Disturbance patterns in a socio-ecological system at multiple scales

Abstract

Ecological systems with hierarchical organization and non-equilibrium dynamics require multiple-scale analyses to comprehend how a system is structured and to formulate hypotheses about regulatory mechanisms. Characteristic scales in real landscapes are determined by, or at least reflect, the spatial patterns and scales of constraining human interactions with the biophysical environment. If the patterns or scales of human actions change, then the constraints change, and the structure and dynamics of the entire socio-ecological system (SES) can change accordingly. Understanding biodiversity in a SES requires understanding how the actions of humans as a keystone species shape the environment across a range of scales. We address this problem by investigating the spatial patterns of human disturbances at multiple scales in a SES in southern Italy. We describe an operational framework to identify multi-scale profiles of short-term anthropogenic disturbances using a moving window algorithm to measure the amount and configuration of disturbance as detected by satellite imagery. Prevailing land uses were found to contribute in different ways to the disturbance gradient at multiple scales, as land uses resulted from other types of biophysical and social controls shaping the region. The resulting profiles were then interpreted with respect to defining critical support regions and scale-dependent models for the assessment and management of disturbances, and for indicating system fragility and resilience of socio-ecological systems in the region. The results suggest support regions and scale intervals where past disturbance has been most likely and clumped – i.e. where fragility is highest and resilience is lowest. We discuss the potential for planning and managing landscape disturbances with a predictable effect on ecological processes.

Introduction

Complexity arises inexorably when we generate descriptions or explanations of ecosystems that simultaneously consider multiple levels of organization or domains of scale (Allen and Starr, 1982). The complexity of ecological systems comprises inherent system properties like the multiplicity of spatial patterns and ecological processes, nonlinear interactions among components, heterogeneity in space and time, and hierarchical organization, but it also depends on the perceptions, interests, and capabilities of the observer (Wu, 1999). Simon (1962) noted that complexity frequently takes the form of a hierarchy, whereby a complex system consists of interrelated subsystems that are in turn composed of their own subsystems. A hierarchy of ecological system levels can emerge during energy dissipation at different focal scales (O’Neill et al., 1986), and any ecological system, at a particular focal scale, appears constrained by the dynamics of larger scale systems (Allen and Starr, 1982). Each level is a domain of scale that can be visualized as a logical subsystem in a simulation model, and it is a region of scale-space where interactions among components have characteristic length and time scales.

The study of scaling is a way to simplify ecological complexity in order to understand the physical and biological mechanisms that regulate biodiversity (Brown et al., 2002). The very concept of biodiversity is inherently multiple-scale, with no preferred scale, and this (tautologically) demands multiple-scale study of the relevant features of the ecosystem. More pertinent is the observation that complex adaptive ecosystems such as the biosphere are self-organizing structures and patterns of interactions that arise from three simple rules (Levin, 1998): (1) sustained diversity and individuality of components; (2) localized interactions among those components, and; (3) an autonomous (self-contained) process that selects among components a subset for replication or enhancement. From these rules system properties emerge such as hierarchical organization and non-equilibrium dynamics (Levin, 1998) that require multiple-scale analysis, not only to comprehend how a system is structured but also to formulate hypotheses about mechanisms regulating the system (Milne, 1998).

The primary role of humans in shaping the environment implies that interpreting the sustainability of biodiversity in socio-ecological systems (SESs) in terms of resilience, adaptability, and transformability (Walker et al., 2004), must be related in some way to the environments that humans create. The terminology of sustainability implies an attractor or at least a basin of attraction (Walker et al., 2004), without which the concept of sustainability is irrelevant. Basins of attraction plausibly represent domains of scale (true attractors) within which interactions among components occur at characteristic length and time scales. According to the pattern – process hypothesis (e.g., Wu and Hobbs, 2002) these characteristic scales in real landscapes are determined by, or at least reflect, the spatial patterns and scales of human interactions with the environment. If the pattern or scale of human actions changes then the environment consequently changes, and the structure and dynamics of the SES can also change accordingly (Gunderson and Holling, 2002). Each SES is a complex system, and no SES can be understood by examining only one component, either social or natural, at one scale (Gunderson and Holling, 2002, Wu and Hobbs, 2002). Consequently, we cannot appropriately deal with a system property like “resilience” (Holling, 1973) unless we consider the entire SES. Displayed or retrospective resilience, as derived from the detection of past changes in the structure of the landscape, is fundamental to define the prospective resilience of a SES since the historical profile reveals a great deal about current system dynamics and how the system might respond to future external shocks (cf. Walker et al., 2002).

Ecological patterns and processes and human activities have interacted in SESs for a long time and are not just “coupled at a single scale.” The human component is increasingly dominating in space and time (O’Neill and Kahn, 2000), thus defining limiting constraints at “higher scales” and altering the natural functioning of ecological processes in absence of human influences. Anthropogenic activities, such as agriculture, industry and urbanization, have radically transformed natural landscapes everywhere around the world, inevitably exerting profound effects on the structure and function of ecosystems (Millenium Ecosystem Assessment, 2003). The dynamic spatial configurations resulting from human appropriation and management of regional landscapes can have a variety of ecological effects within SESs over a wide range of spatial scales. A direct effect is the alteration of ecological processes at local scales through the modification of land cover. For example, converting forest to agriculture land cover alters soil biophysical and chemical properties and associated animal and microbial communities, and agricultural practices such as crop rotation alter the frequency of these disturbances. The spatial configurations of land cover in a region also affect ecological patterns and processes. New land cover types can be juxtaposed and shifted within increasingly fragmented remnant native land cover types, and changes in the structure of the landscape can have disturbing effects on nutrient transport and transformation (Peterjohn and Correll, 1984, Hobbs, 1993), species persistence and biodiversity (Aaviksoo, 1993, Fahrig and Merriam, 1994, Dale et al., 1994, With and Crist, 1995), and invasive species (Fox and Fox, 1986, With, 2004).

Though the term disturbance has been referred to natural causes (Pickett and White, 1985, Romme et al., 1998), here disturbances are relatively small and frequent changes (effects) in the structure of the landscape, as detected by remote sensing and mainly due to human activities (causes), that can reveal a great deal about how humans are affecting ecological patterns and processes.

In this paper, we address the problem of characterizing the spatial patterns of human-driven disturbances at multiple scales in a SES in southern Italy. This is an important first step towards understanding, in the context of complexity theory, how the actions of humans as a keystone species (O’Neill and Kahn, 2000) shape the environment and thereby biodiversity across a range of scales in this region.

The rapid progress made in generating synoptic multi-scale views and explanations of the earth's surface provide an outstanding potential to observe temporal changes in land use pattern as well as the scales of land use pattern (Simmons et al., 1992). Land use changes are signalled by a disturbance of the original land use. Thus, one way to appreciate the interactions between land use patterns and processes is to look at temporal changes in disturbance detected by remote sensing, and how they are associated with different land uses or regions of interest at multiple scales. If disturbance patterns and processes are modified in space and in time, then the adaptability of SESs (Walker et al., 2002), that is the capacity of humans to manage resilience, would change accordingly. That could allow fostering, intentionally, the adaptability of SESs.

Zurlini et al., 2004, Zurlini et al., 2006 suggested that different resilience levels in watersheds are intertwined with different scale domains according to the type and intensity of natural and human disturbances in those watersheds. In this work, we explore the multi-scale patterns of disturbances in an administrative unit (Apulia region) in relation to its land use composition. We hypothesize that differences in multi-scale patterns of disturbance are associated with land uses, because land uses result from other types of biophysical and social controls shaping the regions of interest. We describe an operational framework to identify multi-scale profiles of short-term anthropogenic disturbance patterns using a moving window algorithm to measure the amount and configuration of disturbance as detected by satellite imagery. The resulting profiles are then interpreted with respect to defining critical support regions and scale-dependent models for the assessment and management of disturbances. Results are also discussed in relation to their potential for indicating system fragility and resilience of socio-ecological systems.