Much of the research we do in the Geospatial Analysis Lab (GAL) is on human or wildlife movement, which involves measuring movement over time often using GPS tracking. Dealing with time in GIS is notoriously challenging, a most GISystems are designed to store, analyse and visualize spatial data such as tables, vectors or rasters (i.e. two dimensional spatial information). Time is often integrated into GISystems as an attribute of spatial features, which is the case in our research on GPS tracking feral swine, and other projects. With time as an attribute of spatial data, we can visualize and analyze mobility and movement processes like animal movement in GIS.
One way to deal with time in GIS is to aggregate all features within a time period. When time is stored as a feature attribute, we can plot or analyze all records in the table. In wildlife movement research this allows us to visualize and analyze the space use of individuals over the entire study period. Comparing the overall space use of individuals can tell us about their space use sharing, and is a quick and informative way to detect non-overlapping individuals. If there is no spatial overlap between any individuals over the duration of the time period, there should be little chance of spatiotemporal overlap, especially if there is a significant distance between the most adjacent GPS tracking points of each individual.
Here we see the extent of the GPS tracking points for three individuals. The GPS tracking points of one individual, collar 21965, do not overlap with the tracking points of the other two, therefore it is unlikely these individuals ever overlapped in space and time. Two individuals' (collars 21953 and 21962) tracking points overlap extensively in space, therefore there is potential for them to have been in the same place at the same time, but we cannot yet be sure.
Once we've identified these two might have contacts, we can compare their space use over time, giving us an initial sense of their movement relative to each other. An easily interpretable way to initially incorporate time into our analyses is through animations, which through improvements in GISystems and open source resources are increasingly easy to produce. Animations provide a quick way to visualize the often large volumes of movement data we work with, which is an effective way to interact with the data.
From this movement animation, we can get a better sense of our data, which previously was just a long table of dates, times and positions. This animation allows us to start formulating hypotheses related to the animals' behavior and movement. For instance, we can see these two seem to spend some time apart, then join together and follow each other around for a while. We might hypothesize that these two belong to the same social group, but members of the same social group do not stick together all the time.
Now that we have an initial sense of their behaviour and some hypotheses to investigate, we can do some feature analyses incorporating time. To determine how much of their time these two individuals spend with each other, we can go through the table and identify when these two are in the same place at the same time (i.e. in contact with each other). Measuring contacts is a popular way to begin to analyse mobility patterns and movement behaviour, and allows us to test our hypotheses.
Here, we have a new table of points representing when these two individuals were in the same place at the same time, symbolized in purple. Detecting contacts is much more informative than simple area overlays like bounding geometries shown in the first figure, as we are given a much finer knowledge of the overlap of each individual. Using contact rates, we can begin to look for movement behaviours and phenomenon of interest, such as social structure dynamics.
Working with time in GIS can be tricky due to the language GISystems operate in, but incorporating time into our analyses can give us a finer understanding of the behaviours of interest, such as individual/group space use and sociality.
