New paper: using network analysis to understand ecosystem function in agricultural landscapes


Saunders ME, Rader R (2019) Network modularity influences plant reproduction in a mosaic tropical agroecosystem. Proc. R. Soc. B 286: 20190296. (all data and code available on github)

I’m so excited about our new paper! We use network analysis in a cool new way to understand how pollinator community structure influences ecosystem function in a heterogeneous landscape. Understanding links between structure and function is a core goal of ecological research, but there are still plenty of things we don’t know about these relationships.

In our study, we used a dataset of flower visitor observations on potted flowering plants (Brassica rapa) that were placed in four different land uses across a mosaic agricultural landscape. We collected data on visitation rates, pollinator species, and seed set and used bipartite networks to identify the structure of the pollinator community, and then tested relationships between network metrics and plant reproduction (i.e. ecosystem function).

Bipartite networks are a common analysis tool in ecology, especially in pollination ecology. Plant-pollinator networks are usually constructed by linking flowering plants to the pollinator species that visit them. In our study, we tested a different method, linking pollinators to the sites they were observed at. This method was proposed recently as a novel way to use bipartite networks to address applied conservation problems, because it can reveal how communities are structured at the landscape scale, and how landscape patterns influence community structure.

There are a whole suite of different metrics that can be calculated for bipartite networks, depending on what aspect of network structure you are interested in. We were most interested in connections between structure and function, so we focused on modularity. One relevant node-level metric is the participation coefficient (also called c value), which represents how well the node acts as a connector between modules. Participation coefficients have been shown to have a positive relationship with function in other types of networks, like cognitive and metabolic systems, so we were keen to test this relationship for ecological function. We found a positive relationship in our system – sites with higher participation coefficients had higher rates of plant reproduction, and pollinator species with the highest coefficients appeared to be the most influential species in the network. This is an exciting result that warrants much more exploration!

I think this approach has huge promise for tackling problems around conservation in managed landscapes and understanding ecosystem services supply. These problems need to be solved at the landscape scale, and linking species to sites or habitat types can help identify keystone species that influence structure and function, as well as prioritise different conservation actions for different parts of the landscape.

This paper is a personal highlight for me. It’s the first ‘official’ paper from my current postdoc project and has helped me develop many ideas I’m looking forward to exploring in other systems. I’ve always been interested in network theory, but had only dabbled in the literature and never used network analysis in my own research. The last couple of years have been a crash course in the sometimes confusing world of network metrics, and I’ve loved every minute of it. I still have a lot to learn, but I’m extremely chuffed that I had so much fun and learned so much working on this cool paper. Stay tuned for more insect and ecosystem services network fun!

Figure 3

© Manu Saunders 2019

What do you think?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s