Given a standard set of measurement scales, it is necessary to consider the nature ofthe systems measured using these scales . Ecological systems are fundamentally complex entities because they involve the interaction of many different kinds of subsystems. The problem of recognizing where one system ends and another begins is exacerbated by this complexity. Ecologists have become increasingly aware ofthe fact that this complexity shows some characteristic patterns that allow a certain degree of simplification. A common theme among recent approaches to ecological complexity is the use of hierarchical structures to represent ecological systems.
Nested hierarchies are formed when subsystems are combined together to form larger systems. The properties of any particular focal level are derived from the properties of lower-level entities as well as from the context provided by the higher-level entity in which the focal level is contained. A taxonomy is an ecologically important nested hierarchy. Non-nested hierarchies are formed by functional connections among different levels. Higher levels do not contain lower levels; rather they derive their properties from inputs coming from lower levels. A food web is an example of a non-nested hierarchy. Ecological hierarchies are models used to represent the salient features of a complex ecological system.
The fundamental strategy ofthe hierarchical approach to ecological complexity is to find the appropriate scale or scales to measure the properties of each level in the hierarchical system. Different measurement scales provide complementary information about the complex system. Statistical calculations on such data weight the observations in ways that provide insight into patterns associated with processes operating at different levels in the hierarchy. Appropriately designed experimental and observational studies maximize the ability of researchers to test hypotheses about proposed hierarchical mechanisms responsible for the complex behavior of target systems. Hierarchical theoretical constructs provide a link between ecological theory and data that enhances the ability of ecologists to test hypotheses and develop theories about ecological processes.
If ecological phenomena are organized into hierarchies, then it is often possible to develop models that incorporate effects from several different measurement scales simultaneously. For example, this approach has been applied fruitfully to the study of food webs, where levels in the hierarchy correspond to trophic levels. Usually the temporal dynamics of predators, prey, and herbivores are resolved over similar spans of time, so at least conceptually, measurement of food web dynamical properties can be done at the same scale. Ecological concepts derived from simple multiscale models offood webs show a wide variety of phenomena operating to determine community dynamics. One of the earliest insights of this type was the discovery of 'apparent competition'. Apparent competition occurs when two species share a common predator. Increases in one species may lead to decreases in the other species through the increased impact of predation due to higher predator abundance. The more species and trophic levels included in such models, the more complex the behavior becomes. The existence of such complexity implies that there may be constraints on the number of trophic levels that can exist within a food web. Empirical data of food webs support this interpretation.
The process of developing multiscale models to describe ecological systems requires an understanding of how many scales are needed to make appropriate descriptions. This has resulted in the suggestion that in most situations, there are three relevant scales. The first is the focal scale, that is, the scale at which the components of the system are measured. Conceptually, the next lower level in the hierarchy describes the properties of focal level entities. Measurements are taken at the scale at which the members of focal level entities undergo their dynamics. Finally, the focal level is embedded in a larger, more complex system, and the properties of that system may limit or direct focal level system components toward certain kinds of behavior. As an example, consider a population of organisms living in a particular habitat. If the goal is to describe the dynamics of the population, then that description requires the measurement of the individual organisms that comprise the population. Presumably, the better the descriptions of organismal activities relevant to population change are, the better the description of population change will be. However, the population is embedded in an ecosystem made up of predators, prey, diseases, and abiotic conditions. If the influences of at least some of these factors are included in the description of population dynamics, a clearer representation of the population system is obtained.
Treatment ofan ecological system as a hierarchy requires the ability to recognize and measure different levels of organization relevant to the behavior of the system. This can be particularly difficult in certain kinds of ecological models, particularly those that do not model easily recognized boundaries, such as models of energy flux and nutrient cycling. A useful definition of a boundary among objects at the same hierarchical level is based on the rates of processes within objects. Boundaries among objects are recognized by the existence of a steep rate gradient. That is, the rate of change in space and/or time of the process being modeled shows a marked shift when a boundary has been encountered.
Recognition of the relationships among lower- and higher-level objects depends both on the ability to identify objects at each level, and on the type of hierarchy being represented. In many spatial applications, patterns of spatial autocorrelation identify boundaries among different spatial units. Agglomeration of the minimum spatial units into nested sets of larger units allows the inference of levels or organization in a spatially nested system. Similarly, patterns of temporal autocorrelation may identify rate changes at different temporal scales corresponding to the dynamics of different organizational levels. Spectral analysis of time series can identify harmonics that correspond to different periodicities. Harmonics that occur at different temporal scales may provide evidence of nested processes.
Was this article helpful?