Devil in the detail: New evidence that disturbance does not have a significant impact on the wading birds and shelduck of the Exe estuary This is Part 1 of a summary of the keynote address that I and my colleague, Professor R. A. Stillman (Head of the Life and Environmental Sciences Group in Bournemouth University), gave at the conference arranged by the Estuarine and Coastal Sciences Association (ECSA) that was held in Poole from May 15th-17th 2017. We used data collected by myself on the Exe estuary over the period 20022017 to test the hypothesis that disturbance from human activities has a significant, detrimental effect on those shorebirds – in this instance, mainly waders but also shelducks. These birds feed on the intertidal flats of the estuary when the tide is out and take invertebrates, such as cockles, mussels, clams, ragworms and crabs. I reproduce here the illustrations that we used in the talk.

It is a common-place that proving that something does not have an effect – proving a negative – is challenging, to say the least! The diagram below illustrates the approach we have used since the 1970s to test whether a whole range of human actions have a significant impact on the birds. There is widespread agreement among shorebird ecologists that ‘significant effect’ means decreasing their chances of surviving the winter in good condition or, to put it another way, increasing their mortality rate. The schematic illustration below shows how we tested the ‘disturbance is significant’ hypothesis. As the amount of disturbance increases from left to right, a point will be reached when the birds cannot collect all the energy they need in the time available so that their mortality rate starts to increase above its normal level; this point is the ‘critical threshold’ for disturbance. If the current amount of disturbance falls to the right of this critical threshold, the evidence is consistent with the ‘disturbance is significant’ hypothesis. But if the amount falls to the left of the critical threshold, the evidence does not support the hypothesis – especially, of course, if it falls a long way to the left. [Note that below this threshold, in the ‘ZONE’, there is no point in ‘protecting’ birds from disturbance as it does not affect their mortality rate anyway: this point is illustrated in blue. To claim otherwise is similar to a drugs company selling vitamin C tablets to people whose very healthy diet contains far more of the vitamin than they need anyway! Note also that increasing the amount of disturbance for birds within the ZONE will only have an impact if the increase causes the total amount of disturbance to exceed the critical threshold.]

The rest of this post describes the results of our research to find out where on this schematic graph the amount of disturbance experienced by wading birds and shelduck within the Exe estuary falls. In other words, we test the hypothesis that the current level of disturbance does have a significant effect on these birds. The key issue is the extent to which birds and people overlap in time and in space. Clearly, if birds and people occur together very frequently, it is much more likely that the birds are affected by disturbance than if they do not. The next two illustrations show why, within the Exe estuary itself, there is generally rather little overlap in time and space between wading birds and people. First, when the tide is out, most shorebirds feed on the muddy sediments that people generally avoid. People are seldom seen in winter on the three-quarters of the estuary where mud predominates. People mainly occur in sandy places, such as Cockle Sand, or on easily-accessible mussel beds, such as those at Cockwood.

Second, people walking alongside the estuary do not overlap with the waders on the adjacent mudflats as much as we might suppose. One reason for this is the presence of the railway lines that run alongside each side of the estuary. In many places, this restricts access to the intertidal flats and

to the land alongside . (In general, the birds have become used to the trains themselves, as the next picture illustrates). A second reason is that access is restricted by the terrain and by access rights.

The next illustration shows the places where people and wading birds do overlap in space and time: the ‘overlap areas’. The green lines show where people can walk immediately alongside the intertidal zone where wading birds and shelduck occur. Along the majority of their combined frontage, there is only a risk that birds will be disturbed for about 60-90 minutes at the beginning and end of the exposure period. When the tide is out, most of the birds move downshore to exploit the better feeding areas that occur there and then, of course, they are way beyond the distance at which they might be disturbed by people onshore. There is a similarly limited period of overlap between birds and the people who go out onto the intertidal flats over low tide. As described in the posts in this blog on crab-tiling and angling, most birds have moved away from the overlap areas before the people arrive. And, of course, once birds have been disturbed away, there are fewer or none left for people arriving subsequently to disturb. It is amazing how frequently this rather obvious point is ignored by people assessing the frequency with which birds are disturbed by people!

The next picture shows the numbers of groups of people visiting each ‘overlap’ area during daylight hours during the very dry winter of 2016-17. In many places there were more visits at weekends and during holiday periods than on ordinary weekdays and this has been taken into account.

At the moment, we have two ways of assessing the likely impact of disturbance on the birds. The next picture illustrates the simpler of the two measures: the amount of time and energy expended by the average bird in flying away when disturbed by people. This measure does not include minor effects, such as walking away while continuing to feed which, if included, would increase the estimate somewhat. But on the other hand, it does not allow either for those flights that the birds would have made anyway and which, if included, would decrease the estimates: an example of this would be on an advancing tide when a bird would very shortly have flown anyway to avoid being drowned. The example below is for the oystercatcher which is the species that, on the Exe, has the greatest overlap with people: I have the data to make the same calculation for all the other species but have so far not had enough time to do so. The data show very clearly that disturbance from people causes the average oystercatcher on the estuary to spend only a trivial amount of extra time and energy in flying away from people.

The second measure is much the better of the two, and is illustrated in the next picture. The yellow line shows the extra mortality inflicted on waders and shelduck as the number of groups of people visiting the overlap areas of an estuary increases. The diagram is based on research by Richard

Stillman who used his individual-based model, MORPH, to predict the numbers of groups of people that would need to visit the overlap areas of Southampton Water to increase the mortality rate of waders in winter. MORPH is a powerful technique because it takes into account not only the natural history of each species modelled but also all the most significant consequences for birds of disturbance so far identified: i.e. (i) the increased energy requirements; (ii) the lost foraging time; (iii) the displacement to undisturbed parts of the estuary and the likely increase in competition between birds compressed into a smaller area; and (iv) the ways in which birds can compensate for disturbance by, for example, feeding for longer on the estuary or in fields over high tide. Although the predictions are not for the Exe itself, in terms of its suitability for waders, Southampton Water is not very different to the Exe. The effect of any differences that do exist will be dwarfed by the magnitude of the gap between the numbers of visits to and into the overlap areas of the Exe and the number of visits required on Southampton Water to increase bird mortality. It would take 40 times more visits to occur on the Exe every day for the critical threshold of 30 visits/ha/day to be reached!

This is a very important finding. This value of 30 visits/ha/day is the main evidence used by Natural England to assess when disturbance represents a threat to these birds. A recent Ph.D. project that used a similar model to MORPH in Poole Harbour has shown the critical threshold for the main species that winter there to be about 20 visits/ha/day – not that dissimilar to the value obtained for Southampton Water, and still very much higher than the amount of disturbance I have measured for the Exe estuary. It is therefore clear that the amount of disturbance experienced by waders and shelduck on the Exe is likely to be many times less than the amount that would be required to do them any meaningful harm. This is not say that disturbance, in principle, cannot affect the survival and body condition of shorebirds. Of course it can - but it all depends on its intensity, frequency and duration – just as the EU Directives say. It also depends on other factors too, the most obvious of which is the quality of the food supply available to the birds.

A good example provided by the modelling my team from the Centre of Ecology and Hydrology did many years ago of oystercatchers that winter in the baie de Somme, Picardy, France. Oystercatchers here eat cockles when they are abundant and switch to ragworms when cockles become scarce. Because of their small size and therefore energy content, ragworms are a poorer source of food than cockles. As the next illustration shows, when cockles are abundant, the mortality rate of oystercatchers is not affected by an increase in the frequency with which they are disturbed, even at amounts that are way above the rate that actually occurred in the baie.

We then removed the cockles from the model to explore how disturbance would affect the birds when they were forced to eat ragworms towards the end of the winter: this happens in nature when gales strip the bivalves from the sand. In these simulations, oystercatchers were then feeding on a poorer source of food at the most difficult time of year; i.e. mid- to late winter. The critical threshold is now much lower than when cockles had been abundant throughout the winter. According to these model results, the actual rate of disturbance would then be sufficient to increase the number of oystercatchers starving in late winter. This result from the baie de Somme has implications for the Exe as well. First, the observed rates of disturbance I measured on the Exe are well below the critical threshold in the baie de Somme even when oystercatchers could only feed on ragworms at the end of the winter. Second, the critical threshold for Southampton Water was obtained from just one-third of the estuary where the food supply was relatively poor. Surveys of the food supplies that we had carried out simultaneously on the Exe estuary and in Southampton Water showed that the food supplies on the Exe were rather better. Therefore, if anything, the critical threshold of 30 visits/ha/day obtained from the model for Southampton Water is likely to be too low for the Exe. Accordingly, the gap between the observed amounts of disturbance and the critical threshold may well be even larger on the Exe that it appears to be at the moment. The conclusion is that the hypothesis that wading birds and shelduck are, or are likely to be in the foreseeable future, significantly affected by disturbance is not supported by the data I have collected over the last 15 years of fieldwork and have so far analysed.

Our new evidence strongly supports the conclusion I reached many years ago in my Critique of the Footprint Ecology report (now a post in this blog). On the basis of the evidence I had available at that time, I concluded that it was very unlikely that disturbance had a significant effect on the shorebirds of the Exe estuary. I have a lot more data on wading birds and wildfowl to analyse over the next few months which appear likely to strengthen even further the conclusion given above. The results will be posted on this blog as they emerge.

John Goss-Custard 22nd May 2017