Why Geography is Important – Part I — 12 January 2015
Landscape ecology is fundamentally influenced by geography, and the challenges associated with the field of study are applied, not solely theoretical. One example of this is scale: “There is no single natural scale at which ecological phenomena should be studied; systems generally show characteristic variability on a range of spatial, temporal, and organizational scales.” (Levin, 1992). Everything occurs in a given place at a given time within an organizational structure.
A working solution to this problem is to find an appropriate scale even if a “natural one” does not exist. For health geography, this would include determining whether a phenomenon should be assessed on a a national versus neighborhood level. In ecology, this involves analyzing the specific habitats of a species; some species like urchins occur in specific places (simple selection), and others like cougars cross many different habitats (complementary selection). Crime, as well, involves these two types of selection in its patterns. One example is the problem of Marauder vs. Commuter. Routine activity theory supports the judgement of commuter. Geographic profiling is also used to identify criminals.
A further solution to the problem of scale is to use multi-scale analyses, as the way we look at things changes our perception of what is occurring. For example, you could study the African Monsoon patterns not only at the level of major river basins, but also at the level of pools and vegetation.
Analyzing the geography of a problem is important because of several factors:
- Scale, grain and extent of study area
- Modifiable areal unit problem
- Nature of boundaries of a study area
- Spatial dependence/ heterogeneity
- Whether the data results from processes or inherent spatial nature. The answer to this is usually that the effect is a combination of both.
The following is some useful terminology:
- Grain: “minimum resolution of the data”
- For a raster, this is the cell size. For a field sample, the quadrat size. For imagery, the pixel size. And for vector spatial data, the minimum mapping unit.
- Extent: “scope or domain of data (defined as the size of the study area, typically)”
The are several effects that can occur on the edges of an area. This area can be influenced by its natural neighbors. A researcher can have more confidence in the statistics of an area with many neighbors than one with few, because they will have a better understanding of the phenomena occurring around the study area. From an ecological point of view, continuous landscapes favor some species, while open landscapes favor others. For example, invasive species can thrive on the sunlit open edges of a forested patch
- Scale: “spatial domain of the study” or “level of organization depends on the criteria used to define the system”
The scale of a study will have a huge impact on the results of the study. For example, population-level studies look at conspecific species individual reactions while ecosystem-level look at biotic/abiotic interactions. Changing the scale will also affect the results. For example, if you increase scale, parameters that were constant are now variable. So, if you are looking at nutrient loading in the lower part of a lake, and you increase the scale to include the upstream area, you may find a new source for the nutrients. It is also important to note that scale is not only for spatial factors: time can be a scale. As well, certain things may become apparent as scale changes. These are known as emergent processes.
- Autocorrelation (positive or negative): the relationship between two things based on their location relative to each other.
- “If the presence of some quantity in a sampling unit…makes its presence in neighbouring sampling units…more or less likely” (Cliff and Ord, 1973)
Negative autocorrelation means that objects are dispersed. This includes places such as fire stations that desire maximum dispersal around an area so that all of the places within a study area have equal access to this resource.
Positive autocorrelation means that objects are found together. For example, communicable diseases are found in clustered areas because the germs have spread through contact between individuals.
There may also be random autocorrelation, where no pattern detected, although some clusters may appear.
The non-random distribution of organisms means that most ecological problems have spatial dimension. This may be due to limited dispersal, gene flow, clonal growth, or environmental factors, among others. Inherent, extrinsic, and random forces all determine these patterns. Moran’s I is a “weighted product-moment correlation coefficient, where the weights reflect geographic proximity” to help determine spatial autocorrelation patterns. Values of I larger than 0 indicate positive spatial autocorrelation, while those smaller than 0 indicate negative.
Another important reason for understanding geography is the Modifiable Aerial Unit Problem (MAUP). This includes scale and aggregation problems, and asks how specific or the results are to the data units that were used in the study, and whether they can be generalized.
- Scale effect: “the tendency, within a system of modifiable areal units, for different statistical results to be obtained from the same set of data when the information is grouped at different levels of spatial resolution.” When you change the scale, the grain changes. For example, you change from one individual to a block of individuals.
- Aggregation or Zoning effect: “the variability in statistical results obtained within a set of modifiable units as a function of the various ways these units can be grouped at a given scale.”
One important example of MAUP is gerrymandering, or redrawing electoral districts to get specific people elected. The purpose of census is for accurate population count to determine equal representation. So it could be redistricted if necessary. However, politicians could use this tactic to their advantage to get themselves or others elected.
Often, we do not know how data was aggregated, so we must work with it however it was given to us. This can lead to problems extending the data to draw conclusions from other levels of spatial resolution. Two of these problems are ecological fallacy and individualistic fallacy. Ecological fallacy is the assumption that all individuals within a given group will have the same characteristics. Individualistic fallacy is the opposite: the assumption that the characteristics of a group match that of an individual.
We should note that if something is truly present in the environment, changing the spatial units shouldn’t affect it. Therefore it is useful to investigate different to double check the effects of the spatial units.