1 cnrs, centre national de recherches météorologiques (cnrm / météo-france, toulouse)

American Geophysical Union Meeting, May 22-25, 2007

1 CNRS, Centre National de Recherches Météorologiques (CNRM / Météo-France, Toulouse)

2 CNRS, Centre d’Etudes de la BIOsphere (CESBIO, Toulouse, France)

Aaron Boone1, Patricia deRosnay2, and the ALMIP Working Group

AMMA Land surface Model Intercomparison Project


Goal: Obtain a better understanding of the West African

Monsoon (WAM) on daily to inter-annual timescales. AMMA relies on:

i) the use of extensive observations, notably multi-

sensor remote sensing products, ii) a multi-year field campaign over west Africa, and iii) a coordinated modeling strategy at different spatial

and temporal scales.

AMMA = African Monsoon Multidisciplinary Analysis

* See the AMMA Special Issue of GEWEX News, Feb., 2006


Importance of the Continental Land surface to the WAM

Surface processes which

influence the atmosphere

and hydrology:

• Significant Meridional

Gradients:

- vegetation

- soil moisture

• Large spatial variability:

- surface runoff, infiltration, flooded areas

- surface fluxes


ALMIP Science Questions:

• No one scheme perfect: examine ensemble of models (currently 12): Response of different schemes to scale change

• Develop a mult-model climatology of « realistic » high resolution soil moisture, surface fluxes, water and energy budget

• See what improvements needed in SVATs to simulate processes particular to African climate and surface

• Examine how simple SVATs can simulate spatial distribution and inter-annual variability of vegetation

Break various components of complex coupled system into managable portions which can then provide insight into various processes: 1st step Force LSMs in Offline mode

GSWP

SnowMIP


ORCHIDEE LMD, Paris, France J. Polcher, T. Orgeval

ISBA CNRM / Météo France Toulouse A. Boone

TESSEL ECMWF, Reading, UK A. Beljaars, G. Balsamo

JULES CEH, Wallingford, UK C. Taylor, P. Harris

SETHYS CETP, France C. Ottlé, B. Decharme

CaB UPMC, France A. Ducharne, S. Gascoin

IBIS ISE, Monpellier, France C. Delire

SSib U. Nantes, France, UCLA, USA I. Poccard-Leclercq, Y. Xue

MIKE-SHE U. Copenhagen, Denmark A. Norgaard, I. Sandholt

CLM NCAR, Boulder, CO, USA B. Lamptey

SWAP Inst. Of Water Problems, Moscow, Russia

Y. Gusev, O. Nasonova

NOAH (NCEP, USA) CETP, France C. Ottlé, B. Decharme

Model acronym Institute Contact


Precipitation (mm day-1) 2004 ECOCLIMAP LAI (m2 m-2) month

Land Surface Model Soil-Vegetation Parameters

• ECOCLIMAP – global 1x1 km2, decadal

• Operatioal NWP, Mesoscale research

• Large annual variability (« mirrors » rainfall)

• Single annual cycle


ECMWF Forecast Product:

• Lowest model level atmospheric variables Tair, Qair, Wind, PSurf

• Vertically integrated fluxes SWdown, LWdown, Rainf, CRainf

Merged Forcing:

• ECMWF + Satellite based


• MSG Data: 0.6µm, 0.8µm, 1.6µm

• Solar and View Angles

• Land/Sea Mask

• Cloud Mask (SAF-NWC software)

• Total Column Water Vapour (ECMWF)

• Ozone Content (Climatology)

• Land Surface Albedo: Static Map, later AL product

• [Visibility -> Aerosol Optical Thickness]

• AMMA-SAT 0.05 deg., 30 min.

• Began July, 2005

LAND-SAF

Downwelling radiative flux products

SWdown:

B. Geiger


• SWdown, LWdown radiative fluxes

• MSG (coherent with AMMA-SAT precip.)

• OSI-SAF 0.10 deg., 3h 2004

• LAND-SAF 0.05 deg., 30min 2005+

SAF Radiative flux products:

Same tendency as OSI-SAF: monsoon further N than in ECMWF

Combine with EPSAT: up through 2006


Equation 2

EPSAT-SG(Estimation des Pluies par SATellite – Seconde Génération)F. Chopin, M. Dubois, LMD, Paris, France

Equation 1

Neural NetworkMSG Channels

SRTM Digital Elevation Model

Rainfall probability images (Pr) GPCP1dd rainfall images (Igpcp)

Potential rainfall intensity images (Ip)

EPSAT-SG rainfall est. (Ie) daItaPtaI pre ,,,

Equation 2Estimated Rainfall Intensity at time t during day d and position a:

Final product resolutions:Space resolution : 3 kms

Time resolution : 15 minutes

2A25 TRMM precipitationRadar data

AMMA-SAT Rainfall product:


Monthly Average Precipitation Rates (mm day-1) 2004

A Boone CNRS/CNRM

Recent Study (Jobard et al. 2007): EPSAT best statistical results 2004-6 AGRHYMET Sahelian rain obs

June July Aug. Sep.

EP

SA

TE

CM

WF

RF

E (

CP

C)


Exp 1: Regional

0.5deg, 3h: NWP

2002-2005 (2006)

Exp 2: Regional Merged

0.5deg, 3h: 2004-2005 (2006)

Exp 3: Mesoscale Merged

0.1deg, 1h: 2004-2006

Exp 4: Local 1-4 sites 2004-2006

• Analysis and intercomparison 2004-2005 (2006)

• Repeat Above using « Interactive Vegetation » option

Exp. Name: Resolution Time period

Experiments:


Total Evapotranspiration 2004 (JJAS) in mm day-1

• Exp1 (ECMWF Precip)

• Large inter-model differences





Evapotranspiration Difference:

• Exp2 (Satellite-based forcing) – Exp1 (NWP-based)

• Northward displacement of active precipitation

• 50 to 100% increase in Sahel: region with possibly largest atmospheric feedback


Meridional gradient of Evap:

• Exp 2 vs Exp 1

• Northward shift by several 100 kms

• Implications for NWP model initialization (e.g. ECMWF)

Exp2 2004: Large inter-model scatter quantity atmopshere « feels »


Exp2 JJAS 2004 All-SVATs

Dominant land class:

Coefficient of Variation:

Inter-model scatter

Large in Sahel water stress, low average Evap (where coupling is important ?)

Large in Forested regions Canopy Interception differences important

Product USERS give some estimate of inter-model variability


• Exp2 2004 JJAS Multi-Model AVG

• Water budget terms (mm day-1)

DelSoilMoist

Runoff

Evap

Rainf


DelSoilMoist/Rainf

Runoff/Rainf

Evap/Rainf

Rainf

• Exp2 2004 JJAS Multi-Model AVG

• Water budget terms (mm day-1)


09/84

09/87

09/86

09/00

09/92

09/89Gourma, Mali

• September

• Significant inter-annual vegetation variability

• Can LSMs model this?

* Photos courtesy of CESBIO


Vegetation/Soil Parameters

Atmospheric Forcing

ISBA-Ags

Biomass (LAI)

Turbulent and Radiative fluxes (CO2)

Runoff Irrigation

Interactive mode

Pashtuchak, Peyrillé, Plante, and Boone

Rivers, Lakes… Ocean/Sea

Inputs: Outputs:



Atmospheric Forcing

ISBA-Ags

Biomass (LAI)


Runoff Irrigation

Interactive mode

Peyrillé and Boone

Rivers, Lakes… Ocean/Sea

Coupled (atmosphere) mode


Exp 1 Exp 2

• Correlations between 0.8 and 1.0 over a large part of Sahel (where vegetation signal strongest)

• Exp2 (satellite-based) precipitation improved LAI prediction (and enhanced Evap)

But…

• Montly LAI misses rapid greenup

• Forest regions: MODIS quality questionable AND ISBA-Ags needs work!

• Care must be taken in fully coupled mode!


• ALMIP ongoing: 2006 Forcing to be released and runs using updated forcing (new EPSAT algorithm) to be submitted this summer (Exp2b)

• Complete regional merged and mesoscale atmospheric forcing data (2004): extend through 2006

• Regional scale evaluation data (forcing and for simulations) comparison with AMSR-E, ALMIP-MEB, TRIP 0.5 deg, GRACE…

• 2007-8 Mesoscale Forcing Evaluation of processes using SOP, EOP data

• 1st step towards ALDAS

• Collaboration with WAMME (Lau, Xue), AMMA-CROSS (Guichard, Hourdin, et al)

Parallel actions at CNRM:

• Simulate land surface fluxes, soil moisture, vegetation at various scales

• Compare simulated soil moisture to ARPEGE/assimilation

• NWP/Mesoscale model initialization: CNRM

Perspectives:


Thank you for Your attention…

By M. Nuret, Dakar, Senegal

And to the ALMIP Working Group:Anton Beljaars, ECMWF

Aaron Boone, CNRM

Jan Polcher, LMD

Chris Taylor, CEH

Phil Harris, CEH

Inge Sandholt, U. Copenhagen

Anette Norgaard, U. Copenhagen

Patricia deRosnay, CESBIO

Eric Mougin, CESBIO

Laurent Kergoat, CESBIO

Agnes Ducharne, UTMC

Tristan Orgeval, LMD

Yonkang Xue, UCLA

Isabelle Poccard-Leclercq, U. Nantes

Christine Delire, ISE

Bertrand Decharme, CETP

Catherine Ottle, CETP


EXTRA….


2 Database

Land Parameters: ECOCLIMAP

• Single annual cycle (decad)

• 12 classes 255 types

• Effective or multi-tile schemes

• 1 km, global

Applications:

• used at Météo-France

• to be used at ECMWF

(0.5 deg.)


• Determination of baseline or « minimum » physics (eg. Stomatal resistance)• Importance of sub-grid hydrology• Importance of soil moisture (on surface energy balance: HAPEX MOBILHY)• Better numerical schemes• Chance for extensive validation• Ability of schemes to simulate carbon cycle • Innovations in Sub-grid cold-season processes• Impact of changing scale• Ability of such models to simulate regional / large scale hydrology


2 Database

Atmospheric Forcing:

Data for running an LSM in « offline » mode

Spatial scales for the atmospheric forcing datasets


2 Database

ECOCLIMAP Soil, vegetation 1 km 0.5 deg (monthly climatology)

AMMA-SAT/PRECIP

Precip Probability 3km 0.05 deg, 15min

2004, 2005 (+?)

OSI-SAF SWdown, LWdown 0.10 deg., 3h 2004

LAND-SAF SWdown, LWdown, TSurf

0.05 deg., 30 min 2005+

ECMWF Tair, Qair, Wind, PSurf, Rainf, LWdown, SWdown

0.50 deg., 3h 2001+

AMMA EOP/SOP Fluxes, soil moisture, vegetation (biomass, LAI…)

local 2004 Mali, Dahra, Niger

(2005+)

AMSR Soil moisture (surface)

25 km 2004…

LAND-SAF LST 0.05 deg., 30 min 2005+

AMMA-SAT MODIS LAI 0.01 deg, monthly 2003+

Forcing Data:

Evaluation Data:


ORCHIDEE LMD, Paris, France J. Polcher, T. Orgeval

ISBA CNRM / Météo France Toulouse

A. Boone

TESSEL ECMWF, Reading, UK A. Beljaars

JULES CEH, Wallingford, UK C. Taylor, P. Harris

SETHYS CETP, France C. Ottlé, B. Decharme

NSIPP UPMC, France A. Ducharne

IBIS* ISE, Monpellier, France C. Delire

SSib U. Nantes, France, UCLA, USA

I. Poccard-Leclercq, Y. Xue

MIKE-SHE U. Copenhagen A. Norgaard, I. Sandholt

NOAH (NCEP, USA) CETP, France C. Ottlé, B. Decharme

Model acronym Institute Contact

* Contact us if you wish to participate…

3 ALMIP description


2 Database

Vadidation/Evaluation of atmospheric forcing?

Difference: July-June AMSR-E• Microwave sensor

• « Surface » soil moisture• 50 km, 2 passes daily• Flagged data below 5-10 N• Difference: July – June (averages)


International Land Surface Model Intercomparison Project – Dirmeyer et al.

• 15 LSMs (NWP, GCM, Hydrology…)

• Control Exp. Forcing:

- ISLSCP parameters

- NCEP – hybridized, gauge-corrections

• Global scale, 1 deg resolution

• Forcing: 3h, 1982-1995

• Evapotranspiration (monthly: mm day-1)

• 10-year average: 1986-1995

• Significant differences!!!

GSWP

Global Soil

Wetness Project 2

3 ALMIP description


Long term Observation Period (LOP)

2001 2005 2006 2007

Enhanced Observing Period (EOP)

0

Special Observing Periods (SOP)

2 31

2010

ECMWF-FC Merged ECMWF-FC, AMMA SAT/PRECIP, EOP-SOPForcing

Parameters

2003

ECOCLIMAP, MODIS

On going: - AMMA LSM DB for 2001-2005--- (CNRM)

NCC-DB

(LMD)

1948-2000:

- Analysis of the models sensitivity to the forcing NCC-DB (LMD)

1 Overview


1 Overview

• Influence of soil moisture « memory » on the atmosphere

• Improve regional scale hydrological processes

(sub-grid parameterizations, infiltration, flood planes)

• Determine key processes vegetation functioning (African ecosystems)

• Extend vegetation processes to the regional scale

• Use a multi-scale approach

• Develop a surface flux and soil moisture « climatology » used to study coupling

with atmosphere, improve assimilation systems…

• Intercomparison Project (ALMIP) over region (CESBIO and CNRM)


Equation 2

EPSAT-SG(Estimation des Pluies par SATellite – Seconde Génération)

AT

r

AT

gpcp

ApdadttaP

dadttaI

dcI22

11

,

,

,

Equation 1

Neural NetworkMSG Channels

SRTM Digital Elevation Model

Rainfall probability images (Pr) GPCP1dd rainfall images (Igpcp)

Potential rainfall intensity images (Ip)

EPSAT-SG rainfall estimation (Ie)

A is a disc of about 125kms radiuscA is the centre of Ad is the considered dayT is the period [d-15days,d+15days]dt1 corresponds to 1 daydt2 corresponds to 15 minuteda corresponds to 1 MSG pixel

Equation 1

daItaPtaI pre ,,,

Equation 2Estimated Rainfall Intensity at time t during day d and position a:

Final product resolutions:Space resolution : 3 kms

Time resolution : 15 minutes

2A25 TRMM precipitationRadar data


ECMWF Forecast Precipitation: Hovmoller plots


GSWP

• 15 LSSs, 1 deg. res.

• Evap (average 10 years, -20E to 30 E longitude)

• Inter-LSS Meridional gradient highly variable

3 ALMIP description


• Simulated Offline fluxes: EvapotranspirationECMWF forcing

Merged forcing

Evap: Merged - ECMWF

4 Some Results


ECOCLIMAP Soil, vegetation 1 km 0.5 deg (climatology: decad)

AMMA-SAT/PRECIP Precipitation 3km 0.05 deg, 15min 2004, 2005 (+?)

OSI-SAF SWdown, LWdown 0.10 deg., 3h 2004

LAND-SAF SWdown, LWdown, TSurf 0.05 deg., 30 min 2005+

ECMWF Tair, Qair, Wind, PSurf, Rainf, LWdown, SWdown

0.50 deg., 3h 2001+

AMMA EOP/SOP Fluxes, soil moisture, vegetation (biomass, LAI…)

local 2004 Mali, Dahra, Niger

(2005+)

AMSR Soil moisture (surface) 25 km 2004…

LAND-SAF LST 0.05 deg., 30 min 2005+

AMMA-SAT MODIS LAI 0.01 deg, monthly 2003+

Forcing:

Evaluation:


2 Database

Time (DoY)

Coherence between OSI-SAF and AMMA-SAT data

e.g. Time series:

• Summer 2004

• Downwelling radiative fluxes

• precipitation

Merged AMMA-SAT, SAF and ECMWF Forcing :


Gourma, Mali

Oueme, Benin

Niamey, Niger

CATCH window

Meso-Super Sites:

Regional Scale Domain (0.5deg):



Atmospheric Forcing

ISBA-Ags

Biomass (LAI)


Runoff Irrigation

Interactive mode

MODIS (2005) LAI Exp1 2005 LAI Exp2 2005

Pashtuchak, Peyrillé, Plante, and Boone

1 cnrs, centre national de recherches météorologiques (cnrm / météo-france, toulouse)

Documents