This article is a follow up to my article from 12-days ago, “Dramatic Confirmation of Temperature Change for 1980-2010”, which presented a global map that showed using the AgMERRA dataset temperature trends between 1980 and 2010 that were statistically significant, in some sense showing climate change, or at least the areas that have experienced climate change over the last 30 years.
Jawoo Koo from the International Food Policy Research Institute (IFPRI) suggested that I examine another global dataset, the CRU Time Series, by Phil Jones and Ian Harris, from the University of East Anglia’s Climate Research Unit. This dataset spans the period 1901 to 2012, much longer than that of AgMERRA, which was only 1980 to 2010. CRU has wider geographic coverage, as well, reaching to the northern- and southern-most parts of the earth. Like the AgMERRA dataset, the CRU dataset has a half-degree resolution. However, unlike the AgMERRA dataset, the CRU only has monthly statistics rather than daily weather.
The map above shows the results. They are very similar to those of AgMERRA from the earlier article. This almost had to be the case, since they were using similar reference data to build the datasets. What is perhaps noteworthy is how AgMERRA’s use of satellite data produced differences. One might note, for example, cooling in southern India that was in AgMERRA but not in CRU.
The table shows the results of analysis for every 30-year period covered in the CRU dataset. The dataset has 67,420 land-based gridcells with data. We see that the 1980 to 2010 period had the highest number of gridcells that had statistically significant changes in temperature. We also see that of those with at least 10% statistical significance, almost 92% were positive changes (increases in temperature). This was second highest to the 1970 to 2000 period.
It is noteworthy to point out the temperature rise between 1910 and 1940. This rise has been discussed much in the literature. I found the Skeptical Science website to have a very interesting and well-presented discussion of that particular rise.
In the mean temperature column, we don’t see much of a rise over the span of the dataset, around 0.5 degrees Celsius between 1900 to 1930 and 1980 to 2010. This points to the value in using regressions and computing trends. Much of the change in climate in terms of temperature has taken place after 1970. Indeed, the table suggests that it was more significant after 1980. It is difficult to pick up the full impact of changes as they are taking place, so computing trends helps identify both magnitude and statistical significance of changes.
We see that the 1980 to 2010 period has the largest median change over a 30-year period, with a median rise of 0.77 degrees Celsius in that period. The distribution is not uniform, it is smaller in the tropics and highest in the polar regions, with bands 0 to 30, 30 to 60, and 60 to 90 degrees north and south having medians rises of 0.59, 0.81, and 1.10.
Analyses like this one and its companion from 12-days ago are helpful to identify where it has taken place over the last 30-years, as well as magnitudes of the changes. It allows everyone from farmers to policymakers to take note of changes that are not simply annual variations in the weather but rather indicative of real climatic changes.
– Timothy S. Thomas
University of East Anglia Climatic Research Unit (CRU). [Phil Jones, Ian Harris]. CRU Time Series (TS) high resolution gridded datasets, [Internet]. NCAS British Atmospheric Data Centre, 2008, Date of citation. Available from http://badc.nerc.ac.uk/view/badc.nerc.ac.uk__ATOM__dataent_1256223773328276 Downloaded May 27, 2014.
Ruane, A.C., and R. Goldberg, AgMIP Hybrid Baseline Climate Datasets: Shifted Reanalyses for Gap-filling and Historical Climate Series Estimation (in preparation).