Time-series remote sensing and multi-continental field work reveals that wetlands in drylands can be robust indicators of climate change

Identifying reliable indicators of environmental changes is crucial for effective ecosystem management, particularly in drylands which are prone to climate change impacts. Here, we report on how we are integrating time-series remote sensing, advanced data science techniques, and ground-based observations to identify, map, and assess the sensitivity of a diverse suite of wetlands in drylands to environmental perturbations.  We are particularly interested in potential ‘sentinel wetlands’: natural features that are highly sensitive to subtle climatic changes. These wetlands may act as early warning systems, reflecting the cumulative effects of various climate stressors on their hydrodynamic state.

We have developed a method to automatically map different surface waterbodies (including a range of low- and high-altitude wetlands) and characterise their wetness dynamics at pixel-scale using time-series multispectral satellite data. We have applied the method to drylands spanning three different continents (western and northern India, southwest Spain, Argentinian Patagonia) and validated the mapped wetness dynamics of key features such as floodplain and valley-bottom wetlands, interdunal depressions, playas and pans through extensive field visits (~10 000 km of road trip).

From our field visits, we conclude that not all wetlands are good candidates for serving as sentinel wetlands. The best candidates are those wetlands which are devoid of direct human interventions, sit within endorheic catchments, and are relatively small in size (<10 km²). Each dryland visited hosts several such candidates. We classify these candidates in two categories: controls and targets. Controls are sentinel wetlands with in-situ hydrometeorological data logging stations (e.g. interdunal wetlands in Doñana National Park, Southwest Spain), while targets are the remaining sentinel wetlands that we plan to use as a distributed sensor array. Our field visits reveal that in some wetlands, there has been an increase in wetness frequency in recent years.  In the case of low-altitude wetlands, it is almost exclusively because of human interventions (i.e. these are non-sentinel wetlands) and in the high-altitude wetlands, it is because of increased glacier meltwater supply (i.e. these are sentinel wetlands).  By contrast, most sentinel wetlands in low-altitude regions are exhibiting reduced wetness frequency, in some cases dramatically. The next steps are to monitor and evaluate a wider set of hydrodynamic responses to stressors, including by tracking subtle changes at pixel scale and correlating these changes with local to regional climate.  The results will help further demonstrate how wetlands in drylands can act as robust indicators of climate change.

Knowing the wetness dynamics of sentinel and non-sentinel wetlands will help us to identify and separate the various climate and direct human stressors that might impact future water availability and hence water security in the world’s diverse drylands. This separation is crucial for developing targeted management strategies. By further characterising the sensitivity of sentinel wetlands, our research will enhance predictive models of waterbody responses to climate change and provide actionable insights for sustaining water resources amidst ongoing climate changes.