Use and Misuse of Landscape Indices

Readings Assignment Use and misuse of landscape indices The purpose of this viewpoint paper is to study the fundamental issues that provide considerable challenges to landscape pattern analysis, particularly when it comes to relating patterns to processes, and then offer some solutions. According to the paper, the paper tries to identify three general issues that could be the explanation for this unsatisfactory progress: (1) conceptual flaws in landscape pattern analysis, (2) inherent limitations of landscape indices, and (3) improper use of landscape indices. A. Conceptual flaws in landscape pattern analysis  Unwarranted relationships between pattern and process Sometimes Landscape analysis may have conceptual flaws if it fails to recognize the existence of non-interactive interactions. Thus, some form of “proof of existence” is needed to demonstrate or reliably infer a reciprocal or one-directional relationship between spatial pattern and process. First, if a process of interest does not produce the spatial pattern, the process should not have any inherent relationship to the spatial pattern, and inter- active relationships should not be assumed. “One-directional effects” of spatial pattern on ecological processes may still exist, but additional information is required to confirm them and different interpretation of results may apply. Second, if a process is non-spatial, effects of spatial pattern on the process are unlikely to occur because spatial pattern is not a critical factor to the non-spatial process. In this situation, the value of spatial pattern analysis and spatially explicit modeling may be limited. Third, if the pattern and the process do not operate at similar rates and in the same domains of spatial scales, they cannot have an interactive relationship (Wu 1999).  Ecological irrelevance of landscape indices If landscape indices are ecologically relevant and reflect important attributes of spatial pattern, they can functionally link the dynamics of ecological processes to landscape structure (e.g., Wiens et al. 1993). However, ecological relevance of landscape indices is more often presumed than established, with inadequate supporting empirical evidence in the literature. In the absence of such evidence, landscape indices are mathematical constructs that have no inherent ecological meanings. The validity of landscape analysis will increase if indices are selected according to their ecological relevance rather than the convenience of computer programs.  Confusion between the scales of observation and analysis According to the paper, the scale of observation is determined by the characteristics of the system studied, the questions asked, and the data collection protocols used (e.g., platforms and sensor types for remote sensing data, sampling schemes for field survey data). While the scale of analysis is determined by the original scale of observation and the methods of data transformation, including aggregation, magnification, and resampling. B. Inherent limitations of landscape indices  Variable responses to changes in spatial pattern Landscape indices may not differentiate landscapes with qualitative changes because of their insensitivity and non-uniqueness (Haines-Young and Chopping 1996; Turner et al. 2001). Such limitations make result interpretation difficult and compromise the effectiveness of correlation analysis because they cannot capture or distinguish some of the fundamental changes of landscapes in each situation. A solution to this problem is to assure the ecological relevance of map data by determining how much detail to recognize and what scale to use according to the phenomenon and the objectives of the study. Landscape indices may also be insensitive to functional attributes of landscapes (Wiens 1989). One example of such attributes is the habitat suitability of a landscape as viewed by wildlife species. To generate landscape maps that are relevant for spatial analysis of habitat quality, the paper recommends an approach that combines landscape and species habitat evaluation. The approach involves creation of a habitat rank map for each target species (or a guild) by aggregating adjacent patches with the same habitat rank after ranking each patch as optimal, suitable, marginal, and non-habitat (Hamel 1992; Li et al. 2000). The use of habitat rank maps strikes a balance between speciesspecific information and landscape analysis, and reflects a landscape’s suitability for different species.  Difficulty in interpreting indices The limitations and benefits of landscape measures are still not well understood, which makes analyzing indexes challenging. The best solution is that landscape ecologists understand not only the methods of spatial pattern analysis but also the concepts on which the methods are based (Dale 1999) C. Improper use of landscape indices  Quantifying pattern without considering process Spatial pattern analysis is of limited value if not used to explain structural changes in landscapes and predict how they influence ecological processes. If better understood and practiced, landscape pattern analysis can play useful roles in theory and practice of landscape ecological research (O’Neill et al. 1988; O’Neill et al. 1999; Turner 1990; Li and Reynolds 1994; Gustafson 1998; Turner et al. 2001)  Caveats of correlation analysis with landscape indices correlation analysis with landscape indices can be problematic when the conceptual flaws and inherent limitations discussed in the previous sections are ignored. Although it is not a solution for pseudoreplication, simulation experimentation can be extremely valuable when landscape dynamics can be studied with different scenarios and sufficient replications (e.g., Gardner et al. 1987; Li and Reynolds 1994). D. Concluding remarks The ultimate goal of landscape pattern analysis should be to achieve better explanations and predictions of ecological phenomena based on established relationships between pattern and process. A fundamental challenge in landscape analysis today is to establish sound relationships between spatial pattern and ecological processes (Wu and Hobbs 2002). Specifically, great efforts are needed to address two critical questions: (1) how to use knowledge about spatial heterogeneity to improve explanations and predictions of ecological systems, and (2) how to scale up models and/or relationships across scales with aid of quantifying spatial heterogeneity at multiple scales.