118 WEATHER, CLIMATE, AND SOCIETY VOLUME 4 Interdisciplinary Climate: The Case of the First 50 Years of British Observations in Australia KENDAL MCGUFFIE AND ANN HENDERSON-SELLERS School of Physics and Advanced Materials, University of Technology Sydney, and Department of Environment and Geography, Macquarie University, Sydney, New South Wales, Australia (Manuscript received 25 January 2012, in final form 12 June 2012) ABSTRACT This paper presents the case for improved interdisciplinarity in climate research in the context of assessing and discussing the caution required when utilizing some types of historical climate data. This is done by a case study examining the reliability of the instruments used for collecting weather data in Australia between 1788 and 1840, as well as the observers themselves, during the British settlement of New South Wales. This period is challenging because the instruments were not uniformly calibrated and were created, repaired, and used by a wide variety of people with skills that frequently remain undocumented. Continuing significant efforts to rescue such early instrumental records of climate are likely to be enhanced by more open, interdisciplinary research that encourages discussion of an apparent dichotomy of view about the quantitative value of early single-instrument data between historians of physics (including museum curators) and climate researchers. 1. Interdisciplinary climate efforts to rescue and analyze weather recordings of this vintage, the opinion of some museum curators and his- a. Overview torians of environmental physics is that quantitative study The study of climate is frequently claimed to be ‘‘in- of the meteorological measurements from this period is terdisciplinary’’ (Schmidt and Moyer 2008): a characteris- of questionable value. This is because the instruments tic seen as recent and not always advantageous (Averyt were inadequately calibrated and, despite being built by 2010). The exhortations to climate researchers to embrace craftsmen, were frequently repaired and used by a wide the ever-broadening knowledge pool can be strident, with variety of people with unknown skills. In contrast to these perceived failure to do so blamed for many ills including views, recently rescued late eighteenth-century instru- poor communication of science to the public (Pidgeon and mental records of weather are being concatenated with Fischhoff 2011) and even stalled progress in climate un- modern climatic records, albeit after careful comparisons derstanding itself (Hamilton 2010). In climate research, and evaluation (Frauenfeld et al. 2011). Our goal is to ‘‘interdisciplinary ventures have led to both rapid growth explore these conflicting perceptions of early instrument- in insight and near-catastrophic blunders’’ (McGuffie and derived weather measurements as quantitative climate Henderson-Sellers 2005, p. 66) while its traditional un- records, as an exemplar of the need for increased open- derpinning disciplines are said to carry ‘‘a hubris that they ness in climate research. The case of Australia is chosen seemingly cannot shake off’’ (Hulme 2011, p. 178). because the British settlement was of convicts, their This research concerns the use of instrumental weather guards, and administrators, so that although there was records from the late eighteenth and very early nine- a clear need to understand climate for effective agricul- teenth centuries. Specifically, we examine the case of ture and to nurture water supplies, the invading force had climatically relevant measurements made by the British little or no meteorological training (Phillip 1789). in Australia from 1788 to 1840. Despite many excellent b. Climate reconstruction Historical records of temperature are a constant Corresponding author address: A. Henderson-Sellers, Department of Environment and Geography, Macquarie University, Sydney, source of discussion in climate science (e.g., Nicholls NSW 2109, Australia. 1988; Jones and Henderson-Sellers 1992; Jones et al. E-mail: [email protected] 2001; Wheeler et al. 2010; Bra´zdil et al. 2010). Although DOI: 10.1175/WCAS-D-12-00005.1 Ó 2012 American Meteorological Society Unauthenticated | Downloaded 09/30/21 02:01 PM UTC APRIL 2012 M C G U F F I E A N D H E N D E R S O N - S E L L E R S 119 it is not until the mid-1800s that a consistent global ob- d ‘‘The dates for invention of the thermometer, barom- servational record becomes established, there are many eter, etc. do not mean that they were accurate observations of temperature over the century preceding then—by no means.’’ this that allow, or at least appear to allow, a longer record d ‘‘The crudity of (instrument) manufacture at that in a number of important locations. Many of these records time—when craftsmen had difficulty making an have been examined in detail (e.g., Vinther et al. 2006; even-bore tube—presented real problems.’’ Gergis et al. 2009). However, temperature records in d ‘‘Shipping barometers and thermometers to Australia these early periods have significant problems (e.g., Box meant carrying them through the tropics and that is 2002; Garcia-Herrera et al. 2005a). They represent in- where many broke as the mercury expanded and formation gathered in a totally different social and tech- shattered the glass.’’ nical environment, at a time when the quantification of d ‘‘If the glass and mercury was transported separately, the concepts of heat and cold was a novelty (e.g., Roebuck and assembled on arrival, with no reliable standard 1775) and when the scientific understanding of tempera- barometer to compare with, this was not always the ture aspects of the weather was limited (cf. Golinski answer.’’ 2007). They are from an era when instruments were d ‘‘The problem with barometers carted round the handcrafted and therefore individual and when charac- world is that they could get broken and then have teristics we take for granted today (such as the fixed the glass tube restored—sometimes in London at the points of the temperature scale) were not only in dispute original maker, but possibly in a distant port by some but were very difficult to define (McConnell 2005). Most less skilful person.’’ importantly, the background and education of observers d ‘‘Against what standard have they [the instruments] and interpreters of the observations differed significantly. been adjusted?’’ While other British colonies generally benefitted from d ‘‘The whole idea of assessing the reliability and well-educated (wealthy) invaders, this was not true in the comparability of measurements was hardly con- early period of British settlement of New South Wales. fronted in 1800.’’ Reconstructions of a quantitative record of climate Such (curators’) perceptions of the instruments and ob- are admirable, having a heritage extending over the re- servations from the late eighteenth century are not orig- construction of the central England temperature series inal. The Societas Meteorologica Palatina was founded in by Manley (1953, 1974) through Hubert Lamb’s work Mannheim in 1780, for example, specifically to compare (e.g., Lamb 1995) to the present day [e.g., Atmospheric meteorological observations across Europe. Its activity Circulation Reconstructions over the Earth (ACRE; was immediately curtailed by the extant problem of the Allan et al. 2011), IMPROVE project (Camuffo and total lack of standardization of instruments (e.g., Cappel Jones 2002), and the Millennium European Climate 1980; Cassidy 1985). Project (Wheeler et al. 2009)]. However, there is often no explicit statement about instrument error and un- certainty. This is because uncertainty is frequently un- 2. Historical view of early Australian documented and, due to instrument loss or degradation, meteorological instruments today unknowable. The overall spectrum of quality and a. Uncertainty in meteorological measurements homogeneity issues that affect any historical time series around 1780 are assumed to be included in total uncertainty state- ments or taken as averaged out in composites. Improved Modern temperature measurements use resistance and estimates of the effect of instrument error and inho- other characteristics but, in the eighteenth century, the mogeneity issues are still actively being sought (e.g., property typically used in thermometers was the change Brunet et al. 2010), and time series are periodically of volume of some thermometric substance in response to reevaluated and updated in the light of new information a change in the warmth or cold (Middleton, 1966). (e.g., Parker and Horton 2005). Thermometers in the 1760s–1780s were calibrated by two Our study serves to underline this uncertainty. Our fixed reference points to define a linear relationship be- evaluation involved curators from the United Kingdom’s tween the thermometric property and (what we know as) Science Museum, current and past, as well as experts temperature (e.g., Chang 2004). Considering a simple in meteorological history in Australia and the United liquid-in-glass thermometer, uncertainties can arise from Kingdom. During an investigation of instruments equiv- uncertainties in the chemical composition of the medium alent to those transported to Australia in the period 1788– (e.g., the presence of water in an alcohol thermometer 1840, some museum (instrument) curators made the may introduce some nonlinear behavior); from a de- following comments: pendence of the coefficient of thermal expansion on Unauthenticated | Downloaded 09/30/21 02:01 PM UTC 120 WEATHER, CLIMATE, AND SOCIETY VOLUME 4 TABLE 1.