An analysis of satellite imagery of major river systems in the Philippines has revealed surprising insights into how rivers behave, with significant implications for river management in tropical settings.

An international research team has examined thousands of satellite images to quantify patterns and rates of river mobility, using big data to analyze how rivers change position over time.

In the study published in Nature Communications, researchers used Google Earth Engine to analyze more than 10,000 satellite images and were able to calculate “locational probabilities”—a measure of how often a river occupies specific locations over time. This approach offers a useful way to visualize and quantitatively define the space occupied by river channels over multiple decades.

Lead author Dr. Richard Boothroyd, from the University of Liverpool’s Department of Geography and Planning said, “In the era of big data, digital representations of rivers derived from multi-decade satellite imagery provide new opportunities to study how rivers adjust and evolve.”

“Previous studies have typically focused on just the wetted parts of rivers visible in satellite imagery. However, our approach is different—we’ve looked at the entire active width of rivers, including both water and unvegetated alluvial deposits, to get a comprehensive picture of how rivers adjust.”

The study identified “hotspots” of river mobility, which vary in intensity, size, and location across different river catchments. Rivers were characterized by zones of relative stability interspersed with zones of high mobility, but these patterns vary markedly between systems and even between different reaches of the same river, highlighting the idiosyncratic nature of river mobility.

The study was a collaborative effort between researchers from the University of Glasgow and Liverpool, Brunel University of London, University of Auckland and University of the Philippines Diliman.

Professor Carlos Primo C. David from the National Institute of Geological Sciences at the University of the Philippines Diliman added, “Our findings highlight the striking variability of mobility patterns, each uniquely shaped by the complex interplay of factors across scales.”

The team found they could not identify simple relationships between local factors (such as active channel width, valley floor width, and confinement ratio) and river mobility. No relation between channel pattern type (such as braided or meandering) and rates of adjustment was evident.

Dr. Boothroyd added, “These findings have significant real-world applications. In dynamic landscapes like the Philippines, where rivers pose significant hazards to communities and infrastructure, understanding patterns and rates of river mobility is crucial for developing sustainable management strategies that work with, rather than against, natural river processes.”

The researchers suggest their approach of using satellite-derived locational probabilities could transform how dynamic rivers are studied and managed worldwide, providing a “living database” that can be continually updated as more satellite data becomes available.