On the suitability of ecological in-situ networks for detecting impacts of extreme events

Abstract:

Anomalous hydrometeorological events have the potential to alter land-surface states. Satellite based Earth observations play an important role for detecting and describing the spatio-temporal development of extremes events. However, in-depth investigations of impacts on ecosystem functioning typically requires in-situ observations. The growth of ecological in-situ networks raises the question if we are well prepared to analyze extreme events in today and in the near future. Here we explore the extreme event monitoring capacity of in-situ stations of land-atmosphere flux observations, soil moisture, and other relevant data streams (e.g. Ameriflux, FLUXNET, NEON and ISMN) as well as a range of hypothetical random networks. We focus on extreme events reveled by different Earth observations over the US and Europe. Our results show that small random networks of less than 100 sites per continent would have detected most large-scale extreme events over the past decade, while many small events would have remained unseen. Very large networks with up to 10000 sites would have only moderately improved the detection efficiency, but become several orders of magnitude more efficient in detecting small events. Another fining is that spatial clustering of real-world networks leads to lower detection potentials for extremes less compared to random networks. However, the bottom-up design of NEON could slightly improves the detection efficiency. Further expansions of ecological networks should carefully consider the spatial occurrence probability of extreme events.