Beneath the Surface > August 2014 > Thursday Analysis Update - 14.08.2014

Thursday Analysis Update - 14.08.2014

This is the second Thursday update using the format of a monthly overview of the project progress so far. Office moves, summer holidays and software updates have all resulted in progress being very slow for the last six or seven weeks. However, despite this some progress has been made and this is detailed below...

Assessment of potential relationships between juvenile salmon stocking and the rod catch on the River Carron
The data in this paper has been lagged to give a better representation of the relationship between the numbers of fish stocked and the number of fish in the rod catch. The statistics have been completed – Spearman’s rank correlation, regression analysis and residual plots. To date, all of the analyses are giving a strong relationship. However, the data still needs to be further lagged as the current transformations do not take into account differing proportions of S1 and S2 smolts in the population or the different proportions of S1 and S2 smolts that will mature into grilse or Multi Sea-Winter (MSW) fish. The residuals also indicate better fit between the trend line and the data in the early stocking effort therefore the data will be split in half and pre and post 2004 data subjected to statistical tests separately.
The data in it's current form has now been split into a set of pre-2004 and post-2004 data with the aim of seeing if there is a stronger correlation in the earlier years of the stocking programme than in the later years; a possible indication that the stocking had a greater impact in the early part of the programme which is often seen to be the case. While the analysis of the complete dataset demonstrates a strong correlation, neither the pre-2004 nor the post-2004 datasets show any correlation.
The question of the proportions of S1, S2 and S3 smolts assumed to be present in the data will have to be determined by calculation as no details of an overarching, general proportion has be found in the literature or in communication with other academic institutions. To this end the screw trap and electrofishing data will be used to determine if there is any bimodality in the samples which may indicate differing age classes, this will be combined with scale readings (to be performed be me following training with specialist input where needed).
An initial attempt to calculate proportions of S1, S2, S3 and S4 smolts from the screw trap data generated mixed results – proportions in three years were clearly and easily calculated but beyond these, proportions calculated in one year could not be reliably transferred to other years. This leaves the electrofishing and scale reading data as the only methods likely to generate usable smolt age proportions.
I will be contacting Andrew Duncan in the near future to discuss possible ways of proving whether the lagging transformations performed have improved the fit of the data to the trend and, if so, by how much.

Mark and recapture (Panjet) study
The recapture data has been looked at quite thoroughly in terms of looking at possible correlations and all of the possible biases in the data. So far we have been able to show a weak correlation between the number of fish marked and the number of fish recaptured and between the number of fish marked and total catch for each year. This data also shows that the proportion of the marked fish in the recapture and total catch record stays fairly constant. Analysis shows no significant relationships in the factors that could potentially bias or confound the study – the catch composition does not show greater numbers of males than females or grilse than MSW adults, nor does that data suggest that fish show any preferences in movement direction after initial capture and marking. The analysis did indicate a possible trend in the data towards a reversal from a majority male catch to a majority female catch – this observation will be examined further in a linked, but separate, small scale paper looking to see if proportions of grilse and MSW adults and size of angled fish are also changing in line with the change from male to female fish. I am also looking into development of GIS maps that will show the distribution of where fish were captured and re-captured.

Assessment of River Carron smolt run
So far this data has been organised and graphed to show a three day running mean of smolt numbers against date and river level observations made by Bob. I have already received information from SEPA on the daily flows and rainfall for the Carron which will be related to this smolt run data to see if we can see patterns between rainfall, river level and smolt numbers. Work has also begun on looking at difference between smolt damage in different years and this will be statistically analysed once the organisation and clarification of the data is complete.
Following discussion with Bob Kindness, it has become clear that there is no quantitative data on predator sightings available to compare with the instances of fish damage. Discussions have been had abaout obtaining photographs of damaged fish, although again, there are not many of these available for analysis.
The SEPA data for flow, rainfall, temperature and river level have been recieved and subjected to initial analysis by plotting this information on graphs alongside the smolt number data and moon phase data (see example below).

For a larger image of this graph click here.

Use of coded wire tags to assess the relative impacts of Salmon smolting in the wild or the hatchery
The data for this section has been “cleaned up” to some extent to allow it to be analysed. The data has been further reorganised to allow relationships to be explored but discussion with Andrew Duncan indicates that creation of a mixed model on ‘R’ may be the best way to look for significant patterns and interactions in the data. To that end I am liaising with Andrew to organise attendance on an ‘R’ course covering, or his assistance with, the development of a mixed model and the analysis of the data is on hold until the course of action has been decided upon.
So far no further progress has been made on this as we are waiting for Andrew to return from the college summer vacation before discussing the possibility of him taking over the development of the mixed model entirely as he is likely to be able to generate it faster and more effectively than myself and would avoid the loss of time incurred by attendance on the training course.

Determination of the distribution of juvenile salmon in the River Carron catchment
So far no analysis has been started for the electrofishing data but I have organised a GIS training session with the Inverness College GIS trainer aimed specifically at creating maps that will show the Carron electrofishing data.
The GIS training is progressing but slowly due to office moves, software updates, and holiday and research commitments of both Dr de Raad (GIS trainer) and myself. My decision to effectively start over from first principles rather than trying to repurpose material already available has also meant that the process has taken longer but should produce a more accurate and fit-for-purpose product at the end. The skills obtained in the training will allow comprehensive GIS maps to be generated for many of the sections of this report as the restoration project continues.
The fish obtained by Murray Stark and myself during an electrofishing survey in April have now been processed, and information gathered from this fish will be added into the electrofishing data to help build the picture of the River Carron as well as adding into the genetic picture of the river.

Using genetic analysis to determine the relative abundance of wild and stocked Salmon in the River Carron
The cataloguing of the samples will be complete by 23.07.2014 and the first batches will be sent to the genetics lab by 01.08.2014. The results of these tests will be analysed by Mark Coulson now that he has accepted a place with the RLI. The cataloguing of the samples is now complete. The final decision on which samples will be shipped to the genetics laboratory will be made on Tuesday 19.08.2014 before the samples are then packaged and shipped that afternoon.
Mark Coulson has now joined the team and has already begun analysing 358 samples generated as part of an electrofishing survey carried out by Eric Verspoor and Jonah Tosney in October 2012. When compared to samples taken from fish known to have been stocked by Bob Kindness, these samples will begin to illuminate the relationships between wild and stocked fish in the Carron.

Initial assessment of gravel movements in the River Carron catchment
The analysis in this section has not yet been started but contact has been made with Richard Tipping at Stirling University to obtain copies of three student dissertations based on data collected in the summer of 2013 for inclusion in the final report.
Following discussions between Professor Tipping and Melanie Smith; Professor Tipping has agreed to provide a short summary of the three dissertations for inclusion in this report.

Carron population modelling
This section has not yet been started.

Species other than Atlantic Salmon
This section has not yet been started.

Posted: 8/14/2014 5:43:24 PM by Matthew Curran | with 0 comments

Trackback URL:

Blog post currently doesn't have any comments.