Abstract
Despite their diverse, resilient and dynamic nature, braided rivers are in global decline. They face various pressures, including climate change, invasive species, modification to flow patterns, and channelization. The deterioration of braided river ecosystems is particularly concerning given that their floodplains support a wide array of biodiversity, spanning both terrestrial and aquatic organisms. Consequently, restoration efforts are being dedicated to these threatened environments. However, restoring braided rivers presents a significant challenge, owing to their intricate habitat complexity and spatiotemporal dynamics. To further the understanding of braided river functioning and recovery requirements, some commercial gravel extractors are trialling innovative restoration practices. To determine the feasibility of this approach, I assessed how fish communities respond to a gravel extraction technique intended to be restorative, in comparison to a more commonly used gravel extraction method. Three treatments were compared: beach skimming gravel extraction that aims to restore braided river functioning (RES), pit mining that produces high yields of gravel and creates large ponds (POND), and largely degraded sites that have not undergone recent gravel extraction (UNRES). Field surveys were undertaken monthly over three consecutive months (February - April 2023) on a section of the Aparima River, in Southland, New Zealand. Fish and environmental variables were surveyed at ten sites: three RES, three UNRES, and four POND sites. I analysed habitat variables and the composition, diversity and density of native fish to determine and compare the effects of each treatment on fish communities. Compared to the UNRES treatment, the habitat in the RES treatment was more similar to natural floodplain morphology and functioning, with greater mean riffle habitat, wetted width and number of channels. The POND treatment had high nutrient levels, little or no flow, and was relatively deep in places, distinguishing it from other treatments. Fish species composition showed variation amongst all three treatments, with the greatest difference between POND and RES treatments. Although non-significant amongst treatments, the POND treatment displayed trends of lower native fish diversity, which continued to decrease over time. To investigate population-level variation, samples of the most abundant native fish species, the upland bully (Gobiomorphus breviceps), were collected for dissection. I analysed and compared the diet, parasite infection and growth of upland bullies to determine the effects of each treatment on fish population health. The POND treatment exhibited higher parasite loads, faster growth and different prey composition compared to the RES and UNRES treatments, which were largely similar in these aspects. Gravel resources from braided rivers are in high demand and the ecological impacts of gravel extraction vary depending on the desired outcomes and the methodologies used. The beach skimming technique, developed for restoration purposes, presents an ecologically preferable method of gravel extraction. While it has not yet fully achieved its intended restoration goals, it displays potential to be restorative. Successfully implementing gravel extraction methods that simultaneously contribute to the restoration of floodplains and provide aggregates for human use would bring mutual benefits to both humans and the natural environment.