Spatial models of observer intensity in citizen science

The mechanism by which we observe a phenomenon that unfolds in space and time fundamentally determines what we can learn about that phenomenon. Careful design of the observation mechanism is normally warranted, regardless whether we want to know about how a disease epidemic develops, how rush hour traffic in a city behaves, or how some plant species develops over its annual growth cycle. Volunteering citizen scientists are often involved as a measuring device (aka observation mechanism) because of their intrinsic longevity and the size of the volunteer group. There is, so to say, value in numbers. A fundamental problem with citizen scientists is the timing and location of their actions. This cannot always be steered rigorously from a scientific design. By nature of volunteering, we need to be grateful for their efforts when- and wherever these take place.

A classical case of this type presents itself with the citizen science portal of waarneming.nl. This portal forms the web presence of a large community in The Netherlands and Belgium of observers of natural life, who report on findings of plants and animal of any type. This portal has shown a steady growth in volume of observations, culminating in 7.5 million observation in 2018 for The Netherlands. This is an unprecedented total for a country of the size of The Netherlands, which has resulted in an incredible spatiotemporal observation density.

Without clear understanding of the behaviour of observers seen as a single measuring device, making sense of the data that comes out of those observations is next to impossible. This is the overall motivation for this work. One cannot expect to find observations in places or at times when no observers are present, obviously.

In an ongoing MSc thesis project, the candidate is constructing an explicit space/time model of observer density that aims to account for spatial factors such as land use, temporal factors such as the weekday/weekend/holiday distinction, and spatiotemporal factors such as weather conditions. That model uses as smallest spatial unit 1 km^2.

In the project proposed here, we are keen to find out whether that resulting model also applies outside of The Netherlands. Specifically, we are keen to study the case for Fenno-Scandinavia, which has societal conditions not unlike those of The Netherlands. In these countries, somewhat similar citizen science portals exist for natural history, though because of scale and population density differences there also exist serious differences.

In the project, the candidate will study ways to aggregate spatially, temporally and spatiotemporally the Dutch model. The project will also accumulate data from the foreign portals that determines observer intensity, and it will address building up background data to capture explanatory factors as used in the Dutch model for the country/-tries under study. An evaluation will take place whether the Dutch model applies elsewhere, and if it seems not so, explanation will be sought for differences observed.