Lake Victoria’s Muddy Crisis: Sediment Cores Reveal a Seven-Fold Surge in Erosion

Beneath the surface of Lake Victoria’s Winam Gulf lies a meticulously preserved archive of human transformation, and scientists have now learned to read it with unprecedented precision. A new study published in Environmental Geochemistry and Health has combined lead-210 geochronology, high-resolution sediment geochemistry and geochemical source apportionment modelling to reconstruct more than a century of land-to-lake sediment transfer in western Kenya. The findings are stark: sediment accumulation in the gulf has accelerated dramatically since the 1960s, with the steepest rises occurring after the year 2000. In the Nyando catchment, sedimentation rates increased roughly seven-fold between the 1960s and 2021, while the Sondu-Miriu catchment recorded a three-fold rise over the same period. The research team, led by the British Geological Survey together with Kenyan and British partners, argues that these dated records finally link the timing of sedimentation change to its geographic origins, offering catchment managers a practical roadmap for intervention.

Lake Victoria sustains approximately 42 million people through fisheries, drinking water and agriculture, making the ecological trajectory of the basin a matter of profound regional consequence. Yet the lake has undergone substantial ecological change over the last century under the combined pressures of fishing intensity, land-use transformation and catchment degradation. The Winam Gulf, which receives discharge from five major river systems, has long been identified as a hotspot of land degradation. As early as 2006, the World Agroforestry Centre highlighted severe erosion and sediment delivery from surrounding catchments, warning that urgent action was needed to avert flooding and ecological harm. Despite community engagement programmes and policy initiatives in the intervening years, the new sediment record suggests those warnings went largely unheeded, with the most rapid degradation occurring in the past two decades.

The technical heart of the study is its dating framework. Sediment cores were collected at the mouths of the Nyando, Sondu-Miriu, Awach, Luanda and Kisat rivers, along a transect across the Nyando sediment plume, and from a reference site in the centre of the gulf. Chronologies were built using unsupported lead-210, calculated by subtracting supported lead-210, inferred from lead-214 activity under the assumption of secular equilibrium in the uranium-238 decay series. Ages and dry mass sedimentation rates followed the Constant Rate of Supply model using cumulative unsupported lead-210 inventories. Caesium-137, often used as an independent chronological marker of peak atmospheric fallout in 1963, proved unusable in these equatorial sediments because activities fell below detection limits, a common limitation in East African records owing to low fallout deposition and radioactive decay since peak weapons testing.

The resulting chronologies reveal a consistent inflection point in the 1960s, a period the authors associate with the transition from colonial to independent governance and accelerating land clearance for agriculture. In the Nyando system, sedimentation rose from roughly 0.1 grams per square centimetre per year in the 1960s to a peak of 0.717 grams in 2018, standing at 0.638 grams in 2021, an approximately 700 percent increase. The Sondu-Miriu climbed from 0.122 to 0.382 grams per square centimetre per year, peaking at 0.682 in 2007, coinciding with the commissioning of the first Sondu-Miriu hydroelectric power station. The smaller Luanda and Kisat rivers, which drain areas around the city of Kisumu, recorded 400 to 500 percent increases over the same decades, a trajectory the researchers link in part to the expansion of Kisumu’s urban footprint from 19 to 103 square kilometres between 1969 and 2019.

The geochemical record adds a second dimension to the story. Concentrations of phosphorus, sulphur, calcium and organic matter, estimated by loss-on-ignition at 450 degrees Celsius and measured by triple quadrupole ICP-MS after mixed-acid digestion, remained stable until the 1960s, dipped by 20 to 30 percent, and then surged after 1990, with phosphorus and sulphur nearly doubling within a decade in the Nyando core. Calcium showed a distinct step change around 2005, rising about 25 percent within a single year. Across nearly all cores, rare earth element distributions remained consistent, suggesting a largely stable mineral source, while the shifting chemistry of surface-reactive elements points to increasing mobilisation of agriculturally influenced topsoil. In practical terms, the lake is not simply receiving more dirt; it is receiving more of the nutrient-rich, carbon-bearing surface material on which both farm productivity and aquatic ecosystems depend.

To determine where this sediment was coming from, the team deployed a Frequentist sediment fingerprinting approach using the open-source FingerPro R package, version 2.0. A total of 318 composite riverbed sediment samples, each aggregated from eight to ten subsamples across the channel width, were collected from the Nyando, Sondu-Miriu and Awach catchments in a nested design. Conservative geochemical tracers were selected using a combination of conservativeness index, consensus ranking and consistent tracer selection criteria, and source contributions were quantified with linear variability propagation, a method designed to avoid the biases that non-linear mixing functions introduce under high source variability. The result is a time-resolved provenance reconstruction stretching from 1960 to 2020, effectively turning each dated core layer into a snapshot of catchment sediment supply.

The apportionment results expose how sharply sediment sources can shift in response to land-use change. In the Nyando catchment, contributions from the Nyando-Kipchorian sub-catchment rose rapidly between 2000 and 2005, while the historically erosion-prone Awach Kano and Nyaidho sub-catchment, previously the dominant source, declined in relative terms. Across the plume transect cores, the pattern appeared with a lag that lengthened with distance from the river mouth, tracing the progressive dispersal of material through the gulf. The timing aligns closely with satellite-derived land-cover data showing a 75 percent loss of woodland across the Nyando between 1985 and 2014, most of it between 1996 and 2000, alongside a 1022 percent expansion of urban area. The Tinderet Forest alone lost 26.6 square kilometres of tree cover between 2000 and 2020, a ten percent reduction concentrated within the very sub-catchments identified as rising sediment sources.

In the Sondu-Miriu and Awach systems, similar source shifts tell equally pointed stories. From 2005 onward, the Sondu-Miriu core recorded a growing proportional contribution from one sub-catchment, with the Yurith source trending toward 80 to 90 percent of delivered sediment by 2017, a pattern consistent with deforestation that has removed 32 percent of forest cover over six decades, most notably since 2000. In the Awach, the earlier 2000s sedimentation peak of 0.3 to 0.5 grams per square centimetre per year tracked rising inputs from the lower sub-catchments, where roads, settlements and farmland expanded, while the later 2010 to 2020 peak, reaching 0.9 grams, corresponded to intensifying pressure in the upper catchment, where steep slopes, higher rainfall, timber harvesting and agricultural conversion compound erosion risk. Independent modelling by other researchers reports sediment yield increases of 17 to 33 percent in Awach sub-catchments between 2018 and 2023, corroborating the core-based trends.

The implications reach well beyond academic curiosity. Accelerated sediment transfer represents the mobilisation of nutrient-rich topsoil, organic carbon and associated contaminants from productive landscapes into aquatic systems, threatening soil fertility upstream even as it drives eutrophication, altered nutrient cycling and declining fisheries habitat downstream. Previous work has estimated the cost of non-intervention against soil erosion at 390 million US dollars per year to Kenya’s economy, and the study’s stakeholder consultations identified fragmented governance, monitoring and environmental data as persistent barriers to effective management. The framework’s central promise is precision: rather than spreading scarce mitigation resources uniformly across entire basins, terracing, agroforestry, riparian buffer restoration and cover cropping can be concentrated in the specific sub-catchments shown by the sediment record to be disproportionate contributors. Combined with Earth observation, seasonal monitoring and erosion modelling, this integration of geochronology and fingerprinting offers a replicable template for adaptive catchment management, not only across the Lake Victoria Basin but in the many rapidly changing tropical catchments worldwide where the health of soil, water and food systems remains inseparably linked.

Subject of Research: Pb-210 geochronology and geochemical sediment source apportionment used to reconstruct historical land-to-lake sediment transfers in the Lake Victoria Basin

Article Title: Linking geochronology and source apportionment of sediments to understand land–lake transfers in the Lake Victoria Basin

Article References: Watts, M. J., Humphrey, O. S., Tuffield, L., Gowing, C., Marriott, A. L., Ongore, C. O., Isaboke, J., Osano, O., Blake, W. H., & Aura, C. M. (2026). Linking geochronology and source apportionment of sediments to understand land–lake transfers in the Lake Victoria Basin. Environmental Geochemistry and Health, 48(15), Article 597. https://doi.org/10.1007/s10653-026-03451-x

Image Credits: AI Generated

DOI: 10.1007/s10653-026-03451-x

Keywords: Lake Victoria, soil erosion, sediment cores, Pb-210 geochronology, sediment fingerprinting, source apportionment, Winam Gulf, Nyando catchment, deforestation, eutrophication, land-use change, catchment management