Scope 47- Long-term Ecological Research, An International Perspective

15.1 INTRODUCTION

Ecological research is the investigation of processes and patterns that explain the relationships between living organisms and their environment. Such research involves a multitude of approaches, including laboratory experiments, field observations and tests, theoretical explorations, and the evaluation of conceptual and mathematical models. Explicit ecological research of various kinds has been conducted since before the beginning of this century, and a large and rich literature now exists.

Because of the complexity of the relationships between living organisms and their environment, definitive experiments are rarely possible. Consider, for example, the magnitude of the variation in weather on a daily basis, and then the seasonal, annual, and decadal variations, and finally the long-term changes in climate. Then consider the enormous variation in the genotypes and phenotypes of organisms, and the broad array of their responses to weather and climate conditions. These combinations, superimposed upon the interactions among organisms, lead to exceedingly dynamic ecological systems that make definitive ecological experiments difficult to design and conduct. As a result, specific laws or principles are relatively infrequent in ecological research. Moreover, 'certainties' in ecology are usually defined in the specific conditions under which the expected observations occur and/or are defined in terms of the probability under which the observations will take place.

Many fundamental processes in ecological systems occur over relatively long periods of time. For example, the generation times of some insects are a matter of days, but some species of trees live for hundreds of years. Similarly, soil erosion may occur rapidly with one, possibly unpredictable, storm, while soil-building processes may require centuries. Thus, ecological processes, consisting of both organisms and their environment, are driven by dynamics with periodicities and durations spanning many time and space scales.

The complexity of ecological systems and the various temporal and spatial dimensions have led to the recognition that ecological research must be conducted on long time scales (here long-term ecological research is defined as decades to longer). The argument is that long-term study is required not only to understand ecological complexity, but also to ask important or meaningful questions (Likens, 1983, 1989). Moreover, it is argued that long-term studies are valuable because they encompass unusual or interesting events that allow understanding that would not otherwise be possible. Careful analysis of the published literature, however, suggests that these unusual events have not been a commonly used tool in ecological research (Weatherhead, 1986). Previous evaluations of ecological research have noted that classes of ecological phenomena requiring long-term studies include slow processes, rare events or episodic phenomena, processes with high variability, subtle processes, and complex phenomena (Likens, 1983, 1989; Strayer et al., 1986), and there are examples of each class (Franklin, 1989; Pickett, 1991).

Despite the recognized values that have been derived from previous long-term ecological studies, it is important to consider whether there are alternative approaches for investigating ecological phenomena. That is, a significant portion of our current ecological understanding has arisen because of planned or serendipitous measurements taken over long periods of time. The question arises as to whether there are other methods for arriving at the same understanding.

The chapters in this volume present numerous examples of long-term ecological research that have been conducted throughout the world. In these instances, the authors have described the resulting information, and have usually attributed the understanding to the long-term measurements. For the purpose of this summary discussion, we have selected examples from several general topics. These examples are then examined to see whether there were alternative approaches that would have produced the same information.

15.2 EXAMPLES OF LONG-TERM ECOLOGICAL  RESEARCH

15.2.1 ENVIRONMENTAL QUALITY

Experiments begun in the 1840s to 1860s at Rothamsted have permitted the detection of several long-term changes in agricultural systems (Johnston, 1991). In the last 40 years, there have been large increases in the amounts of cadmium and polynuclear aromatic hydrocarbons in the soil. The experimental treatments allowed differentiation of the sources of these substances (for example, of cadmium from aerial sources, from farmyard manure, and from phosphate fertilizer). These studies indicated that increasing soil organic matter that accompanied manure additions may have contributed to the retention of cadmium in the soil. Also, cadmium did not accumulate in the seeds, but did accumulate on or within the herbage where it could be ingested by herbivores.

The polynuclear aromatic hydrocarbon (PAH) burden in the soil increased five-fold in the last hundred years. Although there are several proposed mechanisms for the loss of PAHs (microbial breakdown, photo-oxidation, vaporization, crop removal and leaching), only very small proportions of the annual input were lost, and the PAHs were shown to have long residence times in the soil (Johnston, 1991 ). Moreover, compounds with complex structures increased in the soil more than those with simpler structures, suggesting that microbial decomposition and soil retention of PAHs may depend upon molecular structure.

In another set of experiments at Rothamsted, nitrogen concentrations were monitored for the past 100 years in the soils of manured and unmanured fields growing wheat, and in fields that were allowed to revert to woodlands. The manured fields demonstrated an increased soil nitrogen content which equilibrated only after approximately 80 years. The soils of the woodlands now have about the same nitrogen content as the manured wheat fields, although the equilibrium point has not yet been reached. It is possible that once the equilibrium point is reached, these forest soils will also lose nitrogen due to leaching, as do the manured wheat fields.

In all three examples, long-term measurements were necessary for detecting trends in the materials of interest. It is difficult to design experiments that would have predicted the specific accumulation rates of cadmium, polynuclear aromatic hydrocarbons, and nitrogen. Cadmium accumulation rates also apparently depend upon organic matter increases, PAH concentrations depend upon several climate-controlled processes and also the chemical structure of the compounds themselves, and even after measuring nitrogen for decades it is impossible to predict the equilibrium concentration in the woodlands. Of the three examples, cadmium may now be the most amenable to prediction since it appears to depend upon input rates and organic matter content. In addition, it should be possible to calculate ingestion rates by herbivores.

15.2.2 SOIL PROCESSES

Threshold-controlled responses have not been considered common in soil dynamics, although Anderson (1991) describes several examples of this phenomenon. In a rare but large rainstorm in Alaska, formerly stable horizons in a Spodosol soil were disrupted, and normally insoluble organo-metallic bodies were dissolved. These materials moved down the profile in association with suspended particulates to form unusually deep B horizons enriched in sesquioxides and humus (Stoner and Ugolini, 1988).

There is a common perception that soil processes occur at relatively slow rates when compared to other parts of ecological systems. This perception may arise because soils are mapped according to standard descriptive characteristics and the subsequent soil classifications are regarded as the result of gradual development that appears stable for long periods of time. This classification scheme has led to the acceptance of soils as a rather stable component of the ecosystem. However, Anderson ( 1991) has shown in Canada that there are dynamic components to the soil and that these dynamic components respond rapidly and may show threshold responses. Specifically, soil salinity at the soil surface changes rapidly depending not only upon agricultural practices, but upon the movement of soil water, i.e. downward flux caused by leaching and upward flux by capillary rise or artesian pressure. The surface soil water dynamics are driven by precipitation and evaporation-processes that themselves change rapidly through the season and are subject to changes in global climate patterns. On the other hand, the total salt content of the entire soil profile changes much more slowly and is controlled by larger-scale changes in regional groundwater flows, often through deep aquifers. Organic matter dynamics in soils include processes of different time scales ranging from rapid changes associated with microbial processes to slow ones controlled by physical sorption to clay or chemically-stabilized humus (Anderson, 1979). Among the most instructive studies of soil organic matter are the long-term crop rotation studies at Rothamsted and elsewhere. As an example of the multiple time scales, the short-term dynamics of microbial biomass over the course of the growing season and the longer-term processes affecting total soil organic matter content were studied in Alberta, Canada (McGill et al., 1986). Soils under a five-year cereal-forage rotation contained 38% more total nitrogen, but 117% more microbial nitrogen than the comparative wheat-fallow soils. The latter soils had average organic matter turnover rates 1.5 to 2.0 faster than the cereal-forage rotation. Furthermore, the faster turnover rates in the wheat-fallow soils could not be accounted for by the average carbon additions and may indicate that native soil organic matter is being mined (Anderson, 1991). Since the total organic matter contents changed very slowly, these conclusions depended upon studies established more than five decades earlier.

Again it is clear that all three questions about soil properties depend upon long-term processes. In the first instance, the unusual event that permitted the leaching of organo-metals would not be understood as unusual if the usual conditions were not known. The surface and subsurface salinity dynamics perhaps could be predicted, but the variability in the climate-driving variables makes prediction tenuous, and validation data would be necessary. In the last case, the small changes in organic matter would only be detectable over a long period of time.

15.2.3 FLUVIAL SYSTEMS

Rivers in human-inhabited regions have undergone two types of major modifications. First, the flow dynamics of these rivers have been modified by control structures on the rivers themselves and on the secondary channels and backwaters. Second, urbanization and intensified agriculture in the watershed have changed water quality in the entire system, fragmented the riparian forests, and affected the flow of organic matter to and from the river. The River Garonne in southern France has been influenced by flood and erosion control devices since the seventeenth century and in the river as it flowed through Toulouse since the twelfth century (Décamps and Fortuné, 1991 ). The results of these human activities include a reduction in the secondary arms of the river, a straightening of the river channel, and stabilization of the river bed. Ecological consequences of these changes include alterations in the successional sequences of the riparian vegetation and of the associated biota, a reduction in the retention of nutrients along the edges of the river, changes in richness of the fish communities, and depression of the water table near the alluvial rivers.

River ecosystems and their responses to human activities represent a special case in the analysis of long-term ecological research. These circumstances arise because of the great inertia to reversal of geomorphic changes (Munn, 1987). These changes are caused both because of the inertial strength of the river ecosystems and because the management of the rivers is so closely tied to socio-economic development. Thus, these fluvial systems respond not only to ecological processes in the usual sense, but also in a cumulative way to local, regional, and global economic conditions. However, unlike a farmer who can simply change the choice of a crop depending upon last year's commodity market, changes in fluvial systems persist for decades and even centuries.

15.2.4 SPECIES RESPONSES

Coastal dunes in Australia, near Myall Lakes National Park, are mined for heavy metals (Westoby, 1991). Small mammal patterns have been investigated on the recovering mined sites (Fox and Fox, 1978, 1984). The abundance of Mus musculus reaches a peak three to five years after the mining, but Pseudomys novaehollandiae becomes more abundant after five to seven years. On sites regenerating after fire, the sequence is compressed and P. novaehollandiae becomes more abundant after only the second year. This is probably because many plant species regenerate vegetatively after fire, but the regeneration, especially of rhizomes and tubers, is  much slower after mining. P. novaehollandiae abundance is positively correlated with vegetation cover.

Experimental manipulations done in concert with these longer-term field measurements confirm that M. musculus declines because it behaviorally avoids P. novaehollandiae. Measuring and predicting the response of organisms to climate change will depend upon the location of the sample sites relative to the ecology of the organisms in question (Heal, 1991 ). For example, repeated annual sampling of the homopteran Neophilaenus lineatus showed that the population regulation was density dependent at a low-altitude site, but density independent and climatically controlled at a more northern high-elevation site (Whittaker, 1971). Thus, the behavior of these two populations was only known from repeated sampling, but just as important was the realization that population control could be density or climate controlled.

These last two examples again depend upon long-term measurements. In the first case, the species response to disturbance depended upon the recovery rates of vegetation and the species-species interactions. These first- and second-phase responses were found because field sampling occurred over a long enough period to detect the change in small mammal dominance; and the associated experimental studies identified the actual mechanisms. The homopteran example might have been predictable from carefully controlled experiments, but again, testing the results of experiments requires a sequence of field observations.

The chapters in this book arise from various ecosystems around the world and they focus on different parts of the ecosystems. Throughout, there is a consistent recognition that long-term studies are necessary to understand ecological processes. The examples given are not idle curiosities, but rather they exemplify many of the most vexing ecological issues facing us today.

15.3 INTERNATIONAL NETWORKS FOR STUDYING

During the Berchtesgaden meeting there was a special effort to identify important long-term ecological questions which might form the basis for studying ecological processes on a global scale. The specific objective was to 'further define the rationale for long-term ecological research and identify important existing and emerging scientific questions which could most appropriately be addressed at long-term ecological research sites, particularly those related to environmental changes at the global scale'. The working group addressed two related aspects: the scientific questions to be addressed at long-term research sites and the networking of those sites.

Understanding the issues of global and regional change depends upon ecological processes at all smaller spatial scales. Achieving the detection, understanding, and prediction of global change and the prescription of human response to global change requires the formation of interactive, real-time networks of ecological studies around the globe. Explicit goals of international and regional networks are (1) to identify common patterns of change, (2) to develop regional and global syntheses of change and a mechanistic understanding of these changes, (3) to partition the variances between global and site-specific forcing functions, and (4) to develop approaches to scaling up (and down) from local to regional to global spatial scales. Networks will accelerate the rate of achieving these goals by reducing the time lags both for information transfer and the planning and conduct of appropriate long-term ecological research.

Networks should be flexible and form around scientists studying similar systems, on the one hand (e.g. those who study grasslands), to those studying markedly different systems (e.g. grasslands, lakes, urban ecosystems, coastal zones). These networks are essential when recognizing the short times over which global change may occur, the need for implementation of new techniques, and the spatial scale and diversity of ecosystems involved in these global problems.

The working group represented many disciplines, study sites, and countries. Discussions soon revealed that research priorities and objectives differed markedly, and that a generic research program could not be forced on sites/countries with different priorities. However, there were concepts that were common to all. For example, all individuals recognized the benefits of long-term research since the fundamental approach was to explain or account for variation in the parameters measured. The resulting measurements form the basis for justifying a global network and extrapolating results to regional and global scales. Only a network of sites can represent the large spatial scales needed to document changes that occur at those scales. The working group envisioned a global scale Geographic Information System (GIS) that could map and project results at appropriate broad scales. Networking allows sites to work on a common time scale and to synchronize efforts at quantifying responses to broad-scale environmental phenomena.

For example, an El Niño event may cause very different responses among different ecological parameters at one site and among different sites. This results from the many interacting factors at a site and the site-specific nature of those interactions. However, the initiation of the different responses may well be caused by the El Niño phenomena, a triggering effect, which could be common at many sites. The network could identify the scale of the triggering phenomena, the times of initiation and conclusion, and the types of ecological responses that were similar and dissimilar. For some sites, the network could separate the triggering response from the subsequent chain of events (succession of events) until a new triggering event occurred. Some regions may respond to certain types of triggering events while others may not. This information may help to identify the significance of various types of constraints on systems, the regional extent of those constraints, and the sequence of ecological phenomena which are typical following changes in those constraints. Networking is essential for a rapid information transfer, and facilitates early recognition of events that are common to large areas.

15.4 PROPOSED EXPERIMENTS

During the second workshop, investigators attempted to build on the results of the  first workshop by identifying important issues that required long-term ecological studies. The discussion involved three biome types: arid/semi-arid, temperate forests, and boreal forest/tundra. Important research questions for the arid/semi-arid biome focused on the stability of these systems and their responses to changing climate and human use. Climate change, pollution, and management issues were deemed most important in boreal forests, particularly the relationships between vegetation and the fluxes of radiatively active gases. In the boreal forest/tundra biome, research topics of highest priority dealt with the soil carbon budget, decomposition-limited nitrogen and phosphorous processes, episodic phenomena such as animal population fluctuations and fires, and permafrost and soil hydrology. Recommended experimental approaches included large 'flagship' studies, networks of smaller experiments, and various manipulation tests depending upon the research question itself.

A significant component of the second workshop involved the use of data management, and observational and communication technologies. These topics are important for organizing and conducting international-level long-term ecological  studies, especially those addressing scientific questions at regional and global scales. As described in Chapter 14, the workshop included televised sessions where scientists from throughout North America participated in the workshop.

15.5 CONCLUSIONS

Over the past century, ecological research has made significant, even remarkable, progress toward providing an understanding of many fundamental ecological processes. A portion of this success derives from the results of 'experiments' in which important variables have been manipulated and compared with controls in the classical sense. Because of the success of these experiments, ecological research has eagerly adopted this approach as the sine qua non for masterful and clever research. However, as described in this volume, many important discoveries have arisen from long-term measurements - both with and without experimental manipulation. Thus, there is great jeopardy in focusing exclusively on just one approach (Taylor, 1989) and losing the value of multiple research strategies. In addition, the natural history observations possible through long-term research projects provide valuable explanations of both specific processes and the evolutionary context of the current observations (Gould, 1989).

The greatest value of this book is the rich array of described long-term ecological studies and the insights provided by the authors. Selecting certain summary points detracts from this richness. However, one of the major purposes of the SCOPE project was to provide recommendations concerning the establishment of international networks for long-term ecological research. It is in this spirit that the following ideas are presented. In many cases, several authors made the same or related points, but at least one citation to a chapter in the book is made for each point.

Long-term ecological studies judged to be successful have been (1) associated with one dedicated leader who was responsible for the project and (2) designed simply, so that the experiments were easy to operate, the data could be used in various ways, and ancillary studies could be developed (Strayer et al., 1986). In addition, many of these successful long-term studies have included one or more of the following characteristics: they involved experimentation as part of the design, had clearly defined objectives, were conducted on protected sites, included provisions for archiving samples for future analysis and for comparing new methodologies, provided short-term justification by a continuing set of publications, responded to recognized policy or society needs, and included modeling and synthesis during the course of the study (Pickett, 1991).

In many cases, data to address long-term ecological phenomena will be acquired from many sources and will cover different time scales. An aspiration of an integrated long-term ecological research approach will be to minimize this ad hoc approach. However, past research should not be ignored since it possesses enormous amounts of information, and significant efforts should be devoted to the incorporation of meaningful and useful information from past studies for addressing specific ecological questions. When the data sets are incomplete or it is difficult to link data with specific management strategies, data from one or more levels may be 'soft coupled' by employing data management techniques combined with expert judgement systems (Grossmann, 1991).

Long-term ecological studies can be most useful for management of natural resources in three ways: (1) by providing records of subtle, chronic changes in ecosystems, (2) providing early warning of the onset of acute changes, and (3) suggesting management strategies for dealing with both of these types of changes (McNaughton and Campbell, 1991).

The substitution of space for time, or the chronosequence, has been used to explore time-dependent processes such as succession. In many cases the results from chronosequences have been misleading, especially because of different initial conditions, and thus long-term studies may be necessary (Pickett, 1991).

Designs for long-term experiments should include estimations of the proportions of variation in an ecological measurement which are attributable to differences between years, to differences between places, or to place-year variations, such as patterns which occur every year but are shifted in space (Westoby, 1991).

Although studying rare events is a valid reason for instituting long-term studies, some rare events are predictable and can be studied with short-term observations. Unique events, such as invasion by exotic species or cases where the system moves to a new set of conditions, are more difficult to predict, and understanding them may require long-term studies (Franklin, 1989; Pickett, 1991).

Long-term studies provide a perspective on rare events, and experimental designs must include sufficient duration to estimate return times, recognizing that return times may change as a result of the phenomenon itself (Westoby, 1991 ).

Long-term ecological studies must provide sufficient time to study second-phase events, such as a species arriving subsequent to the first conditions after disturbance or processes that are quite slow (Westoby, 1991). Not all studies can be extended for a longer duration because of the expense. Thus, the investigator should anticipate some of the likely changes and include these as experimental treatments. Also, modest funding for periodic returns to field sites would be an inexpensive approach for identifying and measuring second-phase and slow processes.

Establishment of a network of long-term ecological sites in any one country must recognize the specific characteristics of the research support structure of that country and, if the network becomes international, of the collection of funding agencies (Callahan, 1991).

The existence and expansion of long-term ecological research studies is constrained by the ability of the scientific community to perform the necessary observations and experiments, the availability of appropriate research sites, and the capacity of the sponsoring agencies to support and administer long-term research projects (Callahan, 1991).

Networks of sites allow more intensive narrow-scale studies to be put in context, and also permit estimates of space and time scales over which aggregated variables become relatively predictable (Westoby, 1991).

If unusual events are considered biologically important, merely monitoring them will not advance knowledge fast enough and, therefore, science funding must be organized so that hypotheses can be tested experimentally during unusual events (Westoby, 1991).

For many problems, the influence of humans is most likely to be important and detectable under conditions which are marginal to the species or system, or which are transitional between states. Thus, selection of sites for long-term measurements on ecotones or threshold sites can be as important as selection of sites which represent more average or typical conditions. Also, selection of sites should include careful analysis of the range of variation to be covered and should build on a planned network of intensive and extensive sites designed to accumulate the expected range of variation (Heal, 1991 ). A great challenge for long-term ecological research is to begin to measure variables over time, on the premise that we must begin collecting a coherent record now in order to have data against which to assess understanding and models of global change that may be developed in 10, 30, or 50 years time (Pickett, 1991; Westoby, 1991). Measuring decade-to-century level ecological phenomena requires serious time and financial commitments to data documentation (Seastedt and Briggs, 1991).

15.6 REFERENCES

Anderson, D. W. ( 1979). Processes of humus formation and transformation in soils of the Canadian Great Plains. Journal of Soil Science, 30, 77-84.

Anderson, D.W. (1991). Long-term ecological research: a pedological perspective. Chapter 7, this volume.

Callahan, J.T. (1991). Long-term ecological research in the United States: a federal perspective. Chapter 2, this volume.

Décamps, H. and Fortuné, M. (1991). Long-term ecological research and fluvial landscapes. Chapter 8, this volume.

Fox, B.J. and Fox, M.D. (1978) Recolonization of coastal heath by Pseudomys novaehollandiae (Muridae) following sand-mining. Australian Journal of Ecology, 3,   447-465.

Fox, B.J. and Fox, M.D. (1984). Small-mammal recolonization of open-forest following sandmining. Australian Journal of Ecology, 9, 241-252.

Franklin, J.F. (1989). Importance and justification of long-term studies in ecology. In Likens, G.E. (Ed.) Long-Term Studies in Ecology. Approaches and Alternative. Springer-Verlag, New York, 3-19.

Gould, S.J. (1989). Wonderful Life. The Burgess Shale and the Nature of History. W.W. Norton, New York.

Grossmann, W.D. (1991). Model- and strategy-driven geographical maps for ecological research and management. Chapter 13, this volume.

Heal, O. W. (1991). The role of study sites in long-term ecological research: a UK experience. Chapter 3, this volume.

Johnston, A.E. (1991). Benefits from long-term ecosystem research: some examples from Rothamsted. Chapter 6, this volume.

Likens, G.E. (1983). A priority for ecological research. Bulletin of the Ecological Society of America, 64, 234-243.

Likens, G.E. (Ed.) (1989). Long-Term Studies in Ecology. Approaches and Alternatives. Springer-Verlag, New York.

McGill, W.B., Cameron, K.R., Robertson, J.A. and Cook, F.D. (1986). Dynamics of soil microbial biomass and water soluble organic C in Breton L after 50 years of cropping to two rotations. Canadian Journal of Soil Science, 66, 1-20.

McNaughton, S.J. and Campbell, K.L.I. (1991). long-term ecological research in African ecosystems. Chapter 10, this volume.

Munn, R.E. (1987). Environmental Prospects for the Next Century: Implications for Long-Term Policy and Research Strategies. Research Report 15. International Institute for Applied Systems Analysis. Laxenburg, Austria.

Pickett, S.T.A. (1991). Long-term studies: past experience and recommendations for the future. Chapter 5, this volume.

Seastedt, T .R. and Briggs, J. M. ( 1991 ). Long-term ecological questions and considerations for taking long-term measurements: lessons from the LTER and FIFE programs on tallgrass prairie. Chapter 9, this volume.

Stoner, M.G. and Ugolini, F.C. (1988). Arctic pedogenesis. 2. Threshold-controlled subsurface leaching episodes. Soil Science, 145, 46-51.

Strayer, D.J., Glitzenstein, S., Jones, C.G., Lolasa, J., Likens, G.E., McDonnell, M.J., Parker G.G. and Pickett, S.T.A. (1986). Long-Term Ecological Studies: An Illustrated Account of Their Design, Operation, and Importance to Ecology. Institute of Ecosystem Studies Occasional Publications 1.

Taylor, L.R. (1989). Objective and experiment in long-term research. In Likens, G.E. (Ed.) Long-Term Studies in Ecology. Approaches and Alternatives. Springer-Verlag, New York, 20-70.

Weatherhead, P.J.( 1986). How unusual are unusual events? American Naturalist, 128, 150- 154.

Westoby, M. (1991). On long-term ecological research in Australia. Chapter II, this volume. 

Whittaker, J.B. (1971). Population changes in Neophilaenus lineatus (1) (Homoptera:Cercopidae) in different parts of its range. Journal of Animal Ecology, 40, 425-443.