National Juvenile Trout Parr Benchmark
What is it:
Electrofishing data are one of the most commonly collected sources of information on the status of stream-dwelling salmonid populations. However, interpretation of these data is challenging because densities vary naturally among habitats. Benchmark models provide an expected density against which observed densities can be compared for assessment purposes. Benchmarks have been developed for salmon (Malcolm et al., 2019a) and trout (Malcolm et al., in prep) using historical electrofishing data and spatial data that acts as a proxy for physical habitat. Benchmark densities can be considered the average expected density for a particular habitat having removed substantial anthropogenic negative impacts identified during analysis. Such benchmarks can thus be considered a target for healthy populations, which is similar in concept to meeting “intrinsic habitat potential” (Burnett et al., 2007). When combined with robustly collected electrofishing data (for example from the National Electrofishing Programme for Scotland - NEPS) these benchmarks are able to provide a fully scalable (site, catchment, region, national) catch independent, juvenile salmonid assessment.
Four benchmarks predictions are provided; 1) Salmon Fry 2) Salmon Parr 3) Trout Fry and 4) Trout Parr
Benchmark densities (fish per m2) are visualised on the log scale. NAs reflect where benchmark densities are not generated (e.g. within lochs, or in rivers where their covariates are outside of the range of data used to fit the benchmark models).
More Information:
In this dataset, benchmark densities were predicted at each river node (point at the start and end of a spatial line feature) for salmon and trout, for fry and parr. Where nodes had the same river order, the edge (spatial line feature or river segment) benchmark was the geometric mean of the two values. Where the downstream node had a higher river order than the upstream node (e.g. a tributary entering a larger river) then upstream benchmark predictions were assigned to the edge to avoid inflating benchmark estimates for the segment. Where benchmark predictions were only available for a single node (e.g. a river source) then this was used alone (see Malcolm et al., 2023 for further details). Predictions are only provided for rivers where the individual covariates (predictor variables) are within the range of data used to fit both benchmark models. This does not account for covariate combinations that may also be out the range of observed data in the benchmark models. The national juvenile salmon density benchmark model reported by Malcolm et al. (2019a) was used to predict salmon densities. A new trout benchmark model, derived following similar procedures was used to predict trout benchmark densities (Malcolm et al., in prep.) Preliminary information on this model can be found in Jackson et al., 2025.
Malcolm, I.A. et al. (in prep) Development of an assessment benchmark for juvenile brown trout (Salmo trutta) using juvenile electrofishing data.
Malcolm I.A., Millidine K.J., Glover R.S., Jackson F.L., Millar C.P., Fryer R.J. (2019) Development of a large-scale juvenile density model to inform the assessment and management of Atlantic salmon (Salmo salar) populations in Scotland. Ecological Indicators 96: 303–316 DOI: 10.1016/J.ECOLIND.2018.09.005
The following reports should be referenced for full details on the National Electrofishing Programme for Scotland (NEPS) survey design and assessment methodologies:
Jackson F.L., Gilbey J., Eagle L.J.B., Fryer R.J., Malcolm I.A. (2025). The status of juvenile Atlantic salmon and brown trout populations in Scotland’s rivers: The National Electrofishing Programme for Scotland (NEPS) 2023. Scottish Marine and Freshwater Science Vol 16 No 1, 107p. DOI: 10.7489/12543-1
Malcolm I.A., Jackson F.L., Millidine K.J., Bacon P.J., McCartney A.G, Fryer R.J. (2023). The National Electrofishing Programme for Scotland (NEPS) 2021. Scottish Marine and Freshwater Science Reports Vol 14, No 2, pp. 63 DOI: 10.7489/12435-1
Very small rivers (Strahler first order rivers on the CEH digital river network) were removed from this dataset. NAs exist where we are unable to make benchmark predictions (e.g. in lochs or in circumstances where we cannot generate the required covariates) or in rivers where their covariates are outside of the range of data used to fit the models.
