Integrating remote sensing and river biogeochemistry to understand catchments controls of riverine carbon and nutrients dynamics for a sustainable water resource management in Kenya

Land use and its cascading effects on riverine biogeochemistry represent one of the most pressing challenges in freshwater science, yet the mechanistic linkages between catchment-scale land cover dynamics and downstream carbon and nutrient fluxes remain poorly constrained in tropical African river systems. This study integrates remote sensing derived land use data with intensive wet and dry season field based river biogeochemical sampling covering inland (Thiba, Chania Sagana) and coastal (Ramisi and Mkurumudzi) catchments in the Kenya. Increasing agricultural land cover was significantly and positively associated with elevated nitrate (NO₃⁻) concentrations, dissolved inorganic carbon (DIC), total suspended matter (TSM), and DIN:SRP molar ratios. The DIN:SRP ratio increased significantly with agricultural cover in inland catchments, with Thiba; having more agricultural and settlement cover consistently exceeding the Redfield ratio of 16:1, indicating a systematic shift towards nitrogen surplus and potential phosphorus limitation of instream primary production. Stable isotope analysis of particulate nitrogen (δ¹⁵N-PN) and streambed sediments (δ¹⁵N-sediments) provided convergent isotopic evidence of agriculturally derived nitrogen enrichment in inland catchments across both seasons, while coastal catchments showed weaker and more seasonally variable isotopic responses. The carbon isotopic composition of particulate organic carbon (δ¹³C-POC) declined significantly with increasing tree cover across inland catchments during the wet season, confirming that forested sub-catchments deliver isotopically light C3 terrestrial organic matter to river channels, while open agricultural areas contribute relatively ¹³C-enriched POC consistent with C4 crop inputs. Carbonate and silicate weathering, traced through positive TA–Ca+Mg and TA–Si relationships, emerged as the primary control on total alkalinity and DIC across inland catchments compounded with soil derived CO2 that was evidenced in the inverse correlation of DIC concentration with δ13C-DIC. Coastal catchments; Ramisi in particular, exhibited pronounced dry season dominance in carbonate weathering signals, reflecting concentrated rock-water contact under baseflow conditions, while wet season dilution from soil derived CO2 weakened these relationships. Riparian buffer of tree cover (1km-100m) was the most critical local-scale moderator of TSM levels while the catchment scale of agricultural land use best predicted the TSM levels in the rivers. Together, these findings demonstrate that integrating remote sensing-derived land cover metrics with multi-seasonal river biogeochemical data provides a powerful framework for disentangling the relative contributions of agricultural intensification, riparian vegetation, hydrology, and lithology to riverine carbon and nutrient dynamics for a sustainable water resource management in Kenya.