Number of clear trap days

Not from NARRS no. NARRS have gone more the way of a regional occupancy model, but some of the supporting DICE research might be applicable.

But, from SARG stats...yes (sort of). This is a version of probable absence, as used for a pre-cursor to reintroductions.

We can generate detectability stats for month of year and met conditions. Probable absence can be calculated from the number of negative results, accounting for the 'minimum' detectability.

'Minimum' is a challenge. For non-mitigation sites, we use the lowest detectability value for viable sites, as we believe this represents the smallest or lowest density populations. I'd guess you'd use the same approach here.

Basically, each survey is not treated the same, but is weighted according to met conditions and month of year. This 'weight' is removed from 1.0 (certain presence) until a threshold for absence is reached.

So far, we can't see any clear correlation between detectability and refugia density, which is counter-intuitive. Probably, the variation in our refugia densities isn't wide enough.

A near-term goal is to write up the approach in a simple manner, which can be exploited in the field (rather than using a computer model - although it could easily be a web-based application).

You also need to determine the target confidence level (nothing is absolute in stats!). For re-introductions, NE have mandated a confidence of absence of 95%, which is really high.

The number of clear trap days to determine probable absence also varies with species. But as a guide, for snake species, at C95%, you're probably looking in excess of 20 clear surveys!

I'd have thought that a slightly lower confidence level was more practicable, and would decrease the number of negative surveys required.

g

Edited by Scale

Vicar wrote: Have been thinking more on this, and I can't see an easy way of incorporating all the Met variables into a simple table. Also, some met variables affect some species more than others. E.g. Vb are more sensitive to cloud cover than Af. I think a half-way house would to create a table of expected detectability (per hour of survey) per species by month of year. This would assume average met data for the month, but could be assured by the met conditions for each survey being reported in an annex of any report. (How I'd love a reptile survey reporting standard!). So, for Ca, surveys in August would carry a high weighting (the peak detectability month), compared with surveys in March (when only half the snakes have emerged from hibernation). So, you'd need fewer clear trap days if surveying in August, and more if the survey was in March. This feels like a better solution than a flat number of clear surveys, and although not perfect, would at least be evidence based. Exactly the same process could be used for initial presence/absence surveys.

This is the thing though that I was trying to point out last night (although I didn't have much time to post the message). The number of clear trap days depends on what species you have present and the time of year that you are doing this.

Tended to find that if you are dealing with slow worms, the number of clear trap days will be lower. 

There is some guidance in the ARG Guidance on reptile translocations which outlines the standards in best practice.

Absolutely Steve, this is why I have big problems with an arbitary figure.

ACO are only one tool in capture works.

By graphing the capture results I get a high degree of confidence that the site is actually being cleared. One can look back over months and see the capture rate reducing until eventually sightings are down to single animals, sooner or later we simply know who's left onsite using ACO and can gauge ACO effectiveness is over. Time for destructive search.

Compare to a sample mitigation where the consultant caught 30 slow worms in a day then no sightings for a week. 7 CTDs, must be time for sign off? Nope, just shows they had a particularly good day and intercepted the proportion of the population using ACO at the time in one hit.

Next time it rains, they'll be more again. It could be 10 or 20 days or more before it rains. If though guidelines just say 10 days to sign off, well you know what's going to happen. Chances are next time the peak will be 20 animals, the next 10 and so on. This is how I work.

I've a feeling we should be looking at the data from capture works to form a model for a degree of confidence. Again though it's not site clearance, it's just the end of ACO effectiveness. So I'm really thinking here CTDs should never be a criteria for sign off. However convenient it might appear. Before we know it clients are telling us that 5 days without sightings means it's over, nope it's over when I say it is

It did though prompt me to look at presence/absence surveys we did in the past.

9 - 18 visits to establish adder presence (not visual spring survey)

5 - 8 visits to establish grassies

3 visits to establish slow worms

1 visit for Zv

It would be kind of crazy though to check ACO for another 20 days at the end of a mitigation when the vegetation is mostly gone and any remaining animals are not using ACO in any case.

It should be Steve, but I'm hearing more and more about avoiding fencing and manipulating habitat in early stages. To me it just seems like carnage for the reasons you have given. God only knows we moan enough when conservation organisations do this kind of thing!

We don't manipulate habitat until ACO are practically not producing, certainly I would want all gravid animals off the site and most if not all the adults. Then we strim between the ACO leaving little islands. By then though I would be looking at just a few individuals left onsite and just a few weeks before calling an end to ACO capture work. Sometimes we find grass snake get 'flushed' under the ACO. Been there all the time not using the ACO but when the covers gone, they hide under them. 

I can't imagine why anyone would think it was a good idea to reduce available habitat early during the works for animals that are going to be stuffed in a bucket and moved. But didn't Jon post something from NE a while back saying this is the way they wanted to go to reduce client cost and time??? (Hopefully I misunderstood it)