MOCA-09 Montr

1
MOCA-09 Montréal, Canada

• Snow-monsoon teleconnections testing competing
  mechanisms using idealized snow forcing
  conditions in a GCM
• Andrew Turner1 Julia Slingo2,1
• 1NCAS-Climate, University of Reading, UK
• 2Met Office, Exeter, UK

2
Historical perspective

• Long history of using Himalayan snow to forecast
  ISM rainfall, dating back to Blanford (1884).
• More recent work examines the influence of snow
  from Himalaya, but also West/North Eurasia, both
  with negative teleconnections.

• Difficulties involve ENSO, snow measure used,
  region of influence
• Himalaya how can such a small region perturb the
  monsoon?
• West/North Eurasia what is the remote mechanism?

3
HadCM3 snow composite difference evolutions
(1050yr control run)
Composite evolutions generated from heavy minus
light snow years followed by weak minus strong
monsoon rainfall under neutral ENSO conditions
MAM snow (kg/m2)
MAM snow (kg/m2)
right using west Eurasia snow index
left using Himalaya snow index
Apr 1.5m temp Z500
Apr 1.5m temp Z500
May 200hPa wind
May 200hPa wind
In HadCM3, weak monsoon summers can be preceded
by heavy Eurasian OR Himalayan snow.
JJA 850hPa wind rain
JJA 850hPa wind rain
4
Aims of this study

• Can the Hadley Centre model simulate snow-monsoon
  teleconnections?
• Which region dominates?
• Can we make a detailed assessment of the
  mechanisms involved?

5
AGCM ensemble experiment design

• Hadley Centre Land-atmosphere model HadAM3
  (3.75x2.5xL30)

• Snow forcing derived from the HadCM3 coupled run
  using climatology with 2s anomalies in FMA snow
  indices over
• Eurasia (30-110E, 50-65N) WNEur
• Himalaya (67.5-100E, 27.5-40N) HimTP
• Climatological SST forcing (to avoid ENSO).

• Experiments initialized 1Nov, 6 month spin-up
  with snow depth updated hourly to chosen forcing.
• 32 member ensemble begun 1Apr for 8 months
  initial conditions from the 15Mar16Apr period.
• Snow no longer constrained free to melt.

6
HimTP ensemble results
April
May
June
Snow amount
These diagrams show ensemble mean differences
between HimTPpos and HimTPneg experiments.
7
HimTP ensemble results

• Himalaya AGCM ensemble results consistent with
  coupled run composites.
• Strong Himalaya snow forcing ? weakened Indian
  monsoon (June).

June
July
August
precip
8
WNEur ensemble results
May
April
June
Snow amount
These diagrams show ensemble mean differences
between WNEurpos and WNEurneg experiments.
9
WNEur ensemble results

• WNEur results contrary to coupled run composites.
• Strong Eurasia snow forcing ? strong Indian
  monsoon (June) due to contamination from induced
  Himalaya anomaly.

HimTP
June
July
August
WNEur

• Both exp. support the Blanford hypothesis.

10
HimTP sensitivity tests

• To test mechanism, use sensitivity tests over
  HimTP
• HimTP1000 (1000kg/m2 snow 4m)
• HimTPzero (0kg/m2)

• Qualitative agreement with standard HimTP
  experiment but larger in magnitude.
• Significant weakening of the early Indian
  monsoon.
• Redistribution of East Asian monsoon rainfall.

11
Effect on tropospheric temperature (600-200hPa,
40-100E)
Measure of reversal of meridional TT gradient
(Xavier et al., 2007) 600-200hPa mass weighted
temperature over 40-100E, difference between
5-35N and 15S-5N regions

• Heavy snow forcing over Himalaya/TP cools
  mid-troposphere 15-40N at Indian longitudes by
  around 3C.

Difference in character between ENSO and snow
effects on the monsoon growing vs. decaying modes
12
The Blanford mechanism in HimTP
13
The effect of snow albedo

• Effect of snow albedo as part of the Blanford
  mechanism is tested in a further version of
  HimTP1000.
• HimTP1000sfa snow albedo over HimTP set to
  snow-free value for that region.
• Albedo 0.67 ? 0.20 averaged over HimTP.

14
The effect of snow albedo
? Downward
15
The effect of snow albedo

• Much more rapid snow melt, reduced reflected
  shortwave, sensible heating over HimTP reduced by
  50 compared to HimTP1000 ensemble.
• Reduced tropospheric cooling compared to
  HimTP1000. Remaining cooling caused by reduced
  upward longwave (insulating effect of snow).

16
The effect of thermodynamic ocean coupling

• Testing the impact of air-sea interaction without
  ENSO present.
• HimTP1000 and HimTPzero ensembles repeated in
  HadAM3 coupled to a 50m mixed layer ocean.

• Main impact of HimTP snow seems delayed until
  July, perhaps related to feedbacks with Indian
  Ocean and developing negative dipole.

17
Summary

• Coupled model (HadCM3) can simulate weak Indian
  monsoon following strong spring snow forcing in
  either HimTP or WNEur regions.
• Tests with HadAM3 AGCM show that HimTP is
  dominant in this model, supporting the Blanford
  hypothesis.
• Snow albedo plays a crucial role.
• Model bias may inhibit teleconnection from
  further north.
• Similar teleconnection in coupled (mixed-layer)
  model air-sea feedbacks need further
  exploration.

18
Thank you!

• a.g.turner_at_reading.ac.uk
• Please see my poster this afternoon
• Uncertainties in future projections of extreme
  precipitation in the Asian monsoon regions