Cheers for that,
out of interest do the sparrows demonstrate shy behavior, do they come out in the open much or stick to cover, what are their food preferences particularely which seed and do you know where they are nesting.
-
Welcome to BirdForum, the internet's largest birding community with thousands of members from all over the world. The forums are dedicated to wild birds, birding, binoculars and equipment and all that goes with it.
Please register for an account to take part in the discussions in the forum, post your pictures in the gallery and more.
Sparrowhawks responsible for House Sparrow decline says scientist (1 Viewer)
- Thread starter John B (not the sloop)
- Start date
Cheers for that,
out of interest do the sparrows demonstrate shy behavior, do they come out in the open much or stick to cover, what are their food preferences particularely which seed and do you know where they are nesting.
occasional
Well-known member
Is there really any clear conflict between your (CPBs) conclusions, and the miscellaneous scientific studies and the various opinions being expressed here ?
It appears to me likely that the principal effect of a "new" predator on an abundant prey species may occur simply because of changes in the behaviour of the prey. (As evidence for this I would cite the appearance of mink in the area where I live. The appearance of mink resulted in declines of 95 to 100 % in the local population of a fair number of species, and the mink could clearly not have eaten that many large birds.)
Thus sparrows may need more cover, or more food close to cover, or more nesting places close to food if sparrowhawks appear.
Perhaps CPBs conclusion should be that the sparrowhawk is the essential factor in the decline. All the other factors will have their effect if you attempt to measure them, but that does not invalidate the conclusions in the original paper.
It appears to me likely that the principal effect of a "new" predator on an abundant prey species may occur simply because of changes in the behaviour of the prey. (As evidence for this I would cite the appearance of mink in the area where I live. The appearance of mink resulted in declines of 95 to 100 % in the local population of a fair number of species, and the mink could clearly not have eaten that many large birds.)
Thus sparrows may need more cover, or more food close to cover, or more nesting places close to food if sparrowhawks appear.
Perhaps CPBs conclusion should be that the sparrowhawk is the essential factor in the decline. All the other factors will have their effect if you attempt to measure them, but that does not invalidate the conclusions in the original paper.
You raise a number of interesting questions. Your point about sparrowhawk being the ‘essential’ factor in the decline is an important one. I don’t dispute that other factors affect the population size of wild animals all the time. All species populations fluctuate, and they do so for reasons relating to their environment. For the most part, however, these fluctuations don’t end up causing a media flap, and I insist that without the effect of sparrowhawk predation, there would have been no ‘great sparrow mystery’ to explain.
The ‘change in behaviour’ argument is the most credible of the critiques put forward by the RSPB/BTO, i.e. not credible at all. Their argument is that the results we attribute to predation related mortality are really just a result of sparrows starting to shun feeding stations after sparrowhawk attacks begin. However, either GBFS data reflect population size or they don’t. If they don’t, the avoidance hypothesis might be true, but this implies no population decline since sparrowhawk incidence explains the decline in all its particulars.
The ‘other factors’ argument is another of the critics’ favourites. They argue that a proper test of the sparrowhawk hypothesis requires that it remains significant when a wide range of other factors are controlled for by including them in any statistical model. However it’s all too easy to tweak this approach to ensure that any given variable will emerge as non-significant, and the recent BTO paper on predation is exemplary in this regard.
The BTO in particular loves multi-factor explanations, because they always need ‘more research’ to tease out all of their implications. This is why I think they’ve jumped the shark in terms of their focus on business development at the expense of scientific rigour. They should really spin off a profit-making environmental consultancy and leave the core monitoring function as a charitable enterprise providing open source data on request.
occasional
Well-known member
The ‘change in behaviour’ argument is the most credible of the critiques put forward by the RSPB/BTO, i.e. not credible at all. Their argument is that the results we attribute to predation related mortality are really just a result of sparrows starting to shun feeding stations after sparrowhawk attacks begin. However, either GBFS data reflect population size or they don’t. If they don’t, the avoidance hypothesis might be true, but this implies no population decline since sparrowhawk incidence explains the decline in all its particulars.
I am not sure that I understand this paragraph.
If the presence of sparrowhawks means that sparrows have to be significantly more careful than previously while searching for insect food for their young, then the appearance of sprawks may well result in increased nest mortality through food shortage as Kate Vincent found, and reduced population size as you suggest.
I am not sure that I understand this paragraph.
If the presence of sparrowhawks means that sparrows have to be significantly more careful than previously while searching for insect food for their young, then the appearance of sprawks may well result in increased nest mortality through food shortage as Kate Vincent found, and reduced population size as you suggest.
This whole argument has become a b bit circular until adequate evidence can be uncovered. Until such time it appears that CPbells results stand firm. Kate vincents research was a very comprehensive and in depth study, and I cannot help but admiire the effoort that went into it. However I must admit that CPbells assertion that it has only proven an already known fact (that food availability will always limit breeding success) is quite a strong point. However he only gets away with this argument because no one has adequately proven that invertibrates have declined particularly in urban areas over the relevant time period, and he genraly does not accept that areas have become 'Less Greener', but again this has not been adequately quantified and correlated.
I find it perfectly feasable that large profit making organisations like RSPB, BTO can manipulate findings to gain more funding. After all alot of people are unwittingly manipulated by buisnesses and goverenment every day, me included.
I find it perfectly feasable that large profit making organisations like RSPB, BTO can manipulate findings to gain more funding. After all alot of people are unwittingly manipulated by buisnesses and goverenment every day, me included.
Predation will be given as much weight as habitat and food when someone can show a proper correlation based on accurate data.
Having now looked into your methodology in more depth...
Firstly, GBFS data:
This survey relies on peak counts over a long period. You might recall a headline last year re Blue Tits declining by 42% in gardens over 40 yrs? This result was from the GBFS http://www.bto.org/news-events/press-releases/feeling-blue-garden-acrobat-takes-tumble
Yet every other survey has shown that Blue Tits have increased strongly over the same period: http://www.bto.org/birdtrends2004/wcrbluti.htm
This is important, as it shows that GBFS can throw up potentially spurious results (What you should have done was confirm the trends using CBC/BBS data before progressing with the model).
Blue Tits are likely to have declined in GBFS due to changes in birdfeeders and food, meaning less birds are visible at one time but NOT that less birds are present in the area overall. So this means that GBFS cannot be relied on to reflect local populations - it can show the opposite population trend to what is actually happening!
Second, you compared urban and rural gardens over time, but your treatment of change (e.g. when rural gardens are built on and become urban) was completely unsuitable - you merely looked at a modern map to see if the garden was urban or rural TODAY, not in 1970. You state that you just lowered the urban threshold in order to account for this. There is a massive problem here, as your study period (1970 onwards) was a period of massive urbanisation (e.g. vast council estate creation, new towns), so it is probable that some of your urban gardens started off as totally rural. Your fudge of lowering the threshold for 'urban' overall does not work, as it still takes no account of this change. You say data was not available to check this - it is. There is ample OS data documenting urbanisation, but you seem not to know about/bothered with it. This is a critical flaw, due to sample sizes (see below).
So that's two big data quality issues. But there is another.
Third, your sample sizes in each Sparrowhawk recovery zone, for rural and urban gardens, are small and heavily unbalanced. E.g. In zone 4 you have about 27 vs 8, in zone 3 about 25 vs 5. Anything less than 30 is considered 'with a pinch of salt', yet most of your samples are below this standard threshold, and you are even unable to say if any of them changed classification over time, which would whittle your samples down even more.
So the data you used:
1. can generate spurious results (that disagree with all other surveys).
2. was not controlled for changes in birdfeeding methods/gardens that may affect peak count (cf Blue Tits).
3. likely contained errors of urban/rural classification, and did not adequately control for landscape change.
4. had small and unbalanced sample sizes.
As such, your patterns of sparrow decline are highly questionable, especially as you attempted to use a complex spatial pattern. In other words, you may have shoe-horned inadequate data into your model, and failed to check if it was accurate.
As such, a model that just generates a correlation is of very limited use, as you cannot be certain that you are correlating 'real' patterns, rather than patterns in the limited and questionable data.
Occasional/Spencer
The point I’m trying to make in the paragraph quoted by occasional is that all of the decline seen in GBFS sparrow data is explained by sparrowhawk presence, and that there is no decline in their absence. If, therefore, RSPB/BTO are right, and the sparrowhawk related decline reflects avoidance, and not population decline, there must have been no decline at all, which is clearly nonsense.
As for the cause of the sparrowhawk-related decline, it’s unlikely that avoidance-related decrease in breeding success is signficant, since breeding success has increased over the period of the population decline. The most parsimonious (i.e. best) explanation is simply sparrowhawks eating sparrows.
Having said that, predator avoidance behaviour has been shown to affect breeding success when the predators are humans. This is probably the source of the correlation between NO2 and breeding success found in the Vincent study, since NO2 is highest along roads, where there is most disturbance of foraging sparrows. This is the basis of the muddled argument about air pollution and insect decline in cities, which is so touted by the RSPB, who quickly forget about the danger of spurious correlation when evidence suits their preconceptions.
Amarillo
Thanks for the detailed response - this is a cross-post. I'll respond when I've had time to absorb your comments.
The point I’m trying to make in the paragraph quoted by occasional is that all of the decline seen in GBFS sparrow data is explained by sparrowhawk presence, and that there is no decline in their absence. If, therefore, RSPB/BTO are right, and the sparrowhawk related decline reflects avoidance, and not population decline, there must have been no decline at all, which is clearly nonsense.
As for the cause of the sparrowhawk-related decline, it’s unlikely that avoidance-related decrease in breeding success is signficant, since breeding success has increased over the period of the population decline. The most parsimonious (i.e. best) explanation is simply sparrowhawks eating sparrows.
Having said that, predator avoidance behaviour has been shown to affect breeding success when the predators are humans. This is probably the source of the correlation between NO2 and breeding success found in the Vincent study, since NO2 is highest along roads, where there is most disturbance of foraging sparrows. This is the basis of the muddled argument about air pollution and insect decline in cities, which is so touted by the RSPB, who quickly forget about the danger of spurious correlation when evidence suits their preconceptions.
Amarillo
Thanks for the detailed response - this is a cross-post. I'll respond when I've had time to absorb your comments.
Amarillo
Thanks again for your critique. It’s a skilful combination of plausibility and rhetorical flourish, but I still find myself unmoved by it.
Ah, but we did. See the last paragraph of the introduction. With hindsight we should have laid this on with a trowel rather than giving it a brief passing mention, as a number of critics elaborated rococo arguments on the assumption sparrow GBFS data do not index population.
We didn’t ‘lower the urban threshold’, but merely used a broad 2-level variable to denote level of urbanisation. We acknowledge that there may be some misclassification at the margins because of development during the data period, but you are exaggerating the potential of this to confound our results.
The period since 1970 has not been a time of ‘massive urbanization’. House building peaked in the late 1960s, and then underwent a precipitous decline. New Town development mainly took place in the post-war period, and the latest of these, such as Milton Keynes and Telford, coincided with the late 1960s peak. Since then, such development that has occurred has been increasingly undertaken by urban infilling, which would not affect our index.
The issue of sample size is not as simple as you make out, but is actually a whole field of study in itself relating to statistical ‘power’. Simply put, this examines the probability that a null hypothesis that is ‘really’ false, will be rejected by a statistical test. Small sample sizes increase the probability that this will not happen, which is known as a ‘type II error’, though this is affected by many things, including sample variance and ‘effect size’ (i.e. difference between sample means). Despite the fact that our sample sizes were not huge, we found widespread and highly significant statistical variation among the contingencies of zone/rurality for both species, which in the context of thorough model-checking means we can be confident that there are ‘real’ differences here. Moreover, our conclusions place little reliance on the contingencies with smaller sample sizes: In fact these tend to be the ones that would confound our arguments were they to represent reliable outcomes, as in the case of the trends in urban zone 4, which were impossible to reproduce using the simulation model (contrary to the argument that the latter could have reproduced 'any' set of sparrow trends).
The analyses presented in the paper were checked by two independent statisticians in the course of peer review, and I was required to respond to literally pages and pages of critique produced by them. Believe me, if there had been a problem with the statistics, we would not have got anywhere near publication.
http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur
Thanks again for your critique. It’s a skilful combination of plausibility and rhetorical flourish, but I still find myself unmoved by it.
Ah, but we did. See the last paragraph of the introduction. With hindsight we should have laid this on with a trowel rather than giving it a brief passing mention, as a number of critics elaborated rococo arguments on the assumption sparrow GBFS data do not index population.
We didn’t ‘lower the urban threshold’, but merely used a broad 2-level variable to denote level of urbanisation. We acknowledge that there may be some misclassification at the margins because of development during the data period, but you are exaggerating the potential of this to confound our results.
The period since 1970 has not been a time of ‘massive urbanization’. House building peaked in the late 1960s, and then underwent a precipitous decline. New Town development mainly took place in the post-war period, and the latest of these, such as Milton Keynes and Telford, coincided with the late 1960s peak. Since then, such development that has occurred has been increasingly undertaken by urban infilling, which would not affect our index.
The issue of sample size is not as simple as you make out, but is actually a whole field of study in itself relating to statistical ‘power’. Simply put, this examines the probability that a null hypothesis that is ‘really’ false, will be rejected by a statistical test. Small sample sizes increase the probability that this will not happen, which is known as a ‘type II error’, though this is affected by many things, including sample variance and ‘effect size’ (i.e. difference between sample means). Despite the fact that our sample sizes were not huge, we found widespread and highly significant statistical variation among the contingencies of zone/rurality for both species, which in the context of thorough model-checking means we can be confident that there are ‘real’ differences here. Moreover, our conclusions place little reliance on the contingencies with smaller sample sizes: In fact these tend to be the ones that would confound our arguments were they to represent reliable outcomes, as in the case of the trends in urban zone 4, which were impossible to reproduce using the simulation model (contrary to the argument that the latter could have reproduced 'any' set of sparrow trends).
The analyses presented in the paper were checked by two independent statisticians in the course of peer review, and I was required to respond to literally pages and pages of critique produced by them. Believe me, if there had been a problem with the statistics, we would not have got anywhere near publication.
http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur
occasional
Well-known member
As for the cause of the sparrowhawk-related decline, it’s unlikely that avoidance-related decrease in breeding success is signficant, since breeding success has increased over the period of the population decline. The most parsimonious (i.e. best) explanation is simply sparrowhawks eating sparrows.
CPB,
I am not sure which particular statistics you are considering, but your dismissal of this particular hypothesis is surely inadequate.
The hypothesis might lead one to suppose that, at any particular place, breeding success would decline as sparrowhawks appeared and then increase as sparrows were confined to safer breeding sites. In particular one would suppose the KV achieved her results because the study was at the appropriate place and time.
Determining how this fluctuation should translate into more general statistics over the whole country would require some significant analysis, so the fact that breeding success has increased does not conflict with the hypothesis.
Given that the "sprawks kill sparrows" theory is in some ways simply a speculative add-on to your basic paper, why not be yet more parsimonious and simply leave open the mechanism by which sprawks influence sparrow populations.
You demonstrate that it is conceivable that sufficient sparrows could be killed to reduce the population, but this appears to arouse scepticism among those who most closely observe both predator and prey.
Would your paper not be more compatible with other peoples observations if this particular part of the thesis was simply left unresolved ?
CPB,
I am not sure which particular statistics you are considering, but your dismissal of this particular hypothesis is surely inadequate.
The hypothesis might lead one to suppose that, at any particular place, breeding success would decline as sparrowhawks appeared and then increase as sparrows were confined to safer breeding sites. In particular one would suppose the KV achieved her results because the study was at the appropriate place and time.
Determining how this fluctuation should translate into more general statistics over the whole country would require some significant analysis, so the fact that breeding success has increased does not conflict with the hypothesis.
Given that the "sprawks kill sparrows" theory is in some ways simply a speculative add-on to your basic paper, why not be yet more parsimonious and simply leave open the mechanism by which sprawks influence sparrow populations.
You demonstrate that it is conceivable that sufficient sparrows could be killed to reduce the population, but this appears to arouse scepticism among those who most closely observe both predator and prey.
Would your paper not be more compatible with other peoples observations if this particular part of the thesis was simply left unresolved ?
pianoman
duck and diver, bobolink and weaver
Another late reply, sorry. This is purely anecdotal, but I do think sparrows are now spending less time on the ground than I remember as a child, when the population overall was higher. Curiously, and this really may be my imagination, their vocalisations seem to have changed to my ears - more varied chirrups and churrs and less of the monotonous chirp - chirp - chirp which passes for a song in HOSP.
They do use my feeder regularly - but only the type which offers an easy landing platform with nearby cover. It has general mixed seed
sorceresselite
Well-known member
We used to have an abundance of house sparrows in the garden, from 2005 to 2008 we never saw any. Then in 2009 we had 2 families and 2010 we had 4 families regulary in the garden and on the feeders. Hopefully 2011 will see an increase, we have a Sparrow hawk in the area, but i do not think you can blame it for the decline. It is more likely to be people removing hedges and replacing them with walls, or putting plastic facias on there houses stopping the birds nesting. There fav food on our feeders are fatty balls.
I’m not sure what the relevance of the Vincent study is to the issue, since it only found poorer nestling condition where aphid density was lower in the nest vicinity, and did not consider Sparrowhawk presence.
The statistics I quote are from the BTO’s DEFRA funded study on sparrows and starlings published in 2002. Both the BTO and DEFRA seem to have taken the report down from their website, which judging from chatter on the bush telegraph might be a preliminary to them raising the white flag on the topic. However we’ll have to wait and see.
Your hypothesis seems to be that sparrows respond with such alacrity to the appearance of sparrowhawks that negligible mortality occurs, but their local ranges are then restricted, resulting in density dependent reduction in breeding success, which then causes a population decline which relieves the density dependent breeding depression. This is not impossible, but then you would have to explain why there has been a decline in adult and first year survival over the period of the decline (also in the withdrawn BTO report). In short we have a sparrowhawk-related population decline in tandem with a decline in adult/first year survival, and no evidence of decline in breeding success temporary or otherwise. This gives us a choice of a simple explanation that fits the evidence (direct predation on adults and first years), or a rather baroque elaboration for which there is no evidence, but for which evidence might conceivably be obtained if we do enough sifting of the data. I know which I prefer, but you have every right to disagree.
Even if we go with your hypothesis, the ultimate explanation is still sparrowhawk predation (though potential rather than actual) which is not compatible with the RSPB/BTO approved orthodoxy of an objective decline in food availability caused by farming methods and urban renewal or pollution or gardening fads depending on who they’re trying to get money out of on any particular day of the week.
Finally, the direct predation hypothesis is not a “speculative add-on”. It’s the hypothesis we started with and from which we derived our critical predictions, and is also backed by direct observation, such as that from the Amsterdam city centre nest discussed way back in the thread, and studies of relative predation risk.
http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur
Vocal support for the voices such as Songbird Survival, a constant stabbing at conservation bodies, every other post jabbing with little one-liners such as "depending on who they’re trying to get money out of on any particular day of the week" and etc, etc, Couple this with obvious flaws in your work and an absolute refusal to consider any of the persons offering alternatives to your work, you come across not as a scientist but so much more as someone with an agenda, someone with an axe to grind.
Hmm, what could it be? You wouldn't be the same CP Bell that described himself as a 'frustrated academic who had worked long hours with little recognition' just after he got sacked from a prestigious conservation organisation for sending rude and abusive e-mails and showing disrespectful treatment of other members of staff?
Considering every other of your posts on this thread complain the RSPB, etc are only out to grab money, is it not rich that you then tried to screw the The Zoological Society of London for £25,000 at an employment tribunal?
Seems you really don't like anybody in the field of conservation - your view on the professional opinion of your employer's senior manager is revealing “Any old rubbish as long as it gets in the papers is a good idea.” Hey, that's pretty much the same as you're saying about the RSPB and BTO, etc.
As your threads, page after page, are a string of attacks on leading conservation bodies in the UK, trying to undermine their integrity, I think it is only balanced to give an indication of where you come from, hardly an unbiased person.
Link here to the relevant article.
.
Hmm, what could it be? You wouldn't be the same CP Bell that described himself as a 'frustrated academic who had worked long hours with little recognition' just after he got sacked from a prestigious conservation organisation for sending rude and abusive e-mails and showing disrespectful treatment of other members of staff?
Considering every other of your posts on this thread complain the RSPB, etc are only out to grab money, is it not rich that you then tried to screw the The Zoological Society of London for £25,000 at an employment tribunal?
Seems you really don't like anybody in the field of conservation - your view on the professional opinion of your employer's senior manager is revealing “Any old rubbish as long as it gets in the papers is a good idea.” Hey, that's pretty much the same as you're saying about the RSPB and BTO, etc.
As your threads, page after page, are a string of attacks on leading conservation bodies in the UK, trying to undermine their integrity, I think it is only balanced to give an indication of where you come from, hardly an unbiased person.
Link here to the relevant article.
.
Last edited:
edenwatcher
Well-known member
Well done Jos for digging this out. Things become much clearer.
Rob
Rob
Jane Turner
Well-known member
It doesn't say what the outcome was.
You compared large-scale garden trends against large-scale national trends. But your Auk paper models patterns at 264 individual sites in 4 zones. So unless you validated the pattern at each site or zone, then your approach was not correct. You reduced the spatial resolution of your comparison down to 1 zone, before extrapolating it up to 4 zones for your analysis, so you assumed that the comparison was the same in all zones, which is a big assumption due to spatial complexity of your analyses. So you still have no good evidence that the patterns you are modelling in the 4 zones reflect the sparrow population in those 4 zones, because the GBFS data has very large holes in it, as can be seen from your Fig 1 (large gaps in data from Scotland, N England, S Wales, Lincs, Zone 1).
You "reduced the categorization to a two-factor threshold" from a 16 factor categorization. But this still means that, in additon to having a question mark over the spatial aspect, your modelling of a spatio-temporal pattern didn't take any account of temporal change at all in one of your categorical fixed effects. It is not an exaggeration to suggest that this might be a problem if the classification of your gardens might have changed over 40 years, as this is not only likely, but also has severe implications for the small sample sizes you used.
The first phase of Milton Keynes was 1971 and it doubled in size during the 1970s (census data). The largest council estate in Europe at the time (in Kingston-upon-Hull) was built from 1968-1984. The number of dwellings in Britain increased by more than 10% in the 1970s and 13% in the 1980s, and increased by 30% between 1972 and 2002. Approximately 3 million dwellings were built during he 1970s alone:
http://www.statistics.gov.uk/STATBAS....asp?vlnk=7315
http://www.parliament.uk/documents/c...9/rp99-111.pdf
Can you provide any evidence to support this claim? "Such development that occurred" amounted to a further 3 million dwellings since the 1980s. These figures do not even include industrial buildings and out of town shopping centres, roads etc. So the potential for landscape change at your 264 sites seems rather large, yet you never attempted to control for this at all, despite rurality being a key element of your model.
But your representation of 'zone' and 'rurality' are not safe. What you are actually saying in your paper is that 8 gardens (zone 4) can represent all urban gardens in the whole of Lincs, (most of) East Anglia, Herts, Beds, Bucks, much of Kent and the East Riding, and none changed in 40 years. And that about 6 urban gardens are representative of all urban areas in Wales, much of Scotland and the SW of England, and none of those changed in 40 years either. That is not even one garden per county, yet you are suggesting that this will give enough statistical power to be confident that the pattern you are modelling is accurate. That seems very optimistic. Your statistical significance was with the small dataset you had, not necessarily the real world, which may explain your result (with more data comes more potential for complexity). In samples that small, one data point can have a very large effect, and if just two of your zone 4 urban gardens had changed, then your data for that zone would be 25% wrong. But you don't know but you never checked.
If you received two reviews (as is normal with Auk), then it was not checked by two statisticians. And unless it was checked by someone with experience in spatial analysis and issues of autocorrelation in complex spatial patterns (as yours is), then you have been lucky in the review process. The model-building and its execution are fine, the problem is the data quality, pattern of data spread, classification of rural/urban, and use of complex zones. In short, the data appears to be of insufficient spatial resolution for the landscape pattern to which it is being applied.
earleybird
Well-known member
This seems to make far more sense to me than House sparrow decline as a direct consequence of predation by Sprawks.
We have an abundance of starlings feeding in our garden all year round with a top count of 32 at one time. Up until 2 years ago they nested in our roof space and that of my neighbours. Since we both got large wasp nests 3x years ago no starlings have nested since. I have no idea if there is any relationship between the two factors but it doesn't seem to have adversely affected starling numbers in my garden.
Neither does the presence of a sprawk nest since 4x years ago in my neighbours garden . I have seen sprawks take 3x birds in my garden ,one a thrush and 2x blackbirds although i'm sure they have taken many more. I have to say that most small species are a lot more 'nervous' and feed more cautiously than before. Bluetits, Greatits and coaltits will all snatch seed or suet and immediately fly back to the cover of an adjacent tree . Whereas sparrows, dunnocks and chaffinch prefer to feed on the ground away from cover ?
The removal of a large hedge/mixed foliage against one of my neighbour's walls has seen an immediate and dramatic drop in numbers of Housesparrows recorded since its removal. Numbers dropping from 30-50+ down to 10-15 !
I have since found another large thick, dense, bush/hedge 3x gardens away which sparrows now use to both congregate prior to and during feeding and also roost .
Last edited:
Jane Turner
Well-known member
There is also the 80s to now trend for slash and burn (flags and decking) garden in existing "urban housing" that was previously good Sparrow feeding territory.
Your tactic here is to say that whatever validation is carried out is insufficient because it should have been even more thorough and done at an even finer scale, even if this implies a scale of comparison that is meaningless, such as on a site by site basis. How would you do this? Compare each GBFS site with the nearest CBC site? Then what? Throw out the ones that aren’t correlated because they don’t index the population? How do you know when CBC and GBFS indices are from different sites?
You’ve got this back to front. We don’t assume the GBFS sites are representative of the zones they are from. The zones are used to classify the sites in a way that expresses the progress of sparrowhawk recolonisation.
Even if you were correct here the problem you are describing is conservative of the null hypothesis, since mis-classification of urban and rural sites will make it less likely that they will differ significantly. The fact remains that we did find significant differences between urban and rural sites, and these match the qualitative trends that can be detected elsewhere. Your attempt to explain these away as artefacts of mis-classification is strained to say the least.
The building of which would have been heavily skewed towards the beginning of the decade, so that the situation converged rapidly with that prevailing post-2000.
I think the burden of proof here is on you since this is a well-known trend. I can personally vouch for the fact that hardly any greenfield development has happened in areas familiar to me from the 1970s.
You keep hammering away at urban zone 4, but as I’ve already pointed out, and as we make clear in the paper, this result completely contradicts our hypothesis, but this is because the small sample renders the result fairly meaningless.
The paper had 4 reviewers, in addition to the managing editor and the journal editor. Two of the four independent reviews were stats focused.
Ok, I admit it. I did the whole thing to get revenge on the conservation industry because I once lost a job somewhere. Oh and by the way, my next research project will be on the migratory behaviour of trolls, to find out why flocks of them seem to mysteriously appear, then disappear again for no apparent reason.
http://www.cpbell.co.uk
http://www.youtube.com/CultoftheAmateur
Similar threads
