Geoscience Reference
In-Depth Information
(equivalent of an r
2
of
0.09 in a rank correlation
with nutrient gradients).
Assessments of the vegetation structure and
functional types provide an alternative to similarity
measures or indicator species indices. The Riparian
Vegetation Index (RVI) (Aguiar
et al
., 2009)
developed for Portugal includes metrics such
as riparian and aquatic species richness, exotic
species richness, nitrophyllous species richness and
riparian integrity. The RVI correlated well (r
2
condition is defined for a site it can be kept
indefinitely. The current approach to developing
reference conditions in the WFD (Wallin
et al
.,
2003) implies that a general (or modelled)
reference condition for a river type can be applied
to a single unique site. This increases the apparent
natural variation and may make small deviations
from a near-natural state harder to detect.
Improving the performance of Mediterranean
macrophyte indices could also be achieved by
increasing the number of aquatic species in the
index, either by expanding the non-vascular
species, especially macroalgae, whose taxonomy
and ecological requirements have yet to be resolved
in Iberia, or by including hygrophyte species found
in the dry river channel. These species indicate
nutrient regime and hydrology in temporary
Mediterranean rivers, but they are usually confined
to the banks of temperate rivers. Plates 15 and 16
show two rivers that have intermittent flow, with
hygrophytes and terrestrial species established in
the channel.
Another way of increasing species numbers is
to combine the macrophytes and diatoms into a
single metric. This combination is permitted in
the WFD, because the biological quality element
to be assessed is defined as 'macrophytes and
phytobenthos' but it is not explicit as to how and at
what stage within bioassessment this combination
should take place. Nonetheless, diatoms are part
of the aquatic flora, with an important role as
bioindicators (Kelly
et al
., 2009).
=
0.56) with a combined pressure gradient in the
south of Portugal, but in the north it performed
less well (r
2
0.31) than other metrics. However,
the RVI was rejected for the WFD because it
used riparian plant species and riparian structure
together with aquatic species.
Thus, all the methods previously attempted have
proved either unsuitable for WFD application,
or performed poorly in the Mediterranean rivers
characteristic of the south of Portugal and Spain.
=
Possible solutions
There are several methods that could be employed
to resolve these problems. First, the definition of
reference conditions at a site could be improved.
Small-scale physical habitat variation is likely to
be a large source of variation for macrophytes
within the WFD, particularly since the comparison
is not directly between two proximate sections
of the same river. However, current modelling
methods are not suitable for predicting the
macrophytes at the habitat scale. Instead, all the
available information could be used to produce
a site-specific reference condition using expert
judgement. The model for predicting the reference
condition could be seen as a first phase, after which
site-specific information, including sediment cores
(Seddon
et al
., this volume) underlying geology
(e.g. an acidic river running over alkaline rocks),
and biogeography, could be used to improve the
accuracy of the predicted reference condition and
allow the uniqueness of the river in its natural
state to be described. Since the same monitoring
networks tend to be retained in subsequent
monitoring periods, once an accurate reference
A combined diatom-
macrophyte index
Within the 373 sites surveyed in Portugal, 105
aquatic macrophyte species and 281 diatom species
were found, giving a total of 386 species. Weighted
Average Partial Least Squared (WA-PLS) (ter
Braak and Juggins, 1993), a more accurate form
of Weighted Averaging (ter Braak and Looman,
1986), was applied to these species along a pressure
gradient formed from the combination of nitrate
and BOD. The species were thus scored to represent
their optimum along the pressure gradient. The
