By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Long duration time-series of atmospheric variables are useful for providing a basis for comparison with modern measurements. These can be obtained from proxy measurements, or by re-examining existing data sources for additional information.
An example of a meteorological data source in widespread use that has probably only been partially exploited is the Campbell-Stokes sunshine recorder (CSSR), invented in the late nineteenth century to provide a measurement of the duration of bright sunlight by making a burn mark on a treated ‘burn card’. Each card provides a continuous record of the state of the sky during daylight hours by recording the absence or the presence (and indeed burning power) of the Sun's rays. Measurement of the lengths of the burns gave the daily sunshine duration, but the burn marks also provide, to some extent, additional embedded information about the state of the sky during the day.
Analogue sunshine recorders are now increasingly being replaced by modern electronic sunshine sensors which measure the sunshine duration in a different way. Because the electronic measurement is solely that of sunshine duration, there is likely to be a temptation to regard the transcription of the sunshine duration from the burn cards as sufficient data with which to form a combined long data series, and as a presumed safe basis on which to destroy the original burn cards. Our purpose here is to emphasise that the additional sky-state information that can be gleaned from the cards, linked with other geophysical archive sources, may be irreplaceable. There are good reasons to seek new information on past cloud properties quantitatively, not least because of the importance of clouds in the Earth's radiation balance (Solomon et al., 2007).
The burn card of the Campbell-Stokes sunshine recorder
The CSSR records sunlight without a mechanical tracking system, through use of a spherical glass lens (Figure 1) to focus the Sun's rays from the variety of possible solar positions. The lens focuses the direct beam onto a treated card placed in a metal holder at the base of the recorder (Met Office, 1982; Stanhill, 2003; Strangeways, 2003).
On cloud-free days, as the relative positions of the Sun and Earth change, a continuous burn is scorched across the length of the card. There is an onset threshold of solar radiation to initiate the burn, which varies slightly and is affected by the state of the card (such as its moisture content: Painter, 1981) and the quality and physical state of the lens (Curtis, 1898). At the onset of the burn – as the solar beam brightens – the card initially begins to mark, then to scorch more substantially, only burning through the card completely in strong sunlight. The length of the burn(s) provides an estimate of the daily summed ‘bright sunshine’ duration, following a standard procedure which allows for intermittent burns (Met Office, 1982). This variable has often been used by seaside resorts in holiday advertisements (e.g. in a 1957 tourism brochure: Fraser, 2009).1
Extra information from the burn card
To retrieve the extra information about the state of the sky that the permanent burn marks may provide, various approaches have been devised. For example, empirical relationships have been derived to estimate the direct-beam irradiant energy I (Jm–2) from the number of sunshine hours (Wright, 1935; Stanhill, 1998). In one novel approach, the total post- exposure mass of 100 burn-cards from Mexico City was shown to correlate well (correlation coefficient r = +0.93) with direct-beam irradiant energy (Galindo Estrada and Fournier D'Albe, 1960). Information has also been obtained about the composition of the atmosphere that the sunlight passes through: such as the number of sunshine hours ‘lost’ due to London smoke pollution in the mid-twentieth century (Cowley, 1976; Hatch, 1981), and estimates of turbidity and aerosol properties (Helmes and Jaenicke, 1984; Horseman et al., 2008).
A different technique, utilising the width of the burn, has been investigated to obtain temporal variations in the direct beam through the day. Strong sunshine (e.g. at the daily maximum) gives a wide burn; the burn width is reduced when the direct-beam intensity is less, such as near dawn and dusk (Figure 2). The arc marked out by the burn mark is smooth on clear days but can show irregularities, such as on 10 May 2008 (Figure 2). The burn starts to become thin (and sometimes disappears) when a cloud covers the Sun, but when the Sun is exposed again then the burn widens once more. This variation of the burn width forms the basis for an estimation of the direct-beam irradiance Sb (Wm–2): a procedure first investigated by Wright (1935). The results originally reported (Figure 3(a)) show the positive relationship between burn width and Sb.
The effectiveness of this method was evaluated using a modern suite of automatic solar radiation instruments at the University of Reading's Atmospheric Observatory (http://www.met.reading.ac.uk/weatherdata/). Burn widths were obtained at 5-minute intervals by making manual measurements on an enlarged copy of the sunshine card from 11 May 2008, and compared with the 5-minute-average Sb measurements made by the Observatory's pyrheliometer (Kipp and Zonen CHP1). Figure 3(b) again shows a positive relationship between direct-beam Sb and burn width w. A relationship is included based on 12 burn cards from Reading during January–May 2008, at 15-minute resolution (Lally, 2008). Interestingly, the data from Reading showed a similar relationship to that found at Kew by Wright (1935), although the burns at Kew were slightly thinner than those at Reading.
The temporal variations in Sb and the measured burn widths match well (Figure 4): the qualitative agreement is clear, and the evidence for the quantitative relationship apparent. Availability of such solar radiation data is potentially useful because surface solar radiation measurements can be used to infer cloud properties (Duchon and Malley, 1999; Harrison et al., 2008), implying in turn the possibility of retrieving cloud information from burn cards. The interpretation of data from partly cloudy skies thus requires further detailed investigation (Campbell, 1879; Pallé and Butler, 2002; Stanhill, 2003), notwithstanding the fact that some basic statistics can be readily obtained from the daily sunshine totals: such as totally overcast or clear days. Finally, observations of burn cards suggest that clouds that are optically thin lead to a surface scorch mark rather than burning fully through the card, implying the possibility of retrieving limited cloud type and/or thickness information.
The CSSR burn cards provide an illustration of the importance of not only conserving the data originally obtained, but also the core geophysical material from which currently unforeseen data may yet be retrieved. The extent of this is clearly difficult to estimate in general, but there is no reason not to expect further serendipity in extracting additional scientific information from existing archive material.
The ever more effective utilisation of historical records can be facilitated through improvements in theoretical models providing new proxy methods, such as those used in determining information on the solar position in Monet's London Series paintings (Baker and Thornes, 2006) or in deriving the diurnal cycle of eighteenth-century air pollution in London from primitive electrostatic measurements (Harrison, 2009). Even sources previously mined extensively for one variable might still yield new information about other variables. An illustrative case is that of a weather diary used as a source for part of the Central England Temperature data series, which also contained indirect solar activity information in the form of auroral observations (Harrison, 2005). A further example is presented by ships' logs made for navigational purposes, but now providing datasets for climatology (Wheeler and García-Herrera, 2008) and paleomagnetism (Jackson et al., 2000).
For the case of the CSSR, which was adopted operationally from the 1880s, measurements exist at some sites for over a century. During 1992, for example, an estimated 303 global stations outside the UK had CSSRs recording data (Stanhill, 2003). There are therefore potentially tens of thousands of station-years of data on the state of the sky which remain under exploited, and on which future cloud retrieval techniques may depend.
Protection of the core geophysical material in perpetuity, to allow for development of better data-retrieval techniques, clearly presents a completely unrestricted commitment. In this respect, practicality will necessarily force some compromises, but there is a sense in which the necessary small amount of continued investment – archiving and conservation – is about maximising the initial investment and effort in the staffing and infrastructure required for the original data acquisition. In a rather humble way, sunshine recorder burn cards offer an atmospheric parallel with space missions which have generated science results well beyond those originally planned because the original data have remained available. At the least therefore, the protection (or destruction) of an irreplaceable geophysical archive – such as that of sunshine recorder burn cards – brings an obligation to consider its potential use beyond that for which it was originally intended.
Ken Spiers provided the sunshine recorder burn cards. Techniques for the digitisation of burn cards have been extensively investigated by John Lally (2008), who also provided the burn-width measurements.
The record total for a calendar month is 383.9 hours at Eastbourne (East Sussex) in July 1911 (Met Office, 2007).