- Top of page
- CASE STUDIES
The internationally important populations of waterbirds that winter in the United Kingdom can face intense pressure from human disturbance as a result of the high urbanization found around many protected coastal or inland wetland sites. Here, I describe and evaluate an approach that has been used to investigate the spatial effects of human disturbance on waterbirds. Rather than directly investigating behavioural responses to individual disturbance events, the presence of features in the landscape associated with disturbance is instead used as a surrogate, with the essential aim being to demonstrate that bird numbers or densities are depressed or their behaviour altered in proximity to areas used by humans. This paper first describes case studies that demonstrate the limitations of the basic inference (i.e. that disturbance influences patterns of waterbird distribution or behaviour) and then how investigations might be strengthened. For conclusions to be sound, it is particularly important that other factors, such as food supply, that might also explain the spatial patterns observed are considered or other corroborative evidence presented. The approach is thus least applicable in the most heterogeneous environments where many factors, perhaps spatially autocorrelated, may explain variation in distribution or behaviour. However, greatest confidence in the validity of conclusions may be gained where studies are able to show (ideally by experimental manipulation) that species’ distributions or behaviour vary temporally in line with the levels of human use of the features examined. Although its aim and scope are thus limited, the use of a landscape approach, provided that it takes into account other factors affecting spatial variation in bird abundance or behaviour, can provide a preliminary assessment of species avoidance of key sources of disturbance that may offer a framework for more detailed investigation.
The coastal and freshwater wetlands of Europe support internationally important numbers of waterbirds in winter (Kershaw & Cranswick 2003, Rehfisch et al. 2003, Collier et al. 2005) with many sites designated as Special Protection Areas owing to their significance for these species (Stroud et al. 2001). However, many important sites are highly urbanized, with habitat having been lost to industry, housing, harbour developments and barrage schemes (Davidson et al. 1991, Burton et al. 2006, Burton in press). With these developments, there has been increased disturbance from, for example, construction work itself, road and rail traffic, and also increased and varied recreational use of wetland sites. Recreational use may include water-based activities such as sailing, water-skiing, jet-skiing and windsurfing, as well as walking, dog-walking, cycling and trail-biking (Davidson & Rothwell 1993).
The effects of these activities on waterbirds have been the subject of considerable study, with research often focused on measuring the behavioural responses shown by individuals to different disturbance stimuli (e.g. Burger 1981, Roberts & Evans 1993, Smit & Visser 1993, Fitzpatrick & Bouchez 1998, Triplet et al. 1998, 2001, Mathers et al. 2000, Lafferty 2001a, 2001b, Thomas et al. 2003). Such studies may provide information on, for example, the proportions of different activities causing disturbance, the distances at which different species take flight, or the time taken for birds to return to a site or resume foraging, though, as noted recently, the level of response shown may depend on the stress that individuals are under (Triplet & Gembarski 2001, McGowan et al. 2002, Stillman & Goss-Custard 2002, Beale & Monaghan 2004). Nevertheless, such information, particularly in relation to the frequency of disturbance, may help to indicate whether individual fitness (e.g. body condition, survival) is potentially impacted by disturbance and in this way help in the development of modelling studies to predict impacts on local populations (West et al. 2002, Goss-Custard et al. 2006, Stillman et al. 2006). Alongside such modelling studies, there is still a need for empirical evidence that disturbance may actually impact bird populations (Hill et al. 1997).
Often, information about the potential impact of human presence is required from a site perspective, for instance by those monitoring the status of sites designated for their waterbird interest. At this level, information may initially just be needed on whether the distributions of birds within a site are affected by particular human activities.
Here, I describe and evaluate an approach that has been used to investigate the spatial effects of human disturbance, whereby the presence of features in the landscape associated with disturbance is used as a surrogate. The premise of these studies is that the abundance, presence/absence or behaviour of birds will vary spatially in relation to certain landscape features because they act as sources of human disturbance. This paper first describes case studies that demonstrate the limitations of the basic inference (i.e. that disturbance influences patterns of waterbird distribution or behaviour) and then how investigations might be strengthened through consideration of other factors that might also explain spatial variation in species abundance or behaviour and temporal changes in species’ responses. The case studies focus on waterbirds, though the approach has been used for other taxa. The limitations of the aim and scope of the approach, as well as its potential role as a preliminary step for further more detailed analyses are discussed.
- Top of page
- CASE STUDIES
Landscape approaches, as described in the case studies above, aim to investigate the spatial effects of human disturbance by demonstrating that bird numbers or densities are depressed or their behaviour altered in proximity to areas used by humans. However, confidence that disturbance is a factor in the patterns observed in such studies may potentially be limited as behavioural responses (notably movements from sources of disturbance) are not usually directly observed. To be able to infer that disturbance is a cause of the spatial variation in bird abundance or behaviour observed it is therefore important that other factors are considered or other corroborative evidence presented.
This is particularly the case in heterogeneous habitats such as estuaries, where variation in factors such as substrate type, water depth and food supply may considerably affect waterbird distributions (Kelsey & Hassall 1989, Nehls & Tiedemann 1993, Velazquez & Navarro 1993, Yates et al. 1993). In such systems the approach would not be applicable unless some attempt were made to account for other causes of variation, and even if this is done it should be noted that some factors may be spatially autocorrelated. Spatial variation in food supply was partially, but indirectly, taken into account in models in the second case study (Burton et al. 2002a) by distinguishing between intertidal count sections that bordered the low tide mark and those that did not and, in the third case study (Burton et al. 2002b), by including the count section itself as an explanatory variable. However, it is clear that such models would be very much improved by more direct measurement of variables influencing food availability, as demonstrated in the study of grassland field use by waders by Milsom et al. (1998), or preferably the food resources themselves. A study of the effects of disturbance on habitat use in Black-tailed Godwits (Gill et al. 2001a), for example, found no effect of the presence of footpaths or marinas on the numbers of birds supported on adjacent intertidal areas once bivalve food supply had been taken into account. By contrast, McKinney et al. (2006), studying waterbirds wintering in Narragansett Bay, Rhode Island, USA, found that landscape characteristics, including the extent of surrounding residential development, explained a significant degree of the variation in species abundance and richness, having allowed for habitat characteristics such as prey density and shoreline configuration.
The use of a landscape approach may also be improved by quantifying the level of use of features such as footpaths and roads. By demonstrating that waterbird species distributed themselves further from drives when visitation levels were high, Klein et al. (1995) provided much stronger evidence that the avoidance of areas close to drives was due to disturbance rather than other coincident factors. Reijnen et al. (1996) similarly demonstrated how reductions in the densities of breeding grassland waterbirds adjacent to roads were positively related to the level of traffic (see also similar studies by van der Zande et al. 1980, Reijnen et al. 1995).
Temporal changes in response, such as found by Burton et al. (2002a) between weekdays and weekends, when the pressure from visitor numbers and recreational activities is likely to be higher, may provide clearer evidence that observed distributional patterns are the result of disturbance. Evans and Warrington (1997) and Evans and Day (2002) similarly showed that wildfowl may redistribute within or between inland waterbodies at weekends in response to recreational or hunting activities. Better evidence still may be obtained where potential new sources of disturbance are introduced to an area or existing sources of disturbance removed and corresponding changes in species’ distributions are observed. In the third case study (Burton et al. 2002b), count sections were only classified as ‘disturbed’ sections during periods of active construction work and thus the ‘disturbance’ effects reported are unlikely to have been the result of differences in food resources across the study area. The approach indeed may be best and most confidently applied in experimental, before-after-control-impact (BACI) studies (Stewart-Oaten et al. 1986) where human presence in the landscape is switched between sites temporally. Madsen (1998a, 1998b) clearly demonstrated how hunting disturbance can influence the distributions of wildfowl by establishing experimental areas where hunting was manipulated prior to the establishment of permanent refuges. Quarry goose and duck species redistributed according to the location of hunting-free areas, whereas non-quarry species did not.
Through the range of studies described, the essential aim is the same, i.e. to demonstrate spatial variation in bird abundance or behaviour in relation to landscape features associated with human activity. This aim is limited and the degree to which it can be concluded rather than merely implied that disturbance is a causal factor of the spatial patterns observed is, as discussed above, first and foremost dependent on the extent to which other explanatory variables are considered. The approach is thus least applicable in the most heterogeneous environments where many factors, some perhaps spatially autocorrelated, may explain variation. Greatest confidence in the validity of conclusions will be gained where studies are able to show (ideally where possible by experimental manipulation) that species’ distributions or behaviour vary temporally in line with the levels of human use of the features examined.
Landscape approaches do not aim to provide any evidence as to whether disturbance may impact individual fitness and thus the size of local populations. Nevertheless, provided that other factors affecting spatial variation in bird abundance or behaviour are taken into account, the use of a landscape approach can offer a framework for more detailed investigation of the mechanisms of disturbance. In a previous study, for example, Gill et al. (1996) used information concerning the reduction in Pink-footed Goose densities with decreasing distance to roads to provide the basis for estimating the numbers of birds that could have been supported by the food resources not exploited as a result of disturbance associated with roads. In particular, by revealing how bird abundance or behaviour may vary spatially in response to landscape features, landscape-based studies may highlight specific sources of disturbance where further investigation would help inform management options. The approach may also provide information on the relative tolerances of different species to the disturbance associated with human landscape features. In the second case study, the relative distances to which models suggested species were affected by footpaths proved a good match to previously reported information on the distances at which species take flight in response to walkers (Burton et al. 2002a). Similar information on the tolerances of different species or species groups to disturbance from ecotourism was obtained by Klein et al. (1995). Site management already employs the use of buffer zones around existing ‘core wildlife areas’, based on the tolerances of key species to human disturbance (Fox & Madsen 1997, Rodgers & Smith 1997, Blumstein et al. 2003). Mathers et al. (2000), for instance, studying Eurasian Wigeon A. penelope and Light-bellied Brent Geese B. b. hrota at Strangford Lough, Northern Ireland, suggested the use of a buffer of 250 m in order to exceed the tolerance distance of the more sensitive of the two species at that site.
It should be noted, however, that the extent to which disturbance does actually affect the distribution of birds within a site and thus their apparent tolerance to different factors will vary both according to the availability of alternative resources and the birds’ own state. Thus, if birds are under stress, e.g. during periods of cold winter weather, and alternative resources are restricted, they may be less easily disturbed than at other times (McGowan et al. 2002, Stillman & Goss-Custard 2002) and as a result show less avoidance of areas associated with human activity, such as construction sites or footpaths, even if the costs for fitness are high (Gill et al. 2001b). As Klein et al. (1995) noted, the information on species’ avoidance responses obtained by landscape approaches would be enhanced by study of the movements of individually marked birds, to show where birds go when disturbed and whether they are able to habituate to human activities.
With increasing numbers of visitors to coastal areas, there is a clear dichotomy for managers of protected areas between the need to protect birds and other wildlife from human disturbance, while still allowing public access to scenic sites and indeed the wildlife the sites support (e.g. Kenchington 1989). Although it may not reflect population-level responses, as noted by Blumstein et al. (2005), the information needed at a site level is often ‘not whether or not a species is negatively affected by human disturbance, but what the probabilities are of any given species using a particular site within a protected area under different levels of disturbance.’ Thus, studies which use landscape approaches to explore the spatial avoidance of key species to sources of disturbance may offer a valuable first step towards site management.