The MTB Impact Myth
The mountain bike is a great iteration in the evolution of the bicycle, opening a whole new world of possibilities as well as challenges. There are places where this is undoubtedly more true than others, and Scotland is, unquestionably, such a place. Not simply because of our long standing tradition of access, but because much of our spectacular landscape lends itself well to what the mountain bike has to offer.
Folk who don't ride mountain bikes sometimes (often?) don't get it. I do. Over the years, I have done a fair bit of rides of all kinds, the trail centres and the local skooshes, the winter mud and darkness, the long days of multiday expeditions into some of the more remote parts of Scotland, sometimes in sunshine, sometimes in torrential rain, the alpine summer holidays in search of dust, the amazing MegaAvalanche extravaganza. I have even earned some titanium bits somewhere along the way.
And like many mountain bikers, I have for years believed that the environmental impact of a person on a mountain bike is more or less identical to that of a person on foot. Well, I have since lost faith in that entrenched dogma, it is not supported by my day to day experience, and as it happens, it is not supported by the research either.
But There are Studies!
The reason most mountain bikers claim the bike has a similar environmental impact as a pedestrian is because back there in the pre-historic times of mountain biking there have been a couple of studies done that came to that conclusion, specifically a 1994 study from Montana by Wilson and Seney, and a 1997 Canadian study by Thurston and Reader. The former of these has little relevance to the Scottish situation, or the issues as they stand today. It examines the impact of different users on pre-existing, hardened mountain trails -- in Scotland the equivalent might be recent hard packed path restoration work, but in this context it makes perhaps sense to talk about cost of maintenance, but not of environmental impact. In addition to that the study is seriously methodologically flawed (see the critique in Pickering, 2010).
The study by Thurston and Reader in 1997 looks at the impact of hikers and bikes on pristine floor of a deciduous forest. The attempt to quantify the impact in this study is rigorous (they essentially count individual blades of grass), but unfortunately the very desire to exactly quantify the impact leads to a set up of the load test that does not at all reflect how bikes are ridden. The test areas are just 4m long, with an extra 0.5m buffer either end (i.e., about 2 bike length), 1m wide, with a marked centre line to follow, as they roll the bikes down:
Bikers traveled at a moderate speed, usually allowing bicycles to roll down lanes with-out pedaling where the slope would allow. Brakes were applied as needed to keep bicycles under control. Over rough terrain, some firm braking, occasional skidding, and some side-to-side movement of the front tire was required to maintain balance until a path developed. Once participants reached the bottom of a lane, they would turn and circle around the nearest end of the block back to the top of the lane to make a second pass.
Considering the detail to which the experiment set up is documented up to this point, the failure to set parameters for the riding is striking. The 'moderate speed' declaration needs to be interpreted in the light of 'side-to-side movement of the front tire was required to maintain balance', and the fact that the runs are just 5m long on not extremely steep slopes -- it is reasonable to conclude that by today's mountain biking standards these test runs are extremely slow.
The 4m length of the test area is entirely inadequate to capture bike behaviour -- there is an implicit assumption that the bike impact is homogeneous along its run, which is simply not the case, mountain bike riding is extremely dynamic. There are no corners, yet the area immediately before and in a corner is where the greatest impact of a bike is generally observed (i.e., braking bumps and skid marks). Similarly there are no uphill runs, for 'bikers could not make uphill passes, even in the lowest of 21 gears' -- this happens to be a normal part of mountain biking, and a realistic scenario would have made the biker to make best effort and then push. There is no defined braking protocol, and the 'firm braking' has to be seen in the context of what has been said about the speed above. The study has also been restricted to dry days only, which eliminates the situation when the ground is most vulnerable.
Not to beat about the bush, this is not how an average mountain biker rides their bike; bikes are invariably ridden fast, braked and cornered hard, and wheels dug in / span on climbs -- the best that can be said about this study is that it represents the most optimistic case of a cautious rider on a dry day, and that in itself should raise a few alarm bells.
Pickering's 2010 survey shows that up to 2010 there has been only one other study on comparative impact between pedestrian and bike traffic done in Australia (Chiu and Kriwoken, 2003) -- this study suffers from similar limitations to its two predecessors; it takes place on a hardened track (a fire road), and the bike load application is not adequately defined beyond 'recreational riding'.
There is a more recent Australian study by Pickering, 2011, that looks on a mountain biking impact on subalpine grass. It uses very similar experimental set as used by Thurston and Reader (in spite of the earlier Pickering survey noting the flaws in the methodology), on an 8 degree slope. This study has also been limited to dry days. As such it again represents the most optimistic case. This study is, however, interesting for two reasons. The results show that mountain biking impact on vegetation scales worse than that of a pedestrian, as by 500 passes the bike lanes were significantly worse off. It further shows that riding up or down hill is considerably more damaging than riding on flat; I shall return to this point later.
So here you go, that's the studies, and, no, the results are not too encouraging. They show that at best mountain bikes can hope for an equivalent impact to pedestrians when ridden with the greatest care in the dry. Nevertheless, this does not stop the mountain biking press from peddling the myth and worse. As recently as November of last year, the MBR has claimed that 'Research reveals walkers do more damage to trails than mountain bikers'. This is complete BS. The basis for the claim is a 2009 USGS study, which, however, was not conducted in a way that would allow meaningful comparison between different user groups. This point is clearly made in the paper:
In contrast, mountain biking, at 3.5 m 3 /km, has the lowest estimated level of soil loss, about 30% as much as on hiking trails. This finding reflects a limited mileage of trails where mountain biking was the predominant use (3.1 km), and these trails received low to moderate levels of use. [emphasis mine]
This point is also reiterated in Pickering's 2010, review of research, but neither of Pickering's papers is mentioned -- one can only wonder whether this is because the MBR writer is completely inept at Googling, or whether it has anything to do with the fact the reality does not fit his agenda (and headline).
Empirical Observations from the Real World
As I have noted above, the main issue with all of the rigorous studies is that they fail to create realistic riding conditions for their experiments. This is understandable. In order to get realistic data the research would have to be carried out on a real trail, considerably longer than the 4m test lanes, using real riders, preferably unaware of the study. This, however, makes quantitative assessment very hard, never mind comparison with other types of use. What we are left with is having to make judgements based on empirical observations in the real world.
So here are some of mine, mostly from the Dumyat hill above Stirling (for those interested, I also have 65 images that capture virtually all of the erosion on the west side of the hill, with comments, at Flickr). About four years ago I picked up an injury that kept me of the bike for the better part of six months. During this time I started hill running to keep the weight from piling on, and got quickly hooked. Much of my initial running was done in the exact same places around Stirling I used to ride, and in the Ochils, with Dumyat, the closest reasonably sized hill to where I live, becoming a regular destination for my midweek runs.
Compared to its neighbours in the Ochils, Dumyat is quite a peculiar hill: a lump of loosely bonded volcanic rocks, covered in a thin layer of soil (at places no more than a couple of inches thick), which in turn is held together by vegetation, primarily grass and some heather. In its natural form, the hill is pretty resilient to erosion. However, when the vegetation is damaged, water quickly washes off the top soil, and once the rock is exposed, it starts to disintegrate rapidly into gravel, forming ever bigger and deeper water channels. By this point the erosion is well beyond the hill's ability to self repair. (Just to be clear on this point, bulk of the erosion on the hill, i.e., the actual moving of soil, is caused by water run off, not by boots or wheels. The character of the rock is such that once exposed it does not require any further mechanical disturbance to erode. This makes the pace and physical scale of the erosion rather large relative the overall visitor numbers.)
The overall pattern of vegetation stripping followed by rapid water erosion is one of the main erosion patterns that can be observed on many of Scotland's more popular hills, except in somewhat accelerated form. Being heavily used by both walkers and bikes, this makes Dumyat a possible useful case study with a view toward the bigger picture.
The thing about running off road, particularly as a novice, is that you are very acutely aware of what is under your feet; far more than either walking or cycling. I did not really set out at any point to investigate erosion on Dumyat, I simply started noticing more clearly than before where the erosion is, and, over time, also how it forms. Over the next few months I made a couple of observations which forced me to change my view on the bike impact.
The first of these was noticing that there are significant differences in the way in which pedestrian and wheeled traffic impacts the vegetation. A pedestrian exerts a primarily downward crushing force. When this force is applied repeatedly in the same location, the area eventually gets de-vegetated, creating a foot shaped hollow, or rather a series of such hollows forming staggered steps. As the pressure continues to be applied, these hollows enlarge, until they form a continuous erosion scar. Notably, while the hollows remain in their discrete state, the area continues to be fairly resistant to water erosion, which only kicks in in earnest once the individual hollows merge.
In contrast, a bicycle wheel primarily exerts a tangential drag force, generated either by accelerating (i.e., pedalling), decelerating (i.e., braking) or centrifugal force (i.e., cornering). Pickering, 2011, as mentioned above, noted that bikes cause considerably more damage when going up or down, than on level. I believe this is the reason why, the wheel does not simply crush vegetation, it pulls on it, tending to damage roots faster. In real off road riding, the bike nearly always generates drag, for freewheeling in a straight line without any braking is a fairly rare occurrence (on Dumyat there is only one longer section that allows for this, and, it also happens to be one that shows very limited signs of erosion). The other important, if self-evident, difference from the pedestrian impact is that the wheel generates a single continuous trace. This, unlike the pedestrian indentations, tends to be subject to additional water run off immediately.
One of the obvious fall-outs from this is that at the start of, and well into, the erosion process we can reliably differentiate the erosion trigger. Over the last 3 years, I have only observed a handful of new areas of erosion that were clearly pedestrian triggered. At the same time, any erosion scar that is narrower than around 45cm can be unambiguously classed as bike-triggered. My conservative estimate is that bikes account for somewhere around 80% of the vegetation stripping on the hill that triggers subsequent water erosion, even though the pedestrian numbers seem considerably higher than the number of bike users.
(The actual numbers are hard to estimate without someone sitting there for a couple of weeks with a clicker. My mate Callum reckons bikes only account for around 11% of the hill users. I expected this number is somewhat understated. My own experience shows that the user make up is very weather dependent, with the walkers being mainly fair weather users; in good weather conditions, a 10% estimate for the bikes might be in the right ballpark. However, on a clagged out autumn day, the pedestrians tend to be reduced mainly to the local hill runners, whose numbers are fairly limited even compared to the local mountain bikers. The biker numbers are less effected by weather, they come for the exceptional quality of the riding, not the views, tend to ride all year around as well as after dark. I think it is reasonable to say that the pedestrians outnumber bikes by several times. This, however, is not a good news for the bikes, in view of what said in the previous paragraph.)
The other observation I made was regarding the pace with which the erosion develops to the 'beyond self repair' state. To my surprise, erosion initiated by pedestrian traffic develops relatively slowly -- because the discrete hollows tend to resist water erosion, it takes a large number of repetitions before a continuous scar forms. In some cases I have observed on Dumyat, as well as elsewhere, the process from the first de-vegetated hollows appearing to the forming of a continuous scar can take a year or more.
The situation with bike initiated erosion is very different. When the ground is saturated with water (which is 6-8 months of the year), it can take just a single, one off, bike track to remove most of the vegetation and kick-start the water erosion process. In one particular place I have observed how a single drag mark in the grass has turned into a 'beyond self-repair' two inch wide deep scar in the matter of weeks, and then rapidly progressed from there.
The empirical observations I have made over the recent years make me believe that the actual bike impact in the typical wet and soggy Scottish conditions is considerably worse than the best case scenarios of the idealised studies. The relative damage is, of course hard to quantify, but an educated guess can be made based on noting visual disturbances caused by a repetitive use by each of the two user groups. Having observed conditions of the ground after both hill running races and mates bike races taking place on similar types of a ground, I am inclined to believe that the impact of a single bike on surface vegetation might be as much as an order of magnitude bigger than that of a pedestrian, i.e., that it takes about 100 strong race field in a hill running race to have the impact of 10 bikes in the same sort of environment.
As far as I am concerned, the relative comparison does not really matter. I have simply stated my observations this way because that is how the mountain biking community chooses to frame the issue. Also, this is not about Dumyat, a rather insignificant hill used for grazing, on the boundary between urban environment on one side and heavily over grazed hills (chunk of which has been recently turned over to commercial forestry) on the other. This puts the erosion somewhat into perspective. This is about discernible patterns of damage, which I am noting with concern at other, more environmentally and culturally significant sites (e.g., Ben Lawers, the Cairngorms). Ridging big hills on bikes has in the recent years become very popular, and I understand why; but this makes it important that the mountain biking community understands the reality of bike impact on the mountain environment, and that it commits to the 'responsible' in Responsible Access. Sometimes we have to adjust the way in which we access the 'wild' places. At times bike is simply not appropriate for a given environment and/or conditions. At times runner is not appropriate either (FWIW, I now tend to stay away from Dumyat during the worst of the wet season when even the fell shoes are just too much).
PS: Well done for making it this far. :)
 Wilson, J.P., Seney, J.P., 1994. Erosional impacts of hikers, horses, motors cycles, and off-road bicycles on mountain trails in Montana. Mountain Research and Development 14, 77–88.
 Thurston, E., Reader, R.J., 2001. Impacts of experimentally applied mountain biking and hiking on vegetation and soils of a deciduous forest. Journal of Environmental Management 27, 397–409.
 Pickering, C.M., Hill, W., Newsome, D, Yu-Fai, L., 2010. Comparing hiking, mountain biking and horse riding impacts on vegetation and soils in Australia and the United States of America. Journal of Environmental Management 91, 551-562.
 Chiu, L., Kriwoken, L., 2003. Managing recreational mountain biking in Wellington Park, Tasmania, Australia. Annals of Leisure Research 6, 339–361.
 Pickering, C.M., Rossi, S., Barros, A., Assessing the impacts of mountain biking and hiking on subalpine grassland in Australia using an experimental protocol. Journal of Environmental Management 92, 3049-3055.
 Olive, N.D., Marion, J.L., 2009. The influence of use-related, environmental, and managerial factors on soil loss from recreational trails. Journal of Environmental Management 90, 1483–1493.