Steven B. Feldstein

Department of Meteorology, The Pennsylvania State University, University Park, Pennsylvania, U.S.A.

Presented at Tel Aviv University, Department of Zoology, Tel Aviv, Israel on May 12, 2010

Atmospheric teleconnections, bird Atmospheric teleconnections, bird migration, and implications for migration, and implications for

probabilistic forecasts of bird strikesprobabilistic forecasts of bird strikes

PREDICTABILITY

Three time scales associated with atmospheric predictability

Deterministic Predictability (Weather) (Useful 3-5 day numerical model forecasts) Numerous studies show a linkage between weather, i.e., storms, precipitation, fronts, etc., and bird migration. Van Belle et al. (2007) 3-day forecast of bird

migration intensity.

Extended-Range Predictability (1 week to 1 month timescale) Predictability mostly poor (because of chaos), except perhaps when teleconnction patterns are excited. Linkage to bird migration associated with storms that accompany teleconnection patterns (Elkins (2008).

Monthly and Seasonal Predictability (Climate, > 1month, average of weather) (Closely linked to seasonal cycle and boundary forcing) (Useful ensemble forecasts.) Linkage between El Nino/Southern Oscillation and bird migration (Hameed et al. 2009).

El Nino/Southern Oscillation and Bird Migration at Attu, Alaska

From 1980-2000, each year, about 100 birders visited Attu, Island to add Asian bird species to the North American bird lists Island, Alaska

to add Asian birds to their North American bird

Common Sandpiper Long-toed Stint Rustic Bunting

AIPD = Asian Individuals Per Day (North American species excluded from total) (Hameed et al. 2009)

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.QuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

El Nino/Southern Oscillation

ENSO, storms, and bird migration

ENSO affects

bird migration

through its

influence on the

latitude of the

jet and the

storms that follow

the jet, i.e. ENSO

alters the

environment

through which

storms propagate.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

The dominant Northern Hemisphere teleconnection patterns

North Atlantic Oscillation Pacific/North American pattern

Teleconnections evolve on a 7-10 day time scale (longer than weather time scale).

They alter storm path and storm intensity.Climate Prediction Center

Siberian Vagrants and the PNA Siberian Vagrants and the PNA Teleconnection PatternTeleconnection Pattern

Records of all observations (locations and dates in California) of Dusky Warbler, White Wagtail, Red-throated Pipit and other Siberian vagrants

Composite PNA index

Dusky Warbler (PNA=- 0.42, n=15, p<0.1)

White Wagtail (PNA=-0.26, n=22, p<0.02)

Red-throated Pipit (PNA=-0.38, n=79, p<0.01)

(Steven Feldstein, Peter Pyle, Steve Mlodinow, Richard Erickson, Jim Tietz)

Anomalous wind vectors associated with Dusky Warbler records in CaliforniaAnomalous wind vectors associated with Dusky Warbler records in California

Lag 0 DaysLag -1 Days

Lag -2 DaysLag -3 Days

Nearctic Vagrants and the North Atlantic OscillationNearctic Vagrants and the North Atlantic Oscillation

(Elkins 2008, British Birds)

Circumglobal Teleconnection Pattern

Time-averaged

V over

persistent event

(lag -6 to lag +9 days)

Correlation with EOF1 =0.83 Correlation with EOF1 =-0.72

Wave packets associated with SL precip

300

wet dry

EOF1

300-hPa geopotential height evolution - Middle Eastern precipitation

Feldstein and Dayan (2008)

-6 days -4 days -2 days

0 days +2 days +4 days

+5 days +7days +9 days

Composite

analysisEvolution of

300-hPa height

field determines

the wind, T,

T, P, P, and

rainfall,

variables which

influence bird

migration.

Wave packet

first seen over

NE Pacific.

Wave field

persists for 2

weeks. This

slow evolution

may allow for

a 7-day

weather

forecast for

Israel?

Israeli Migrants and the Circumglobal Teleconnection Israeli Migrants and the Circumglobal Teleconnection Pattern (CTP)?Pattern (CTP)?

QUESTIONS: Does the CTP influence bird migration over Israel?

Can the CTP be used to forecast the bird migration intensity or bird strike frequency with a

1-7 day lead time (fall season)? Beyond 3-4 days, is a forecast of bird migration

intensity with a probabilistic model based upon the CTP better than that based upon a deterministic linear regression model (e.g., Van Belle et al. 2007)?

16 North Pacific sea level pressure cluster patterns16 North Pacific sea level pressure cluster patterns

Example of

cluster

analysis

Tropical Convection Associated with Tropical Convection Associated with the Madden-Julian Oscillation (MJO)the Madden-Julian Oscillation (MJO)

Phase 1

Phase 2

Phase 3Phase 4

Phase 5

Phase 6

Phase 7

Phase 8

Time between Phases ~ 6 days

180 ۫° 60 ۫°W20 ۫°E

Dominant intraseaonal oscillation in the tropics

MJO cycle: 30-60 days

Shading OLR

Time between phases ~ 6 days

From Wheeler and Hendon (2004)

From Wheeler and Hendon (2004)

Frequency of occurrence for each cluster pattern and MJO phaseFrequency of occurrence for each cluster pattern and MJO phase

1-7 day Forecast of Anomalous Bird Migration intensity in Israel1-7 day Forecast of Anomalous Bird Migration intensity in Israel

Phase Number = location in Israel

Pattern Number = cluster pattern

Lag = 1 to 7 days (Feldstein

and Dayan 2008)

Color denotes anomalous

bird migration intensity

determined from composites of

the daily bird migration intensity

for each pattern number

Cluster analysis with 300-hPa meridional wind (1-7 day variability (CTP) dominated by the upper tropospheric flow and it

also determines the lower tropospheric flow where birds are observed)

(Cluster patterns represent slowly-evolving component of flow, small number of patterns with large spatial scales)

Analysis performed at separate locations (airports, radar stations, etc.) in Israel

Analysis can be performed separately for soaring (raptors) and powered flight (shorebirds) migrants

Conditional probabilities based on the accumulation of migrants during previous days (bad weather)

1-7 Day Probabilistic Bird Migration Intensity Forecast

Seasonal Forecast of Anomalous Bird Migration Seasonal Forecast of Anomalous Bird Migration Intensity in IsraelIntensity in Israel

mean is the anomalous seasonal meridional wind

is cluster pattern c

is the frequency of cluster pattern c

Obtain forecast of seasonal mean meridional wind

Determine the cluster pattern which has the largest projection onto

Estimate seasonal mean bird migration intensity (above average, average, below

average) in terms of the most frequently occurring cluster pattern.

Combined Probabilistic/Deterministic Bird Combined Probabilistic/Deterministic Bird Migration Intensity ForecastMigration Intensity Forecast

FF = (w1F1 + w2F2)/(w1 + w2)

F1 = probabilistic forecast based on cluster patterns

F2 = linear regression forecast based on deterministic

weather forecast model (e. g., Van Belle et al. 2007).

W = P(B|A) A= forecasted bird migration intensity

B= observed bird migration intensity

Presumably Presumably F1 (F2 ) forecast is better for longer (shorter) lead times.

Bayesian forecast also possible using multimodel ensemble approach

Bird migration related to (a) weather, (b) teleconnections (PNA, NAO),

and (c) climate (ENSO)

Forecast of bird migration intensity based upon cluster analysis of the 300-hPa

meridional wind

Using weights, can combine probabilistic forecast with deterministic linear regression

forecast

The technique can be extended to seasonal bird migration intensity forecasts.

ConclusionsConclusions