The central American mid-summer drought
(MSD, veranillo or canicula)
Brian Mapes
University of Miami
midsummer drought
30-year climatology
Key WestKey West
Miami
MSD
1999
(
?
What causes the MSD?• Local, fussy:
– The ingredients for convection get worse (moisture, instability, lifting mechanisms).
• Global, vague: – Solar declination angle ultimately drives all
aspects of the seasonal cycle.– Continent-ocean contrasts involved
• Somewhere in between lies a usefuluseful truth...A
What causes the MSD?
• Regional: – North Atlantic Subtropical High (NASH or
Bermuda High) & tradewinds are enhanced and extended west in mid-summer.
late May-early June
July
Midsummer nose on N. Atlantic Subtropical High (NASH)
dry
Carib easterly jet
VeracruzVeracruz
ValparaisoValparaiso
wet
wet
CaymanCayman
upwind upwind slopesslopeswetter wetter inin midsummermidsummer
midsummer drought
Key WestKey West
Miami
Valparaiso, FL
CaymanCaymanMSD broader in
time than VeracruzVeracruz
CMAP
Individual years - still discernable
(in big regional Central American box average)
intraseasonal canmask it a bit
9 year mean
CMAP rainfall
TAIWAN 25N
CUBA 25N
SE BRAZIL 25S
MADAGASCAR 25S
May
Nov.
Nov.
May
MSD happens in other places too
Where are those boxes? W. subtropical oceans
High HighHighHigh
What causes the MSD?• Regional:
– NASH (Bermuda High) & tradewinds are enhanced and extended west in mid-summer.
• Locally, less rain is due to:– frictional divergence or other subsidence mech.?
– flow direction brings drier airmass?
– wind leads to lower SST (albeit via enhanced LHF)...?
– capping inversion strength/height?
– different meso/synoptic disturbance characteristics?
• But what causes the NASH/trades to vary?
A broader view of the summertime oceanic subtropical highs
• Meteorology 101:– Hot air rises, cool air sinks.
• Hydrostatic sinking is all about PGF to push horizontal branch of circulation (divergence)
• e.g. Hadley cell: eq. trough, subtropical H belt
• In summer, subtrop. oceans cooler than land
=>Highs over the ocean basins=>Highs over the ocean basins
Annual mean SLP
July SLP (same color scale)
January SLP (same color scale)
coolingcooling coolingcooling
• Meteorology 301: stratified
– Heated air rises, cooled air tends to sink. • heating rate = rate of flow across isentropes = vertical
velocity (in coordinates)
– E. subtropical oceans: cooling (even in winter)• little latent heat release (rain), much rad. cooling (stratocu
decks under dry upper troposphere)
“The Usual Explanation”
• Hoskins (1996): • “the usual explanation for the subtropical
anticyclones (radiative cooling)” in summer is “inadequate”
• “monsoon latent heat release over the neighboring continent” is essential
2001
Model experiments, forced by realistic heating (above), in realistic JJA zonal mean flow, with real topography
Rodwell and Hoskins 2001
Asian Heating Q_Asia
887 (Q_Asia, mts)
Pure heating effect: 887 (Q_Asia, mts)-887 (mts only)
RossbyRossby
KelvinKelvin
??
Try NAm and NPac Heating:
887 forced by mountains + NAm Q
887 forced by mtns + NAm Q + NPac Q
Rossby
Kelvin
??
Red: global Q
Small et al. 2007 concur: imposed heating anomalies in the region can explain most of the flow anomalies
• some westward influence from E. Atl
• MSD negative heating drives NASH/trade enhancement
• positive feedback
positive feedback
• Easterlies / High drive (negative) rain anomalies
• (negative) latent heating anomalies drive easteries/ High
• An amplified, twitchy system - but what drives it, to make this climatological feature?
One hypothesis:
MSD
wet midsummer tradewind surge
One hypothesis:
Drivers: 2. Mean flow changes in midsummer? Chen, Hoerling & Dole 2001
Heating
Eddy Z1000
w/o shear
Eddy Z1000
July u(y,p)
u300, zonal mean Jul-Aug time slice
Westerlies retreat to >30N in midsummerQuickTime™ and aTIFF (LZW) decompressorare needed to see this picture.
<0
time slice
easterlies protrudeto 30N suddenlyin mid summer
WHY, in terms of [u] budget? • Not f[v]: ~barotropic; [v](t) wrong
• [u’v’]: Tilted TUTTs, Tibetan High, Transients?
Eq
60N
J F M A M J J A S O N D
Drivers: 3. Influences via the midlatitude westerlies?
west coast thermal gradient
Hudson Baycontributes too
2006
Japan’s midsummer
drought mechanism: the
Bonin High
“Formation mechanism of the Bonin High in
August”(Enomoto, Hoskins, & Matsuda 2003)
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
positive feedback• An amplified, twitchy system - what drives it, to make
this climatological feature?
1. Local SST cycle hypothesis?– not well supported by obs
2. Global easterlies in midsummer?– but where do they come from?
3. Midlatitude wave pathways?
4. Asian monsoon? – via Rodwell-Hoskins’ wraparound Kelvin wave?
Clues from climate models• Do free-running, full-physics climate models
have the MSD?– Short answer: yes
• What drives it there?– Global, vague:
• solar declination, like all aspects of seasons• continent-ocean contrasts• Experiment!
Observations
IPRC/ ECHAM model:
quite good
SLP
IPRC GCM rain
July Aug
CMAP rain
GCM SLP
1
5
2
8
Experiment: global monsoon system on steroids
• Increase continent-ocean contrasts– perpetual June (and every other month):
• fix SST and solar decl. for each day of year• run model for many days, so atm+land come
into equilibrium
– Warmest NH now in June, not July– Continents hotter in summer (dried out?)
– What happens to monsoons (and MSD)?
Northern continents get
hotter in summer
Dec
Jan
Our MSD is a major part of the
global rainfall
response
Annual mean rainfall changeAnnual mean rainfall change
IndianIndianOceanOceanwetterwetter
early onsetearly onset
AmericasAmericasdrierdrier
in MSDin MSD
Dec
Jan
Conclusions (so far)
• The Central American MSD is a significant aspect of the seasonal cycle, both regionally and globally.
• Useful for understanding phys. mechanisms– how exactly does the NASH affect rainfall/convection
• U. for u. summer teleconnections– from Asia, zonal [u] easterlies, waves on westerlies...
• A “natural mode” of the regional system - relevant to other types of S-I climate variability?
Future work
• Starting fall 07 in earnest (thanks NSF)
• Atlantic summer climate reading group– me, 2-3 students on project, other experts
(please!)
• where to start?
software demo
HF anomalies (total - annual - semiannual):
Yes, the Bonin High and 2 other stationary waves are evident, but also a zonally elongated u
anomaly spanning all Asia-WPAC
time slice
0 m/s 2-4 -2
u200 HF anomaly plot, July-Aug
HF wavelet analysis
100d
10d
total HF
Japan
Jul Aug
Interannual consistency of the Asian HF jet anomaly ~1 Aug
Low-level heating dominant. Influences from monsoons and planetary waves, but mainly local
feedbacks. Eastern-basin oriented.
ISM
WNPM
EASM
May 1 Jun 1 Jul 1
(2nd active)
EASMNAM
ENASM?
May 1 Jun 1 Jul 1