The rise of the generalists (BTO) (1 Viewer)

I'm not a scientist, so ignore me if you like.
But I'm willing to bet that this "sweeping generalisation" is true sometimes, not other times.
Some species must "move northwards" because their former habitat no longer allows them to breed successfully. And some may indeed be out-competed by their southern relatives moving northwards.
The amount of science needed to study this is huge, and I have no idea how much has been / is being done already.

CPBell said:

For the record, I’ve received replies to the mail posted above from David Currie and Martin Sykes, the editors of Global Ecology and Biogeography, providing assurances about the peer-review process for the article. However I’ve received no response or acknowledgement from the author. If this is still the case by the middle of next week, I’ll call Lund University to find out whether there’s any intention to respond, and post the outcome here.

http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur

Click to expand...

Why don't you either

1) publish a rebuttal
2) contact one of your former co-authors who was also an author on the paper?

The lead author in this case may not be the best person to go hounding...

ZanderII said:

Why don't you either

1) publish a rebuttal
2) contact one of your former co-authors who was also an author on the paper?

Click to expand...

And how about

3) Have a re-read of your 'critique' and see that it's actually baseless and rather silly... Your argument is that most of the species classified as 'specialist' are waterbirds, and that the trends observed could be caused by a climate-driven shift (decline) in waterbird populations. Are waterbirds not specialists now? I don't see how this waterbird angle contradicts the general message of the paper - that habitat specialists (including waterbirds) are declining whilst generalists increase.

Your other criticism was that the negative trend in the 'specialism index' could still arise if specialists are increasing, as long as generalists are increasing quicker. So you're saying that their conclusions are false because there's a chance that everything in the UK is currently increasing... Does that sound reasonable, seriously?

ZanderII said:

Why don't you either

1) publish a rebuttal
2) contact one of your former co-authors who was also an author on the paper?

The lead author in this case may not be the best person to go hounding...

Click to expand...

When a multi-authored paper is published in a peer-reviewed journal, one of the authors is nominated as contact for correspondence. This is not always the lead author, but in this case it is, since Catherine Davey was nominated as correspondence author. It’s entirely appropriate, therefore, that I try to contact her if I think there is a serious flaw in the publication – indeed, it would be a dereliction not to do so.

The editor of Global Ecology and Biogeography has suggested that I submit a correspondence piece setting out my concerns, and I’ll consider doing so. However it would be premature to do this without giving the author an opportunity to respond to the criticisms I’ve put forward. It may be that she has not replied because she is away on fieldwork or simply on vacation. However, if it turns out that she does not wish to respond, I’ll be happy to respect that. There is no question of anyone being ‘hounded’.

'roy said:

And how about

3) Have a re-read of your 'critique' and see that it's actually baseless and rather silly... Your argument is that most of the species classified as 'specialist' are waterbirds, and that the trends observed could be caused by a climate-driven shift (decline) in waterbird populations. Are waterbirds not specialists now? I don't see how this waterbird angle contradicts the general message of the paper - that habitat specialists (including waterbirds) are declining whilst generalists increase.

Click to expand...

Waterbirds are specialists by virtue of their dependence on water, and it may be this specific relationship, rather than specialism per se that has caused the observed patterns to emerge. To establish such a relationship it would be necessary to show that the patterns occur independently of such confounding variables, e.g. do they hold within the set of waterbirds or within the set of terrestrial birds? The fact that waterbird trends are less favourable than those in terrestrial birds tells us nothing about whether generalist terrestrial birds are doing better than terrestrial specialists.

'roy said:

Your other criticism was that the negative trend in the 'specialism index' could still arise if specialists are increasing, as long as generalists are increasing quicker. So you're saying that their conclusions are false because there's a chance that everything in the UK is currently increasing... Does that sound reasonable, seriously?

Click to expand...

That is not what I’m saying, but your misinterpretation of it. My point was that the inference around which the paper’s discussion revolves is formally invalid. For the CSI index to decline, all that is required is that the average population trend among generalists is more favourable than that among specialists. An average trend across species can be increasing when many of the species are declining.

http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur

CPBell said:

Waterbirds are specialists by virtue of their dependence on water, and it may be this specific relationship, rather than specialism per se that has caused the observed patterns to emerge.

Click to expand...

Ok, you've lost me here. Why does "dependence on water" not count as a specialism? Why does their status as 'waterbirds' diminish the general message, which is that 'specialists' are losing ground and 'generalists' are gaining ground in the present situation? Is the precise mechanism by which it's happening really that important?

CPBell said:

For the CSI index to decline, all that is required is that the average population trend among generalists is more favourable than that among specialists. An average trend across species can be increasing when many of the species are declining.

Click to expand...

Not sure how I've misinterpreted you. It seemed like your criticism was that the authors cannot conclude that specialists are decreasing on average, because they didn't measure it directly in the paper. My point was that it's extremely unlikely that specialists are increasing on average, given what we know of bird population trends in the UK. What am I missing?

'roy said:

What am I missing?

Click to expand...

A massive scientific 'conspiracy theory' complex.

'roy said:

Ok, you've lost me here. Why does "dependence on water" not count as a specialism? Why does their status as 'waterbirds' diminish the general message, which is that 'specialists' are losing ground and 'generalists' are gaining ground in the present situation? Is the precise mechanism by which it's happening really that important?

Click to expand...

Dependence on water may be the only ‘specialism’ contributing to the pattern. Its relationship with their ‘specialisation index’ is so strong that the derived patterns in community specialisation would still emerge even if every other form of specialisation were completely uncorrelated with the relevant variables (time, temperature, rainfall).

Your reasoning goes like this:

(i) Waterbirds are undergoing relative decline
(ii) Waterbirds are specialists
(iii) Therefore specialists are undergoing relative decline

Compare this to:

(i) Waterbirds have webbed feet
(ii) Waterbirds are specialists
(iii) Therefore specialists have webbed feet

Comprende? (If not you’re on your own).

'roy said:

Not sure how I've misinterpreted you. It seemed like your criticism was that the authors cannot conclude that specialists are decreasing on average, because they didn't measure it directly in the paper. My point was that it's extremely unlikely that specialists are increasing on average, given what we know of bird population trends in the UK. What am I missing?

Click to expand...

No, your point was that for my critique to be valid everything has to be increasing. You are now saying that it’s unlikely that specialists are increasing on average ‘given what we know’. The argument therefore becomes, ‘generalist trends are more favourable than those among specialists, therefore specialists must be declining because we know specialists are declining’.

Specialists may be declining (though I see no reason to assume this), but Davey et al. have presented no additional evidence to this effect. They not only fail to present evidence in favour of their conclusion that generalists cause specialist decline, but also fail to a to establish that there is a problem to solve in the first place.

http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur

CPBell said:

Dependence on water may be the only ‘specialism’ contributing to the pattern. Its relationship with their ‘specialisation index’ is so strong that the derived patterns in community specialisation would still emerge even if every other form of specialisation were completely uncorrelated with the relevant variables (time, temperature, rainfall).

Click to expand...

Since when did a generalisation have to be universally true in order to be useful?

And what exactly is your alternative hypothesis to explain the pattern? Something causing a major range shift in waterbirds that is unrelated to climate and habitat specialism? Something sufficiently pronounced that it is the overwhelmingly dominant across the 5,000 odd BBS squares?

CPBell said:

Your reasoning goes like this:

(i) Waterbirds are undergoing relative decline
(ii) Waterbirds are specialists
(iii) Therefore specialists are undergoing relative decline

Compare this to:

(i) Waterbirds have webbed feet
(ii) Waterbirds are specialists
(iii) Therefore specialists have webbed feet

Comprende? (If not you’re on your own).

Click to expand...

There's a huge and extremely important difference between these scenarios. In the first case (specialists), there's a strong theoretical background underpinning the hypothesis that's being asserted. In the second case (webbing), there's absolutely no theory to support the hypothesis, it's just silly. Correlative studies that test a strong theory are scientifically worthwhile. This is why so many correlative studies get published (but the results must always be taken with a pinch of salt). Correlations that have no theoretical underpinning have no value. It's a big difference.

But anyway, I don't know why I'm bothering as I it's overwhelmingly clear you're not going to change your mind about anything. No more from me on this. I hope the authors themselves get back to you and provide some more meat for you to chew on.

Best of luck in your future scientific endeavours!

Actually, I can't resist making one more point:

CPBell said:

Specialists may be declining (though I see no reason to assume this), but Davey et al. have presented no additional evidence to this effect. They not only fail to present evidence in favour of their conclusion that generalists cause specialist decline, but also fail to a to establish that there is a problem to solve in the first place.

Click to expand...

They show a strong correlative pattern supporting the notion that generalists are succeeding where specialists are failing. But yes, it's true that they don't present clear direct evidence supporting their proposed mechanism. Terrible crime that.

It reminds me of a certain paper I saw a while back on sparrowhawks and house sparrow decline... As I recall, that one made quite a big deal out of a broad correlative pattern, although it provided absolutely no evidence in support of the mechanism that was proposed. No evidence that predation rates on house sparrows were driving population declines, or even that sparrowhawks were actually predating that many sparrows. Just a broad correlation.

Very lax these BTO ornithologists...

'roy said:

Since when did a generalisation have to be universally true in order to be useful?

And what exactly is your alternative hypothesis to explain the pattern? Something causing a major range shift in waterbirds that is unrelated to climate and habitat specialism? Something sufficiently pronounced that it is the overwhelmingly dominant across the 5,000 odd BBS squares?

Click to expand...

A generalisation doesn’t have to be universally true, but it does have to be general, and my argument here is that the evidence presented does not support this proposition. If you read the 5th paragraph in the letter quoted above, you will see that I assume that climate change is responsible for the observed patterns. The point I’m making is that the patterns tell us nothing about its effect on community specialisation.

'roy said:

There's a huge and extremely important difference between these scenarios. In the first case (specialists), there's a strong theoretical background underpinning the hypothesis that's being asserted. In the second case (webbing), there's absolutely no theory to support the hypothesis, it's just silly. Correlative studies that test a strong theory are scientifically worthwhile. This is why so many correlative studies get published (but the results must always be taken with a pinch of salt). Correlations that have no theoretical underpinning have no value. It's a big difference.

Click to expand...

Again you are missing the point being made here. The second example is intentionally ridiculous to bring out the fact that the form of the syllogism (which both examples share) is formally invalid. It matters not a jot whether or not the conclusion in the first example has a ‘strong theoretical background’, since we are trying to decide whether it is supported by the information contained in the two premises. Once more, your reasoning is question begging: i.e. the results say nothing about whether climate change affects community specialisation, but you conclude that is does because you think you know this already.

'roy said:

Actually, I can't resist making one more point:

They show a strong correlative pattern supporting the notion that generalists are succeeding where specialists are failing. But yes, it's true that they don't present clear direct evidence supporting their proposed mechanism. Terrible crime that.

It reminds me of a certain paper I saw a while back on sparrowhawks and house sparrow decline... As I recall, that one made quite a big deal out of a broad correlative pattern, although it provided absolutely no evidence in support of the mechanism that was proposed. No evidence that predation rates on house sparrows were driving population declines, or even that sparrowhawks were actually predating that many sparrows. Just a broad correlation.

Very lax these BTO ornithologists...

Click to expand...

Epic fail. Davey et al. do not present any strong correlations concerning generalists and specialists, since their approach doesn’t control for confounding variables. The approach I used in my House Sparrow decline study was specifically designed to address this problem, such that any hidden variable causing sparrow decline would need to have a spatio –temporal pattern identical to that of Sparrowhawk population recovery to achieve similar explanatory power.

This difference should be quite clear to anyone who has read both articles, which I presume you’ve done since you have such strong opinions?

http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur

You see that bit of flob on the floor? That's you that is.

Professor Dave - do you have any statistically significant data to support your hypothesis that there is correlation between said flob and any previous posters?

Just askin', like.

CPBell said:

Epic fail. Davey et al. do not present any strong correlations concerning generalists and specialists, since their approach doesn’t control for confounding variables. The approach I used in my House Sparrow decline study was specifically designed to address this problem, such that any hidden variable causing sparrow decline would need to have a spatio –temporal pattern identical to that of Sparrowhawk population recovery to achieve similar explanatory power.

Click to expand...

Oh no! I've failed epically. You've got me bang to rights there. Except, wait a minute, you've still not told us what this 'confounding variable' for Davey's study might actually be?

And in your sparrowhawk study, you never actually report a goodness-of-fit statistic for your sparrowhawk model - all you do is compare it to the performance of a chronological trend model. All that tells us is that the sparrowhawk model provides a better explanation than an simple geographically universal trend. It doesn't tell us how close the relationship actually is. Any number of other models could fit better, but you don't test any other models. So no, you certainly can't claim that you controlled for confounding variables in your study.

Look, I quite like your sparrowhawk paper, and I think the pattern you found has some merit. But you've got to acknowledge that it's just a correlative study (and a univariate one at that), and it can't be used to prove anything. Just like Davey et al's. They provide evidence to support a hypothesis, not prove it.

Your attempts to aggrandize your own work and diminish that of others are, well, childish is the best word I think.

No more from me, for reals this time. I'll let you have the last word, as you always do...

CPBell said:

For the record, I’ve received replies to the mail posted above from David Currie and Martin Sykes, the editors of Global Ecology and Biogeography, providing assurances about the peer-review process for the article. However I’ve received no response or acknowledgement from the author. If this is still the case by the middle of next week, I’ll call Lund University to find out whether there’s any intention to respond, and post the outcome here.

Click to expand...

I managed to speak to Catherine Davey yesterday, and I’ve uploaded the conversation to my Youtube channel. I repeated the two main criticisms set out in the email sent last week which were (i) The specialisation index used is not independent of taxon and broader ecology, and is heavily weighted towards waterbirds at higher values, so that derived trends in community specialisation may reflect differential impact of climate change on waterbirds v terrestrial birds (ii) Even if the specialisation index were valid, it would not be possible to conclude that generalists are affecting specialists, not least because no data are presented to show that specialists are declining.

Ms Davey defends these criticisms as follows: (i) Numbers of waterbirds counted in the survey are small compared to terrestrial birds, so are unlikely to have a significant effect on the community specialisation index (ii) Work done by the BTO elsewhere indicates that specialists are declining.

Taking the second response first, I would say that this throws into sharp relief a difficulty with this kind of work. The article presents no new evidence in favour of its major conclusion, relying on work done elsewhere. However the article has been widely publicised as providing new and additional evidence for a process of wide conservation significance, and no doubt will be widely cited as doing so. In other words, we will see further articles in the future that rely on this paper as having provided solid evidence. There is an obvious analogy with ‘toxic’ financial derivatives – the BTO as the Lehman Brothers of ornithology.

A quick check of the numbers presented in appendix S1 confirms that waterbirds are indeed much less common than terrestrial birds in the BBS dataset used. Mean counts are 5400 and 58000, while median counts are 675 and 5098 for waterbirds and terrestrial birds respectively, so it seems safe to assume that terrestrial birds are ten times as frequent as waterbirds. Let’s examine the implications of this for the Community Specialisation Index.

Take a simple community consisting of two species, a thrush with a specialisation index of 0.5, indicating generalisation, and a duck with an index of 2.0, indicating specialisation. If the community consists of one duck and ten thrushes we calculate community specialisation as follows:

CSI =((0.5x10)+(2.0x1))/(10+1) = 0.636

Let’s say climate change favours the thrush without affecting the duck, so that the number of thrushes increases to 11. CSI will decline to 0.625, so despite the overwhelming prevalence of terrestrial birds, the differential effect of climate change is still reflected in CSI. One might argue that the change is small, 0.01 for a 10% increase in terrestrial birds, but the change demonstrated in the article is also small – a decline from ca. 0.99 to ca. 0.97 over a twelve year period. On this basis I would say that the argument from rarity of waterbirds fails to stand up.

http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur

CPBell said:

There is an obvious analogy with ‘toxic’ financial derivatives – the BTO as the Lehman Brothers of ornithology.

Click to expand...

get a grip.

abagguley said:

Professor Dave - do you have any statistically significant data to support your hypothesis that there is correlation between said flob and any previous posters?

Just askin', like.

Click to expand...

Professor Adrian, the mistake is all mine. Thought 'twas the best YouTube videos thread.

'roy said:

No more from me, for reals this time. I'll let you have the last word, as you always do...

Click to expand...

Don't duck out yet James, I think you had him on the ropes at one point. :t:

As an alternative to assessing the change over time in the balance of specialists/generalists, it would be interesting to compare the change over time in homogeneity between BBS squares (i.e. how similar are the bird communities in different squares). Has such an analysis been published?