In a cluster of 48 artificial stream channels at The Arboretum in Flagstaff, Arizona, researchers spent two years feeding isotope-labeled leaves to flowing water and watching where the carbon went. The answer, published in the journal Ecosphere by lead author Tanner J. Williamson and colleagues, was troubling: as water temperatures climbed, microbes broke down the leaves faster but wasted more of the carbon as CO2 instead of converting it into biomass. Less microbial biomass meant less food for caddisflies, the shredding insects that form a critical link between dead leaves and the trout, salamanders, and birds higher up the food chain.
“We could literally trace the carbon atoms from the leaf to the microbe to the insect or out into the air,” said Williamson, an ecologist at the time based at Northern Arizona University. “What surprised us was how consistently warming tipped that balance toward the atmosphere.”
The experiment offers some of the most direct evidence yet that warming rivers do not just stress aquatic life through heat alone. They also rewire the basic energy economy of freshwater ecosystems, diverting carbon away from living tissue and into the atmosphere.
A controlled test with an isotope tracer
The Flagstaff team designed their stream chambers to mimic the conditions of a natural mountain creek: flowing water, real sediment, and native microbial communities colonizing the leaf packs. The key variable was temperature, raised in controlled increments across the 48 replicates. By tagging the leaves with a rare carbon isotope, the researchers could trace every milligram of carbon to its destination: microbial cells, caddisfly tissue, or respired CO2.
The results showed that warming reduced what ecologists call carbon-use efficiency, the fraction of consumed carbon that microbes lock into their own bodies rather than exhale. When efficiency drops, the math is simple but consequential. More carbon leaves the stream as a greenhouse gas. Less carbon is available to feed the invertebrates that underpin the food web. Over two years and dozens of replicates, the pattern held.
River heatwaves are already intensifying worldwide
The temperature increases tested in Flagstaff are not hypothetical. A separate global analysis published in Nature Communications documents that river heatwaves, periods when water temperatures spike well above historical norms, are growing more intense, longer-lasting, and more widespread. The study examined roughly 3,000 river sites across six continents and projects that under continued greenhouse gas emissions, up to 4.4 billion people living near rivers could face increasingly severe thermal extremes in the waterways they depend on for drinking water, irrigation, and fisheries.
“The scale of the problem is what gets overlooked,” said Gabriel Singer, an aquatic ecologist at the Leibniz Institute of Freshwater Ecology and Inland Fisheries, who was not involved in the Flagstaff study but has researched carbon cycling in rivers. “Inland waters cover only about 1 percent of Earth’s surface, yet they emit roughly 2 to 3 gigatons of carbon per year. If warming pushes that number higher, it matters for global budgets.”
Earlier experimental work reinforces the biological stakes. A foundational 2009 study in PLoS Biology demonstrated that warming shifts the metabolic balance of food webs, giving heterotrophs (organisms that consume organic matter) an advantage over primary producers. When that balance tips, the entire architecture of who eats what in a stream can change.
What scientists still do not know
The Flagstaff experiment was designed to isolate temperature as a variable, and it did that well. But real rivers are messier. Sediment type, flow regime, organic matter supply, and the particular species that process leaf litter all vary enormously from one watershed to the next. A Nature Communications study on streambed sediments found that greenhouse gas production in response to warming depends heavily on grain size, organic content, and underlying geology, and that responses can be nonlinear. In some cases, small temperature increases triggered disproportionately large CO2 releases once certain thresholds were crossed. Whether the efficiency losses measured in Flagstaff’s controlled chambers scale up the same way in sediment-rich lowland rivers or bedrock-dominated mountain channels remains an open question.
No published study has yet connected observed river heatwave events directly to measured changes in CO2 emissions at the watershed scale. The global heatwave projections and the laboratory decomposition results point in the same direction, but the bridge between them is built on inference, not integrated field data. Likewise, while oxygen loss and warming almost certainly interact (lower oxygen can push microbial communities toward anaerobic pathways that produce methane alongside CO2), no study in the current evidence base quantifies that interaction in a natural river setting.
Geography matters, too. The Flagstaff work focused on organisms and conditions typical of temperate mountain streams. Tropical rivers, Arctic-fed systems, and lowland floodplains each host different decomposer communities that may respond to warming on different curves. Expanding the experimental approach to those settings is a clear next step.
Why it matters for river management
Inland waters already play a larger role in the global carbon cycle than their surface area suggests. Rivers, lakes, and wetlands collectively transport and transform vast quantities of carbon moving from land to ocean and atmosphere, a point established by large-scale analyses over the past decade. The Flagstaff results suggest that as rivers warm, they may become even stronger net sources of CO2, creating a feedback loop: climate change heats streams, heated streams release more carbon, and that carbon accelerates further warming.
For the people who manage watersheds, the findings sharpen the case for interventions that are already well understood but inconsistently applied. Riparian buffers of trees and shrubs can shade smaller streams and moderate temperature spikes. Protecting groundwater inputs helps buffer thermal extremes by mixing cooler subsurface water into surface flows. Reducing nutrient pollution limits the algal blooms that drive nighttime oxygen crashes, which become more dangerous as water warms. None of these measures will reverse global temperature trends, but they can preserve pockets of thermal and biological resilience within river networks.
From lab channels to living watersheds
The next phase of research will need to move from controlled chambers to whole watersheds, tracking carbon, temperature, oxygen, and food-web productivity together in real time. That work is technically demanding and expensive, but the Flagstaff experiment has laid a clear foundation: warming shifts carbon out of food webs and into the air, and the mechanism is measurable.
For communities that depend on healthy rivers for drinking water, recreation, and fisheries, the practical takeaway as of April 2026 is straightforward. Thermal pollution and climate-driven warming are not background conditions to monitor passively. They are active forces reshaping how rivers function, from the microbial films on a submerged leaf to the trout that rise in a cold-water pool. Managing for that reality is becoming as much a climate issue as it is an ecological one.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
