The jury is in on magpies and sparrowhawks

Today sees the publication of two interesting and important studies of the impacts of predator species on their prey's populations.  I'll discuss them both here in this blog.

This post is about the analysis carried out by the BTO (with the University of St Andrews and the Game and Wildlife Conservation Trust) on the huge datasets of the Common Birds Census and the Breeding Bird Survey.

The study looked to see whether predators such as sparrowhawk, magpie, kestrel and buzzard have any discernible impacts on the populations of potential prey such as songbirds.  If this study is the jury deliberating on the question of whether predators are guilty for songbird declines then the verdict is not guilty!

This doesn’t come as a great surprise – it repeats previous findings and is biologically realistic.

Is it quite as clear as ‘Not guilty’?  Almost, but not quite. There might be three reasons for treating the findings with some caution: bad dataset, bad analysis and weak approach.  Let me explain.

Is the dataset a good dataset?  The dataset can’t be bettered in the UK or Europe, and hardly in the world.  It is a long term set of data stretching back for decades with good geographic coverage, good species coverage, good standardisation of methods, high sample size etc etc etc.  No dataset is perfect, but there is hardly a better dataset on long term national bird trends than this.

Is it a good piece of analysis?  It pains me to say that I can’t really comment on this because, despite being a scientist by training, things have moved on sufficiently that I struggle fully to understand the details of this analysis!  However, the authors are very clever and this paper is published in a high quality journal and has been peer-reviewed so we can trust that there are no big errors in it.

Is the approach strong?  By this I mean if there really were an impact of predators on songbird populations would this dataset and analysis pick it up? The answer is yes but since it is a study of correlations it is not as strong an approach as an experiment.  But few would want to start slaughtering predators on a massive and widespread scale to investigate this question more rigorously – there are some though!

So it’s a good study, which will allow a bit of quibbling to continue, but which absolutely knocks on the head the idea, proposed by some, that predators are the main reason for the decline in songbird populations.

Let me try to summarise a complicated story.  To do this I am going to ignore the results from the analysis of the Breeding Bird Survey and from including the great spotted woodpecker as a predator.  I do this to make the story simpler but it doesn’t mean I have missed out the juicy or inconvenient bits – the paper is published, you can go and see for yourselves!

There are actually more positive correlations than negative ones – which suggests that the more prey there are, the more of their predators too, which would make a lot of sense.  Magpies, jays, carrion crows and buzzards look pretty innocent of any impacts on songbird populations.  If one were to imagine any predator having an impact at all from these results then it would be kestrel and sparrowhawk. Let’s look at sparrowhawk in more detail.

Of the 29 prey species, there is some evidence of sparrowhawks having a negative impact on the populations of 18 of them.  However, of the 29 species, the results for only seven prey species are statistically significantly correlated with sparrowhawk numbers – some positively and some negatively.  The seven prey species are:  lapwing, yellow wagtail, robin, nuthatch, tree sparrow, bullfinch and reed bunting.  The two most statistically significant results are that sparrowhawks depress tree sparrow numbers and increase yellow wagtail numbers.  I find it easier to believe the former than I do the latter – but who knows?  And although I can imagine that sparrowhawks might depress both reed bunting and bullfinch numbers I find it more difficult to believe that they increase robin numbers or decrease nuthatch numbers.  And a whole host of regular sparrowhawk prey species (blue tit, great tit, blackbird, song thrush, starling and house sparrow) are unaffected, it seems, by sparrowhawk numbers.

This analysis could have come out very differently – it could have shown clear across-the-board negative correlations between predators and their prey populations.  Or it could have shown just a small number of clear negative correlations for some predators and some prey which were biologically realistic.  Neither happened.  So let’s stop fingering predators as the cause of widespread songbird declines and look to the real causes – unsustainable agriculture, pollution, climate change and the overall pressure of our species on the natural world.

But let’s also wonder what we would have done if this analysis had shown that predators reduce the populations of their prey.  I think the answer is not much!  If that’s what happens then let’s get the habitats right – as we have at Hope Farm – so that predators and their prey can coexist and let’s have many more of both!