|dc.description.abstract||Grassland farming systems are an important contributor to the New Zealand economy, however recent expansions in dairying have lead to increasing concern about its effects on water quality. Phosphorus (P) and sediment losses from intensively farmed dairy pastures can impair surface water quality. Previous research has shown that loading from agricultural catchments is mainly from diffuse sources. However, an alternate hypothesis is that there are small areas within fields that can act as point sources during storm events, due to high concentrations of P in the soil from excreta. It is important to locate and understand source areas of P in agricultural catchments to effectively target mitigation strategies and decrease losses to surface waters. This thesis investigated the relationships between identified sources of P and sediment in grazed catchments, and the transport and transformation processes of P at different spatial scales and during different seasons.
One of the first steps in mitigating the loss of P and sediment was to determine where in a field the potential for P loss was greatest. Once the sources of P were found at the field scale (< 10 ha) the next step was to investigate the importance of the sources at the farm (10-500 ha) and catchment scale (> 500 ha). This study began by comparing P export in overland flow from grazed pasture with areas that receive elevated P inputs and stock traffic (e.g. gateway, water trough, stream crossing and cattle lane). Intact soil blocks were removed in a preliminary investigation of sources areas. Phosphorus loss from the sites was in the order: trough > crossing > gateway > pasture. Total P losses from the trough averaged 4.20 mg P/m² while the pasture exported 0.78 mg P/m². In addition, runoff from lane soil was measured with total P averaging 5.98 mg P/m²; however, the method used was different from the other soil blocks. The data suggested that locating and minimizing the size of these source areas in fields has the potential to significantly decrease P loss to surface waters.
To explore the importance of these sources at a greater scale, the soil physical properties and surface runoff from pasture, a laneway and around a watering trough, together with subsurface flows from pasture and catchment discharge, were measured in a small dairy catchment (4.1 ha). Soil measured around the trough and in the laneway was found to be enriched in Olsen P (56 and 201 mg P/kg, respectively) compared with the pasture (24 mg P/kg), as well as having a greater bulk density resulting from more frequent treading by animals. Use of the lane and trough during grazing greatly enhanced dissolved P losses from plots via dung. On a catchment scale, surface transport processes and sources were reflected in stream loads by the dominance of particulate P lost. Sub-surface flow was found to be an important contributor of discharge and likely P losses, which warrants further investigation. The scaling up of runoff plot data suggested that the laneway contributed up to 89% of the dissolved reactive P (DRP) load when surface runoff was likely. This represents a substantial source of P loss on dairy farms. In addition, the variation of sources and transport processes with season adds another aspect to the critical source area (CSA) concept, and suggests that, given the loss during summer and high algal availability of dissolved P, mitigation strategies should target decreasing dissolved P loss from the laneway.
Spatial and temporal variations in material flow paths, in-stream processes and differing land uses was then measured in a larger catchment using a nested catchment approach. Flow, P and sediment concentrations were measured at 8 sites (5 - 8,000 ha) in the Silver Stream (mixed forest/pastoral) catchment. The hypothesis was that P and sediment concentrations and loads would be a reflection of land use or flow regime, and that this would be a further reflection of the sediment equilibrium P concentration (EPC₀) of suspended or bed sediments. Dissolved reactive P was found to be unrelated to streambed EPC₀ during base flow. The mean DRP concentration under forest (0.006 mg/L) was lower than DRP measured from the grazed catchments (0.09-0.019 mg/L), but similar to the outlet of the catchment, suggesting dilution of P input from the grazed catchment area. During storms, 54-57% of particulate P was exported at the 5-10 ha scale compared with 85-90% at the >300ha scale. The findings of this study suggest that management to decrease P inputs should focus on strategies that target erosion processes and direct stock access to streams, as well as improving grazing management during sensitive times of year.
A Bayesian Network was constructed using results from this study and published data to create a whole farm TP export model. From this network the relative importance of various sources, both point and non-point, could be seen with different management options. Point sources (lanes and stock access to waterways) have the potential to significantly affect TP loads at the farm scale. However, under current practice, loading from diffuse sources is more important than from point sources. Of the various sources, the load from dairy farm effluent was found to be the greatest potential contributor to total diffuse loading.
The key findings of this work were the quantification of the trough and lane as potential source areas, and the effects of these sources with increasing scale. Transport of the particulate fraction of P from diffuse sources was greatest during storm events and found even at the catchment scale. In contrast, the dissolved P fraction from point sources was subject to transformation processes and transport during storm events was only observed at the field or farm scale. One of the remaining questions is what has the greatest influence on P transformations at different scales and with different land uses.||en