Scaling terrestrial biogeochemical processes contrasting intact and model experimental systems
Planet Earth is undergoing enormous change, clearly discernable on time scales of decades to centuries. This is largely a result of human activities, especially the emission of greenhouse gases and pollutants, the clearing of global forests, the urbanization of agricultural cropland, and many other extensive modifications in the land surface. Significantly, these rates of change are unprecedented in the Earth's geological history (Committee on Global Change 1988), and one of the great scientific challenges of the 21st century is to forecast future behaviors of global ecosystems under the constant pressure of human insults (Clark et al. 2001). This requires that we better understand feedbacks and interactions of the major patterns and processes of the key components of planet Earth: the atmosphere, oceans, freshwater, rocks, soils, and biosphere. In an attempt to meet this challenge, an interdisciplinary approach to studying systems dynamics on a planetary-scale has emerged, known as Earth System Science (ESS) (Schellnhuber 1999, Lawton 2001). Biogeochemical processes of terrestrial ecosystems are at the core of ESS research (Schellnhuber 1999). Hence, considerable effort has been invested towards understanding the relative importance of biotic and abiotic regulators and controllers. Of special interest are how natural and human-induced perturbations may affect the rates and directions of biogeochemical processes in terrestrial ecosystems, especially in terms of potential feedbacks to climate systems (Walker and Steffen 1996, Pielke 2002). Given the complexity of global systems and their many interconnections, one of the main scientific challenges of ESS is to document change, diagnose underlying causes, and develop plausible projections of how natural variability and human actions may affect global biogeochemical cycles in the future. With regard to the latter, once we have the requisite quantitative understanding of process rates, as well as a detailed understanding of key regulatory mechanisms, the goal is to extrapolate findings obtained at one temporal and spatial scale to another. Typically extrapolations for terrestrial ecosystems are done using mathematical models (see Goudriaan et al. 1999, Prinn et al. 1999, Wu and Li, Chapters 1 and 2). However, in this chapter we focus on extrapolations via empirical experimentation: we discuss experimental designs that inform about process rates and regulatory factors at spatial and temporal scales greater than the one on which the experiment is conducted. Paradoxically, while most experiments are in fact intended to further understanding and knowledge at scales beyond the ones at which they are actually being conducted (e.g., 1 m2 plots or a forested watershed), most fail to incorporate the spatial and temporal scale considerations necessary to justify such an extrapolation (Gardner et al. 2001). A number of issues are germane to this discussion, including the specific characteristics of the factors under investigation, the importance of nonlinear responses, the type of treatment imposed (e.g., step vs. gradual), and whether the goal is spatial or temporal extrapolation. We discuss experiments conducted using two general types of systems: intact systems and model systems. Our objective is to compare and contrast these approaches in the context of their potential for contributing to our predictive understanding of process rates and their regulators in terrestrial biogeochemical cycles. In this chapter we will show that: (1) intact ecosystem experiments can provide process rates, mechanistic understanding and absolute/relative treatment effects suitable for direct extrapolation, but rarely do; and (2) model ecosystem experiments can provide the sign (positive or negative) of treatment effects and insights into their mechanistic basis. However, data obtained on process rates and absolute/relative treatment effects are not suitable for extrapolation. We concur with Gardner et al. (2001) that there is a need for much greater "scale awareness" in ecology, especially with regard to the role of experimental design and execution. Our primary objective is to raise awareness of the importance of spatial and temporal scale considerations in the design and interpretation of experiments, so that findings at the scale of an experimental plot and duration may be extrapolated with known confidence. © 2006 Springer.
