RAUK - Archived Forum - Evaluating population sizes and capture e

I agree totally !

I've been chatting this over with Chris G-O at HCT. I have a bee in my bonnet re the assessment of likely non-presence of a species at a site too. How many visits without a sighting is sufficient? My personal view is much in line with Lee, applying statistical methods to legacy records.

The only reliable method re pop estimates I have come up with so far are any of the various mark/recapture methods. But the shortcommings of this approach are manifest. Clearly there is likely to be a two-tier approach to this; a 'quick and dirty' estimation methodology (general applications/consultants) and a more rigorous (and more accurate?) approach (herpetologists ?)

I would be very keen to actively support any widespread attempt to crack these issues.

My only slight concern with a workshop is that; a significant amount of work is necessary to crack this. So long as a workshop agrees general guidance and the will in principle to reach a common goal, and perhaps the adoption method, that's fine. Agreeing on 'numbers' without this pre-work would attain consistency, but is unlikely to be particularly useful (scientifically).

Comments ?

Well there are currently mitigation guidelines being written for English Nature by Cresswell Associates - I think they may be finished - but I do know that they want to provide more information on assessing population sizes from survey results -

You left out the conservation bodies in the list of the two tier approach - there is a third lower level - where reptiles arent even bothered with - for instance in many nature conservation management plans - reptiles oh they can look after themselves......bring on the bulldozers (bovine or otherwise!)

Still wouldnt reptile surveys have to go on for several years or decades to try and answer the question - how big is the population size? wouldnt reptiles undergo the same population 'highs' and 'lows' as amphibians and other wildlife species -

The approach to the different species would be different from each other - for instance the snakes would need to be considered differently to the lizards etc

JC

PS it would make a good workshop for the next HGBI Herpetofauna Workers Meeting

Hi all

No doubt this is something that is needed. But it is appallingly complex. I have a pretty good "feel" for Common Lizards - but there's no science - just years of experience saying "habitat looks like thus, lizards seen X thus population Y"! Not a lot of use.

With Sand Lizards for example, it is far easier to spot them and to evaluate populations in heath than in dune. With intensive surveying in heath you can get to know individuals and come up with a very good estimate. With large dune complexes you count yourself remarkably lucky to ever see the same individual lizard twice or more on separate visits.

An example is in the Merseyside dunes where we are constantly increasing the estimated population - what we don't know is whether this reflects climate change extending the population, increased monitoring, improved monitoring skills or a combination of all these factors and possibly more.

Basically, every estimate we make with these populations has been blown out of the water by further monitoring - however scientific we try to make it!

Precisely what I was trying to indicate by example, Lee.

There are simply so many variables to take into account - we don't even know how to evaluate them. For example surveyor "skill" - what baseline do we measure from, how do we set a scale on it, how do we apply it to an individual (without causing internecine warfare ) and so on.

Inevitably some species/habitats are easier than others - the adder hibernaculum being a case in point - but even then - how confident can we be of its accuracy? Having said this almost every method we currently use can be out by several orders of magnitude - even getting +/- 50% would be a huge improvement in many cases.

I have a feeling that pop estimation methods can't progress much further without some confidence in the ground truth for a selection of sites.

Were it possible to know the population by species, for even a handfull of sites; then various methods could be tested (by species) and appropriate calibration or estimations applied to more simplified methodologies. Clearly, the more sites where this can be achieved the better, but even a small number initially would provide a good working premise.

Next comes the thorny question of HOW to we obtain such a ground truth? My thinking is that it is likely to involve detailed work over a number of years, which is why veterans' work, such as Tony Phelps(no offence intended Tony ), where sites have been monitored closely over a good number of years is so important.

I can't see a 'quick-fix' for this one.

I'm new to this, so apologies if this is old hat:

Imagine a bag containing an unknown number of white marbles. Now add, for example, 5 black marbles which are otherwise identical to the white ones. Next spend some time withdrawing, noting the colour of the withdrawn marble, then putting it back in the bag. You can make an increasingly accurate estimate of the total population from the emerging ratio of drawn black and white marbles. If, for example, after 100 withdrawals and replacements the tally is 50 black and 50 white, we can conclude that there are probably 5 white marbles.

The key is to have a known number of recognisable individual reptiles or black marbles!. The withdrawals and replacements are of course sightings.

Hope this helps .

Jim,

Good standard scientific approach that ! Unfortunately, it relies somewhat upon the assumption that 'detection' probability is similar, and that you WILL detect every marble, given enough time. (although I take your point about the estimate improving over time). The other issue in my mind is how many surveys are sufficient ? Adders are relatively easy to distinguish between individuals, but I have a harder time with some of the other species.

Some of our little blighters (I'm thinking of Ca), I personally find very difficult to detect visually, so tins are needed. We do see many reptile species under tins, but are these the individuals who 'like' tins ? what is the proportion of individuals who use tins ?

Nevermind the negative tone above, I actually agree whoeheartedly with the principle you suggest. Yes this is a tough nut to crack, but far from impossible to improve upon what we currently have.

Detection is key, just to fuel the debate, I offer something I've been working on, and would be grateful for comment. (this is just the headlines - lots of hidden subtleties). I'm trying to generate a simple mathematical model to assist with detection.

Hi Steve

I think you do have the beginnings of a potential mathematical model here. Unfortunately its limitation appears to be its link to species that use refugia. I suppose to an extent, with La for example, focii might give a similar input. There is still a potential weakness in that such focii are not necessarily always found - nor do we know what proportion of a given population use them. Conversely, with the refugia using species such as Ca you have effectively eliminated variability in the surveyors so at least one of the "unknowns" has been dealt with. You also raise an interesting question regarding tin usage by herps. Is it, as you say, simply certain individuals that use them. Certainly with Lv I believe that any/all individuals will use them - albeit usually basking on top rather than underneath! Nonetheless in a recent Lv/Af rescue I estimated populations based on sightings on and under tins as 250/100 in 1 Ha in four visits in good conditions multiplying individuals see by 5. Animals actually captured and removed was 450/200 - so what the hell do I know!

Hi Lee

I must agree your first point, also addressed in your final paragraph. I must admit I have so little opportunity to see Lv in the wild nowadays that the generally carefully selected weather conditions late in the season would definitely produce a bias towards refugia for additional warmth. And there is no doubt about variability of population density whatever the species.

And, indeed, I stand corrected re the surveyor experience. I only get tin placement wrong intentionally (as in I've got so many I'll chuck one here in case the lizards don't agree with me) but I must admit it would be very easy to do so. In fact, in the rescue I mentioned I had a couple of inexperienced helpers at the tin placing stage. After carefully describing and demostrating appropriate locations it felt like I spent the rest of the day saying "No, not there because.....".

Makes a helluva job with Steve's model, or indeed any other when it is filled with nothing but variables.

Cheers guys,

this is all useful debate. To be honest, I was waiting for the fixed factors comment . its supposed to represent factors beyond the control of the surveyor, the terminology is a military hangover.

As for placement of refugia, its what I mean by physical placement, although as Chris suggests, it's well worth placing tins in less obvious places. I'm a firm believer in that if you only look for targets in the places you expect to see them, then you'll only detect them where you expect. As Gemma has said before, nature doesn't follow many rules closely (or should that be absolutely?).

Part of the rationale with building any model is to expose the issues and develop understanding, its pretty much a given that any model will be wrong, its just a case of how much practical value it produces somewhere along its development.

You have a military hangover? How unpleasant!

Further to Gemma's previous comment my parallel one which I often use is "To the best of my knowledge the lizards (or snakes, or whatever) seldom read the books by the experts." Certainly most of us have come across various herps in atypical habitat and behaving in atypical manner at some time or another.

My chief frustration in this debate thus far is that it is so easy to pick holes in someone else's ideas - but I feel terribly guilty doing so because I'm damned if I can come up with anything better! Or, indeed, anything!

The single greatest problem that I can perceive is that we do need a fixed reference from which we can work. But all the potentials appear themselves to be governed by a number of variables.

For what its worth we loosely estimate Sand Lizard populations on heath as around three times the number of individuals seen in at least four visits in otimal conditions. Even more loosely, in dune, experience so far suggests that the factor should be more like seven. But even that is difficult because the nature of the habitat is such that we might never be able to confirm our estimate.

It does, however, seem to me, that if we could assemble and analyse the results of various rescues, together with all possible details from preceding surveys of the rescue site, habitat type etc we might have something to work on. After all this is the only occasion on which we can be sure that a reasonably substantial plot of land actually surrenders all - or at least the majority - of its animals.

Dear all

I would agree that rescue sites seem to be the only real opportunity to try and answer these questions - but would the secretive reptile mitigation industry give up their records or do we need to start up a reporting scheme for consultants to report their data and findings.......ohhh another part of the NARRS project??

There are several projects looking at reptile mitigation - Cresswell Associates are looking to compare survey results with total captures - along with habitat criteria - I also know that the numbers of questionaires being returned is quite low - incidently not many reptile records or even crested newt records are passed onto local recorders - not even the BRC!!! - does make you wonder where the records are?

I feel that this debate is needed to try and push reptile conservation rather than 'protection' - translocating thousands of reptiles is clearly not conservation - rather the focus should be on habitat protection, creation, enhancements and monitoring - Research should also be included in reptile mitigation schemes - an agreed format of recording data across the country would help - Reptile habitat suitability index, standard survey methods etc

JC

Barry, obviously you fired up an interesting debate but you are right, we have strayed from your initial question. Returning to your specific point while what you say sounds good in theory I am unsure it holds good in fact. I am, for example, unsure of the origin of the HGBI figures you quote. If you hunt around enough you will find density figures here and there BUT most of them are either informed guesses, specific to only one habitat type/climate etc or simply quotes of the other two.

Now, having said this, from nothing but personal experience (informed guess!) I would not disagree greatly with any of the figures you quote. Less than 20 Common Lizards per hectare is certainly a very low density population - and, by definition this includes 5 (which is a very, very low density!). So far as Adders are concerned these are found at much lower densities and over wider areas. Thus, 15 adders over 10 hectares represents 1.5 per hectare - certainly both a low density and less than 2 !

Conversely most of those 15 adders might be stuffed into 1 hectare and most of the 20 lizards stuffed into 20 x 20 metres within the hectare. This is one respect in which even measuring density as animals per hectare can be grossly misleading. Clearly, if 15 of the Common lizards were in prime habitat of 400 sq metres we would have a population density there of 375 per hectare which represents a reasonably dense population (and yes, I know a population fairly near me that does broadly match those figures).

I am, in fact, less than convinced that we can improve significantly on these specific definitions. The real problem is that which this thread has people trying to address - not what constitutes as low/medium/high density - but how, other than by destructive search can we calculate density.

So, I think that we are actually moving towards what you want - or at least trying to. If we had a reliable method of assessing populations (and note I say populations rather than population density) then we would be in a position to re-define low/medium/high density - if necessary.

Lee, you are quite correct in that I am seeking a "detectability constant". Still a variable constant but one that varies within definable parameters so could become an absolute in each specific locale. I recognise that it might not be truly scientific and disciplined - but nonetheless better than what we have.

Sadly, as Jon says, there do seem to be a number of members of the mitigation/consultancy business who do not seem to be prepared to share this information. Generally, and unsurprisingly, these seem to be the ones who regard it as money for old rope and have little care or concern for the animals themselves. So, while I think my proposal has some validity, as Cresswells are discovering it may still not be wholly practical.