Modelling Catchment-Scale Responses to Climate Change - Climate Change Impacts on Freshwater Ecosystems - page 246

To evaluate agreement between the model runs, a 'consensus' parameter was
calculated as the first axis of a principal components analysis of the predictions
for each species. The mean consensus between models was 57% (Table 10.3), but
values for each species differed considerably. It ranged from 31% for Gasterosteus
aculeatus (three-spined stickleback) to 70% for Gobio gobio (gudgeon). Generally,
the rarer species had lower consensus values. Table 10.3 also shows the sources
of variability in projections for each fish species. For the entire data set, 70% of
the variability was due to the species distribution models, 24% to the GCMs and
only 6% to the emission scenarios. However, the pattern varied somewhat
between species, and for some, such as the chub ( Leuciscus cephalus ), the choice
of GCM was a more important source.
Given the dominant role of the species distribution model in the fish species
projections, a single GCM (HadCM3) combined with a single emission scenario
(A1FI) was arbitrarily chosen to assess the potential impacts of climate change on
stream fish assemblages (see Buisson et al . 2008). These predictions are shown in
Fig. 10.1 as changes in the probability of occurrence of each species. Buisson
et al . (2008) found that all 35 fish species would be positively or negatively
affected by climate change. On average, changes in the probability of occurrence
ranged from −36.6% for Salmo trutta fario (brown trout) to +44.6% for
Leuciscus souffia (varione). Those gaining the most appear to be warm-water
species or those with a large range of thermal tolerance. Most negatively affected
were two cold-water species - Cottus gobio (bullhead) and Salmo trutta fario
(brown trout). On average, however, the 35 fish species would change their
probability of occurrence by +5.6% by 2080, indicating a slightly positive
response to climate change. Even though a species increases overall, local
extinctions in some areas might occur, compensated by colonizations of new
sites. For some species, the changes in spatial distribution were calculated and are
illustrated for three representative species in Fig. 10.2: Barbus barbus (barbel),
Esox lucius (pike) and Salmo trutta fario (brown trout).
Barbel, Barbus barbus , a rheophilic species relatively common in French
streams, was predicted to expand its range greatly under climate change. The
consensual model did not predict any extinctions and forecast that it could
colonize a large number of areas where it does not currently occur, for example,
NW France, the Pyrenees, the Massif Central and the Jura Mountains. Pike ( Esox
lucius ), a predatory species living among dense vegetation, was predicted to move
to new suitable habitats mainly in eastern France and in mountainous areas,
whereas it could suffer from local extinctions in the western part of France where
it currently occurs in many sites. Finally, brown trout Salmo trutta fario , a species
characteristic of cold and well-oxygenated streams, was predicted to be the most
severely affected by climate change. By 2080, its distribution would be restricted
to the most upstream sites of the mountainous regions and some streams of NW
France. The results thus predict that cold-water species living in headwater
streams would suffer the most deleterious effect of climate change as their
potential distributional area reduced, while other fish species occurring currently
downstream would be able to expand their range upstream.
Overall, these results were consistent with those obtained in North America,
which predicted a decrease in salmonid distribution and contrasting results for