-
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.
Norfolk birding (143 Viewers)
- Thread starter Edward woodwood
- Start date
willowgrouse
Well-known member
Bird recording
HD - Thanks for your thoughts, as you are clearly a poster of sensible stuff on this forum I have considered them carefully...
Oops! I fear this does not bode well for its adoption by NBMR
This is of course the truth and could be an area where BT can produce some useful data. I admire your data handling skills (thank you for taking the time to do this), but my 'tongue in cheek' reference to avocet recording was designed to make a point, one that I feel is still relevant, that maps with dots mainly reflect observer coverage. I feel I need to point out that your distribution map for Blackbird clearly shows exactly this: It reflects the places where birders go birding, or actually live. All the large conurbations are clearly defined, as is the BTO BT recording team location. Interestingly there are a number of large gaps where well-known shooting estates are situated. I could believe this if it were a raptor map! So, are we to assume that game-keepers are shooting blackbirds as well? Of course not. Its just that 'us plebs' are not allowed in so there are no records.
In addition. your monthly distribution map appears to show observer activity more clearly than anything useful. Are blackbirds really at a low ebb in mid- to late-summer? Of course not, they are theoretically at their peak in the post-breeding period. This is not reflected in the data because birders are not out birding (hopefully they are out recording butterflies for you).
Whilst I can see the benefits for you of a distribution map for recording nomadic 'whites' year-on-year, I can't see its use as a bird recording tool - other than over an extensive period - that warrants its publication annually at a time when funds for producing conservation data are at an all-time low.
Wholeheatedly agree! This is, I suspect, a major point of concern for the county bird recording team.
Once again HD, thanks for your constructive and considered input. With your kind offer of assistance I sense that some communication across recording means and methods may be beneficial to all.
HD - Thanks for your thoughts, as you are clearly a poster of sensible stuff on this forum I have considered them carefully...
Oops! I fear this does not bode well for its adoption by NBMR
This is of course the truth and could be an area where BT can produce some useful data. I admire your data handling skills (thank you for taking the time to do this), but my 'tongue in cheek' reference to avocet recording was designed to make a point, one that I feel is still relevant, that maps with dots mainly reflect observer coverage. I feel I need to point out that your distribution map for Blackbird clearly shows exactly this: It reflects the places where birders go birding, or actually live. All the large conurbations are clearly defined, as is the BTO BT recording team location. Interestingly there are a number of large gaps where well-known shooting estates are situated. I could believe this if it were a raptor map! So, are we to assume that game-keepers are shooting blackbirds as well? Of course not. Its just that 'us plebs' are not allowed in so there are no records.
In addition. your monthly distribution map appears to show observer activity more clearly than anything useful. Are blackbirds really at a low ebb in mid- to late-summer? Of course not, they are theoretically at their peak in the post-breeding period. This is not reflected in the data because birders are not out birding (hopefully they are out recording butterflies for you).
Whilst I can see the benefits for you of a distribution map for recording nomadic 'whites' year-on-year, I can't see its use as a bird recording tool - other than over an extensive period - that warrants its publication annually at a time when funds for producing conservation data are at an all-time low.
Wholeheatedly agree! This is, I suspect, a major point of concern for the county bird recording team.
Once again HD, thanks for your constructive and considered input. With your kind offer of assistance I sense that some communication across recording means and methods may be beneficial to all.
crapbirder
Well-known member
Norfolk Records.
Not sure if I can meet Willowgrouse's criteria for sensible posting on this forum but here goes.
Why should I contribute any records to an organisation which to the outsider, appears prone to internal strife and disunited over the usefulness of BirdTrack.
Not sure if I can meet Willowgrouse's criteria for sensible posting on this forum but here goes.
Why should I contribute any records to an organisation which to the outsider, appears prone to internal strife and disunited over the usefulness of BirdTrack.
Nick Moran
Well-known member
Good point, Ryan. My big worry is the "that I deemed useful for the report" element. If the NBMR database is the database for bird records in Norfolk, it has been starved (subjectively, with the best will in the world) of 96% of your records for 2015. Even if only a small proportion of those are 'useful' in the future, there's no telling how they might be used for good.
Last edited:
Nick Moran
Well-known member
There are many people who gladly give up many hours of their own time for the betterment of bird recording in Norfolk, and we all owe them a huge debt of gratitude.
I'm very lucky to have a paid role in the field of bird recording, and feel immensely fortunate to be in that position (most of the time!). Promoting BirdTrack is clearly an important part of my job. Regardless of that, I'm a very active local birder, with Norfolk records going back to my childhood visits in the 1980s. One selfish reason bird recording in Norfolk matters greatly to me – apart from the fact I've lived in or near it for half my adult life – is that I've invested many hours of my own time putting all my Norfolk data (>87,500 records to date) into BirdTrack. It'll be much more than a great shame if that information does not fully contribute to research and conservation at a local level, as well as at national and international ones (which I'm already satisfied it does).
The idea of some form of standardised metric is that you can immediately make meaningful between-year comparisons. Apologies: I should have included the following year's account for the same species to highlight this.
Reading the NBMR 2014 account for Blackbird is little different from reading the equivalent for Ring Ouzel. They both list notable counts at coastal sites, in the ‘migration periods’. The main difference is that the Blackbird account includes the total number of territories at one site. I find it difficult to see what this might tell us about Blackbird in the long term. I strongly believe that one or two simple, consistent metrics would enhance the accounts of common species and prevent them from reading like the accounts of scarce and rare species (which in my opinion, is not the best approach for common species, migratory or otherwise).
HarassedDad answers this far more eloquently than I can. The beauty of both the volume of BirdTrack data for Norfolk and the concept of complete lists is that these (and many other) observer-effort related biases are or can be accounted for. Unfortunately that isn't the case if you're restricted to 94 ‘casual’ Blackbird records in a given year; at that end of the spectrum, it's hard to say anything meaningful or scientifically robust because there are so few data to play with.
Bird distributions can and do change regardless of (apparent) habitat availability. Saying that it is not much use listing how many squares a species was recorded in seems at odds with the approach of the Norfolk Bird Atlases and national Atlases over the years, all of which use similar metrics as one of the ways to document distribution changes.
There's mounting evidence that with enough unstructured* data, signals about population changes may be detectable – and potentially very rapidly, too. That is not to say that there'll be a time when we don't need the kind of data provided through the likes of the Breeding Bird Survey and Nest Record Scheme – on the contrary! – but that there's room for and value in both.
*A BirdTrack complete list with a start and end time, counts of individuals and breeding evidence has quite a lot of structure.
Last edited:
Nick Moran
Well-known member
Many thanks for flagging this (and for all the other sensible, reasoned points you make). I'll look into it; I'm sure there'll be a simple explanation and remedy.
Last edited:
Nick Moran
Well-known member
This week BTO launched an appeal to raise funds to help understand and account for these and other biases in the BirdTrack data set through a major research project. Needless to say (I hope!) the overall goal is to mobilise the BirdTrack data set to do more for bird conservation. Please consider supporting it.
In the last 24 hours, four of us were involved in discussions with two of BTO's top analysts about the very issue of how to account of the bias of "where birders go birding, or actually live". It can be done, and this project will see it happen.
It may not have anything to do with birders not being out birding because the UK-wide (and Norfolk) reporting rates – % of complete lists that include Blackbird – show a similar drop in late summer and these are independent of observer effort. In other words, it doesn't matter if 10, 100 or 1,000 birders log complete lists; if only 6, 60 or 600 of them log Blackbird, the reporting rate is 60%.
A more likely explanation is that the detectability of Blackbird changes seasonally...and that in itself raises all sorts of very interesting questions. It also has potential implications about what you can 'read into' high counts and other one-off phenomena.
I've attached the historical UK-wide reporting rate for Blackbird for comparison with HarassedDad's 'number of reports' chart.
Attachments
Last edited:
James Emerson
Norwich Birder
Thanks to those who have posted information and opinions so far, and of course to everyone involved in producing the Norfolk Bird & Mammal Report.
The discussion has mostly focussed on the practicalities and benefits of observer selected vs whole sets of data. Whilst it is clearly important that data should be manageable for those who give their time to prepare the bird report and enter records into the county database, it is also important that the end result is representative of the county's avifauna as far as is possible.
To that end, it would be interested to know what proportion of the birders actively recording in Norfolk submit their records via the approved method, and those that solely use BirdTrack. Rare birds (i.e. description species) are already robustly dealt with by the county records committee, and common resident birds represent difficulties in writing anything more than a general summary, but this still leaves a large tranch in the middle, including migratory species and declining residents. These can be some of the most interesting species to read about, and records may be important to researchers in the future even if not included specifically in the bird report itself. If say, three quarters of active recorders submit at least their key records via the desired route, then that gives a good representation, but what if that were to decrease to only a half of birders over the next few years, i.e. the bird report is accurate with regards to rare and common species but only represents 50% of records of the rest? At that point (and I'm picking 50% as an arbritrary figure, I'm sure other people would have a different opinion on how much would be unrepresentative) it would seem inconceivable to ignore an extra 50% of sightings in BirdTrack and available.
I realise that it would be difficult to put figures on the number of birders submitting directly vs BirdTrack only due to people using both methods, but I hope that the number of people submitting records, and the overall number of records submitted, are monitored by someone year on year so that if we do start to see a decrease in people sending in spreadsheets then the use of at least some BirdTrack data can be reconsidered.
Regards,
James
The discussion has mostly focussed on the practicalities and benefits of observer selected vs whole sets of data. Whilst it is clearly important that data should be manageable for those who give their time to prepare the bird report and enter records into the county database, it is also important that the end result is representative of the county's avifauna as far as is possible.
To that end, it would be interested to know what proportion of the birders actively recording in Norfolk submit their records via the approved method, and those that solely use BirdTrack. Rare birds (i.e. description species) are already robustly dealt with by the county records committee, and common resident birds represent difficulties in writing anything more than a general summary, but this still leaves a large tranch in the middle, including migratory species and declining residents. These can be some of the most interesting species to read about, and records may be important to researchers in the future even if not included specifically in the bird report itself. If say, three quarters of active recorders submit at least their key records via the desired route, then that gives a good representation, but what if that were to decrease to only a half of birders over the next few years, i.e. the bird report is accurate with regards to rare and common species but only represents 50% of records of the rest? At that point (and I'm picking 50% as an arbritrary figure, I'm sure other people would have a different opinion on how much would be unrepresentative) it would seem inconceivable to ignore an extra 50% of sightings in BirdTrack and available.
I realise that it would be difficult to put figures on the number of birders submitting directly vs BirdTrack only due to people using both methods, but I hope that the number of people submitting records, and the overall number of records submitted, are monitored by someone year on year so that if we do start to see a decrease in people sending in spreadsheets then the use of at least some BirdTrack data can be reconsidered.
Regards,
James
Last edited:
A lot of my data was just common species saying present, no numbers and 300+ times in Hemsby. It doesn't give a huge amount of information really. We also have to remember that there is space constraint for the report and it is not possible to produce maps/graphs/tables for every species. It was decided that the BBS data was most suitable and with your help we produced some excellent graphs showing declines and increases in some common species over a 30 year period with reasonably constant effort.
Going on to complete lists, I don't feel they give anymore a true reflection of bird status in Norfolk than what is in the bird report at the moment. Firstly, HarrasedDad kindly made the map for Blackbird and if someone outside of Norfolk birding looked at that they would assume that Blackbirds don't occur in two thirds of Norfolk which is blatantly wrong. Surely it is better to cherry pick some interesting sightings of the species rather than produce misleading maps?
On a more personal note, looking at my records and my complete list %'s do not reflect the bird life in Hemsby. I seawatch a lot and thus enter seawatch complete lists but it then means that common garden birds are only on c60% of my complete lists, suggesting that they are not in the area 40% of the time which is incorrect. For complete lists to work there has to be a way to filter out seawatching records etc (maybe there is now but when I asked in the past there wasn't).
You have also said that by not using all my data it makes 95% of my records inaccessible but again this is untrue, they are still in Birdtrack and available to use by anyone and if I'm being honest I always assumed that the BTO would use the data to write papers, reports, atlas's etc and not volunteers writing a annual report which is meant to be a brief summary of the years birds not a scientific journal.
We also have to take in to account that the vast majority of birders in Norfolk (East Norfolk anyway) don't use Birdtrack.
Apologises for this long post, I wasn't going to get involved and I will still be using Birdtrack (I've got a lot of data entry ahead of me from the past couple months!) and these points are my own personal point of view on teh debate.
Nick Moran
Well-known member
Definitely not suggesting or advocating that for every species; could be as simple/space-limited as a numerical metric in the header row for the majority of common species.
Agree that the exemplar map provided could be misleading. As I said, we're working on accounting for the bias in observer coverage, so there's scope for such maps to be get much closer to what Atlas work can yield. However, it is not comparing like-with-like to say it is better to cherry pick unless you propose mapping the cherry-picked records (which I don't think you are).
In general terms, I'm not a fan of cherry picking because of the unpredictable nature of change, and the current rate of change for many species: all too easy to throw out the baby with the bathwater.
We may be talking at cross-purposes though: I see 6,000 Blackbird records as the potential to offer one piece of information, whereas others see it as 5,906 extra bits of information that need processing manually.
You can mark seawatches as a such via the 'seawatch' count type. There isn't currently a way for individuals to filter out such count types but it can be done via the database. If you've already been using this, we could switch all 'Hemsby seawatch' counts to a separate location from your non-seawatch data so that you could generate more habitat-specific reporting rates by site. Email me if you want to discuss that idea further.
Sorry: that was me not expressing myself properly. I'd actually removed the 'not accessible' bit before I saw your reply because I realised it could be misinterpreted. I was just thinking along the lines of the Norfolk bird database; if a consultant came to NBIS about a planning application in Hemsby, would all your records be available automatically? If not, is there a danger that incorrect assumptions could be made about what exists for a given area? It was these kind of 'data flow' issues I was failing to explain; something we're working on through the National Biodiversity Network. Again, please accept my apologies.
Last edited:
HarassedDad
Norfolk County Butterfly Recorder
I fell prey to unconscious bias of course. I'm used to a single database holding all records. Of course in birds there are multiple sets of records from multiple schemes run by multiple groups. And there's no requirement that a county bird report use all the data: it doesn't go away and can be used in other research and other publications.
On the points about the usefulness of casual records. With the right statistical techniques, and a large enough data set all sorts of information can be coaxed out of casual records. Since records contain grid references, and NBIS have a GIS based habitat map, we can assign records to habitats, and select all records from urban areas for example, or heathlands. We can then select only those sites that have ten years of records. You can then deal with a consistent data set that exists independent of recorder effort. Or you can work with all records but control for increasing recorder effort by weighting earlier datasets.
And of course you can use the data without necessarily including it in the final report: I plot sightings across the year for each species when I'm writing the butterfly report - but they don't get printed because space is limited. But when I write that the second brood of swallowtails was poor to non-existent last year I'm basing it on this chart of the number of people reporting sightings per day. Different people may go out on different days, more people may look for swallowtails one year than another - but you can't see them if they're not there, and so the aggregate of all those individual choices results in a chart which is different to what it would be if there were more around in August.
It's like how you can derive the value of Pi by throwing frozen hot-dogs around - and the more you throw the more accurate your value becomes.
http://www.wikihow.com/Calculate-Pi-by-Throwing-Frozen-Hot-Dogs
(and now you've learnt something - a first for one of my posts
)
On the points about the usefulness of casual records. With the right statistical techniques, and a large enough data set all sorts of information can be coaxed out of casual records. Since records contain grid references, and NBIS have a GIS based habitat map, we can assign records to habitats, and select all records from urban areas for example, or heathlands. We can then select only those sites that have ten years of records. You can then deal with a consistent data set that exists independent of recorder effort. Or you can work with all records but control for increasing recorder effort by weighting earlier datasets.
And of course you can use the data without necessarily including it in the final report: I plot sightings across the year for each species when I'm writing the butterfly report - but they don't get printed because space is limited. But when I write that the second brood of swallowtails was poor to non-existent last year I'm basing it on this chart of the number of people reporting sightings per day. Different people may go out on different days, more people may look for swallowtails one year than another - but you can't see them if they're not there, and so the aggregate of all those individual choices results in a chart which is different to what it would be if there were more around in August.
It's like how you can derive the value of Pi by throwing frozen hot-dogs around - and the more you throw the more accurate your value becomes.
http://www.wikihow.com/Calculate-Pi-by-Throwing-Frozen-Hot-Dogs
(and now you've learnt something - a first for one of my posts
)Attachments
I fell prey to unconscious bias of course. I'm used to a single database holding all records. Of course in birds there are multiple sets of records from multiple schemes run by multiple groups. And there's no requirement that a county bird report use all the data: it doesn't go away and can be used in other research and other publications.
On the points about the usefulness of casual records. With the right statistical techniques, and a large enough data set all sorts of information can be coaxed out of casual records. Since records contain grid references, and NBIS have a GIS based habitat map, we can assign records to habitats, and select all records from urban areas for example, or heathlands. We can then select only those sites that have ten years of records. You can then deal with a consistent data set that exists independent of recorder effort. Or you can work with all records but control for increasing recorder effort by weighting earlier datasets.
And of course you can use the data without necessarily including it in the final report: I plot sightings across the year for each species when I'm writing the butterfly report - but they don't get printed because space is limited. But when I write that the second brood of swallowtails was poor to non-existent last year I'm basing it on this chart of the number of people reporting sightings per day. Different people may go out on different days, more people may look for swallowtails one year than another - but you can't see them if they're not there, and so the aggregate of all those individual choices results in a chart which is different to what it would be if there were more around in August.
It's like how you can derive the value of Pi by throwing frozen hot-dogs around - and the more you throw the more accurate your value becomes.
http://www.wikihow.com/Calculate-Pi-by-Throwing-Frozen-Hot-Dogs
(and now you've learnt something - a first for one of my posts
Firstly I'd like to thank Nick for his latest response, I look forward to seeing the results of the work they are doing to counter bias etc.
I'm finding it difficult to be convinced by this more is better attitude from some on here. As someone who has had to deal with large amounts of bird survey data through my work I was always reiterating to my surveyors that quality was the key not quantity. The example given here, Blackbirds, shows that despite there being 100 times the amount of data in Birdtrack all that has done is provide misleading information. I'd prefer to keep it simple rather than have a misleading report.
For all the people asking why Norfolk isn't using Birdtrack when a lot of counties are, I thought I would look into this more and researched a few reports I could get hold of and again the stats probably don't tell the full story. Its all well and good to say that XX% of counties use it but how do they use it? Below is exert from an annual report for Wren:
"On Birdtrack there were 763 records from 171 sites but sadly very few
contributors provided any usable counts or specifically mentioned any breeding information".
In an effort to bulk up a common species account all it has done is highlight that Birdtrack has given them a lot of data but little can be taken from it.In my view quantity isn't always the best but that is just my view.
Again, I'd like reiterate that I use Birdtrack a lot and don't want to discourage people from using it but for 1/2 an hour of your time you could both contribute to the NBMR with your Birdtrack data, its very easy to do. There were long discussions last year and it was decided for the moment to use BBS data to highlight trends and I hope people found the graphs in the last report useful?
crapbirder
Well-known member
Bird Recording.
As one almost totally lacking in electronic data handling skills, may I add my appreciation to those who have so freely and comprehensively given their time and expertise on both sides of this debate. While their efforts have dispelled some of the opinions I held, others have been reinforced.
Have I been sufficiently convinced by either argument how best to submit bird records? - only to the extent that I remain sceptical of the value of any records, regardless of the means of submission, collected at random by observers of unknown ability.
As one almost totally lacking in electronic data handling skills, may I add my appreciation to those who have so freely and comprehensively given their time and expertise on both sides of this debate. While their efforts have dispelled some of the opinions I held, others have been reinforced.
Have I been sufficiently convinced by either argument how best to submit bird records? - only to the extent that I remain sceptical of the value of any records, regardless of the means of submission, collected at random by observers of unknown ability.
Last edited:
Nick Moran
Well-known member
There's a full spectrum of attitudes towards BirdTrack data by Bird Report writers. This example paints a negative picture on face value, hence supports your point.
Without knowing the background and context though (i.e. how much that county is using BirdTrack data and promoting the system), it's potentially misleading. The comment may well have been an attempt to 'prod' local BirdTrack users into adding more value to their Wren records by including breeding evidence (albeit using negative rather than positive wording).
Irrespective of the local situation regarding use of BirdTrack data, please do add highest breeding evidence for all species if you're using BirdTrack because the records are contributing to the ongoing second European Breeding Bird Atlas.
Last edited:
Titchwell May 6th
Today highlights
Wood sandpiper - 5 on grazing meadow this morning only
Short eared owl - 1 hunting over the Volunteer Marsh this morning
Cuckoo - 1 singing from dead trees
Red crested pochard - 4 in reedbed
Med gull - 1 on fresh marsh
Hobby - 2 hunting over fresh marsh
Red kite - 1 west
Little stint - 2 on fresh marsh still
Common sandpiper - 2 on fresh marsh
Paul
Today highlights
Wood sandpiper - 5 on grazing meadow this morning only
Short eared owl - 1 hunting over the Volunteer Marsh this morning
Cuckoo - 1 singing from dead trees
Red crested pochard - 4 in reedbed
Med gull - 1 on fresh marsh
Hobby - 2 hunting over fresh marsh
Red kite - 1 west
Little stint - 2 on fresh marsh still
Common sandpiper - 2 on fresh marsh
Paul
HarassedDad
Norfolk County Butterfly Recorder
If there's one point everyone involved would agree on, it's that, however they're submitted, if at all possible they should be electronic. Retyping paper records is a deeply depressing experience. (But I'd still rather have them on paper than not at all)
As to "observers of unknown ability" - that's why you need humans in the loop to make judgments about the quality of the data - computers can't do it all.
MICK SAUNT
Well-known member
Crapbirder
Firstly, knowing that you are long-term contributor to the NBMR, I sincerely hope that you continue to contribute your records.
Many words have appeared on this forum over the past few days over the merits or otherwise of how the NBMR records and utilises the sightings submitted by observers. The unanimous view of the recording team is still that a system where EVERY record that is submitted to us is assessed (even if it does not subsequently appear in the report) is vastly superior to a 'Wild West' free-for-all system where no meaningful vetting or validation takes place and where quality clearly takes a back seat to quantity.
It seems that even the much-vaunted ability of that system to produce distribution maps and data for individual species turn out to be little more than maps showing the distribution of birdwatchers!
Anyway, many thanks for your comments.
Regards
Mick
crapbirder
Well-known member
Dawn Balmer
Well-known member
Anyone been for the Dusky Warbler today? Is it showing as well as singing? Fantastic find!
Thanks
Dawn
Thanks
Dawn
