Climatic and anthropogenic controls on sediment-, flood-, and carbon dynamics in the Wei River Basin, Chinese Loess Plateau

Abstract

As a critical component of the Earth’s surface system, the formation and evolution of fluvial systems are affected by both internal factors, such as tectonic forces, and external drivers, such as climate change and anthropogenic disturbances. Human activities have profoundly altered fluvial systems, challenging their ecological integrity and long-term sustainability. Key anthropogenic stressors, including damming, land-use change and sediment mining, have significantly modified fluvial morphology, catchment hydrology, and associated ecosystem services. Understanding fluvial system trajectories under changing human pressures and climate changes has emerged as a key research focus in geomorphology. Although the magnitude and scope of anthropogenic impacts on river systems have been well-documented in recent centuries, the long-term (millennial-scale) imprint of human activities since early agriculture on catchment dynamics remains poorly quantified. The middle reach of the Yellow River, a cradle of Chinese agricultural civilization, provides an ideal setting to investigate long-term fluvial responses to climatic and anthropogenic forcings. This dissertation integrates landscape evolution modeling, machine learning, and historical land-cover reconstructions to quantitatively assess how millennial-scale land-use changes have reshaped the processes of water and sediment and thus geomorphological and ecological functioning of river catchments. This dissertation includes four scientific chapters. Chapter 2 and 3 develop a fluvial evolution model to simulate spatial-temporal trends in discharge and sediment load in the Wei River catchment. Building on these simulations, Chapter 4 and 5 evaluate human-induced geomorphic impacts on fluvial flooding and catchment carbon fluxes. To be specific, chapter 2 couples a landscape evolution model with an evapotranspiration model to simulate discharge and sediment load dynamics influenced by changes in anthropogenic land use and climate in the Beiluo River, a major tributary of the Wei River, from the mid-Holocene (8000 BP) to the early industrial era (100 BP). Model parameters are calibrated using modern climatic and hydrological data, and reconstructions of paleo-climate and anthropogenic land cover are served as inputs. By integrating these models, we quantify the relative contributions of climatic forcing and anthropogenic disturbances to long-term hydrological and geomorphological changes in the catchment. Furthermore, the simulated results enable an assessment of the accuracy and associated uncertainties in the fluvial evolution model, providing critical insights into model reliability and limitations. Chapter 3 extends the model framework to the entire Wei River catchment, incorporating spatial paleo-climate variability and crop-specific evapotranspiration effects. This chapter quantitatively analyzes the no-linear response of fluvial systems to climate change as a result of historical land-use modifications. In Chapter 4, we investigate the impacts of climate change, early agricultural practices, and intrinsic fluvial dynamics on paleo-flooding episodes in the Wei River catchment, as well as the downstream propagation of flood events from the Wei River to the lower Yellow River between 5000 and 100 BP. High temporal-resolution reconstructions (~20-year intervals) of precipitation, temperature, and historical land-use data are employed to capture decadal-scale variations in river morphology and soil moisture dynamics. These datasets enable the identification of dominant drivers behind flood episodes and its spatial pattern. In chapter 5, we couple the simulated spatial erosion and deposition (from Chapter 4) with machine learning-based reconstructions of soil organic carbon and net primary production dynamics to reconstruct carbon flux trajectories in the Wei River catchment over the past 5000 years. Five distinct carbon flux pathways (associated with erosion, transport, deposition, etc.) are quantified to assess the role of historical land-use change in modulating soil carbon cycling in the drainage system. The results demonstrate how the non-steady-state erosion regime, shaped by millennia of agricultural activity, determines whether this landscape functions as a carbon sink or source, with profound implications for global carbon cycling.