openLandscapes - Aspects of Ecosystem Dynamics

University of Kiel, Ecology Centre, Msc Environmental Science, a seminar paper
Status: completed (2009)

Abstract

Ecosystems have a very complex dynamics. It is the “story” of the enormous number of different life forms and the life conditions. As increasing natural disasters like the global climate change the dynamic research area is more accelerated. Because of the complexity and unique attributes in ecosystem research, various interdisciplinary approaches are ongoing.

Ecological succession is the changing process of biological communities over time, which occurs in new, unoccupied habitat or recently disturbed area. Because succession theories are rooted in plant communities, they are limited to be generalized. Later orientor theories are noticed for describing growth and development processes. But still theories are deficient to describe ecosystem fully. Starting from Holling’s four phase diagram, disturbance and decay processes are regarded as normal components. In this context, the remnants of recently disturbed area are the bases of new development periods.

Key words: ecosystem dynamics, ecological modelling, orientors, succession theory, bifurcation, logistic growth, Holling's diagram, Bénard cell .

Content

1. Introduction

The global climate change is attracting our notice as one of core problems over almost all of the society. It is caused by many factors like greenhouse gases, solar radiation flux, ocean dynamics and so on. Human’s erroneous behaviors utilizing natural resources are the main problem. In the name of economic development, human changed ecosystem and it influenced climate status which can be result in the catastrophe to threat human. Thus we must understand why the climate has been changed, how it will affect our society and what we need to act for minimizing the disasters. The patterns and processes of system between components and their relationship are the objectives of ecosystem dynamics. In this paper we would like to have a look about the basic concepts of ecosystem dynamics.

Ecosystem dynamics tells the “story” of biotic organisms surrounded by abiotic circumstances. Organisms themselves have many different probabilities to survive, to give a birth, to die and to do other activities. Also the changes of habitat conditions over temporal scales (so-called “succession” processes) are impacting organism’s abundance and composition. So for us, it is not easy to define an ecosystem and create theories to describe such a complex system. As the term ‘ecosystem’ is coined, we want to analyse it in comprehensible ways. During the last several decades, many ecologists tried to work out theories from interdisciplinary studies and these are continuously developed, thereby often correcting past errors.

Complex system modeling is one of the interdisciplinary efforts, that are using computational tools under mathematical structures. Although modeling cannot explain ecosystem in a perfect way, it can capture the essential features. Data availability is also a big problem in modeling procedures, however we can simplify the model depending on the quality and quantity of the data. Also continuous development of hardware and software help us to examine ecosystem more clearly. Previsously many ecologists have stressed about the topics of ecosystem growth and development. But it was still deficient to explain ecosystem and from Holling’s adaptive cycle they started to consider the other focal processes, disturbance and decay. Under this view, disturbance and decay are the necessary steps of ecosystems to develop in a sustainable way.

We will start with the concepts of succession theory and simple mathematical models, and later will draw growth and development with the necessary backgrounds such as thermodynamics and self-organization. After introducing Holling’s adaptive cycle, we will complete dynamic concepts with disturbance and decay.

back

2. Succession theory

Ecological succession is the changing process of biological communities over time, which occurs in new, unoccupied habitat or recently disturbed area. In the past, some ecologists thought the successional process has more or less directions which can be predictable like generally fast growing pioneer species will be soon replaced by competitive colonists along a predictable path towards a relatively stable climax. Later this view has been argued that climate factors, spatial heterogeneity of physical factors, and the adaptations of species determine successional processes which are harder to predict (Westman, 1985). The view of the climax concept has been modified by several ecologists during past several decades. Frederic Clements insists the monoclimax theory that ecosystem is a deterministic model which has only one single equilibrium. He showed the example of the convergence of floristic composition and emphasized the stable endpoint through the predictable and orderly change. In 1926 Henry Gleason objected Clements’s theory that the type of vegetation was more accidental, variable and not orderly, taking one example: seed migrated from one place to another by winds, storms and other disturbances (Merchant, 2002). Another critic, Arthur Tansley said the floristic convergence is only partial during the succession and there may be more alternative stable endpoints while Clements defined those as subclimaxes. In 1953 Whittaker developed further that the continuum/gradient of each regional environment could lead vegetation towards different climax patterns (Westman, 1985).

Figure 1. Succession and Forest Change (Michigan State University Extension, http://mff.dsisd.net/Environment/PICS/Succession.jpg).

Ecologists generated various succession typologies by identifying patterns, processes, forces or mechanisms. Two of the most lasting typologies are primary or secondary succession and autogenic or allogenic succession. First, primary succession is a vegetation development on newly formed or exposed substrate that was previously unvegetated contains no seed bank, no organic matter from prior vegetation. On the other hand secondary succession occurs on vegetated substrate before the disturbance destroyed the plant life. There are developed soils, which supported previous vegetation, and a biological legacy like a seed bank and organic matter. Second, autogenic succession is vegetation change due to internal forces and mechanisms, for example competition, shade generation, and soil modification while allogenic succession is driven by external forces like climate change (Glenn-Kewin et al., 1992).

back

3. Mathematical models: introduction

As a result of several efforts to formulate mathematical models in ecology, the general relationship regardig biotic communities can be drawn during successional processes, with the population size [N], the rate of reproduction [r], and the time [t]. Since the existing population gives a birth, simply the net growth rate of a population [dN/dt] is proportional to the existing population size [N] and the rate of reproduction [r]:

Equation (1) assumes no intraspecific competition, for example the growth of pioneer species in a period of earlier succession processes. If the reproduction rate [r] is constant and [r] > 1, the net growth rate [dN/dt] will increase then, the population size [N] will grow exponentially. In reality, however, there is a carrying capacity [K] under the given circumstances like limited food, habitat, and other necessaties available within an ecosystem. By multiplying the intraspecific competition [1-N/K] as a limiting factor to the equation (1), equation (2) is generated. This is known as the logistic equation and a population is increasing until reaching the carrying capacity [K], shaped sigmoidal (Begon et al., 2006). The logistic equation interprets late succession processes.

Figure 2. Exponential and sigmoidal increase (dashed line) in density (N) with time for models of continuous breeding. The equation giving sigmoidal increase is the logistic equation (Begon et al, 2006).

Equation (1) and (2) are assuming that the rate of increase of a population[dN/dt] is affected by the present population size [N] instantly. What if this assumption is simply modifed that the existing population size [Nt] is giving influence to the next period population size [Nt+1]? Thus, due to the ‘time-lag’, the population size in the next period [Nt+1] is the summation of [Nt] and the rate of increase [dN/dt] and following equation (3) is formulated:

Until now, we regarded the reproduction rate [r] as a constant value. In reality, however, the reproduction rate [r] is also continuously changing. As changing the [r] value, we can observe several dynamic patterns, either converging to one stable equilibrium or fluctuating non-equilibrium dynamics. Figure 3 shows 4 representative dynamic patterns: chaos, limit cycle around a carrying capacity [K], damped oscillations, and monotonous increase. The first two types are non-equilibrium dynamics and the later two types have a stable equilibrium(Begon et al., 2006).

Figure 3. Population dynamics variation with changing [r] (Begon et al, 2006).

The various dynamic patterns as the value [r] is changing, can be viewed briefly by a bifurcation diagram. Figure 4 shows that when the [r] value is less than 2 it converges to one stable equilibrium point. But, from the point where the [r] value is 2 it bifurcates into two different stable points, which means the population oscillates stably in a 2-point cycle. As the [r] value is increasing, bifurcations are arising infinitely (Jørgensen, 2000).

Figure 4. The hierarchy of stable fixed points of periods 1, 2, 4, 8... 2n, as the parameter r increases. The y-axis indicates relative values. (Jørgensen, 2000).

Based on these basic models, further mathematical approaches are ongoing for describing the more accurate ecological phenomenon. Recently, stochastic elements are examined for better fitting in site development process. Together with the development of software and hardware, sampled and stored data in a relatively long-term period makes it possible to deal with the quantitative approach. Although perfect embodiment of nature is not possible, quantitative modeling still gives the essential point of view in the research.

back

4. Growth and development

Ecosystem theories have firstly focused on growth and development processes orientating in pioneer stages and moving to conservation stages on the denuded area. In this ecological successional stage, the compositions and structures of biotic communities are changing from r-selected to K-selected species. In fact the ecological succession procedures are the most important functions that determine the respective system status. In the evolutionary theory, the origin of life appeared when lightening stuck a primordial soup. After several reactions, organisms could develop continuously by storing information into DNA and consuming matter and energy (Jørgensen, 2001). That explains why ecosystem is open to material and energy.

During the last years, self-organization and the laws of thermodynamics from network theory, information theory, and cybernetics have been developed to elucidate growth and development processes in a new perspective (Müller et al., 1998). According to the second law of thermodynamics, entropy is always increasing in closed systems, so living systems are going to the death in the end. But the earth is not a closed system since it has a permanent input of energy from solar radiation. Therefore, living systems can develop continuously. This can be interpreted that the system is moving further away from and thermodynamic equilibrium, i.e. non-equilibrium status, for example plants are growing by nutrient uptake through roots from disordered molecules or atoms in soil (Kay, 2000).

Figure 5. Benard cells: hexagonal structures after exceeding a critical point of heating (Mccrone, 2007, http://www.dichotomistic.com/images/Benard%20cell.jpg).

Müller and Nielsen (2000) have illustrated self-organization with the example of the BÉNARD experiment. When a fluid in the pan is heated from below, a heat gradient will be formed and conduction processes among molecule will occur. After reaching a critical point, this unordered conduction will be changed to highly organized convection and hexagonal structures will be formed. If the energy input exceeds a further critical point, the system will be moving into chaos, in other words self-organization is only maintaining in a certain energetic range. So self-organization can be defined as the spontaneous formation of ordered spatio-temporal and functional structures (e.g. convection cells) from microscopic disorder (e.g. non-structured liquids on the pan).

Figure 6. Self-organized creation of convection cells (BÉNARD cells) in a liquid that is heated from below and some thermodynamic features of the structural formation (after Schneider and Kay 1994). The upper part of the scheme describes heat conduction, the corresponding heat gradients and the heat flux for both processes, convection and conduction. The lower part illustrates the conditions while heat convection appears. The right figure shows that the efficiency of the gradient degradation is much higher in the self-organized case (Müller and Nielsen, 2000).

In ecosystem there is a certain direction that we can observe regularly, such as organisms are growing through increasing physical size, or vegetation structures are developing towards the climax status. The term “orientor” has been coined in this context: we can observe the phenomenon that ecosystems are moving towards certain stages. Bossel (1998) defined the term orientor to be “used to denote normative concepts that direct the behavior and the development of systems in general.” Another definition is that orientors are aspects, notions, properties, or dimensions of systems which can be used as criteria to describe and evaluate the system’s developmental stage (Bossel, 1992). Ref. “Orientor wiki”

In the ecological modeling approach, “goal functions” are assumed to measure given properties or tendencies of ecosystems, emerging as a result of self-organization processes in their development (Marques 1998). The direction is indicated by orientors, however, in reality there is no specific, clear goal itself. Therefore the goal function terminology should be limited to modeling work. Bass (1998) added the term “attractors” to describe the stable points in the state space toward which the system is heading. He said that there are several semi-stable points which can be destabilized at different times, but these points can be subsumed by a larger attractor which remains stable. So when we move to a larger scale the system might still be developing into certain direction.

Figure 7. A generalized model of developmental trends of ecological state variables, indicators or orientors. As a consequence of conceptual uncertainty in ecological systems the tendency is directed toward a range of potential states. (Müller and Jørgensen, 2000).

Succession theory firstly attended vegetation dynamics, so the conclusions derived from this theory have limitations to being adapted to the overall ecosystem dynamics while the system-based orientors can include the whole ensemble of organismic and abiotic subsystems and their interrelations (Jørgensen et al. 2007a). According to Bass (1998), orientors assess the development level of the ecosystem and classify different ecosystems.

Still the concept of growth is deficient for describing the phenomenon of ecosystem. Once the environment cranks up, the remnants of the system are reorganized and started development processes again. Holling suggests the four phases including release and reorganization phases which are differentiated from the existed view, and now those two phases are generally accepted as components of ecosystem processes.

back

5. Holling’s Adaptive cycle

Holling’s adaptive cycle model additionally involves disturbance as a normal procedure of ecological development. As Figure 8 shows, Holling’s adaptive cycle describes ecosystem dynamics as a sequence of four stages: the growth or pioneer stage [r], the conservation stage [K], collapse or release [W], and the reorganization stage [a]. The two depicted dimensions are the Potentials which determine possible range options and the Connectedness between internal variables and processes, which is a degree of flexibility or rigidity by external disturbances.

Figure 8. A stylized representation of the four ecosystem functions (r, K, W, a) and the flow of events among them. The arrows show the speed of that flow in the cycle, where short, closely spaced arrows indicate a slowly changing situation and long arrows indicate a rapidly changing situation (Holling and Gunderson, 2002).

The [r] stage is describing exploitation dominated by r-strategists, in which the rapid colonization of recently disturbed areas is emphasized while conservation is characterized by K-strategists, where slow accumulation and storage of energy and material are emphasized (Holling and Gunderson, 2002). The phase from [r] to [K] is called the foreloop. After accumulating biomass and nutrients, the system can become fragile (overconnected), so abiotic or biotic changes such as forest fires, droughts, insect pests, or intense pulses of grazing provoke the system release function [W]. In a reorganization phase [a], the systems reorganize nutrients so that they become available for the next phase of exploitation. The phase from [W] to [a] is called the backloop. These processes occur in uneven time durations. The process moves slowly from [r] to [K], very rapidly to [W], rapidly to [a] and rapidly back to [r] (Holling and Gunderson, 2002).

back

6. Disturbance and decay

Disturbances play a crucial role in ecosystems. When the system is growing the accumulated nutrients and information are only available for existing elements and this can hinder new species to enter the system. At this stage, the mutual connection of the mature system is too strong, so that it can become very sensitive to disturbances if they exceed certain thresholds: flips and bifurcations occur and the system’s trajectories are irreversibly changed (Jørgensen et al., 2007b).

According to the level of disturbances, existing structures and functions of ecosystem can be totally destroyed, losing energy, nutrients and information and the stable equilibriums are moved. Sometimes the proper degree of disturbances, the continuing and moderate level of disruptions, which ecosystem can manage to bear, leads to a more resilient system (Samson and Knopf, 1996). Intermediate disturbance theory which has been the main point of disturbance theory suggests that the disturbances of intermediate frequency, intensity, and amplitude, the resident species and the pioneer species cohabit, which favors a better species richness. In another word, competitive exclusion by the dominant species arises under the low level of disturbance and when the disturbances are too frequent or intensive then only colonizing species can be survived (Wikipedia, 2009). Theses creative destructions are advantageous to replace existed components like energy and nutrients provided by saprophagous organisms sometimes are higher than those provided by phytophagous organisms (Jørgensen, 2007b).

Figure 9. Interrelationship between frequencies and magnitudes of perturbations and disturbances, after White and Jentsch (2001).

The between collapse and rebound from disturbances is depending on the resilience of system and on the strength of the impact, e.g. originating in newly introduced elements. Ref. “Resilience Wiki” Magnitudes, specifities and severities of disturbances are considered regarding the impact. Namely, these disturbances do not totally destroy ecosystem, rather they partially impact ecosystem, so that the remained structure can be reorganized used in the next exploitation phase.

Samson and Knopf(1996) gave an example about fire climax systems. Pine forests of the Yellowstone National Park in its unmanaged state burned over extensive areas relatively often, however this high fire frequency caused lack of fuel wood, therefore enormous fires to destroy forests could not occur. While frequent fires released nutrients accumulated in forest system and supported a new growth, huge fires hugely damaged the forests under managed states and reorganization was not easy. Therefore natural disturbances are setting successional cycles for long term ecosystem resilience and integrity (see Figure 10).

Figure 10. Long-term succession of ecosystems, indicated on different scales: small-scale disturbances may support the development of the overall system. (Jørgensen et al., 2007b)

Long Term Ecological Research is ongoing for the purpose of understanding a diverse array of ecosystems at multiple spatial and temporal scales and creating well-designed and documented long-term observations, experiments and samples for future generations in different aspects. Kiel University has also invested on the beech forest and arable lands of Bornhöved Lake district from the various sides of meteorological, hydrological, biological and pedological area.

back

7. Conclusion

Ecosystems are open to matter, energy and information. Their combinations make ecosystems unique with respect to other systems, so that thermodynamic laws which are generally applied to many other natural or social science subjects should be modified to ecosystem problems. There is a permanent energy input from solar radiation, and it drives the ecosystem to develop from thermodynamic equilibrium. In the initial stages, many researches have focused on growth and development. But starting with Holling’s adaptive cycle, disturbance and decay are finally regarded as necessary factors to sustain ecosystem functions.

Traditionally succession theories are noticed as the only meaningful factors through vegetation dynamics, so they had have a limitation to be generalized. Overcomming the limitation of succession theories, orientor theory suggests the tendency or direction of ecosystem development. From growth to conservation, ecosystems store nutrients and information until reaching the constraints provided by the prevailing conditions. It is also coincident with mathematical approach discussed above. More mature states need more energy to maintain the structure and function of ecosystem and in these mature states, the system can become overconnected and can easily break down after exogeneous and endogeneous disturbances due to a reduced adaptability. Depending on the degree of disturbances, some of them can provide a positive effect on the ecosystem so called, creative destruction.

The Greek philosopher, Heraclitus said that ‘Nothing is permanent but change.’ This is the just exact expression to describe dynamic patterns and processes since the ecosystem dynamics is more or less ambiguous to clearly define. But the environmental issues including global climate change are already too severe to leave in a passive attitude. More strong actions based on objective data are needed to persuade economic-oriented decision making-processes. Dynamic research will give the better understanding of ecosystem by finding certain relationships between factors or several stable attractors to give management directions. The research is developing under interdisciplinary framework and the better objective data also available. Long-term data is being continuously measured and stored, GIS and Satellites are producing higher resolution grids, and applications and the development of modeling are progressing.

back

References

Bass, B., 1988. Goal Functions in the Holling Figure-Eight Model. In: F. Müller and M. Leupelt, Editors, Eco targets, goal functions, and orientors, springer-Verlag, Berlin, pp.194-208.
Begon, M., Townsend, C.R., Harper, J.L., 2006. Intraspecific competition. In: Michael Begon, Colin R. Townsend and John L. Harper, Editors, Ecology: From individuals to ecosystems, Blackwell Publishing, 4th ed., Oxford, pp. 132-162.
Bossel, H., 1992. Modellbildung und simulation–Konzepte, Verfahren und Modelle zum Verhalten dynamicscher Systeme. Viwweg-Verlag, Braunschweig, pp. 400.
Bossel, H., 1998. Emergence of basic orientors. In: F. Müller and M. Leupelt, Editors, Eco targets, goal functions, and orientors, springer-Verlag, Berlin, Heidelberg, pp.19-33.
Glenn-Kewin, D.C., Peet, R.K., Veblen, T.T., 1992. Successional dichotomies and population-based process. In: David C. Glenn-Lewin, R.K. Peet, and Thomas T. Veblen, Editors, Plant succession: thory and prediction, Chapman & Hall, London, pp. 12-17.
Holling, C.S., Gunderson, L., Ludwig, D., 2002a. In Quest of a Theory of Adaptive Change. In: L.H. Gundersson and C.S. Holling, Editors, Panarchy: Understanding Transformations in Human and Natural Systems, Island Press, Washington, DC, pp. 3-24.
Holling, C.S., Gunderson, L., Ludwig, D., 2002b. Holling, C. S., L. Gunderson, and G. Peterson, Sustainability and Panarchies. In: L.H. Gundersson and C.S. Holling, Editors, Panarchy: Understanding Transformations in Human and Natural Systems, Island Press, Washington, DC, pp. 63-102.
Holling, C.S., Gunderson, L., 2002. Resilience and adaptive cycles. In: L.H. Gundersson and C.S. Holling, Editors, Panarchy: Understanding Transformations in Human and Natural Systems, Island Press, Washington, DC, pp. 25-62.
Jørgensen, S.E., 2001. Exergy as limiting factor. In: Sven.E. Jørgensen, Editor, Thermodynamics and Ecological Modeling, CRC Press LLC, Boca Raton, Florida, pp. 300.
Jørgensen, S.E., Müller, F., 2000. Ecosystems as chaotic systems, In: S.E. Joergensen and F. Müller, Editors, Handbook of ecosystem theories and management, CRC Press LLC, Boca Raton, pp. 411-426.
Jørgensen, S.E., Fath, B.D., Bastianoni, S., Marques, J.C., Müller, F., Nielsen, S.N., Patten, B.C., Tiezzi, E., Ulanowicz, R.E., 2007a. Ecosystems have complex dynamics (Growth and development). In: S.E. Jørgensen and B.D. Fath, Editors, A NEW ECOLOGY: SYSTEMS PERSPECTIVE, Elsevier, Amsterdam, pp. 103-142.
Jørgensen, S.E., Fath, B.D., Bastianoni, S., Marques, J.C., Müller, F., Nielsen, S.N., Patten, B.C., Tiezzi, E., Ulanowicz, R.E., 2007b. Ecosystems have complex dynamics – disturbance and decay. In: S.E. Jørgensen and B.D. Fath, Editors, A NEW ECOLOGY: SYSTEMS PERSPECTIVE, Elsevier, Amsterdam, pp. 143-166.
Kay, J., 2000. Ecosystems as Self-organising Holarchic Open Systems: Narratives and the Second Law of Thermodynamics. In: S.E. Joergensen and F. Müller, Editors, Handbook of ecosystem theories and management, CRC Press LLC, Boca Raton, pp. 135-160.
Marques, J.C., Pardal, M.A., Nielson, S.N., Jørgensen, S.E., 1998. Thermodynamic Orientors: Excergy as a Holistic Ecosystem Indicator: A Case Study, In: F. Müller and M. Leupelt, Editors, Eco targets, goal functions, and orientors, Springer-Verlag, Berlin, pp. 87-101.
Mccrone, J., 2007. Fractals< Thermodynamics< Hierarchies, Dichotomistic. [Online] URL: http://www.dichotomistic.com/hierarchies_fractals.html
Merchant, C., 2002. The economic approach to ecology. In: Carolyn Merchant, Editor, The Columbia guide to American environmental history, Columbia university press, New York Chichester, West Sussex, pp. 167-168.
Michigan State University Extension. Succession and forest change. Michigan forests forever teachers guide. [Online] URL: http://mff.dsisd.net/Environment/PICS/Succession.jpg
Müller, F., Jørgensen, S.E., 2000. Ecological Orientors: A path to Environmental Applications of Ecosystem theories. In: S.E. Joergensen and F. Müller, Editors, Handbook of ecosystem theories and management, CRC Press LLC, Boca Raton, pp. 562.
Müller, F., Nielsen, S.N., 2000. Ecosystems as subjects of self-organising processes. In: S.E. Joergensen and F. Müller, Editors, Handbook of ecosystem theories and management, CRC Press LLC, Boca Raton, pp. 177-194.
Müller, F., Leupelt, M., Reiche, E.W., Breckling, B., 1998. Targets, goals and orientors. In: F. Müller and M. Leupelt, Editors, Eco targets, goal functions, and orientors, springer-Verlag, Berlin, Heidelberg, pp.3-10.
Samson, F.B., Knopf, F.L., 1996. Modeling complex ecological economic systems. In: Fred B. Samson and Fritz L. Knopf, Editors, Ecosystem management: selected readings, Springer, New York, pp. 148-163.
Schneider, E.D., Kay, J., 1994. Life as a manifestation of the second law of thermodynamics. In: Mikulecky, D. and Whitten M., Editors, Jour. of Mathematical and Computer Modeling, Volume 19, pp. 25-48.
Westman, W.E., 1985. Succession and resilience of ecosystems. In: Walter E. Westman, Editor, Ecology, impact assessment, and environmental planning, John Wiley & Sons, Inc., New York, pp. 480-488
White, P.S., Jentsch, A., 2001. The Search for Generality in Studies of Disturbance and Ecosystem Dynamics, Progress. In: Botany, Vol. 62, Springer-Verlag, Berlin, Heidelberg.
Wikipedia, 2009. The Free Encyclopedia. Creative Commons Attribution-ShareAlike License. [Online] URL: http://en.wikipedia.org/wiki/Intermediate_disturbance_hypothesis

back

Useful links

Alexei Sharov, Quantitative population ecology, Department of Entology, Virginia Tech, Blacksburg, VA. http://home.comcast.net/~sharov/PopEcol/lec9/chaos.html
Thomas T. Veblen, Traditional succession and climax concepts, University of Colorado, http://www.colorado.edu/geography/courses/geog_4371_f04/handouts/16TraditionalSuccession.html
University of Idaho, College of Natural Resources, Student learning center, Succession, http://www.cnr.uidaho.edu/learn/ecology/lessons/lesson05/5_1.htm
The US Long term ecological research, http://www.lternet.edu/

back