A great deal of effort continues to be invested in habitat-related studies of wildlife. In the United States, federal land management agencies in particular have focused on developing formalized procedures for evaluating habitat for wildlife (Morrison et al. 1998). The procedures that have been adopted rely on models derived by species experts, who in constructing these models tend to rely more on experience than on empirical data (Schamberger and Krohn 1982; Thomas 1982). Therefore, the models are really hypotheses in need of testing. However, because these models hypothesize explicit relationships between habitat attributes and animal populations (so-called habitat suitability indices), they cannot be rejected or accepted in normal scientific fashion; that is, none of the relationships are likely to be exactly correct. Thus it seems inappropriate to suggest, as is common parlance, that they should (or even could) be "verified" or "validated."
Brooks (1997) and Morrison et al. (1998) proposed steps for verifying or validating habitat suitability models. Unfortunately, these authors and most others writing on this subject have misused these terms. Verification means establishment of truth and validation technically refers to establishment of legitimacy (i.e., in the case of a model, showing that there are no logical or mathematical flaws; Oreskes et al. 1994). The general misapplication of these terms in reference to model testing is not merely a semantic issue, but rather a real misrepresentation of accomplishment. Because of the complexity of natural systems, habitat models invariably exclude some relevant parameters and presume relationships that are not exactly or not at all correct. There is simply no way to perfectly model these sorts of open systems (i.e., systems in which all variables and relationships are not known). In a well-reasoned discussion of the subject, Oreskes et al. (1994:643) argued, "A match between predicted and obtained output does not verify an open model. . . . If a model fails to reproduce observed data, then we know that the model is faulty in some way, but the reverse is never the case." Assuming so is a logical fallacy called affirming the consequent. "Numerical models are a form of highly complex scientific hypothesis [unlike simple null models that we are accustomed to testing]; . . . verification is impossible." The utility of models is to guide further study or help make predictions and decisions regarding complicated systems; thus they warrant testing, but that testing should be viewed as a never-ending process of refinement, properly called benchmarking or calibration. Given the basis of habitat suitability models and the complexity of their many interacting variables, it is likely that any such model could be improved through rigorous testing.
Several attempts have been made to test such models. Often this process is circular, involving just another panel of experts making qualitative assessments (O'Neil et al. 1988). In other cases, models have been tested using results of a study on habitat use (Lancia et al. 1982) or use relative to availability (Thomasma et al. 1991; Powell et al. 1997), with the inevitable associated shortcomings discussed in detail in this chapter. In some instances, habitat management prescriptions based on "common knowledge" or expert opinion have, through collection of better data, been proven faulty (Brown and Batzli 1984; Bart 1995; Beyer et al. 1996). I found one case in which model-derived habitat scores for individual home ranges were compared with reproduction, juvenile growth rates, and home range size, but no significant relationships were observed (Hirsch and Haufler 1993).
Often, models have been tested by comparing habitat-specific densities to model predictions. However, even if a model explains a significant portion of the variation in density (Cook and Irwin 1985), the data collected to test (or purportedly validate) the model could be better used to modify it or build new one (Roseberry and Woolf 1998). In most cases, habitat models have proved to be poor predictors of animal density, indicating either defectiveness of the model, lack of a clear habitat—density relationship, or effects of other confounding factors, such as hunting pressure, which are also habitat-related (Bart et al. 1984; Laymon and Barrett 1986; Robel et al. 1993; Rempel et al. 1997). Bender et al. (1996) found that if the variance around the estimated values of the model inputs were taken into account (a process that is not commonly done), suitability scores for a variety of habitats that appeared very disparate were not significantly different; that is, the parameter estimates were not precise enough to even enable the model to be tested.
A major inherent but generally unstated (maybe unrecognized) assumption of habitat suitability models is that high-quality habitats (i.e., habitats that confer high fitness) are in fact suitable (i.e., able to sustain a population; Kellner et al. 1992). Explicit tests of this assumption are rarely conducted, yet counterexamples exist. In one case, a habitat suitability model for Florida scrub-jays (Aphelocoma coerulescens) correlated well with demographic performance (reproduction, survival, and density), but most of the area was found to be a population sink where mortality exceeded reproduction (Breininger et al. 1998). Apparently, as birds competed to occupy the best habitats, they were less alert to predators, and thus suffered high mortality. In another example, Kirsch (1996) found that interior least terns (Sterna antillarum, least terns nesting in noncoastal areas) selected high-quality nesting habitats, but possibly because of disturbance, their productivity was not sufficient to maintain population size (i.e., the nesting habitats were unsuitable). Lomolino and Chan-nell (1995, 1998) observed that remnant populations of endangered mammals often occur near the periphery of their former ranges; because the periphery of the range represents the edge of suitable habitat, studies of habitat suitability of endangered mammals based on habitat use in existing populations are likely to be misleading. These are not gratifying results.
It has not been for lack of effort, expense, or analytical developments that relationships between habitats and population growth often elude detection. It is simply the complexity of the interactions between animals and their environment that make such relationships deceptively difficult to understand. A case in point is the spotted owl, which has undergone intense scrutiny because of its threatened status and apparent proclivity for old-growth forest in a region where the economy is tied largely to timber; despite a plethora of habitat studies, significant debate persists among ecologists as to the critical habitat requirements of this species, and why it prefers old-growth forest (Forsman et al. 1984; Carey et al. 1992; Rosenberg et al. 1994; Carey 1995).
I am not suggesting that ecologists have not been creative in their efforts. However, substantive flaws in the most commonly used techniques for studying wildlife-habitat relationships apparently have not been widely recognized. I believe it is time to reconsider the ways these techniques are used in evaluating habitat quality.
Studies of the use of habitat have merit, but also many limitations. Habitats in which an animal spends a large proportion of its time are clearly selected among others available. Even if widely available, frequently used habitats are certainly not "selected against" or "avoided" (except maybe among life forms that have no memory of where they were). However, frequent use suggests nothing about habitat importance or substitutability in terms of fitness. Correspondingly, infrequent use may not be indicative of lack of suitability. A habitat may be used infrequently because it serves little value, because its value can be extracted in a short amount of time, because it is not readily available, or because access is constrained by threats (social pressures, competition, predation) or physical barriers. A high use:availability ratio might suggest (correctly) that such a habitat is more important than indicated by its infrequent use.
Studies of habitat use would benefit greatly from replication. Significant insights might be gained from comparisons of habitat use and use:availability among individuals, among groups of individuals in different portions of a study area, among study areas, among time periods, and so on. If individuals are disparate in their use or apparent selection of habitats (Holbrook et al. 1987; Ehlinger 1990; Donazar et al. 1993; Boitani et al. 1994; Macdonald and Courtenay 1996), inferences regarding habitat quality become more equivocal. In contrast, if the data are partitioned and the subsets show consistent patterns (e.g., use:availability ratios for each habitat are similar among individuals despite large differences in availability within individual home ranges) or if curvilinear relationships between use and availability can be ascertained (figure 4.4), inferences are strengthened; even so, studies of this sort provide only a superficial understanding of the effects of habitat on population dynamics. Certainly no strong prescriptions for habitat manipulation are warranted from interpretations of selection based solely on observed patterns of habitat use.
Site attribute studies tend to provide stronger inferences about habitat selection. However, in identifying myriad habitat characteristics that are apparently preferred, even for activities that impinge on the animal's survival and reproduction, these studies still cannot assume that population growth would be significantly higher in more "ideal" settings. Controlled experiments can help sort out the factors important in the animal's selection processes (Danell et al. 1991; Parrish 1995) and can thus provide a better understanding of its behavior, but without corresponding demographic measurements, the importance of various habitat components in terms of their contribution to the animal's fitness cannot be appraised. I agree with Kirsch (1996:37-38): "Unfortunately, proximate habitat features may not indicate habitat suitability, nor do they reveal the possible selective pressures that influence habitat selection in a system. One must measure components of fitness, determine factors that influence fitness, and relate fitness and factors influencing fitness to habitats or habitat features."
Demographic response studies are the only means of truly evaluating the relative importance and suitability of habitats for supporting animal popula-
Was this article helpful?