Interaction of Climate Change and Eutrophication - Climate Change Impacts on Freshwater Ecosystems - page 135

and may reflect conditions in ponds rather than lakes (though ponds and other small
water bodies dominate the world's area of freshwaters and have been unjustifiably
derogated). The same arguments apply to short self-contained experimental streams,
though warming experiments in such systems have yet to be attempted. Given the
open nature of streams with their unidirectional flow and often extensive associated
wetlands, experiments involving artificial heating are very difficult and expensive
to perform (Hogg & Williams 1996), with substantial constraints on both
experimental scale and replication. A compromise is to use natural settings that
provide the contrasted conditions required for testing a given hypothesis. For
example, to conduct experiments on temperature effects, areas of geothermal
activity provide excellent opportunities for adopting this paired-site approach.
Thus, we took advantage in Euro-limpacs of adjacent streams and wetlands of
different temperature regimes created in the actively volcanic landscapes of
Iceland.
Stream studies
There is a wealth of information on stream macro-invertebrates in relation to
temperature (e.g. Ward & Stanford 1982; Ward 1992), whether effects of
temperature are direct or mediated by changed hydrological regimes of streams
(see Chapter 4). Recent analyses of long-term data sets have also suggested major
effects of temperature on the species distributions of invertebrates and on local
community structure (Daufresne et al . 2003; Mouton & Daufresne 2006;
Burgmer et al . 2007). Furthermore, predictions of community change in response
to rising temperatures have been supported by recent invasions of terrestrial
species from lower latitudes and altitudes coinciding with global warming
(Parmesan et al . 1999; Hickling et al . 2006).
The streams at Hengill in Iceland are almost ideal model systems to investigate
the foundations of such field observations, as they provide a relevant temperature
gradient that is largely unconfounded by other environmental variables (Friberg
et al . 2009).
The Hengill geothermal area (64°03'N, 21°18'W) of south-western Iceland
(Fig. 6.9) is one of the most extensive geothermal fields on the island, with a total
coverage of 110km 2 (Saemundsson 1967; Gunnlaugsson & Gíslason 2005).
Precipitation, which seeps through the porous volcanic bedrock, is among the
highest in Iceland and feeds the numerous streams emerging from the mountain
ridges at an elevation of 300-500 m a.s.l. The study site where the paired-stream
experiments were carried out is located in the valleys of Miðdalur and Innstidalur
at 350-420 m a.s.l. These valleys have lush grassland, wetland patches and streams
with a range in temperature from 6 to 42°C. The streams are primarily
groundwater fed with little or no influence from volcanic gases. As a result, pH
is comparable among streams differing in water temperature, as are other chemical
variables which could have a negative effect on the biota. We investigated 10
streams which drained individual sub-catchments and experienced a summer
temperature range of 7-23 °C (Fig. 6.10).
In the streams investigated, structural and functional attributes of macro-
invertebrates responded to temperature changes in similar ways as observed
elsewhere along nutrient enrichment gradients (e.g. Pascoal et al . 2003;