Understanding The Effectiveness Of Vegetated Streamside Management Zones For Protecting Water Quality

We set out to improve understanding of the effectiveness of streamside management zones (SMZs) for protecting water quality in landscapes dominated by agriculture. We conducted a paired-catchment experiment that included water quality monitoring before and after the establishment of a forest plantation as an SMZ on cleared farmland that was used for extensive grazing. In a second study, we monitored water quality during the harvesting of a 20-year-old plantation in an SMZ. We found concentrations of bacteria, sediment and phosphate were lower in the buffered paired catchment, but that lower nitrogen concentrations could not be attributed to the intervention. Harvesting caused no appreciable increase in sediment delivery to the stream and we found it to be a minor source compared to other disturbances (road drainage and cattle disturbance).Simulation of hillslope processes and stream flow illustrated that uptake by SMZ vegetation was a more important nitrate-mitigating process than denitrification. Because this model thoroughly integrates climate and within-soil processes, its usefulness should also be tested for other nutrients and chemicals. For example, the observed decrease in phosphate concentrations in stream water due to the SMZ treatment was probably predictable. However, additional modelling approaches are needed to simulate SMZ effects on sediment and bacteria.Key factors contributing to the success of the studies were the cooperation of land owners who understood the research aims and were willing to collaborate with on-ground actions essential to the study, convergence of the interests and resources of multiple partners, the proximity of research staff and laboratory to the field site, early peer review of aims and methods, and experimental designs that controlled many potentially confounding factors (e.g. farm management, roads, catchment geomorphology). Some of our results and interpretations were compromised by limited pre-treatment data during a period of high inter-annual variability in rainfall and flow. Longer-term datasets would have improved assignment of measured water quality changes to the watershed management interventions. Apart from climate variability, these changes were also driven by high variability in other natural characteristics (e.g. size, soils and topography of each catchment), multiple interacting drivers of change, long time-lags between intervention and response, and non-linear processes driving undefined thresholds of change. Our key results and limitations are presented as a guide to planning future research.