Geoscience Reference
In-Depth Information
water budget. It may also place the trees under stress
through lack of water, thus leading to an altered
canopy.
There are different statistical techniques that
address the spatial distribution issues, and with the
growth in use of
Geographic Information Systems
(GIS)
it is often a relatively trivial matter to do the
calculation. As with any computational task it is
important to have a good knowledge of how the
technique works so that any shortcomings are fully
understood. Three techniques are described here:
Thiessen's polygons
, the
hypsometric method
and the
isohyetal method
. These methods are
explored further in a Case Study on p. 31.
Stemflow
The normal method of measuring stemflow is to
place collars around a tree trunk that capture all
the water flowing down the trunk. On trees with
smooth bark this may be relatively simple but is
very difficult on rough bark such as found on many
conifers. It is important that the collars are sealed
to the tree so that no water can flow underneath
and that they are large enough to hold all the water
flowing down the trunk. The collars should be
sloped to one side so that the water can be collected
or measured in a tipping-bucket rain gauge.
Maintenance of the collars is very important as they
easily clog up or become appropriate resting places
for forest fauna such as slugs!
Thiessen's polygons
Thiessen was an American engineer working around
the start of the twentieth century who devised a
simple method of overcoming an uneven distri-
bution of rain gauges within a catchment (very
much the norm). Essentially Thiessen's polygons
attach a representative area to each rain gauge.
The size of the representative area (a polygon) is
based on how close each gauge is to the others
surrounding it.
MOVING FROM POINT
MEASUREMENT TO SPATIALLY
DISTRIBUTED ESTIMATION
a
1
r
1
The measurement techniques described here have
all concentrated on measuring rainfall at a precise
location (or at least over an extremely small area). In
reality the hydrologist needs to know how much
precipitation has fallen over a far larger area, usually
a catchment. To move from point measurements to
a spatially distributed estimation it is necessary
to employ some form of spatial averaging. The
spatial averaging must attempt to account for
an uneven spread of rain gauges in the catchment
and the various factors that we know influence
rainfall distribution (e.g. altitude, aspect and slope).
A simple arithmetic mean will only work where a
catchment is sampled by uniformly spaced rain
gauges and where there is no diversity in topog-
raphy. If these conditions were ever truly met then
it is unlikely that there would be more than one rain
gauge sampling the area. Hence it is very rare to use
a simple averaging technique.
a
3
a
2
r
2
r
3
a
4
a
5
r
4
r
5
r
6
a
6
Figure 2.12
Thiessen's polygons for a series of rain
gauges (
r
i
) within an imaginary catchment. The area of
each polygon is denoted as
a
i
. Locations of rain gauges
are indicated by bullet points.
