fieldherping.eu

Here some data from different papers (I already tried to "translate" everything into the same "format"!!!):

Schwarz, A. (1997): Möglichkeiten der Ermittlung von Raumnutzung und Populationsdichte bei der Kreuzotter (Vipera b. berus L.).- In: Henle, K. & Veith, M. (Hrsg.): Naturschutzrelevante Methoden der Feldherpetologie, Mertensiella 7: 247-260

total area: 368 ha
potential habitat: 147 ha
used habitat: 45 ha
-
1,8 adult animals (counted) / ha (used habitat)


Thomas, B. (2004): Die Kreuzotter (Vipera b. berus [L.]) im Toten Moor in der Region Hannover.- In: Joger, U. & Wollesen, R. (Hrsg.): Verbreitung, Ökologie und Schutz der Kreuzotter (Vipera berus [Linnaeus, 1758]), Mertensiella 15: 175-185

total area: 480 ha
potential habitat: 215 ha
used habitat: 75 ha
0,4 adult animals (counted) / ha (potential habitat)
1 adult animals (counted) / ha (used habitat)


Weinmann, K.., Beck, C., Madl, R., Penner, J., Sound, P., Wollesen, R. & Joger, U. (2004): Zur Ökologie und Raum-Zeit-Einbindung einer Kreuzotterpopulation (Vipera berus [L.)) im hessischen Spessart.- In: Joger, U. & Wollesen, R. (Hrsg.): Verbreitung, Ökologie und Schutz der Kreuzotter (Vipera berus [Linnaeus, 1758]), Mertensiella 15: 197-212

total area: 250 ha
potential habitat: 16,7 ha
-
4,1 adult animals (median from counted animals between 1998 and 2003) / ha (potential habitat)
-


Wollesen, R. (2000): Zur Ökologie der Kreuzotter (Vipera berus berus L:) an einem anthropogen beeinflußten Sekundärstandort.- Faunistisch-ökologische Mitteilungen 8: 9-59

-
potential habitat: 28 ha
used habitat: 4 ha
1 adult animals (counted) / ha (potential habitat)
6,8 adult animals (counted) / ha (used habitat)


Wollesen, R. & Schwarze, M. (2004): Vergleichende Betrachtungen zweier linearer Kreuzotter-Habitate (Vipera berus [Linnaeus, 1758] in der norddeutschen Tiefebene.- In: Joger, U. & Wollesen, R. (Hrsg.): Verbreitung, Ökologie und Schutz der Kreuzotter (Vipera berus [Linnaeus, 1758]), Mertensiella 15: 164-174


-
potential habitat: 17 ha
used habitat: 1,6 ha
0,7 - 1 adult animals (counted 2001 and 2002) / ha (potential habitat)
6,9 - 10,6 adult animals (counted 2001 and 2002) / ha (used habitat)



Looking at my own data (status 2011)...

total area: I simply draw a polygon around all known spots (staus 2011)
potential habitat: Everything inside the polygon that looks suitable for me to find berus ;-D
used habitat: I draw a circle with R=10m around every spot I found an animal, dont ask me what all the others did to measure this!!!
animals / ha: median of the counted* adult animals from 2006 to 2011 / potential habitat or known used area from 2006-2011
* counted means what I found in this year + all animals that I found before and after that year but not in the year!


Pop A

total area: 396 ha
potential habitat: 29,9 ha
known used area 2006-2011: 8,2 ha
1,9 adult animals (counted) / ha (potential habitat)
7,0 adult animals (counted) / ha (used habitat)


Pop A

total area: 147 ha
potential habitat: 27,1 ha
known used area 2006-2011: 8,9 ha
3,4 adult animals (counted) / ha (potential habitat)
10,4 adult animals (counted) / ha (used habitat)


But looking at one subpopulation with a terrifying decrease:
using the peak year (2008) with the known used habitat from 2006-2008 I get 17,7 Ind. /ha
using the worst year (2012) with the known used habitat vom 2006-2012 I get 1.6 Ind. /ha

that really hurts

Daniel Bohle wrote:but if you estimate the population size, these 10 dead animals are not recaptured and the more animals you have that you never recapture the bigger is you population size.

That's where that variance will go up too.

Daniel Bohle wrote:

in real if you have 100 animals and 10 of them are dead at the end of they year you simply have 90 left
but if you estimate the population size, these 10 dead animals are not recaptured and the more animals you have that you never recapture the bigger is you population size.
if I take smaller intervall i get the problems with the different detectability...

you can get software packages for the analysis of this kind of data (I think there is one called MARK that is widely used), which account for mortality in the model.

You are right about defining the size of the the study area though. I think in big interconnected landscapes like your berus forest, this is particularly difficult. In more fragmented populations in the UK it is probably easier.

Great, Liam!

http://www.phidot.org/software/

Thanks for the link Jeroen

It seems MARK allows you to run open or closed population models which should make density calculations more easy as you probably don't need to be quite as exact in defining the study site.

Daniel Bohle wrote:Here some data from different papers (I already tried to "translate" everything into the same "format"!!!)

Thank you for these Daniel, great data and seems to be within the normal 'limits' from what I understand (as in, seems to me that none of the densities stated are unrealistic). I think that the programme MARK seems to be the way to go.

Jeroen Speybroeck wrote: Looks nice. I just think one should never consider population size estimates without some sort of deviance/variance/confidence interval indication - the values by themselves can be misleading (in
relation to their error bars). ...

Fot this particular figures the error bars were not given in the thesis. I can ask the author why.
(But Daniel also provided none in his rather extensive list...)

Do plans exist for this thesis to result in published papers?

Yes, that's the idea.

On a side note, how do you deal with sensitive nature of studied populations? I hear more often
that editors seem to insist on providing precise coordinates...

...and the authors do provide them, if pressed. I have no idea how to solve the problem.

Liam Russell wrote:It seems MARK allows you to run open or closed population models which should make density calculations more easy as you probably don't need to be quite as exact in defining the study site.

The programme MARK version 6.1 was used in the Zagreb case. That's all I know about it.

Berislav Horvatic wrote:(But Daniel also provided none in his rather extensive list...)

Indeed, it's a principle I'm trying to address, not a person.

Berislav Horvatic wrote:Fot this particular figures the error bars were not given in the thesis.

I saw that also in a thesis that was just published on a Flemish berus population.

Jeroen Speybroeck wrote:

Berislav Horvatic wrote:Fot this particular figures the error bars were not given in the thesis.

I saw that also in a thesis that was just published on a Flemish berus population.

The population density estimate as obtained by each of the methods/procedures is a single number
and has no error bars, since the result is not a set of estimates. In short, "a measure of deviance to
EACH n°/ha", which you asked for, does not exist (= is not given by the procedures).
Using several methods gives several different estimates, which can then be compared among themselves,
to get the feeling how much they differ. But it makes no sense to calculate the average value of these
results and the corresponding SE, since each of the methods obtains its result in a different way.
(Anyhow, that's NOT what you asked for.)