Ecological monitoring through harmonizing existing data: Lessons from the boreal avian modelling project

To accomplish the objectives of a long-term ecological monitoring program (LTEM), repurposing research data collected by other researchers is an alternative to original data collection. The Boreal Avian Modelling (BAM) Project is a 10-year-old project that has integrated the data from >100 avian point-count studies encompassing thousands of point-count surveys, and harmonized across data sets to account for heterogeneity induced by methodological and other differences. The BAM project faced the classic data-management challenges any LTEM must deal with, as well as special challenges involved with harmonizing so many disparate data sources. We created a data system consisting of 4 components: Archive (to preserve each contributor's data), Avian Database (harmonized point-count data), Biophysical Database (spatially explicit environmental covariates), and Software Tools library (linking the other components and providing analysis capability). This system has allowed the project to answer many questions about boreal birds; we believe it to be successful enough to merit consideration for use in monitoring other taxa. We have learned a number of lessons that will guide the project as it moves forward. These include the importance of creating a data protocol, the critical importance of high-quality metadata, and the need for a flexible design that accommodates changes in field techniques. One of the challenges the BAM team faced—gaining access to relevant data sets—may become easier with the increased expectation by journals and funding agencies that documenting and preserving research data be a standard part of scientific research.