Sin ttulo de diapositiva

1
A FIRST APPROACH MODELING IBERÁ WETLANDS USING
VIC HYDROLOGIC MODEL Ricardo Vidal 1 - Inés
Camilloni 2 1 Departamento de Ciencias de la
Atmósfera y los Océanos - Facultad de Ciencias
Exactas y Naturales- UBA 2 Centro de
Investigaciones del Mar y la Atmósfera (CIMA) -
CONICET vidal_at_cima.fcen.uba.ar
INTRODUCTION OBJECTIVE Most of La Plata basin
and central-eastern Argentina has become
increasingly vulnerable to floods and droughts
with important social, economic and environmental
consequences. A detailed knowledge of the
hydrological variability in the region as a
consequence of climate variability and change
will be an important tool for the development of
adaptation strategies for the conservation of the
region. The Iberá wetlands are considered one
of the most important wetlands of the world due
to its high level of conservation and
biodiversity. They are located in the
central-northern province of Corrientes in
northeastern Argentina, covering approximately
15000 Km². Location of
Iberá wetlands. The objective of this work is to
evaluate the ability of the VIC hydrologic model
to represent the Corrientes River streamflows
downstream of Iberá wetlands in order to be used
as a tool for estimating future environmental
conditions in the region under different climate
scenarios. It is important to note that VIC
model was not designed specifically to simulate
streamflows in flooded areas, such as Iberá
wetlands. For this reason the main objective at
this time is simply to evaluate the ability of
the model to reproduce the mean streamflows of
the river.

• METODOLOGY DATASETS
• 
  
  

The VIC Model The Variable Infiltration Capacity
(VIC) macroscale hydrologic model was developed
at University of Washington, USA (Liang et al
1994) and it has already been succesfully applied
to La Plata basin (Su et al 2005) and to the
Uruguay basin (Saurral 2007). In this work the
VIC model was applied to Corrientes basin, which
includes Iberá wetlands, with closure point at
the gauge station Los Laureles (29.76ºS,
59.22ºW). Basically the model divides the basin
into an equally-spaced grid and it calculates the
hydrologic balance equations for every point at
each time step. To do that it takes into account
three sub-superficial layers and the vegetal
canopy. Furthermore, every grid point is
subdivided into subregions depending on the
different types of vegetal covers that exists in
that point. To evaluate the evapotranspiration it
considers the Penman-Monteith formula. .
Datasets Soil type dataset Soil Map of the
World (UNESCO). Spatial resolution 0.5º lat x
0.5º lon. Vegetal cover type dataset
University of Meryland (Hansen et al 2000) from
setellite estimations between 1992 -1993, spatial
resolution of 1 Km. Meteorological dataset
- Precipitation 221 daily gauge stations
(National Climatic Data Center NOAA, Instituto
Nacional de Tecnología Agropecuaria (INTA),
Subsecretaría de Recursos Hídricos de la Nación,
EVARSA, Servicio Meteorológico Nacional (SMN))
- Max/Min Temperatures 37 daily data stations
(National Climatic Data Center NOAA, Instituto
Nacional de Tecnología Agropecuaria (INTA),
Servicio Meteorológico Nacional (SMN)) -
Wind data Daily mean 10m wind velocity from
NCEP/NCAR reanalysis All datasets were gridded
at the model resolution by Krigging interpolation
method.
Model parameters Temporal resolution Daily
Spatial resolution 0.125º latitud x 0.125º
longitude Integration mode Water Balance
When VIC ends the integration, results are
given in the form of daily and monthly water
fluxes at each grid point, which need to be
routed to obtain the streamflow at the closure
point of the basin. Routing Model requires a
Digital Elevation Model (DEM) that accounts for
the elevation of each grid point of the domain.
In this case GLOBE (Global Land One-Kilometer
Base Elevation) data was used at original spatial
resolution of 1 Km. (Hastings et al 1990) and
then interpolated to the VIC model resolution.
RESULTS The calibration period for the model was
1994-1997. The model parametrs were selected
according to the best fit to observations B
(variable infiltration curve parameter)
0.001 DSmax (maximun velocity of baseflow) 2.5
m/s DS (fraction of DSmax where non-linear
baseflow begins) 0.001 WS (fraction of soil
mosture where non-linear baseflow occurs)
0.5 The validation period was1999-2001. 1998 was
not included in this period to avoid forcing the
model to an extreme El Niño event when severe
floods occured all over the region. Simulated
(MOD) and observed (OBS) monthly mean streamflows
time series are shown for both periods. 3-months
backward running average of the monthly
simulated streamflows are also presented.

In spite of the main objective of the simulation
which is to reproduce the average streamflow in
the validation period, the E estimator (Nash and
Sutcliffe (1970) was calculated as a measure of
the quality of the adjustment. This parameter has
the form where
Qot measured streamflow
at time t
Qmt simulated streamflow at time t


observed average streamflow Calibration is
considered to be fine if E gt 0.5. The
percentage of variance of the observed series
that is explained by the model (R2) was also
calculated. The following table summarizes the
results

• DISCUSSION
• In general terms, theVIC model simulates the
  average streamflow in the validation period
  similarly than during the calibration period.
• Simulated monthly streamflow series explain 55
  of the variance of the observed monthly
  streamflow series in the calibration period, but
  it decreases to 35 in the validation period.
• Due to the inherent limitation of the Routing
  Model to simulate the acumulation of water that
  is produced in wetlands systems such as Iberá, it
  tends to run off the precipitated water into the
  basin in a cuasi-instantaneous way simulating
  spurious streamflow peaks at the closure point.
  For that reason, the E estimator results lower
  than the 0.5 threshold.
• By considering three months backward running
  average of the monthly modeled series, model
  performance is enhanced which is reflected in
  incresing values of E estimator.
• FUTURE WORK
• To analyze the relationship between Iberá
  wetlands water level and the measured streamflow
  at the closure point of the basin.
• To study the variability of Iberá wetlands water
  level in response to different future climate
  scenarios by simulating the mean streamflow of
  the Corrientes River.
• VIC model successfully simulates the mean
  streamflow at the closure point of Corrientes
  basin, but it does not represent properly the
  monthly variability due to the impossibility of
  the Routing Model to acumulate the precipitated
  water in a lacunal area. Consequently, the
  routing scheme needs to be modified by adding the
  possibility of acumulate water up to a terrain
  dependent threshold and to rout the surplus. This
  way, the monthly variability of the streamflow at
  the closure point will be better simulated and it
  will be also possible to infer the water level of
  the lagoons of the Iberá wetlands system.

ACKNOWLEDGMENTS The authors thank Ramiro Saurral
for his advices and important tips on working
with the VIC model.
REFERENCES Hansen et al, 2000 Global land cover
classification at 1 km resolution using a
decision tree classifier. Int. J. Remote Sensing,
21, 1331-1364. Hastings et al, 1999 The Global
Land One-Kilometer Base Elevation (GLOBE) Digital
Elevation Model, Versión 1.0. National
Geophysical Data Center NOAA. Liang et al,
1994 A simple hidrologically based model of land
surface water energy fluxes for GSMs. J. Geophys.
Res. 99, 14415-14428 Saurral R. 2007 Impactos de
las variaciones climáticas en la cuenca del río
Uruguay. Tesis de Licenciatura en Ciencias de la
Atmósfera.- UBA Su et al, 2005 Modeling of land
surface processes in La Plata basin. Proceedings
of the AGU fall meeting, San Francisco, CA, USA..