Splitting Hares – Part 2

Last year, a paper was published by the highly respected Scottish ecologist Dr Adam Watson – celebrated as the most widely published ecologist in Europe before he sadly passed away – and Professor Jeremy Wilson, head of the RSPB’s Centre for Conservation Science in Scotland. This paper, published in the Journal of Applied Ecology and titled “Seven decades of mountain hare counts show severe declines where high‐yield recreational game bird hunting is practised, was a mini bombshell, documenting headline-grabbing declines in mountain hares by as much as 99% since the 1950s.

But perhaps what was most controversial was where those declines were reported to have taken place, as Watson and Wilson reported that moorland managed for grouse shooting (sites with more burning) had become associated with the highest rate of decline since 1999. They further suggested that the most likely explanation for the recent precipitous declines was the deliberate targeting of mountain hares by gamekeepers following intensive culls.

Mountain hares are culled on grouse moors because of the suspicion that they may act as reservoirs for the tick-borne louping ill virus (LIV) which, as the theory goes, may impact on grouse chick survival and hence profits. But even this tenuous justification is suspect, as there has never been any empirical link made between hare culls and improved grouse densities.

At a time of increasing public awareness regarding the management of our uplands for driven grouse shooting, the Scottish government has been subject to growing pressure on the specific issue of hare culls, with 36 Parliamentary Questions on mountain hares being asked in the Scottish Parliament between May 2005 and September 2017, and recently a broad coalition of organisations, including RSPB Scotland, Scottish Wildlife Trust, Scottish Raptor Study Group, Badenoch and Strathspey Conservation Group, Cairngorms Campaign, National Trust for Scotland, Royal Zoological Society of Scotland, Mammal Society, John Muir Trust and Mountaineering Scotland have joined OneKind in calling for a moratorium on the hare culls. In response, and following publication of Watson and Wilson’s explosive paper, First Minister Nicola Sturgeon has promised that her government will “be exploring all available options to prevent the mass culling of mountain hares.”

Screen Shot 2019-04-19 at 17.15.34

Fig. 1 Declines in mountain hare abundance on moorland sites managed for driven grouse shooting since 1954. Note zero abundance does not equate to extirpation as the surveys did not set out to count all hares, merely to get a measure of relative density (Watson and Wilson 2018).

But voices from within the shooting community immediately expressed skepticism about Watson and Wilson’s findings, with the Scottish Gamekeepers’ Organisation stating that the reported declines did not tarry with what their members were observing on the ground and Tim Baynes, director of the Scottish Moorland Group, saying there was no evidence “to substantiate claims that periodic culls are endangering mountain hare populations”. At the same time the Game and Wildlife Conservation Trust pointed to the fact that that their own records, in the form of the National Gamebag Census (NGC), did not show any significant decline in hare numbers.

Admittedly, the survey sites monitored by Watson and Wilson and the NGC were not the same, so one explanation for the disagreement is that there has been genuine variation in the rate of decline across these different sites, but another explanation is that the per capita death rate has been increasing – that a greater proportion of the hare population is being killed each year. This explanation for the discrepancy between the two datasets was one I proposed at the time – noting that because the NGC only records dead animals, not living ones, it was quite possible for increasing hunting effort to generate apparently stable numbers of dead animals even while the population itself was declining. Indeed this scenario gains credibility when you consider that grouse moor management has been intensifying in recent years, as illustrated by the increasing use of muirburn in the hills (the number of burns in Scotland increased by 11% per annum between 2001-2011).

Now however, a new paper has been published, authored by a collection of scientists from the Game and Wildlife Conservation Trust, that apparently directly contradicts last year’s paper from Watson and Wilson. In this paper Dr Hesford and his colleagues state that rather than exhibiting declines, mountain hare abundance indices “were higher and relatively stable on moors where driven grouse shooting was practised relative to lower indices and declines on moors where grouse were either walked-up or not shot.”

And they went further, stating that “Watson and Wilson (2018) used a method for counting mountain hares that was not consistent, employindogs at some study sites but relying on visual observations only at others” before citing a recent study that  “suggested that diurnal transects without a trained dog would be unlikely to producreliable indices of mountain hare abundance”. They concluded that “the estimated population trends presented in Watson and Wilson (2018) should be treated with considerable caution.” In scientific circles this is about as disparaging as it gets. 

So what’s going on? In this Trumpian era, where we are all encouraged to select from a menu of “alternative facts”, we now have the regrettable situation where those who wish to believe that driven grouse shooting is causing declines in mountain hares can point to one peer-reviewed scientific paper, while those who wish to defend shooting can point to another. Both sides want their viewpoint to be supported by the ultimate arbiter – science – but they can’t both be right. Are mountain hares really more abundant on grouse moors than other habitats in the UK? Are they in decline? And are culls driving that decline? In my view, the answer to all these questions is probably yes.

Let’s consider the abundance question first. Hesford et al. report that mountain hare numbers are higher on driven grouse moors (where grouse arflushed by beaters toward a stationary line of hunters) than on moors where grouse shooting was walked up (where hunters walk in a line, using dogs to flush grouse ahead of them), while hare numbers under both these management regimes were higher than on moorland  where there was no shooting at all.

For anyone familiar with Scottish moorland, this is not really that surprising and Watson and Wilson never made any claims to the contrary. Although mountain hares are not the intended beneficiaries of grouse moor management, mountain hares thrive in the same conditions that allow red grouse to reach such extremely high densities. Regular burns maintain the carefully cultivated patchwork of tender young heather plants mixed with sheltering stands of older heather, all in the absence of almost any mammalian or avian predators – the former being controlled by mostly legal means, while the latter remain suspiciously suppressed due to a suspected culture of illegal persecution.

Hesford et al. do not compare the number of hares on driven grouse moors to the number of hares that can occur in woodland – the hares’ natural habitat across most of their near-circumpolar range – and the possibility exists that hares may reach even higher densities in certain woodland types. We know that hares thrive in young plantations where they enjoy nibbling on the vulnerable saplings, but they do significantly less well in mature plantations of the sort that now cover much of our uplands, and prior to 1999 this was where Watson and Wilson documented their greatest declines.

Indeed, Hesford et al. recorded hare numbers on driven grouse moors in their Highland survey blocks that were 35 times higher than on moors where there was no shooting, whilst even on walked-up moors they were 15 times higher. On the other hand, there was no significant effect of management intensity on hare numbers at all in Tayside.

But, what about the second question. Are these grouse moor hare populations “relatively stable”, as Hesford et al. claim, or in decline as Watson and Wilson suggest?

The first thing to ask is what does relatively stable mean? Hesford et al. state that hare numbers “fluctuated over time in quasi-cyclical manner, fluctuations being more pronounced where hares were more abundant, i.e. on driven grouse moors.” In other words, hare numbers went up and down on all the moors they surveyed, but both those increases and decreases were more dramatic on driven grouse moors. 

Hesford et al. speculate that this yo-yoing is likely due to “density dependent factors” including resource competition and parasitism, whereby as the hares become more abundant they may suffer more intense competition with each other for food and increased exposure to parasites, until a tipping point is reached and their numbers begin to drop, relieving the competition and decreasing their parasite load until their numbers are able to recover. Such dramatic cycling is relatively rare in nature, and Hesford et al. admit that “a lack of cyclicity in hare abundance indices away from driven grouse moors may result from predators suppressing hare numbers.

Thus, in effect, predators dampen these oscillations by limiting the extremes of population growth and subsequent collapse. But where predators have been removed on driven grouse moors it is left to disease and sometimes starvation to curb booming populations. And severe fluctuations are unlikely to be good for the hares in the long term; a wildly fluctuating population runs a greater risk of bottoming out, of suffering a period of decline that is too prolonged or too steep, so that it ultimately results in local extinction.

But what about Watson and Wilson’s claims that hares have declined so much, and that recently these declines have occurred most rapidly on grouse moors? Watson and Wilson’s data derive from seven decades of spring counts, but Hesford et al. were critical of their methodology, stating that Watson and Wilson “used a method for counting mountain hares that was not consistent.”

In fact Watson and Wilson both acknowledge and explain this inconsistency in their paper, stating that “at some sites, hares were counted with the aid of one or more setter dogs, but at others, especially at alpine sites where vegetation was prostrate and too short to provide any concealment, dogs were not used.”

Furthermore, the survey method they used on any one site was applied consistently, while all the moorland sites – where grouse shooting takes place and where Watson and Wilson documented the greatest declines – were surveyed with dogs, just like Hesford et al., invalidating this criticism. Watson and Wilson also provide much more detail in describing their methodology than Hesford et al. do, including ensuring that surveys were carried out within the same time window, when there was no more than 30% snow cover and when wind speed was below Beaufort force 5.

Hesford et al. collected data over far fewer years than Watson and Wilson, but make the unsupported claim that “the most accurate information we have on long-term trends in mountain hare indices in Scotland have been derived from harvest data collected since 1961 by the Game & Wildlife Conservation Trusts (GWCT) National Gamebag Census.” They then go on to claim that these data, “together with sightings from Breeding Bird Surveys collated by the British Trust for Ornithology (BTO), show no significant change in hare hunting bag records or sightings in recent years (1995-2012).” 

It is not clear what justification Hesford et al. have for their claim that the National Gamebag Census (NGC) represents the most accurate long term population data available, nor is it clear that the BTO’s Breeding Bird Survey (BBS) records always support the NGC data. For example, BBS data has recently revealed a 60% decline in rabbit numbers, while the National Gamebag Census has recorded no significant change in rabbit numbers.

Worse, the literature Hesford et al. cite to support their claim that there has been no significant change in hare sightings in recent years does no such thing. Separately the BBS data and the NGC data recorded -26% and -36% decreases in mountain hare numbers between 1995 and 2009, and although neither decline was considered significant on its own, a joint model in the report cited by Hesford et al. actually showed this decrease to be significant overall.

Screen Shot 2019-04-19 at 17.20.36

Fig. 2 Joint plot of NGC and BBS data showing the significant decline of -28% in mountain hare numbers between 1995 and 2009 (Noble et al. 2012).

Thus the GWCT’s own data, when combined with the BBS data, appears to corroborate Watson and Wilson’s analysis that, while numbers do fluctuate, mountain hare numbers are in long-term decline.

Furthermore, while Hesford et al. were critical of Watson and Wilson’s methods, their counts of mountain hares were undertaken within sample blocks that were selected by an apparently arbitrary process. In their own words blocks were selected to be representative of the habitat for that moor as well as for their accessibility by four-wheel drive vehicles” but no details are given regarding how this qualitative assessment was achieved, nor do the authors describe any checks to see whether the blocks which were so conveniently sited for vehicular access were truly representative of the wider moor. 

I am no statistician, but I also struggled to marry some of the average annual rates of change reported by Hesford et al. with the accompanying graphical figures. For instance, in the Highland region (see Fig. 3 below), Hesford et al. calculate an annual increase in mountain hare abundance of 4.9% on driven grouse moors (the u-shaped dark grey curve that appears to end at a lower level than it began!), but an annual decrease of -6.6% on walked-up grouse moors (the middle light grey curve) and finally an annual increase of 7.6% on the moorland sites that were not shot (the bottom black line). Look at the graph for yourself and see if you can make sense of those numbers.

Screen Shot 2019-04-19 at 22.57.58

Fig. 3 Average mountain hare density (hares per km²) ± SE. Dark grey line on driven grouse moors, the lighter grey line for walked-up grouse moors and the black line for moors that were not shot over in the Highland region. (Hesford et al. 2019)

What is clear is that mountain hares are capable of exhibiting localised explosions in their population (see Fig. 4 below), and the dramatic increases recorded by Hesford et al. on the three driven grouse moors in the Grampian region that they surveyed between 2010 and 2017 may explain the incredulity of some of those in the shooting community who doubt Watson and Wilson’s claims of a more general longer term decline.

Screen Shot 2019-04-19 at 17.24.06

Fig. 4 Average mountain hare density (hares per km²) ± SE. The dark grey line represents driven grouse moors and the lighter grey line records walked-up grouse moors in the Grampian region. (Hesford et al. 2019)

In fact, Watson and Wilson themselves were careful to emphasise that “the severe recent decline on grouse moors in our study area should not be assumed to be replicated across all grouse moors or other habitats occupied by mountain hares in the UK” even as they identified culling of hares as the most “parsimonious explanation” for the severe declines they have recorded since 1999.

And since Watson and Wilson’s data extends over a much longer time period, it would seem reckless to dismiss such an unparalleled dataset without taking the declines it documents seriously. Indeed, even Hesford et al. acknowledge that their much shorter 16-year study “probably captures only part of longer-term population cycles.”

Meanwhile, let’s not forget that GWCT’s own NGC data revealed evidence of a significant downward trend in mountain hare numbers when it was combined with BTO’s BBS data, while another study published in 2018 by Massimo et al. recorded that “statistically significant declines were evident in 34% of the mountain hare’s range” with “a large area characterised by severe abundance declines, indicating an emerging conservation issue for this species”, all of which supports Watson and Wilson’s contention that mountain hares are in decline. 

Hesford et al. do shed some interesting light on the role of gamekeepers. Perhaps surprisingly, hare numbers were lowest in Tayside where gamekeeper densities tended to be highest, actually adding credibility to Watson and Wilson’s suggestion that culls by gamekeepers might be driving the most severe declines, although Hesford et al. choose to highlight the spread of mature plantations in this region when speculating about the cause of low hare numbers. Elsewhere, gamekeepers may have a more positive influence on hare numbers, and where the intensity and extent of heather burning was greatest – in the Grampians – hares were at their most abundant.

Hesford et al. conclude that on driven grouse moors the “benefits to hares from fewer predatorand better foraging opportunities may outweigh dis-benefits from sporting harvests and tick-related culls. An estimated total of 16,763 km² of Scotland* is managed for grouse shooting (Hudso1992), representing approximately 30% of the total upland area. Thus, it seems likely that driven grouse shooting provides a net conservation benefit to Scotland mountain hare population.”

*Hesford et al. got their numbers wrong here, citing Hudson’s (1992) estimate that “16,763 km² of Scotland” was managed for grouse shooting when in fact this estimate pertained to the whole of the UK. Exactly how much of Scotland is managed for grouse shooting remains subject to debate.

That strikes me as an ambitious claim, and a conclusion that will attract some raised eyebrows, particularly since it is in such stark contrast to the most “parsimonious explanation” identified by Watson and Wilson just a year earlier – that excessively zealous culling was most likely to blame for significant declines, while Hesford et al. completely ignore the other studies that have documented mountain hare declines.

Questions have been asked about the methodologies employed in both studies, but for now, with two contradictory datasets, we can perhaps only be sure of a few things, namely that driven grouse moors drive more extreme cycles in hare populations and hares are culled in large numbers on those moors.

It also seems likely that hares are typically more numerous on driven grouse moors where the environment is managed in a way that suits them so well, but there is no question that the moors are managed for the hares’ benefit. And it also seems likely, on the balance of evidence available, that mountain hares are in a state of long term decline, albeit with that trend complicated by localised exceptions and periodic fluctuations in hare abundance.

After all, hare culls for LIV control are designed to reduce hare densities severely, to less than 5 hares per square kilometre, or 95–97.5% less than the highest recorded densities, while some estates aim to keep hare densities as low as 0.3 hares per square kilometre. Moreover, culls typically take place in late winter and early spring when mountain hare populations have the least opportunity to recover.

In any case, wouldn’t it be better to let a more natural balance re-establish itself? Reducing the frequency and extent of muirburn in our national parks while shifting the emphasis away from predator control and back towards the self-willed primacy of nature might mean fewer hares in some areas, but it would also remove any need for culls. We might also then witness an increase in overall biodiversity, while mountain hares could establish smaller but more stable, sustainable populations, free from culls and less at risk from the disease outbreaks and parasite burdens characteristic of populations on driven grouse moors. Sometimes less is more.


Hugh Webster is an independent conservationist and author of The Blue Hare, a fable about mountain hares and the restorative power of wilderness.







3 thoughts on “Splitting Hares – Part 2

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s