Impacts of post-Soviet land-use transformation on sediment dynamics in the Western Caucasus

• Riverine sediment transport dynamics over the last five decades in the western Greater Caucasus were investigated. • This study focuses on the Krasnodar Reservoir, and using repeated bathymetric surveys and gauging station measurements. • A progressive increase in erosion rates from 1975 to 1993–2005 was followed by a decline that persists to the present day. • The collapse of the USSR and subsequent land-use changes were identified as primary drivers of sediment flux alteration. • The Top-Kriging modelling was used to extend sediment records and predict in ungauged basins, demonstrating good accuracy. This article investigates sediment transport dynamics in the western Greater Caucasus region over the last 50 years, focusing on the Krasnodar Reservoir, the largest artificial reservoir in the Western Caucasus. Built between 1973 and 1975 on the Kuban River, the reservoir’s catchment has experienced significant hydrological and land-use alterations over the past several decades. Using two independent approaches — repeated bathymetric surveys and gauging station measurements — we analysed interannual sediment dynamics from the reservoir’s commissioning date to the present. We extended existing suspended sediment yield ( SSY ) records using a Top-Kriging modelling approach. Our findings revealed a gradual increase in erosion rates from 1975 to 1993–2005, followed by a decrease. This shift in trend aligns with the collapse of the Soviet Union, which led to significant land-use changes, particularly in arable and pastoral lands. Statistical analysis indicated that precipitation patterns played a secondary role in these dynamics. The critical period for changes in sediment dynamics appears to be between 1995 and 2000, coinciding with major alterations in cropland area. Additionally, this study highlights the feasibility of predicting sediment yield at a semi-regional scale using limited yet representative datasets. The Top-Kriging modelling technique effectively predicted SSY in ungauged basins, with good accuracy ( NSE = 0.63 ) achieved not only during cross-validation but also when compared with independent reservoir sedimentation records.