Why bother going outside: the role of observational studies in understanding biodiversity-ecosystem function relationships

The role of observation in the design, execution, and interpretation of BEF relationships

Observation of nature is a fundamental element of ecology and environmental science that provides inspiration, fires our imagination, challenges us to define useful experiments, and helps us select relevant theory. However, there is no single natural scale at which ecological phenomena can be observed and our own perceptual filters influence the patterns we observe (Levin 1992). Furthermore, mechanisms and patterns do not necessarily occur at the same scales. Patterns can be constrained by large-scale environmental forcing—e.g. wave energy can influence the strength of adult/juve- nile interactions in bivalves (Thrush et al. 2000)— or emerge as a result of the collective activities of organisms interacting on fine scales—e.g. the hummock and hollow patterns on some mudflats generated by the influence of microphytobenthos on sediment stability (Weerman et al. 2010). Nevertheless, it is important to try to understand the mechanisms that underlie observable patterns. This is a major theme in the biodiversity- ecosystem functioning (BEF) literature, where ecologists are seeking to clarify the contribution of biodiversity to the delivery of ecosystem goods and services.

Biodiversity loss is a systemic response to stress or disturbance and not simply a random process in

a homogeneous landscape, as is assumed in many models and controlled experiments. In this chapter, we discuss the role of observational studies; we classify observational studies as empirical research that does not involve manipulative experimentation. This can include a broad range of approaches from observations of natural history, behaviour, and bio-physical interactions, to the descriptions of patterns in space and time, to correlative studies and explicit broad-scale hypothesis tests of patterns and relationships. Our intention is to encourage a more integrated empirical approach to BEF research by emphasizing what is gained by field observation compared with controlled experiments in studying these patterns and processes. We seek to emphasize the merits of combined and iterative approaches using both experimental and observational studies. We draw on a long history of experimental marine ecology to demonstrate the value of observational studies as a route to providing context, generality, and mechanisms to scale up our results. Where possible we illustrate with examples from the BEF literature, but we also draw on key species-ecosystem function studies to illustrate the value of more integrative approaches.

Observational studies illustrate how function can be driven by changes in size, density and spatial pattern of organisms. Marine soft-sediment habitats deliver a number of ecosystem functions associated with nutrient and carbon cycling, sediment trapping/transport/stability, habitat

Marine Biodiversity and Ecosystem Functioning. First Edition. Edited by Martin Solan, Rebecca J. Aspden, and David M. Paterson. © Oxford University Press 2012. Published 2012 by Oxford University Press.

formation, provision of settlement sites and refu- gia, productivity, and resilience. We discuss the heterogeneous nature of seafloor landscapes and the potential influences of this variability on BEF relationships. Next, we discuss scaling issues for experiments, in particular the issue of generality in determining the significance of biodiversity- ecosystem function relations across seafloor landscapes. Finally, we emphasize the importance of observational studies in providing a context for model development, verification, and testing.

The very nature of BEF experiments makes them complicated and difficult to perform because they involve measuring interactions between biodiversity and various bio-geo-physico-chemical processes. When designing an experiment that seeks to identify the role of particular combinations of species in affecting an ecosystem function, it does not take more than a few species before the number of possible combinations swamps the most ardent experimentalist (Naeem 2008) . In aquatic studies, this complexity is often reduced using aquarium experiments or mesocosms as 'model systems'. Mesocosm studies simplify nature by using a limited suit of species in a controlled environment—e.g. buckets or plastic containers. Often the 'ecosystem response' is limited to one or a few related functions. These model system studies can be practical, tractable, and rigorously designed, and can be useful for isolating some of the mechanisms underpinning BEF relationships (Benton et al. 2007). However, the process of simplification can hide valuable information about the workings of natural systems and the applicability of mechanisms at varying scales. These implications must be carefully considered when we assess the ecological relevance of these model systems (Petersen et al. 2009).

There are two basic, but strongly overlapping, reasons to improve our understanding of BEF relationships: to understand how ecosystems work, and to apply this knowledge to resource management, conservation, and environmental economics (Stachowicz et al. 2007). When scientific information is applied to environmental issues, it is imperative that we understand its relevance and generality. In fact, Srivastava and Vellend (2005) have questioned the value of justifying the conservation of biodiversity on the basis of maintaining ecosystem functions because of the continued uncertainty over the relevance and generality of BEF relationships. The weight of evidence for biodiversity to be important in defining ecosystem function, at least for some functions and in some locations, is growing (Caliman et al. 2010), but Srivastava and Vellend

(2005) highlight the need for rigorous and integrative science if general principles are to be found. There is the potential for tight coupling between biodiversity and ecosystem function, though these relationships are not necessarily uni-directional (Figure 14.1). BEF studies have tended to emphasize the way some particular element of biodiversity —e.g. genotypic, or phenotypic diversity, species richness, or some measure of functional diversity— influences some measure(s) of ecosystem functional performance as a cause and effect relationship, rather than as a feedback loop. When feedback process are important, designing cause and effect experiments requires careful consideration of the processes involved and the scales over which they interact. In terms of recognizing the value of observational studies, BEF studies have much to learn from past research focused on the role of key species or functional groups in influencing ecosystem functions.

To understand the generality and limits of extrapolation of model systems, and to build a more integrative science, observations from the real world are essential (Table 14.1). For example, many experiments lack the spatial and temporal heterogeneity within replicates that can vary functional responses and allow the complementarity among species to be expressed as increased functional performance. We know from observational studies and experiments that heterogeneity is an important element of the functionality of ecosystems (Legendre and Fortin 1989; Legendre 1993; Seitz et al. 2001; Huffaker, 1958). Yet heterogeneity occurs across a range of space and time scales. The importance of specific scales of variability to individual species will depend on the traits of those species—e.g. size, mobility and resource specificity (Wiens, 1989; Wiens et al. 1995; Hewitt et al. 1996). However, only small scales of spatial heterogeneity have been directly

Figure 14.1 BEF relationships are not simple cause-effect relationships. They should be viewed as feedback loop relationships. This chapter argues for developing a process of iterating between (a) observation of patterns, (b) local measurement and manipulation, and (c) formulation of general relationships. See also Plate 6.

Table 14.1 The merits of mesocosm experiments versus observation versus an integrative approach

manipulated in BEF experiments (see Bulling et al. 2008; Dyson et al. 2007; Jones and Frid 2009). Just as spatial heterogeneity and extent can be an important element in defining function, temporal variability and extent can influence the strength of diversity effects and allow complex interactions to emerge (Stachowicz et al.; 2008, Suttle et al. 2007) . Many processes in ecology are not directly amenable to traditional randomized, replicated, small-scale experiments (Carpenter 1996; Doak et al. 2008). Observation can help us design and interpret our experimental studies and test the generality of our new findings and theories. Integration can be achieved by testing predictions based on theory or small-scale experiments with broad-scale observational studies.