1 Research programmes in ecology: implications for scientific production and framework for biodiversity management policies analysis F. Burel, J. Baudry, A. Ricroch, N. Allsop 1- Introduction Lakatos developed a framework based on the idea that science within a research programme progresses by continuous replacement of one theory by another (Lakatos, 1980). Repeated observations using the positive heuristic drive the process. A research programme is defined by a hard core of knowledge protected by ad hoc hypotheses and a positive heuristic. The hard core of knowledge cannot be refuted by the methods used in the programme. Successive corroborations of theory increase the protective belt of the “hard core” of knowledge. The positive heuristic is the long term research policy that anticipates possible refutations and develops theory to harden the “hard core” of knowledge. It generates a sequence of investigations stimulated by theory. The term ecology, proposed in 1866 by the German biologist Haeckel, denotes the study of relations of living things with their environment. Etymologically, it combines the Greek words oïkos and logos and signifies the science of the habitat. Ecological science is a complex subject because it concerns the study of all forms of life over the expanse of time that life has existed on earth, and all the environmental relationships in which life is present. Ecology relies on physics, chemistry, geology, molecular biology and the biology of organisms to explain certain phenomena. It is mainly concerned with organisms, groups of organisms, interaction between organisms and between their environments, and complexes made of the physical, chemical and biological components called ecosystems. Ecology aims at discovering pattern and process in natural networks. Networks or systems comprise organic and inorganic elements in dynamic relationships. Ecology is a discipline of vast scope. It ranges from interest in how organisms affect the chemistry of the entire earth, to how a particular trait of an organism adapts to its local environment (Keller and Golley, 2000). Ecology has evolved a rich diversity of active subdisciplines, such as autoecological (ecology of one individual), population, community, ecosystem, landscape and global ecology (Burel and Baudry, 2003). Ecology concentrates on the relationships between organisms and their environments. It focuses on patterns of distribution, patterns of abundance, factors that determine the range of environments that organisms occupy and that determine how abundant organisms are within those ranges and functional interactions between co-occurring organisms. Two main research lines have emerged and develop, the population and the ecosystem (Pickett et al., 2007). The primary distinction between these research lines is their focus on organisms on one hand and on materials and energy flow on the other. The contrasting foci are reflected in the definitions given in ecological textbooks that represent these two paradigms. Ecosystem ecology: the study of structure and function of nature (Odum, 1971); Population ecology: the study of the interactions that determine the distribution and abundance of organisms (Krebs, 2001) or the study of the natural environment, particularly the interrelationships between organisms and their surroundings (Ricklefs and Miller, 2000). Since goals differ between the two approaches, as well as the scale of investigation and methods, there is a gap between them, and there is a need for integration. 2 Despite the large variety of subdisciplines and different research lines which characterize research in ecology one may wonder if there is or not some unity in addressing questions and setting observations or experiments or if this diversity leads to a wide array of approaches, concepts and research programmes. By reviewing the main trends in the recent history of ecology, based on data from South Africa, Brazil and France we identified two main research programmes which used to be, and are still in opposition, and that we named “autonomy” and “independence”. After defining them we will attempt to give a clear overview of their hard core and ad hoc hypotheses. We will conclude with a critical analysis of this classification looking at their use in environmental policy making. Considering environmental policies, it is relevant to know how scientific information on which they are based has been produced, and if this process is consistent with the framework policy makers are working with. In Europe most of the policies oriented toward the maintenance of biodiversity in agricultural areas impose an obligation of means by regulating agricultural practices but with no required evaluation of biodiversity improvement. This suppose that the effects of the practices are known (i.e. increase a population of plants, birds or an ecosystem service) considering the variability of environmental conditions. Even if scientific information is available, the model of mechanisms relating causes and effects may vary greatly according to the research programme as we will see. 2- Autonomy and dependence two dominant research programmes in ecology Two opposite positions are found in research programmes in ecology. In the first, oldest and long dominant position, ecological units, as defined by the scientist for his/her investigations, are autonomous and can be understood with no consideration of their surrounding either in space or during ecological succession. The main attribute of these units is that they evolve toward some sort of equilibrium. In the opposing standpoint ecological entities have a behaviour strongly dependent on their surrounding. They do not tend toward any kind of equilibrium. (Pickett et al., 2007) write that the first programme has become outdated as since the early 1990’s because evidence demonstrates that the “equilibrium” state does not exist in most cases. Nevertheless, we assume that it is still, explicitly or implicitly, a position that is frequent if not dominant in many areas, but primarily in nature management and biodiversity conservation. After presenting the two programs, we will see that practitioners (policy makers and land managers) often navigate between the two positions, which may be a source of trouble. Contrasting the two research programs: Autonomous: the properties of and changes in an ecological entity (population/ community, ecosystem) depend solely of its own characteristics. Each entity tends toward equilibrium, the most common view for vegetation is that it follows a deterministic successional pathwayto a climax stete where it remains stable (Clements, 1936). Systems in equilibrium are essentially closed, self regulating, possess a stable equilibrium point or stable equilibrium cycle , have deterministic dynamics, are virtually free of disturbance and independent of human influence . As the importance of disturbances in vegetation dynamics became obvious, an ad hoc hypothesis was proposed: the climax as a shifting mosaic (Bormann and Likens, 1979a, Bormann and Likens, 1979b). . 3 The concept is difficult to test empirically, but can emerge from simulations under certain assumptions (Turner et al., 2003). Biodiversity plot experiment as those conducted by Tilman (Tilman et al., 1997, Tilman, 2000) in Minnesota are part of this research program. The aim of this study was to test if a greater number of plant species leads to greater community productivity. In the experiment, 245 plots, each 9 m x 9 m, were assigned randomly to have from 1 to 16 prairie plant species, with the species composition of each plot being separately chosen at random. They were distributed regularly in the same large field and considered by the authors as independent plots in which only the number and nature of the species play a role in determining plant productivity. They showed that species composition and plant diversity were both strong determinants of ecosystem functioning. This is an example of method where all experimental units are considered autonomous whatever the context (proximity of other pots, history and heterogeneity of the field etc.). Dependent: the properties of and changes in an ecological entity (population/ community, ecosystem) depend of both its own characteristics and the “landscape context” or hierarchy it is embedded. It has been recognized in many systems, for example in wildlife biology or landscape ecology that the fluxes across boundaries are important and these start to be recognized as something real in succession ecology for instance. Successional pathways are considered to be site and situation dependent. The lack of rigid determination is encapsulated in the term “contingency” which means that the current state of an ecological or evolutionary system is dependent on the specific conditions that have occurred from time to time in the past or on the order of events that affected the trajectory. For example the abundance of parasitoids of Meligethes aeneus in oil seed rape fields has been shown to increase with the proportion of non-crop areas surrounding crops. Percentage of herbivory decreases on the same landscape gradient, while total parasitism increases (Thies et al., 2003). This case shows clearly that at this scale of investigation composition and nature of insect community as well as the processes of herbivory and parasitism depend on the nature of the surrounding landscape (Figure 1). Fig. 1: Relations between rates of herbivory, total parasitism and specific parasitism of Meligathes aeneus and proportion of non crop areas in the landscape context (Thies et al. 2003). Main concepts related to the two programmes (the core) 4 Autonomous: is characterized by equilibrium, stability and existence of optimal values. In plant ecology the climax (Clements, 1936) has been the dominant model for plant succession and is still used by scientists and especially by conservationist and landscape planners. Climax is defined as a biological community of plants and animals which has reached a steady state. This equilibrium occurs because the climax community is composed of species best adapted to average conditions in that area. It supports the idea of a single climatic climax, which is defined in relation to regional climate. Another important concept is that of homogeneity. All studied systems from experimental plots to agricultural fields or natural ecosystems are viewed as homogeneous entities with no boundary effect. Dependent: is characterized by its dynamics, the existence of trade-off, non-linearity of many processes determining thresholds; discontinuities allow one to identify hierarchical levels; there is no « optimal value » and the context plays a major role. This program derives from Gleason’s (Gleason, 1926) view of the “individualistic concept of vegetation. It has been summarized: “A change must occur within the environment, creating the opportunity for new species to utilize the change (i.e. light gaps in rain forests, flood plain flooded). What grows there next (plant succession) depends upon what individuals have been dormant, waiting for specific conditions to occur. So, the way a specific area develops is random, dependent upon many variables and conditions in regards to the individual species present at the time. It is generally unpredictable”. (www.goshen.edu/bio/Musci/HAGleason.ppt). Heterogenity and complexity are recognized explicitly. For example in Yellowstone National Park, USA, discontinuities in the vegetation cover depend on different disturbances operating at their own scales (Romme, 1982, Turner et al., 2003). Pinus contorta var. latifolia forest cover 80% of the Park and there are patches of Picea engelmanii. At high altitude Pseudotsugu menziesii is present and at low altitude Populus tremuloïdes and Pinus flexilis are present and willows develop in wet areas. This mosaic changes through time in response to fire. Romme (1982) studied a 73 km² watershed and could reconstruct fire history over a period of 350 years from charcoal sequences. A series of fires differing in size and frequency, small disturbances being more frequent than large ones, and fire history influencing the severity of fires, create the shifting mosaic and its heterogeneity. This ecological system depends on fire disturbance, there is no equilibrium state but a dynamic of disturbance which sustains the diversity of plants and animals found in the Park. Ecological systems studied Autonomous: research is done on finite and discrete objects which are part of the diversity and hierarchy of living organism e.g. plants, animals, micro-organisms, and individuals, populations, communities. They may also be interactions among organisms such as predator/prey, host/parasite, competition, facilitation etc.; or ecological processes (ecosystems services) such as pollination, production and decomposition. Dependent: studied ecological systems have an explicit spatial dimension and there is an infinity of ecological systems as scale varies. There may be small plots of vegetation within a grassland or deciduous forest. They are characterized by their size, their boundaries and their ecological functioning. Those systems are organized within an ecological hierarchy. Theory predicts that levels of hierarchy form discontinuities in space and time. At each level, 5 processes are studied at different scales to identify ecological response scales (Allen and Hoekstra, 1992). For example communities of carabid beetles depend at low levels of the hierarchy at the interaction between individual movements and the structure of their environment. At intermediate ones they depend on population dynamics at a larger time scale. Biogeographical constraints define the potential pool of species at a higher regional and long term level of organization (Figure 2). The development of landscape ecology is linked to the possibilities it offers to demonstrate scale dependencies in ecological patterns and processes. Biogeographical extent space Population dynamics Activity, movement time Fig. 2: Scale dependencies of communities of carabid beetles (Millan de la Pena et al. 2003) At a low level of organization of one individual hedgerow in an agricultural landscape, abundance of carabids depends on the farmer’s management practices over years. Herbaceous layer of hedgerows were traditionally mowed, but as working power of farms decreased, current practices are either abandonment, use of herbicide, fire or grazing. If brambles and ferns are allowed to grow with no cutting, or herbicide spraying is the dominant practice over several years abundance of large, specialist species is far lower than if grazing is the dominant practice (Aviron, 2003). At a higher level, the landscape scale, abundance of large carabid beetles decreases with the length of hedgerows and connectivity of the hedgerow network (Millan de la Pena et al., 2003). Ad hoc hypotheses Autonomous: Most experiments which refer to this research programme are carried out at small spatial and temporal scales, the first ad hoc hypothesis we identified is that in this research programme processes are not scale dependent. The methods used are based on experiments, replicates and data are analysed with statistics, and another ad hoc hypothesis is that those data do not depend on the initial state of the studied system. Human activities are considered as external factors. If for example in an agricultural experiment a plot is disturbed by unexpected farming practices (mowing or trampling of a successional plot) data are removed from the set to be analysed. All events that are not part of the experiment protocol, but are part of the variability of farming activities, are considered as external disturbances. The ad hoc hypothesis is there that experiment excludes human activities, if they use it as a parameters their complexity if highly reduced by using surrogates. 6 For exmple grazing is often mimicked by mowing but this does not encounter for trampling, or animal dejection. Dependent: Systems studied cover a large range of scales, from several fields to a region in space, they include long term monitoring (decades) in time and long term processes (regional or landscape dispersal). To account for complexity most of the time an ad hoc hypothesis is that density of organisms is not considered. Methods are based on observation and experiments for lower hierarchical levels, and models and scenarios to test hypotheses and predict behaviour at higher levels. The ad hoc hypothesis is that the model and scenarios do not take into account the dynamics of environmental conditions. Research is characterized by pluri- and inter-disciplinary approaches and the role of human activities on complex systems is explicitly recognized and studied. But to deal with complexity studies consider and rank a limited number of parameters and interactions. The current state of the debate Why is research conducted within the autonomous research programmes still important and dominant in many ecological journals? Several reasons can be advanced experimental approach both in population ecology (eco-physiology, evolutionary ecology) and ecosystem ecology use of micro-, mesocosms and ecotrons few data and methods exist to integrate human activities in ecological research the Popper’s view of science (hypothesis testing) is still dominant; it requires replicates, difficult to set up in complex systems. This goes along with the dominance of inferential statistics. The advances in two questions can illustrate the importance of changing perspectives: 1) what are the driving factors of biodiversity in field margins? And 2) How can one assess the risk of spread of pollen from GM plants? In both cases, the increase of spatial scale of investigation has led to new insights; this supports the “dependence” program. For field margins, the consideration of the landscape context and then the farming system context show that these important variables that explain differences in species composition (Le Coeur et al., 2002). The time scale is also important as time lags are frequent (Petit and Burel, 1998). For pollen from GM plants, (Lavigne et al., 2008) demonstrate, that not only the distance between GMO and non-GMO fields matters, the total area of GMO crop drives the background pollen composition. In both cases, the policy implications are obvious: the management of field margin biodiversity cannot be based solely on margin scale practices and the potential for crosspollination increases cannot be controlled only by regulating distance between GMO and nonGMO fields. Scientific results and policy making Clements (Clements, 1935) already advocated the use of ecology as a basis to manage natural resources and this is till a subject of research (Liu and Taylor, 2002). We must make a distinction between the way that ecologists propose that their science be used to designe policies and the way policy makers refer to ecology. 7 The idea of the balance of nature (Egerton, 1973) has influenced and continue to influence what and how science is used by managers and planners (Hull et al., 2002 ). The idea that biodiversity is manageable in discrete, autonomous, integrated entities is appealing and still very widespread. So, many international treaties and conventions stressed the international importance of various entities (Washington-Convention., 1973, Bonn-convention, 1979); Birds Directive, 1979; (Bern-Convention); Habitat Directive 1992), This is the “habitat” approach of the EU Natura 2000 policy. As the same time, the concepts of “corridor” and “biological continuity” have been put forward in land planning (Hobbs, 1992, Johnson et al., 1992). One can see some contradiction between these two approaches combining stable mapable units and a view of landscapes as places for changes and movement. The result is that corridors are often represented by structural units, rather than functional ones. The Natura 2000 policy is an interesting case as it is recent and strongly influences nature conservation in Europe. It is based on the Habitat Directive (COUNCIL DIRECTIVE 92/43/EEC) for which the following definitions are adopted: “(a) conservation means a series of measures required to maintain or restore the natural habitats and the populations of species of wild fauna and flora at a favourable status (…); (b) natural habitats means terrestrial or aquatic areas distinguished by geographic, abiotic and biotic features, whether entirely natural or semi-natural;… (f) habitat of a species means an environment defined by specific abiotic and biotic factors, in which the species lives at any stage of its biological cycle; (j) site means a geographically defined area whose extent is clearly delineated;” The sites of Natura 2000 network are “sites hosting the natural habitat types listed in Annex I and habitats of the species listed in Annex II” The reference to “autonomous” units is clear. A habitat is a holistic closed unit, which the use of phytosociological units in the list of “habitats” reinforces. The sites must have a management plan that is a “patch scale” plan. Changes in the spatial distribution of the species of interest cannot be taken into account by a procedure of adaptation management. In contrast, the corridor initiatives in conservation, for example the Greater Cederberg Biodiversity Corridor in South Africa (Barodien, 2005), goes beyond the concept of conserving a representative sample of all species and the habitat in which they occur. This is viewed as providing only for biodiversity pattern as it currently exists. Corridors are supposed to accommodate spatial and temporal changes in biodiversity processes. This suggests that an ecological research programme based on the precepts of dependence will provide better evidence for management strategies than an autonomous one. Conclusion Ecologists proposing to support management and policy, and those wishing to use ecological evidence need to be aware of the spatial and temporal constraints of investigations framed within an autonomous research programme. There is a need for integration between experiments conducted within the autonomous framework, with a short range of spatiotemporal scales. There is also a need to couple empirical studies with theoretical models in the proper investigation of the consequences of spatial heterogeneity and various scales of time. . 8 References Bern-Convention Convention on the Conservation of European Wildlife and Natural Habitats Gleason, H. A. (1926) The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club, 53, 7-26. Clements, F. E. (1935) Experimental Ecology in the Public Service. Ecology, 16, V. Clements, F. E. (1936) Nature and structure of the climax. The Journal of Ecology, 24, 252284. Odum, E. P. (1971) Fundamentals of ecology, Philadelphia, W.B. Saunders Co. Egerton, F. N. (1973) Changing concepts of the balance of nature. Quaterly review of biology 48, 322-350. Washington-Convention. (1973) Convention on International Trade in Endangered Species CITES Bonn-convention (1979) Convention on the Conservation of Migratory Species of Wild Animals Bormann, F. H. & Likens, G. E. (1979a) Patterns and Process in a Forested Ecosystem. , New York, Springer-Verlag. Bormann, F. H. & Likens, G. E. (1979b) Catastrophic disturbance and the steady state in northern hardwood forests. . Amer. Scientist., -. Lakatos, I. (1980) The methodology of scientific research programmes, Cambridge, Cambridge University Press. Romme, W. H. (1982) Fire and landscape diversity in Subalpine Forest of Yellowstone National Park. Ecology Monograph, 52, 199-221. Allen, T. H. & Hoekstra, T. W. (1992) Toward a unified ecology, New-York, Columbia University Press. Hobbs, R. J. (1992) The role of corridors in conservation: solution or bandwagon? TREE,, 7, 389-392. Johnson, A. R., Wiens, J. A. & Milne, B. T. (1992) Animal movements and population dynamics in heterogeneous landscapes. Landscape Ecology, 63-75. Tilman, D., Knops, J., Wedin, D., Reich, P., Ritchie, M. E. & Siemann, E. (1997) The influence of functional. diversity and composition on ecosystem processes. 277, 1300–1302 (1997). Science, 1300-1302. Petit, S. & Burel, F. (1998) Effects of landscape dynamics on the metapopulation of a ground beetle (Coleoptera, carabidae) in a hedgerow network. Agriculture, Ecosystem and Environment, 69, 243-252. Keller, D. R. & Golley, F. B. (2000) The philosophy of ecology: from science to synthesis, Athens, Georgia, USA, University of Georgia Press. Ricklefs, R. E. & Miller, G. L. (2000) Ecology, fourth edition, W.H. Freeman. Tilman, D. (2000) Causes, consequences and ethics of biodiversity. . Nature, 405, 208-211. Krebs, C. J. (2001) Ecology: the experimental analysis of distribution and abundance, 5th edition., Chicago, University of Chicago Press. Le Coeur, D., Baudry, J., Burel, F. & Thenail, C. (2002) Why and how we should study field boundaries biodiversity in an agrarian landscape context. Agriculture, Ecosystems & Environment, 89, 23-40. Liu, J. & Taylor, W. W. (Eds.) (2002) Integrating landscape ecology into natural resource management, Cambridge ; New York, Cambridge University Press. Hull, R. B., Robertson, D. P., Richert, D., Seekamp, E. & Buhyoff, G. J. (2002 ) Assumptions about Ecological Scale and Nature Knowing Best Hiding in Environmental Decisions. Conservation Ecology, 6, 12. 9 Aviron, S. (2003) Effets des activités agricoles et des éléments permanents sur la biodiversité : l’exemple des carabes et des papillons rhopalocères. . université de Rennes 1. Burel, F. & Baudry, J. (2003) Landscape ecology : concepts, methods, and applications, Enfield, N.H., Science Publishers. Millan de la Pena, N., Butet, A., Delettre, Y., Morant, P. & Burel, F. (2003) Landscape context and carabid beetles (Coleoptera: Carabidae) communities of hedgerows in western France. Agriculture, Ecosystems & Environment, 94, 59-72. Thies, C., Steffan Dewenter, I. & Tscharntke, T. (2003) Effects of landscape context on herbivory and parasitism at different spatial scales. Oikos, 101, 18-25. Turner, M. G., Romme, W. H. & Tinker, D. B. (2003) Surprises and lessons from the 1988 Yellowstone fires. Frontiers in Ecology and the Environment, 1, 351-358. Pickett, M., Kolasa, J. & Clive, G. J. (2007) Ecological understanding: the nature of theory and the theory of nature. Second edition, Burlington, MA, Academic Press. Lavigne, C., Klein, E. K., Mari, J.-F., Le Ber, F., Adamczyk, K., Monod, H. & Angevin, F. (2008) How do genetically modified (GM) crops contribute to background levels of GM pollen in an agricultural landscape? Journal of Applied Ecology, 45, 1104-1113. Barodien, G. ( 2005) Greater Cederberg Biodiversity Corridor Biodiversity Spatial Plan Technical Report. Le Cap, CapeNature.