This is not a peer-reviewed article.

21st Century Watershed Technology: Improving Water Quality and Environment,

Proceedings of the 29 March - 3 April 2008 Conference,
Concepcion, Chile, Publication date 29 March 2008
ASABE Publication Number 701P0208cd.
Eds. E.W. Tollner and A. Saleh

Modelling hydrological response to climate change;

experiences from two south – central Chilean Watersheds
Alejandra Stehr1, Patrick Debels1,2, Jose Luis Arumi3
1 Centre for Environmental Sciences EULA-CHILE, University of Concepción, PO Box 160-C,
Concepción, Chile
2 Centre for Environmental Studies CIEMA-UNI, National Engineering University, PO Box 5595,
Managua, Nicaragua
3 Department of Water Resources, Facultad de Ingeniería Agrícola, University of Concepción, Avenida
Vicente Méndez 595, Chillán, Chile.
Abstract. Over the last decades a progressive change in world-wide climatic conditions has been
observed and described by the scientific community. Associated to these ongoing changes, reports on the
devastating effects of floods and prolonged periods of droughts are increasingly brought to the attention of
the public opinion. Also, for considerable parts of the Chilean territory, important changes in climatic
conditions have been projected for the decades to come. This situation has encouraged society to take
measures in order to attempt to control such changes, as well as to adapt to those changes that cannot
eagerly be prevented through world-wide mitigation actions. From a water resources management
perspective, the implementation of specific measures at the level of the river basin asks for the availability
of tools that allow managers to take decisions in an informed way. Considering that the Biobío Basin
(24.371 km2) is of high strategic importance for Chilean development, advances in the general
understanding and capacity to describe and predict, in a spatially explicit way, the impact of climate and
climate change on (sub-)basin hydrology is urgently needed. The Soil and Water Assessment Tool (SWAT)
was chosen to model the hydrology of two sub-basins of the Biobío, the Vergara Basin (4.265 km2), located
in the south-eastern part of Biobío, and the Lonquimay Basin (455 km2), located within the Andean
Mountain Range. First, the model was calibrated and validated for both sub-basins, and then, the response of
basin hydrology to different plausible climate change scenarios was modelled. Even when results indicate a
wide array of potential impacts in terms of the magnitude of a future reduction of monthly mean discharges,
such reductions appear to be highly plausible as future conditions of river discharge. In the event of such
reductions, important adaptation measures will be required.
Keywords. Climate change, hydrological modelling, SWAT, Chile, Biobío.
Introduction
During the last five decades the worldwide water consume has increased remarkably, due to the
explosive world population growth, a non preceded improvement in living standards in several places, as
well as an enormous economic development (Xu & Singh, 2004). Associated to this, serious environmental,
political and social issues have arisen in many regions, preventing a sustainable socioeconomic
development. At this context, climatic change threatens with turning the situation even more complex, since
it can strongly disturb water resources availability in medium and long terms, due to the concomitant
changes in precipitation distribution and temperature rising. Among the most important physical changes
expected to be happen due to climatic change are a growth in the frequency and severity of extreme climatic
events, such as warmth waves and droughts, added to an increase in precipitations intensity (Stott et al.,
2004) There has also been announced important changes for flow timing and water volume of rivers at
different study areas (Whitfield et al., 2002). At basins that have a relevant snow contribution at present, the
temperatures increase will lead to an increase in the precipitation/snow ratio during the cold months (Gleick,
1986; López-Moreno & Nogués-Bravo, 2005).
For Chile, Hulme & Sheard (1999) expect a precipitation decrease from 6% to 18% by 2080 (compared
with 1961-1990 period; scenarios B1 low A1 high) as a consequence of climatic change. Specifically, the
Chile Climatic Variability Study for the XXI Century (CONAMA-DGF, 2006) estimates that temperature
for the 2071 - 2100 period will be dominated by rising temperature at all Biobío basin sectors. Seasonal
warming would be higher in summer, exceeding 5ºC at some mountains spots. Also, a decrease in
precipitations is expected, being higher at the mountains than at seaside.
In order to manage efficiently the water resources it is necessary an appropriate understanding of the
precipitation – runoff processes, in mid and long terms, as well as its impact in the basin’s hydrological
balance. In this way, the development and application of a hydrological model can be of great importance
(Yang et al., 2005). In order to consider climatic change, the classical flow study is substituted by
hydrological models based on a physical description of the most relevant processes, which relate the
climatic variables with the flows (Dankers & Christensen, 2005; Vandenberghe et al., 2005). It is worth
mentioning that using hydrological models is a rather low cost way to evaluate the short and long term
effects of human activities on basin’s hydrology and to make decisions on basin’s management, provided
that there is a minimum amount of data available to guarantee a correct model performance. Low cost tools
availability to assist decision making is especially important in developing countries, where resources for
this effects are typically scarce.
The objective of this work is to model the hydrological response to climate change at two sub-basins of
the Biobío River. SWAT model (Soil & Water Assessment Tool; (Arnold et al., 1998: Neitsch et al., 2002ª
2002b)), which is a physically based and spatially distributed hydrological model, was used to assess historic
and future scenarios at both basins..
Study Area
The Biobío basin (24.371 km2) is the third largest Chilean basin and it covers about 3% of the national
territory. It is placed at south central Chile (36º45’ – 38º49’ S and 71º00’ – 73º20’W). The basin extends
from the Andes mountains at the East, to the Pacific Ocean at the West, and is influenced by the southern
mild climate as well as central Chile Mediterranean’s climate. This basin constitutes the most important
country’s forestry center, holds a big part of the country’s agricultural soil and also plays a remarkable role
in the country’s electricity (hydrologic) supply. This paper focuses on two sub-basins of Biobío; the first one
(Lonquimay) placed at the Andes mountains sector, and the second one (Vergara) placed at the Central
Valley and foothills area (Figure 1). The Lonquimay River basin covers an area of 455 km2, has a
pluvionival hydrological regime, with maximum and minimum flows during June and March, respectively,
and a snow peak flow in October. The Vergara River basin covers an area of 4.265 km2, has a
predominantly pluvial regime, with maximum and minimum flows occurring during July and February,
respectively.

Figure 1. Lccation of Vergara and Lonquimay River basins.

Methodology
The development of this research was carried out in four stages:
1. Calibration and validation of the hydrological model using observed time series of meteorology and
flows.
2. Climatic scenario generation for the zone on study.
3. Basin’s hydrological characteristics simulation using perturbed climatic series (scenarios) and the
calibrated hydrological model.
4. Comparison of actual with future (considering climate change) basin’s hydrology .
Following, the outstanding components for stages one and two will be briefly explain. Afterwards the
results will be presented.

The SWAT model

The SWAT (Arnold et al., 1998, Di Luzio et al., 2002) model was used to accomplish the hydrological
modeling at both selected sub-basins. It was developed during the 90’s at the United States Agricultural
Department and due to its flexibility has been used at many places throughout the world (USA, Europe,
India, New Zealand, Chile, etc; (Abu El-Nasr et al., 2005; Eckhardt 2005; Govender & Everson 2005; Cao
et al., 2006; Tripathi et al., 2006; Stehr et al., in press). Within the SWAT conceptual framework, the
representation of the hydrology of a basin is divided into two major parts: the land phase of the hydrological
cycle; and the routing of runoff through the river network. The contribution to the main flow on each sub-
basin is controlled by the land phase, while the routing phase determines the water movement through the
river network up to the different control points and the basin outlet. A complete description of the SWAT
model can be found in Neitsch et al. (2002a).
This model accepts changes in precipitation and temperature factors, allowing the simulation of potential
climatic scenarios. In order to do so, a simple methodology is provided to adjust precipitation and
temperature.
⎛ adjust pcp ⎞
Pday = Pday ∗ ⎜⎜1 + ⎟ (1)
⎝ 100 ⎟⎠
Tmax = Tmax + adjust tmp Tmin = Tmin + adjust tmp (2)
Where Pday is the precipitation over a sub-basin for a given day, adjustpcp is the expected change in
precipitation (as percentage), Tmax and Tmin are the maximum and minimum daily temperature, respectively,
and adjusttmp is the expected temperature change. The adjustment terms can be modified monthly in order to
simulate seasonal differences on climatic changes conditions.

Climatic change scenarios

The scenarios were generated using:
1. MAGGIC/SCENGEN (Model for the assessment of Greenhouse-gas Induce Climate Change/Scenario
Generator): Software that considers the greenhouse-gas, aerosols and sulphuric dioxide emissions as
input and estimates average global temperatures, sea level rise and regional climate as output. The latest
version (v4.1; Wigley 2003a, Wigley 2003b) handles all the emission scenarios (SRES) and allows the
research on variability changes and to make probabilistic projections. SCENGEN is a regionalization
algorithm that uses a scaling method to deliver climate output and changes information interpolated at a
5º x 5º grid.
2. The Climatic Variability Study in Chile for de XXI Century (CONAMA-DGF (2006)): It used
HadCM3/HadAM3 models, with HadAM3 forcing the regional simulations generated with the PRECIS
model, with a 25 km spacial resolution. The model output variables (average, maximum and minimum
temperature and precipitation) are available for each one of the geographic dominium points (18ºS –
57ªS and 62ªO – 85ªO). We took the predictions of precipitation, temperature and expected changes for
the end of the XXI century for the study area. Thus, 3 daily 30 years long time series are available
(actual climate and 2 (A2 and B2 scenarios) for the future climate (2070-2100)) for each grid point and
variable.
Within these study 44 future climate scenarios were used, 2 scenarios from local RCM and 42 scenarios
generated by MAGIC/SCENGEN (7 GCMs*6 SRES) including greenhouse-gas and aerosols, changed were
modeled for a 30 years’ time window, centered in 2085, considering scenarios of annual and monthly
changes. When disturbing SWAT with the obtained changes, the special and temporary precipitation
distribution is assumed as remaining constant and that changes are only related with total amount
(precipitation)/absolute values (temperature) as a first approach. The base line was taken from RCM
CONAMA – DGF (1961-1990). This was chosen in order to compare the differences in results obtained
from perturbing a hydrological model with variations coming from GCM versus the series obtained from a
local RCM.
Results and Analysis
Calibration and Validation of the SWAT model
 The 90 m-resolution topography data from the Shuttle Radar Topography Mission (SRTM DEM, final
version) was used as a basis for the construction of the model application. Available national data was used
to characterize land use and soil type. The meteorological data (precipitation and temperature) were
obtained from the National Water Databank of the Chilean General Water Directorate. A major description
about the data used can be found in Stehr et al. (accepted). Before calibration, a sensitivity analysis was
made on both basins (Van Griensven et al., 2006) in order to determine the parameters to which the model
results are most sensitive, after this taking the eight most sensible parameters an automated calibration
procedure implemented in SWAT2005 called PARASOL (Parameter Solution Method; Van Griensven &
Bauwens, 2003) was used.
Calibration and validation were made at monthly level and for model’s behavior evaluation; three
different statistic indicators were used (Nash-Stucliffe (EF) efficiency, PBIAS and R2). Moreover, the
agreement between observed versus modeled graphics was also evaluated. A model is considered to have
better behavior when EF and R2 are closer to 1, PBIAS closer to 0, and graphics show correspondence
between modeled and observed flows. These results show a good model behavior for both basins (Table 1)
during calibration, and was satisfactory for Lonquimay and good for Vergara during the validation stage.
Also, a good correspondence between modeled and observed flows, except for some high flows, can be seen
in Figure 2.

Figure 2. Calibration and validation (monthly results) for (a) Vergara Basin (b) Lonquimay Basin.

Table 1 Statistical indicators of model performance (monthly output)

Vergara* Lonquimay
Calibration Validation Calibration Validation
EF 0.93 0.93 0.81 0.56
PBIAS 11.78 2.77 0.87 0.57
R2 0.96 0.93 4.88 7.86
* This parameters were obtained at the DGA station Vergara at Tijeral.
Results from scenario generation
Figure 3a shows the results obtained (annual variations) from applying MAGICC/SCENGEN. In the
study area (35º-40ºS, 70º - 75ºW), the model forecasts an increase in temperatures and a decrease in
precipitation for all six marker scenarios combined with 7 GCM, with respect to the referential period,
although the change magnitudes differ remarkably for the different scenarios originated from different
GCM. With data series obtained from RCM CONAMA – DGF (2006), the average annual change variation
was estimated for a 30 years period (2071 – 2100), both for the precipitation (variation %) and temperature
(magnitude change), in order to compare if the magnitude of the predicted changes correspond to those
estimated by MAGICC/SCENGEN (Figure 3b). It can be seen that the changes predicted for Chile by RCM
are larger than the changes predicted by MAGICC/SCENGEN, both for A2 and B2 scenarios considering
the same starting GCM, but all of them are placed at the second quadrant (-/+) being inside the envelope
made by the 42 scenarios. On the other side, most of the scenarios predicts a greater precipitation decreases
than those expected by Hulme and Sheard (1999) (6 – 18% up to 2080 (scenarios B1 low A1 high)).
 Figure 3. a) MAGICC/SCENGEN mean output for a 30-years time slice centred around 2085; output is
given in terms of deviations of temperature (∆T) and precipitation (%) with respect to the 1961-1990
reference period for the 5º x 5º grid cell covering the study area (35º-40ºS, 70º-75ºW); greenhouse-gas +
aerosols. b) Mean annual temperature and precipitation variations starting from RCM accomplished by
CONAMA – DGF, 2071 – 2100 period; Vergara River basin sector.
From the Climatic Variability Study in Chile for the XXI Century (CONAMA – DGF 2006), there is an
estimation of average temperatures increase and a precipitationprecipitation decrease (Table 2) in the period
2071 – 2100, for both selected sub-basins. precipitation

Table 2. Mean annual change expected for temperatures and precipitations at the Vergara and
Lonquimay basins (2071 – 2100).
Vergara Lonquimay
A2 B2 A2 B2
%∆PP ∆T (ºC) %∆PP ∆T (ºC) %∆PP ∆T (ºC) %∆PP ∆T (ºC)
Anual -33.5 2.9 -21.9 2.0 -27.7 3.5 -15.4 2.4

Results from the scenario modelling

Figure 4 shows that SWAT model predicts an almost lineal relationship (R2 = 0.93 on both basins)
between the mean annual average flow change and the mean annual precipitation change for the 30 years
period modeled. For the Vergara basin it can be seen that a change at the average percentage of annual
precipitations makes a greater change than at the Lonquimay River basin’s flow. This is an expected result,
considering the change factor methodology used for the scenario generation combined with the actual
hydrological model configuration.

Figure 4. Annual average precipitation change versus flow for Lonquimay (blue) and Vergara (red)
basins
 Figure 5. Flow duration curve for: base scenario; generated scenarios thru MAGICC/SCENGEN
(average, most favorable situation (Max) and most unfavorable (Min)) and two RCM scenarios for June and
November months. (a)Vergara (b)Lonquimay
Comparing both basins (Figure 5) it can be appreciated that for the month of June at Vergara River
basin, there is a difference between the envelope form by the flow duration obtained from the RCM A2 and
B2 scenario and the base flow, which is not produced at the Lonquimay, where the flow duration curves
generated for flows obtained from RCM A2 and B2 scenario and base are almost coincidental.
Even during June, both basins have a possible flow increase prediction, which does not occur in
November, where all scenarios show a flow decrease, especially at Lonquimay basin where this decrease is
greater due to a decrease of the melted snow, as shown at Figure 6.

Figure 6. Monthly averages for modeled 30 years, flow, precipitation, snowfall and snow melt for (a)
base scenario (b) scenario obtained from RCM CONAMA –DGF (2006).

At the Lonquimay basin, a decrease in flow with respect to base scenario can be assessed during spring
and summer months at all scenarios, which is closely related with the decrease of winter snowfall (Table 3),
which decrease at all scenarios. This makes the maximum flows to decrease and delay two or three months
due to snow melting. This is of great relevance since this basin is located at the mountains sector, upstream
of two important hydroelectric reservoirs (13% of national established capacity at 2007). Figure 6 shows
actual situation versus predicted one by RCM CONAMA- DGF (2006), which assesses a precipitation
decrease and shows that the difference between precipitation and snowfall increases at the A2 scenario
(most extreme). It also shows that the winter flow stays within range, but there is a curve lowering due to the
decrease in the melting peak.

Table 3. Precipitation and snowfall ratio during cold months, base scenario, most favorable and most
unfavorable scenario
Month Base Favorable Unfavorable
June 1.72 2.27 1.87
July 1.26 1.86 2.28
August 1.15 1.62 1.82
Discussion
The available actual results regarding the hydrological impact on the Vergara and Lonquimay basins,
obtained from a previously calibrated and validated hydrological model and produced on a series of possible
climate change scenarios, show a wide range of potential impacts in terms of monthly average flow
decreases. It has to be remark that observed variations are larger during the spring and summer months. The
wide range of output differences obtained by the simulated scenarios is a consequence of the great
uncertainty that actually persists regarding the way that future greenhouse-gases will take place, as well as
which is the world and regional climatic system’s answer going to be facing these emission scenarios. This
uncertainty is shown by the great number of GCMs that have been used, as well as for the emission
scenarios modeled. This scenarios diversity may cause divergences/convergences with different models for
different regions worldwide (Ruosteenoja et al., 2003), as well as at regional/local level. In the Biobio basin
case, a major convergence was observed regarding the trend of changes, but holding significant differences
in the magnitude they are expressed.
Based on SWAT results, the hydrology of Vergara River basin seems to be sensible to rainfall changes,
moreover the percentage flow decrease is typically higher than the corresponding precipitation percentage
decrease. The Lonquimay basin as well, is highly sensible to precipitation and temperature changes, since
the increase of these mean less snow and confirms what Gleick (1986) and López-Moreno & Nogués-Bravo
(2005) stated when they mention that there will be an increase in the precipitation/snow ratio during winter
months and a decrease of the snow season.
Another uncertainty factor is related to the model’s performance. The uncertainty on the results may be
somewhat lowered with an increase on the quality of the input data used for modeling (precipitation and
temperature time series, as well as their special variability). For instance, in the Lonquimay basin would be
very useful to install meteorological stations placed higher than 2000 masl. They could generate input series
regarding liquid and solid precipitation as well as density and height of the snow pack and other
meteorological variables (wind, radiation, temperature, etc.). Previous statement takes more relevance if we
consider that the hydroelectric generation reservoirs are supplied mainly by waters coming from the higher
zone mountains, where the accumulation and melting contribute to the intra-annual flow variation
regulation. Local prediction capacity advances and the evaluation of impacts of possible scenario changes
regarding availability and distribution of water resources provides real capacity of having a better future
resources managing, and contribute in that way to a better use of it, as well as to avoid and prevent conflicts
among users.

Conclusion
Applying a plausible climatic change scenarios series at sub-basins Lonquimay and Vergara by means of
a previously calibrated and validated hydrological model shows that considerable average monthly flow
decrease are part of a highly probable scenario for the XXI century, for the Chilean (South-)Central zone
basins and sub-basins. Under such conditions, adaptation measures for resources managing would be
necessary.
The obtained results for 2 sub-basins modeled (20% of Biobio basin total area) allow to state a first
qualitative interpretation about the potential impacts of climate changes on the availability of water
resources in the basin. This shows that the climatic change may be added as an additional stress factor inside
the Biobio basin. When the combined impacts of flow decrease due to climatic change, water transferences
to adjacent basins, rises of human water consumption patterns, increase in basin’s industrial activity and
potential impacts of changes in soil use within it make a contrast with increasing environmental
consciousness (for instance, new rules on water quality for ecosystem’s protection) the need of taking
measures to mitigate comes clear, such as basin’s water use rationalization and a consensuated decision
making among the different stakeholders involved in development and future water use scenarios. In this
aspect, Chilean government has advanced the implementation of a national basins strategy which constitutes
a first step with potential to contribute to the optimization of the use of water resources during next decades.
 Acknowledgements. The present research was conducted in the framework of the TWINBAS and
TWINLATIN projects, which were both financed by the European Community through its Sixth Framework
Programme for Research and Technological Development (Priority Area “Global Change and Ecosystems”;
Contract N° 505287 and Contract N° 018436). The authors wish to express their sincere gratitude to the
Chilean General Water Directorate DGA, in particular to the staff of the Biobío Division; the Chilean
Meteorological Directorate (DMC), as well as to all other data providers that have not been explicitly
mentioned above.

References
Abu El-Nasr, A., Arnold, J. G., Feyen, J. & Berlamont, J. (2005) Modelling the hydrology of a
catchment using a distributed and a semi-distributed model. Hydrol. Processes 19, 573-587.
Arnold, J. G., Srinivasan, R., Muttiah, R. S. & Williams, J. R. (1998) Large area hydrologic modeling
and assessment - Part I: model development. JAWRA 34(1), 73-89.
Cao, W., Bowden, W. B., Davie, T. & Fenemor, A. (2006) Multi-variable and multi-site calibration and
validation of SWAT in a large mountainous catchment with high spatial variability. Hydrol. Processes 20,
1057–1073.
CONAMA-DGF (2006). Estudio de la variabilidad climática en Chile para el siglo XXI, Comisión
Nacional del Medio Ambiente. Departamento de Geofísica. Facultad de Ciencias. Físicas y Matemáticas.
Universidad de Chile.
Dankers, R. & Christensen, O. B. (2005) Climate change impact on snowcoverage, evaporation and river
discharge in the sub-arctic tana basin, Northern Fennoscandia. Clim Change 69, 367 – 392.
Di Luzio, M., Srinivasan, R., Arnold, J. C. & Neitsch, S. L. (2002). ArcView Interface for SWAT2000
User’s Guide. Texas, Texas Water Resources Institute, College Station
Eckhardt, K. (2005) How to construct recursive digital filters for baseflow separation. Hydrol. Processes
19, 507-515.
Gleick, P. H. (1986) Methods for evaluating the regional hydrologic impacts of global climatic changes.
J. Hydrol 88, 97-116.
Govender, M. & Everson, C. S. (2005) Modelling streamflow from two small South African
experimental catchments using the SWAT model. Hydrol. Processes 19, 683-692.
Hulme, M. & Sheard, N. (1999) Climate change Scenarios for Argentina. Climate change Scenarios for
Argentina. Norwich, UK: 6pp.
IPCC (2001). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change [Houghton, J.T.,Y. Ding, D.J.
Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A. Johnson (eds.)], Cambridge
University Press, Cambridge, United Kingdom and New York, NY, USA, 881pp.
López-Moreno, J. I. & Nogués-Bravo, D. (2005) A generalized additive model for the spatial
distribution of snowpack in the Spanish Pyrenees. Hydro Process 19(16), 3167-3176.
Neitsch, S.L., Arnold, J.C., Kiniry, J.R., Williams, J.R. & King, K.W. (2002a). Soil and Water
Assessment Tool Theoretical Documentation. Version 2000. Texas, Texas Water Resources Institute,
College Station.
Neitsch, S.L., Arnold, J.C., Kiniry, J.R., Williams, J.R. & King, K.W. (2002b) Soil and Water
Assessment Tool User's Manual. Version 2000. Texas, Texas Water Resources Institute, College Station.
Nurmohamed, R., Naipal, S. & De Smedt, F. (2007) Modeling hydrological response of the Upper
Suriname river basin to climate change. J Spatial Hydrol 7(1), 1 - 22.
Stehr, A., Debels, P., Romero, F. & Alcayaga, H. (in press) Hydrological modelling with SWAT under
limited conditions of data availability: evaluation of results from a Chilean case study. Hydrolog Sci J.
Stott, P. A., Stone, D. A. & Allen, M. R. (2004) Human contribution to the European heatwave of 2003.
Nature 432, 610-614.
Tripathi, M. P., Raghuwanshi, N. S. & Rao, G. P. (2006) Effect of watershed subdivision on simulation
of waterbalance components. Hydrol. Processes 20, 1137–1156.
Van Griensven, A. & Bauwens, W. (2003) Multiobjective autocalibration for semidistributed water
quality models. Water Resour. Res. 39(12), 1348.
Van Griensven, A., Meixner, T., Grunwald, S., Bishop, T. & Sirinivasan, R. (2006) A global sensitivity
analysis tool for the parameters of multi-variable catchment models. J. Hydrol. 324, 10-23.
Vandenberghe, V., Van Griensven, A., Bauwens, W. & Vanrolleghem, P. A. (2005) Propagation of
uncertainty in diffuse pollution into water quality predictions: application to the River Dender in Flanders,
Belgium. Water Sci. Technol. 51, 347-354.
 Whitfield, P., Reynolds, C. J. & Cannon, A. J. (2002) Modelling streamflow in present and future
climates: Examples from the Georgia Basin, British Columbia. Can. Water Resour. J. 27, 427-456.
Wigley, T. M. L. (2003a) MAGICC/SCENGEN 4.1: Technical Manual, National Center for
Atmospheric Research, Colorado, USA, October 2003, 14 pp. MAGICC/SCENGEN 4.1: Technical Manual,
National Center for Atmospheric Research, Colorado, USA, October 2003, 14 pp.
Wigley, T. M. L. (2003b) MAGICC/SCENGEN 4.1: User Manual, National Center for Atmospheric
Research, Colorado, USA, October 2003, 23 pp. MAGICC/SCENGEN 4.1: User Manual, National Center
for Atmospheric Research, Colorado, USA, October 2003, 23 pp.
Xu, C.-Y. & Singh, V. P. (2004) Review on Regional Water Resources Assessment Models under
Stationary and Changing Climate. Water Resour Manag 18, 591 -612.
Yang, T.-C., Yu, P.-S. & Chen, C.-C. (2005) Long-term runoff forecasting by combining hydrological
models and meteorological records. Hydrol. Processes 19, 1967-1981.