Transcript 2005_soutenance_Thanh

Ph.D. Defense
Modeling the continental hydrologic
cycle: interannual variability and trends.
Comparison with observations
Thanh NGO-DUC
Modélisation des bilans hydrologiques continentaux:
variabilité interannuelle et tendances.
Comparaison aux observations
23/09/2005
Thanh NGO-DUC
Page ‹#›
Introduction
Our climate is changing, which direct
consequences for the Earth.
Variations in greenhouse gases, aerosols, and
land use/cover force changes in climate…
Bolivie, 1983 ©IRD photo Denis Wirrmann
…but, most of consequences of climate
change are realized through the water
cycle : flood, drought, sea level rise, etc.
Brésil, 1997 ©IRD/ photo Bernard Osès
23/09/2005
Thanh NGO-DUC
Page ‹#›
Introduction
Water cycle
http://www.wilkes.edu
×1012m3/yr
Solar
heat
water vapor
Net mouvement
of water by wind
water vapor
40
Precip.
111
Evaporation
Precip.
387
Evaporation
427
71
GRACE
Ocean
flow of water
40
?
surface water and
ground water
model
Water exchanged volume estimated by Baumgartner et Reichel (1975)
This thesis aims to study the variability of continental hydrologic
cycle by using numerical models and observations.
23/09/2005
Thanh NGO-DUC
Page ‹#›
Introduction
The ORCHIDEE land surface model (LSM)
ORCHIDEE is the new land-surface scheme of the IPSL. It is composed of:
SECHIBA : surface energy and water balances
STOMATE : surface biochemical processes
LPJ
: dynamical evolution of the vegetation and the carbon budget
Inclusion of a routing scheme, which
routes the water to the oceans through
a cascade of linear reservoirs.
2m
Only SECHIBA is used.
river discharge
ORCHIDEE: Organising Carbon and Hydrology in Dynamic EcosystEms
SECHIBA : Schématisation des Echanges Hydriques à l’Interface entre la Biosphère et l’Atmosphère
STOMATE : Saclay Toulouse Orsay Model for the Analysis of Terrestrial Ecosystems
LPJ: Lund –Postdam-Jena
23/09/2005
Thanh NGO-DUC
Page ‹#›
Plan
Thesis
Seasonal/interannual
variations using
Topex/Posédion,
Decadal/interdecadal
timescales,
the NCC forcing data set,
continental scale
basin/continental scale
off-line simulation
1987-1988
coupled simulation
1997-1998
Ngo-Duc et al.
(JGR, 2005a)
Construction &
validation
Ngo-Duc et al.
(JGR, 2005b)
Seasonal variations
using GRACE,
basin scale
Applications
land water &
sea level
Ngo-Duc et al.
(GRL, 2005c)
(NCC:NCEP/NCAR (National Centers for Environmental Prediction/National
Center for Atmospheric Research) Corrected by CRU (Climate Research Unit)
23/09/2005
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
Topex/Poséidon (T/P)
Launched: 10/08/1992
Orbit: quasi-circular, 66°,
1336 km of altitude
Principe of altimetry
Radar altimeters transmit signals to
Earth, and receive the echo from the
sea surface.

Measuring the time span between
sending and receiving of the pulses
allows to determine the height of the
satellite above sea-level.

The distance from the satellite to the
mean earth ellipsoid is known.

the height of the sea-level above
the ellipsoid
23/09/2005
http://www.aviso.oceanobs.com
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
Causes of sea level variations:
thermal expansion of the oceans (steric effect)
water mass exchanged with other reservoirs
8.0
mm
4.0
NCEP/NCAR vapor
0.0
- 4.0
- 8.0
T/P
Ishii steric
T/P-steric-vapor
mean seasonal variations 1993-1998,
expressed in sea level equivalent
23/09/2005
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
a) forced simulation for 1987 & 1988
ISLSCP-I (International Satellite Land-Surface Climatology Project, Initiative I)
produced the atmospheric forcing over the continents for 1987 and 1988.
12.0
mean seasonal cycle for the
period 1993-1998
T/P derived value
(T/P-steric-vapor)
8.0
The model outputs are comparable to
the observations (phase, amplitude)
The differences could be due to :
 incompatibility of the compared
periods
 data/model uncertainties
4.0
0.0
- 4.0
- 8.0
ORCHIDEE forced by ISLSCP-I
- 12.0
In the simulation, there are interannual
variations between 1987 and 1988
1987
1988
continental water variations expressed
in sea level equivalent (mm)
next study: use GCM simulations, for 1997 and 1998
23/09/2005
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
b) coupled simulation for 1997 & 1998
Ngo-duc, T., K. Laval, J. Polcher and A. Cazenave (JGR, 2005a)
AMIP Simulation
•
●
LMD GCM, version 3.3 ; 96×72×19
Forced by SST form 1979 to 1999
mm
AMIP
AMIP
T/P derived value
1997
1998
Contribution of continental water to sea level variations
AMIP
23/09/2005
: Atmospheric Model Intercomparison Project
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
b) coupled simulation for 1997 & 1998
Seasonal variations of tropical continental water
expressed in terms of equivalent sea level
4.0
mm
2.0
0.0
- 2.0
- 4.0
A major part of the interannual variability of the continental water
storage comes from the strong variability of precipitation on the tropical
continents.
Our results don’t agree with the analysis of Chen et al. (2002) who attributed
the contrast between 1997 and 1998 to a change in snow cover at high latitudes.
23/09/2005
Thanh NGO-DUC
Page ‹#›
I. Topex/Poséidon & ORCHIDEE
c) Limitations and perspectives
Limitations of this part
forced simulation: the period of the ISLSCP-I forcing data is incomparable
with the Topex/Poséidon

coupled simulation: uncertainty of the precipitation fields, in particular when
looking at geographical details.

next study: forced simulation over a long period
23/09/2005
Thanh NGO-DUC
Page ‹#›
Plan
Thesis
Seasonal/interannual
variations using
Topex/Posédion,
Decadal/interdecadal
timescales,
the NCC forcing data set,
continental scale
basin/continental scale
off-line simulation
1987-1988
coupled simulation
1997-1998
Ngo-Duc et al.
(JGR, 2005a)
23/09/2005
Construction &
validation
Ngo-Duc et al.
(JGR, 2005b)
Seasonal variations
using GRACE,
basin scale
Applications
land water &
sea level
Ngo-Duc et al.
(GRL, 2005c)
Thanh NGO-DUC
Page ‹#›
II. The NCC atmospheric forcing data
a. Construction
Ngo-duc, T., J. Polcher and K. Laval (JGR, 2005)
NCEP/NCAR Reanalysis 6-hourly, ~1.875°, 1948-present
NCEP
NPRE
Interpolation to the grid 1°×1°, differences in
elevation between the grids were taken into account
CRU (Climate Research Unit) precipitation
0.5°×0.5°, 1901-2000
CRU temperature
Specific humidity, pressure, precipitation
NCRU
Radiation: SRB (Surface Radiation Budget)
NCC
(NCEP Corrected by CRU)
6-hourly, 1°x1°, 1948-2000
http://dods.lmd.jussieu.fr/cgi-bin/nph-dods/Dods/NCC/ (~40GB)
23/09/2005
Thanh NGO-DUC
Page ‹#›
II. The NCC atmospheric forcing data
b. Validation
The world's 10 biggest rivers (by
the estimated river mouth flow rate)
Station Obidos
55.51°W, 1.95°S
Observed discharge
 GRDC (Global Runoff Data Center)
 Data at UCAR (the University Corporation for Atmospheric Research)
23/09/2005
Thanh NGO-DUC
Page ‹#›
II. The NCC atmospheric forcing data
b. Validation
NCC
OBS
NCEP
 quality of the NCC forcing
data is improved compared
to NCEP/NCAR
 high flow in simulated
mean seasonal signal is too
early
NCC
OBS
 the interannual signal is
well described by the NCC
experiment
NCEP
23/09/2005
Thanh NGO-DUC
Page ‹#›
II. The NCC atmospheric forcing data
b. Validation
Standard deviation
Standard deviation
Taylor diagram
(Taylor, 2001)
OBS
OBS
Standard deviation
1. Amazon
2. Congo
3. Orinoco
4. Changjiang
5. Brahmaputra
NCEP
NCRU
6. Mississippi
7. Yenisey
8. Parana
9. Lena
10. Mekong
NPRE
NCC
OBS
23/09/2005
Thanh NGO-DUC
The quality of
forcing data is
improved after
each
adjustment.
Precipitation: most
important improvement
Temperature:
significant
effect only at high
latitudes
Radiation: improves
discharge amplitudes
Page ‹#›
II. The NCC atmospheric forcing data
b. Validation
Standard deviation
GSWP2
NCC
1. Amazon
2. Congo
3. Orinoco
4. Changjiang
5. Mississippi
6. Yenisey
7. Parana
OBS
Comparison between
NCC and GSWP2
(Global Soil Wetness Project)
Discharge is better
simulated using NCC
than GSWP2.
Standard deviation
Standard deviation
series
mean seasonal signal
23/09/2005
anomaly
Thanh NGO-DUC
Page ‹#›
II.c. Applications of NCC
Effects of land water storage on global mean sea level
Effects of land water storage on global mean sea level over the past 50-yrs
Ngo-Duc, T., K. Laval, J. Polcher, A. Lombard et A. Cazenave (GRL, 2005)
Over the past 50 yrs, the rate of global mean sea level rise was on the order of 1.8
mm/yr [Church et al., 2004], where:
Thermal expansion contributes ~ 0.4 mm/yr [Lombard et al., 2005]
Mountain glaciers melting accounts for ~ 0.4 mm/yr [Meier and Duygerov, 2002]
Greenland & Antarctica melting provide ~ 0.5 mm/yr [Thomas et al., 2004]
What is the land water contribution?
23/09/2005
Thanh NGO-DUC
Page ‹#›
II.c. Applications of NCC
Effects of land water storage on global mean sea level
greatest variation is
associated with ground
water, followed by soil
moisture
no significant trend was
detected
strong decadal variability
driven by precipitation, strong
decrease in the beginning of
1970s
5-yr moving average of water reservoirs changes
expressed as equivalent global sea level anomalies.
23/09/2005
Thanh NGO-DUC
agreement between
ORCHIDEE and LaD.
(Land Dynamics LSM of
GFDL)
Page ‹#›
II.c. Applications of NCC
Effects of land water storage on global mean sea level
the strong decrease of the
global signal in the early 1970s is
due to changes in the Amazon
basin
during the past 50 yrs, the
northern tropical Africa lost water
to the benefit of the oceans
regions 2 and 3 seem to be anticorrelated (-0.78), suggesting
a possible teleconnection
mechanism
5-yr moving average time series of changes in
land water storage for the six study regions
23/09/2005
Thanh NGO-DUC
Page ‹#›
II.c. Applications of NCC
Effects of land water storage on global mean sea level
Relations between land water and thermosteric sea level fluctuations
R=-0.84
+ Tocéan
+Eocean, (P,E,R)land
+ Vocéan
+ Mcontinents
- Mocéan
These results do not confirm the
suggestions of Gregory et al. [2004] that
decadal fluctuations in ocean heat
content are artifacts of the interpolation
processes of raw hydrological data.
23/09/2005
+
-
sea level rise
oceans warmer
continents wetter
negative feedback to sea level
Thanh NGO-DUC
Page ‹#›
Plan
Thesis
Seasonal/interannual
variations using
Topex/Posédion,
Decadal/interdecadal
timescales,
the NCC forcing data set,
continental scale
basin/continental scale
off-line simulation
1987-1988
coupled simulation
1997-1998
Ngo-Duc et al.
(JGR, 2005a)
23/09/2005
Construction &
validation
Ngo-Duc et al.
(JGR, 2005b)
Seasonal variations
using GRACE,
basin scale
Applications
land water &
sea level
Ngo-Duc et al.
(GRL, 2005c)
Thanh NGO-DUC
Page ‹#›
III. GRACE & ORCHIDEE
GRACE Mission (Gravity Recovery And Climate Experiment)
© NASA
GRACE, twin satellites launched in March 2002, are making detailed
measurements of Earth's gravity field. They study the movement of water
over the surface of the Earth with a level of detail never before possible.
23/09/2005
Thanh NGO-DUC
Page ‹#›
III. GRACE & ORCHIDEE
GRACE
LaD model
Seasonal variations of continental water
April/May - November 2002
How do ORCHIDEE results compare?
Figure provided by Ramillien, G., LEGOS
23/09/2005
Thanh NGO-DUC
Page ‹#›
III. GRACE & ORCHIDEE
Numerical experiments
- Built a new atmospheric forcing from 2001 to 2003, named NCMAP
(NCEP/NCAR-NCC-CMAP): constrained by monthly CMAP
precipitation.
- experiment ORCHIDEE-1: ORCHIDEE without routing scheme,
forced by NCMAP
- experiment ORCHIDEE-2: ORCHIDEE with routing scheme,
forced by NCMAP
CMAP
23/09/2005
: CPC (Climate Prediction Center) Merged Analysis of Precipitation
Thanh NGO-DUC
Page ‹#›
III. GRACE & ORCHIDEE
GRACE
ORCHIDEE with routing
ORCHIDEE without routing
Seasonal variations of
continental water
April/May – November 2002
The routing scheme much improves the signals
over tropical basins
23/09/2005
Thanh NGO-DUC
Page ‹#›
III. GRACE & ORCHIDEE
Variations of water stock
over the 8 tropical basins
GRACE
ORCHIDEE
with routing
ORCHIDEE
without routing
The routing scheme much
improves the signals over
tropical basins
23/09/2005
Thanh NGO-DUC
Page ‹#›
Conclusions
ORCHIDEE is able to reproduce seasonal and interannual
variations of continental water reservoirs
The important role of the tropical regions in the variability of the
climate was underlined
The NCC data set was found to be reliable in the validations
On studying the variability of land water storage, an hypothesis
was proposed: when the oceans are warmer, the continents will be
wetter, leading to a negative feedback on sea level changes
The role of the routing scheme on simulating land water
storage was shown
23/09/2005
Thanh NGO-DUC
Page ‹#›
Future directions
Merica and north tropical
study the anti-correlation between South A
Africa
examine land water storage at interannual/d
GRACE and other LSMs, forcing data sets
look at smaller scale: soil moistu
Ecadal timescale using
Re (T. D’Orgeval)
Change
study anthropogenic impact : irrIgation, floodplains, dams, …
study the relations with climate
Soil moisture index comparison
Observations: Global soil moisture
data bank [Robock et al., 2000]
23/09/2005
Thanh NGO-DUC

Plant  available soil moisture
Soil moisture capacity
Page ‹#›