1. Introduction
Quantifying and understanding catchment sediment yields (SY , t km−2 y−1; i.e., the amount of sediments exported from a river system per unit of time and catchment area) has been a central research theme for many decades. Over this time, it became increasingly clear that humans have a strong and rapidly growing impact on SY (Walling and Fang, 2003; Syvitski et al. , 2005; Montgomery, 2007; Borrelli et al. , 2017). Quantifying and understanding these impacts is not only a key research challenge in hydrology and fluvial geomorphology (Hoffmann et al. , 2010; Tarolli, 2016; Poesen, 2018) but also of great societal/economic relevance and necessary to fully understand anthropogenic impacts on carbon fluxes and the Earth System as a whole (e.g., Oost et al. , 2007; Galy et al. , 2015).
From earlier work, it became evident that human impacts on sediment fluxes are highly scale-dependent (Walling, 1988; Walling and Fang, 2003). Hillslopes and small catchments typically demonstrate a much stronger and faster response to land-use changes than larger river systems (Dearing et al. , 2006; Montgomery, 2007; Montanheret al. , 2018). This is also apparent at a global scale: estimates of impacts of human land cover changes on hillslope erosion rates (Oostet al. , 2007; Borrelli et al. , 2017) are about ten times larger than the corresponding impact on the sediment flux to the oceans (Syvitski et al. , 2005). While the general mechanisms explaining this scale-dependency are known (e.g. Trimble, 1999), its characteristics and relation to other environmental factors remain poorly quantified and comprehended (Dearing et al. , 2006; Hoffmann et al. , 2010; Tarolli, 2016; Poesen, 2018). Nevertheless, understanding the sensitivity and scale dependency of river systems to human disturbances is crucial for effective catchment and land management strategies (Trimble, 1999; Vanmaercke et al. , 2011; de Vente et al. , 2013; Poesen, 2018) and strongly links to ongoing debates about the role of (historic) land use as the primary driver of soil erosion and land degradation in the Mediterranean and other regions (Cox et al. , 2010; García-Ruiz, 2010; Dusaret al. , 2011; Vanacker et al. , 2014). Detailed catchment sediment budget studies or lake sediment analyses can provide valuable insights into the history and degree of human impacts on SY(Trimble, 1999; Dearing et al. , 2006; Hoffmann et al. , 2010; Dusar et al. , 2011; Golosov et al. , 2021; Ivanovet al. , 2021).
However, the human factor is not the only one controlling the sediment flux — variations in SY are driven by a wide range of natural factors, including geomorphic, tectonic, climatic, and biotic factors (e.g. Syvitski and Milliman, 2007). For example, it was shown that interactions between lithology and seismicity could exert strong controls on erosion and catchment denudation rates via the effect of rock fracturation (Molnar et al. , 2007; Portenga and Bierman, 2011; Vanmaercke et al. , 2014b, 2017). While some recent works revealed no significant climatic impact on natural SY (Vanmaerckeet al. , 2014a), this factor will likely be more relevant at a temporal scale for mountain regions (Carretier et al. , 2013; Jeffery et al. , 2014).
To deepen the understanding of the impacts of land use and climate change on sediment load, we explore mechanisms of the suspended sediment yield formation in the Northern Caucasus during the Anthropocene (Waterset al. , 2014, 2016).
The collapse of the Soviet Union in 1991 has led to significant land reforms in Russia (Ioffe et al. , 2004; Golosov et al. , 2018), including the Caucasus region (Hartvigsen, 2014). Agricultural land previously owned by the State and used for large-scale farming was privatized in the early 1990s by distributing ownership rights to large state farms among former collective farm members (Hartvigsen, 2014). Overall, the economy’s restructuring has led to agricultural land abandonment in the former Soviet Union (Lesiv et al. , 2018). Recent studies show that Northern Caucasus has cropland loss byca. 8% in 2015 compared to 1987 (Buchner et al. , 2020). However, common for former Soviet Union forest recovery on abandoned agricultural fields (Griffiths et al. , 2014) has resulted in forest gain of 6% in the Northern Caucasus (Buchner et al. , 2020).
The Northern Caucasus has experienced climate changes over the past decades, with summer temperature increased by 0.5-0.7°C over the past 30 years (Toropov et al. , 2019). The recent warming over the Caucasus Mountains has substantially impacted the glaciers, leading to losses at an average of 0.46% of the glacierized area per year (Tielidze and Wheate, 2018). However, the water cycle’s associated intensification did not cause any significant change in the precipitation regime (Toropov et al. , 2019). Recent studies on the North Caucasus rivers (Rets et al. , 2020) show that June runoff increased by 1.1-9.1% per decade for the last 70 years, while August’s runoff from highly glacierized catchments has decreased by 1.0-6.3% per decade. In contrast, the August runoff of non-glacierized catchments has increased by 1.5–11.5% per decade. With the growth of interest in environmental change over the Caucasus Mountains (Tielidze and Wheate, 2018; Toropov et al. , 2019; Rets et al. , 2020), it is crucial to consider the extent to which sediment flux is changing, as an important index of the functioning of the earth system mainly in response to landuse and climate changes (Walling and Fang, 2003).
Based on previous studies relating to sediment flux response to climate change (Walling and Fang, 2003; Li et al. , 2020; Zhang et al. , 2020), we hypothesized that in mountain and high-mountain catchments of the Northern Caucasus, the suspended sediment discharge values (SSD , [kg s−1]) have been decreasing since the beginning of the Anthropocene in ca. 1945. To test this hypothesis, we analyzed catchment suspended sediment yields calculated using observed hydrological data. We explore how sediment flux of various river basins with different land-use/landcover and glacier cover changes over time. We mainly focused on small catchments (A < 103 km2) as they are typically more sensitive to human impacts (Walling, 1983; Dearing and Jones, 2003; Vanmaercke et al. , 2015) and climatic change (Oswoodet al. , 1992; Moore et al. , 2009).