Topographic complexity and terrestrial biotic response to high-latitude climate change: Variance is as important as the mean

Abstract

Climate changes that occur as a result of the increased greenhouse effect are expected to be greatest at high latitudes (Cattle and Crossley 1995; Watson et al. 2001), and resultant earth-system responses are likely to be dramatic (Weller 1998). For example, even the small increase in temperature witnessed over the past few decades has melted permafrost in many places, causing major changes in soil drainage, chemistry, and biology, as well as damage to roads, buildings, airports, and pipelines (Williams 1995; Weller 1998; Osterkamp and Romanovsky 1999). The expectation of dramatic environmental change has stimulated numerous empirical and model-based studies addressing both ecosystem responses to predicted changes in temperature and CO2 concentration and how these responses feed back to the climate system (e.g. Sellers et al 1997; Kane and Reeburgh 1998; Arft et al. 1999; Chapin et al. 2000; Eugster et al. 2000; Rupp et al. 2001; Levis et al. 1999, 2000; Friedlingstein et al 2003; Ichii et al. 2003; Jones et al. 2003; Kaplan et al. 2003; McGuire et al. 2002, 2003). Nearly all studies of environmental change are based on modelling and scaling-up activities that use mean values or properties as input. Yet it is well known that the values of climatic parameters vary locally due to topographic influences (e.g. Hungerford et al. 1989; Young et al. 1997). Elevation has well documented effects on temperature and precipitation. Microclimate is also influenced by the radiation load, which varies in response to variation in aspect of slopes ('aspect') and steepness of slopes ('slope'). Hence, rough (mountainous or hilly) landscapes generate a mosaic of diverse microclimates, with direct effects on organism physiology and ecology and biotic community structure and function (Knapp 1985; Young et al. 1997, Porter et al. 2002). Implicit in this variation is variation in ecosystem processes. For example, an increase in temperature will probably lead to a release of CO2 from many high-latitude soils, at least transiently, generating a positive feedback and further warming (Chapin et al. 2000; Jones et al. 2003; Knorr et al. 2005). However, carbon sources and sinks are finely balanced, and the sign of the projected change may be sensitive to scaling-up procedures (Chapin et al. 2000). While analysis of means is practical and necessary in current modelling, ignoring variation about the mean may leave out complexity of potentially great ecological importance. Most studies of biotic response to climate depend on the mean environmental parameters because these are the only data that are widely available. However, all terrestrial biotic response to climate change is mediated by the local microclimate actually experienced by the organism. Some of the ecological complexity that exists on the landscape due to small-scale variation in topographic setting has been studied using dense arrays of micro-loggers; results indicate the potential importance of local microclimate variation for both modelling and empirical studies of terrestrial biotic responses to environmental change (Edwards and Armbruster 1989; Wesser and Armbruster 1991; Lloyd et al 1995; Rae 2003). Large-scale patterns in elevation-generated microclimate variation are determined by the location of mountain ranges. In contrast, patterns in the amount of variation induced by aspect and slope appear to have a latitudinal trend. For reasons explained below, aspect/slopeinduced variation in microclimate appears to peak near the Arctic and Antarctic Circles. Thus, the sensitivities of vegetation and ecological processes to small-scale climate variation may be especially great in the Arctic and Antarctic.