The impact of disturbance on the behaviour and energy budgets of Brent Geese Branta b. bernicla

Bird Study ISSN: 0006-3657 (Print) 1944-6705 (Online) Journal homepage: http://www.tandfonline.com/loi/tbis20 The impact of disturbance on the behaviour and energy budgets of Brent Geese Branta b. bernicla R. Riddington, M. Hassall, S.J. Lane, P.A. Turner & R. Walters To cite this article: R. Riddington, M. Hassall, S.J. Lane, P.A. Turner & R. Walters (1996) The impact of disturbance on the behaviour and energy budgets of Brent Geese Branta b. bernicla, Bird Study, 43:3, 269-279, DOI: 10.1080/00063659609461019 To link to this article: https://doi.org/10.1080/00063659609461019 Published online: 29 Mar 2010. Submit your article to this journal Article views: 592 View related articles Citing articles: 36 View citing articles Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalinformation?journalcode=tbis20 Download by: [46.3.193.172] Date: 31 December 2017, At: 11:34

Bird Study (1996) 43, 269 279 The impact of disturbance on the behaviour and energy budgets of Brent Geese Branta b. bernicla R. RIDDINGTON*, M. HASSALL, S. J. LANE, P. A. TURNER and R. WALTERS School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK We studied disturbance factors that caused Brent Geese along a stretch of the north Norfolk coast to take flight. Patterns of activity and undisturbed flight were also quantified, to derive energy budgets. The most frequent source of disturbance was pedestrians. Those disturbances resulting in greatest energy expenditure were also of human origin, but tended to be mechanised (e.g. aircraft, gunfire). Activity budgets for high- and low-disturbance days were compared, which showed that birds feed less and are more vigilant when disturbance is greater. On grass pastures, 68.9% of flight was a response to disturbance, which increased estimated hourly energy expenditure (HEE) from 32.2 to 35.6 J/h (10.8%). On exceptionally disturbed days, HEE increased to more than 44.0 J/h. As a consequence of disturbance, geese may need to feed at night for up to an hour in mid-winter, to balance their daily energy budget. Unless they are able to feed easily at night, disturbance may be one of the primary factors influencing local distribution of Brent Geese. Management implications are discussed. Disturbance may be defined as discrete events that disrupt ecosystem, community or population structures, or in some way alter resource levels. 1 Disturbance may be natural (such as fire, flooding) or of human origin, but the latter is currently the cause for greatest concern. This paper examines the impact of disturbance on wintering Dark-bellied Brent Geese. Approximately half the world population of Dark-bellied Brent Geese, winters in the UK, which confers international responsibility to protect this population. 2 When considering disturbance of wintering bird populations, several issues are important; in particular, changes in local distribution and how these may relate to total food supply, foraging efficiency and compensation for increased energy expenditure due to flight. 3 Geese are heavy birds, and hence flight imposes considerable energetic costs. They are herbivorous, feeding in winter on grass foliage, *Correspondence author at the Fair Isle Bird Observatory, Shetland, Fair Isle ZE2 9JU, UK. saltmarsh plants, intertidal algae and newly emergent crops such as barley and wheat. Compared to carnivorous or invertebrate feeders, their food sources are nutrient-poor and difficult to digest: consequently, energy assimilation is relatively inefficient (Lane, unpublished data). Further constraints are imposed (for day-time feeders) by short daylength, adverse weather and the need for a safe night roost site. Lack of growth, depletion and seasonal die-back cause food supplies to decline through the winter faster than they are replenished. 4,5 The need to balance daily energy budgets therefore imposes a considerable restriction on behaviour. Consequently, disturbance may have a substantial impact on population distribution and individual survival, while there may also be consequences for reproductive output. These constraints mean that wintering geese need to feed for a large proportion of the day to satisfy energy requirements. Choice of feeding site is determined by trade-offs between energy gain (from food intake) and energy expenditure 1996 British Trust for Ornithology

270 R. Riddington et al. (through foraging, maintenance and movement to and from feeding sites). Since the response to disturbance often involves flight, which energetically is by far the most expensive of normal daily activities, 6 disturbance has potentially two main effects : (i) decreasing time available for foraging and (ii) increasing energy expenditure as a result of extra flight. Possible responses to disturbance include : (i) changing feeding site and/or diet, if alternatives are available; (ii) increasing the amount of time spent foraging; (iii) increasing intake or assimilation rate; and/or (iv) increasing the level of night-time feeding (if disturbance is lower at night). If none of these options is available, geese may incur an energy deficit and lose weight. 7 By affecting both the intake and expenditure sides of the energy equation, disturbance is potentially an important factor for many goose populations. Snow Geese Chen caerulescens staging in Quebec were estimated to suffer a decrease of 4 51% in daylight foraging time due to disturbance, with a loss of up to 2.9% of hourly metabolizable energy intake and an increase in hourly energy expenditure of up to 5.3%. 8 An average seasonal disturbance rate of 0.5 events/hour was estimated to cause a reduction of 20.4% in the energy reserves of staging Snow Geese. 9 In Britain, wintering Brent Geese more than doubled flying time due to disturbance, losing 4.9 11.7% of daytime feeding; 10 daytime energy expenditure of disturbed Brent Geese increased by 31%. 11 Disturbance is therefore a critical factor affecting migrating and wintering geese, and it is important to integrate the costs of maninduced disturbance when calculating energy budgets. 10 This paper addresses the following questions. What are the main causes of disturbance to wintering Brent Geese, and how do they respond to them? What effect does disturbance have on daily activity patterns and energy consumption? How does disturbance affect ways in which geese may overcome constraints imposed by their environment? What are the implications of this study for refuge management? METHODS Study area This study was carried out on the north Norfolk coast (Fig. 1), an area which covers the home range of a wintering population of Brent Geese (currently 3000 4500 individuals), which roosts on Scolt Head island. 13 Disturbance and activity data were collected Figure 1. Study site (after Summers & Critchley, 1990). 13

Disturbance behaviour and energy budgets of Brent Geese 271 between November 1992 and March 1993 on saltmarsh and grass pastures. As flocks of Brent Geese in winter 1992 93 consisted almost exclusively of adults, it is assumed that there was almost complete breeding failure in 1992. 14 During winter, Brent Geese in the study area usually feed on saltmarsh for a few hours in the morning, before flying to grass pastures where they feed until dusk. 13 Dawn to dusk watches were carried out every three days on the waning moon cycle during January, February and March. This gave five dawn to dusk watches for each of the first three months of the year. In addition, goose activity and disturbance events were recorded as a routine part of fieldwork on 19 other days between November and March, when the geese were not monitored from dawn to dusk. Observations were carried out, using tripod-mounted telescopes, from public footpaths in the area. Observers were usually more than 100 m from the geese; at this distance we recorded no increase in vigilant behaviour of the geese. Disturbance A disturbance event was defined when 75% or more of the birds in a flock suddenly became alert, shortly followed by all or part of the flock taking flight. Disturbed flight was distinguished from natural flight by the criterion that, for a disturbed flight, birds take off almost simultaneously. 15 For all disturbance events observed, the following information was recorded: (i) flock size; (ii) number disturbed; (iii) nature of response (whether they returned to the same site, an adjacent area, or left the area completely); (iv) the disturbance source, defined as pedestrians (with or without dogs), vehicles or boats, aircraft, gunfire, mammals (including stock; not dogs), birds and unknown ; and (v) duration of the resulting flight. For large flocks, the time of both the first and last birds to take off and land was recorded, and a mean of the two taken. The number of goose-minutes of flight due to disturbance was calculated for each event by multiplying the number of birds by the time in flight. The magnitude of response to disturbance events was assessed by comparing the mean duration of flight per event caused by each disturbance source, which reflects relative energy expended in response to different sources. Activity budgets Diurnal activity was monitored every 15 20 minutes by scanning the flock and recording the behaviour of at least 100 birds, or the entire flock if smaller than this. 16 The birds behaviour was divided into the following categories: feeding, vigilant, preening, roosting, aggressive behaviour, drinking and other. Birds were sampled from both the edge and centre of flocks, as individuals in different parts of a flock tend to behave differently. 17 Flight The time spent in flight (excluding the morning and evening flights from and to the roost) was calculated on 15 days when continuous dawnto-dusk watches were carried out. Once a flock was located, a vantage point overlooking the study area was chosen and the flight times of the birds recorded. If birds flew out of sight, duration of their flight could be estimated accurately, as their destination (and average flight times to these areas) would be known. Activity and disturbance were recorded simultaneously, so the proportion of flight due to disturbance could be calculated directly. Activity time budgets can be converted to energy expenditure by using a simple formula: Energy consumed = (Cost of the activity basal metabolic rate) % of time occupied in that activity. 14 Based on figures in Wooley & Owen, 18 costs for various activities (as multiples of basal metabolic rate, BMR) can be estimated as follows: resting, 1.1 ; preening, 1.6 ; feeding, 1.7 ; vigilance, 2.1. We estimated aggressive encounters to be approximately 6.4 BMR. 14 In addition to these, flight can be estimated at anywhere between 8.0 and 14.0 BMR. 16,18 20 Based on Tucker, 22 we calculated flight costs as a multiple of BMR for Brent Geese in our study area, timed whilst flying known distances, to be 14.9 BMR; we have therefore used the multiple of 14.0 as a conservative estimate of flight cost that is within the accepted range of values. From Lasiewski & Dawson, 23 BMR can be calculated as 17.1 J/h, assuming an average mass for an adult Brent Goose of 1365 g. 24 From these figures it is possible to estimate energy expenditure and the increase resulting from disturbance.

272 R. Riddington et al. 60 a 15 b 50 Number of events 40 30 20 Number of events 10 5 10 0 Figure 2. Frequency of disturbance sources affecting wintering Brent Geese grazing on (a) grass pastures and (b) saltmarsh, Norfolk, UK, November 1992 to March 1993; 179 disturbance events observed. RESULTS Disturbance pedestrians vehicles aircraft gunshots Source mammals birds unknown A total of 179 disturbance events were recorded, 89.4% on pastures and 10.6% on saltmarsh areas. The mean disturbance rate was 0.83 events/hour, 0.91 per hour on pastures (n = 72), compared to 0.53 per hour on saltmarsh (n = 11) (χ 1 2 = 2.38, ns). A total of 160 disturbance events were observed on pastures, the most frequent source being pedestrians (31.9%), followed by aircraft (19.4%), then unknown factors (18.1%) (Fig. 2a). Non-human causes triggered a relatively low proportion of disturbance events: mammals and birds combined accounted for just 20.6% of all disturbance events. Most bird disturbance was caused by raptors (in particular Hen Harriers Circus cyaneus, but occasionally also Sparrowhawks Accipiter nisus and Merlins Falco columbarius when they flew low and fast above the flock), but also by large, slow, raptor-like birds, particularly Grey Herons Ardea cinerea. Disturbance from mammals was primarily a response to hares Lepus europaeus, which, when running through the flock, would often cause the geese to fly short distances. 0 pedestrians vehicles gunshots Source birds unknown Flight duration in response to aircraft and gunshots was much higher than that in response to other sources; response was lowest to non-human sources, i.e. birds and mammals (Table 1). The percentage of each flock disturbed was recorded; there were no significant differences between disturbance sources but, on average, a higher percentage of the flock responded to aircraft and vehicles compared with mammals and birds, which again drew the least response. The total sum of goose-minutes in flight as a result of disturbance sources, which takes into account both magnitude and frequency of events and the number of geese affected (Table 1), shows pedestrians, aircraft and unknowns as high impact causal factors, followed to a lesser extent by gunshots. The low position of mammals and birds again suggests that these are not perceived by geese as a particularly threatening source of disturbance. To investigate whether there was any relationship between disturbance and date (which might occur if geese habituated to certain disturbance sources, or if the need to increase food intake late in the season caused them to become more tolerant of disturbance), distur-

Disturbance behaviour and energy budgets of Brent Geese 273 Table 1. Magnitude of response of wintering Brent Geese to disturbance events, grass pastures, Norfolk, UK, 1992 1993 Mean no. seconds Total goose-minutes Percentage of flight per event of disturbed flight flock disturbed Pedestrians 69.5 ± 7.2 51.593 79.9 ± 5.5 Vehicles 71.3 ± 23.1 9.533 100 ± 0 Aircraft 102.7 ± 9.9 48.912 85.8 ± 4.6 Gunshots 90.1 ± 24.4 33.629 79.4 ± 9.7 Mammals 49.5 ± 12.9 13. 211 67.0 ± 12.4 Birds 56.4 ± 10.4 17.180 74.9 ± 8.5 Unknown 68.5 ± 8.0 47. 559 83.1 ± 5.9 Statistics: one-factor ANOVA F = 2.62, P = 0.019 F = 0.716, P = 0.64 (Difference significant at P = 0.05 (Data are arcsin transformed) Aircraft vs pedestrians, mammals, birds & unknown bance flight duration was plotted against date, for dawn to dusk observation days on pastures. There was no relationship between date and disturbance (r 14 = 0.148, ns). On saltmarsh, where 19 disturbance events were observed, pedestrians were the most frequent source of disturbance (63%), followed by boats (21%, Fig. 2b). No other disturbance sources were recorded more than once. Pedestrian disturbance resulted in a mean of 113.4 ± 27.3 (se), seconds of disturbance per observed event and a total of 5153 gooseminutes of disturbance, making it the most important cause of disturbance on this habitat. Activity Brent Geese in our study area spent most of their time feeding: 79.1% on pastures, 69.6% on saltmarsh (Fig. 3). This difference between habitats was significant (t = 3.10, df = 142, P = 0.0024). On pastures, the majority of non-feeding time was devoted to vigilant behaviour (18.4%), with little time invested in other activities such as preening (1.8%) or roosting (0.2%). This was also true for saltmarsh, although here birds spent significantly more time being vigilant (21.2%: t = 2.03, df = 148, P = 0.044), roosting (2.4%: t = 3.91, df = 132, P = 0.0001) and preening (5.4%: t = 2.77, df = 127, P = 0.007). Geese on saltmarsh also spent less time in flight than those on pasture, although the difference was not significant (1.6% versus 2.4%; t = 1.762, df = 23, P = 0.091). There was no clear relationship between date and percentage time feeding in either habitat (pastures: y = 0.009x + 78.691, r 2 = 0.003, P = 0.747; saltmarsh : y = -0.17x + 67.14, r 2 = 0.001, P = 0.913), suggesting that neither daylength nor season affects proportion of time spent feeding. On pastures, there was a significant negative relationship between the proportion of time spent feeding and time of day, measured as hours before sunset, r 954 = 0.167, P = 0.001. This suggests that geese feed more intensively as dusk approaches. We do not have sufficient data to determine whether geese feed more intensively at dawn on the saltmarsh. Birds spent proportionately (though not significantly) more time feeding and less time vigilant in large flocks, which is possibly an effect of birds in small flocks needing to spend more time alert in order to maintain a given mean level of vigilance. Impact of disturbance on activity To estimate the effect of disturbance on diurnal activity patterns, activity budgets recorded for pastures on low disturbance days were compared with those on high disturbance days (there were insufficient observation-hours on saltmarsh to achieve this distinction). Only dawn-to-dusk days were used in this analysis, when both disturbance and activity were monitored for most of the day. Observation time of geese on pastures on these days varied between 3.5 and 6.6 hours (mean 5.3 ± 0.25

274 R. Riddington et al. feeding vigilant preening roosting aggression drinking other flying 0 20 40 60 80 Frequency Figure 3. Mean activity budgets of wintering Brent Geese, grazing on grass pastures, Norfolk, UK, November 1992 to March 1993; 19 disturbance events observed. First seven categories describe non-flight activity budgets; hours). They were divided into high (mean level of disturbance > 80 seconds/hour) and low (mean level of disturbance < 80 seconds/hour) disturbance days, based on the amount of disturbance per hour of observation. This gave eight high disturbance days (mean 132.0 ± 15.8 s/h; range 87.4 217.1 s/h) and seven low disturbance days (mean 44.6 ± 6.9 s/h; range 18.4 67.6 s/h); high and low disturbance days were spread fairly evenly throughout the study period, with examples of both from each moon cycle of observations. Geese spent less time feeding and roosting on high disturbance days, and more time in vigilant behaviour and preening (Table 2). The difference in time spent feeding (t = 1.94, df = 258, P = 0.054) may be biologically significant. Since these activity budgets do not include time in flight, and this is greater on high disturbance days (Table 2), the real differences in the Table 2. Mean activity budgets (% time) of wintering Brent Geese in Norfolk, UK, 1992 93 High disturbance days (n = 8) Low disturbance days (n = 7) Activity Mean ± se Mean ± se Feeding 78.4 0.73 80.2 0.76 Vigilant 18.9 0.69 17.8 0.73 Preening 1.9 0.22 1.4 0.18 Roosting 0.2 0.06 0.2 0.08 Aggression 0.6 0.08 0.3 0.05 Drinking 0.1 0.05 0.1 0.04 Other 0.0 0.01 0.0 Flying 3.1 0.60 1.6 0.37 Days with mean recorded levels of disturbance > 80 s/h are classified as high disturbance days; low disturbance days are below this threshold.

Disturbance behaviour and energy budgets of Brent Geese 275 proportion of the day spent feeding are slightly greater than this, calculated to be 75.8% versus 78.8% (t = 0.958, df = 13, ns). Flight The proportion of time spent in flight due to disturbance differed substantially between pasture and saltmarsh (Fig. 4). On pastures, geese spent an average of 2.4% of their time in flight, of which 68.9% was due to disturbance. On saltmarsh, geese spent less time in flight, 1.3%, but little of this (2.5%) was due to disturbance. Consequently, without disturbance, geese would spend less time flying on pastures (0.75%) compared to saltmarsh (1.25%). At our study site there is an additional cost to feeding on pasture, as geese have an additional 3.13 minutes of flight (on average) to reach the nearer site (Burnham Norton) and an additional 4.15 minutes of flight to reach the farther site (Burnham Overy) used for feeding, compared to saltmarsh. As the amount of time that geese spend on each habitat in our study area varies with daylength and tide, the increase in energy expenditure due to disturbance was calculated in terms of hourly energy expenditure (HEE) (Table 3). These results give confidence to our estimates of the cost of various activities, as for Brent Geese the average daily energy expenditure (DEE) should be approximately 841.4 J/day, equivalent to 2.05 BMR. 12 The estimates of HEE in Table 3 range between 1.94 and 2.19 BMR. The impact of disturbance on energy expenditure of geese on saltmarsh is negligible; only 2.5% of flight, on average, is Time in flight (%) Pasture Habitat Saltmarsh Figure 4. Percentage of time spent in flight by wintering Brent Geese, due to disturbance; Norfolk, UK, 1992 1993., Undisturbed;, disturbed. due to disturbance and this is responsible for a 0.013 J/h difference in energy expenditure. In contrast, on pastures, where an average of 68.9% of flight results from disturbance, there was an increase in HEE from 33.8 to 37.5 J/h (10.9%). (This of course does not take into account the decrease in food intake due to less time spent grazing). On average, since Brent Geese wintering in our study area spend about 5 6 hours per day feeding on pastures, the increase in flight due to disturbance represents an increased energy expenditure of around 18.4 22.2 J. Given that the intake rate of feeding Table 3. Mean hourly energy expenditure (J/h) grouped by activity, of wintering Brent Geese, Norfolk, 1992 1993 Activity Pasture Pasture Saltmarsh (high disturbance) (low disturbance) Feeding 22.1 22.9 20.0 Vigilant 6.6 6.4 7.5 Preening 0.5 0.4 1.5 Roosting 0.04 0.04 0.5 Aggression 0.8 0.3 0.5 Drinking 0.04 0.04 0.04 Flying 7.5 3.8 3.1 Sum 37.5 33.8 33.2

276 R. Riddington et al. Brent Geese on grass pastures is approximately 124.7 J/h (Lane, unpublished data), the geese would need to spend, on average, an extra 9 11 minutes per day grazing on pastures. An extra 15 19 minutes of saltmarsh feeding would be required to make up this deficit, because we found intake rates to be lower on the saltmarsh (approximately 71.6 J/h). On some days, the amount of feeding needed to compensate for the increase in energy expenditure due to disturbance may be even higher. For example, on 20 March 1993, during a 5.5- hour observation period on the pastures at Burnham Overy, a flock of around 400 Brent Geese spent an average of 4 minutes in flight every hour. Most of this (99%) was attributable to known disturbance events and to compensate for the increase in HEE (10.8 J/h) would require an extra 29 minutes of feeding per day on pasture, or an extra 50 minutes feeding per day on saltmarsh (given that the geese spent 5.5 hours on the pastures that day). Such an increase would almost certainly not be made up that day, unless the birds fed during the night. DISCUSSION Disturbance In our study area, 0.83 disturbance events per hour were recorded, similar to that for wintering Brent Geese in Essex, UK (0.74/h), 10 although less than reported in North America for staging Snow Geese (1.26/h; 15 1.4 2.6/h 25 ) or spring-staging Brent Geese in the Wadden Sea (2.19/h 26 ). As we only measured disturbances severe enough to cause geese to take flight, these figures represent a minimum figure for disturbance impact. In our study area, the choice of grazing habitats for wintering Brent Geese is effectively saltmarsh or pasture (scaring of geese from arable crops is highly efficient and birds are never allowed to feed there for any length of time). We found disturbance rate to be slightly (although not significantly) higher on pastures. Geese also spent more time in flight as a result of disturbance on pastures (Fig. 4), and these findings suggest that disturbance impact was greater on pastures. This may reflect easier public access, and also that pastures are grazed later in the day, when some forms of disturbance are more common (particularly those relating to human recreation). Human-initiated events are by far the most important cause of disturbance, and non-human disturbance is usually small, in both habitats. Pedestrians are the most common cause of disturbance in both habitats (mostly walkers, bird watchers, bait-diggers or wildfowlers), the majority along public paths. However, at least on pastures, the overall impact of this source (taking into account time in flight and proportion of the flock affected) was similar to or less than mechanized forms of human disturbance, e.g. shooting, aircraft and vehicles (Table 1). Unknown factors contribute substantially but our impression is that often there is no extrinsic cause. If so, and these are not true disturbance events, the overall rate of disturbance falls to 0.68/h. On saltmarsh, pedestrians were clearly the most important source of disturbance. This finding is biased by the timing of our observations (which reflect the behaviour of the geese). For example, no aircraft disturbance was recorded on the saltmarsh, which suggests that aircraft in this area are only active at certain times of day. The effect of flock size on disturbance is also important. Large flocks are likely to be more wary, and less tolerant of disturbance than small flocks, simply because there is a greater chance of large flocks containing jumpy members which are liable to startle the rest of the flock. 10,15,27 Geese feeding on saltmarsh tended to do so in smaller flocks, mostly fewer than 200, as opposed to on the pastures, where they fed mostly in large flocks of up to several thousand individuals. We found no evidence that geese habituate to disturbance. Similarly, the geese did not habituate to repeated scaring from adjacent arable crops, but avoided these areas almost completely, unless disturbance levels on the pastures were exceptionally high (Lane, unpubl.). Activity Behaviour patterns were similar for both habitats, with the majority of time spent grazing; however, this was significantly lower on saltmarsh, with a corresponding increase in vigilant behaviour, preening and roosting. Geese on saltmarsh also spent less time in flight, but this can be explained by differences

Disturbance behaviour and energy budgets of Brent Geese 277 in disturbance (Fig. 4). Our observations suggest that time spent feeding varied throughout the day, with a lull in feeding activity around midday and an increase towards dusk. This has been reported in other species of wintering geese (e.g. Bean Geese Anser fabilis 28 and White-fronted Geese Anser albifrons 29 ), perhaps so that digestion of a full gut can occur during the long winter night. However, variation in the percentage time spent feeding through the day was rather small (hourly means ranging from 75.9 to 82.4%), suggesting that these geese feed at, or close to, their maximum possible rate throughout the day. The lack of any relationship between percent feeding and date further suggests that Brent Geese feed at or near their maximum rate throughout the winter (it might be expected that the geese would feed more intensively during the shorter daylight hours of mid-winter); and/or that they are foraging at night to a greater or lesser extent. Results from a parallel study on this Brent Goose flock suggests that, on average, geese feed for almost 20% of the night, on saltmarshes adjacent to the roost. 30 Clearly, this flock can compensate somewhat for disturbance by night-time feeding, although this may not always be possible (e.g. on neap tides, as at night birds tend to feed around the high tide). Flocks in other areas, however, may not be able to feed so easily at night, in close proximity to their roost. Impact of disturbance on behaviour Activity budgets on high and low disturbance days showed a broadly similar pattern, rather than a radical change of behaviour in response to disturbance. However, as expected, birds spent less time feeding on days when disturbance was high, with a corresponding increase in vigilant behaviours. Disturbance clearly increases flight time (Fig. 4), and this increases energy expenditure (see below). In extreme cases, disturbance caused birds to leave their feeding site, but usually the birds would return to where they were disturbed. In many cases, geese disturbed from the pastures at Burnham Norton would fly to nearby freshwater pools to drink and bathe, from where they usually returned to their original feeding area within 20 30 minutes. Occasionally, disturbance would result in birds switching to another habitat (e.g. forsaking pastures for the rest of the day, and returning to saltmarsh). While disturbance may have been the proximal cue for birds to switch, we cannot say whether or not this was the ultimate cause. Flight Disturbance has, potentially, a considerable impact on the energy expenditure of Brent Geese in our study area (Table 3). We calculated that, on average, disturbance increased hourly energy expenditure (HEE) by 10.8%, which over five hours would require an additional 8.3 minutes of feeding on pastures, or an extra 14.6 minutes of feeding on saltmarsh to compensate (equivalent to an extra 1.7 and 2.9 min/h of feeding, respectively). On days in our study area when disturbance was greatest, the increase in HEE could be as much as 38.5% (12.4 J), which over five hours requires an extra 29.8 minutes feeding on pasture, or 52.1 minutes on saltmarsh, to make up the shortfall. Although it should be remembered that the figures used to calculate energy expenditure are fairly crude estimates, it is more likely that the energetic cost of flight (estimated at 14 BMR) is underestimated, as more energy is required for take-off than level flight. Frequent disturbance may therefore be even more costly than we have estimated. Since our data suggest that the geese are already feeding at or close to the maximum possible rate during the day, the consequences of such increased energy demand may be critical during mid- or latewinter when daylength is at its shortest, food reserves have been substantially depleted and regrowth of plants is lowest. A net energy deficit was reported for Barnacle Geese Branta leucopsis on Islay in December and January, 8 and this was attributed to the fact that, because the population did not feed at night, they were simply unable to feed enough to maintain a positive energy budget during short midwinter days. The main impact of disturbance is therefore not that it reduces the time spent feeding (and thus food intake per unit time is lower when birds are disturbed), but that it increases the amount of time spent in flight, which is extremely expensive in terms of energetic cost. We suggest that unless birds are able to feed easily at night to compensate for daytime disturbance, then disturbance may be one of the primary factors influencing the local

278 R. Riddington et al. distribution of Brent Geese within their wintering areas in Britain. Management implications On goose refuges, we suggest that human access should be restricted to clearly defined routes, well away from from grazing areas. However, regulating disturbance from highimpact sources (e.g. aircraft), not only within the refuge but also on adjacent areas, may be far more effective in providing an undisturbed site. Although we have no data to support this, we believe that imposing a minimum height for aircraft would be very effective, possibly no lower than 500 m. 26 Since geese tend to land in the same or an adjacent area when disturbed (in our study area, <20% of all disturbed flocks left the area completely) we suggest that a single, large refuge may be more effective than two separate, smaller sites, because the birds would not have to fly between them. Proximity to roost is also very important. The applicability of these measures will vary with region, according to disturbance magnitude and frequency. For example, if an area is subject to prolonged and intense disturbance, such as aerial spraying, rather than intermittent events, such as walkers or jet aircraft, then smaller, separate refuges may be more effective. ACKNOWLEDGEMENTS We are grateful to Viscount Coke for permission to work on the Holkham estate, R. Harrold and M.E.S. Rooney (English Nature) for their cooperation through the project, Dr J. Vickery for comments on a first draft of the manuscript, and W.C. Christie for help with fieldwork. The work was funded by the Agriculture and Food Research Council, grant AG 83/569 and the Natural Environment Research Council, grant GR 9/827, awarded to MH. Referees B.S. Ebbinge and A.D. Fox made constructive comments on an earlier draft of the paper. REFERENCES 1. White, P. S., & Pickett, S.T.A. (1985) Natural disturbance and patch dynamics: an introduction. In The Ecology of Natural Disturbance and Patch Dynamics (eds S.T.A. Pickett & P.S. White), pp. 3 15. Academic Press, New York. 2. Owen, M. (1990) The damage-conservation interface illustrated by geese. Ibis, 132, 238 252. 3. Hockin, D., Ounsted, M., Gorman, M., Hill, D., Keller, V. & Barker, M.A. (1992) Examination of the effects of disturbance on birds with reference to its importance in ecological assessments. J. Enironv. Manage., 36, 253 286. 4. Drent, R.H. & Prins H.H.T. (1987) The herbivore as a prisoner of its food supply. Geobotany, 10, 131 147. 5. Patterson, I.J. (1991) Conflict between geese and agriculture; does goose grazing cause damage to crops? Ardea, 79, 179 186. 6. Tucker, V.A. (1969) The energetics of bird flight. Sci. Am., 220 (5), 70 78. 7. Owen, M., Wells, R.L. & Black, J.M. (1992) Energy budgets of wintering Barnacle Geese: the effects of declining food resources. Ornis Scand. 23, 451 458. 8. Belanger, L. & Bedard, J. (1990) Energetic cost of man-induced disturbance to staging Snow Geese. J. Wildl. Manage., 54, 36 41. 9. Davis, R.A. & Wiseley, A.N. (1974) Normal behaviour of snow geese on the Yukon-Alaska North slope and the effects of aircraft-induced disturbance on this behaviour, September, 1973. In Gunn, W.W.H., Richardson, W.J., Schweinsburg, R.E. & Wright, T.D., eds. Studies on Snow Geese and Waterfowl in the Northwest Territories, Yukon Territory and Alaska, 1974. Can. Arct. Gas Stud. Ltd., Biol. Rep. Series 27. 10. Owens, N.W. (1977) Responses of wintering Brent Geese to human disturbance. Wildfowl, 28, 5 14. 11. White-Robinson, R. (1982) Inland and saltmarsh feeding of wintering Brent Geese in Essex. Wildfowl, 33, 113 118. 12. Drent, R., Ebbinge, B.S. & Weijand, B. (1979) Balancing the energy budgets of Arctic-breeding geese throughout the annual cycle: a progress report. Verh. Orn. Ges. Bayern, 23, 239 264. 13. Summers, R.W. & Critchley, C.N.R. (1990) Use of grassland and field selection by Brent Geese. J. Appl. Ecol., 27, 834 846. 14. Summers, R.W. (1986) Breeding production of Dark-bellied Brent Geese Branta bernicla bernicla in relation to lemming cycles. Bird Study, 33, 105 108. 15. Belanger, L. & Bedard, J. (1989) Responses of staging greater Snow Geese to disturbance. J. Wildl. Manage., 53, 713 719. 16. Altmann, J. (1973) Observational study of behaviour: sampling methods. Behaviour, 49, 227 267. 17. Black, J.M., Carbone, C., Wells, R.L. & Owen, M. (1992) Foraging dynamics in goose flocks: the cost of living on the edge. Anim. Behav., 44, 41 50. 18. Wooley, J.B. & Owen, R.B. (1978) Energy costs of activity and daily energy expenditure in the Black Duck. J. Wildl. Manage., 42, 739 745. 19. Lasiewski, R.C. (1963) Oxygen consumption of

Disturbance behaviour and energy budgets of Brent Geese 279 torpid, resting active and flying hummingbirds. Physiol. Zool., 36, 122 140. 20. LeFebvre, E. A. (1964) The use of D 2 O 18 for measuring energy metabolism in Columbia livia at rest and in flight. Auk, 81, 403 416. 21. Tucker, V.A. (1968) Respiratory exchange and evaporative water loss in the flying budgerigar. J. Exp. Biol., 48, 67 87. 22. Tucker, V.A. (1971) Flight energetics in birds. Am. Zool., 11, 115 124. 23. Lasiewski, R.C. & Dawson, W.R. (1967) A reexamination of the relation between standard metabolic rate and body weight in birds. Condor, 69, 13 23. 24. Cramp, S. & Simmons, K.E.L., eds. (1977) Birds of the Western Palearctic, Vol. 1, Ostriches Ducks. Oxford University Press, Oxford. 25. Prevett, J.P. & MacInnes, C.D. (1980) Family and other social groups in Snow Geese. Wildl. Monogr., 71, 45pp. 26. Stock, M. (1992) Effects of man-induced disturbance on staging Brent Geese. Netherlands Institute for Sea Research Publication Series No. 20. 289 293. 27. Madsen, J. (1985) Impact of disturbance on field utilization of Pink-footed Geese in West Jutland, Denmark. Biol. Conserv., 33, 53 66. 28. Allport, G. (1991) The feeding ecology and habitat requirements of overwintering western taiga Bean Geese. Unpubl. Ph. D. thesis, University of East Anglia. 29. Owen, M. (1972) Some factors affecting food intake and selection in White-fronted Geese. J. Anim. Ecol., 41, 79 92. 30. Lane, S.J. & Hassall, M. (1995). Nocturnal feeding by Dark-bellied Brent Geese Branta bernicla bernicla. Ibis, 138, 291 297. (MS received 21 March 1995; revised MS accepted 4 October 1995)