Keeping an eye out – How we “observe” biodiversity from space


Biodiversity is a complex term but it essentially encompasses life in all its variety, ranging from individual genes to entire ecosystems. The legendary biologist Edward O. Wilson defines biodiversity as “the totality of all inherited variation in the life forms of Earth”. The loss of biodiversity due to human activities and its negative effect on ecosystems are well documented. The United Nations Sustainable Development Goal 15 is, among other things, focused on halting biodiversity loss, and the Convention on Biological Diversity has developed five Strategic Goals and twenty Targets to protect and conserve global biodiversity. Reliable and regular monitoring is necessary in order to understand the what, how, when and why of biodiversity loss. This is essential both for preventing further biodiversity losses and restoring it to healthy levels.

Satellites remote sensing is a powerful tool in this regard and has a long history of being implemented in biodiversity conservation. However, it’s only been in the last decade or so that satellite applications in biodiversity research have really accelerated due to a confluence of factors. One factor is the development of essential biodiversity variables (EBVs) that support the objectives of the Convention on Biological Diversity and from the beginning satellite remote sensing was expected to be a critical tool for the derivation of EBVs. A second factor is the launch of several new satellites with varying complementary capabilities to map and characterize the surface of the Earth.

Our research group at Lund University’s Centre for Environmental and Climate Science and our collaborators have been taking advantage of these new resources. For starters, it is well-known that increased intensification of land use negatively impacts biodiversity through, for example, increased agricultural inputs like fertilizers and the use of pesticides, herbicides and fungicides. But, monitoring land-use intensity is a different matter entirely and there is no efficient way to do that. To help bridge this gap, we designed a study that explores whether it is possible to link agricultural land-use intensity and biodiversity, and potentially monitor how the former affects the latter using satellite technology.

We use data from the Sentinel-2 satellites to create a suite of indicators that provide us with information about cereal crops in conventional and organic farms in the southern Swedish region of Scania. Among other things, these indicators include the timing of crop growth and how productive each field is in terms of biomass. Because each Sentinel-2 pixel is 10 meters long and wide, we were able to create indicators for each field across all our sites in Scania. We found through statistical testing that the satellite indicators were strongly correlated with cereal yields reported by farmers as well as with field-collected data on biodiversity (insect pollinators and flowering plants). We also found that conventionally-managed farms had an average of 15% higher productivity than organic ones. Of course, the fact that conventional farms are more intensely managed than organic farms is well-known, but it is good that we can capture this difference with satellite data because it allows us to monitor large areas more efficiently.

Another interesting finding of our study is that our field-collected biodiversity data were strongly linked to the portion of satellite indicators that were related to crop productivity, which are in turn linked to land-use intensity. Simply put, our results suggest that we have the potential to monitor the health of the agricultural ecosystem across large areas at a level of detail that is ecologically relevant even for small-scale organisms such as insect pollinators. We hope that we put on a convincing case that the data provided by the Sentinel-2 satellites really has the potential to provides us with information that helps make biodiversity assessments more efficient.

Ta del av information om behandlingen av dina personuppgifter