Title: Long-lasting locally coupled anomalies
1Long-lasting locally coupled anomalies
EMC Seminar Thurs. 29 Jan 2004
M. Peña1, E. Kalnay1 and M. Cai2 1University of
Maryland 2 Florida State University
2Content
- Phase relationship of anomalies (e.g., cyclonic
over warm) and forcing direction. - Preferential phase relationship of long-lasting
Ocean-Atmos and Land-Atmos coupled anomalies. - Consequences of assuming ocean-driving (AMIP)
scenarios to the simulation and prediction of
coupled anomalies.
3Definitions
- Anomaly departure from the annual cycle that
exceeds a threshold. The threshold used is one
fourth of the local standard deviation. - Duration of an anomaly time interval that the
variable continuously exceeded the threshold. - Locally Coupled anomalies coincide spatially and
temporally in the two media.
4Anomalies over the ocean
Zonally averaged number of anomalies
colors denote the number of cases
5Coupled anomalies
colors denote the number of cases
6Simplified Mo and Kalnay Dynamical Rule
Diagnosing the driving direction
7Driving direction
NCEP Reanalysis
Atmosphere-driving anomaly cases
Ocean-driving anomaly cases
Anomalies lasting at least 15 days
Atmosphere-driving cases dominate the
extratropics!
8Driving direction
ECMWF Reanalysis (14 years)
Atmosphere-driving anomaly cases
Ocean-driving anomaly cases
Anomalies lasting at least 15 days
9Correlation SST and 850 hPa CV
Extratropics Atmos leads Ocean
10Leading mode of covariability
North Pacific
Slide 10
500 hPa height
SST
40ºN
Deser and Timlin 1997
11North Pacific Basin-wide modes of covariability
Slide 9
Atmos-driving Phase relationship
12Diagnostic rule
- The geographic distribution of the forcing
direction obtained from the diagnostic rule is - Independent on the reanalysis data set used
- Very similar in the daily, five-day average, and
monthly data (no shown). - Consistent with Lag-lead correlation technique
- Consistent with basin-wide modes of variability
- Knowing the simultaneous phase relationship in a
given time we can estimate the local forcing
direction.
13Normal and Abnormal ocean-atmos coupling
Normal Coupling
In the extratropics anomalies with cyclonic over
cold and anticyclonic over warm relationship
tend to last longer. In the tropics cyclonic
over warm and anticyclonic over cold tend to
last longer.
Abnormally coupled anomalies have shorter
duration
14AMIP one-way interaction scheme
- Atmospheric GCM run with prescribed observed SST
(AMIP runs) are usually assumed to be the upper
limit for potential predictive skill (perfect
SST). - However, they assume (incorrectly) that the ocean
always forces the atmosphere.
15AMIP simulation of anomalies
Ignoring the atmospheres feedback leads to
Number of cases of AMIP simulated minus Reanalysis
Fewer long-lasting simulated anomalies than
observed in the extratropics
More long-lasting simulated anomalies in the
tropics
16AMIP simulation of anomalies
Reanalysis minus AMIP cases
AMIP ignores these cases in the extratropics
17Normal coupling and AMIP run
- Atmosphere-driving in the extratropics and
ocean-driving in the tropics constitute the
normal coupling. - There are abnormal coupled anomalies
(atmosphere-driving in the tropics and
ocean-driving in the extratropics), but they are
invariably short-lived. - As a result, in AMIP runs (where the ocean is
always forcing the atmosphere) there are too many
long lasting anomalies in the tropics, and too
few in the extratropics
18Analysis of Medium Range Forecasts
- Skill of prediction of the 500hPa height.
- CDCs Reforecast data
- Daily 21-forecast runs with a frozen version of
the NCEP MRF. - Initialized with the Reanalysis data.
- SST anomalies at the initial time are kept
constant through the run. - Anomaly Correlation Forecast (ACF) measures how
close the forecast pattern verifies with the
observed pattern in a given region.
19Measuring Skill and Forcing direction
Days of Reliable Forecast Interval from the
Integrations Initial Time to the time when ACF
0.5
Fraction area of Ocean-driving phase
relationship diagnosed (in the Reanalysis) prior
to the Initial Time.
20MRF and Forcing direction
Northern Pacific
21500 hPa composites
Atmos-driving
Ocean-driving
H
L
L
H
L
H
PNA pattern
- PNA Pattern
22Skill of AMIP simulation
- Measure the error of simulating high amplitude
anomalies wrt reanalysis in the monthly data. - Coupling index, CI
- CI -1 if all the grid-points in a given area
experience atmosphere-driving - CI1 if all the grid-points experience
ocean-driving
23Skill of AMIP simulation
AMIP error
In the extratropics, errors in AMIP are larger in
atmosphere-driving situations In the tropics,
errors are larger in ocean-driving
24Skill of AMIP simulation
Cross-correlation CI and simulation error
Regions where atmosphere-driving situations
hurt the skill of simulation
25Skill and forcing direction in Northern Pacific
- When atmosphere-driving situations dominate
- ? decrease in skill of simulation/prediction of
AMIP-like models. - ? increase in skill of persistence forecast
- The opposite happens when the ocean-driving
situations dominate. - Predominant atmosphere-driving or
ocean-driving situations resemble a negative or
positive PNA pattern, suggesting local
extratropical coupling is an important root of
the PNA.
26Land-Atmosphere coupling
Distribution of 15-days or longer-lasting
anomalies
Long lasting anomalies have a preferential local
phase relationship between vorticity and surface
temperature
27Local phase relationship and Soil Moisture
Relationship Low-level circulation-Skin
Temperature
Soil Moisture
28Cross-correlation 850hPa CV and ST
Monthly
Daily
29Energy- and moisture-limited regions
Considering the ST-Precipitation feedback The
vegetative and wet region of the SE US requires
large amounts of energy to induce significant
changes in ST. The very dry region of the SW US
requires large amounts of moisture to produce
changes in the precipitation regime.
30Regional differences over the U.S.
31Land-Atmosphere coupling
- There is a preferential phase relationship of
long-lasting anomalies over the continents. - Over the very wet and very dry regions, the
normal coupling is characterized by cyclonic
over cold anticyclonic over warm. - Over regions to the east of high range mountains,
such as the Rockies, the normal coupling is
cyclonic over warm and anticyclonic over cold.
32Thanks !