Transcript Monte Carlo Analysis of Land Surface Parameters using - U

Analysis of TRMM and gauge rainfall in the La Plata River Basin in
South America for hydrologic predictions
H33A-0976
Eleonora Demaria1, Juan Valdés2, Daniel Rodriguez3 and Matej Durcik4
1Department
email: [email protected]
of Hydrology and Water Resources, University of Arizona, 1133 E James E. Rogers Way, Tucson, AZ 85721, United States
2Department of Civil Engineering and Engineering Mechanics, University of Arizona, Tucson, AZ 85721, United States
3Centro de Previsàõ de Tempo e Estudos Climàticos, Rodovia Presidente Dutra, Km 40, Cachoeira Paulista, SP 12630, Brazil
4SAHRA-University of Arizona, Marshall Bldg. 845 N. Park Ave., Room 545. P.O.Box 210158-B, Tucson, AZ 85721, United States
Results and Discussion
Motivation
The CRA method was used to evaluate how well TRMM 3B42.v6 captures the spatial structure of
ten MCCs that occurred over the basin in the summer of 2002-03. A total of 140 MCCs have been
identified over the basin during that period using the methodology described by Salio et. al.,
2006. MCCs ranged in size from 50,000 to 1,138,00 km2.
January
23rd,
TRMM Prec. @ 12 UTC
2003 - 02:00 UTC
100%
80%
60%
0.0 %
1.4 %
98.6%
10%
1.1 %
0%
Volume
Pattern
 TRMM 3B42.v6 represents the maximum rain
rate closely but it fails to accurately estimate
the rain volume of the complexes.
Storm Displacement in degrees
Longitude
1
0
-5
-4
-3
-2
-1
0
-1
-2
1
2
3
 Positive BIAS larger
5%
6.3 %
9.3 %
7.7 %
 The heaviest rains tend to be forecast to the
north of where they actually occurred.
3
Northwards
displacement
2
4
5
 The horizontal displacements of the center of
maximum rain intensity can surpass 100 km.
Misplacement of the storm center has great
impact on hydrologic predictions, agriculture
uses and are invaluable for evacuation
efforts.
-5
 Cumulative
precipitation shows
better agreement
during the wet
season over both
basins
 Largest component of the error is due to
small differences in the structure of forecast
and observed fields.
-4
TRMM estimates were temporal aggregated at 12 UTC to obtained daily totals over each grid cell. Spatial
maps of satellite errors were created as the difference between TRMM 3B42.v6 and observations for daily
and seasonal data.
Observed daily precipitation from SALLJEX were linearly interpolated to a 0.25 degree grid cell. The
magnitude, spatial correlation and horizontal displacement of ten MCCs that occurred over the basin
during the SALLJEX field experiment are evaluated with the Contiguous Rain Area (CRA) method. CRA,
developed by Ebert and McBride (2000), horizontally displaces the forecast over the observed field until
the best match is obtained based on the minimization of the Mean Squared Error (MSE). The method
allows a decomposition of the total MSE into components due to: a) Displacement ; b) Volume and c)
Pattern.
2.6 %
20%
-3
•TRMM precipitation estimates (3B42v.6 post-real time) @ 0.25 degree (Huffman 2007) - Period 19982006
•Daily gridded observed precipitation available @ 0.25 degree (Liebmann and Allured, 2005)
•Daily observed precipitation SALLJEX (South America Low Level Jet Experiment) extended rain gauge
network. Period November 15th 2002-February 15th 2003.
1.1 %
30%
4
Data and Methodology
5%
40%
5
Source: Climate Change in the La Plata Basin, Barros, et al.
Error Decomposition
Displacement error
Volume error
Pattern error
50%
Upper Paraná Basin
Summer (DJF) maximum
precipitation linked to
monsoon activity
Iguazu Basin
maximum precipitation Fall
(SON) and Spring (MAM)
Su et al., (under review) shows that streamflows simulated with TRMM 3B42.v6 precipitation
overestimate by a 22% streamflows simulated with observed fields over the Upper Paraná basin. The
bias in the Iguazu Basin is 9%. Spatially averaged daily precipitation over the two basins show:
Gauge rainfall
Displacement
Accumulated precipitation (mm)
u
n
d
e
r
70%
11
/1
11 8/
/2 02
11 2/
/2 02
11 6/
/3 02
0
12 /0
/4 2
12 /0
/ 2
12 8/
/1 02
12 2/
/1 02
12 6/
/2 02
12 0/
/2 02
12 4/
/2 02
8/
1/ 02
1/
1/ 03
5/
1/ 03
9
1/ /0
13 3
1/ /0
17 3
1/ /0
21 3
1/ /0
25 3
1/ /0
29 3
/
2/ 03
2/
03
La Plata River Basin
second largest river basin in South America
covers 3.2 x 106 km2
mean annual streamflow ~ 21,000 m3/s (outlet)
total precipitation ~ 2000 mm
shared by Argentina, Bolivia, Brazil, Paraguay and Uruguay
50% population of those countries reside in the basin
generates 70% of GPD of the five countries
Gauge Prec. @ 12 UTC
Area ~ 510,000 km2
90%








o
v
e
r
Satellite estimates underestimate observed precipitation in the region of maximum precipitation  the
Iguazu River Basin. Conversely satellite tends to overestimate observations on the northeast of the
basin  the Upper Paraná basin and over part of the southeast. Due to lack of rain gauges in the west
part of the basin biases were not computed.
We selected ten MCCs with maximum
areas larger than 350,000 km2
Study Area
Spatial distribution of daily differences (BIAS) between TRMM 3B42.v6 and observations for each season
Fields processed
with CRA
Predicted rainfall (shifted)
The accurate estimation of rainfall is crucial for flood and drought monitoring, hydroelectric
generation, navigation and water resources management. In developing regions of the world,
available hydrometeorological information is obtained from a coarse network of precipitation
and streamflow gauges. In this context, satellite estimated precipitation constitutes an
invaluable source to be used for flood prediction.
Emergent economies in South America heavily depend on their water resources for economic
development. For this reason, it is necessary to quantify the discrepancies existent between
satellite products and gauge measurements in order to understand how errors in satellitederived precipitation could affect simulated hydrological fluxes. In the La Plata River Basin,
Mesoscale Convective Complexes (MCCs) are the dominant source of precipitation contributing
as much as 80% of the total precipitation in some areas. Their accurate spatial representation
by satellite products is crucial since horizontal displacements in the location of storms have a
great impact on streamflows forecasted with hydrologic models.
The objective of this study is twofold: identify regions of the basin where satellite
estimates over/underestimate precipitation amounts and intensities; and, verify to
what extent satellite precipitation estimates reproduce the spatial properties of
Meso Convective Complexes over the La Plata River Basin, South America.
in the Upper Paraná
Basin than in the
Iguazu Basin. It can
be related to rain
gauge density,
larger area and
different rainfall
genesis
References
Latitude
Future work
We plan to extend the CRA analysis using more MCCs. A similar analysis will be performed with
smaller storms that occurred over the Upper Parana and Iguazu basins during the TRMM 3B42.v6
period.
The next step is to implement a semi-distributed model over the La Plata Basin (VIC model). We will
bias-correct satellite estimates and drive VIC with satellite precipitation, observed precipitation and
bias-corrected precipitation to asses the impact of errors on streamflow simulations.
We will compute BIAS and RMSE between TRMM and rain gauge precipitation at each rain gauge
location to see the impact of interpolation schemes on the differences. Categories based on rain
intensity will be used to compute the errors.
Barros V., R. Clarke and P. Silva Dias. Climate Change in the La Plata Basin. Project SGP II 057- Inter American Institute on Global Change
(IAI)
Ebert E. E. and J.L. McBride, 2000. Verification of precipitation in weather systems: determination of systematic errors, J Hydro, 179-202.
Huffman G.J. et al, 2007. The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-Global, Multiyear, Combined-Sensor Precipitation
Estimates at Fine Scales. J Hydrometeorology, 8, Issue 1, 38-55.
Liebmann B and D. Allured, 2005. Daily Precipitation Grids for South America. BAMS, 1567-1670.
Salio P. et al, 2006. Mesoscale Convective Systems over Southeastern South America and their relationship with the South American LowLevel jet. M. Weather Review, 135, 1290-1309.
Su F. et al., (2007 submitted to J. Hydrometeorology). Evaluation of TRMM Multi-satellite Precipitation Analysis (TMPA) and its utility in
hydrologic prediction in La Plata Basin.
Acknowledgements
This project has been funded by NASA Grant # NNG04GA79G to the University of Arizona and partially by SAHRA. We would
like to thank the Institute for Study of Planet Earth (ISPE) at the University of Arizona for covering some travel expenses.
Particular thanks go to Dr. Paola Salio from the Universidad de Buenos Aires (Argentina) who provided us with the list of
MCCs used in this paper. Many thanks to Hoori Ajami for her help with ARCGIS. Dr. Beth Ebert from the Bureau of
Meteorology Research Centre (Australia) patiently answered all our questions about CRA. We greatly appreciate Julio CañonBarriga constant help with MATLAB and discussing the results.