Geoscience Reference
In-Depth Information
of discrete features (e.g. the proportion of 500 m
river lengths where trees were present). Statistics
for England and Wales were calculated in three
stages: (i) estimates of the extent or prevalence of
different features were calculated initially within
each stratum (i.e. 10 km grid square); (ii) each
stratum mean was weighted according to the 'size'
of the stratum (in this case the total river length
within the 10 km square); and (iii) the weighted
means were combined to obtain the average
for England and Wales. This produced the well-
known Horvitz-Thompson estimator (Gregoire and
Valentine, 2007). Two different display options
are illustrated for England and Wales. The first
displays the location of features, summarizing
results for each 10 km square in the familiar
form of 'dot' maps (e.g. Figures 3.4 and 3.5). The
second interpolates the raw data using ordinary
kriging, to produce a smoothed map of river
characteristics. Trans-Gaussian and indicator forms
of kriging were used as necessary to handle non-
normally distributed quantitative variables and
presence-absence data respectively (Cressie, 1993).
Both approaches to mapping present the data in
a form which non-specialists and decision-makers
can easily understand because broad patterns are
evident (e.g. Figure 3.2). They also illustrate
the value of stratifying data collection using the
Ordnance Survey grid, consistent with biological
surveillance information in the UK (e.g. Gibbons
et al
., 1993).
This chapter is limited to a descriptive overview,
so six attributes of particular relevance to river
management and conservation are described: (i)
the extent of river channel modification, because
of its ecological impact on habitat structure; (ii)
riverside tree distribution, because this influences
riparian habitat and the extent of channel shading;
(iii) the distribution of three invasive non-native
plant species, because of their impact on local
flora; (iv) the occurrence of large woody debris,
owing to its importance as a temporary habitat
and local influence on morphological processes;
(v) the extent of channel silting, as a surrogate
indicator of excessive erosion and sediment input;
and (vi) the distribution of in-channel bars, because
of
their
fluvial
morphological
and
ecological
importance.
Results
Context
The overall physical characteristics for rivers and
streams in 2007-2008 are based on data from
both the core and the minor watercourse samples
(
n
=
4849; Table 3.1). The overwhelming majority
of sites (81%) had a water width of 5.0 m or
less, confirming the predominance of streams and
small rivers in the landscape of England and Wales
(Table 3.2).
River channel modification
The HMS score represents the extent and severity
of artificial modification to the river banks and
channel (Raven
et al
., 1998a). Examples include:
bank and river-bed (channel) resectioned; bank
and channel reinforcement; culverts; bridges;
fords; weirs; dams; instream flow deflectors;
drainage outfalls; embankments; artificial two-
stage channels and damage caused by the trampling
of river banks by livestock (Environment Agency,
2003).
Based on the 4849 sample sites, it is estimated
that 11% of the length of rivers and streams
had a near-natural channel form (HMC class 1)
in 2007-2008; more than 40% was severely
modified (HMC class '5') as a result of one or
more artificial factors (Figure 3.2a). Bank and
river-bed resectioning and bank reinforcement
are major contributory factors in the HMS scores.
An estimated 43% of river length was resectioned
(Figure 3.2b) - a modification usually resulting
from channel widening and deepening to increase
flood conveyance. An estimated 8% of river
length had channel or bank reinforcement -
an activity carried out mainly to protect roads,
railways and buildings from erosion. Where it
occurred, resectioning tended to be extensive (i.e.
affecting
≥
33% of river length) within a 500 m
site, whereas reinforcement was more localized,
frequently affecting
10% of a 500 m site where