Bird stomach contents analysis - final report: Goosander and Cormorant diet on four Scottish rivers 2019 to 2020

Appendix 5: The timing of bird sampling and the smolt run

Given the 'patchy' nature of the material available in the present study, this Appendix provides a preliminary exploration of the sampling dates for Goosanders and Cormorants on the river Tweed in relation to information on the timing of the smolt run there in spring 2019. It does not explore issues of spatial abundance which might also be important in understanding dietary findings - as birds were sampled from a limited number of locations, in the absence of information on the distribution of smolts within any of the study river systems.

Collecting bird samples is obviously constrained by many issues, and so it is not surprising that they were often aggregated – generally coming from particular locations and particular times during the season (for full details see Appendix 3). Sampling thus occurs at times and places when predation on smolts is thought to be at its maximum and/or when collection opportunities are most appropriate/favourable. This patchiness – and the confidence associated with interpreting the present findings - was further compromised by the necessarily small samples available. Here, samples of each species were limited to 12 bird stomachs with food, and so in practice the number of birds available from each river during the smolt run was only 12-14 birds.

There is thus an important issue in relation to matching when the majority of the bird samples were taken and the timing of the smolt run. The Tweed Foundation were able to provide some information on the timing of the smolt run which could help investigate this point.

To provide an indication of smolt movements, James Hunt (JH) of The Tweed Foundation kindly provided some raw data from the Gala Water smolt trap, a tributary of the Tweed located 70 km upstream of the estuary. In 2018, when the trap was first installed, the main run started on the 6th May with almost no smolts counted until then, presumably as a consequence of an exceptionally cold spell of weather affecting water temperature. In 2019, the main run on the Gala Water started earlier on the 22nd April, although the first smolts were trapped on 30th March and the last on the 7th June that year. The main 2019 smolt run started on the 3rd May, again after a cold spring.

The raw daily catch data from 2018-2019 (number of smolts/day) were amalgamated to show the relative proportion of the catch on a weekly basis throughout the season. This temporal smolt run pattern is shown alongside the numbers of birds sampled on the Tweed, for each calendar week (Figure AP5.1).

Gala Water trap data show that smolts were caught during April but suggest the main run started on the 22nd April in 2019. Taking the graphical interpretation of the smolt run data at face value for now (see trap efficiency/river flow comment below), data suggest that the birds sampled in the first half of April might have been before the main smolt run whilst those at the end of May might have been towards the end/after the main run (Figure AP5.1).

Some words of caution are needed here though. As well as the actual 'magnitude/scale' of the run, the amount of smolts caught in the trap is also affected by water flow. The smolt trap gives only a partial count because when the river is high, trap efficiency will be lower. Indeed James' interpretation of the "relatively small" peak on the 5th/6th April (i.e. in week commencing 02 Apr in Figure AP5.1) is that the river was high on these days, so trap efficiency would be lower. Thus, although Figure AP5.1 shows the temporal distribution of smolts trapped, it has not been 'corrected' for river flow/trap efficiency, which could change the picture. JH's "suspicion is there is a smaller annual run of smolts in early April, but the main run is late April."

Such speculation highlights the imprecision with which such comparisons can currently being made, and also highlights that further bird samples are needed to better understand the dynamics of their diet in relation to the timing and magnitude of the smolt run and evidence of the scale of smolt consumption. The increasing availability of smolt run data from traps does suggest the possibility of better understanding bird diet in relation to the timing of the run in specific years, and might be an aid to refining the timing of samples. Smolt trap data are probably available from at least the rivers Tweed, Dee, and Spey.

A further caveat is that the relationship between the timing of smolt catches in the Gala Water trap and the actual pattern of the smolt run in locations on the main stem of the Tweed where birds were sampled is currently not clear. Further investigation may illuminate this. To paraphrase JH, smolts migrated out of the Gala Water relatively early in 2019, either as a consequence of the tagging process or because they were early-running smolts. Also, 2019 was a dry year and so there was no rise in water level and flow that would help fish out to sea. The 2020 smolt tagging study on the Tweed provided further insights and showed that smolts were typically in the main river until mid- to late-May that year. However, it is not clear how this fits with the curve sketched in Figure AP5.1 and the interpretation offered above.

It is interesting to note that if Gala Water data truly reflects the smolt run on the entire Tweed catchment (although this is not clear), then around half of the Goosanders were sampled before the main runs commenced. Nevertheless, given the smolt run itself lasts for several weeks and aggregations of fish are presumably moving down the system throughout this period, there should be some broad accordance between bird diet and the 'availability' of smolts during the run. It seems that stomach samples are thus unlikely to be completely unrepresentative of the prey consumed at this time of year. However, further research is required into the possible mismatch of bird samples and the timing of biological events such as the smolt run, as well as the efficiency – and wider representativeness – of smolt trap data.

Data from smolt trapping on the Tweed suggest the main smolt run is in May (at least in the last couple of years for which data are available). Many bird stomach samples from the 1990's were from March-April, and although considered by all at the time to be essentially representing the smolt run, it would be interesting to understand how representative of the main smolt run they might be considered now. Of course, such retrospective exploration might not be possible and/or the timing of the smolt run in the 1990s might have been earlier than it is now, but it is not clear whether any data might be available to help explore this issue. Again, this lack of clear evidence suggests that more intensive bird diet sampling is required to explore dietary changes alongside concurrent smolt run timing data.

Given that there is some finer resolution location and date information available for the R. Tweed bird samples from the 1990s, it might be possible to explore this alongside the samples in the present study to (i) undertake a fine-scale analysis of temporal dietary composition throughout the sampling periods available to us so far, and to (ii) perform some form of power analysis to examine issues of sample size (i.e. numbers of stomachs with food) in relation to the resulting length frequency estimates of fish therein, particularly Salmon. This could also be linked to an exploration of the potential influence of sampling location (i.e. beat/stretch) and used to plan a more intensive sampling programme for evidence gathering of bird diet in relation to the smolt run on Scottish rivers. This, in turn, should be linked to further studies of smolt movements and distribution within study rivers. It is also vital to know where birds are actually foraging rather than the locations where they were sampled.