Inspired by birds, informed by science: providing the evidence-base for policy decisions

Conservation often requires difficult decisions to prioritise one species over another. A current example of this is the plight of migratory salmonid fish (Salmon and, more recently, Sea Trout) in Wales. The problems facing these fish are complex, including changes in water quality, sedimentation and water flow in the freshwater environment and the impact of climate change at sea. Welsh Salmon populations are all declining and the species is designated ‘at risk’ or ‘potentially at risk’, even in the six Special Areas of Conservation (SAC) designated for them under the EU Habitats Directive. They are also a key feature of healthy freshwaters, and financially valuable to fishery managers, angling societies and the Welsh economy. Although a suite of pressures drives their decline, predation by birds presents a highly visible cause of salmonid mortality, and some stakeholders argue that lethally controlling fish-eating birds could have a positive impact on salmonid populations.

UK law protects all wild bird species, meaning a licence is necessary for any type of control. With a ’general licence’, it is legal to kill certain species including Carrion Crow and Woodpigeon, in order to mitigate a carefully-defined list of harms, such as significant economic damage, disease transmission risk, public safety or for the conservation of another ecologically-linked species. For other species, ‘specific licences’ are required to mitigate such harms, permitting the removal of a set number of individuals only. The control of Cormorants and Goosanders primarily falls under the second category, and embodies a trade-off between the conservation of the salmonid fish and birds. Licensed lethal control of both Cormorant and Goosander has already been happening in Wales for several years, but with fish populations still declining, Natural Resources Wales (NRW) recognised the need for all stakeholders to understand the implications of the conservation trade-off when making any new decisions. NRW convened a stakeholder group (Wales Fish-eating Birds Advisory Group) in 2018 for this purpose, and BTO was asked to participate as a scientific advisor. 

Winter populations of fish-eating birds

The Group built a shared understanding of both fish and bird populations, identifying knowledge gaps and issuing contracts to address them. As part of this, the Group asked BTO Cymru to research the current and potential future population status of Cormorant and Goosander in Wales.

The first task was to design a survey, gathering data to inform current population estimates. This was challenging: the species are highly mobile, have poorly understood distributions and also have recent experience of human disturbance, including lethal control. The second task was to deliver that comprehensive survey across 10 major Welsh rivers and a representative sample of stillwaters.

After conducting the survey, we combined the survey data with data collected by BTO/JNCC/RSPB Wetland Bird Survey (WeBS) volunteers to calculate the size of the overwintering populations of Cormorant and Goosander during the winter of 2020-21.

We estimated there to be between 2,580 and 3,259 Cormorants overwintering in Wales. Around 40% of these were found on rivers, and 50% were found in estuaries. Our estimate was approximately half of previous suggestions derived from UK-wide estimates made by the Avian Population Estimates Panel (APEP). At the UK scale, the Cormorant population is made up of two subspecies (coastal-breeding Phalacrocorax carbo carbo and inland-breeding Phalacrocorax carbo sinensis), and the overall population is increasing. However, the Welsh population consists almost exclusively of the carbo subspecies, and is currently fairly stable in numbers but progressively decreasing in productivity. This means fewer chicks raised each year per adult, which could cause declines in future. As the previous estimates used UK-wide APEP data to estimate Welsh Cormorant populations, the estimate may have been inflated.   

Our estimates for the overwintering Goosander population were slightly smaller, at 1,223 to 1,750 individuals, and were in line with previous estimates.  The vast majority of Goosander were found on stillwater (almost 30%) or on rivers (around 70%), with only a few on estuaries.

We estimated there to be between 2,580 and 3,259 Cormorants overwintering in Wales, and 1,223 to 1,750 Goosander.  

Neither species was evenly distributed across rivers, either along the length of the rivers or between different rivers. Numbers were higher near the river mouth than the source, and most birds were recorded on a few specific rivers. We found the greatest number of Cormorants on the River Dee, and Goosanders on the River Wye. 

Spring populations and the smolt run

The following spring (2021), the Fish-eating Birds Advisory Group contracted BTO to conduct an additional survey of the River Usk, a key river for spawning (breeding) Salmon which had the third- and second-largest numbers of overwintering Cormorants and Goosanders respectively. This survey was timed to coincide with the ‘smolt run’, a critical period in the life-cycle of salmonids, when young fish (smolts) undergo a profound metabolic shift to live in saltwater, and migrate en masse downstream to the sea. Smolts migrate overnight and rest at ‘pinch-points’ - places where the flow of the river is obstructed, like weirs, bridges or at natural barriers - during the day. During this migration, any smolt mortality means that fewer adults will return later to spawn. Smolts are especially vulnerable to predation at pinch-points, meaning opportunistic Cormorants and Goosander may have their most direct impact on salmonid populations at these points in migration.

The survey in spring found birds present in different densities and locations than those recorded in the winter survey. Cormorants were recorded at a much lower density, probably because the carbo subspecies moves back to breed at coastal colonies in spring.  Although Goosander numbers were similar to those recorded in winter, the smolt-run coincided with a change in population composition: between the start and end of the survey, males departed, migrating to traditional sites (most likely in Scandinavia) to moult, and were replaced by the first broods of young birds, which were accompanied by females.

Alongside the spring Usk re-survey, BTO designed a further survey to investigate predation risk at known pinch-points on the river, comparing these with nearby unobstructed stretches of river. The survey was conducted by Afonydd Cymru volunteers, and found that Goosanders gathered in greater numbers, and invested greater fishing effort, at pinch-points during the smolt run than on other unobstructed sections of the river, or anywhere after the run’s conclusion. However, there was very little evidence to support targeted predation or high-intensity predation risk at pinch-points during the smolt run from Cormorants, or by other fish-eating species including Grey Heron and Little Egret.

Our third report modelled the future population trends of Cormorant and Goosander in Wales, to examine the potential outcomes of a range of scenarios of lethal control.  

Evidence-based decision-making

Sometimes, policy development requires difficult questions to be thought through fully before decisions can be made. In this case, the Advisory Group wished to understand what the implications of lethal control might be for Welsh populations of fish-eating birds. Our third report drew on the results of the winter and spring surveys, along with established population change data from the BTO/JNCC/RSPB Breeding Bird Survey (BBS) and the expert knowledge of a panel of BTO researchers, to model the future population trends of Cormorant and Goosander in Wales. We used this demographic modelling approach to examine the potential outcomes of a range of scenarios of lethal control.

Constructing a single demographic model carries a significant risk of error, because, at the population level, no two sites and no two birds will experience the same conditions nor live the same life history. To reduce this risk, we combined every possible set of the full range of feasible demographic parameters (like the number of offspring raised per adult per year, and the number of adults surviving each year) and generated 62,500 different population projections for each species. The truth was in there somewhere, and we used BBS data to select the subset of these population projections that best matched observed population trends, and could therefore plausibly be true in Wales. 

Once we had selected these models, we applied different Cormorant and Goosander management scenarios, ranging from a 0% removal ‘business as usual’ scenario to an extreme case where 15% of the population of each species was removed each year. Cormorant populations were estimated to decline in future, even in the absence of lethal control (the 0% ‘business as usual’ scenario). In fact, 52% of the plausible models suggested a decline of 25% or over (which would trigger what is known as a ‘WeBS alert’) over a 25 year timescale. The proportion of models associated with WeBS alerts increased as the percentage of the population to be controlled increased, as might be expected. By contrast, the models suggested Goosander populations were on the rise, with 94% of plausible models predicting population increases under the 0% ‘business as usual’ scenario. However, higher levels of lethal control resulted in declines becoming more probable, including declines severe enough to trigger a WeBS alert.

It is personally and professionally challenging to work so closely with a decision-making process that, depending on NRW’s ultimate decision, may result in the lethal control of some Cormorants and Goosanders.  

Outcomes - where next?

BTO’s design, survey and analytical work was written up as a suite of NRW Public Evidence reports, and (alongside several other reports from other organisations) presented to the Advisory Group for their consideration in framing recommendations to NRW about a new policy approach. At this stage, BTO’s role as scientific advisor included clarifying the outcomes (and limits) of the modelling, and answering any remaining questions about ecology or statistical confidence in the results, but we were not involved in deciding the Group’s recommendations to NRW.

The Group recommended that continued, or extended, lethal control of both Cormorant and Goosander was necessary to promote the recovery of salmonid populations in Wales. They also recommended several other actions, including continued monitoring and non-lethal management of predation risk (actions not involving fish-eating birds were outside the remit of this Advisory Group). The final stage is the responsibility of NRW, as the competent licensing authority, to make its own decision on whether to accept the recommendations of this document, and if so, what level of lethal control to apply for each species. These decisions will again be made with reference to the results of our work, especially the modelling of future population trends under different scenarios of lethal control.

It is personally and professionally challenging to work so closely with a decision-making process that, depending on NRW’s ultimate decision, may result in the lethal control of some Cormorants and Goosanders. BTO staff and supporters understand that conservation can involve difficult choices. But we should be proud that, whatever decision is reached by NRW, the best possible evidence concerning the population status of these birds in Wales, and a careful analysis of the likely impacts of lethal control, was available to the decision-makers. The recommendations of the Advisory Group, and NRW’s decision, will be made with a clear and unbiased understanding of the risks, or otherwise, to the birds we are so wholly invested in. And that is an encouraging thought.