Adder Dispersal Distances

My current take on available literature suggests that adder can disperse over
distances of ~ 2km (principally between hibernacula and summer foraging sites).
In reality, the majority of such movements are likely to be much less than this -
particularly give the fragmented nature of today's countryside. What experiences
or opinions do other folks have?

You can stop reading this rather long post now! For nerds interested in database
stuff I have attempted to explain how I will be using this info below

I'm attempting to pin down potential movement distances for all species as part
of my attempts to better define favourable conservation status in Kent. This
follows on from a meeting with NE and HCT that Steve (Vicar) and John Baker also
attended.

Still a way to go, but results to date are encouraging. KRAG have now been
promised full access to the Kent habitat data - provided we can justify that
access. Before making the request, I'm trying to put some of the background
processes in place to demonstrate how the data will be analysed.

One of those processes has been to calculate nearest neigbour distances for all
records in the database. The database has been able to calculate distance
between records for some time, but this information is calculated on the fly and
not saved. I have now added a script that calculates nearest neighbour for each
record and saves the result in a separate field. Unfortunately having more than
22,000 records on the system means that there are several thousand records for
each species and this process takes a little while. Each of the species records has
to calculate distances to every other of the records of that species to find out
which one is the nearest

One application of nearest neighbour is to illustrate metapopulation ranges. As
an example of this, I have identified sites that have at least one other adder
record within 0.1 - 2 km, and then plotted those sites on Google Earth with a
circle of 2 km radius. Other validated records that do not have a neighbour
within 2km are plotted with a circle of 500 m (just to put them on the map and
differentiate them from the main metapopulation - these records could also be
plotted with a 2 km radius). The results are useful for identifying areas where
survey effort may be better directed. See the attached graphic:





In trying to understand metapopulations (and defining favourable conservation
status), I aim to look at the habitat that is available within each of these defined
polygons.

It may be better to consider range as 'potential range' and give distances that are
defined by the time of year. For example, maximum dispersal distance is likely to
be between hibernacula and foraging site. It may therefore only be relevant to
plot distances (whether they are 2 km or another value) for records generated up
to the end of April. Records collected during this period are likely to be for
animals situated close to their hibernacula. Another possibility is defining
potential dispersal distances by the number of sites that are situated within a
predefined distance - the more sites located within the target range, the larger
the potential maximum range plotted (up to a maximum limit that equates with
the target range). The latter approach may help to control for populations where
the factors limiting dispersal are unknown - sites with good dispersal potential
are likely to have more records situated nearby. In fact I quite like the latter
approach...    

Clearly there is more work to do and I'm likely to be a very busy boy over the
next few months!

Caleb wrote: You could also reject anything that's obviously outside the range you're interested in; if N-S or E-W distance is >2km, then you don't even need to consider it as a possible neighbour.

True, but setting the maximum distance can be problematic. For adder the
most isolated population is currently 3.3 km and this figure varies for
species (10.6 km for marsh frog). I do try to speed up the analysis using the
most efficient algorithm that my brain can deal with and since I am
interested in the distance to records that are not situated on the same spot,
I also exclude duplicate 6 figure grid references (I chose 6 figure rather than
8 or 10 to maximise the number of included records).

But I want my results now !   

armata wrote: Neos don't disperse until their second or third year

This I had not heard!

Question...

I've often seen neo adders hunting lizards at Thursley on the wet bog areas. Presumably the hibernaculum isn't very close by, as I would have thought that flooding would be an issue.

My conclusion was that neos may also migrate to summer grounds (following scent trails).

Maybe there is a hibernaculum close by? or maybe animals behave differently in different environments?

I guess it's hard to say without definitive mark-recapture.

armata wrote: Male, female immature, neos display different dispersal. e.g. breeding females stick around the hib area. Neos don't disperse until their second or third year and return to hibernation sites at random i.e. following conspecific scent trails. (Genetic stability maintained throughout an area thus).

I've just exchanged emails with Gareth discussing this very issue. A
problem is the number of available records for different lifestages. For
example the db holds 1139 adder records, but only 227 of them are for
immatures.

Whilst I accept that different lifestages will have different dispersal
distances, ultimately what I am hoping to illustrate is the maximum
potential dispersal from each site (as defined by a 6 fig grid reference). It
*may* therefore be appropriate to use the overall maximum dispersal
distance for the initial analysis.

Of course there is no reason why the db cannot hold the additional
dispersal data and the analysis updated once sufficient records become
available.

I am at this very moment trying to get to grips with this issue myself Suzi.
In a nutshell, the larger and more connected the site the better - but
what we really need is a HSI that is directed towards reptiles.

To date my efforts on development of a reptile HSI have been thwarted
by the issue of site definition. For amphibians, HSI's are usually applied to
a pond with each pond representing a discrete 'site' that is relatively easy
to define (so much so that I have recently created a new table in my db to
hold waterbody as site data, but I digress...).

Defining a reptile 'site' can be very problematic and the habitat factors
that need to be quantified are likely to vary quite considerably across
large sites.

Rather than attempting to define HSI based on 'site', it may therefore be
more useful to define HSI based on recording unit. Of course, this
approach may result in area becoming a fixed variable (e.g. hectare
blocks of land). I'm not sure if this is useful and it certainly wouldn't help
answer your question. It may therefore be acceptable to attempt to define
the recording unit as a clearly demarcated block of habitat (e.g. a
grassland field or even a patch of scrubby grassland in the middle of a
large expanse of arable). The larger the 'site', the more of these habitat
units there are likely to be.

An alternative approach is to define reptile HSIs in broader terms at the
landscape level - but individual site area could be lost in such an
approach. In effect this is what I am trying to achieve with the favourable
conservation assessment of adder in Kent - available suitable habitat vs
occupancy of said areas etc.

It may be worthwhile spinning off your question into a new thread - what
is the smallest discrete block of land in which adder has been recorded?