Science and research

UK Upland Waters Monitoring Network

Follow us on Twitter

UK UWMN Fish Sampling Methodology

fish picture from Bill Beamont

Annual electric fishing surveys have been employed to assess the abundance, age structure and standing stock of salmonid populations at each stream site and at the outflow streams immediately downstream from each lake site. Fishing takes place between mid-September and mid-October. Three local laboratories take part in the fish component of the Network (see collaborators). Marine Scotland and ENSIS/ECRC provide central coordination of the fish surveys.

The sampling procedure involves isolating within a 500 m section, three 50 m reaches of the stream with stop nets. Where topographical conditions permit these reaches are at least 50 m apart. Each 50 m reach is fished using electric fishing apparatus, usually from the downstream to the upstream net. The fishing is repeated three times, or more if no clear drop off in numbers has occurred. The fish are anaesthetised and the fork length and weight of each individual fish are recorded. All fish are retained in holding boxes in the stream and returned alive at the cessation of electric fishing.

The size of the fish population is calculated from the Exact Maximum Likelihood estimate method (Carle and Strub 1978) modified from the method devised by Zippin (1956). The method assumes that the catch per unit of effort is proportional to the population number. A series of samples taken with the same effort from a closed population should therefore show a decline in number per sample. It is important that a constant and relatively large proportion of the fish present in the reach are caught at each fishing and that the effort is roughly constant at each sampling. This latter assumption is tested by a Chi square test and invalid estimates can therefore be determined.

Catch data for trout, and where present salmon, are stratified into 0-group and >0-group cohorts and population estimates calculated. Estimates include population number and density for each reach of each site and for the combined reaches of each site. Estimates of the error associated with the estimates are also calculated. Where the population is small (less than 30) the calculated standard error of the population may be too narrow and hence unreliable. The possibility of small-sample error and the validity of the Chi square tests are annotated in the table of results. Calculations are carried out on computer using the "Remove" computer programme (Clarke 1988).

It is possible to determine age of smaller fish from length-frequency distributions, but scales are taken from the larger fish (>65mm length)and from a selection of the smaller ones to determine age. Scales are taken from the flank of the fish just behind and below the dorsal fin. Historically quality control in the determination of fish age was coordinated by the Centre for Ecology & Hydrology Dorset and by a scale exchange exercise among the fish scientists involved in the Network.

Additional information is recorded on the presence/absence and distribution of non-salmonid species.

Along each 50 m fishing reach a detailed survey of the stream habitat is made utilising the "HABSCORE III" format. This assesses the physical make up of the stream and in particular the habitat suitability for juvenile trout and salmon.

Data from the annual surveys are transferred to the central biological database. Fish catches at each site are summarised by age/class, length and weight and together with the population estimates are recorded in the UK UWMN annual data reports.


Carle, F. L. & Strub, M. R. (1978) A New Method for Estimating Population Size from Removal Data. Biometrics, 34, 621-630.

Clarke, R. T. (1988) Remove - a program for population assessment. Institute of Freshwater Ecology, Internal Paper.

Zippin, C. (1956) An Evaluation of the Removal Method of Estimating Animal Populations. Biometrics, 12, 163-189.

Page last modified:
Page published: 12 March 2010