Before we project how Ascension Island’s climate is likely to change in future, it is first necessary to understand what the climate is like now and how it has varied over recent time. As part of the CRACAB project, Phil Jones and Dave Lister from the University of East Anglia’s Climatic Research Unit (CRU) have been painstakingly assembling available rainfall and temperature data from Ascension Island to construct the first comprehensive, long-term climate timeseries for the Island.

Reconstructing contemporary climate from direct observations is important for several reasons. It allows us to model recent climate change that has already taken place; it helps us to validate and bias-adjust predictions made by global and regional general circulation models (upon which climate projections are based); and it provides the baseline against which future changes are measured. In recent decades, the UK Meteorological Office (UKMO) has operated a permanent meteorological station on Ascension Island that has maintained daily records of temperature and rainfall. Going back further, however, records become sparser and more disparate and we have had to consult a wide range of archives and data sources in order to assemble reasonably comprehensive timeseries.

Some very early temperature records from 1815, 1834 and 1835 are mentioned in the scientific literature, but more continuous recording of air temperatures began at the main settlement of Georgetown (also referred to as ‘Garrison’) in August 1854 until April 1859, albeit with some missing months. Temperature measurements were also summarised from HMS Tortoise for July 1860 to October 1865. The latter are clearly cooler than those in the 1850s, so indicative that they were taken on board a ship anchored offshore. Records of air temperature measurements then recommenced in the late 1910s at Georgetown and are almost continuous there since March 1923 until the mid-1970s. An airfield was built on the island in 1942 by the United States, about 5-6km south of Georgetown, and continuous records from there are available from September 1957, with some earlier data for September 1942 to May 1947. This airfield site is known as Wide Awake Field (referred to as Cat Hill, by the USA) with a second set of measurements from 1989 referred as Airhead (by the UKMO). Figure 1 shows the annual average temperatures from the early 1920s compared to sea-surface temperatures from around the island in a Hadley Centre dataset. The two series are completely independent and the agreement is excellent on the year-to-year timescale, except for some years during WW2

Figure 1: Comparison of the composite Ascension Island annual surface air temperature series developed here with the sea surface temperature series (from HadISST) for the square of ocean where Ascension Island is located.

Putting together a long precipitation record was slightly easier as more of the source records from the island had been collated on the island. Precipitation measurements were taken at the Garrison site and they recommenced at the Georgetown site in 1899, with annual totals for 1895 to 1898. These continued until the end of 1975, but commenced again for 1985-2011. Gaps occur between 1914 and 1923. An airfield site for precipitation starts in 1962 (referred to as Pan-Am/CSR) and continues till 2010. A second gauge starts recording at Airhead in 1984 and continues to the present.

Figure 2: Annual Precipitation totals for Ascension Island from 1899 to present. The composite series combines Georgetown data with the more recent records from the Airfield locations.

For both variables (air temperature and precipitation) we have spent much time trying to access the original records held by the UK Met Office (UKMO) and United States Air Force (from datasets available from the National Centres for Environmental Information, NCEI in Asheville, NC).  Data for the two locations (Georgetown and Wideawake) in well-known global datasets (those at NCEI and also at CRU) often have the same values for the two sites, particularly during the late 1960s and much of the 1970s. Data from UKMO are missing for the entire period 1977 to 1988, which is surprising considering the importance the island played in the Falklands War in 1982. To overcome this, NCEI sent us their hourly data for their airfield location, which was vital to fill this gap.

We plan to write a scientific paper on the work in the coming months which will explain the methodology and data sources used in greater detail.