean’s dominant role over the atmosphere, land, or cryosphere comes from its high heat capacity and ability to remove heat from the sea surface by currents and mixing. The longest interval over which instrumental records of subsurface global-scale temperature can be compared is the 135 years between the voyage of HMS Challenger³ (1872–1876) and the modern data set of the Argo Programme⁴ (2004–2010). Argo’s unprecedented global coverage permits its comparison with any earlier measurements. This, the first global-scale comparison of Challenger and modern data, shows spatial mean warming at the surface of 0.59 °C±0.12, consistent with previous estimates⁵ of globally averaged sea surface temperature increase. Below the surface the mean warming decreases to 0.39 °C±0.18 at 366 m (200 fathoms) and 0.12 °C±0.07 at 914 m (500 fathoms). The 0.33 °C±0.14 average temperature difference from 0 to 700 m is twice the value observed globally in that depth range over the past 50 years⁶, implying a centennial timescale for the present rate of global warming. Warming in the Atlantic Ocean is stronger than in the Pacific. Systematic errors in the Challenger data mean that these temperature changes are a lower bound on the actual values. This study underlines the scientific significance of the Challenger expedition and the modern Argo Programme and indicates that globally the oceans have been warming at least since the late-nineteenth or early-twentieth century.

Main

• Main•
 
• Methods•
 
• References•
 
• Acknowledgements•
 
• Author information•
 
• Supplementary information

The voyage of HMS Challenger^7, 8, 1872–1876, was the first globe-circling study of the oceans, obtaining multidisciplinary data along a 69,000-nautical-mile track. “One of the objects of the Expedition was to collect information as to the distribution of temperature in the waters of the ocean  …  not only at the surface, but at the bottom, and at intermediate depths”³. The thermal stratification of the oceans was described for the first time from about 300 temperature profiles made using pressure-protected thermometers. The Challenger temperature data set was still prominent in large-scale maps and analyses even into the 1940s (ref. 9). Nothing remotely comparable to the Challenger expedition was undertaken until the 1920s–1950s, when theMeteor¹⁰, Discovery¹¹, Discovery II¹¹, and Atlantis¹² systematically explored the Atlantic and Southern oceans.

Although the Challenger temperature profiles were global in scale, as they were made along the vessel’s track they were not global in the sense of areal sampling. The modern-day Argo Programme, by contrast, is the first globally and synoptically sampled data set of temperature and salinity. Argo’s free-drifting profiling floats collect more than 100,000 temperature/salinity profiles per year, nominally every 3° of latitude and longitude, every 10 days and to depths as great as 1,980 m.

When qualitatively comparing features of the Challenger transect from New York to Bermuda to St Thomas with nearby tracks sampled in the 1950s, C. Wunsch observed “One is hard pressed to detect any significant differences on the large scale”¹³. Now, with an added 50-year interval, a quantitative comparison is made by interpolating Argo data¹⁴ to the location and depth of eachChallenger measurement and to the same time of year, to minimize seasonal sampling bias in theChallenger data set (Fig. 1). Warming is predominant from the sea surface to below 1,800 m. The largest values are in the Gulf Stream (about 38° N), indicating that the current is at a higher latitude in the Argo data than in the Challenger data. Obviously, these local differences may represent any timescale in the 135-year interval—from a transient meander of the Gulf Stream in 1873 to a long-term change in the current’s latitude. Similarly, regional to ocean-scale differences may be affected by interannual to decadal^15, 16 variability, including in the deep ocean¹⁷, and hence our Challenger-to-Argo difference based on stations along the Challenger track must be viewed with caution.

Figure 1: New York–St Thomas transect differences.

Background contours indicate mean temperature (2004–2010) from Argo data¹⁴ along theChallenger’s New York to Bermuda to St Thomas transect. Colour spots show where Argo values are warmer (red), unchanged (white), or cooler (blue) than Challenger, with magnitudes according to the colour scale.

• Full size image (95 KB)

• Figures/tables index
• Next

Seasonally adjusted Argo-minus-Challenger differences reveal warming in both the Atlantic and Pacific oceans (Fig. 2a). The Challenger made only a few stations in the Indian Ocean, all at high southern latitudes, so that region is omitted here. Out of 273 Challenger temperature stations analysed, the Argo-era sea surface temperature (SST) is higher at 212. The mean SST difference is 1.0 °C±0.11 for the Atlantic and 0.41 °C±0.09 for the Pacific Ocean. As the Challenger’s sampling was more intensive in the Atlantic and the warming may be greater in that ocean, we estimate the global difference as the area-weighted mean of the Atlantic and Pacific values, 0.59 °C±0.12. There are extensive historical measurements of SST, providing context for the Argo-minus-Challengercomparison. A time series of reconstructed global mean SST from 1856 to the present day⁵ indicates a cooling of SST from 1880 to 1910, with larger warming since 1910. The overall warming⁵ between the Challenger and Argo eras of about 0.5 °C is consistent with the Argo-minus-Challenger estimate, given the sampling errors.

Figure 2: Spatial differences.

Background contours indicate the mean temperature (2004–2010) from Argo data¹⁴ at the sea surface (a), 366 m (b) and 914 m (c). Colour spots indicate the Argo-minus-Challenger temperature difference, as in Fig. 1; the colour scale is shown above.

• Full size image (194 KB)

• Previous
• Figures/tables index
• Next

Most of Challenger’s subsurface temperature measurements were made using Six’s (Miller–Casella) thermometers. These are maximum/minimum thermometers, with the mercury column displacing a sliding index to record the maximum or minimum temperature, and are fitted with an external bulb to remove the influence of pressure. These instruments were used in the initial belief that temperature decreased monotonically with increasing depth, an assumption discovered to be incorrect during the voyage. Other types of thermometer were used less often, including reversing thermometers that became commonplace later. The Six’s thermometers were graduated in increments of 1° F (0.56 °C) and “the length occupied by one degree (F) could not easily have been subdivided beyond a quarter”³. Hence the temperature, which was recorded to a precision of 0.1° F (0.06 °C), had a reading accuracy of about 0.14 °C. In the report of results¹⁸, the type and serial number of thermometers used on each station are not specified. The sounding line was 8-mm-diameter hemp, with a bottom weight of 25–75 kg (ref. 3). During the measurements the line was “…kept quite perpendicular for 5 min…” and after recovery the thermometers were “…carefully read and registered…and …corrected for errors of zero point…and a curve of temperatures drawn”. It is noted that if there were outliers “…the temperatures at those depths were taken again”³. TheChallenger also deployed water-filled and mercury-filled piezometers, constructed like unprotected Six’s thermometers with one end open³. Together with data from the protected thermometers, these could be used to estimate depth. However, the ratio of temperature to depth sensitivity of these instruments was 1 °C for 783 m of depth change³, so they were useful for correcting only large errors in near-bottom depths. The Challenger data listings¹⁸ do not explicitly state that the fathom (fm) values are uncorrected line-out, but this is evident because the 100-fm and other evenly spaced increments in the data records were obtainable only by measuring and marking the sounding line.

Three sources of systematic error are considered in the Challenger subsurface data. First, taking depth from line-out overestimates the true depth of the thermometer, resulting in a warm bias in the recorded temperature. “If there be a current of any appreciable force, the sounding line begins to wander about, and has to be followed by the ship…an operation of considerable delicacy, even in good weather”³. Second, before the Challenger’s voyage, laboratory measurements of pressure effects on the Challenger thermometers had been made erroneously¹⁹. The post-voyage analysis by P. Tait showed that the actual compression effect on the protected glass thermometers was about 0.04 °C km⁻¹ (0.3° F per 2,500 fm; ref. 19), much less than the prevoyage estimates. We therefore used the raw temperature data¹⁸ rather than the overcorrected version listed in otherChallenger reports. Finally, the Challenger thermometers were mounted in their frames using vulcanite, compression warming of which might be transferred to the glass, causing a small warm bias in the reading¹⁹. Thus, the errors in depth and temperature all tend to make the Challengertemperatures systematically warm at the recorded depths. A small number of temperature measurements were discarded in our analysis. Stations at high southern latitude were excluded owing to the shallow-temperature minimum found there and at a few other locations where Argo indicates a temperature minimum, making them incompatible with the use of maximum/minimum thermometers. The lack of high-latitude and Indian Ocean stations could produce a sampling error in global averages, as multidecadal ocean warming is known to have been strong in the Southern Ocean since the 1930s (ref. 20) as well as having substantial basin-to-basin differences¹⁷. Error bars on our estimates of globally averaged temperature differences are discussed in the Methodssection.

Proceeding downwards to 366 m (200 fm, Fig. 2b) and 914 m (500 fm, Fig. 2c), the pattern of mostly warm differences persists in both oceans, diminishing in magnitude with depth. The global average temperature difference (ocean area weighted) decreases to 0.39 °C±0.18 at 366 m and 0.12 °C±0.07 at 914 m, reaching zero at about 1,500 m (Fig. 3).

Figure 3: Globally averaged difference.

Mean Argo-minus-Challenger temperature difference ±1 s.e.m. The black line is a simple mean over all stations with data at 183-m (100-fm) intervals. The red line uses values for the Atlantic and Pacific oceans in a weighted mean, with weights proportional to the area of the two oceans. The blue line applies the Tait pressure correction¹⁹ (−0.04 °C km⁻¹) to the weighted mean. Error estimates are described in Methods.

• Full size image (75 KB)

• Previous
• Figures/tables index

For the upper 700 m, the ocean area-weighted difference, using only those stations with samples every 183 m, is 0.33 °C±0.14, corresponding to a heat gain of 1×10⁹ J m⁻². This increases to 1.3×10⁹ J m⁻² for 0–1,500 m, or 0.3 W m⁻² of ocean surface area, averaged over the 135-year interval. The average differences, 0–700 m, are 0.58 °C±0.12 for the Atlantic and 0.22 °C±0.11 for the Pacific Ocean. The Tait pressure correction¹⁹, equivalent to −0.04 °C km⁻¹ would increase these values by only 4%. The temperature bias caused by depth errors is difficult to assess, but may be significant at locations such as the equatorial Pacific, where the strong subsurface shear of the Equatorial Undercurrent, not known in Challenger’s time, would cause a slant in the line. Evidence of this bias can be seen (Fig. 2a,b) where temperature differences change from positive to negative between the sea surface and 366 m at near-equatorial Pacific stations. A systematic overestimate of 1% in the depth of Challenger measurements would result in a warm bias in the 0–700 m average temperature of about 0.05 °C.

The Challenger temperature measurements are known to be far from perfect and were the subject of controversy as instanced in the correspondence between J. Murray and W. Leighton Jordan in the late 1880s (ref. 21). However, the data were collected with great care and attention, and the large temperature changes over the subsequent 135 years are revealed by comparing the Challenger and Argo data sets.

We find that the modern upper ocean is substantially warmer than the ocean measured by HMS Challenger in the 1870s and that the warming signal is global in extent. Challenger obtained enough measurements of temperature for statistical confidence at about the 95% level in the mean temperature differences and the nature of systematic errors in the Challenger data makes these differences a lower bound on the true values. Moreover, comparisons with other temperature records including global SST (ref. 5), extensive subsurface data in the Atlantic as early as the 1920s (ref. 22) and global subsurface data over the past 50 years⁶, all indicate that the warming has occurred on the centennial timescale rather than being limited to recent decades. From 1969 to 2009, globally distributed temperature measurements, 0–700 m, showed warming of an average of 0.17 °C (ref. 6), with the Atlantic Ocean warming more strongly (0.30 °C) than the Pacific (0.12 °C). The larger temperature change observed between the Challenger expedition and Argo Programme, both globally (0.33 °C±0.14, 0–700 m) and separately in the Atlantic (0.58 °C±0.12) and Pacific (0.22 °C±0.11), therefore seems to be associated with the longer timescale of a century or more.

The implications of centennial-scale warming of the subsurface oceans extend beyond the climate system’s energy imbalance. Thermal expansion is a substantial contributor to global sea-level rise^23, 24, 25 and extending the record length of subsurface temperature can help in the understanding of the centennial timescale in sea-level rise^26, 27. Furthermore, changes in subsurface temperature and in SST are closely related. SST is important in determining air–sea exchanges of heat and increasing SST is linked to increasing rates of evaporation, and hence precipitation, in the global hydrological cycle^28, 29. The long-term increase of SST should be understood in the context of changes in both temperature and salinity extending deep into the water column.

Enormous advances in ocean-observing technology have occurred from the time of the Challenger, when about 300 deep-ocean temperature profiles were acquired over three-and-a-half years by a ship with more than 200 crew on board, to today’s Argo Programme, obtaining more than 100,000 temperature profiles annually by autonomous instrumentation. The Challenger data set was a landmark achievement in many respects. With regard to climate and climate change, Challengernot only described the basic temperature stratification of the oceans, but provided a valuable baseline of nineteenth-century ocean temperature that, along with the modern Argo data set, establishes a lower bound on centennial-scale global ocean warming.

Methods

• Main•
 
• Methods•
 
• References•
 
• Acknowledgements•
 
• Author information•
 
• Supplementary information

Consecutive Challenger stations were typically spaced 100–300 km apart (Fig. 1) and separated by a few days to months. From the standpoint of mesoscale eddy noise, the temperature data might be considered to be independent from station to station. However, regional variability in temperature and heat content on interannual³⁰ to decadal¹⁷ timescales is noise in the context of our 135-year Challenger-to-Argo difference. To estimate the reduction in the number of independent data points, we divided the multiyear Challenger track into seven continuous segments, four in the Atlantic and three in the Pacific, and calculated the along-track autocorrelation of the Challenger-to-Argo temperature differences as a function of the number of stations of separation. The sample autocorrelation has a narrow peak at all depths, but a low tail extending from three to five stations before decreasing to zero. Taking twice the integral of the autocorrelation gives a correlation scale of three stations at the sea surface, decreasing to two at 500 fm. For simplicity, we use three stations as the correlation scale at all depths. The standard error (s.e.m.) of the Challenger-to-Argo difference was then estimated as the standard deviation (s.d.) divided by the square root of the number of degrees of freedom (NDF), where NDF was the number of stations divided by three. Table 1 lists the s.d. of the temperature difference at each depth together with the number of stations. For the differences in 0–700 m average temperature (and difference in heat content) we used the s.d. of 0–700 m average temperatures, and again NDF is one-third of the number of stations that have data every 100 fm over this depth range. Errors listed in the manuscript are one s.e.m.