|dc.description.abstract||Due to anthropogenic activity, many rivers have been degraded and have lost structural diversity. Rivers restorations are commonly undertaken in order to reinstate the lost habitat heterogeneity. In New Zealand, freshwater systems are mainly threatened by agricultural land use. Streams running through agricultural land have often being artificially straightened, therefore resulting in a decrease in riparian vegetation and an increase in fine sediment and nutrients. Waituna Creek, a tributary of Waituna Lagoon in Southland, has all these characteristics. Therefore, there have been efforts to reduce the effect of agriculture within the catchment, with the largest effort being a physical instream restoration.
The Waituna Creek restoration project focused on restoring the physical aspects of the stream. This work required the use of a digger to widen the stream banks and install large wooden logs. However, this kind of work can have undesired effects on the stream bed by compacting and/or re-suspending existing fine sediment, or by further increasing fine sediment cover. Stressors such as excess fine sediment often have negative effects on macroinvertebrate communities, but the short term effects of physical instream works that can further stress macroinvertebrate communities are not well studied. In the present study, the first aim was to determine the short term effects of physical instream restoration works on the macroinvertebrate community within Waituna Creek.
In New Zealand, the standard method used for the annual State of the Environment monitoring in streams is quantitative Surber sampling. This method is not often suitable for soft-bottomed, slow-flowing streams such as Waituna Creek. Therefore, the second aim of the present study was to compare the standard Surber sampling approach, with semi-quantitative kick-net sampling, in order to determine the better-suited method for Waituna Creek.
I monitored the macroinvertebrate communities of Waituna Creek using a BACI design, by collecting samples at replicated Control and Impact sites, before (360 days) and shortly after (2 days) the restoration works, using the two different sampling approaches. Surber sampling involved three 0.1-m2 samples taken from riffle areas within each 40-m study reach. The kick-net procedure involved sampling 10 locations allocated to equally represent the different microhabitats present within the 40-m reach. To assess the effect of the restoration works, a variety of stream macroinvertebrate community health indexes were used, as well as the relative abundances of invertebrate taxa commonly found.
For the community index data collected using both sampling methods, the majority of macroinvertebrate community indexes (5 of 9) showed no Year x Site interaction, indicating no effect of the restoration works. The four EPT taxon richness metrics that did show a positive Year x Site effect also had a confounding effect of a drought in the summer of 2017/2018 (the period leading up to the collection of the after samples), so it cannot be confirmed whether the positive trends were caused by the restoration works. The results of the common taxa showed that for the Surber data, five of 10 common taxa found had a Year x Site interaction, with all five implying a negative effect of the restoration works. The kick-net data showed five of 14 common taxa with Year x Site interactions, with three taxa implying a negative effect of the restoration while the other two implied a positive effect. Considering a new taxon-specific sediment tolerance metric developed in 2017, there were also implications that some of these common taxa were being affected by increased sediment cover, a possible consequence of the restoration or the drought.
Overall, there appeared to be little effect of the physical restoration works on the invertebrate community indexes. This may be because the Waituna Creek invertebrate community as a whole is already fairly resistant to the possible additional of fine sediment added by the restoration works. Most of the common invertebrate data did not show any effect of the restoration. However, there were still several taxa that showed predominantly negative effects, which may be due to increased sedimentation from the restoration works or the drought.
The two sampling methods compared well for the community level metrics but differed more often for the common taxa, likely due to the different microhabitats of the creek sampled. The Surber method focused on the riffle areas of the creek whereas the kick-net method also sampled other important microhabitats such as areas with macrophytes and woody debris. Therefore, I recommend the semi-quantitative kick-net sampling method for routine monitoring of macroinvertebrate community in lowland soft-bottomed streams.||