Written by Jonathan Reynolds, Head of Predation Control Studies as part of a two-part series. Read part two here.
Part one - The thought process
This week sees the publication of a very important paper by GWCT scientists, one that has been a long time in the making. First author Tom Porteus worked with detailed day-to-day records of fox control activity and success recorded for us by gamekeepers in the 1990s. Through a very sophisticated analysis he has succeeded in reconstructing the dynamics of fox numbers over several years on each shooting estate. We can now appreciate visually the extent to which fox numbers were knocked back by culling. Magically, at the same time, he has quantified processes that are normally hidden or near-impossible to measure – such as the production of cubs within the estate, and immigration – both of which counteract the impact of culling.
The background to this paper goes back some decades and is rooted in controversy over the importance of predation and thus of predator control. Predator removal experiments like Salisbury Plain and Otterburn showed that suppression of common generalist predators including foxes can increase breeding success and population growth of ground-nesting gamebirds. Note the word ‘can’. Because they were well-designed experiments, those studies had unambiguous conclusions, but their general relevance could be questioned. It was because the relevance of Salisbury Plain to the uplands was questioned, that Otterburn was needed as well. And both these studies were on military training areas. What happens in more intensively farmed regions; or in regions where there are more or fewer predators; or where terrain or vegetation or climate or public access make predator control more difficult; or where the gamekeeper is simply less proficient?
Sceptics have argued that in wider practice predator control may not be as effective as in those high-profile experiments, or in GWCT demonstrations, Allerton and Royston and where gamekeepers were strongly motivated. In particular, it is often suggested that rapid replacement of culled animals through immigration will prevent any worthwhile impact on predator density.
In the early 1990s, when this story began, predator control strategy and methods had not been subjected to any scientific study. Our sole source of data was the National Gamebag Census (NGC), a collection of annual estate records which since 1961 had specifically included predators. The NGC showed a steady 3-fold increase in annual fox bags between 1961 and 1991. There is no detail in NGC records of how these bags were achieved (e.g. effort or methods), and this makes interpretation hazardous. Did the increase reflect a genuine increase in the number of foxes, or was it just that gamekeepers were killing a greater proportion of the same fox population? In either case, the sustained increase in bags suggested that at a national scale this was a harvest rather than control. It said nothing about effectiveness at the estate-scale.
During that 30-year period the use of spot-lamp and rifle for fox control (‘lamping’) had become steadily more prevalent, with better equipment emerging all the time. (Currently the spot-lamp is being superseded by thermal and night vision equipment.) Lamping success was always greatest in the autumn after breeding and after harvest, when visibility was good, the fox population was at its highest, and two-thirds of the population were naïve juveniles. The long nights of winter obviously favoured the method too. This seasonality suited the protection of reared gamebirds after they were released, but we suspected it was one factor drawing gamekeepers away from the protection of breeding wild gamebirds. In spring, when fox numbers were lowest, and summer, when nights were brief and crops limited visibility, lamping was less successful. Another disincentive to control through the nesting season was the withdrawal in 1987 of Cymag, which had been widely used to destroy litters of cubs at the breeding earth.
So, how do you measure the impact of an ongoing cull within a relatively small area like a shooting estate? For avian predators like crows or magpies, you could just count them. For mink, we have the mink raft which tells you where there are - and where there aren’t - any mink left. But species like the fox are a biologist’s nightmare. The tried and tested methods of wildlife research are either too slow, too insensitive, or are already being used by gamekeepers to remove the animals you are trying to monitor! In the 1980s I had a taste of doing fox research on a gamekeepered estate. It is an impossible situation, both practically and in terms of human relationships.
In the mid-1990s, I somehow persuaded 74 indulgent gamekeepers to record details of all their fox control activity in a pro-forma diary, for at least 12 months, documenting the hours they spent on fox control and what success they had. Actually they were really interested to do this, but their dedication to keep records in such detail to help improve knowledge deserves resounding applause. The knowledge of fox control practice that we gained in this way has been invaluable in many ways since, and we may not have exhausted its value yet.
The thing that most fascinated me was the data from lamping, which told us both how many foxes were seen on each excursion (i.e. opportunity), and how many of those were shot (i.e. success). Two extreme cases illustrate how this helps to explain bag size. One gamekeeper on an 800 ha estate in East Anglia went out lamping twice a week and rarely saw a fox (despite a flat landscape). His annual cull – by lamping and other methods combined – was also low, about 20 foxes. It just seemed that there were very few foxes available to be culled. Another gamekeeper on a 600 ha estate in the south of England went out lamping equally often, and every time saw 1 to 6 foxes, typically shooting a third of those seen, amounting to an annual cull of 60 foxes. Despite this relentless cull, the number he saw each night showed no sign of a decline. From bag size alone, one might have concluded that the latter case was the more effective, but the reverse was true.
These examples fitted both with NGC data and with subjective experience. NGC data showed a much slower rise of fox bags in East Anglia compared with southern England, starting from a much lower level in the early 1960s. So there seemed to be regional differences, but it wasn’t clear what caused them. It was conceivable that where management for shooting was a prevalent land-use, culling itself had a long-term suppressive impact on fox numbers across regions the size of a county or two. This led on to our “three-region study” in which we compared fox density, fox breeding performance, and fox culling in each of three contrasting county-sized regions (2 in England, 1 in Wales). We showed that widespread, intensive culling was almost certainly responsible for low fox density across two of the three regions.
We still needed to understand how effective fox control was on individual estates. We expected differences in effort and skill, but there were also major differences in strategy. Some gamekeepers used fox snares, others didn’t. Some put a lot of effort into finding and dispatching litters of cubs, others didn’t. Some gamekeepers concentrated effort in winter, others in spring. On Salisbury Plain, our own gamekeeper seemed scarcely to sleep in spring and summer but eased back substantially for autumn and winter. Intuitively one can also see that the lasting effect of removing foxes will depend on how quickly they are replaced. But unless the cull was very intensive, we couldn’t be sure that they were replaced by immigration from outside the estate rather than cub production by overlooked foxes within it. The average annual cull recorded in the NGC (about 3 per km2 in 1991) gave no clue either way.
It was fairly obvious that the lamping data held more untapped information, and for that reason we kept the recording scheme alive for several successive years. Every time a fox was culled – by whatever method – it should have affected the number of foxes seen on the next lamping excursion. Thus each removal was potentially an ‘informative perturbation’, which could cast light both on the number of foxes present on the estate and on the speed of replacement. We just need the analytical skills to extract this information. “Just”! It took several years for Tom Porteus to acquire these valuable skills, learning from experts in modelling fisheries, where again harvest and effort data are the only clues to what may be going on beneath the surface. How was it done? See here.
Foxes, and what we can do about them, have been preoccupations in GWCT research since 1985. It has been a slow journey which is far from over, but we have methodically built up evidence to address successive questions. This has required a wide variety of research techniques, which we have generally had to develop these for the purpose. As a result our expertise in this field is unique. Complex computer models that reconstruct the unseen processes taking place during fox culling are the latest examples of this.