River Waves: A New Framework Turns Flow Data Into Ecological Predictions

Rivers are among the most altered ecosystems on Earth, and their biodiversity is collapsing under the combined pressure of dams, water abstraction, pollution and climate change. For decades, ecologists have relied on a handful of celebrated concepts to explain how river ecosystems work, but those ideas have never quite fit together. Now, a team of researchers led by Henry H. Hansen of Karlstad University in Sweden has published, in the journal Nature Water, a quantitative framework that promises to change that. By treating a river’s discharge as a physical wave and mapping the area and shape of that wave onto established ecological theory, the framework converts the routine flow measurements collected daily at thousands of gauging stations worldwide into concrete predictions about which ecological processes dominate a river at any given moment, and how those processes might shift under restoration or continued fragmentation.

The intellectual problem the framework addresses is a long-standing one. Three foundational concepts have shaped river ecology since the late twentieth century. The river continuum concept, proposed in 1980, describes a river as a longitudinal gradient of energy and materials flowing from headwater source to mouth. The flood pulse concept, articulated in 1989, emphasizes the lateral exchange between a river channel and its floodplain during overbank flooding. The riverine productivity model, from 1994, focuses instead on local, in-stream production as the dominant energy source. Each concept was developed for particular kinds of rivers, such as temperate, near-pristine or forested headwater streams, and each assumes a particular dimension of the ecosystem, whether longitudinal, lateral, vertical or temporal. As a result, they are conceptually disjointed, and it has remained unclear how to apply any of them to the majority of the world’s rivers, whose natural flow regimes are now altered by dams and other infrastructure.

The River Wave Concept, introduced by Paul Humphries and colleagues in 2014, was designed as a unifying metatheory. Its central insight is that these three concepts are not competing explanations but distinct functional states that occur at different positions on a hydrological wave. When discharge is reframed as a wave defined by its intensity, amplitude, wavelength and frequency, the area under a hydrograph indicates which concept best describes the river’s dominant processes, while the shape of the hydrograph indicates the potential ecological response. During baseflow conditions at wave troughs, low discharge confines energetic pathways to localized, in-stream processes dominated by autochthonous production, which is the domain of the riverine productivity model. As discharge rises within the banks, longitudinal transport accelerates and upstream-derived materials move downstream, matching the river continuum concept. At wave crests, overbank flows laterally reconnect the channel with its floodplain, and the flood pulse concept takes over. Until now, however, the idea remained largely theoretical because no quantitative tools existed to map it onto real-world hydrological data.

The new framework closes that gap with a data-driven workflow. It takes multi-year discharge records from a gauge station and statistically estimates the discharge breakpoints that separate the three conceptual states. The delineation rests on flood-frequency analysis: any discharge exceeding the mean annual flood, equivalent to a 2.33-year recurrence interval under the theory of largest values and conservatively rounded to a two-year interval in dammed systems, marks the onset of flood pulse conditions. The boundary between the riverine productivity model and the river continuum concept is set at the median discharge, approximating a one-year recurrence interval. The team implemented these calculations with a Bayesian flood-frequency approach that fits a log-normal distribution to flood events. In parallel, the framework extracts wave metrics from the shape of the hydrograph, including peak flood timing, baseflow consistency and flashiness, and links them to organism patterns across space. A joint analysis combining both streams, using cluster analysis, then synthesizes the outputs into management-relevant river classifications.

To demonstrate the approach, the researchers applied it to 165 Swedish gauging stations, drawing on observed fragmented flow records from the Swedish Meteorological and Hydrological Institute and on modelled free-flowing regimes generated by the S-HYPE hydrological model. Three index rivers illustrate the range of patterns. The Vindelälven, a large northern river that remains nearly free-flowing, shows the signature of a snowpack-driven water year, with stable winter discharge and regular, predictably timed spring floods. The Mörrumsån in the south, where a main-stem hydropower plant was removed in 2021 but upstream dams still operate, shifts between riverine productivity model conditions and floodplain-associated conditions that vary in both magnitude and timing. The Rönne Å, regulated for hydropower, displays yet another pattern. In each visualization, every day of the record is colour-coded by its dominant concept, revealing how the timing of ecological processes varies within and among years, with cascading consequences for the rest of the ecosystem.

Comparing the fragmented and free-flowing datasets across all 165 stations revealed systematic shifts in ecosystem functioning. The riverine productivity model was more common under modelled free-flowing conditions, particularly during the warmer spring and summer months, suggesting more local production and nutrient inputs. Conditions associated with the flood pulse concept and the river continuum concept were more frequent under the observed fragmented regime, indicating greater contributions from riparian zones and upstream reaches. The timing of these shifts varied seasonally: the day-of-year range of roughly 100 to 200 showed inverse relationships between concepts, while the end of the water year, beyond day 300, showed the most similar conditions between the two datasets. The beginning of the water year showed the greatest variation, suggesting river conditions then are not biased toward any particular theory. Regression models fitted to the concept proportions had adjusted R-squared values ranging from 0.61 to 0.73, indicating that a substantial share of the variation in free-flowing proportions is explained by the fragmented observations.

The wave metrics told an equally striking story about fish. The team linked three metrics to the life history strategies of Winemiller and Rose: equilibrium strategists, with high juvenile survivorship, thrive under stable flows; periodic strategists, with high fecundity, depend on regular, predictable floods; and opportunistic strategists, with short generation times, exploit highly disturbed environments. Baseflow consistency was measured as the skewness of the baseflow index, peak flood timing as the coefficient of variation of annual flood dates, and flashiness as the average Richards–Baker flashiness index. Ternary plots of these metrics showed that under modelled free-flowing conditions, Swedish rivers cluster tightly: consistent flood timing between days 80 and 100, moderate baseflow consistency and modest flashiness, predicting a spatially uniform balance of periodic and equilibrium fish with little room for opportunists. Under observed fragmented conditions, by contrast, the patterns split strongly by latitude and catchment size, with periodic fish favoured in the north, equilibrium fish in the south and opportunists potentially colonizing a few smaller, flashier catchments. The authors suggest that the reduced suitability for periodic species such as salmon could help explain their decline in the country.

Perhaps the most surprising result came from the joint cluster analysis. For the observed fragmented regime, the hierarchical clustering, validated by gap statistics and k-means approaches, identified three optimal river types nationwide. For the modelled free-flowing regime, only two clusters emerged. In other words, restoring all of Sweden’s rivers to free-flowing conditions would make the country’s rivers more similar to one another, not more heterogeneous, challenging the common assumption that free-flowing rivers are naturally more varied than fragmented ones. The authors note that Sweden’s granite-dominated boreal landscape probably limits natural variation in river characteristics, and that a small subset of rivers retains similar wave characteristics regardless of fragmentation. Most rivers, however, would merge into shared wave profiles once connectivity returned. This finding arrives at a critical moment: Sweden is undergoing a nationwide hydropower-relicensing process, and the European Union’s biodiversity strategy for 2030 calls for restoring 25,000 kilometres of free-flowing rivers. The framework suggests managers should anticipate one class of rivers that barely responds to restoration and another that changes substantially yet converges with the rest of the landscape.

The authors are careful to stress that their predictions are hypotheses awaiting field validation. Relating the framework’s outputs to real fish observations over time will be essential to confirm that the wave metrics genuinely capture the ecological responses they are meant to proxy, and the framework’s assumptions should be tested against other stressors such as drought, water abstraction and non-point-source pollution. Future extensions could incorporate landscape data such as digital elevation models to refine the discharge-state relationships, track individual water years rather than long-term averages, and expand the approach to other biomes as gauging networks grow worldwide. Still, the significance of the work is hard to overstate. It offers a scale-independent tool that translates the most routinely collected hydrological data on Earth into a unified, testable picture of river ecosystem dynamics, finally bridging the gap between a century of river theory and the practical decisions that will determine whether the world’s rivers recover or continue to unravel.

Subject of Research: A data-driven framework unifying river ecosystem concepts through hydrological wave analysis to predict ecological change in rivers

Article Title: A River Wave framework to unify river ecosystem concepts and predict ecological change

Article References: Hansen, H. H., Erickson, J., Lindmark, M., McPhan, L., Humphries, P., & Bergman, E. (2026). A River Wave framework to unify river ecosystem concepts and predict ecological change. Nature Water. https://doi.org/10.1038/s44221-026-00721-x

Image Credits: AI Generated

DOI: 10.1038/s44221-026-00721-x

Keywords: river ecology, River Wave Concept, hydrology, flow regime, river continuum concept, flood pulse concept, riverine productivity model, fish life history, dam fragmentation, river restoration, Sweden, Nature Water