Tuesday, 22 November 2016

Atlas 2020 Coverage in V.C. 23

Botanists recording meadow resplendent with devil's-bit
scabious and the fruiting heads of saw-wort,
Otmoor SSSI, August 2016. Image by B. Spence
This is the first of what I intend to be a semi-regular update on progress with recording for the BSBI Atlas 2020 project in the vice county of Oxfordshire (V.C. 23). Perhaps it is a little dense for the blog's first post!

The Atlas 2020 project aims to update Atlas 2000, and will be based on records from 2km x 2km Ordnance Survey grid squares (tetrads) collected between 2000 and 2019. I'm anticipating records from a few more recorders, but as I've been itching to present something since I agreed to take on the job of recorder, here for what it's worth is a preliminary analysis based on what I've received so far (about 17,000 records, some recent, some old). These data are also not yet in the BSBI Distribution Database (DDb), so this is my own analysis. I hope it gives an idea of where we are and some sense of what needs doing. If you would like to know how your own patch is doing then please get in touch.

The Numbers

The pair of maps below show how the county's doing: each coloured square represents a tetrad, with the colour indicating the number of taxa recorded in that square since 2000 in V.C. 23; the larger labelled squares are 10km squares (hectads). The map on the left shows the full picture, while the other only tetrads with more than 100 taxa. I am using number of taxa as a crude measure of surveyor effort; as a thorough session recording a tetrad almost always throws up at least 100 taxa even in rather uninteresting places, 100 seems like a reasonable threshold for considering a square as having received an acceptable level of recording. These squares and their colours represent 2035 individual taxa and 71,358 individual records (albeit before validation etc.).

Qualitatively, some parts of Oxfordshire are obviously doing well. The west of the county is filling in nicely, and this year Jonathan Shanklin, the joint county recorder for Cambridgeshire (V.C. 29), has exerted himself commendably and filled in a large area in the north. Sue Helm's regular square bashing sessions are producing some good results, and the green patch north of Oxford is the work of myself and several diligent botanists during recording meetings I've organised. The Chilterns in the south-east are showing some good signs of progress also. The swathes of red in intervening areas show that a start has been made in the less populous and perhaps less interesting areas of the county, but there's work still to be done there.

Guidance for Atlas 2020 recommends a sample of one-in-five (20%) V.C. tetrads within each hectad, i.e. five tetrads for hectads with all twenty five tetrads in the V.C., four for those with twenty tetrads etc. Supposedly this sampling rate should pick up 80% of previously recorded taxa in the hectad if the tetrads are well chosen, but I won't be able to analyse this efficiently until I get everything into the DDb.

Thirty eight hectads intersect the V.C., only eleven of them fully, and these contain 597 tetrads in the V.C. - so far there are records from 342 tetrads or 57%. However, to achieve the above sampling rate we need to record 239 tetrads or 40% of the total. This is rather more than 20% of the total number of V.C. tetrads as there are so many partial hectads (as can be seen from the map). So far 76% of hectads (29 of 38) have passed the one-in-five sampling rate, but if one excludes hectads with fewer than five tetrads this increases to 83% (25 of 30). Only 47% (18 of 38) have more than one-in-five tetrads with more than 100 taxa recorded, 57% if marginal hectads are excluded (17 of 30) - it'd be desirable to get this statistic up to 100%. I certainly think this is achievable; the hectads SU68, SU69, SU78, SU79, SP22 and SP40 are almost there. The main issue is with boundary hectads, though SP42 is a notable exception.

Hectads with ≥ 20% tetrads with more than
0 taxa100 taxa200 taxa300 taxa400 taxa

In contrast to this, the mean number of taxa per tetrad is respectably over 100 at 112.2. However, the standard deviation is high at 95.6, reflecting the large degree of variation in survey effort, with a quarter of tetrads with fewer than 20 records and also about 5% with over 300. Aiming higher, the proportion of hectads meeting the one-in-five sampling rate with higher numbers of taxa is shown in the table above. I don't know if we'll get the >200 taxa figure very close to 100% in the next three years but it'd be good to improve this statistic. A sample of five tetrads chosen for their range of habitats should be able to produce over 200 taxa each.

Future Work

I think the lessons to take from the above are two-fold. First, we're making good progress in terms of overall coverage of the county assessed against the BSBI's advised sampling rate, but boundary hectads are not adequately covered. This is understandable: some boundary hectads are very interesting, e.g. those along the Chilterns escarpment, but they are a long way away for many of us. If you live in such far-flung places and would like to take on a boundary hectad then let me know!

Second, there is a large variation in survey effort. Myself, I'm conscious of probably over-recording SP51 and I'm sure my time could be better employed spreading the effort, but it's fun getting to know an area thoroughly, especially if it's as interesting as Otmoor. A positive way forward, therefore, would be to see a consolidation of the >100 taxa tetrad map, with tetrads revisited to get them over the 200 taxa mark. Once I'm able to look at the re-recording rate I might be able to give a better idea of where this should be readily achievable and I shall continue to think about this.


  1. Great work on the Atlas recording!

  2. Thanks, Camilla! Well done everyone and keep it up.