How to monitor ecological communities cost-efficiently: The example of plant–pollinator networks

Conservation practitioners often lack tools to monitor functioning of communities because time and monetary constraints create a gap between the optimal monitoring methods and the practical needs in conservation. Interaction networks provide a framework that has proven useful in ecological research. However, they are considered time consuming and too expensive for conservation purposes. We investigate whether it is possible to sample interaction networks cost-efficiently and whether a compromise exists between data quality and amount of resources required to sample the data by using a highly resolved mutualistic plant–pollinator network sampled over two years in Norway. The dataset was resampled with decreasing sampling intensity to simulate decreasing monitoring costs and we investigated the cost-efficiency of these monitoring regimes. The success in monitoring community structure varied largely with sampling intensity and the descriptor investigated. One major result was that a large proportion of the functionally most important species in the community, both plants and insects, could be identified with relatively little sampling. For example, monitoring only in “peak-season”, which costs ca. 20% relative to full monitoring, resulted in recording of 70% (in 2003) or 85% (in 2004) of the top 20 most functionally important pollinator species. Also, peak-season monitoring resulted in relatively precise estimates of several network descriptors. We present a first estimation of the full cost (travel time, sampling time and taxonomic services) of constructing pollination networks with different sampling effort. We recommend monitoring plant–pollinator networks in temperate regions during peak-season to cost-efficiently collect data for practical habitat management of ecosystem functioning.