Linking soil erosion processes with arable weed seedbank dynamics to inform sustainable cropping

Accelerated soil erosion affects sustainable food production through the degradation of arable soils resulting in lower crop yields and compromising biodiversity. Over the past 50 years, the weed seedbank has been declining due to farming intensification, increased herbicide use and weed suppression through competitive autumn crop planting. However, there is less recognition of the potential of soil erosion affecting the weed seedbank. This thesis contributes to an improved understanding of the effect of geomorphological processes (soil erosion) on biological systems (weed seedbank) in arable ecosystems. The first investigation assessed whether the management of farm machinery field tramlines would decrease soil erosion rates and effect the movement of weed seeds. Over three winter seasons, eroded material was collected by a network of Gerlach Troughs. The results showed that tramline management with a spiked harrow decreased soil and seed loss by 93.9% and 86.56% respectively, compared to regular tyre tramlines. Analysis of seed data to runoff and sediment load found seeds were transported along with sediment (r²=0.62) rather than runoff (r2=0.2) over the long term. In addition, tramline management significantly affected the number of seed species transported (p<0.001), which was found to relate to seed morphologies. Overall, tramlines cause 0.01% - 0.32% seed fluxes annually depending on management. These findings have implications for farmers to protect tramlines from erosion and displace seeds through management thereby, preventing the loss of biodiversity within the field. The second investigation looked at the movement of weed seeds at the field scale by erosion through the use of a radionuclide (¹³⁷Cs) tracer. A single field was sampled for seedbank and soil cores taken for ¹³⁷Cs analysis in two sub field grids. The results indicated weak relationships between seedbank densities and erosion. The weak relationships in the grids (r² =0.13, p =0.029 in 2011 and r² =0.12, p = 0.036) were due to land management contributing to spatial variability within seedbank abundance and composition. Individual species showed mixed responses to erosion rates. The findings indicate farmers need to consider management strategies at field scale to effectively manage erosion and seedbanks because seedbank losses of between 2 – 2.5 % annually within the field which is linked to field scale sediment budgets. The third investigation looked at specific environmental controls that would affect soil erosion and seedbanks. This was achieved by using a portable rainfall simulator on plots containing either seeds from the natural seedbank or spiked with seeds. The key control was the presence of crop/vegetation cover in affecting erosion rates (p<0.001) and seed movement (p = 0.001). The presence of crop cover resulted in low erosion rates but a greater loss of seeds compared to plots with no crop cover. This was linked to vegetation cover providing a protective environment for weeds to grow and produce additional seeds via seed rain. Ground cover prevents erosion but also highlighted seed movement was higher than on bare soil due to a greater availability. This means that surface wash is more important than rainfall in causing seed transport. For spiked plots, more seeds were displaced in short (3 minute) events compared with long (6 minute) events (p = 0.04). This shows protecting the soil and seedbank from rainfall detachment is crucial to preventing transport of sediment and seeds that could enter other transport pathways (e.g. tramlines, rills, gullies). The fourth investigation looked into the processes and impacts of soil erosion on seedbanks at the catchment scale. This was done by establishing a monitoring station at the outlet of an arable catchment for one year to monitor discharge, suspended solids and seed flux. The results of monitoring in 2012 found seeds numbers were positively related to discharge (observed r² = 0.62, p<0.001; observed plus modelled r²=0.50, p <0.001) and sediment load (observed r² = 0.64, p<0.001; observed plus modelled r2= 0.89, p<0.001). Seed species had poor negative relationships with discharge (observed r² = 0.03, p=0.357; observed plus modelled r²=0.11, p = 0.017) and sediment load (observed r² <0.001, p=0.352; observed plus modelled r²=0.14, p=0.004). An initial estimate of losses from the catchment was around 0.008 – 0.027% of the weed seedbank. Interestingly, there appeared to be a trend in the abundance of seed collected relating to patterns of farming activity within the catchment. This finding has management implications as there is evidence, for the first time, of arable weed seeds being exported from the catchments, which could affect other agricultural land and ecosystems downstream. The findings of these four investigations showed that the effect of soil erosion on the seedbank is connected at different spatial scales. Scope for future work is to improve the understanding of the role of seed morphologies, land management and field scale processes affecting the transportability of seeds by erosion processes.