03 Zubieta - Modelado hidrologico diario de la cuenca

Daily hydrological modeling in the Peruvian Amazon basin
using rainfall satellite estimates
Eng. Ricardo Zubieta Barragán
Espinoza Jhan Carlo , Lavado Waldo , Saavedra Miguel
5th ORE-HYBAM Scientific Meeting
INTRODUCTION
The Amazon basin extends
and from the Andes to the Atlantic
Ocean, covering approximately 6 000
000 km2. Its fresh water contribution
to the global ocean is 15% and its
average discharge at the delta is 209
000 m3/s (Molinier et al.,2009)
878,305 km2
(To the location
station Tabatinga) ,
average maximum
flow 49,000 m3/s(
station Tamshiyacu)
Fuente: El comercio
Predictions from rainfall–runoff models are
often unsatisfactory because spatial
variability in rainfall is poorly represented
in regions where data are scarce
furthermore the catch of conventional
raingauges is representative of only a small
radius around the instrument..
satellite estimates can be an alternative
source of data for rainfall-runoff simulation,
by improving the spatial distribution of
rainfall.
One of the more promising applications of
satellite-based rainfall estimates is the
coupling of rainfall measured in real time,
meteorological forecasts and rainfall–
runoff models for flow forecast, since there
are very few real-time raingauges in most of
the Perú
Rainfall regimes in eight stations in the Peruvian and
Ecuadorian Amazon basin., (Espinoza et al., 2009).
Main Objective
Evaluation of the usefulness of the
estimates of rainfall, TRMM, CMORPH,
PERSIANN and Reanalysis as an input
variable from the modeling rainfall - runoff
in the Peruvian Amazon basin using a
distributed hydrological model.
Source: SRTM
Modelo hidrológico concentrado ( SACRAMENTO, GR2M)
distribuido (VIC, MGB) --- principal diferencia : The spatial variability of land by
Distributed hydrological models unlike
vegetation, soils.
concentrates has been little studied and
applied in Peru.
On the other hand are also
used
semi
distributed
models (WEAP, SWAT,
HEC-HMS).
Figuras : Mantaro River
Basin
fuente proclim
MGB -IPH(Collischonn, 2001)
Modelo de Grandes Bacias
Daily hydrological
modeling
Collischonn 2007
The law of conservation of mass,
or principle of mass conservation,
PCi
ETi,j
Cell i
Minicuencas
EIi,j
Pi,j



Wm



Dsup i,j



Wi, j 



Qsup
Qsub
Qbas
Dint i,j
Dbas i,j
Minicuencas
Cell
downstream
aguas
of cell i
abajo
River
channel
Pixel and GRU
discretization
Model grid cells
Data used
METODOLOGY
•Caudales observados de la red OREHYBAM
•Reanalysis NCEP-NCAR
(Humedad Relativa, Vel. Viento,
Radicación solar, Temperatura del
aire)
•SRTM (The Shuttle Radar
Topography Mission).
•Mapa de cobertura vegetal del Perú
(INRENA)
•Mapa de suelos del Perú (INRENA)
•Mapa de tipos de vegetación del
Ecuador continental (INEFAN 1999)
•Mapa general de suelos del Ecuador
(Sociedad Ecuatoriana de la Ciencia
del Suelo 1986)
•Mapa de uso de suelo de Colombia
(IGAC, 2002)
•Mapa de suelos de Colombia (IGAC,
2005) .
Adaptado de
Collischonn 2007
Estimates of precipitation
from remote sensing techniques
TMPA : (Huffman et al., 2010)
GPROF
(Kumerow
at al., 1996)
{
{
NESDIS (Zao and
Weng 2002) IWP
GOES Precipitation
Index (GPI; Arkin
And Meisner 1987)
Calib
TMI
Adj
TCI (Miller 1972, Krajewski and
Smith (1991)) TCI (PR-TMI)
TMPA
CMORPH
PERSIANN
5000
0
0
0
4500
−2
−2
−2
3500
4500
4500
4000
−4
5000
5000
4000
4000
−4
3500
−4
3500
−6
3000
−6
3000
−6
3000
−8
2500
−8
2500
−8
2500
2000
−10
2000
−10
1500
−12
−12
1000
−14
500
−80
500
−76
−74
−72
−70
−80
500
−16
−16
−78
1000
1000
−14
−16
1500
1500
−12
−14
2000
−10
−78
−76
−74
−72
−70
−80
−78
−76
−74
−72
−70
Precipitación media multianual 2003-2009
Elaboración propia
RESULTS
Modelling Geographic
Information Systems
1. Delimitación de cuencas
semiautomatizada a diferentes detalles
En Software GIS
RESULTADOS
Obtaining HRU BY
GEOGRAPHIC INFORMATION SYSTEMS
Procesamiento de Mapas temáticos
Uso de suelo
Tipo de suelo
Unidad de
Respuesta
Hidrológica
Imagen Landsat (a); Índice de área foliar (b) y albedo de superficie (c),
estimado por satélite.
RESULTS
RESULTADOS
TRMM
CMORPH
PERSIANN
REQUENA
TRMM
PERSIANN
CMORPH
SAN REGIS
TRMM
CMORPH
PERSIANN
Efficiency coefficients Nash - Suttcliffe for estimated precipitation products.
Performance data from Reanalysis precipitation
Number River
1 Napo
2 Napo
3 Ucayali
4 Marañon
Station
Area
Nueva Loja
San Sebastian
Km 105
NCEP DOE
< 0.1
< 0.1
< 0.1
ERA
INTERIM
< 0.1
< 0.1
< 0.1
22068 < 0.1
< 0.1
< 0.1
NCEP NCAR
4331 < 0.1
5311 < 0.1
9635 < 0.1
23857 < 0.1
< 0.1
< 0.1
6 Napo
7 Marañon
8 Marañon
Santiago
Francisco de
Orellana
Nuevo
Rocafuerte
Nueva York
Chazuta
30428 < 0.1
39634 < 0.1
68685 < 0.1
< 0.1
< 0.1
0.37
< 0.1
0.23
0.55
9 Marañon
Borja
92302 < 0.1
< 0.1
< 0.1
5 Napo
10 Napo
100169 < 0.1
< 0.1
< 0.1
11 Ucayali
12 Ucayali
Bellavista
Santa Rosa de
Ucayali
Pucallpa
191159 0.2
260418 0.2
0.52
0.65
0.59
0.69
13 Ucayali
14 Marañon
Requena
San Regis
350215 0.19
359883 < 0.1
0.61
< 0.1
0.61
0.64
15 Amazon
16 Amazon
17 Amazon
Iquitos Enapu
Tamshiyacu
Nazareth
682970 0.11
699381 0.11
877763 < 0.1
0.1
0.14
< 0.1
0.63
0.67
0.63
18 Amazon
Tabatinga
878141 0.2
< 0.1
0.71
Conclusions
In this work, the performance by MGB-IPH modeling of discharge values in the main
tributaries of the Amazon Andes basin of Perú and Ecuador using six precipitation datasets at
the daily time step over , as input to rainfall-runoff models was evaluated by comparing with
observed discharge, focusing on the use of several high (TMPA, CMORPH, PERSIANN) and low
(NCEP, NCEP-DOE, ERA INTERIM)
The modeling using precipitation satellite is able to reproduce hydrographs for large river in
the Peruvian Amazon, these results suggest that the product 3B42 v7 can be used as input to
a hydrological model of rainfall - runoff in the Marañon and Ucayali basins.
The results show a clear opposition in the performance of the model in the basins located
between the north and south of the tropical regions of the Peruvian Amazon, and similar
conditions for satellite precipitation products and Reanalysis used, this model shows the
difficulty of hydrographs represent observed in regions closer to the equator, characterized by
weak seasonal variability, achieving performance levels NS <0.4.
Generally, results shows that not tend to be better when evaluated at gauging stations
controlling large drainage areas. This is not clear and can be due to rainfall data quality
difficult for the model to represent different streamflow regimes and the confluence of rivers
upstream whose performance is much lower or higher than others.
GRACIAS
THANK YOU
OBRIGADO
[email protected]
Agradecimientos :