Electrofishing Data Analysis Tool for NEPS Collaborators
Analysis of data from the National Electrofishing Programme for Scotland (NEPS) depends on a series of complex models and modelling approaches developed by Marine Scotland Science. These models are often not readily accessible to local fisheries managers, requiring significant computing power, statistical and coding expertise. However, in recent years, it has become increasingly possible to make complex outputs from the R statistical environment available through user friendly computer programmes known as R Shiny Apps. Marine Scotland Science has developed such an application (tool) to allow local fisheries managers, collaborating under the NEPS programme, to apply some of the models developed by Marine Scotland Science to locally collected electrofishing data.
While reliable catchment or regional assessments of fish abundance, status or trends depend on a representative (statistical) survey design (e.g. NEPS), substantial electrofishing data is often collected at a local level where the interest is frequently focussed on the enumeration of fish at the level of indivdual electrofishing sites. This new electrofishing tool allows local fisheries managers (that are also NEPS collaborators) to upload electrofishing count data (i.e. the number of a particular species and life-stage of fish caught on a particular pass of an electrofishing survey) to obtain estimates of the total density of salmon and trout fry and parr at a particular site, regardless of whether they were caught during electrofishing.
This approach also allows managers to obtain broadly comparable estimates of abundance from 1-pass or 3-pass electrofishing data provided that there are sufficient relevant 3-pass data available to parameterise the underlying capture probbaility model, and that the 1-pass data are comparable in method and effort to the first pass of a multi-pass fishing.
The resulting density estimates can be plotted spatially on simplified maps of the river network. In the case of salmon, estimated densities can also be compared to a benchmark (expected density) for the site.
The electrofishing tool is underpinned by a new national capture probability model for Scotland that has been extended from previous models to include both salmon and trout (Figure 1). The model was developed using methods described previously by Millar et al. (2015). The tool works as follows:
- NEPS colaborators upload information on site name (must be unique for each uploaded dataset), location (British National Grid x and y coordinates), area (wetted or bankfull), date of sampling, and the number of fish caught of each species (salmon or trout) and lifestage (fry 0 years, parr>0 years) on each pass into a template spreadsheet that contains appropriate column headers for the capture probability model
- The x and y coordinates are used to snap electrofishing locations to the digital river network (DRN)
- Spatial covariates (Altitude, Gradient, Upstream Catchment Area, Hydrometric Area) and benchmark salmon density predictions are extracted from the nearest up and downstream “node” locations on the DRN and averaged
- Spatial covariates, date of sampling (Year and Day of Year; DoY), electrofishing pass and Trust Team (a factor representing a time delimited combination of organisation and staff composition) are used to predict the capture probability for each combination of site, date, species, lifestage and pass
- The density of fish at a site is estimated from the total counts across all passes, cumulative capture probability and site area (see Malcolm et al., 2019 for more details)
- The resulting density estimates can then be plotted on maps (either as observed densities or, in the case of salmon, performance against the benchmark) or exported for use elsewhere.
There is a problem
Thanks for your feedback