BSBI Distribution Database > message board

Tetrads - well recorded or not??

Any queries about using the BSBI database

Tetrads - well recorded or not??

by DaveLVC75 » Thu Aug 11, 2016 10:39 am

This may already have been explained in a newsletter etc., but -

I find the Atlas 2020 Coverage summary information that you can access on the Distributional Database a bit perplexing.

Firstly there is the – sometimes fairly significant - time lag that seems to exist between records being synched from Mapmate to the hub and their then impacting on the summary data. But that is probably to be expected with limited staff resources etc. to work on things, so I’m not overly bothered about that one.

What is far more confusing however is how the summary tables determine whether a tetrad is “well recorded” or not.

Firstly – unless I’m missing something - it seems that unless 100% of historical records are refound, a tetrad is deemed poorly recorded. Which makes no account of species being lost to that area over time or new ones coming in.

Secondly –

To take an example in our VC75 - Ayrshire – a tetrad like NS40P has a total of 264 species recorded. As far as the data suggests, it wasn’t visited prior to 2000 – so there is a 100% post-2000 find rate. Yet it is deemed to be poorly recorded on the system.

Yet a tetrad like NS42S – again seemingly unvisited prior to 2000 – has a mere 105 species recorded, again therefore with a 100% post-2000 find rate, and it is deemed well recorded.

Am I – as is more than likely - missing something obvious here?

Because at present, a couple of people I’ve spoken to are not putting a lot of faith in some aspects of the data summaries presented on the system. (Although obviously the hard numbers are useful regardless of qualitative measures such as whether something is well recorded or not.)

I’d be very grateful for peoples’ thoughts.


Many thanks,

Dave Lang
Joint VCR
VC75 Ayrshire

Tel. 07766 505253
DaveLVC75
 
Posts: 1
Joined: Mon Jan 05, 2015 3:58 pm
name: Dave Lang

Re: Tetrads - well recorded or not??

by admin » Thu Aug 11, 2016 11:40 am

Tom Humphrey
Database Officer, Botanical Society of Britain and Ireland (BSBI)
c/o Centre for Ecology and Hydrology,Maclean Building, Crowmarsh Gifford, Wallingford, Oxon, OX10 8BB, UK.
tom.humphrey@bsbi.org
User avatar
admin
 
Posts: 438
Joined: Tue Nov 20, 2012 4:16 pm
name: Tom Humphrey

Re: Tetrads - well recorded or not??

by Angus » Thu Aug 11, 2016 1:00 pm

If you look at the section above the data tables for the individual tetrads, where you can set the parameters for a 'well-surveyed tetrad', you will see a check-box 'Include count of new (post 2000) finds when assessing survey coverage'. If you tick this box you will find that tetrads such as NS40P change to well-recorded. But you have to tick the box every time you look at the table, as it defaults to unchecked.

This solves the worst of the anomolies, which arise from the requirement that a certain percentage of old records should be refound. The default is 75%, and you can alter this if you wish, but any positive value will result in a failure to achieve well-recorded status in every case where only a few rarities were recorded in the tetrad pre-2000 and none of them re-found, no matter how many new finds were made. That is, unless you tick the box as described above.
Angus
 
Posts: 14
Joined: Wed Mar 13, 2013 4:38 pm
name: Angus Hannah

Re: Tetrads - well recorded or not??

by AndyAmphlett » Fri Aug 12, 2016 8:33 am

Hi Dave,

as you will have seen, the 'Hectad and tetrad recording coverage' data that the DDb presents to VCRs via the My County > Atlas 2020 survey coverage tabs, is based on a spreadsheet that I created back in 2014 (with a finalised version in January 2015). This spreadsheet is still available via the link on the 'Atlas 2020 survey coverage' page. A number of VCRs found this useful, and Tom made similar results available on VCR's my county pages. This removed the need for VCRs to faff about running DDb queries and manipulating an Excel file.

In implementing the on-line, DDb version of the original spreadsheet, Tom made a couple of improvements, and a third change added sophistication to the analysis, but has caused some confusion. The improvements are that the calculation of Species not refound since 2000, Species refound since 2000 and Newly found species since 2000 are done differently in the DDb version than in the original spreadsheet. Tom's implementation is better than my earlier version. The second, very significant improvement, is that results of these three queries are linked to the DDb report.

Where the DDb version differs from the original spreadsheet, is in how (at default settings) it indicates if a tetrad is 'well recorded'. Below I show the options boxes for setting up the parameters for the the DDb to calculate 'hectad and tetrad rcording coverage'.

Image

As Angus indicated, for the DDb results to match the original spreadsheet, you MUST tick the 'Include count of new (post 2000) finds when assessing survey coverage'. Personally, I think the default should be that this option is ticked, but just remember to tick that option. Once this option is ticked the 'well recorded' test is met when the minimum species per tetrad target is met AND the percentage threshold is met. Assuming you have ticked the 'Include count of new (post 2000) finds when assessing survey coverage' box, then the percentage threshold target is NOT a re-recording target; it is the percentage of total taxa ever recorded in the tetrad that have been recorded since 2000. By default, this option is not ticked, and then the percentage threshold test is, indeed, a test of re-recording, ie. the percentage of taxa recorded pre-2000, that have been found since 2000. This option, to test recording coverage in two different ways is, potentially, useful to VCRs in different situations, but has caused puzzlement. I hope this clarification helps.

Assessing if a tetrad is 'well recorded' can be done in a multitude of ways, and the approach adopted here is (as long as you tick the 'Include count of new (post 2000) finds when assessing survey coverage' box), the one I came up with. To be indicated as 'well recorded' a tetrad must meet BOTH of two criteria; that the number of taxa ever recorded exceeds a threshold, and that the percentage of the all time taxa total recorded since 2000 exceeds a threshold.

I can not stress too much that the default thresholds are just that, and the VCR must decide what levels to set them at. They are not one size fits all vice-counties. Because the DDb results update in real time, it is quite instructive to play with the values and observe how results change. For many vice-counties, I would recommend that what the DDb page refers to as the 'percentage re-find threshold' is reduced from 75% to a slightly lower value, eg 65%. The minimum species per tetrad threshold needs to be set with regard to local conditions. In a well recorded vc in the south of England, the threshold should be much higher than 100. In the north of Scotland, 100 is not a bad average target.

The other two options are to set the target number of tetrads per whole hectad, and whether to include tetads with less than 1km2 within the vc (ie part tetrads). 5 tetrads is the BSBI's suggested minimum, but, again, in a very well recorded vc with lost of active recorders, 5 is too low, and it should be increased.

Note, that hectad scale recording is summarised on the same page (rows in yellow), and I would recommend that users look at these results carefully, as a good aid to planning survey work. And finally, you can highlight, copy and paste the results into a spreadsheet (though on my pc not all the links are retained).

I hope this helps to clarify how to use this tool. But it is just a tool, so please make use of it in ways that meet local requirements and preferences.

Andy.
AndyAmphlett
 
Posts: 408
Joined: Fri Nov 23, 2012 9:32 am
name: Andy Amphlett


Return to Help and support

cron