Geoscience Reference
In-Depth Information
concentrations (Tørseth & Semb 1997). The main threats to aquatic ecosystems
in the catchment are acidification and nitrogen enrichment, with substantial
leaching of atmospheric-derived N from soils to surface waters already happening.
The question addressed in this study was whether climate change was likely to
alter the situation so that more (or less) nitrogen leached into the water, and
whether nitrate concentrations in the rivers, lakes and receiving coastal waters
were likely to increase or decrease and, if so, by how much.
Data from two GCMs - ECHAM4 from the Max Planck Institute, Germany,
and HadAm3H from the Hadley Centre, the United Kingdom, driven with two
scenarios of greenhouse gas emissions (IS92a and A2, respectively) - were
dynamically downscaled (see above) to a spatial resolution of 55km 2 × 55 km 2
and a temporal resolution of 6h using a model related to that used to perform
weather forecasting in Norway. The downscaled predictions for control periods
were adjusted to match observed data in the catchment before being used to
make climate change prognoses. These were for different periods: 2030-49 for
the MPI IS92a scenario and 2071-2100 for the Hadley A2 scenario (hereafter
called 'MPI' and 'Had', respectively). Predicted changes in nitrogen deposition
due to presently agreed legislation were also included in the forecast scenarios.
The climate change data were then used to drive four models linked to assess the
effects on hydrology and nitrogen concentrations and fluxes in the river and its
coastal fjord (Kaste et al . 2006; Fig. 10.4). These models were the hydrological
model HBV (Sælthun 1996), the water quality models MAGIC (Cosby et al .
2001) and INCA-N (Whitehead et al . 1998a, b; Wade et al . 2002a) and the NIVA
Fjord models (Bjerkeng 1994). The flow of data between models is shown in Fig.
10.4. HBV was used to predict flows and other hydrological variables needed to
run INCA, MAGIC to calculate N retention in INCA, INCA to calculate stream
nitrate concentrations and the Fjord model to assess the biological consequences
in coastal waters. The HBV, INCA and MAGIC models were calibrated initially
to small sub-catchments within the main catchment area, before extending the
calibration to the whole Bjerkreim basin. After calibration, the models were
tested against observed data from a control period, in which they performed
acceptably, though the means were simulated better than the seasonal patterns,
and extreme events and unusual meteorological conditions were more difficult
still to simulate (Kaste et al . 2006).
The two downscaled climate scenarios projected a temperature increase in the
study region of about 1 °C with MPI and about 3 °C with Had, and both predicted
increased winter precipitation. Projections of summer and autumn precipitation
were quite different between the two models, however: a slight increase with
MPI and a significant decrease with Had. Because different models were used for
different periods, it was not possible to determine whether this was a model
effect or a function of time. Because of higher winter temperatures, the HBV
model predicted a dramatic reduction of snow accumulation in the upper parts
of the catchment for both the climate change scenarios. This, in turn, led to
higher run-off during winter and lower run-off during spring snowmelt. With the
Had scenario, run-off in summer and early autumn is substantially reduced as a
result of reduced precipitation, increased temperatures and, thereby, increased
evapotranspiration. MAGIC and INCA models predicted no major changes in
Search WWH ::




Custom Search