Title: Mechanisms of land-atmosphere in the Sahel
1Mechanisms of land-atmosphere in the Sahel
- Christopher Taylor
- Centre for Ecology and Hydrology, Wallingford,
U.K. - Richard Ellis, Phil Harris (CEH)
- Doug Parker (Leeds)
2Outline
- Soil moisture - rainfall feedbacks on daily
timescales - Satellite analysis
- Aircraft observations (AMMA)
- A dry case
- A wet case
3Soil moisture rainfall feedbacks
Shows where climate models sensitive to soil
moisture Large coupling strength implies soil
moisture has significant impact on precipitation
i.e. feedback possible Large variations between
models - models dont represent basic
processes well. Do we have observations to
judge models by? Focus on West African hotspot
Koster et al, Science 2004
4How strong should coupling be?
- What are mechanisms?
- Are our parameterisations suitable?
5Daily Variability in Surface Fluxes in Sahel
- Evaporation limited by soil moisture so fluxes
very sensitive to rainfall - For several days after rain
- large evaporation rates direct from soil
- low sensible heat flux
- low surface temperature
Observations from savanna site at the start of
the 1990 wet season (Gash et al)
6Does daily surface variability matter in a GCM?
Power spectra of simulated rainfall in HadAM3
Variations in surface fluxes on short timescales
feed-back on simulated rainfall.
Taylor and Clark, QJRMS (2001)
7Impact of soil moisture on afternoon convection
12 June 2000 2215
13 June
Meteosat 7 TIR
Polarisation ratio TMI
Wet soil
In this single case, extent of convective system
influenced by soil moisture Convection avoids
wet soil
8Results from 108 cases
- Over 50 cases similar to example shown
- 33 less cloud over wet soil than nearby drier
zones - Initiation over wet soil strongly suppressed (2
cases) - Suggests a negative soil moisture precipitation
feedback for initiating storms (cf Taylor and
Lebel 1998) - Potential mechanisms?
Cold cloud extent 13 June
Taylor and Ellis, GRL 2006
9Aircraft ObservationsAfrican Monsoon
Multidisciplinary Analyses
Special Observing Period during 2006 Wet
Season Focussed observations at multiple ground
sites and with 5 aircraft, including NERC/Met
Office BAe146 5 week deployment in Niamey, Niger
10A dry case study 1 August 2006
1700 UTC 31 July
12Z Aug 1
Meteosat thermal infra-red
Niamey
Initiating storm
11Global View
12Flight over storm track 18 hours later
1000 km
Storm track
Flight track
Polarisation ratio anomalies from TRMM Spatial
resolution 50 km
13Land Surface Temperature Anomalies
Extract mean diurnal cycle to obtain Land Surface
Temperature Anomaly (LSTA)
500 km
Cold (wet)
Warm (dry)
White no data (cloud or river)
14(No Transcript)
15Aircraft data within planetary boundary layer
(PBL)
Land surface temperature anomaly (satellite)
Wettest soils
Generally very good correlation between satellite
surface data and PBL at fine scale weak heating
from wet soilgtcool PBL
16Aircraft data within planetary boundary layer
(PBL)
Similar story for specific humidity High values
above wet surface
17Vertical profile data (dropsondes)
Wet soil
Pressure
Dry soil
PBL twice as deep over dry soil as wet, and
markedly drier and cooler. More inhibition to
convection over wet soil. In fact, no significant
convection on this afternoon along track.
18An impact on low level winds?
If surface heating contrasts large enough, might
expect a sea-breeze type response i.e.
convergence over dry (hot) surfaces
So surface gradients ARE strong enough to induce
circulations.
19Low level wind vectors
Land surface temperature anomaly
Analysis suggests that soil moisture patterns
strong enough to induce sea-breeze type
circulations. Can they cause further storms on
more favourable days?
20A wet case study 31 July 2006
wet
dry
wet
Had similar flight planned previous
afternoon Very dry surface bounded by wet areas
21Storm initiation during flight
Aircraft track
System developed very rapidly over dry soil as we
approached.
22Storm initiation
Clouds over dry soil
Due to convective inhibition or convergent winds?
23Early evolution of storm
Shading land surface temperature
(reddry) Contours cloud from visible channel
Storm develops along wet-dry surface
contrast Signature of triggering by circulation
rather than thermodynamic profiles
24Current work in AMMA
- Quantifying surface fluxes (ALMIP)
- Best available met forcing
- Surface flux obs to calibrate models
- Assimilation of LST data
- Feedbacks on convective initiation
- Role of circulations and/or thermodynamic
profiles - MCS feedbacks
- Sign and strength of feedback
- Key space scales
- Intraseasonal feedbacks
- Wet/dry spells
- Interannual memory
- vegetation
- Observational diagnostics to test atmospheric
models
25Hombori Tondo (Mali) from UK BAe146. Photo Doug
Parker
26(No Transcript)
27(No Transcript)
28Soil moisture and monsoon dynamics
- Intraseasonal variability in West African
rainfall - Large-scale wetting/drying 15 day cycle
- Cause and effect?
Atmospheric warming
Satellite soil moisture
Surface heating (W/m2)
T 925hPa (ECMWF)
29Cause and effect lagged relationships
Composite data based on surface wetting
TMI wetness
ERA40 Temperature anomalies
Satellite cold cloud
Additional daytime cooling at 925hPa day 0 and
day 1 - shows soil moisture leads to cooling in
ECMWF analyses
30Wet v Dry Spells
Shading surface heating Contours 925hPa
Temperature
- During wet spells, cool high develops across
Sahel - Dynamic response to soil moisture consistent with
forcing of variability - Studentship with UEA looking at feedbacks in GCM
31Convective scale feedbacks
Rain gauge data from HAPEX-Sahel
- From observations, found tendency of rain within
squall lines to be heavier in locations that have
been recently wetted - Linked to a positive feedback between soil
moisture and rainfall at scales of only 10 - 15
km (Taylor and Lebel, MWR 1998)
32Modelling Impact of Moisture Anomalies on
Convection
Used cloud-resolving model (RAMS) to assess
impact of humidity on cloud-scale dynamics within
squall line. Run large ensembles.
Unexpected sensitivity of feedbacks to length
scale, convection sensitive to fine scale
variability (Clark et al 2003 QJRMS, 2004 JHMet)
33Synoptic Scale Surface Variability
Warm
Cool
Screened TIR anomalies are well-organised at
large scale (1-2000 km) in N. Sahel
34Synoptic Scale Surface Variability
- Alternate warm (dry) and cool (wet) surface
anomalies travel westwards across the Sahel
Black lines cold cloud
Cool surface features appear after rain
Day
Longitude
35Impact of Synoptic Surface Variability on
Atmosphere?
1000 km
Produced composite hotspot from 2000 wet season
to assess feedback of surface on atmosphere.
Southerlies
Observational analyses suggest
higher atmospheric temperatures lower surface
pressure
Anomaly
TIR C
vortex develops
subsequent cold cloud (rainfall) modulated
Northerlies
Degrees longitude
Taylor et al QJRMS 2005
36Identifying Wet Soil From Satellite
- Several possibilities for detecting soil moisture
from space - Passive microwave (10.65 GHz) from TRMM Microwave
Imager to infer wet soil (high evaporation) after
recent rain
Rainfall (bars) and TRMM polarisation ratio
(asterisks) in Banizoumbou region (Niger)
Soil drying after rain
Rainfall data courtesy of T. Lebel (IRD)
37Thermal Data
Meteosat Second Generation provides data every 15
mins at high spatial resolution (3 km) Land
surface temperature products produced by LandSAF
in near real time