Future directions in disturbance research
The major questions identified for future research are: what determines patterns of human disturbance; how can we determine population-level responses to disturbance; are there general rules for predicting how important disturbance will be; how important are disturbance-derived ecological traps; what is the interaction between predation and disturbance; when does habituation occur; how do physiological responses to disturbance affect population size; what is the evidence for changes in access impacting upon populations; what are the positive consequences of access to the countryside; how important is habitat-specific disturbance; which measures reduce human impact and how can large-scale planning minimize the impact of disturbance?
UK policy-makers were asked in 2005 to identify the 100 ecological questions that they considered to be of greatest importance. The agreed final list included: ‘what are the impacts of recreational activities upon biodiversity?’, ‘what are the consequences for biodiversity of fragmentation by development and infrastructure?’ and ‘what are the ecological impacts on semi-natural habitats and ecosystems of adjacent large developments (e.g. housing and airports)?’ (Sutherland et al. 2006). Thus, there is clear demand for further research on human disturbance.
The overarching practical issues are impact and mitigation. Disturbance research, including that presented in these proceedings, focuses upon the impact that disturbance has on individuals and populations. Quantifying impact is important for identifying problems and showing where actions to mitigate undesirable impacts may be required. Thus, the outcome of many planning decisions and inquiries depends upon the impact of a proposed development. Understanding the effectiveness of mitigation measures is also of considerable applied importance in determining how planners, managers and reserve wardens can reduce the impact of disturbance, yet their study has attracted remarkably little research attention.
The appropriate approaches used in studying disturbance will depend upon the question being asked. Many studies of disturbance deal with measures of behavioural change, such as the frequency of leaving the nest, the time spent vigilant or the time in flight. Such information can be useful when asking questions such as whether birds respond more strongly to people when they are with dogs or whether noise is important. For studying whether disturbance is an important conservation issue it is necessary to determine the consequences for populations. This can be achieved by studying the consequences for nesting success, mortality or changes in distribution. Determining such impact is usually the most realistic objective. However, if possible, it is preferable to be able to assess the impact on numbers as the conventional currency used by conservationists is population size and this is the measure usually used in legislation.
In this paper I suggest 12 important research questions.
What determines patterns of human disturbance?
In order to predict and plan it is necessary to understand who goes where. Understanding people's preferences makes it possible to devise sensible interventions to adjust human behaviour for conservation benefit (Bathe 2007, Underhill-Day & Liley 2007). It should be possible to produce quantitative predictions as to how changes, for example in car park provision or signposts, are likely to alter patterns of human use. This requires social science research on a hierarchy of questions at national, site and within-site levels including the following:
National participation in an activity, such as orienteering, paintballing or wildfowling
How many are undertaking that activity?
Are the numbers undertaking that activity increasing or decreasing and at what rate?
What demographic, social and economic factors influence the uptake of that activity?
How often and when do they participate in the activity and what influences this?
Use of a given site
How many people who are interested in a given activity are within the catchment of that site? Of course the catchment size will differ for different activities.
What determines whether they will visit that site rather than alternative sites? In theory it might be possible for new housing developments to incorporate areas that are sufficiently attractive that they reduce the impact on nearby sensitive areas.
Behaviour within the site
How do people behave within a site?
What routes do they take and what do they do and where?
What affects their behaviour within the site?
An important question is whether rights of increased access result in more individuals visiting the countryside. The Countryside & Rights of Way Act 2000 allows open access to most areas of common, down, heath, mountain and moor in England and Wales. It is, as yet, unclear whether this will result in greater overall use of the countryside, whether this will result in similar levels of use but with that use distributed differently or whether the current patterns of use will largely persist.
How can we determine population-level responses to disturbance?
Two means of determining a local population response to disturbance are correlations and experiments. Correlations show the relationship between disturbance levels and bird abundance usually across a number of sites with different levels of disturbance, but sometimes on a smaller number of sites where disturbance events or intensity varies temporally. A complication is that there are often confounding habitat variables. These can just be a source of additional variation, which can complicate the analysis. More seriously, both bird abundance and the intensity of disturbance may vary between habitats, a situation which requires careful analysis to avoid statistical artefacts.
Experiments can be a very powerful technique. Thus, Madsen (1998) showed experimentally that restricting hunting to certain Danish wetlands results in considerable increases in the waterfowl populations in the refuges. A problem with both correlations and experiments is that they could just show a local response, such as localized movement, while the large-scale population response may be quite different (Gill et al. 2001).
A challenge is that many disturbance events are relatively small scale but the cumulative impact of such events needs to be assessed. For example, disturbance on waterbodies may have little impact where there is an adjacent undisturbed site but more serious for isolated sites or where there is simultaneous disturbance across the available sites.
Another approach to determining a population response to disturbance is to create population models. These require determining how density-dependent and density-independent processes determine population size. Once this is understood it is possible to incorporate the demographic consequence of disturbance and assess the consequences for population size (Sutherland 1998, Liley & Sutherland 2007, Mallord et al. 2007).
The usual challenge in creating population models is to assess the nature and scale of density dependence. There are four main means of determining density dependence: using time-series analysis (which is fraught with problems; Freckleton et al. 2006), relating birth or death rates to density over time (excellent but requires a considerable amount of data and is confounded if there are other changes), using experiments (in theory the best approach but usually impractically difficult) or using game-theory population models.
Game-theory population models require an understanding of both the decisions that individuals make and the consequences of these decisions (Sutherland 2006). For example, Stillman et al. (2007) use models based on competition for winter food amongst Eurasian Oystercatchers Haematopus ostralegus to understand the nature of density-dependent mortality and then incorporate the costs of human disturbance. Similarly, Liley and Sutherland (2007) use models of territorial behaviour to predict the nature of density dependence in the breeding season and then consider the impact of disturbance through nest loss by trampling and avoidance of otherwise good quality sites.
Are there general rules for predicting how important disturbance will be?
There is now a series of case studies quantifying the impact of disturbance on birds, such as those within these proceedings. It is obviously unrealistic to carry out field studies for every disturbance issue for every species of concern. It would thus be useful to have short cuts or a general guide that could predict the likely sensitivity to disturbance of a particular species to a particular disturbance. One possible approach would be to carry out comparative analyses to understand which groups of species are most sensitive to broad categories of activity. An alternative approach would be to examine the features that make a species sensitive to disturbance.
How important are disturbance-derived ecological traps?
Ecological traps are areas or habitats that species select but are actually population sinks. An example would be breeding birds selecting apparently high-quality bright green grassland that is actually highly fertilized and destined to be mown repeatedly for silage. As ecological traps attract birds even at low densities they can drive populations to extinction (Kokko & Sutherland 2001, Battin 2004). Mallord et al. (2007) showed that Woodlarks Lullula arborea avoided disturbed areas even though disturbance actually had negligible consequences for the birds’ fitness. There is strong competition amongst Woodlarks, so that at high densities birds have low breeding success. Woodlarks thus have higher breeding success in the disturbed areas they try to avoid than they do in undisturbed areas, where the Woodlark density is much higher. The undisturbed sites thus act as ecological traps. In this case it might be possible to reduce the impact of disturbance by modifying Woodlark settlement patterns, for example by reducing human presence during the settlement period.
It is important to understand how important disturbance-derived ecological traps are. The potential candidates for study include those species that have been shown to be less likely to settle in areas with high human disturbance: Stone-curlew Burhinus oedicnemus (Taylor & Green 2007), Ringed Plover Charadrius hiaticula (Liley & Sutherland 2007) and Golden Plover Pluvialis apricaria (Pearce-Higgins et al. 2007).
What is the interaction between predation and disturbance?
The impact of disturbance may be mediated through predation. The traditional concern is that predators, such as gulls or corvids, remove eggs or young when adults are disturbed from the nest (e.g. Langston et al. 2007). However it is also possible that disturbance could reduce predation rates if the predators are themselves sensitive to disturbance. There is thus a need for further studies on the interaction between predator behaviour and disturbance. This could lead to the ability to manipulate disturbance to reduce predation risk or reduce predation risk in areas with high levels of disturbance.
When does habituation occur?
If habituation towards disturbance occurs then the impact of disturbance is likely to be reduced. It would be useful to be able to predict when individuals are likely to become habituated or if there are ways of increasing habituation.
How do physiological responses to disturbance affect population size?
Many studies measure the physiological consequences of disturbance, such as changes in heart rate, hormone levels or energy expenditure. Although they might be interpreted as signs of stress, which should be avoided, it is usually difficult to interpret such information and use it as a base for practical actions. For example, what change in heart rate do we treat as a serious problem? The challenge is to relate such physiological responses to changes in demography. Beale (2007) showed that increased heart rates are linked to increased breeding failure in Black-legged Kittiwakes Rissa tridactyla. It would be useful to understand the mechanism by which such a link occurs. Is the link causal or does disturbance both increase heart rate and also increase the likelihood of abandonment? Can physiological measures provide a more efficient means than changes in breeding success in identifying potential problems? One approach is to understand the costs of these responses. The model of Stillman et al. (2007) incorporated increased costs through increased energy expenditure, which could be linked to changes in population size.
What is the evidence for changes in access impacting upon populations?
There are innumerable occasions in which access arrangements or levels have changed on a site as a result of adding or removing paths, car parks or the right to access. It would be invaluable to document how wildlife populations have changed in these cases. It would also be very useful to carry out such work to determine the responses to any changes in visitor numbers resulting from the Countryside and Rights of Way Act in the UK. It would also be useful to quantify the consequences of single short-term events, such as the site being used for an orienteering event. Anecdotal information, such as was described during the closing of the countryside in response to the outbreak of foot and mouth disease in the UK, is of little use as it tends to be highly biased towards reporting positive increases in wildlife observations that tend to be attributed to changes in access but in reality could be linked to other changes or just be due to chance. Data collected systematically in response to changes in access would provide a source that could be used to provide detailed analyses of the consequences of different activities on different populations. Ideally such data would be matched by another site where the disturbance levels stayed constant to allow for weather-related changes.
What are the positive consequences of access to the countryside?
There is a common tendency among those concerned with nature conservation to consider increased access to the countryside in a negative manner. As well as the general principle that we should not interfere with the decisions of others unless there are good reasons for doing so, there are benefits to access to the countryside including the health benefits (Pretty et al. 2003). It is also likely that greater access rights may result in increased interest in the countryside and thus concern for conservation. A greater quantification of the consequences of access for economics, health and happiness would thus be useful. Disturbance could, theoretically, even improve habitat quality in some specific cases by reducing overgrazing or by scaring specific predators.
How important is habitat-specific disturbance?
Murison et al. (2007) show that the severity of the impact of disturbance on Dartford Warblers Sylvia undata on heathland varies greatly between habitats, apparently as dogs readily enter land dominated by Heather Calluna vulgaris but not the various species of gorse Ulex sp. Linear habitats, such as the beaches studied by Liley and Sutherland (2007), could well be more susceptible. This has considerable applied importance in planning and modifying patterns of access and determining which sites are particularly vulnerable to disturbance. It would be useful to know how widespread habitat-specific disturbance is.
Which measures reduce human impact?
Although a wide range of measures are available to reduce the possible impacts of disturbance on wildlife, remarkably little evidence is available as to what is effective. A good example is the work of Pearce-Higgins et al. (2007), which shows that improving the quality of the Pennine Way footpath increases the number of walkers but reduces the number of individuals that leave the path, so reducing the impact on the surrounding birds.
The options for interventions include educational material, signs, footpath diversions, access point restrictions, habitat manipulations and zonation of potentially disturbing activities. There are a considerable number of specific questions for which there seems to be no information. For example, within the Countryside and Rights of Way Act dogs are only allowed in certain seasons if kept on a lead under 2 m long. Does having a sign stating that dogs must be kept on a lead result in more dogs being on leads than no sign? Is having a sign explaining why such a measure is necessary better than a sign just stating the law? Such questions are of widespread interest and would not be difficult to answer. They would make good student projects.
There is a need for evidence on the effectiveness of the main approaches. Sutherland et al. (2004) suggest that there is a need for the collation of the effectiveness of management. The website conservationevidence.com collates such evidence and it would be invaluable for those involved in managing access to document the interventions and the consequences.
How can large-scale planning minimize the impact of disturbance?
By planning on a larger scale than the sensitive area it should be possible to reduce the impact in such areas by understanding the spatial variation in sensitive features and access levels. Where is disturbance critical? Where are the impacts minimal? Where are there options for manipulating the impacts and what are the likely consequences? From such information it should be possible to consider the implications of modifying parking provisions, footpaths or zones for different activities.
As an example, there are proposals for building large numbers of homes in southern England. English Nature has questioned the building of houses near to heaths as these will lead to increased disturbance unless alternative recreation sites are provided.
The combination of theoretical developments and detailed field studies has meant that the study of disturbance has progressed markedly over the last decade, as is evident from these proceedings. Current work has concentrated upon showing whether disturbance has an impact but the focus is likely to change towards finding the most effective means of mitigation and visitor control.
The solution for minimizing the impact of disturbance is usually likely to be a combination of (1) reducing the numbers of visitors to sensitive sites (including by providing alternatives), (2) manipulating visitors on the site (for example by positioning footpaths or ensuring dogs are kept on leads at sensitive periods) and (3) manipulating the condition of the site (for example by ensuring that habitats in which disturbance is less of a problem are adjacent to areas that have more disturbance). Carrying out such a programme requires greater knowledge on impact and mitigation in order to determine the precise problems and solutions.
I thank Allan Drewitt for organizing such an interesting conference and the participants in the discussion for useful comments and thoughts. Andy Brown, Allan Drewitt and Rhys Green made useful comments. Our work on disturbance has been funded by the Countryside Agency, English Nature, RSPB and NERC.