An updated analysis of homogeneous temperature data at Pacific Island stations

Abstract

In this paper we report on the development of a new homogeneous dataset for Pacific Island station temperatures, and the variability and change that these data show. We have worked closely with meteorological services in the Pacific and East Timor to collate and rehabilitate historical climate records, incorporating homogeneity assessment and adjustment where appropriate to produce reliable analyses of temperature trends. Overall, these homogeneity adjustments create no systematic bias compared to the raw data since the early 1960s, but improve both spatial and temporal consistency at the individual station level for both variability and trends. The updated homogeneous records indicate that mean temperatures have increased at all available Pacific island stations over 1961–2010, with rates ranging from +0.04 to +0.33 ºC/decade, across the network as a whole. The strongest warming trends are found in Papua New Guinea and French Polynesia, while none of the homogeneous station records show cooling over the past 70, 50 or 30 years. Trends in maximum and minimum temperatures are very similar to those of mean temperature at most stations, apart from in Fiji, Tonga and Niue where there is a tendency for greater warming for maximum temperatures. The amount of warming in the wet and dry seasons is also similar at most stations, although stations in Fiji, Tonga and Niue display more warming during the wet season. A simple regional mean of Pacific island series shows a trend of +0.16 °C/decade over 1961–2010 and indicates that 2010 was the warmest year on record averaged across the stations available at time of analysis. This regional mean station series compares well with a regional mean series extracted from a global gridded dataset for the last 50 years. Overall, the warming in Pacific Island temperature records over the past half-century is consistent with that expected from humaninduced global warming and is large relative to natural interannual variability associated with factors such as the El Niño-Southern Oscillation.