Sediment and Wood Dynamics in Forests Managed for Timber Production in New South Wales, Australia

Abstract

Over the next several decades there may be substantial pressure to extract timber from within harvest exclusion zones, or buffer strips along headwater streams to meet the projected increase in the global demand for forest products. Forest management activities have the potential to significantly modify woody debris loads in buffer strips, and thus the availability of material for recruitment into headwater streams. Therefore, it is important that forest management practices in these headwater systems take account of the role wood supplied from buffer strips play in the development of channel structure and sediment retention capacity. Most of the research on the effects of timber harvesting in buffer strips on sediment dynamics and wood accumulation in small headwater streams has been carried out in the Northern Hemisphere, in particular the Pacific Northwest of North America. Australia, in contrast to these regions, is tectonically inactive, with rivers that are characterised by highly variable flow regimes, and low sediment loads. Consequently it is likely that the drivers of erosion and sedimentation and woody debris recruitment are different. Where studies have been conducted in Australia, they have mostly been on larger systems in catchments not typical of the small headwater basins and ephemeral channels where the majority of timber harvesting in the coastal forest of New South Wales occurs. The main aim of this study is therefore to advance our understanding of the physical interactions between buffer strips and small zero order ephemeral headwater channels, and the potential impact of timber harvesting on these interactions. To address this main aim, five related research aims were proposed and focussed on: 1) assessing the characteristics of large wood pieces and their geomorphic significance in the channels; 2) identification of how, and from what distance the LW pieces were recruited to the channels; 3) measuring the impact of harvesting in the zero order basins and buffer strips along the channels; and 4) modelling the effect on channel erosion and sediment transport of in-channel obstructions created by large wood. To achieve this, five zero order channels in a native forest on the south coast of New South Wales, Australia, were instrumented with rain gauges and weirs to monitor rainfall, streamflow, turbidity, and suspended sediment concentration over a seven and a half year period. After a five year calibration period, four of the catchments were harvested, while one remained as a control. In the harvested catchments there were two replicated treatments; harvesting in the 10 m buffers, and no harvesting in the 10 m buffers. Two major methods were applied to identify interactions in the zero order channels between buffer strips and in-channel large wood obstructions, and the potential impact of timber harvesting on these interactions. The first was analysis of the effect of timber harvesting using a range of statistical techniques including generalised linear models and maximum likelihood estimation. The second method invoked a cellular landscape and river reach process model to simulate the role of channel obstructions created by large wood on channel erosion and sediment transport. The process model was parameterised and calibrated against an observed discharge and suspended sediment record. Following calibration, 12 discharge and large wood obstruction scenarios were carried out to model erosion and suspended sediment transport over a timber harvesting cycle. Results of the statistical analysis showed that harvesting had a significant impact on runoff and suspended sediment yield relative to the control, but did not have a significant impact on turbidity or suspended sediment concentration. There was no significant difference in post-harvest runoff, turbidity, suspended sediment concentration, or suspended sediment yield between the treatments. In the treatment where there was no harvesting in buffers, the significant increase in suspended sediment yield dissipated after 18 months and was accompanied by a hydrological recovery in peak flows toward pre-harvest values. In the treatment where there was harvesting in the buffers, there was no hydrological recovery in peak flows, nor was there a decrease in suspended sediment yields 18 months after harvesting. Erosion rates in the harvested catchments over the 32 month post-harvest period were three times greater than the control erosion rate, but were still less than the sustainable erosion rate recommended for Australia, and were largely in balance with regional estimates of soil production. The majority of large wood interacting with the channels, and forming obstructions behind which sediment was stored had diameters ≤ 30 cm. Most of this wood was being delivered from the channel margins within the buffer strips. Most of the sediment within the zero order channels was stored on the channel bed, and was an order of magnitude greater than the amount of sediment stored behind obstructions. Of the sediment stored behind obstructions, large wood provided the majority of storage sites. The use of the process model to increase and extend the observed discharge record provided some useful insights into the temporal and spatial pattern of erosion, sediment storage and transport, and complimented the statistical analysis of harvesting impacts by extending these processes over a harvesting cycle. The benefit of retaining buffers alongside the zero order channels was demonstrated by a substantial reduction in channel erosion and suspended sediment yield obtained through the inclusion of in-channel large wood obstructions in the cellular landscape and river reach model. Therefore it is important that harvesting practices maintain sufficient volume of standing wood in these buffer strips alongside zero order channels to ensure ongoing recruitment of large wood to the channel network. The management implications of this study are not just relevant for coastal New South Wales, but also for other catchments where their hydroclimatologies may produce similar results. Show more