Forecasting drifting objects
Transcript of Forecasting drifting objects
Stochastic Forecasting of DriftingShips and Smaller Objects
Dr Øyvind BreivikNorwegian Meteorological Institute
Kjell RøangChristian Michelsen Research
Objective• To generate search areas for the Norwegian
Search and Rescue Service based on thebest available wind and current information
• Make search areas as small as possible, butnot smaller
Challenge
Other uses
• Containers• Mine-like objects• Objects at large (buoys, ROVs)
The Uncertainties Involved
• Where and when did the accident takeplace?
• Which object should we look for (life raft,person in water, …)?
• What are the wind conditions like in thearea?
• What are the surface currents in the area?
Search Maths
POS = POD x POC
• POS: Probability of success (do we findwhat we are looking for?)
• POD: Probability of detection (the keeneyes of the rescuers)
• POC: Probability of containment (are wesearching in the right place?), our business
Forces on a Drifting Object
• Wind (leeway)• Surface current• Wave motion (damping and excitation)
The motion of an object of arbitrary shape isdifficult to model, approximations areneeded
Wind to Wind to
WindDirection
From
WindDirection
From
Wind to
Lα = -25o Lα = +25o
RWD = -135o RWD = +135o
Leeway DriftDirection
Leeway DriftDirection
Leeway divergence
• Leeway divergencesends objects at an anglerelative to downwind
• Symmetry allows stabledrift left and right ofdownwind (little jibingis observed).
? Diverging search areaswith time
DownwindApproximations
Wind speed and object drift is approximatelylinearly relatedDifferent objects drift differently ...
Life raft with drogue
Undrogued life rafterror alexperiment
dσ
0 5 10 15 20 25−30
−20
−10
0
10
20
30
Wind Speed to 10−m height (m/s)
Cro
ssw
ind
Com
pone
nt o
f Lee
way
(cm
/s)
Maritime Life Rafts with Deep Ballast System and Canopy(4−6 person capacity)
RIGHT of DownwindLEFT of Downwind
CrosswindCrosswinderror alexperiment
cσ
0 5 10 15 20 25−20
0
20
40
60
80
100 O(100) ensemble of random maritime life raft drift properties
10 m wind speed [m/s]
Dow
nwin
d co
mpo
nent
of l
eew
ay [c
m/s
]
Empirical leeway modelEmpirical leeway model
' walk)(random windPerturbed
10 wWW +=
),0(onsperturbatileeway Constant
σε N∈
offset ,2/'slope ,10/'
εε
+=+=
bbaa
''[cm/s]leeway Crosswind
ccc baL ±±= W
''[cm/s]leeway Downwind
ddd baL += W
Archive and prognoses7 day archive combined with 60 h prognoses of• Wind from 20 km res. atmospheric model (HIRLAM 20)• Surface currents from 4 km res. ocean model with tides and wind forcing (POM)
Wind and current close upWind and current close up
AWI 0203
Longitude
Latit
ude
Drifting benthic lander, 2002−03−13T11 UTC to 2002−03−17T17 UTC
30’ 15oE 30’ 16oE 30’ 17oE 30’ 18oE
55’
72oN
5’
10’
15’
Ensemble mean trajectory
Observed trajectory
Pickup +104 h
Drift start 2002−03−13T11 UTC
13T12
15T0616T04
16T09
16T15
17T01
Icelandic exercise 2003
Longitude
Latit
ude
Iceland−Faroe exercise 2003−02−06T11 UTC to 2003−02−07T11 UTC
20’ 10’ 13oW 50’
42’
45’
65oN 48.00’
51’
54’
57’
Liferaft pickup +24 h
Convex hull of ensemble +24 h (search area)
Ensemble mean trajectory
Observed raft trajectory
Liferaft release position 2003−02−06T11 UTC
Faroe exercise 2004
Longitude
Latit
ude
Iceland−Faroe exercise 2004−05−03T22 UTC to 2004−05−04T14 UTC
30’ 15’ 12oW 45’ 30’ 62oN
10’
20’
30’
40’ Liferaft release position 2004−05−03T22 UTC
Liferaft pickup +16h
Convex hull of ensemble (search area)
Manual search areas
Operational use of LeeWay model
• Service operational from 2002• Simulation started by web order form• Result transferred by e-mail• Result presented in JRCC decision support
system SARA
System architecture
Wind and Current
DriftSimulator
Web order ResultPresentation
E-mail Transport
GRIB format
Web order, old
Web order, New
Results, sailboat
Results, PIW and Life raft
Other simulators
• Oil drift• Ship drift
Web order oil drift simulations
Oil drift
Web order, ship drift simulations
Ship drift, 1 simulation
Ship drift, ensemble model
Other services
• Weather forecasts overlay (MetOc)• AIS/VMS• Weather Observations
Weather forecast
AIS
Basestations PortsGraticule
Legend
Weather Observations
Weather Observations