2020

Hedger, Richard; Bergan, Morten A.; Blumentrath, Stefan; Eloranta, Antti P.

We developed a new method to map and evaluate the impact of potential natural and artificial migration hindrances on the spatial distribution of sea trout (Salmo trutta) within stream networks. A stream network was derived from a 1 m2 spatial resolution LiDAR-based Digital Terrain Model (DTM), using part of Trondheim Region as a test case. Algorithms were developed to identify potential artificial migration hindrances (stream crossings and culverts) from the DTM, and to correct the DTM to enable generation of a terrain-derived stream network that followed the topography better than manually-digitized stream networks. Stream slope was computed at multiple-spatial scales throughout the terrain-derived network because steep slopes can be a potential natural migration hindrance. Potential migration hindrances were then quantified across the network from (1) the positions of crossings and culverts (using information generated from the DTM alongside GIS databases) and (2) stream slope metrics. The impact of potential migration hindrances on the spatial distribution of sea trout was determined by analysing the relationship between these stream network properties and the prevalence of sea trout across Trondheim Region, as determined by electro-fishing surveys conducted by Trondheim Kommune, NINA and NIVA. Models showed that prevalence was negatively related to the number of crossings and culverts downstream of the electrofishing site. However, no effect of slope was identified, and the predictive power of models was low. The terrain derivation-based approach developed here offered high local accuracy, but was computationally intensive, and suffered from potential confounding effects, and investigation of the effect of stream network properties on sea trout prevalence was limited by the quantity and quality of available data. This study has shown that a GIS-based approach, reliant on semi-automated processing of high-resolution DTM data, and integrated with GIS data, can be used to construct a stream network showing potential migration hindrances for fish populations. Further, there is potential for applying this approach over a wider geographical area and in different freshwater applications.