Analysis of Ecological Networks

The study of ecological networks has expanded greatly in the last 15 years. This makes sense, species are “connected” with one another in many different ways and in many different contexts. Even ecological communities can be thought of as being connected. Beyond representing a reality in nature, ecological networks and the application of a network framework and analysis to ecological data is highly useful. It can reveal meaningful pattern in a way unlike any other statistical/analytical method.

My interest in ecological networks stems from separate collaborations with Giovanni Strona (University of Helsinki) and Dan McGarvey (Virginia Commonwealth University). Giovanni and I have developed new statistical methods to analyze structure in networks, and conducted other studies that utilize a network approach (see Publications list). Dan and I analyzed species co-occurrence networks developed from large continental databases of the distribution and abundance of stream and river fishes in the USA.

In the broadest context, a network consists of nodes connected with each other through linkages (edges) and sometimes internal structuring in the form of modules (subsets of the overall network consisting of highly connected nodes). A linkage almost always entails that the two linked nodes share a feature, property, or entity in common. For example, in species co-occurrence networks, two linked species (nodes) co-occur at a greater-than-randomly-expected number of survey sites, hence they share locations in common. In bipartite networks such as plant-pollinator, two linked pollinator species may share many of the same plant species, and vice versa. Two nodes always share something. Given this perspective it should be possible to apply network-thinking and build networks to investigate a wide array of ecological pattern and process particularly when the nodes are species or particular places in nature.