Month: April 2017

This week, landscape ecology and the application of GIS within the field was explored. Landscape ecology can be defined as a field that explores the importance of spatial patterning on the dynamics of interacting ecosystems. We learned with landscape ecology the landscape is the most important unit of study, where an area is spatially heterogeneous in at least one factor of interest, and is the scale at which management decisions and human effects are most commonly considered. Landscape ecology also has a fundamental assumption that the location of things can have important consequences on the surrounding environment. If spatial autocorrelation is shown by objects or events, patterns can develop from a response to an environmental factor (first-order process) or interactions between objects themselves (second-order process). Stationarity and scale are also important to landscape analysis as it measures how well patterns stay the same over space.

Within this week, the overall impact geography can have on data was touched upon. Geography is crucial because the geography of a problem must be considered to produce an effective analysis and many geographical elements of a study such as scale, boundaries and spatial dependence can impact results produced. In regards to landscape ecology, there isn’t a “natural” scale at which ecological studies should be undertaken. While most ecological phenomena have a spatial temporal domain, the definition most appropriate scale must be considered carefully. The Modifiable Areal Unit problem (MAUP) was again reviewed, which describes how scale or boundaries impact the data being collected.

MAUP must be considered in all GIS analyses to evaluate potential sources of error and uncertainty. One concept which highlights the impact geography can have on an analysis or social data collection is Gerrymandering, which refers to the practice of manipulating electoral district boundaries in order to create zones where a particular party would have a political advantage. The Simpson’s paradox was also explained as a confounding variable that is not taken into consideration in the analysis as some variables vary in correlation with another (e.g. high unemployment and other socio-economic characteristics) making it difficult to obtain a reliable estimate of true correlation between two variables.

Within our first week of GEOG 479, we went over the general structure of the course and the topics we will be covering in the future. We will be exploring three broad topics-landscape ecology, health and crime, finding geographic commonalities in their spatial analysis. Learning to see geography as a common factor in the analysis of subjects that at first seem too distant to be related seems to be the takeaway point of GIScience.

FRAGSTATS was also introduced this week. Fragstats is a software program that computes landscape metrics for categorical map designs. We will be using it in our first lab to see how patterns and interactions of a landscape changes over time. Fragstats is a useful program for the study of landscape patterns and mosaic. A program that is able to do such an analysis is critical because landscapes do not exist in isolation. They exist within context regardless of the scale defined, housed in a larger landscape within larger landscapes. While keeping patterns, processes, places, people and perspectives in mind, we will analyze landscapes and changes to them in next week’s lab using the background skills we went over this week.