Managing wildlife populations often requires a lot of data. Scientists need information on things like the spatial distribution of a population, how individuals move, and birth and death rates. However, it can be difficult to make these estimates because of the limited resources available to collect data across large regions, ranges of movement, and the long time periods over which population dynamics occur.
Citizen science data, collected by individuals outside of the science community, has a chance to help address this problem. To alleviate the issue of sparse data in inferring wildlife population patterns and processes, authors of a recent publication in Ecology developed an integrated population model (IPM) that, for the first time, incorporates citizen science data. Generally, IPMs unite multiple sources of data collected from independent studies in order to estimate parameters of a focal population. This allows for multiple datasets to inform the estimation of parameters more precisely and accurately than with only one dataset. However, IPMs have been limited to the generally small amounts of detailed sampling that researchers alone can accomplish. Now, the framework of the new IPM incorporates species-level data collected opportunistically from citizen science with individual-level data collected systematically, thus expanding the extent of inference because volunteers and members of the public can collect data over much larger spatial regions and longer timeframes.
In the paper, authors conducted simulations to evaluate how estimates of population abundance and survival, recruitment, and population growth rates can improve with varying additional amounts of opportunistic, citizen science data. The authors also applied the IPM to a case study on American black bears in southeastern New York, USA in which limited spatial capture-recapture sampling were supplemented with trail camera data submitted by citizen scientists.
The authors found that adding citizen science data increased the precision and accuracy of parameter estimates. Importantly, estimates in some seasons of multi-season simulations were as precise and unbiased as when only sparser, but demographically more informative, individual-level data were available.
Given how cost-effective citizen science data can be to collect, these findings suggest that long-term population monitoring may benefit from employing both systematic and opportunistic sampling methods, and collecting more opportunistic data and less systematic data. Furthermore, in some cases, increasing the number of opportunistic sampling locations improved abundance estimates more so than increasing their detection probabilities, providing important initial guidance for allocating effort in citizen science sampling.
Opportunistic datasets are increasingly available. The new IPM provides a first example and framework for how to expand the spatiotemporal extent of population sampling and inferences beyond what has been traditionally possible for identifying patterns and processes important to wildlife population management and conservation.