Introduction

Gully erosion is globally recognised as a significant source of soil loss and can directly influence the water quality of downstream aquatic ecosystems by the presence of fine suspended sediment (<63 µm) and suspended sand (63-2000 µm) (Vercruysse et al., 2017; Walling, 2006; Walling et al., 2016). The impact to water quality from gully erosion can vary depending on the particle size distribution and sediment yield. For example, small gullies that form in recently tilled agricultural land can affect local waterways (Capra et al., 2002), whereas larger landscape-scale gullies can impact the water quality of connected waterways for hundreds of kilometres downstream (e.g., >40% of the sediment polluting the Great Barrier Reef Marine Park, in the coastal waters of Northern Australia, is generated by gully erosion that occurs hundreds of kilometres upstream) (Brooks et al., 2013; Olley et al., 2013; Wilkinson et al., 2015).
Gullies are often formed by water flowing over land at a sufficient velocity to incise the soil and form a deep exposed gap, typically in areas with poor soil cohesion (Casalí et al., 2009). Over time the area of gullies expands due to active erosion caused by repeated flow events associated with intense rainfall (Poesen et al., 2003). Gully erosion is a natural process, however, the rate of erosion can drastically increase as a consequence of anthropogenic land use change (e.g., installation of roads and other infrastructure or through agricultural activities, such as land clearing and livestock grazing) (Nyssen et al., 2002; Wilkinson et al., 2018).
To date, suspended sediment monitoring in gullies has been conducted using methods designed for rivers and streams (e.g., automated discrete samplers (autosamplers), rising stage samplers (RS samplers), and turbidity loggers) (Baker et al., 2016; Bartley et al., 2017; Caitcheon et al., 2012; Nistor et al., 2005). However, gully systems present a unique set of hydrological and operational challenges for which these various sampling and measurement methods have not been thoroughly evaluated. The method considered to be the most accurate for measuring suspended sediment, flow-proportional manual sampling (Horowitz et al., 2008; Perks, 2014), is often infeasible or unsafe to conduct in gully systems due to the unpredictability of rainfall events, the remoteness of many gully landscapes, and the instability of gully channels and banks.
Currently, the operational challenges associated with measuring water quality in gullies are overcome using autonomous sampling or surrogate measurement techniques (autosamplers, RS samplers, and turbidity loggers) and remote sensing methods (time lapse cameras or light detection and ranging (LiDAR) techniques) (Casalí et al., 2009; Castillo et al., 2012). These methods are established monitoring techniques with well understood capabilities and limitations (Table 1). Many of these methods, however, are too expensive to implement over the large spatial network of actively eroding gullies within a catchment (e.g., autosamplers and turbidity loggers), and others provide incomplete information when deployed in isolation (e.g., RS samplers and time-lapse cameras). To address this gap, our study includes the recently developed Pumped Active Suspended Sediment (PASS) sampler; an automated, time-integrated, and in situ sampling device, as a low-cost approach that could be used to monitor gully erosion over large spatial scales (Doriean et al., 2019; Nunny 1985; Phillips et al., 2000). As this work is primarily focussed on approaches that measure water quality associated with suspended sediment (i.e., suspended sediment concentration and particle size distribution) remote sensing techniques (i.e., LiDAR) that are used to estimate soil loss using landscape scale volumetric analysis will not be included in this evaluation.
Here we aim to systematically evaluate and compare the capabilities and limitations of a variety of suspended sediment monitoring methods (i.e., autosampler, RS sampler, turbidity logger, and PASS sampler) when applied to gully systems. The methods are compared under controlled laboratory conditions and in the field to assess the relative ability of each method to provide accurate measurements of suspended sediment concentration and particle size distribution.