Climate Change: Defining Reference Conditions and Restoring Freshwater Ecosystems - Climate Change Impacts on Freshwater Ecosystems - page 209

were not constrained by the use of a limited number of a priori selected variables.
Realization that not only environmental factors but also biological communities
are best described as gradients and not discrete entities has also been proposed by
Hawkins and Vinson (2000). These authors argued that since assemblages
generally vary continuously along environmental gradients, the classification
methods that seek to place sites into discrete categories are fundamentally limited
compared with approaches that recognize biological continua.
In conclusion, there are now indications that classifications based solely on
large-scale (e.g. landscape-level) predictors are not sufficient to capture the fine-
scale variability of aquatic communities. In other words, fixed typologies, such as
system-A classification, may provide a useful framework for setting ecological
targets, but need to be augmented by the use of more site-specific predictors such
as in-stream water chemistry and substratum. A predictive approach using
continuous environmental and taxonomic data is a robust method for establishing
reference conditions and one that should be considered. Lastly, regardless of the
method used to determine the reference condition, assessments of ecological
quality are of limited value without knowledge of their precision and confidence
and these need to be quantified.
Historical information and palaeo-records
For some aquatic system types, it is now difficult to find good examples of
reference sites and this is especially true of European shallow lowland lakes,
many of which have been long disturbed (e.g. Bradshaw & Rasmussen 2004;
Leira et al . 2006) and particularly during the 20th century (e.g. Bennion et al .
2004). In such cases, it is necessary to use historical information to determine the
reference conditions. Unfortunately, long-term data sets are rare for most
ecosystems, and where they do exist, monitoring programmes tend to have begun
after disturbance has occurred. However, for lakes, palaeoecological techniques
can be employed, whereby the remains of aquatic organisms preserved in sediment
cores are examined to provide historical information and hence to define
reference conditions (Bennion & Battarbee 2007).
In a few rare instances, where historical data sets are of sufficient length,
contemporary time series and palaeolimnological methods can be used in
combination to describe the past state with increased precision. One such example
is provided by Groby Pool, a small (0.12 km 2 ) shallow (mean depth < 1.1 m) lake
in England that has experienced nutrient enrichment over at least the last two
centuries (Sayer et al . 1999; Davidson et al . 2005). This lake has attracted numerous
botanists since the mid-1700s, and an extensive botanical record has been collated
from several sources including four county floras (1850, 1886, 1933 and 1988),
UK museum herbaria, journal articles, notebook entries, unpublished manuscripts/
sketches, academic theses and various reports of conservation bodies and local
interest groups. A second example is Loch Leven, a key study site in Euro-limpacs,
which is a significantly larger (13.3 km 2 ) but relatively shallow (mean depth 3.9 m)
lowland lake in Scotland. The site underwent early nutrient enrichment in the
mid-1800s associated with the lowering of the lake, followed by a second
eutrophication phase during the mid-1900s when phosphorus-rich effluent from
a large woollen mill discharged to the loch (Carvalho & Kirika 2003). The lake