January 20, 2012

The Changing Influence of Time

Filed under: Climate Changes

Since anyone first heard of global warming, we have been told that a warmer world would result in higher moisture levels in the atmosphere, and intensification of the hydrological cycle, and any number of negative consequences could result (e.g., more floods, more intense storms). A warmer world would almost surely mean more evapotranspiration (ET) – hard to disagree on that front.

Well, in the real warmer world, it appears that things aren’t that straightforward (but what’s new about that!).

A 2010 article in Nature entitled “Recent decline in the global land evapotranspiration trend due to limited moisture supply” was produced by 33 scientists from every corner of the planet. Furthermore, funding for the study came from every corner of the world as well.

Jung et al. begin noting “Climate change is expected to intensify the hydrological cycle and to alter evapotranspiration, with implications for ecosystem services and feedback to regional and global climate. Evapotranspiration changes may already be under way, but direct observational constraints are lacking at the global scale. Until such evidence is available, changes in the water cycle on land—a key diagnostic criterion of the effects of climate change and variability—remain uncertain.” They further state “Acceleration or intensification of the hydrological cycle with global warming is a long-standing paradigm in climate research, but direct observational evidence of a positive trend in global ET is still lacking”. Once again, we see great uncertainty acknowledged for one of the key pillars in the climate change issue!

The team used a multi-prone approach to estimate ET from land areas all over the planet, and the results are seen below. They report that “Our results suggest that global annual evapotranspiration increased on average by 7.1 ± 1.0 millimetres per year per decade from1982 to 1997.” The finding is described as “consistent with the expected ‘acceleration’ of the hydrological cycle caused by an increased evaporative demand associated with rising radiative forcing and temperatures.”

Just when everything looked to fit with a warming world, they discovered that the trend has reversed since that time. Amazingly, they find that “The trend of the median global land-ET anomalies derived from these models becomes negative during 1998–2008 (-7.9 mm per year per decade, P < 0.05).” The downward trend in the most recent decades is of a similar magnitude as the upward trend in the previous 15 years. Jung et al. conclude “It is hard to evaluate whether this is part of a natural climate oscillation, or a climate-change signal in which land evapotranspiration becomes more and more supply-limited in the long term. The latter would imply that there is a limit to energy- and temperature-driven acceleration of the terrestrial hydrological cycle, and that it may have been reached.”

Figure 1. Annual global land ET anomalies; error bars indicate one standard deviation. Numbers at the bottom show the number of models available each year (figure from Jung et al., 2010)

Shortly after the Jung et al. paper appearing in the literature, a second paper was published on this same subject. This paper appeared in the Journal of Geophysical Research and was written by four scientists with leading institutions in Maryland, Texas, China, and Switzerland. Funding for their effort was provided by NASA, the US National Science Foundation, and the National Research Program of China.

Like Jung et al., Wang et al. also focused on global terrestrial evapotranspiration (ET) noting “How terrestrial ET has varied in the past and what caused the variations, however, have remained quite uncertain”. Wang et al. gathered data from 1982 to 2002 for 1120 stations around the world, and they calculated ET using a popular and widely used set of equations. They found that “Over the period analyzed ET for global land increased by 0.6 Wm−2 per decade equal to 1.2Wm−2 during the study period (about 2.2% in relative value since global averaged land ET is ∼55 W m−2) or 15 mm yr−1 in water flux.” The two papers are quite consistent – Jung et al. would also have seen an increase if they had limited their study to 1982 to 2002. These papers show how results are very sensitive to the beginning and ending dates of the study period.

Another paper in the Journal of Geophysical Research focused on the issue of differing trends depending on starting and ending dates. The work was conducted by three scientists with the University of Alaska and the University of Chicago and the work was supported financially by a National Science Foundation Arctic System Science grant.

Bone et al. begin noting “Quantifying temperature trends across multiple decades in Alaska is an essential component for informing policy on climate change in the region.” We agree, but we knew this would get interesting as the scientists identify three questions for the research including “(1) How sensitive are temperature trend estimates in Alaska to reference start dates? (2) To what degree do methods vary with respect to estimating temperature change in Alaska? and (3) How do different reference start dates and statistical methods respond to climatic events that impact Alaska’s temperature?”

To answer these questions, the team examined five different methods for quantifying temperature trends at 10 weather stations in Alaska. They explain “Five statistical methods were employed to describe temperature change: (1) a 5‐year running average, (2) a 10‐year running average, (3) a 5‐year Hamming filter, (4) a 10‐year Hamming filter, and (5) a linear best‐fit model. The Hamming filter method, employed by IPCC for estimating global temperature trends, uses a neighborhood function that weights observations based on their location within the window. It is similar to a running average in that a set of observations is used to estimate the average temperature at a specific date, but observations farther from the date being estimated have less influence.” No problem here – these are all methods commonly used in this type of research. They discovered that the different methods indeed produced different trend results, but frankly, the differences were not so great.

The really big differences came when they started changing the beginning date for the analyses; the consistently ended the time series in 2003. They conclude “The results from this analysis demonstrate that, with some methods, the discrepancy in temperature trend estimates between consecutive start dates can be larger than the overall temperature change reported for the second half of the 20th century.” They would begin the time series using different years as the start date, including consecutive years, and as seen in their statement above, the difference in the calculated trends could be as great as the overall trend for the past 50 years. Not surprisingly, they found that the trend differences were most striking if one of the starting years was defined by an extreme (high or low) temperature value.

They conclude “These findings emphasize that sensitivity analyses should be an essential component in estimating multidecadal temperature trends and that comparing estimates derived from different methods should be performed with caution. Furthermore, the ability to describe temperature change using current methods may be compromised given the increase in temperature extremes in contemporary climate change.”

A great lesson is learned from these three recent papers – scientists can make decisions on ending and starting dates and that decision alone can have a profound impact on the results of the study (we detailed another prominent example here in our back pages). This is a good thing to keep in mind when assessing the practical (and statistical) significance of any results being presented.


Bone, C., L. Alessa, A. Kliskey, and M. Altaweel. 2010. Influence of statistical methods and reference dates on describing temperature change in Alaska. Journal of Geophysical Research, 115, D19122, doi:10.1029/2010JD014289.

Jung, M., et al. 2010. Recent decline in the global land evapotranspiration trend due to limited moisture supply. Nature, 467, 951-954.

Wang, K., R.E. Dickinson, M. Wild, and S. Liang. 2010. Evidence for decadal variation in global terrestrial evapotranspiration between 1982 and 2002: 2. Results. Journal of Geophysical Research, 115, D20113, doi:10.1029/2010JD013847.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress