4 4. DISCUSSION AND IMPLICATIONS

A fundamental point that emerges from the data presented here is that albedo and thermal conductivity can exert a significant influence on maximum stone surface temperatures and surface-subsurface temperature gradients. While this effect seems obvious, relevant empirical evidence is rare, and such properties are not generally taken into account in considerations of stone weathering under either field or laboratory conditions.

In terms of conservation studies, the data presented have several implications. First, microenvironmental conditions experienced at and just below stone surfaces can be more variable and more, or indeed less, extreme than macroenvironmental data, particularly air temperature, might suggest. Consequently, inferences regarding the nature and potential effectiveness of stone deterioration processes drawn solely from meteorological observations should be guarded. The view that detailed environmental monitoring should inform the development of conservation programs and management strategies at cultural sites (Thorn 1994; Thorn and Piper 1996) is reinforced by the data reported here. Particular care should be taken not to generalize where more than one stone type is present at sites under investigation.

Second, results from the direct and indirect heating experiments raise questions concerning the usefulness of laboratory-based stone weathering simulations and durability tests, not only of the stone itself but of any surface treatments applied to the stone surface. Invariably, indirect heating methods are used and result in test specimens' being cycled through “lithologically indiscriminate” temperature regimes. Stone types with a high albedo (such as the Portland limestone reported here) are, therefore, likely to be subjected to high surface and subsurface temperatures that they may rarely, if ever, experience under natural conditions of exposure. In contrast, other lithologies (such as the basalt), which frequently reach high temperatures in the natural environment, may not attain these in laboratory tests. As a consequence, test blocks used in wetting and drying tests may dry at a slower rate and may not dry out completely over the duration of a test cycle. Where salts are present, a slower drying rate may result in crystallization at the surface as an efflorescence, compared to rapid heating or drying under natural conditions, which is more likely to induce crystallization at depth. Crystallization at depth is more likely to produce phenomena such as flaking and scaling (Smith and McGreevy 1988), whereas surface crystallization is more commonly associated with granular disaggregation. Obviously, incomplete drying of the samples will also limit the overall crystallization of any salts within a stone.

The significance of these observations lies in their implications for (1) the relative effectiveness of mechanical weathering processes, and (2) the overall assessment of stone durability through exposure to these processes in oven-based tests.

Laboratory evidence reported in the open literature points to a positive relationship between temperature and the effectiveness of mechanical weathering such as salt crystallization (Price 1978; Sperling and Cooke 1985; Davison 1986; Goudie 1993). High stone temperatures may enhance breakdown by increasing salt crystallization pressures (Winkler and Singer 1972; Sperling and Cooke 1980) and concentrating salts in surface and near-surface layers. Consequently, in oven-based tests, indirect heating may contribute to more serious deterioration of, for example, samples with a relatively high albedo because they will be forcibly cycled through uncharacteristically high surface and subsurface temperature regimes. Conversely, salt crystallization may appear to be less effective in lithologies with a low albedo because the high surface-subsurface temperature frequently experienced in the natural environment may not be replicated under laboratory conditions. However, given that the majority of laboratory tests employ quite high temperatures (Marschner 1978; Ross and Butlin 1989, crystallization test), it is probable that most have tended to “overweather” high albedo stone types.

Disparities are often reported between the durability of stone as indicated by laboratory tests and behavior once exposed to “real world” conditions. Given that stone deterioration reflects a complex interaction between stone properties and environment, such disparities may well derive from the use of unrealistic parameters and/or the exclusion of relevant components in experimental design: failure to include the effect of thermal properties through the use of indirect heating methods is an example. Any comparative assessment of susceptibility to, for example, salt weathering, made on the basis of such conditions will not provide a true measure of weathering effectiveness or stone susceptibility. If direct heating of specimens is not possible, each stone type should be assessed separately using indirect heating regimes that reflect temperatures experienced under natural conditions, i.e., they should be both site- and stone-specific. Only through the inclusion of all properties that control the interaction of stone with other deterioration factors can more accurate predictions of behavior under a given set of conditions be made.

Finally, and specifically, the results of the soiling experiments add to the conservation debate as to whether features such as the pollution-related black crusts and stains that typically develop on building surfaces infrequently exposed to rainwash should be removed. Previous arguments have centered around questions of aesthetics, historical integrity, cost, the need for consolidation of cleaned stone, and the role of crusts as repositories of salts that can migrate into the underlying stone (Maxwell 1992). Our data suggest that, for some stone types, removal of dark surface stains and soiling could at least reduce the degree of mechanical stressing that they experience as a result of temperature fluctuations, especially rapid short-term changes. Under such conditions, temperature fluctuations are most marked in the outer millimeter or so—the precise zone in which granular disaggregation occurs. This finding is of particular relevance to lithologies that comprise a variety of minerals with different thermo-elastic properties, but it also has implications for homogeneous varieties, such as marble, in which grain orientation is variable (Yatsu 1988; Lewin 1990).

Much scope remains for the collection of microenvironmental data from “soiled” and “clean” stone surfaces under “real world” conditions. Furthermore, in natural environments (unpolluted), especially in hot climates, the effect of albedo change may also be an important factor in stone deterioration. Development of rock varnishes (Dorn and Oberlander 1982; Whalley et al. 1984), weathering rinds (Anderson and Anderson 1981; Dorn et al. 1991), iron staining (Dorn 1998) and algal and lichen growth (Satterwhite et al. 1985; Urz� et al. 1993) can all alter stone surface coloration. Although the causes may differ from those prevalent in urban areas, the overall effect should be the same, i.e., more extreme temperature regimes at the stone surface.