Poiesis & Praxis

, Volume 3, Issue 1, pp 83–105

Analysing biodiversity: the necessity of interdisciplinary trends in the development of ecological theory

Authors

    • UFT, Department of General and Theoretical EcologyUniversity of Bremen
  • Hauke Reuter
    • UFT, Department of General and Theoretical EcologyUniversity of Bremen
Focus

DOI: 10.1007/s10202-004-0072-7

Cite this article as:
Breckling, B. & Reuter, H. Poiesis Prax (2004) 3: 83. doi:10.1007/s10202-004-0072-7
  • 39 Views

Abstract

Technological advancement has an ambivalent character concerning the impact on biodiversity. It accounts for major detrimental environmental impacts and aggravates threads to biodiversity. On the other hand, from an application perspective of environmental science, there are technical advancements, which increase the potential of analysis, detection and monitoring of environmental changes and open a wider spectrum of sustainable use strategies.

The concept of biodiversity emerged in the last two decades as a political issue to protect the structural and functional basis of earthbound life. In this respect, it represents a great challenge for science, in particular for ecology, which is the scientific discipline mainly involved in contributing to understand biodiversity issues.

In this paper, we state a strong necessity for ecologists to work in close connection with other disciplines and within their own discipline across the different organisation levels. Each level has some specific properties to which ecological terminology has been adapted, and joint views are necessary to understand complex networks. In this context, ecological theory provides the background to analyse biological complexity and the relationship of structure and dynamics on different integration levels and provides the interface to mediate social and political issues.

Important features of new technologies for advances in ecological theory refer to (1) an increase in information processing capacities, (2) more efficient automatic data acquisition and device operation, and (3) an increase in resolution (grain and extent).

One crucial consideration we analyse is the trend that aquantitative development in one particular discipline may open a new potential forqualitative advancement in other disciplines when the quantitative advancement is applied in a new disciplinary context.

We illustrate these qualitative developments that are based on technological advancements and which helped to advance ecological theory qualitatively with two examples: (1) The underlying mechanisms causing regularly oscillating rodent populations are subject to a decade long discussion in ecology. Using the possibilities of modern information processing, it is possible to represent the discussed hypotheses in an integrative object oriented model and analyse how the underlying causal net works. (2) The second example originates from biosafety research dealing with the environmental impact of genetically modified organisms (GMO). The project GenEERA develops a complex up-scaling procedure from below field-level information to the landscape scale in order to investigate spread and persistence of GM oil seed rape (Brassica napus) under different scenarios. The approach gives an example, how ecological modelling can be used to combine different information levels to derive conclusions on a higher spatial scale.

In an overall conclusion we relate the described approaches to a wider system analytical context in which we interpret theory developments and biodiversity issues with a system theoretical description of growth processes. We obtain the view that in self-organising systems there is a tendency for autonomous development which tends to be dominating far away from capacity limitations. However, while approaching capacity limitations, a tendency towards closer coupling of internal and external cause-effect networks emerges. We also find that the relation of biodiversity, ecosystem services, and social dynamics can be interpreted in this framework. In this context, the demand for closer interdisciplinary cooperation to solve existing problems appears as an indication of emerging capacity limitations (or the reaching of saturation levels) both, in the theoretical as well as in the (bio-) physical domain.

Zusammenfassung

Technischer Fortschritt ist von zweierlei Belang hinsichtlich Biodiversität: Einerseits ist solcher Fortschritt für beträchtliche Umweltschäden verantwortlich; andererseits gibt es technische Errungenschaften, die hinsichtlich der Anwendung von Umweltwissenschaften das Potential zur Analyse, Entdeckung und Überwachung von Umweltveränderungen vergrößern und ein breiteres Spektrum nachhaltiger Nutzungsstrategien eröffnen.

Der Begriff Biodiversität ist in den letzten beiden Jahrzehnten als politisches Thema in Erscheinung getreten, wo es um den Schutz der strukturellen und funktionellen Basis des erdgebundenen Lebens ging. In dieser Hinsicht stellt Biodiversität eine große Herausforderung an die Wissenschaft dar, insbesondere an die Ökologie, die sich als wissenschaftliche Disziplin hauptsächlich damit befasst, zum Verständnis von Problemen in Zusammenhang mit Biodiversität beizutragen.

In diesem Beitrag weisen wir auf die starke Notwendigkeit hin, dass Ökologen eng mit anderen Disziplinen und innerhalb ihrer eigenen Disziplin quer durch die verschiedenen Organisationsebenen zusammenarbeiten. Jede Ebene hat ihre spezifischen Eigenschaften mit ihrer entsprechend angepassten ökologischen Terminologie, doch um komplexe Netzwerke zu verstehen sind gemeinsame Sichtweisen nötig. In diesem Zusammenhang stellt ökologische Theorie den Hintergrund für die Analyse biologischer Komplexität und der Beziehung zwischen Struktur und Dynamik auf verschiedenen Integrationsebenen, während sie zugleich als Schnittstelle zur Vermittlung zu sozialen und politischen Fragen dient.

Als wichtige Beiträge neuer Techniken zu Fortschritten in ökologischer Theorie sind zu nennen (1) die Zunahme an Informationsverarbeitungskapazitäten, (2) effizientere automatische Datenerfassung und (3) verbesserte Auflösung (Feinheit und Ausdehnung).

Als zentralen Punkt analysieren wir den Trend, dass einequantitative Entwicklung in einer bestimmten Disziplin ein neues Potential fürqualitativen Fortschritt in anderen Disziplinen eröffnen kann, sobald die quantitative Fortentwicklung in einem neuen fachlichen Zusammenhang eingesetzt wird.

Wir illustrieren solche qualitativen Entwicklungen, die auf technischen Fortschritten gründen und zu neuen Entwicklungen in ökologischer Theorie beigetragen haben, anhand zweier Beispiele: (1) Die Mechanismen, die den regelmäßigen Oszillationen von Nagerpopulationen zugrunde liegen, sind seit zehn Jahren Gegenstand ökologischer Diskussionen. Die Nutzung der Möglichkeiten der modernen Datenverarbeitung ermöglicht es nun die diskutierten Hypothesen mithilfe eines integrativen, objektorientierten Modells darzustellen und zu analysieren, wie das zugrunde liegende kausale Netz funktioniert. (2) Das andere Beispiel stammt aus der Biosicherheitsforschung, die sich mit den Umwelteffekten genetisch modifizierter Organismen (GMO) befasst. Das GenEERA-Projekt entwickelt ein komplexes Verfahren zur Hochrechnung von kleinskaligen Daten zu ganzen Landschaften, um damit die Ausbreitung und Beständigkeit von GM-Raps (Brassica napus) unter verschiedenen Szenarien zu untersuchen. Dieser Ansatz kann als Beispiel dienen, wie mithilfe ökologischer Modelle verschiedene Informationsebenen kombiniert werden können, um zu Aufschlüssen zu gelangen, die auch im größeren räumlichen Maßstab Gültigkeit haben.

Resümierend stellen wir die beschriebenen Ansätze in einen breiteren systemanalytischen Kontext, in dem wir theoretische Entwicklungen und Biodiversitätsfragen mithilfe einer systemtheoretischen Beschreibung von Wachstumsprozessen interpretieren. Wir kommen zu der Sicht, dass selbstorganisierende Systeme eine Tendenz zu autonomer Entwicklung zeigen, die fern von Kapazitätsgrenzen dominiert. In der Nähe von Kapazitätsgrenzen zeigt sich dagegen eine Tendenz zu engerer Kopplung zwischen internen und externen Ursache-Wirkung-Netzwerken. Nach unserem Befund lässt sich auch die Beziehung zwischen Biodiversität, Ökosystemdiensten und sozialer Dynamik in diesem Rahmen interpretieren. In diesem Kontext erscheint uns die Forderung nach engerer interdisziplinärer Kooperation zur Lösung bestehender Probleme als ein Zeichen aufkommender Kapazitätsgrenzen (oder des Erreichens von Sättigungsniveaus) sowohl in der Theorie als auch in der (bio-)physikalischen Welt.

Résumé

Le progrès technologique possède un caractère ambivalent pour ce qui est de son impact sur la biodiversité. Il est responsable de préjudices majeurs sur l’environnement et accroît les menaces pesant sur la biodiversité. D’un autre côté, du point de vue des applications des sciences de l’environnement, certains progrès techniques accroissent le potentiel d’analyse, de détection et de contrôle des changements écologiques, et débouchent sur un éventail plus large de stratégies pour un usage fondé sur la durabilité.

Le concept de biodiversité est né au cours des deux dernières décennies comme thématique politique visant à protéger la base structurelle et fonctionnelle de la vie sur terre. À cet égard, il constitue un important défi pour la science, en particulier pour l’écologie, qui est la discipline scientifique la plus impliquée dans la tâche de comprendre les questions de biodiversité.

Dans le présent article, nous affirmons qu’il est absolument nécessaire que les écologistes travaillent en relation étroite avec d’autres disciplines, et dans leur propre discipline à tous les niveaux d’organisation. À chaque niveau se trouvent des caractéristiques spécifiques auxquelles la terminologie écologique a été adaptée, et il faut unir les différentes perspectives pour parvenir à la compréhension des réseaux complexes. Dans ce contexte, la théorie écologique fournit la base à l’analyse de la complexité biologique et de la relation entre structure et éléments dynamiques à différents niveaux d’intégration, et constitue une interface avec les thématiques sociales et politiques.

Les grandes caractéristiques des nouvelles technologies pour faire progresser la théorie écologique concernent (1) davantage de capacités de traitement de l’information, (2) une collecte automatique des données et des systèmes de traitement plus efficaces, et (3) une résolution accrue (grain et extension).

L’un des éléments majeurs de notre analyse est la tendance selon laquelle un développementquantitatif dans une discipline particulière peut dégager un nouveau potentiel pour des progrèsqualitatifs dans d’autres disciplines, quand le développement quantitatif est appliqué au contexte d’une autre disciplinaire.

Nous illustrons par deux exemples ces développements qualitatifs basés sur les progrès technologiques, qui contribuent à faire progresser qualitativement la théorie écologique : (1) les mécanismes sous-jacents provoquant des oscillations régulières dans les populations de rongeurs font l’objet d’un débat écologique depuis dix ans. En recourant aux possibilités du traitement moderne des informations, il est possible de représenter les hypothèses en jeu dans un modèle intégrateur orienté sur l’objet, et d’analyser les réseaux causals sous-jacents. (2) L’autre exemple est issu de la recherche sur la biosécurité et concerne les conséquences écologiques des organismes génétiquement modifiés (OGM). Le projet GenEERA développe une procédure complexe de gradation allant de l’information issue du niveau inférieur sur le terrain à l’échelle d’une aire plus vaste, dans le but d’étudier la diffusion et la persistance du colza oléagineux GM (Brassica napus) selon différents scénarios. L’approche illustre comment utiliser les modèles écologiques pour combiner différents niveaux d’information et en faire dériver des conclusions à une échelle spatiale plus élevée.

Dans une conclusion générale, nous plaçons les approches décrites dans un contexte analytique plus vaste, dans lequel nous interprétons les développements de la théorie et les thèmes de la biodiversité au moyen d’une description théorique systémique des processus de croissance. Nous en arrivons à la conclusion que les systèmes auto-organisés présentent une tendance au développement autonome qui tend à être dominante loin des limites de capacités. À proximité des limites de capacités toutefois, il apparaît une tendance à une association plus étroite entre les réseaux internes et externes de causes et d’effets. Nous avons constaté que la relation entre la biodiversité, les services à l’écosystème et la dynamique sociale peut également être interprétée selon cette grille. Dans ce contexte, la demande d’une coopération interdisciplinaire plus étroite pour résoudre les problèmes existants semble indiquer l’émergence de limites de capacités (ou le fait que des niveaux de saturation sont atteints) tant dans le domaine théorique que dans le domaine (bio)physique.

1 Introduction

The concept of biodiversity originated in the debate of world wide species loss and appropriate measures to stop this process. It constitutes a major challenge for ecology and ecological theory. Biodiversity research aims to understand the different facets of the ‘web of life’ and, among others, attempts to accumulate the scientific basis for management and conservation of natural systems.

In this context, technical developments have an ambivalent character. They influence the way humans organise and operate their interactions with nature. By enabling rationalisation and structural changes in agriculture, medicine, and industrial production, new technologies have direct and indirect effects on ecological processes. Developments and applications of new technologies alter the quality and scale of impact. This involves ecological processes ranging from the molecular up to the global scale.

Human impacts on natural systems and biodiversity are one of the topics theoretical ecology deals with. The complexity of ecological interactions and the number of components to be considered led to controversial discussions of biodiversity related issues (Loreau 2000). In addition, the collection, categorisation and evaluation of biodiversity information represents an enormous challenge for the ecological community as large amounts of data have to be collected and managed. In this context, technical developments, in particular in molecular biology and in information processing, provide important tools for ecological theory to analyse and understand biodiversity related processes.

1.1 Biodiversity as an issue in ecology

The analysis and understanding of processes involved in biodiversity is an issue which goes by definition beyond the listing of different organisms. It implies a variety of different aspects in science as well as in the political and social perception and its motivations to handle environmental issues. The biodiversity concept encompasses three main organisational levels:
  • diversity of genes: research on this level focuses on the analysis and comprehension of the molecular basis of live processes

  • diversity of organisms: its investigation relates to behaviour and life history traits as well evolutionary processes and population dynamics

  • diversity of ecosystems and landscapes: approaches to understand this level comprise, e.g. ecological functions, and species composition.

This organisation is reflected in the definition used in the Convention on Biological Diversity (CBD):

“Biological diversity” means the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems. (CBD 1992, Art. 2)

Biodiversity emerged as a political issue posing the task to protect the structural integrity as well as the functional dimension of the “web of life”. In this context ecology (in conjunction with taxonomy and systematics) provides the scientific basis to analyse and understand the relations which increase or decrease, protect or endanger biodiversity. In an applied context, ecological concepts are used in conservation biology. Besides describing the components of life on different organisational levels, the description of the interaction of these components and how their states and dynamics emerge in a self-organised way, is an elementary part of dealing with biodiversity issues (Noss 1990). Various scientific debates aim at an understanding of different aspects involved, e.g. relating biodiversity with stability aspects (Lehman and Tilman 2000; Ives and Hughes 2002), ecosystem function (Mikola and Setala 1998; Duffy 2002), disturbances (Silver et al. 1996; Mackey and Currie 2000), and an evolutionary context (Signor 1994). The results of these discussions provide the theoretical background to approach more practical issues such as biological invasions, effects of land use practices, and nature conservation as well as specific potential problems such as the environmental impact of genetically modified organisms (GMO).

1.2 Ecology requires work across organisation levels

The scientific details of information gathered in ecology concentrate on specific organisation levels such as the individual organism, the population, the ecosystem, or the landscape, and investigate the context in which the focal unit(s) are embedded (Hölker and Breckling 2002). Each of these organisation levels has its specific properties. Figure 1 shows the sections which make up the structure of ecology. The criteria of how ecology is structured as a science follow the requirements to capture the specific relations to be dealt with and reflect the achievements to structure the methodological access that is available. Ecological approaches and ecological terminology are adapted to deal with the properties relevant on the various levels. Respective properties on a certain level may not be found on higher or lower organisation levels or may play a different role there. Hierarchy theory provides an explanation of how the phenomena on a particular level come about (Allen and Starr 1982; O’Neill et al. 1986). Its application requires: (1) an understanding of bottom up-causation, i.e. how the phenomenon is brought about by the interaction of basic components, and (2) an understanding of top-down control, i.e. how the higher level context imposes constraints.
https://static-content.springer.com/image/art%3A10.1007%2Fs10202-004-0072-7/MediaObjects/10202_2004_Article_72_Fig1_HTML.gif
Fig. 1

Organisation levels in ecology (Hölker & Breckling 2002). Each level has it specific properties which may not be found on other levels. In relation to these properties specific topics are subject of investigations. Focusing on one of these levels does not necessarily provide experience in others.

Hierarchy theory distinguishes between the different levels of organisation, their specific characteristics and fundamental relationships. While there are properties to be found equally relevant on different levels, there are also level-specific properties—otherwise a distinction from higher or lower levels would be obsolete (Hölker and Breckling 2002).

In order to understand ecological relations, it is generally not useful to focus on one particular level, to derive applicable paradigms on that focal level and then simply generalise to other levels. This could easily lead to confusion and misunderstanding. Though this point might be considered as trivial, in practice, it can be quite an obstacle. “Interdisciplinary” communication is required to deal with cross-level problems: terminologies, approaches and paradigms are widely level specific, since levels are defined according to properties which make them in a qualitative way distinguishable from others.

To solve scientific tasks frequently causes the need to communicate beyond the limits of personal experiences, preferences and expertise. Problems in communication and understanding do not only emerge when social implications of ecological changes are discussed and competence from science and humanities needs to be involved. Problems arise already e.g. when eco-physiologists, who have studied the body functions of an organism, have no background to deal with the behavioural properties of that species on the landscape level. Another example is when scientists involved in matter flux studies on the ecosystem level usually are neither able nor interested in determining (and name!) the particular species of organisms whose interactions and life history processes bring about the particular fluxes and contribute to the overall ecosystem steady state. To bring together separate views and methods can frequently spark off new paradigms and is a major source of theoretical progress. For example, ecological modelling started with an adoption of formalisms originally developed to handle physico-chemical dynamics (Lotka 1926). Fractal geometry (Mandelbrot 1982) largely inspired landscape ecology and analysis (Milne 1991, 1992; Gardner and Turner 1991). Object oriented programming languages (Dahl et al. 1968; Krasner 1983) enabled individual-based modelling as a new unifying approach in ecology (Huston et al. 1988; DeAngelis and Gross 1992).

In this paper we want to deal with illustrations of cross-level approaches in ecology, and we will discuss chances and rewards to work across integration levels. We will give an overview of quantitative technical progress in information processing and technical handling of ecological dynamics and their conceptual implications on ecological theory formation. These will be illustrated by two examples, from which we finally derive our conclusions on the interdisciplinary context of biodiversity management implications.

2 Development tendencies in ecological theory driven by technical innovation

Technical innovation does not always brings something entirely and qualitatively new. Often, innovation just extends available capacities. In many ecological contexts, the potential to answer scientific questions is limited by (working) capacity. What happens, if thesequantitative transitions brought about by technical innovation achieved in whatever discipline enter ecological application?Qualitative changes may be induced, as tasks which previously could not be handled may now be accessed. This is what we want to focus on. Quantitative developments in one discipline can have the potential to trigger qualitative changes in other fields. This applies in particular to an interdisciplinary context.

We will illustrate the impact of technical developments for ecology with regard to: (1) developments extending capacities in information processing, (2) developments in automatic data acquisition, and (3) developments in increasing resolution depth. In many cases, the three aspects occur in combination, which increases their power to expand disciplinary potentials.

2.1 Developments extending capacities in information processing

The technical development in information processing is characterised by remarkable extents in processing speed and storage capacities of data. Over the years, computer processor speed and storage capacity have been continuously extended, increasing approximately by a factor of two every 18 month as predicted by Moore (1965). Faster computers and larger storage space are just quantitative extensions. Using these increasing capacities to handle problems in other disciplines, the additional capacity may allow tackling questions which were previously unmanageable. In particular, the study of self-organisation phenomena in ecology has benefited from this extension (e.g. Reuter and Breckling 1994). The representation of dynamics resulting from the interaction of a large number of structurally distinguishable units, which cannot be summarised to homogeneous wholes, especially profits from an extension of available processing capacities. Extending processing capacities in particular allows the investigation of phenomena, which involve numeric complexity. Self-organisation and numeric complexity are inherent not only to many ecological phenomena but are also involved in a large number of concepts in other disciplines. The following examples illustrate the importance of handling complex phenomena in various contexts.

2.1.1 The travelling salesman problem

The ‘travelling salesman problem’ describes the simple task to find the shortest route to visit a larger number of places In mathematical terms: Given a set ofn nodes and distances for each pair of nodes, find a roundtrip of minimal total length visiting each node exactly once. The distance from nodei to nodej is the same as from nodej to nodei. In formal analogy, this type of problem frequently occurs also in other contexts and is not just a mathematical exercise. In the development of mathematical theory the travelling salesman task was fundamental in studying problems, which cannot be solved by a generalisation. Instead a step-by-step attempt is required to compare all possible alternatives of a specific case. The number of alternative paths depends combinatorically on the number of places to be visited and involves very large numbers. In fact, it is still very illustrative to deal with this problem to provide a basic understanding of phenomena that can only be solved in an iterative approximation (http://www.math.princeton.edu/tsp/index.html).

2.1.2 Genetic algorithms

Genetic algorithms provide a strategy to deal with numerically complex phenomena in terms of random procedures, generating potential solutions, which are compared according to their suitability (Holland 1975; Goldberg 1989). Using the obtained solutions to generate proximate new ones turned out to be surprisingly efficient to solve many complex optimisation tasks (among which is also the travelling salesman problem). Interestingly, it has also turned out, that using the principles of mutation, selection and analogies to some cellular genetic mechanisms (crossing over), speeds up the finding of close-to-optimum solutions. This means that mimicking evolutionary processes in an algorithmic form can be used to handle complex adaptation and optimisation tasks (Schwefel 1995). Due to the fact that the basic operations have to be repeated many times, the approach crucially depends on computational capacities.

2.1.3 Cellular automata

The cellular automaton (CA) concept (Wolfram 1984; Toffoli and Margolus 1987) opens new possibilities to deal with spatial heterogeneity and connected dynamics. CAs may be utilised to investigate how global spatial patterns emerge from a multitude of repeated interactions of local units. A CA operates on a grid base. Essential is the definition of the cell states, the cell neighbourhood, and the rules to change the states. The rules define under which conditions (considering the states of a cell and the states of neighbouring cells) transitions between different states occur in a particular cell. In order to obtain successive global states, each cell of the grid has to be updated according to the rules, starting with an initial configuration. Cellular automata are examples of systems, which may have an extremely complex behaviour resulting from a set of very simple rules. In many situations it is not possible to forecast the outcome of an initial state other than going through a step-by-step application of the rules. Application examples in the ecological context relate, e.g. to disease spread (Thulke et al. 1999), heterogeneous habitat structures (Sole and Manrubia 1995; Middelhoff and Breckling 1997), plant competition (Matsinos and Troumbis 2002) and pattern formation (Czárán and Hoekstra 2003; van der Laan 1996).

2.1.4 Individual-based models (IBM)

The domain of IBM is the analysis and explanation of complex population patterns which are important for many biodiversity related issues. The approach is frequently used to link activity and behavioural processes on the level of single organisms with the development and spatial configuration of populations and the dynamics of communities. Individual-based models can be set up starting from a generally applicable basic scheme, which is successively refined to match a particular application case. IBMs describe the interaction of basic units (often but not necessarily the individual organism). These units consist of delimited programme structures containing variables describing the potential states of the unit, rules to change these variables and a structure providing a regular update (often called the life loop) (Breckling et al. 1997; Reuter 2001). During simulation, copies of these units may be created with specific values for the variables, leading to an individual life history of each represented organism despite the common rules. Higher level dynamics (e.g. the population) emerge from the repeated interaction of the basic units. The development of the individual-based modelling approach proceeded in close connection with the emergence of Object Oriented Programming (OOP) in computer science. The approach is very similar to the autonomously acting agents in Artificial Life applications (e.g. Langton 1988). The first attempts to apply this individual approach using OOP-techniques in ecology did not find much attention (Kaiser 1976; Hogeweg 1980; Seitz 1984). As a new paradigm for ecological modelling, the approach arose when object-oriented programming became popular through implementations in the programming languages SMALLTALK (Krasner 1983; Goldberg and Robson 1983) and C++ (Stoustrup 1991). At the same time, Lomnicki (1988) outlined the principal advantages of modelling individuals to extend ecological knowledge and Huston et al. (1988) illustrated these new possibilities with plausible examples from ecological investigations relating to different ecological processes. The breakthrough in the 1990s led to a large number of reviews and applications (e.g. DeAngelis and Gross 1992; Judson 1994; Grimm 1999; Breckling 2002). These cover the range of important ecological processes such as reproduction (Wolff 1994; Reuter and Breckling 1999), dispersal (Crist and Wiens 1995; Jopp et al. 1998), food search behaviour (deVries and Biesmeijer 1998), the emergence of social structure (Hogeweg and Hesper 1991) and also population management strategies (Liu et al. 1995). Individual based models thus extend the potential of ecological models to represent self-organisation resulting from interactions on and between several levels such as physiological processes, individual behaviour, social interaction and population dynamics. They present a tool to cope with spatial heterogeneity and complex ecological interaction networks with variable structures.

All these techniques, which are of considerable importance for ecology (and thus in a wider sense also for the understanding of biodiversity phenomena) have largely and qualitatively benefited from quantitative extensions to handle large, structured data sets.

2.2 Developments in automatic data acquisition and device operation

In information processing, the capacities and means of data acquisition and automatic processing are extended. This allows collection and handling of data quantities, which cannot be analysed by manual processing. In particular, long range and large-scale environmental analyses benefit from this development. Automatic data loggers provide data of a quantity (and quality), which is not achievable “by hand”. The investigation of climate-dependent processes was largely advanced by device automation. Biodiversity research also profits from automatisation. With respect to the biodiversity census, Wilson (2003) proposes a global species inventory on the basis of modern information technologies. In this context, Edwards and Morse (1995) summarise the state-of-the-art in computer aided species identification, while Weeks and Gaston (1997) emphasise the potential of automatic image analysis to identify species. Taxonomy and systematics currently undergo a basic methodological transformation process by involving electronic approaches. The Global Biodiversity Information Facility www.gbif.org is a major organiser of this process, see the organisations strategic plan (GBIF 2003).

Remote sensing represents another field which often combines large scale data with automatic processing. For example, Corine Landcover, a data basis for habitat types covering the whole EU with a resolution of 25 ha, has been developed by analysing satellite images (Aubrecht 1998).

In combination, the mentioned developments largely extend the investigation window. In particular, the progress which has been made in the last decades is impressive. While, for example, genetic analyses were extremely time-consuming and thus expensive, automation has reduced handling time and cost to a factor of 1/100 or 1/1000.

2.3 Developments to increase resolution (grain and extent)

Quantitative technical advancement brings about the possibility of increased resolution. For ecological applications, this can mean acquiring records on larger areas, longer times, and more fine-scale details. Environmental analysis and documentation using remote sensing has largely benefited from this development (Turner et al. 2003). This often overrides previous limitations, which forced looking at single cases and taking them as “representative” without being able to qualify the representativeness. The gain in capacity for handling large collectives such as automatic data logging, tracking of individuals with telemetric devices allows an increase of parallelism which is substantial as ecological dynamics often have high individual variability.

Besides applications on the global, biome and landscape level, these changes involve also the molecular level. They include, for example, automatic genetic sequencing and synchronous analyses of thousands of proteins (proteomics) or metabolites (metabolomics ). Advancements on these levels implicitly allow tackling questions on higher levels, which were inaccessible so far. These especially concern population structures, evolutionary dynamics and history. Examples from population ecology which were inaccessible or only accessible with large effort include kin relationship and spatial population structure (Billot et al. 2003) and the relationship between social and genetic structure in social Hymnoptera (Oberstadt and Heinze 2003).

For an application perspective, we can state that quantitative extents in data acquisition, data handling, automation and higher resolution in grain and extent improve detection and monitoring of environmental changes. This development comes together with parallel increases in environmental impact strength and extent of environmental interference. While the potential to generate environmental impacts is the more dynamic factor, the application of the detection and control potential often follows. While a general surveillance technology may be available, its implementation and application depends on the perceived necessity of monitoring or management, which implies a preceding risk or impact. Control measures depend on the social norm-setting process, i.e. the development of the regulatory adaptations to the new potential.The Starlink scandal in the US shows the dilemma. The approval of genetically modified organisms generates new monitoring requirements. In that case, a transgenic maize variety in 1998 approved only for animal feed and not for human consumption was found to contaminate many processed foods. Molecular monitoring unveiled the deficits in food testing and regulation efficiency (White 2001).

In the following two chapters we exemplify, how technical developments were applied to gain new insight in ecological processes as well as advancing the formation of ecological theory.

3 An example from modelling: community diversity of small rodents

Regularly oscillating small mammals populations are an example of how a long discussed issue in community ecology can be tackled with a new modelling approach relying on extended computing power. Community interactions of small rodents have fascinated ecologists for many years due to their abrupt changes in population numbers and their impact on the local biocenosis. From an economical point of view they are interesting because of the immense damages, which population outbreaks can cause for agricultural production. Most of all, the regularly oscillating rodent populations in Scandinavia gave rise to a long lasting controversy in ecology (e.g. Chitty 1960; Rosenzweig and Abramsky 1980; Stenseth 1999; Oli 2003). The causes for these population cycles have been subject of discussions among theoretically and empirically working ecologists for many decades. As empirical work is spatially and temporally restricted, computer models have been developed in order to analyse large scale effects resulting from complex interactions in variable cause-effect networks. These approaches usually utilised conventional equation based models, which operate on the population level (e.g. Hanski and Korpimaeki 1995; Turchin and Hanski 2001). For this analytical approach, computational power and data handling capacities were not limiting. However, these analytical models operate on a single integration level and thus are not able to depict and analyse effects for the self-organisation processes which result from trans-level interactions. Furthermore, they could only include a limited set of hypotheses. For example, Hanski and Korpimaeki (1995) do not consider spatial interaction and feed-back processes between the rodents and their resources. Employing the individual-based modelling approach, however, the potential is different. The models generate enormous quantities of data when tracking the states and actions of a large number of individuals. The handling of these large data quantities makes the approach dependent on high speed data processing and large storage capacities.

The model example described small rodent communities as a set of autonomously acting agents with a detailed life history and behavioural repertoire in a food web composed of three trophic levels—rodent food, rodents and predators (Reuter 2001). The modelled rodents acted in a simulation environment containing the spatial arrangement of habitats under seasonally changing conditions. It thus extended previous modelling attempts in integrating the most relevant factors known to account for the dynamics of rodents.

The model was applied to regularly oscillating rodent populations in Scandinavia in order to analyse the contribution of the involved factors for the resulting dynamics. The scenario set-up included two rodent species (Microtus agrestis, Clethrionomys glareolus) and two predators with different ecological functions, the weaselMustela nivalis as a specialised predator and the owlAsio otus as a migratory bird.

A decisive question in the debate on rodent cycles concerns the sudden decline in population numbers of rodents in spring and summer after a peak year. The model gave new insights into the processes driving these abrupt population crashes. Both lack of resources as well as predation pressure contributed to about 90% of mortality, but no pattern could be found when relating either cause with the properties of the respective cycle. For example, no correlation of either of these factors was found with the number of predators or rodents, or the time lag between the respective cycle phase. The trophic control (the difference between mortality rates due to either of these factors) varied unpredictably and chaotically in the model (Fig. 2). Empirical investigation limited to short time intervals or long term studies taking samples with large intervals have not been able to identify an irregular change in the contribution of the involved factors. They searched for a kind of stable ranking of the investigated factors according to their relevance.
https://static-content.springer.com/image/art%3A10.1007%2Fs10202-004-0072-7/MediaObjects/10202_2004_Article_72_Fig2.jpg
Fig. 2

Trophic control in the simulation model of oscillating rodent populations. Control varied chaotically between longer phases of top down and bottom up limitation. This kind of variation explains contradictory empirical results from temporally restricted investigations, which base on comparatively shorter observation periods (Reuter 2001).

The model results shed an important light on the controversy of the driving forces of the cycles, in particular on the empirical results which have either identified predators, or food shortage as the main impact. As the empirical investigation is limited to comparatively narrower extents of spatial resolution and observation time, most of the data cover only specific situations. So it is no wonder that generalisations from these data yielded contradictory results, since the observation range appears to be too narrow to unveil the entire dynamic. From the model we concluded that top down and bottom up effects were both involved, and that the involvement varied in an unpredictable way.

4 An example from environmental regulation: assessing potential effects of GMO-commercialisation

A second example illustrates the investigation of the environmental impact of genetically modified (GM) plants. The production of these plants became feasible as a result of biotechnological advancement. Oilseed rape, Brassica napus, (Fig. 3) is one of the species to which genetic modifications were applied. Commercialisation in Europe is currently in a process of notification: the developers seek admission from the competent authorities and have to provide data on plant characteristics and risk analysis. In parallel, biosafety research is employed to analyse potential environmental impact, which may result from an introduction of this GM crop into European environments.
https://static-content.springer.com/image/art%3A10.1007%2Fs10202-004-0072-7/MediaObjects/10202_2004_Article_72_Fig3.jpg
Fig. 3

Flowering oilseed rape (Brassica napus) growing wild on a field margin

Setting up experiments on a larger scale than the scale of a single farm is rather difficult. Therefore, we cannot sufficiently anticipate the cumulated effects on biodiversity and ecosystem processes using direct empirical tests. Large scale implications can be relevant. The motivation to reduce uncertainty as far as possible is prompted by the fact that GM plants are self-replicable units, which are very hard to retrieve once established in the wild. A reproducible experimental approach is not feasible at the landscape scale. In order to close this gap, the GenEERA project (Breckling et al. 2003) integrated lower level information to scale up the variability of local process to the landscape level in a modelling approach. This included an extrapolation of the potential interaction of GM oilseed rape(B. napus) with conventional oilseed crops as well as with feral populations and other potential hybridisation partners (Table 1).
Table 1

Integration levels linked in biosafety research to estimate the fate of transgenes in the wild—the example of oilseed rape. To answer the questions at a particular level, the processes on the levels below need to be understood. Biosafety research proceeded from the molecular level towards landscape processes. Since experiments on the landscape level are most difficult to set-up, the knowledge on that level (most relevant for commercialisation) is the least well-investigated

Organisation level

Question

Finding during Biosafety research

Molecular level

Is the gene expression of transgenes in varieties under notification reasonably stable?

Yes—to the level of agricultural practicability

Is the genetic constitution in varieties under notification satisfyingly stable?

Genetic instabilities occur but are rare events

 

Does the molecular stability hold for variable environmental conditions?

Surprisingly, it was found that a switch off of the transgenes (“gene silencing”) frequently occurs after particular plant virus infections

 

Organismic level

Is the morphology of the plants normal?

Yes—comparable to other cultivated varieties

Are the plants reasonably productive?

Yes—comparable to other cultivated varieties

 

Population level

Can the plants grow beyond intended cultivation?

Yes, due to harvest losses causing volunteer growth

 

Yes, due to transport losses along roadsides causing feral growth

 

Can the plants persist in the wild?

Most populations disappear after a short while, but occasional reproductive success combined with the potential of long seed dormancy (>10 years) will allow existence as intermittent populations

 

Are there indications of long-term persistence?

In vegetation surveys it was found that the distribution pattern of feral oilseed conforms to the one of non-cultivated wild relatives

 

Ecosystem and landscape level

Is outcrossing to other wild relatives likely to occur?

Potentially yes—but detailed investigations are widely lacking

To what extent will unwanted interference of transgene crops and conventional crops occur (field-to-field contamination)?

Detailed investigations are widely lacking, mainly qualitative estimations of pollen transfer rates between fields are available, indicating that cross-pollination will occur

 

Will oilseed rape or any hybridisation product cause any ecologically detrimental effect?

This is widely unknown

 
The applied approach combined information on different integration levels, from the molecular and eco-physiological level, up to geographical information and remote sensing data at the landscape level (Fig. 4). The aim was to estimate the potential distribution dynamics for larger regions. The investigation area of the project was Northern Germany. First results concerned regional phenological implications, cultivation density by remote sensing and feral population density estimations based on vegetation surveys (Breckling et al. 2003; Middelhoff and Breckling 2003). It can be expected that, in case of commercialisation, cross-pollination between neighbouring conventional and transgenic fields will take place. The results indicate that persistence time of transgenic varieties as volunteers on the farmland and outside cultivation areas is very long so that a return to conventional agricultural practice (i.e. exit scenarios) will be very cost intensive.
https://static-content.springer.com/image/art%3A10.1007%2Fs10202-004-0072-7/MediaObjects/10202_2004_Article_72_Fig4.jpg
Fig. 4

Landsat 7 ETM satellite image from the area around Bremen (Northern Germany). The image shows oilseed rape fields highlighted by a buffer of 250 m. Information on this scale was used in the GenEERA project to identify regions most susceptible for dispersal and for outcrossing. Additionally, this spatial information was important for scenario development. Image processing was done by Hendrik Laue.

The approach illustrates how ecological modelling can be used to combine different information levels to derive conclusions on a higher scale. Continuation of the analysis will allow prediction of region-specific risk levels depending on the combination of cultivation density, climatic factors, pollen and seed dispersal and land-use structure.

5 Implications in a globalising context: the role of theory in transient dynamics

As we have demonstrated for relevant developments in ecology, quantitative technical changes emerging in other disciplines can provide the tools to allow tackling of questions that were qualitatively unmanageable before. Since it is not quite uncommon that one scientific discipline uses methods developed in other fields of science, we have to consider that the qualitative boundaries of what is answerable in a particular discipline may be influenced by quantitative developments in others. Ecological modelling is a very important basis for theory development in ecology. We have seen that advancements in handling model complexity received major stimuli by quantitative developments in information processing. This encourages the hypothesis that quantitative-qualitative relationships could also be relevant in other comparable fields of science. However, it remains a topic in itself to discuss the extent and contribution of this type of process to dynamics of theory in science (and humanities).

Though this issue cannot be exhausted in this paper, we would like to complement the question with an argumentation pattern derived from the analysis of abstract growth and saturation formalisms applicable to ecological as well as social issues. The formalisation of growth patterns is a part of systems analysis (Imboden and Koch 2003). Systems analysis focuses on concepts with discipline transgressing applicability. Growth and saturation dynamics are some of the issues covered. The resulting formalisms are widely used for the description of biological and ecological processes but not limited to this field of science. The principle of shifting limitations (Breckling 1990) was derived as a network implication in limited growth processes. In the following, we investigate its implications for theory dynamics with regard to biodiversity and biodiversity management in an interdisciplinary context.

6 The principle of shifting limitations

To analyse the processes of growth and saturation, and the underlying dynamics, is one of the basic issues in science. The dynamic theory as a fundament of classical systems analysis (e.g. Forrester 1972; Patten and Odum 1981) provides a particular interpretation, distinguishingvariables which are subject to change and thedriving forces behind. Complex circular interaction networks can give rise to self-organising and self-reproducing entities, constitutinginternal driving forces causing growth as a result of interaction (Prigogine 1980). To understand the dynamics it is of interest to capture the interplay of driving forces and limiting factors which reduce their efficiency.

Biological growth processes, e.g. of a population in a given environment, can serve as an illustration. A driving factor is what facilitates and triggers increase, a limiting factor is what prevents the speed of growth accelerating to infinity. In the underlying interaction networks we usually find various factors involved which will have a direct or indirect impact on growth. Provided optimal external conditions (resources), growth only depends on the internal self-organisation potential. Among the internal factors there is also at least one that is limiting the speed of growth (if not, the speed would tend to become infinite at least if the produced entities continue to perform the same growth pattern—this causes exponential increase). Since self-organisation depends on external supply of matter and energy, there is an interaction of internal and external factors. Frequently we find situations in which external limiting factors are effective. Growth dynamics which are determined predominantly by internal factors on the other side exist as a transient phenomenon only. If each new entity continues to grow at a constant speed, an exponential increase results. Because of the exponential self-expansion due to unlimited self-reproduction, the required external supply becomes limiting in long run. When the growth process approaches such an overall capacity limitation, the speed of growth necessarily changes and successively decreases. The minimum-factor principle, postulated by Liebig (1840) states that under (externally) limited conditions growth can be increased by supplying the limiting factor.

In a classical system dynamic description of such a process, the growth parameters are specified as constants. In the absence of an external capacity limitation, growth processes are controlled by internal factors (such as the maximum birth rate per individual). Capacity limitations operate in the form of external constraints. They gain importance for the dynamic process as growth proceeds. The simplest form of modelling this process is to assume only one constant to represent the internal growth capacity in which models frequently named “r” for reproduction and one constant to represent external limitations symbolised by “K” for capacity. Models of this type were first proposed by Verhulst (1838, logistic growth). Growth brings the particular variable closer towards the capacity. Thus, as the process proceeds, the capacity gains increasing relevance for the ongoing dynamics, until it retards growth completely as the capacity is reached.

This means that during the development towards a capacity limit the control of the ongoing dynamics partially transits from a more internal to a more external control. This type of transition implies a decrease of systems autonomy and thus a decrease of lower-level self-regulation potential: What was regulated in the beginning (far away from capacity limitation) by internal processes, will become regulated by an interplay of internal and external factors.

If we extend the consideration to systems which involve an interplay of a larger number of factors, we find that the situation is, in principle, comparable. Quantitative changes, e.g. population size, bring about differences in the contribution of these factors to growth control. Those ruling the internal dynamics will prevail far from saturation levels. When a saturation level or a capacity limit is approached, external control “overrules” the internal.

Now we relate the abstract growth considerations of system dynamics to ecosphere-anthroposphere relations and its scientific analyses.

As a thesis to stimulate further discussion we would like to hypothesise that in the process of quantitative extension of scientific means (faster computers, better data loggers, higher resolving satellites, etc.) theendogeneous developmental dynamics of a discipline may be most pronounced in an early stage of the extension process. According to the views of Kuhn (1962), the establishment of a new paradigm creates new space which is successively explored and “filled”. Later on in a maturing stage of an established paradigm, external impact and interrelations with other disciplines gain importance. Looking back to the development of ecological modelling we find an exemplification. After an import of the physical dynamic paradigm in ecology using differential equations (Pearl 1924; Lotka 1926), an “autonomous” exploration in ecology successively followed and found its culmination in the International Biological Programme (e.g. Newbould 1967; Ricker 1968). Later, the methodology was employed to an increasing extent also in interdisciplinary studies. The Club of Rome (Meadows et al. 1972) integrating economic and ecological dynamics in the “world model”, gave a first paradigmatic example in which, however, the variables were so highly aggregated that it was hardly possible to relate some of the model specification parameter to anything empirically tangible. To date, e.g. in the context of the analysis of global warming, well parameterised relations of economic processes, energy consumption, CO2 emission and atmospheric processes are under development. As new modelling paradigms, going beyond the classical dynamics fuzzy sets, fractals, cellular automata and object oriented models were introduced to ecology, opening new and extended capacities.

The thesis of theory dynamics is that a new horizon emerging is at first explored due to its immanent, discipline-internal implications and then integrated in networking structures. Applied to science as a whole it would mean that the more science matures, the more vital will be the role of interactions and interdisciplinary connections as an impetus of disciplinary development. This would also mean that mature science will be a highly interdisciplinary and coupled science; but it will not only be the coupling of different scientific disciplines which should get closer in a process of growth and maturation.

6.1 The decrease of autonomy and the need for coupling

Let us finally interpret what the abstract model considerations allow concerning the relations of biodiversity dynamics, the implications for social development and the interaction of society and its natural resource basis, in particular the so-called “ecosystem services” (Daily et al. 1997). These services largely depend on biodiversity.

In an initial phase of a successive growth process, the social system (as a self-organising interaction network) can extend the acquisition of natural goods. Society depends on natural processes but can develop in a comparably autonomous way following internal dynamics as long as natural goods are not substantially limiting. While the human induced matter turnover in the biosphere is much smaller than the turnover in the natural biogeochemical cycles, both develop relatively independent. Man’s impact on nature is within the buffer capacity. Coming close to the order of magnitude of the natural transition cycles, the types of limitation change, both “partners” tend to lose autonomy. The impact of society alters the states of natural systems in a structuring way and imposes altered boundary conditions. On the other side, social expansion (as far as its dependence on ecosystem services is concerned) faces previously irrelevant but now meaningful limitations, i.e. the state of each of the systems tends to depend closely on the state of the other and thus, the systems tend to react more as a coupled whole. Alterations in one part expand the potential to also affect the other. This can have apparent management implications. To maintain desirable states of the coupled systems now requires higher intensity of regulation, a higher complexity, a higher synchrony and better flexibility of organising adaptation. It requires mature science, in which social and scientific cognition work closely together.

Technical extension helps a lot to meet these requirements in a world where anthropogenic matter cycles approximate the order of magnitude of natural ones. According to Vitousek et al. (1986), more than 40% of the terrestrial primary productivity of the biosphere are acquired, appropriated or directed by human intervention. Not only does the complexity of the tasks increase but also the requirement to synchronously regulate processes on diverse levels of integration. This calls for a high increase of the potential of interdisciplinary mediation, communicative abilities, personal experience in different fields and efficient management and decision structures. We can easily interpret the call for interdisciplinarity, which is frequently heard from political institutions and from funding bodies as a necessary response to the requirements of closer coupling and the trend of autonomy loss in the partial systems.

Biodiversity is currently decreasing with a rate roughly 100 to 1000 times larger than the average natural rates valid for the time before human activity became relevant (Lawton and May 1995). Even the mass extinction events of geological time were transitions usually taking longer time spans for manifestation than what we witness today as a result of human intervention. This insight is the driving force behind pushing this topic onto the political agenda. More precisely, it is the insight that this rate of loss may result in the loss of significant economic benefits or even of essential ecosystem services. As we see the currently limited success to decrease the rate in biodiversity losses, we have to admit that the technical extension to monitor and the so-far achieved priorisation for biodiversity protection are still not on a level to assure long-term sustainability. An important means of adaptation is the organisation of cross-regulation of the dynamic potential in science, humanities, society and nature.

Copyright information

© Springer-Verlag 2004