Geocomputation

Connections to other Fields in Geography and in other Disciplines

The problems that geocomputational methods solve span the breadth of geography, including applications in geomorphology, ecology, urban geography, transportation geography, climatology, and medical geography.

In addition to the ties to application domains in which geocomputational techniques and methods are used, the field of geocomputation has connections with several other methodological subdisciplines within geography, including remote sensing, geographic information science, spatial analysis, and geovisualization.

Many data sources for geocomputational data structures and algorithms come from remote sensing, so many of the issues that arise in remote sensing are also dealt with in geocomputation. A prominent issue from remote sensing is the size of data sets that can be involved in this research. Remote sensing satellite systems can generate up to gigabytes of data each day. To best use data sets of this size, especially those that are continually growing in size, data structures allowing for efficient compression of the data and algorithms that can equally rapidly analyze raster data sets of that size need to be developed.

Geocomputation connects with the broader field of geographic information science (GIScience) by contributing to the computational aspects of GIScience. Sample topics in this overlap include data and process modeling with algorithms such as cellular automata and agent-based simulations, how best to represent and analyze uncertainty in spatial data, what data structures and algorithms are effective for representing and analyzing spatiotemporal data, and contributing to the computational aspects of concept mapping and ontologies.

As data sets grow in size and complexity, many of the techniques developed in spatial analysis are entering the geocomputational realm. As data sets get bigger, the statistics used to analyze them must be performed on a computer. The connection between spatial analysis and geocomputation is more substantial than simple application of previously developed methods; the greater processing power of computers permits new forms of analysis on these data sets. Statistical methods, such as Moran's I autocorrelation statistic and regression, can be adapted for spatial data to become the local Moran's I statistic and geographically weighted regression. Making even more use of the computer's processing power, artificial intelligence and data mining techniques developed in statistics and computer science are being adapted and applied to geographic data.

Geocomputation has connections with geovisualization and the newly emerging field of geovisual analytics. Geocomputational methods contribute to these fields through the data compression, summarization, and filtering methods needed to reduce very large data sets to a size that is viewable on a computer monitor. Additionally, as ways of visualizing the results from the analytical methods developed in geocomputation are developed, geocomputation contributes to geovisual analytics.

...