UNDERWATERSURVEY 129
of the control points are then considered to be fixed, so
the addition of subsequent detailed measurements cannot
alter the positions of the control points. Measurements
from four or more control points can be used to position
detail points on features such as artefacts and structure.
Survey work on land is based on triangles because the
tools used for land survey, such as theodolites, measure
angles. Survey under water is largely done with tape-
measures, measuring distances rather than angles. If three
points are set out on land and the angles between them
measured and added together, the total should add up to
180 degrees. If the total is not 180 degrees, then the dif-
ference is a measure of how good the angle measurements
were. Unfortunately, this does not work with tape-
measurements, as three distance measurements will almost
always fit together, so a ‘braced quadrilateral’ is used
instead. The braced quadrilateral or ‘quad’ is made from
four control points in a square or rectangular shape, with
the sides and both the diagonal distances measured
(figure 14.12). If the positions of the points are plotted
onto a piece of paper, the four points can be positioned
using five of the six measurements and the sixth mea-
surement is the check. In most cases the check measure-
ment will not fit perfectly and the size of the difference
gives an idea of how well the other measurements fit
together. If the check measurement is significantly differ-
ent to the value expected, there is a mistake in one of
the six measurements.
A complete control network can be as simple as just
these four points. However, if tape-measurements are
kept to less than 15 m (49 ft), this can only cover a small
area of a typical site. To cover a larger area, a number of
quads joined at the edges are used (figure 14.13). Avoid
placing control points less than 5 m (16 ft) apart, as this
will not improve the survey but will waste a lot of time
in positioning the extra points. Install the control points
high up so they have a clear line-of-sight to as much of
the site as possible but note that high points with a good
line-of-sight are often the most easily damaged.
Once a couple of quads are joined then it is possible
to measure between points in adjacent quads. The net-
work in figure 14.13 has the barest minimum number of
measurements to be able to compute the positions of the
points in three dimensions. If a mistake occurs in one
of the measurements, it will be difficult to find and fix.
By adding more measurements, as in figure 14.14, more
information about the positions of the points is available,
so any measurements that are mistakes should be easier
to find. Sets of quads can be extended to cover large areas
but, on very large sites, they can be time-consuming to
set up. In some cases, it is easier to treat groups of fea-
tures more than 30 m (98 ft) apart as separate sites.
The best network shapes for surveying are circles and
ellipses (figure 14.15). Circular shapes can be used but not
many sites are circular, so in most cases an elliptical net-
work is required. The ratio between the length and width
of the ellipse should ideally be less than 2:1 (less than twice
as long as it is wide). To achieve this sometimes requires
extra ‘outrigger’ bracing points to be placed either side of
the site. It is essential that all control points for a site be
connected together into one network. Separating parts of
the site into smaller, unconnected networks will cause prob-
lems when drawing up. Where the extents of the site are
Figure 14.12 Braced quadrilateral (3-D survey) not entirely known, the control point network may need
Figure 14.13 Joining quads (3-D survey)Figure 14.14 Measuring between quads (3-D survey)