Landscape Pattern Analysis

The concept of metapopulation, a population of subpopulations that go extinct locally and recolonize regionally, resembles the theory of island biogeography in that both consider extinction and colonization as the two key processes. However, the former is concerned with population dynamics and species persistence while the latter focuses primarily on species diversity and turnover. Also, sources for species colonization in most metapopulations are neighboring habitat patches that themselves are subject to local extinctions.

The classic (or Levins) metapopulation models are commonly known as “patch-occupancy” models in which the proportion of habitat patches occupied by a species is modeled as a function of local extinction and inter-patch colonization (Fahrig 2007). These models assume that there are an infinite number of identical habitat patches in the landscape, and that within-patch population dynamics and the landscape matrix are not important to metapopulation dynamics. The classic metapopulation models are not really spatial models. More spatially sophisticated metapopulation modeling approaches have been developed in the past several decades. For example, many population models based on diffusion-reaction equations, which consider both local population processes and patch attributes (e.g., size, relative distance to other patches), are relevant to the study of metapopulation dynamics. However, these are quasi-spatial models which can not explicitly consider the locations and geospatial relations of habitat patches and the heterogeneity of the landscape matrix. The prevailing metapopulation modeling approach now is the so-called spatially realistic metapopulation models that incorporate the effects of habitat patch size and isolation on extinction and colonization rates into the classic metapopulation models. While these models are spatially realistic, like the classic models they are concerned only with two states of habitat patches: presence and absence of a species under study, not with population processes within habitat patches. Also, the heterogeneity of the landscape matrix is usually ignored in spatially realistic metapopulation models.

Metapopulation theory has been increasingly used in conservation biology in the past three decades, replacing the prominent role of island biogeography theory. However, its use for the practice of biodiversity conservation is limited by its species-specific focus and inadequate consideration of the heterogeneity of landscape matrix and socioeconomic processes. In reality, populations neither live in habitat patches that can always be neatly delineated nor reside in a homogeneous landscape matrix. Rather, they are situated in heterogeneous and dynamically complex landscapes that are shaped by a myriad of physical, biological, and socioeconomic processes. Thus, the metapopulation approach is useful, but certainly not adequate for achieving the overall goal of conserving all levels of biodiversity.

In reality landscape fragmentation simultaneously leads to habitat loss and habitat isolation. These changes can certainly affect the demographic and genetic processes of populations. A great number of theoretical and empirical studies have been carried out to understand how habitat fragmentation affects population dynamics and species persistence in the past several decades. This section provides an overview of the major findings up to date.

Findings of the effects of landscape fragmentation on population dynamics and species persistence have been, more often than not, incongruent because of several reasons. First, the term “landscape (or habitat) fragmentation” is often used to denote both habitat loss and habitat isolation (i.e., habitat fragmentation per se), and consequently the effects of the two factors are confounded in the results of such studies. Second, various measures that reflect different aspects of landscape pattern at different scales have been used to quantify habitat fragmentation. Some measures focus on habitat loss, others are indicative of changes in habitat configuration, and still others are mixtures of both. Also, habitat fragmentation is measured either at the scale of individuals patches (as in most metapopulation models) or the scale of the entire landscape (as in most landscape ecological studies). Third, different theories and models have different assumptions about what is important in fragmented landscapes in terms of population dynamics and species persistence, and these differences in assumptions often translate into discrepancies in results. Nevertheless, studies in recent decades have produced several important findings.

The relative effects of habitat loss and habitat isolation have been one of the central topics. Increasing empirical evidence indicates that habitat loss usually has much stronger effects on population dynamics and species persistence than habitat isolation. In general, the effects of habitat isolation tend to be stronger when the total amount of habitat in the landscape is small and when the species under consideration have limited dispersal abilities. The effects of habitat loss are consistently negative whereas those of habitat isolation can be either negative or positive depending on the idiosyncrasies of the landscape pattern (e.g., the spatial configuration of habitat patches) and the species under consideration (e.g., abilities for local competition and regional dispersal). The negative effects of habitat loss are easier to understand because the removal of habitat usually results in reduction in the number of species, the abundance of populations, and the carrying capacity of the landscape.

The effects of fragmentation per se, however, are more complex because the outcome depends on how the species responds to the specific features of the fragmented habitat and altered interactions with other species in the landscape. The negative effects of habitat isolation may be caused by the disruption of dispersals, increased local extinction rates in small patches, and detrimental edge effects. The positive effects of habitat isolation may be attributable to relaxed interspecific competition, reduced predation, and disrupted spreading of disturbances. However, it is important to note that the effects of spatial patchiness occurring naturally are different from the effects of habitat fragmentation by human activities. In the latter situation, species usually do not have enough time to adapt to the newly changed environment, and thus positive fragmentation effects are less likely, especially, for non-edge species.

The size of habitat patches has significant effects on population dynamics and species persistence simply because large patches tend to have larger populations (thus with lower extinction probabilities) and more species (due to both pure area effect and higher habitat diversity). In general, patch size has strong positive effects on interior species that require sufficiently large and relatively stable habitat. As patch size increases, the relative area of edge habitat decreases, resulting in negative effects of patch size on edge species. For generalist species that do not distinguish between edge and interior habitat, the effects of patch size usually are insignificant. The effects of patch size on population dynamics and species persistence may also vary with species that have different behavioral characteristics. For example, some studies have suggested that more mobile or dispersive species would be less strongly affected by landscape fragmentation. However, recent studies show that the opposite may be true when more mobile species suffer severe dispersal mortality in the landscape matrix (Fahrig 2003, 2007). Other patch characteristics such as shape, orientation, and boundary conditions can also affect population processes. Their effects seem less significant than those of patch size, and usually are even harder to generalize across different species and habitat types.

Landscape connectivity, which is conversely related to habitat isolation, plays a crucial role in maintaining population abundance and species persistence by affecting the movement of organisms and propagules, dispersal mortality, and gene flows. Studies from landscape ecology based on percolation theory have suggested that, as habitat area decreases to some critical value, landscape connectivity drops abruptly, indicating a possible extinction threshold for species with limited dispersal ability or high dispersal mortality (With 2004). This finding corroborates the hypothesis that the effects of habitat isolation on population and species dynamics tend to be more important with decreasing habitat amount in the landscape. Thus, landscape connectivity exhibits threshold behavior and is species- or process-specific. Corridors, as a means of increasing habitat connectivity, can promote species persistence (by enhancing re-colonization) and genetic integrity (by preventing genetic drift and bottleneck effects). However, corridors may also increase the spread of diseases and other disturbance agents across the landscape.