593
1 INTRODUCTION
Despite the technological advances observed in
maritime transport, the average annual number of
maritime accidents and incidents, is still very high
2,676 [18]. The largest maritime disasters in the Baltic
Sea, which occurred during the storms above 7 B, in
peacetime, involved two Ro-Pax ferries: Estonia which
sank in 1994 in storm conditions of 7-8 B, with large
number of victims and Jan Heweliusz in 1993 in storm
conditions of 12 B.
SAR services operate under the time pressure,
requiring a quick and effective response to emergency
situations. The most important factor influencing SAR
mission success is the time to search, detect and pick
up the survivors, which must be shorter than their
survival time [3][27].
Assuming the optimal technical and organisational
conditions for conducting the search and rescue
operations for people in life-threatening at sea, the
effectiveness of the maritime SAR (Search and Rescue)
operations depend on the knowledge of
hydrometeorological conditions affecting the drift of
search objects, predicted datum (the most probable
reference position of search object), predicted search
area – area with the highest probability of search
object containment and search object detection
including coverage patterns [2][9]19].
The use of multiple unmanned vehicles in search
and rescue missions, cooperating with the surface
Determining the Search Object Datum in a Semi-
Enclosed Sea. A Case Study for the Southern Baltic Sea
Z. Burciu
Gdynia Maritime University, Gdynia, Poland
ABSTRACT: One of the most important issues related to the success of SAR (Search and Rescue) action at sea is
prediction of the datum – the most probable position of the search object, dependent on its drift - movement
caused by wind and water current. The paper presents an empirical model of search object drift, based on
laboratory tests, intended for implementation in a decision support system dedicated to SAR services.
Verification of the model was carried out in real sea environment conditions during the international maritime
SAR exercises carried out in the Southern Baltic Sea. The main objective of the exercises was determination of
the datum of selected life rafts. The results of sea trials compared to the positions determined using different
DSSs (decision support systems), used by SAR services, showed significant differences between them. The
models of the surface water current and leeway of search objects, presented in the paper, allowed to determine
the reference positions of life rafts closest to the positions observed during the sea exercises. The presented
empirical model allowed to predict much smaller search area than the area determined based on the IAMSAR
Manual (International Aeronautical and Maritime Search and Rescue Manual) recommendations. The major
conclusion from the presented research for the Southern Baltic is the need to take into account the local
conditions of the semi-enclosed sea and standardize the methods used by cooperating national SAR services.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 20
Number 3
September 2026
DOI: 10.12716/1001.20.03.07
594
rescue units, can reduce the time needed for SAR
operations and detection of the missing persons
[19][20][23], however the correct prediction of the
datum is the main factor shortening the search time
and increasing the probability of people survival
[1][12][24]. 2022).
The generally accepted methods for determining
the drift of search objects, contained in the
recommendations of the IAMSAR Manual
(International Aeronautical and Maritime Search and
Rescue Manual), approved by the ICAO (the
International Civil Aviation Organization) and IMO
(International Maritime Organization) Maritime
Safety Committee are based on the sea trials
conducted in the ocean [7][6][16]. The
recommendations for the management and
coordination of the search and rescue operations,
based on these methods, enable the SAR action
coordinators and Ship Masters to determine the
datum and search area during the SAR operations in
weather conditions up to 7 B.
The recent advanced models as presented by [3][5],
based on the DQN (Deep Q Network) algorithm,
combining the perceptual capabilities of deep learning
with the decision-making capabilities of
reinforcement learning, also use the leeway speed
model, recommended by the IAMSAR Manual,
proposed by [6].
[25][26], who analyzed the field experiments and
laboratory simulations, reported in the literature, on
life raft leeway characteristics, confirmed the
importance of a correct modeling the real sea
conditions. The research presented in this paper
allowed to determine the datum and search area,
based on the sea trials in the Southern Baltic and
experimentally predicted search objects leeway in
higher weather conditions than 7 B, up to the limit,
related to the life raft safety function [13].
2 EXPERIMENTAL DETERMINATION OF
SEARCH OBJECTS LEEWAY IN THE REAL
CONDITIONS OF THE SOUTHERN BALTIC SEA
The research on the drift of search objects, conducted
at the Gdynia Maritime University in the Southern
Baltic Sea [10][12] showed the significant
discrepancies with the recommendations of IAMSAR
Manual. The reason for the discrepancy was mainly
related to the local conditions, prevailing in the semi-
enclosed sea, compared to the ocean, where the
research presented in IAMSAR Manual was
conducted. The new approach enabled the
development of drift models for life rafts and PIW
(person in the water), and determination of the search
area in weather conditions up to 11-12 B [10][13][11].
The trajectory of a drifting object in a two-
dimensional coordinate system can be calculated from
the system expressed in Eq. (1) [8].
( )
( )
b
/ , ,
/ v , ,
b
dx dt u x y t
dy dt x y t
=
=
(1)
where: x(t), y(t) are the positions of the centre of
gravity of the solid body in a fixed rectangular
coordinate system and ub, vb are the components of the
velocity in the x and y directions, respectively.
The forces acting on the search object and the
velocity of the search object can be defined in form of
Eq. (2), on the basis of Newton's second and third
laws [8].
'
O N W R
dv
M F F F F
dt
= + + +
(2)
where: M' is the total mass (mass of the object + added
mass of water), v - velocity, FO – calm water resistance
force, FN – wind pressure force, FW – additional water
resistance in waves, FR – additional forces (Coriolis,
Stokes).
To determine the motions of life rafts in the real
sea conditions the real scale experiments were
conducted in the Southern Baltic Sea with 6, 10 and
20-person life rafts, produced by Stomil Grudziadz
and real scale model of PIW. The visualization of sea
trials of the 10 person life raft are presented in Fig. 1.
Figure 1. Visualization of sea trials of the 10 person life raft
in the real sea conditions.
The measuring equipment installed on board the
life raft is presented in Fig. 2.
Figure 2. Measuring equipment installed in the life raft.
The model of PIW during the sea trials is
presented in Fig. 3.
Figure 3. PIW model during trials in the real sea conditions.
595
The measuring equipment installed inside and on
the PIW model is presented in Fig. 4.
Figure 4. Measuring equipment installed inside and on the
PIW model.
During the trials conducted in the weather
conditions above 11 B, the research life raft with the
equipment was lost, therefore it was decided to
conduct the experiments in laboratory conditions.
2.1 Laboratory tests on hydrodynamic and aerodynamic
impacts generated on a life raft
The hydrodynamic and aerodynamic tests of 6, 10 and
20-person pneumatic life rafts, produced by Stomil
Grudziadz, were conducted in the towing tank of the
Maritime Advanced Research Centre (CTO SA) in
Gdansk, Poland and in the wind tunnel at the Low
Speed Aerodynamics Laboratory of the Institute of
Aviation in Warsaw, Poland.
The 6-person life raft during the tests in the towing
tank is presented in Fig. 5.
Figure 5. Towing tank tests of the 6-person life raft [11].
The 6-person life raft during the tests in the wind
tunnel is presented in Fig. 6
Figure 6. Wind tunnel tests of the 6-person life raft [11].
The tests allowed to determine the FN and FO
forces. Assuming that an object is carried by the wind
in a straight line with a constant speed relative to the
water, the Newton's first law of motion can be applied
Eq. (3) [8][10].
NO
FF=
(3)
Research conducted in the towing tank made it
possible to determine the life raft calm water
resistance curve FO(v). An example of the resistance
curve for the 6-person life raft occupied by 2 survivors
and regression equation of the resistance function are
presented in Fig. 7 and Eq. (4).
Figure 7. Calm water resistance curve for 6 persons life raft
occupied by 2 survivors.
2
0
7.61 45.9 192,10F v v= − +
(4)
where FO (N) is the life raft calm water resistance, v
(m/s) is the life raft velocity.
The general model of the life raft calm water
resistance FO (N) dependent on the life raft velocity v
(m/s) is expressed in form of Eq. (5).
2
0 0 1 2
a a aF v v= + +
(5)
where a0, a1, a2 are the coefficients 6, 10 and 20-person
life rafts, occupied by 10%, 50% and 100% of
maximum number of survivors. The coefficients are
presented in Table 1.
Table 1. Coefficients of the hydrodynamic force FO.
Type of life
raft
Life raft
occupancy
a0
a1
a2
6 person
10 %
0,754
7,892
118,671
50 %
10,371
-73,897
196,791
100 %
8,368
-43,569
189,496
10 person
10 %
7,388
-30,440
258,830
50 %
3,028
7,711
237,813
100 %
7,193
-41,947
274,490
20 person
10 %
3,725
54,036
119,578
50 %
5,404
-31,157
239,841
100 %
4,457
-37,354
294,345
The general model of the aerodynamic force FN
(N), dependent on the wind velocity vW (m/s) for three
types of life rafts is expressed in form of Eq. (6).
2
0 1 2
b b b
Nw
F v v= + +
(6)
where b0, b1, b2 are the coefficients, for 6, 10 and 20-
person life rafts presented in Table 2.
596
Table 2. Coefficients for the wind pressure force FN.
Type of life raft
b0
b1
b2
6 person
28,591
-4,216
0,937
10 person
-18,461
3,550
0,815
20 person
-57,865
5,770
1,321
Based on the assumption of equilibrium of forces
acting on the life raft at a constant velocity and using
the models describing the calm water resistance force
FO, and the wind pressure force FN, the regression
models describing the life raft leeway velocity, in
terms of wind velocity, were determined.
The obtained characteristics were compared with
the data available in the IAMSAR Manual [16] (Fig. 8),
used to predict the reference position and search area
during the SAR operations, determined by [6],
representing the leeway velocity of various drifting
objects. The characteristics of life rafts, marked in red
in Fig 8: A – the leeway velocity for life rafts with no
ballast, with canopy and drogue, with probable
leeway velocity errors 0.35 kts (knots, kts=NM/h=0.514
m/s, NM - Nautical Mile=1852 m),
B – leeway speed for life rafts with no ballast, no
canopy and no drogue, with probable leeway velocity
errors 0.25 kts, were compared with the characteristics
obtained from the laboratory tests, presented in
(Fig. 9).
Figure 8. Velocity of leeway (maked “Leeway” in the
original drawing) of drifting objects in terms of wind
velocity up to 34 kts. Developed on the basis of IAMSAR
Manual [16]
The Figures 9 a) and b) present the determined
leeway speed of different types of life rafts without
drogue and with drogue respectively, for the weather
conditions up to 11 B. The leeway characteristic A and
B from Fig. 8, for the wind speed up to 7 B, have been
superimposed on these characteristics.
Figure 9. Determined characteristics of leeway speed of life
rafts: a) without drogue, b) with drogue. A, B –
characteristics from Fig. 8; Life raft 20-2 – 20-person life raft
occupied by 2 survivors; Life raft 20-20 – 20-person life raft
fully occupied; Life raft 10-1 – 10-person life raft occupied
by 1 survivor; Life raft 10-10 – 10-person life raft fully
occupied; Life raft 6-1 – 6-person life raft occupied by 1
survivor; Life raft 6-6 – 6-person life raft fully occupied.
3 DETERMINATION OF SURFACE CURRENT
VELOCITY
The real measurements of wind-generated surface
water currents in a layer of 0.5 m below the water
surface, in the Southern Baltic Sea, allowed to
determine the surface current velocities in terms of the
wind velocity. The measurements were conducted
using the hydrometeorological buoy designed at the
Gdynia Maritime University (Fig.10).
Figure 10. Measurements of the surface water current, from
the left: hydrometeorological buoy, anchoring method,
measuring buoy installed in the Southern Baltic Sea.
The results showed the significant differences in
values of the surface water current direction Kc and
velocity vc compared to the recommendations of
IAMSAR Manual [16].
The statistical analysis of the measurements
allowed for determination of the surface water current
velocity filed, related to the wind direction, as a
function of the wind speed and wind speed
thresholds at which the reverse currents on the left
and right side of the wind direction disappear.
597
The maximum surface current velocity equal to
1.2 kts was recorded during the sea trials for the wind
speeds of 80 kts (over 12 B).
The surface water current velocity fields related to
the wind direction, in terms of wind speed in the
range of 0 to 50 kts, observed in the Southern Baltic
Sea, are presented in Fig. 11.
Figure 11. Surface water current velocity fields in terms of
wind speed observed in the Southern Baltic Sea.
The general conclusions from the measurements
are presented in Table 3.
Table 3. General conclusions from the measurements of the
surface water currents in the Southern Baltic Sea.
vw
(kts)
General conclusions
0 - 5
The influence of the history of changes in hydrometeorological
elements is clearly visible.
5 - 20
There is an increasing influence of wind friction with the
water surface on the distribution of surface current direction.
20 -
25
The difference between the current direction and the wind
direction, occurring as a result of wind friction, does not
exceed 90° to the left of the wind direction.
> 25
There are no back currents.
The models of wind-generated surface water
current velocity, related to the wind direction, in
terms of the wind velocity, have been determined
separately for the wind velocities less or equal to 18
kts and for the wind velocities greater than 18 kts
(Table 4).
Table 4. Wind generated surface water current velocity
models in terms of wind speed vw, wind direction Kw and
surface current direction Kc.
vw (kts)
≤18
> 18
Surface
current
direction
To the left of
the wind
direction
To the right
of the wind
direction
To the left of
the wind
direction
To the right of
the wind
direction
Kc-Kw (deg)
–180° to 0°
0° to 180°
103/(1.66 –
0.56 vw)
-
5.34+3587.98/vw
vc (kts)
0 - 0.35
(0.23 +
0.01vw)
2
±
0.17
The comparison of wind-generated surface water
current velocities, recommended by IAMSAR Manual
[16] for the Latitude greater than 10° N, with direction
30° to the right from the wind direction and velocities
observed in the Southern Baltic Sea, and difference
between them Δvc are presented in Table 5.
Table 5. The maximum wind-generated surface water
current speeds recommended by IAMSAR Manual and
observed in the Southern Baltic Sea, in terms of wind
speeds.
vw (kts)
vc (kts)
Δvc (kts)
IAMSAR
the Southern
Baltic Sea
0 - 18
0 – 0,65
0.0 – 0.35
0.3
23
0.80
0.37
0.43
28
1.00
0.39
0.61
34
1.20
0.41
0.59
40
–
0.45
–
50
–
0.54
–
The surface water current velocities of 1.2 kts were
recorded during the sea trials at wind speeds of 80 kts
(over 12 B). The knowledge of the local surface water
currents and their proper consideration when
planning SAR action, enables more accurate
determination of datum and search area, which
significantly reduce the time needed to find survivors.
4 DETERMINATION OF THE DATUM AND
SEARCH AREA DURING SAR ACTION
The basic parameters to determine the datum are the
search object leeway velocity vL, total water current
velocity vC and time t elapsed since receiving the
information of LKP − the last known position of the
search object.
Determination of the search object drift velocity vD
is a sum of leeway and total water current velocities
(Fig. 12).
Figure 12. Determination of the life raft drift velocity vD.
The datum is determined at the distance equal to
the search object drift from LKP. The drift is a product
of vD and time t. Knowing the drift from LKP, the
assumed search area during the immediate SAR
action, as recommended by IAMSAR Manual [16], is a
square circumscribed on a circle with a radius R of
10 NM (Fig. 13).
598
Figure 13. Determination of the datum and search
area according to IAMSAR Manual.
In the proposed model of search area
determination the parameters used to define the
search area include the leeway vector, area dependent
on the water current velocity field and the areas of
possible errors WDE (Wind Direction Error) and WVE
(Wind Velocity Error), dependent on the wind
direction and wind velocity, respectively. The sea
current, if present in the actual theatre of SAR
operation, can be additionally taken into account.
An example of the search area determined for a life
raft, vW=23 kts, in the Southern Baltic Sea, obtained
using GMU DSS (Decision Support System developed
by GMU), is presented in Fig. 14. Due to the influence
of a near-surface current field and the direct impact of
the wind on the life raft, the shape of the presented
search area is deformed.
From the initial circular area, we obtain an ellipse
with axes located on the resultant wind direction:
semi-axis l1 in form of Eq. (7) and semi-axis l2,
perpendicular to the wind direction, in form of Eq. (8).
( )
1 c
l t v OVE t WVE= + +
(7)
where: l1 is the semi-axis of the search area ellipse on
the wind direction; vc is the current velocity; t is the
time of drift; OVE (Occupation Velocity Error) is the
life raft’s leeway velocity error, dependent on the life
raft's occupation, expressed by Eq. (8), and Eq. (9) for
the example 6-person and 10-person life rafts with a
drogue and without a drogue respectively; WVE is the
wind velocity error expressed by Eq. (10) for a life raft
with a drogue and Eq.(11) for the life raft without a
drogue.
( )
8
n
0
a
n
w
n
OVE v
=
=
(8)
where an, n=0,...,4 are the coefficients for 6-person and
10-person life rafts with a drogue, given in Table 6.
Table 6. Coefficients of Eq. 8.
n
an
life raft with a drogue
6-person
10-person
0
-0,00355
-0,02526
1
-7,491·10
-4
-5,482·10
-4
2
2,6072·10
-4
9,049·10
-5
3
9,588·10
-7
6,255·10
-7
4
-1,669·10
-7
-5,168·10
-8
5
-3,352·10
-10
-1,973·10
-10
6
3,8913·10
-11
1,0877·10
-11
7
5,1718·10
-14
2,8465·10
-14
8
-4,5235·10
-15
-1,1984·10
-15
( )
4
n
0
a
n
w
n
OVE v
=
=
(9)
where an, n=0,...,4 are the coefficients for 6-person and
10-person life rafts without a drogue, given in Table 7.
Table 7. Coefficients of Eq. 9.
n
an
life raft without a drogue
6-person
10-person
0
-0,00584
-0,4138
1
-1,2244·10
-3
8,2539·10
-3
2
4,272·10
-4
1,388·10
-4
3
8,803·10
-7
5,773·10
-7
4
-1,5397·10
-7
-4,861·10
-8
0.05
W
WVE v t=
(10)
0.08
W
WVE v t=
(11)
where ΔvW, based on the observations, in
deterministic model is assumed equal to 2 knots.
In the random model wind velocity error is
expressed in Eq. (12) and Eq. (13) for a life raft with a
drogue and without a drogue respectively:
0.05WVE t
=
(12)
0.08WVE t
=
(13)
where β is the random variable with a continuous
uniform distribution on the interval <-2, 2>.
The semi-axis l2 of the ellipse, expressed by Eq(14),
depends on the divergence between the total current
and wind directions Kc - Kw, OVE, time of drift t and
WDE.
( )
( )
2
sin
cw
l t K K OVE t WDE
= − + +
(14)
where: l2 is the semi-axis of the search area ellipse on
the direction perpendicular to the wind, Kc and Kw are
current and wind directions, WDE is the Wind
Direction Error expressed by Eq. (15) in deterministic
approach and Eq. (16) in the random approach:
( )
sin 10WDE v t=
(15)
( )
sinWDE v t
=
(16)
where α is the random variable with a continuous
uniform distribution on the interval <-10⁰, 10⁰>.
599
Search area determination for vW=23 kts, without
sea current, dependent on the wind-generated surface
water current field in the conditions of the Southern
Baltic Sea is presented in Fig. 14.
Figure 14. Search area determination for vW=23 kts, without
sea current, dependent on wind-generated surface water
currents field (blue area) in the conditions of the Southern
Baltic Sea (Burciu, 2003).
The actual search area is a rectangle circumscribed
on this ellipse (Fig.15).
Figure 15. Square search area applied in SAR operation,
determined for vw=23 kts, dependent on the wind-generated
surface water current field (blue area) in the conditions of
the Southern Baltic Sea Developed on the basis of [10].
5 VERIFICATION OF SEARCH AREA
DETERMINATION METHOD
5.1 Comparison of the method developed by GMU with
the method recommended in IAMSAR Manual
The method of the search area determination
developed by GMU have been verified by comparing
it with the method recommended in IAMSAR Manual
[16]. The examples of the life raft datum and search
area determination, using both methods, for the 6-
person life raft with drogue and 10-person life raft
without drogue, with unknown occupation, are
presented in Fig. 16 and Fig. 17 respectively.
Figure 16. Graphical representation of datum and search
area determined by GMU DSS (Datum UMG) in relation to
the datum determined according to IAMSAR Manual
(Datum IAMSAR) for 6-person life raft with drogue.
Figure 17. Graphical representation of datum and search
area determined by GMU DSS (Datum UMG) in relation to
the datum determined according to IAMSAR Manual
(Datum IAMSAR) for 10-person life raft without drogue.
The distances between datums were 9.8 NM and
4.8 NM for the 6-person and 10-person life rafts
respectively. The results of datum determination are
summarised in Table 6.
Table 6. The results of datum determination for 6-person
and 10-person life rafts, with unknown occupation, with
and without drogue, using the method recommended by
IAMSAR Manual and GMU DSS.
Life raft type
6 person
10 person
Drogue
Yes
No
Time of drift (h)
14
8
7 7
Wind direction Kw (°)
240 260
240
Wind velocity vw (kts)
24 21
24
LKP
φ
55°21.0’ N
55°20.7’ N
λ
017°12.9’
E
017°12.7’E
Real position
φ
55°25.4’ N
55°26.5’ N
λ
017°27.8’
E
017°24.3’
E
Datum Position
IAMSAR
Manual
φ
55°23.5’N
55°23.5’N
λ
017°50.5’E
017°50.5’E
GMU
DSS
φ
55°24.1’N
55°24.1’N
λ
017°31.1’E
017°31.1’E
Distance between datum and
real position (NM)
IAMSAR
Manual
GMU DSS
Distance between datums
(NM)
9.8
4.8
Datum
LKP
l1
l2
Area
of water
currents
Search
Area
LKP
Wind
direction
error
Wind
speed
error
Datum
LKP
LKP
Datum
UMG
Datum
IAMSAR
Datum
UMG
LKP
Datum
IAMSAR
600
The comparison of search areas determined by
GMU DSS for the 10-person life raft without drogue
(Fig.17) and based on IAMSAR Manual is presented
in Fig. 18.
Figure 18. Search area determined according to IAMSAR
Manual (red square) and developed using by GMU DSS
(green square).
The area determined according to IAMSAR
Manual is about 9 times bigger.
5.2 Verification of the method based on the field
experiments
The GMU method was verified using the data from
the filed experiments conducted on 26 March 2024, in
the Bay of Gdansk, NE of the Vistula Estuary, during
joint SAR exercises of SAR services from four Baltic
Sea countries. Two 10-person life rafts with different
occupation, with a drogue and without a drogue were
involved. The results of these sea exercises are
presented in Table 7 and Fig. 19.
Table 7. Results of sea exercises of drift of 10-person life
rafts.
10 person life raft
A
B
No. of survivors
9
5
Drogue
Yes
No
LKP
φ
54° 23.79’N
54° 23.79’N
λ
019 02.35’E
019 02.35’E
Final position
φ
54° 26.20'
N
54° 26.70' N
λ
019° 06.60'
E
019° 06.70' E
Distance from LKP
(NM)
3.4
3.9
Direction
045°
042°
Velocity (kts)
average
0.5
Loss of signal after 1.5
hours
maximum
1
1.3
Drift of life rafts observed during the sea exercises
is presented in Fig. 19.
Figure 19. Drift of life rafts observed during the sea
exercises: a). life raft A with drogue, occupied by 9
survivors; b). life raft B without drogue, occupied by 5
survivors.
Verification of the GMU method based on the
comparison of final life rafts’ positions and datums
determined by GMU DSS and decision support
systems used by the MRCCs (Maritime Rescue
Coordination Centres) from Baltic States participating
in the sea exercises is presented in Fig. 20 and
summarised in Table 8.
Table 8 presents the last known positions (LKP)
actual final positions of life rafts A and B, and
datums determined by MRCCs of Baltic countries.
Different DSS systems allowed for different data
entry. For DSS MRCC I, II, and IV, it was not specified
whether the life raft was with or without a drogue. In
the case of GMU DSS, hydrometeorological data
obtained from the OSC (On Scene Commander) report
and the SatBaltic system (SatBaltic) were taken into
account.
601
Figure 20. Positions determined by MRCC stations and
GMU DSS in relation to the real final positions of life rafts A
and B.
Table 8. Positions of life rafts recorded during the exercises.
Datums determined by MRCCs and GMU DSS.
The exercises conducted in the South Baltic Sea
took place in 2-3 B conditions, for time of drift equal
to 7 hours. The maximum difference of the datums
was 3 NM. It is expected that in severe weather
conditions the dispersion of positions determined by
the MRCCs would be greater.
6 CONCLUSIONS
When planning the SAR operations in semi-close seas,
the recommendations of the IAMSAR Manual
regarding the wind currents and leeway, should be
applied with caution. They can interfere with the
determination of the datum. The presented test results
and computations determining the wind current areas
and leeway, specified by GMU DSS, enable more
accurate prediction of the datum and search area with
a measurably smaller surface, included in the larger
area determined by IAMSAR.
The values of wind surface water current and
leeway parameters, given in IAMSAR Manual are
available only for values up to 7 B and they do not
distinguish between the sizes of life rafts. GMU DSS
enables to determine the datum for values up to 11 B.
The exercises show that using different prediction
decision support systems based on the different
methods, used by MRCCs, different results of the
final positions of life rafts can be determined. The
datum is influenced by the diversity of
hydrometeorological forecasts from different
meteorological stations, what was presented in Table
8 and is confirmed by Kilic et al. (2025) who believe
that the discrepancies in the datum position, despite
known hydrometeorological conditions, may be
confirmed by different assessments of environmental
conditions.
The planned integration of the decision support
systems with an artificial intelligence technique, can
be a step forward to improve the effectiveness of SAR
operations [18]. The proper modelling is especially
important when the autonomous vehicles are
involved in the operation [15][17]. Additional
consideration of the characteristics of both the
manned and autonomous search units, participating
in the operation, included in DSS will increase the
SAR action effectiveness and probability of finding
the alive survivors [14][19][21].
The regional cooperation in implementation of the
SAR missions should increase the effectiveness of
SAR operations [27]. Since the IAMSAR Manual
recommendations are not the recommendations
precise enough for the Southern Baltic Sea, MRCCs
should consider standardizing the procedures for
determining the search areas in this region,
commencing a search in a smaller area designated
within a large area consistent with IAMSAR Manual.
ACKNOWLEDGEMENT
The author would like to thank the Polish SAR services for
the opportunity to use the data from the international SAR
exercises in the Southern Baltic Sea.
FUNDING
This research was funded by the Gdynia Maritime
University Grant WN/2026/PZ/08.
REFERENCES
[1] Abramowicz-Gerigk, T., Burciu, Z., Przybyl, W.,
Plebankiewicz, I., Januszko, A., 2020. The impact of the
improved search object detection on the SAR action
success probability in maritime transport. Sensors, 20,
3962. doi:10.3390/s20143962.
[2] Agbissoh Otote, D., Li, B., Ai, B., Gao, S., Chen, X., Lv,
G., 2019. A Decision-Making Algorithm for Maritime
Search and Rescue Plan. Sustainability 11(7), 2084.
https://doi.org/10.3390/su11072084.
[3] Ai, B., Jia, M., Xu, H., Xu, J., Wen, Z., Li, B., Zhang, D.,
2021. Coverage Path Planning for Maritime Search and
Rescue Using Reinforcement Learning. Ocean Eng. 241,
110098. https://doi.org/10.1016/j.oceaneng.2021.110098.
[4] Ai, B., Li, B., Gao, S., Xu, J., Shang, H., 2019. An
Intelligent Decision Algorithm for the Generation of
Maritime Search and Rescue Emergency Response Plans.
IEEE Access 7, 155835–155850. https://doi.org/10.1109/
ACCESS. 2019.2949366
[5] Ai, B., Caia, H., Zhua, Q., Fenga, C., Xub, N., 2025. A
maritime search and rescue path planning method based
on improved DQN. International Journal of Digital Earth
18, 2, 2562555. https://doi.org/10. 1080/17538947.,
2025.2562555.
LKP Life rafts: A and B
φ =54°23.79’N
λ =019° 02.35’E
Real final position
A
φ= 54° 26.20'N
λ= 019° 06.60'E
B
φ =54° 26.70'N
λ= 019° 06.70'E
Datum I MRCC
1930 UTC
φ =54° 24.48’N
λ =019° 03.42’E
Unknow drogue
Datum II MRCC
1600 UTC
φ =54° 25.54’N
λ = 019° 03.06’E
Unknow drogue
Datum III MRCC
1730 UTC
φ =54° 26.30’N
λ =019° 02.24’E
With drogue
φ =54° 27.00’N
λ =019° 01.48’E
Without Drogue
Datum IV MRCC
1630 UTC
φ =54° 27.27’N
λ =019 02,56’E
Unknow drogue
GMU DSS 1630 UTC
Hydrometeorological
conditions
from OSC reports
φ =54° 25.89’N
φ =54° 26.64N
λ =019° 03.68’E
λ =019° 02.9’E
Life raft A
Life raft B
GMU DSS 1630 UTC
Hydrometeorological
conditions
from SatBaltic System
φ =54° 26.10’N
φ =54° 27.02’N
λ =019° 03.93E
λ =019° 03.04’’E
Life raft A
Life raft B
1
602
[6] Allen, A. A., Plourde, J.V., 1999. Life rafts, survival craft
and person in the water (PIWs). Review of Leeway: Field
Experiments and Implementation.1999. U.S. Coast
Guard and Development Center, Report No. CG-D-08-
99. https://apps.dtic.mil/sti/html/tr/ ADA366414/.
[7] Allen, A., Fitzgeral, R. B., 1997. The Leeway of an Open
Boat and Three Life Rafts in Heavy Weather.
dhttps://www.researchgate.net/publication/235104790_T
he_Leeway_of
_an_Open_Boat_and_Three_Life_Rafts_in_Heavy_Weat
her.
[8] Anderson E. , Odulo, A., Spaulding, M., 1998. Report No.
CG-D-06-99, Modeling of Leeway Drift. U.S. Coast
Guard Research and Development Center.
https://apps.dtic.mil/sti/tr/pdf/ADA363933.pdf.
[9] Breivik, Ø., Allen, A., 2008. An Operational Search and
Rescue Model for the Norwegian Sea and the North Sea.
Journal of Marine Systems, Maritime Rapid
Environmental Assessment 69 (1), 99–113.
https://doi.org/10.1016/j.jmarsys.2007.02.010.
[10] Burciu, Z., 2003. Modeling search areas in terms of the
safety of human transport at sea. Publishing House of
the Warsaw University of Technology, Warsaw.
[11] Burciu, Z., 2007.The disturbances of the life raft leeway
induced by the fluctuations of wind direction in the life
raft coordinate system. Polish Maritime Research,
Special issue. Available online
https://yadda.icm.edu.pl/baztech/element/bwmeta1.elem
ent.baztech-4c3281d7-957c-44f0-bf2a-e4fc6e2b1c4a.
[12] Burciu, Z., 2012.Reliability of rescue operations in
maritime transport. Publishing House of the Warsaw
University of Technology, Warsaw.
[13] Burciu Z., Grabski F., 2011. The experimental and
theoretical study of life raft safety under strong wind.
Reliability Engineering & System Safety, 96, 11, 1456-
1461.
[14] Gerigk, M.K.; Gerigk M. 2023. Application of
unmanned USV surface and AUV underwater maritime
platforms for the monitoring of offshore structures at
sea. Scientific Papers of the Maritime University of
Szczecin, 76 (148), 89-100. https://doi.org/: 10.17402/590.
[15] Hangyu L., Miao, F., Mei, X., 2025. Facilitating Multi-
UAVs application for rescue in complex 3D sea wind
offshore environment: A scalable Multi-UAVs
collaborative path planning method based on improved
coatis optimization algorithm. Ocean Eng., 324, 30,
120701. https://doi.org/10.1016/j.oceaneng.2025.120701.
[16] IAMSAR Manual, 2025. International Aeronautical and
Maritime Search and Rescue Manual - Volume II -
Mission Co-Ordination. 2025 Edition. International
Maritime Organization and International Civil Aviation
Organization.
[17] Junmin, M., Hu, T., Chen, P., Chen, L., 2021.
Cooperative MASS Path Planning for Marine Man
Overboard Search. Ocean Eng. 235: 109376.
https://doi.org/10.1016/j.oceaneng.2021.109376.
[18] Kilic, K. I., Maity, S., Sung, I., Nielsen, P., 2025.
Challenges and AI-driven solutions in maritime search
and rescue planning: A comprehensive literature review.
Marine Policy, 178, 106692.
https://doi.org/10.1016/j.marpol.2025.106692
[19] Larsen, A. K., Kilic, K. I., Ladefoged, M. B., Sung I.,
2025. A novel path-finding approach for maritime
search-and-rescue missions incorporating dynamic
probability of a target location. Engineering
Optimization, 57, 11, 3521–3540. https://doi.org/
10.1080/0305215X.2024.2446590.
[20] Liu, L., Shan, Q., Xu, Q., 2024. USVS Path Planning for
Maritime Search and Rescue Based on POS-DQN:
Probability of Success-Deep Q-Network. J. Mar. Sci. Eng.
12(7), 1158. https://doi.org/10.3390/jmse12071158.
[21] Yue, M., Li, B., Huang, W., Fan Q., 2023. An Improved
NSGA-II Based on Multi-task Optimization for Multi-
uAV Maritime Search and Rescue under Severe
Weather. Journal of Marine Science and Engineering 11
(4), 4. https://doi.org/10.3390/ jmse11040781.
[22] SatBaltyk system. https://satbaltyk.iopan.pl/.
[23] Taipalmaa, J., Raitoharju, J., Queralta, |J. P.,
Westerlund, T., Gabbouj, M. On Automatic Person-in-
Water Detection for Marine SAR Operations. IEE Access,
VOL.12, 2024, doi: 10.1109/ACCESS.2024.3386640,
https://ieeexplore.ieee.org/stamp/stamp.jsp?
tp=&arnumber=10495041.
[24] Tipton, M., McCormack, E., Elliott, G., Cisternelli, M.,
Allen, A., Turner, A. C., 2022. Survival time and search
time in water: Past, present and future. Journal of
Thermal Biology 110, 103349.
https://doi.org/10.1016/j.jtherbio.2022.103349.
[25] Tu, H., Xia, K., Mu, L., Chen, X., Wang, X., 2022.
Predicting Drift Characteristics of Life Rafts: Case Study
of Field Experiments in the South China Sea. Ocean Eng.
262, 112158. https://doi.org/10.1016/j.oceaneng2022.
[26] Wu, J., Cheng, L., Chu, S., 2023. Modeling the Leeway
Drift Characteristics of Persons-in-water at a SeaArea
Scale in the Seas of China. Ocean Eng. 270, 113444.
https://doi.org/10.1016/j.oceaneng.2022.113444.
[27] Wang, Y., Qu, X., Li, Z., 2025. Optimizing pre-
occurrence maritime search and rescue system: A
dynamic location-allocation model with considering
spatiotemporal accessibility. Ocean Eng. 337, 1, 121964.
https://doi.org/10.1016/j.oceaneng.2025.121964
