10 Hydrological Impacts of Climate Changes in Romania
331
logical regime of the Valea Cerbului River. The scenario data were obtained within
the ESEMBLE project, at Max–Planck Institute (Germany) using the ECHAM5
global climatic model. They were retrieved from the CERA database (Climate and
Environmental Retrieval and Archive) and refer to the B1, A2 and A1B scenario
families (for the period 2001–2065) and to the reference period C20 (1961–2000)
[178]. We considered a medium time horizon which is generally more appropriate to
make and implement regional and local policies and adapted management measures.
The data on air temperature and precipitation for both the reference (1961–1991)
and scenario (2001–2065) periods, were subjected to bias corrections (delta change
method, as described in [179]).
The spatial data were used as part of the input in the hydrological model and
consists of data on altimetry, stream channels, rock and soil types, land cover etc.,
retrieved from the thematic maps available at different scales (1:25,000, 1:50,000),
vectorized, then processed in GIS environment and/or with the specific tools provided
by the model’s developers.
The changes in the flow data series recorded at Bu¸ steni g.s. were investigated
based on frequency and trend analysis for the period 1961–2012. The significance of
linear trends was established using the Mann-Kendall test and Sen’s slope method,
developed on an Excel template [180].
The changes in the hydrological and climatic parameters, as projected by the
emissions scenarios, were estimated using WaSiM-ETh (Water Balance Simulation
Model), a fully distributed hydrological model that simulates all the components
of the water balance. The required input data are climatic data series (e.g., daily
data on air temperature, precipitation, sunshine duration, wind velocity, air relative
humidity) and gridded spatial data [181]. The model was calibrated and validated
using historical data series and with the PEST (Parameter Estimation and Uncertainty
Analysis) software [182]. The objective functions used to assess its efficiency were
the Nash-Sutcliffe Efficiency Coefficient (NSE = 0.80 in calibration, and 0.60 in
validation) and the Pearson Correlation Coefficient (r = 0.90 and 0.86 for calibration
and validation, respectively).
10.5.2 Observed Changes in Valea Cerbului River Flow
The streamflow changes are mainly caused by climatic and anthropogenic factors.
Because the Valea Cerbului Catchment is partially included in the Bucegi Natural
Park, the human pressures are quite limited. The intakes for the water supply of
Bus , teni town are insignificant and do not affect the flow regime [183]. Therefore,
the hydrological alteration can mainly be the result of climate changes.
A previous study [177] have shown for the period 1950–2010 a general upward
trend in the variability of monthly average discharges. Except for April and May,
increasing trends were founded in all the months, statistically significant from August
to December (with the level of significance α = 0.01 in September and October and
α = 0.05 in August, November and December). A significant decreasing trend was
detected in May (for α = 0.01). At the seasonal scale, upward trends were identified
331
logical regime of the Valea Cerbului River. The scenario data were obtained within
the ESEMBLE project, at Max–Planck Institute (Germany) using the ECHAM5
global climatic model. They were retrieved from the CERA database (Climate and
Environmental Retrieval and Archive) and refer to the B1, A2 and A1B scenario
families (for the period 2001–2065) and to the reference period C20 (1961–2000)
[178]. We considered a medium time horizon which is generally more appropriate to
make and implement regional and local policies and adapted management measures.
The data on air temperature and precipitation for both the reference (1961–1991)
and scenario (2001–2065) periods, were subjected to bias corrections (delta change
method, as described in [179]).
The spatial data were used as part of the input in the hydrological model and
consists of data on altimetry, stream channels, rock and soil types, land cover etc.,
retrieved from the thematic maps available at different scales (1:25,000, 1:50,000),
vectorized, then processed in GIS environment and/or with the specific tools provided
by the model’s developers.
The changes in the flow data series recorded at Bu¸ steni g.s. were investigated
based on frequency and trend analysis for the period 1961–2012. The significance of
linear trends was established using the Mann-Kendall test and Sen’s slope method,
developed on an Excel template [180].
The changes in the hydrological and climatic parameters, as projected by the
emissions scenarios, were estimated using WaSiM-ETh (Water Balance Simulation
Model), a fully distributed hydrological model that simulates all the components
of the water balance. The required input data are climatic data series (e.g., daily
data on air temperature, precipitation, sunshine duration, wind velocity, air relative
humidity) and gridded spatial data [181]. The model was calibrated and validated
using historical data series and with the PEST (Parameter Estimation and Uncertainty
Analysis) software [182]. The objective functions used to assess its efficiency were
the Nash-Sutcliffe Efficiency Coefficient (NSE = 0.80 in calibration, and 0.60 in
validation) and the Pearson Correlation Coefficient (r = 0.90 and 0.86 for calibration
and validation, respectively).
10.5.2 Observed Changes in Valea Cerbului River Flow
The streamflow changes are mainly caused by climatic and anthropogenic factors.
Because the Valea Cerbului Catchment is partially included in the Bucegi Natural
Park, the human pressures are quite limited. The intakes for the water supply of
Bus , teni town are insignificant and do not affect the flow regime [183]. Therefore,
the hydrological alteration can mainly be the result of climate changes.
A previous study [177] have shown for the period 1950–2010 a general upward
trend in the variability of monthly average discharges. Except for April and May,
increasing trends were founded in all the months, statistically significant from August
to December (with the level of significance α = 0.01 in September and October and
α = 0.05 in August, November and December). A significant decreasing trend was
detected in May (for α = 0.01). At the seasonal scale, upward trends were identified