Sorry for any duplicates...
NHBC&EA (2000) give a guidance on the number of sampling locations we are
looking at when a SI is carried out in a house development area. Part2 Step
5 (page 29) reads: "Typical densities of sampling grids can vary from 50m
to 100m centres for exploratory investigations [...] and from 20 to 25m for
detailed investigations. Where contamination is known to be localised, or
the presence of highly contaminated hot spots is suspected, sampling grids
of 10m centres may be necessary. Sampling at this density may be also
appropriate where it is intended to relate the size of the grid in the main
investigation to the size of garden areas. If this approach is taken [...]
data from at least one sampling position within or close to the boundary of
the plot will be available. However it is unlikely that sampling of an
entire site would be required at such a density".
I have a couple of observations regarding this guidance:
First:
It is conceptually different searching for hot-spots, from sampling to
assess the overall chemical nature of a large area. In the first case it is
obvious that the size of the sampling grid must be related to the likely
size of the hot spot (N.B. likely means that it could well be bigger than
one garden and potentially refined in case a hot-spot is hit), whereas in
the second case only a sufficient number of samples (therefore of sampling
locations) need to be chosen for the investigation. In the latter case a
more detailed investigation is triggered in case initial SI results (e.g.
95% one-sided Upper Confidence Limit of the mean) show evidence for the
site-wide average concentration to be greater than the SGV (or SSAC).
In both cases statistical procedures can help to assess the most
appropriate number of sampling locations (and therefore the dimension of
the sampling grid) taking into account our prior knowledge of the site.
Second:
Sometimes site investigations have more than one purpose (for example a
site investigation might aim at hitting and characterising the extension of
hot spots AND assess the overall chemical nature of the site at the same
time). In these cases statistical procedures can assist in determining the
appropriate number of sampling locations.
In other words there isn't a unique answer to all the questions on
contaminated land investigation, but there is a cost effective answer to
each specific case. This is usually guided by a statistical principle.
At a first glance the case presented by Wendy Lilico seems to require a
combined solution (if a source proves to be contaminated then search for a
hotspot).
Statistical guidances, in my experience, point you in the right direction,
but are a bit vague and sometimes require a stronger statistical background
to enable an appropriate answer to these 'combined' problems and sometimes
it would be appropriate to seek expert guidance.
Regards
Paolo Masi
Paolo MASI
Specialist in Applied Statistics
___________________________________
ESI Ltd
Soil and Groundwater Specialists
New Zealand House
160, Abbey Foregate
Shrewsbury SY2 6FD UK
Telephone: +44 (0)1743 276100 Fax: +44 (0)1743 248600
email [log in to unmask]
<http://www.esinternational.com>
This email and any files transmitted with it are confidential and intended
solely for the use of the individual or entity to whom they are addressed.
They may also be legally privileged. If you have received this email in
error please notify us by immediate reply and destroy any copies
|