Why does study design matter?
Science can be a powerful tool when used correctly. It allows us to better understand the world we live in or how to make the most of the resources available to us, but faulty experimental design can undermine our effort by taking us down the wrong path. Just how if you don’t follow a recipe correctly the end result won’t be a tasty meal but a probable indigestion, conducting an experiment without a sound experimental design doesn’t lead us to the truth, but to wrong and sometimes dangerous conclusions. The catch, here, is that scientists rarely have a ready recipe to follow*. Rather, what they do is alike to reading many different recipes (similar studies) and figure how to combine them based on the ingredients (variables) they will work with. To better understand what this means, let’s delve deeper into a study on how predators impact ecosystems, and unpack what about this recipe makes it so good. First of all, it’s important to understand what we mean by the ‘landscape of fear’…
*The exception to this is experimental replication, which is when a team of scientists replicates step by step the design of a different team to verify if they can reach the same conclusions. Such process is very important in science, allowing us to confirm the reliability of key studies.
The landscape of fear: how predators benefit an ecosystem even without hunting
Scotland has seen extensive deforestation at the hands of humans over the past few millennia, as well as a proliferation of deer aided by the hunting industry and the absence of an apex predator1. As a consequence, efforts to restore the Caledonian Pine Forest in favour of threatened species such as pine martens, twinflowers, and pine hoverflies must fight against thousands of hungry deer roaming the barren hills. Saplings can be protected with a mix of fences and annual deer culls to try and mimic predation, but these techniques don’t solve the underlying issue of a lack of natural control on deer numbers. In synthesis: a real predator is needed. Europe has bears, wolves, and lynxes, and you may be surprised to learn that they lived in the UK before being hunted to extinction not so very long ago – lynxes around 1300 years ago, while wolves as recently as 300 years ago2.
If you follow any conservation debate surrounding large carnivores, you have probably heard that reducing grazing pressure is one of the main benefits that their return can have in an area. To put it simply, grazing pressure is the relationship between how much vegetation herbivores in a certain area need to eat and the amount of vegetation available in said area. In a healthy ecosystem, grazing helps controlling the growth of certain plant species, preventing them from doing too well at the expense of others, and thus promotes plant diversity. But when herbivores eat faster than the surrounding vegetation can recover, this can have devastating effects, as we’ve seen in Scotland3-4 .
Much work is being carried out by organisations such as The Missing Lynx Project to investigate the potential reintroduction of an apex predator, but how exactly will a few lynxes help us with managing an estimated one million deer5? Surely, they can’t predate more than are already culled by humans on a yearly basis. But while you would be right thinking that, predation alone, mimicked by deer stalking, isn’t enough to deter overgrazing. What we need is a ‘landscape of fear’.
The Landscape of Fear
Let’s put it this way: what would you do if every time you went into the kitchen for a snack there was a chance you died? Surely you would stay there as little as possible. Well, for herbivores, the landscape can look like a series of kitchens one right next to the other, with each habitat offering different snacks and a different level of safety.
In summary: the presence of predators has the capacity to prevent overgrazing not only by lowering deer numbers, but especially by preventing herbivores from spending too much time on any given area – or kitchen. This is achieved by creating a situation in which deer have to balance their need for feeding and for vigilance; a trade-off that is part of any healthy ecosystem, and gives saplings the fighting chance they need to grow.
All good so far, right? But how will deer know that a predator is around if it can’t see it? And could simulating this landscape of fear really produce results? A recent study performed in Germany can help us illustrate exactly that – and we’ll discuss just what about this study’s recipe makes it work so well!
The study
We are now in the Bavarian Forest National Park, at the boarder with the Czech Republic; a landscape where red and roe deer face attacks from lynxes and wolves. A team of scientists led by Suzanne van Beeck Calkoen and Walter Di Nicola wanted to find out whether the reason for the forest’s health lay not only in the predators’ impact on deer numbers, but on how they made them behave…
Here scientists simulated the presence of large carnivores by using urine and later scat as olfactory cues (smells). These were used on a total of 44 plots scattered across the National Park, with the vegetation of each plot spritzed with one of four treatments: water (the control), cow urine/poo (to compare herbivore with carnivore), wolf urine/scat, and finally bobcat urine/scat (unfortunately Eurasian lynx urine was unavailable – must be in high demand – but bobcat was predicted to have the same effect due to the close evolutionary relationship between the two species). In the research world, experiments are often a compromise between what an ideal experimental design would look like and what a team can feasibly achieve.
This diversion from the perfect experiment doesn’t undermine the validity of their findings, but acknowledging the ideal scenario and motivating the reasons behind this change can help fellow scientist understand how they can tackle similar issues or how future experiments can be improved.
The plots were far enough apart to prevent each one from affecting the smell of its neighbour, but they were close enough together to have similar vegetation. This meant that the scientists didn’t need to worry that any difference in predator behaviour between conditions was actually due to differences in vegetation – or rather, this variable had been ‘controlled for’. Compared with lab experiments, where the environment one works in can be tightly controlled, field experiments must deal with a lot of variables. The more you can account for, the better you can pinpoint those that really matter for your results.
Did scientists expect deer to have any reaction to the smell of water? Of course not, which is why this plot served as a ‘control’. In this case, to ensure that the scientist’s presence (and smell) alone was not the reason behind a change in deer behaviour – but that this change was driven by the evidence that predators had been passing through the area. Cow urine was used to compare the impact of herbivore smell to carnivore smell, while using lynx and wolf urine was meant to compare two very different predators. Lynx are solitary hunters which rely on ambushes to take down a prey, while wolves hunt in packs, running after their target and taking it down through team effort. Would deer be more cautious of a potential ambush, which might be uncomfortably close, or a pack of wolves nearby?
The reaction of deer visiting the various plots was then observed across multiple months and seasons, painting a clear picture of their behaviour. This was achieved by planting native saplings in each plot and monitoring how intensively they were fed on by deer – after all, a deer will only graze when it is relaxed and not fearing predation.
Results & Conclusions
What this study found confirmed what conservationists have been arguing for years: the landscape of fear truly works. Not only did both red and roe deer spent significantly less time in areas where they could smell the presence of lynx, but this avoidance was also more pronounced during dawn and dusk, suggesting that deer in the National Park are familiar with their predators’ ecology and know to be more careful in times when they are most active. This is further confirmed by how mothers with young were found to be especially vigilant on winter nights; a time when snow on the ground makes lynxes even more deadly as they approach silently towards their prey.
As expected, deer showed no response to cow urine or water. But what was surprising was that also showed no response to wolf urine. What could be the reason? Perhaps it is the difference in hunting behaviour between wolf and lynx; perhaps it is due to the fact that wolves were only reintroduced to the national park in 2015 after being hunted to extinction in the mid 20th century. Given their recent return, Bavarian deer may simply not find wolves as scary as the historically ever-present lynx. We cannot know unless further research is done, but this discovery brings to our attention a small puzzle on deer ecology which we would otherwise not be aware of. While science communication is filled with the most exciting and headline-worth discoveries, results such as these are just as important in building up our knowledge of natural processes. The authors admit that there are still unsolved questions here – this is what good science is made of, and what truly drives us to continue trying to improve our understanding of the world we live in!
And finally: could the use of predator urine work in the UK without reintroducing large carnivores alongside it? It definitely seems feasible. After all, similar methods using the false presence of a predator are already in place. In British Columbia, for example, just the sound of barking dogs has been used to deter raccoons from foraging, benefitting native populations of crabs and fish6. So, as long as we can prevent this red herring from being discovered, implementing something similar in landscapes like Scotland could prevent herbivores from undermining our attempts at reforestation while the debate around predator reintroduction moves forward.
References
Primary Paper
van Beeck Calkoen, S., Di Nicola, W., Smit, C., Kuijper, D. P., and Heurich, M. (2026).
Experimental evidence for large carnivore risk cues reducing deer browsing intensity in a temperate forest. Journal of Applied Ecology, 63(1), e70267. https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2664.70267
Other References
1. Damask, K. O. (2026). Habitat use and abundance of mesocarnivores and deer along Eastern Loch Lomond (Doctoral dissertation, University of Glasgow). https://theses.gla.ac.uk/85787/
2. Wilson, S., and Campera, M. (2024). The perspectives of key stakeholders on the reintroduction of apex predators to the United Kingdom. Ecologies, 5(1), 52-67. https://www.mdpi.com/2673-4133/5/1/4
3. Ye, K., Li, J., Liang, Z., Zhao, X., Luo, H., Chen, Z., et al. (2025). Overgrazing impacts plant species diversity in alpine wetlands indirectly by altering its environmental dependency. Global Ecology and Conservation (e03532). https://www.sciencedirect.com/science/article/pii/S2351989425001337
4. Beck, J., Hernández, D., Pasari, J. and Zavaleta, E. (2015), Grazing maintains native plant diversity and promotes community stability in an annual grassland. Ecological Applications, 25: 1259-1270. https://doi.org/10.1890/14-1093.1
5. Deer Working Group (2020). The management of Wild deer in Scotland. https://www.gov.scot/publications/management-wild-deer-scotland/
6. Suraci, J., Clinchy, M., Dill, L. et al. (2016). Fear of large carnivores causes a trophic cascade. Nat Communications 7: 10698. https://doi.org/10.1038/ncomms10698
