FutureWaves™:

a Real-Time Ship Motion Forecasting System employing Advanced Wave-Sensing Radar

J.G. Kusters, K.L. Cockrell, B.S.H. Connell, J.P. Rudzinsky, and V.J. Vinciullo
Applied Physical Sciences Corporation, USA
[email protected]

Abstract — We present the design, prototype, and performance

results of a standalone system that predicts vessel motion time-
series based upon wide-area ocean-surface radar operational-
observations. It is based on development conducted with
sponsorship from the United States Office of Naval Research
(ONR) under the Environmental and Ship Motion Forecasting
(ESMF) program. The key inputs are Doppler-radar
measurements of the ocean surface, the vessel characteristics,
and measured vessel motions. The principal outputs are a
statistical representation of expected vessel motions for various
courses and speeds and a precise time-specific prediction of the
vessel motions for several minutes. The system also provides
detailed, real-time wave characterization in the form of a self-
Fig. 1. USNS Montford Point (T-ESD-1) demonstrates skin-to-skin
calibrating energy power spectrum, sea heights for various operations and Landing Craft Air Cushion (LCAC) recovery while
frequency (wave period) bands, and trends of the ocean attached to the USNS Bob Hope (T-AKR 300) (U.S. Navy photo by Mass
environment, as well as trends of the measured resultant vessel Communication Specialist 1st Class Shannon E. Renfroe/Released)
motions. Meteorological and Oceanographic (METOC)
numerical models and direct buoy observations can also be used Motivated by the complex skin-to-skin (S2S) operations
by the statistical ship-motion forecaster to provide vessel required for military sea-basing, shown in Fig. 1, the ESMF
operational planning guidance for hours and days ahead. The
program set out to determine if detailed coherent resolution of
developmental prototype system was tested and demonstrated on
four different USNS vessels in support of naval operations on two
an ocean field and the resultant precise prediction of vessel
occasions in 2015 and again in July 2016. These demonstrations motions was feasible [1]. The Expeditionary Transfer Dock
have borne out the relevance and utility of such a system in (ESD) is designed to serve as the interface between Large
enhancing safety of critically sensitive operations and in Medium-Speed Roll-on/roll-off (LMSR) ships and Landing
expanding operational envelopes where the ocean environment Craft Air Cushion (LCAC) surface connectors [2]. ESD allows
can have significant impact. Coined “FutureWaves™”, the for the transfer of equipment via a connecting Vehicle Transfer
system provides the ship master and crew with the advanced Ramp (VTR) at sea in non-anchorage depths and delivery from
warning information necessary to assess and safely respond to a over the horizon through restricted access environments, which
challenging ocean environment. Mariners have long relied upon when combined with the mobility of the sea-base, enhances the
seaman’s eye to assess the seas and warn of impending waves and safety of the assets in potentially hostile environments [3].
potential hazards. FutureWaves™ revolutionizes this essential
seafaring skill by leveraging modern processing strength with an Derived from the ESMF effort, FutureWaves™ is a method
inventive approach that provides a cutting-edge capability that is and system for sensing, decomposing, and reconstructing a
key to the safe execution of crucial offshore operations involving wide-area wave-field in the proximity of waterborne vessels
the transfer of personnel and equipment at sea. that is subsequently used to predict specific wave events at or
Keywords— wave radar, real-time wave characterization, vessel
motion prediction, deck motion coherent prediction, phase-resolved
wave fields, workability

I. INTRODUCTION
Conducting operations and work at sea is an inherently
risky proposition. FutureWaves™ seeks to reduce risk and
improve safety, as well as the operational availability, or
workability, by providing not only detailed information about
the ocean surface and incident wave field, but also precise
time-specific predictions of when conditions will remain within
defined thresholds or exceed operational limits. It can also be
used for research and studies where detailed, coherent (phase
and amplitude) knowledge is required of the ocean surface over Fig. 2. Simplified multi-modal wave-field with various wave groups
a wide area of interest. spatiotemporally interacting with the vessel along the intended track.

Sponsored by Office of Naval Research (ONR) Division 331,

Ship Systems and Engineering Research.
near the vessels in the very near future (out to minutes) and the
resultant time-domain motions of the vessels. Fig. 2 illustrates
a simplified multi-modal wave-field and how various wave
groups will spatiotemporally impact the vessel. Phase-resolved
measurement of these wave groups and subsequent forcing for
a real-time vessel motion model is the function of the
FutureWaves™ system. The time and geographic
(spatiotemporal) location of specific forecasted waves (phase-
resolved ocean surface height) is reconstructed and used as
information for user operational guidance along with the
spectral (direction, period and energy) information via a
display or sent directly to an external device or system where it Fig. 3. FutreWaves™ System Diagram
can be utilized for further processing where spatiotemporal or wave radar, a meteorological sensor, and the ship motion
spectral wave information is required. sensor (SMS) for 6DOF motions, heading, and location. These
inputs are sent to the wave signal processing algorithm, shown
II. SYSTEM DESCRIPTION in yellow, which generates a two-dimensional power spectrum
The system employs one or more marine X-Band (9.2GHz) (2DPS) of the wave-field and the amplitude and phase of the
Doppler radars to measure the ocean surface orbital radial- discrete spectral components. The two core algorithms for
velocities through the Bragg backscattering in an area around vessel motions are shown in red. The Course & Speed
the vessels out to several thousand meters, a geolocation Recommendation Model (CSRM) can receive input from the
position and heading device [e.g. Global Positioning System live wave-radar 2DPS as well as available buoy data and/or
(GPS)], an orientation device [e.g. Inertial Measurement Unit gridded METOC forecast ocean wave spectra data, both shown
in green, and then uses the spectral wave-field information to
(IMU)] that provides six degree-of-freedom (6DOF)
generate statistical motion predictions for each degree-of-
information, and a processor for computation. The measured
freedom for various courses and speeds. The Ship Motion
wave orbital radial-velocities are used to calculate detailed
Forecast (SMF) model receives initial vessel state conditions
wave-field amplitude and phase frequency components which
from the SMS and the wave components from the wave signal
are then used to determine and predict the incident wave
processing algorithm and generates a coherent prediction for
interaction along the ship’s intended (dead-reckoning) track
and along the ship’s hull. The phase-resolved wave-field is each degree-of-freedom for up to three (3) minutes with our
described by discrete spectral components that are used as current prototype system. The forecast length (time) is largely
input to a time-domain ship motion simulation based on the dependent on the wave-radar range and the ability to see ‘up-
vessel’s characteristics such as length, beam, and tonnage. The wave’ in the ocean wave-field in the vicinity of the vessel.
wave measurement process fits the wave field in the region of This is required in order to determine the phase and timing of
the ship with a set of discrete wave components and their future incident waves on the vessel. For typical sea-states with
complex amplitudes [4]. Ship motion sensors aboard each of a swell component period on order of up to 20 seconds, the
the ships provide the 6DOF state history of the ships. The ship wave group velocities are roughly 16 m/s (~30 knots),
motion prediction approach must therefore be able to accept therefore the wave-radar must ‘see’ on order of three (3)
discrete wave components and the ship state history as inputs. kilometers to provide a three (3) minute coherent phase-
The system’s real-time forecasting requirements translate to a resolved wave forecast.
requirement to process the wave measurement and ship motion An example of the live wave-radar 2DPS and the
simulations within a few seconds. We have developed a corresponding CSRM output for a single degree-of-freedom
reduced-order model (ROM) for ship seakeeping [5] based on (roll in this case) are shown in Fig. 4. The coherent prediction
the Cummins approach [6], and have integrated this into our for a single degree of freedom is shown in Fig. 5. For
system to produce the necessary computational speed. The development purposes we display detailed signal processing
Cummins equation is cast to represent motions of multiple (N) 2DPS CSRM
ships as a 6×N degree-of-freedom system represented by 6×N
equations. The formulation uses the state as an initial condition
in a time-domain solution, and the state history is used in a
convolution with pre-calculated impulse-response functions. A
pre-calculated wave-forcing database allows mapping of the
discrete wave components to modal forcing of the ship system. Course

Resultant vessel motions for different courses and speeds are

calculated for the recommendation model, and the actual vessel
motions for the current course and speed are predicted for Speed
several minutes into the future.

A. System Functional Diagram Fig. 4. Two-Dimensional Power Spectrum (2DPS) of the wave-field and
corresponding Course & Speed Recommendation Model (CSRM) for single
Fig. 3 is a functional diagram of the FutureWaves™ degree-of-freedom (roll in this case).
system. Shown in blue are the shipboard inputs from the
 B. Advanced Wave Sensing Radar (AWSR)
As previously mentioned radar range capability and
performance are critical for system forecasting performance.
For the ESMF effort and in conjunction with the Applied
Physics Lab at the University of Washington (APL-UW), we
designed and built a customized wave radar capable of meeting
system requirements [7]. This Advanced Wave Sensing Radar
(AWSR) was to meet the following goals:
 Accurate Doppler sensing to 5 km, enabling a 5 minute
wave forecasting capability in all environmental and
ship speed scenarios of interest.
 Azimuthal scanning interval of less than 3 seconds
providing sufficient sampling criteria for wave fields
whose content is largely for periods > 6 seconds.
 Azimuthal resolution of ~2.5 degrees and range
resolution of ~7.5 meters.
The AWSR achieves these goals by drawing on many of
the design features of APL-UW’s previous ship-board CORAR
system [8, 9]. These features include:
 Vertical Transmit and Receive (VV) Polarization:
substantially improve scattering performance compared
Fig. 5. ‘Micro-Forecast’ Display for a single degree-of-freedom to standard marine radar HH- polarization.
prediciton (roll in this case).  Coherent Doppler Radar Capability: Stable oscillator
information of forecasted motions that would not likely be and 3kW amplifier permit high-fidelity measurement of
intended for an operator display, such as cross-correlation of Doppler shift.
the forecast to the measured signal, average-angle  Four antennas with a switching network which allows
measurements, cross-correlation lag, rms ratio, and spectral for a slower rotation rate (increased radar dwell) while
comparison between measured and predicted signals. We have maintaining the required sampling rate.
coined this the ‘Micro-Forecast’ display and are striving to The AWSR implemented and built for ESMF is shown in
reflect the system performance in real-time on a single display. Fig. 6. It is constructed in two main components, a Deck Top
The time-domain prediction for the selected degree-of-freedom Unit (DTU) and Mast Top Unit (MTU) which allows the MTU
is shown in green with the ‘now’ time or zero-time (0) is in the to be separate from the bulkier electronic components and
center of the time-domain plot. The measured motion is shown possibly placed in a more ideal location. The DTU contains
to the left in red with a user-selected forecast time (in this case the RF power amplifier, RF/IF circuitry, a GPS disciplined
60 seconds) shown to the left also in green for comparison to oscillator, a processing and control unit, and power supply. A
the measured motion. This allows a quick review of amplitude photograph of the MTU is shown on the right of the figure as it
and phase prediction for the selected forecast time. The was installed on board T-ESD-2. The MTU contains the four
forecast signal in green to the right displays the current forecast antennas, RF circuity (Circulator, BPF, LNA), RF switching,
with the average signal envelope shown in white, and the slip ring, rotator, and position encoder. The MTU and DTU
envelope with one-sigma (standard deviation) added shown in are connected via 3-cm RF waveguide and associated power
violet. The intent of this is to determine threshold exceedance and data cables.
of a particular parameter or correspondingly quiescent periods
of prediction where operations can occur and a confidence
(sigma) level for the prediction. The waterfall display below
the time-domain signal shows how each new generated
forecast compares to previous ones. When a new forecast is
obtained and displayed in the time-domain plot, the previous
forecast is shifted left by the processing lag time (~1.5 seconds
in the present system) and down a raster line. Hence, when a
signal peak is being forecast on subsequent updated forecasts it
appears as a bright vertical line in the waterfall display. In the
example shown the fuzziness of some vertical lines indicates
variance in subsequent forecast, and the spectral comparison
reflects a magnitude and period discrepancy for roll in this
case. The waterfall display to the left of the now-time also
displays the measured motion so that it can also be compared
to several forecasts. In this manner the system analyst can
quickly assess live system performance. Fig. 6. Advacned Wave Sensing Radar (AWSR)
 III. AT-SEA TESTING coupled S2S operations, however the system was fully
The ESMF program conducted two major at-sea test in functional for single-ship operations and relevant operating
May and October 2015 and an additional ‘experimental’ guidance based on statistical motions was provided for S2S
operational support period in July 2016. Prototype systems operations. It was during these test periods that we gained
were temporarily installed onboard the USNS MONTFORD valuable insight with environmentally sensitive and critical
POINT (T-ESD-1), USNS JOHN GLENN (T-ESD-2), USNS sea-basing operations that will benefit from a FutureWaves™
DAHL (T-AKR-312), and USNS BOB HOPE (T-AKR-300). system.
Representative installations for an LMSR and an ESD are
shown in Fig. 7 and Fig. 8. The May 2015 test was in A. CULEBRA-KOA 2015, Hawaiian Opareas east of Oahu,
conjunction with the Culebra-Koa 2015 (CK15) exercise off May 2015
Oahu [10]. CK15 was a U.S. Pacific Fleet training exercise USNS MONTFORD POINT (T-ESD-1)
designed to demonstrate and increase joint proficiency in USNS DAHL (T-AKR-312)
expeditionary operations. This was the first test of ESMF on
the intended relevant naval platforms and included temporary Primary objectives of ESMF participation in CK15 were to:
system installations on both T-ESD-1 and T-AKR-312. The  Collect data to support subsequent system development,
October 2015 testing was in support of PMS-385 bow-thruster evaluation, and demonstration.
testing for T-ESD-2 and included installations on that ship and  Demonstrate the system's capability to support the
T-AKR-300. The July 2016 ‘experimental’ operational operator for challenging operations associated with sea-
support included a single-ship ESMF system installed on T- basing.
ESD-2. The lack of a system on the second S2S ship did The test focused on verifying new aspects of the ESMF
preclude coherent time-specific motion predictions during Multi-Ship System including:
 Large-Ship Operations
 Dual/Multi Radar Implementation
 Multi-ship data communications
 Multi-sourced Wave Reconstruction
 Reduced Order Model and in situ tuning processes for
forces & moments due to ship-to-ship connectors

Fig. 7. USNS BOB HOPE (T-AKR-300) [top], Wave-Radar Layout Fig. 8. USNS MONTFORD POINT (T-ESD-1) [top], Wave-Radar
[middle], as installed on House-Top [bottom] Layout [middle], system component locations [bottom].
 The ESMF program had intended CK15 to be a multi-ship Course and Speed Recommendation Model (CSRM). For the
data collection with S2S operations demonstrating the utility of ESMF test day ONR requested the ships establish a S2S
the system. While the system was exercised frequently in configuration and then to vary the base course to compare
single-ship mode, and used extensively by ship operators to CSRM predictions to observed motions. Single ship validation
understand sea states and resultant ship motions, the S2S for the CSRM were also coordinated for the ESD on October 19,
operations were precluded by equipment issues and excessive 2015. Fig. 9 provides an example directional wave spectra using
sea states (>SS3) beyond the present operating limitations for the FutureWaves™ system and the corresponding CSRM output
ESD/LMSR S2S operations. Nevertheless, an abundance of for single ship calculations. Four headings at two-knots covered
data was collected that provided further insight into system variation in the predicted roll response. Two factors influence
development, in particular understanding of operator utility. the quality of a CSRM prediction, power spectrum input and
Despite no S2S evolutions, CK15 was a success for the ESMF ship motion model accuracy. The crosses in the top-right of the
program. We were able to operate on two ships simultaneously figure show four of the course/speed combinations that were
and observe limiting conditions for sea-base operations. ESMF executed, and the bottom of the figure shows how the CSRM
program accomplishments for CK15 were: value compared to the measured roll. While more extensive
validation is still required, the results show promise and the
 Conducted first at-sea test of the two-ship system potential utility of FutureWaves™ and the CSRM tool.
 Collected several runs of individual ship data
(ESD and LMSR)
 Tested two-ship system architecture
 Integrated UM/OSU CORR radar and components into a
single combined system
 Under certain conditions, demonstrated that phase
resolved forecasting of large ship motions can be
achieved with high performance
 Received valuable feedback from ESD Captain and
Mooring Master (Ship Handling Expert)
 Demonstrated utility/operation of system displays and
planning tools
 Provided information to ESD Captain which supported
operational planning
 Recognized value of the system to ESD/LMSR sea-
basing operations
Conclusions drawn after CK15 were that the system can
provide the equivalent of a wave buoy ‘array’ for characterizing
sea conditions; the system consolidates METOC, wave, ship
motion, and geo-sit data into a single ‘data-center’ display; and
that although the ‘data-center’ aspects of the system appear most
appealing to the operators, the coherent prediction and statistical
forecasting capabilities were seen as valuable and applicable.
Fig. 9. FutureWaves™ Wave Energy Spectrum, Course and Speed
B. JG15 – Southern California Opareas, October 2015 Recommendations Model (CSRM) [top], and comparison to observed
motions [bottom].
USNS JOHN GLENN (T-ESD-2)
USNS BOB HOPE (T-AKR-300)
Since S2S operations did not occur during the initial multi- C. PALS-16, Southern California Opareas, July 2016
ship ESMF test (CK15), a second opportunity was sought out.
This opportunity presented itself in the fall of 2015 as PMS-385 USNS JOHN GLENN (T-ESD-2)
was scheduling a bow-thruster test for the second ESD ship, USNS GYSGT FRED W. STOCKHAM (T-AK 3017)*
USNS JOHN GLENN (T-ESD-2). ONR named the test ‘JG15’ *system not installed
for reference. JG15 took place off of Oceanside CA from With the success of the 2015 sea testing, APS was contracted
October 18 to 26, 2015. Significant sea heights were generally to re-install the developmental prototype system and conduct
a meter or less with the wind exhibiting speeds from 0 to 10 ‘experimental’ support for the PALS-16 naval exercises held
knots and a diurnal variation in direction for most of the testing offshore near Camp Pendleton, CA. The U.S. Pacific Command
period. This presented challenging conditions for radar-based Amphibious Leaders Symposium 2016 (PALS-16) brings
wave retrieval during the times of low wind (<5 knots) together senior leaders of allied and partner nations to discuss
conditions, typically in the early morning. key aspects of amphibious operations, crisis response, sea-
While JG15 was for the most part a PMS-385 controlled test basing, capability development and interoperability [11]. The
specifically for bow-thruster testing, ONR was afforded the prototype system employed was a variant of the ESMF Fly-
opportunity to have a dedicated day for ESMF testing. For this Away-Kit (FAK) which APS is under tasking from ONR to
day, October 25, 2016, the test team focused on validation of the develop for rapid deployment capability. The system installed
for PALS-16 included a single AWSR installed on the foremast,
a single IMU/GPS, a bridge display, and a streamlined processor
suite in the ship’s service trunk.
Overall the system performed very well and anecdotally
provided useful guidance to the ship’s master, bridge watch, and
onboard ship handling experts (mooring masters). An example
user-interface from the system bridge display is shown in Fig.
10. The system was frequently consulted throughout the exercise
and accolades were received regarding its benefit to safety of
operations.

Fig. 11. Location of ship/radar and nearest deep-water CDIP buoy (#045)
during the JG15 at-sea system test.

the southeast of the ship’s operations area as shown in Fig. 11,

however #045 was in deeper water and chosen for the
comparisons. The locations of the shipboard radars used for
comparison are also shown in the figure and demonstrate that
some differences are possibly due to the geographic separation.
Fig. 12 shows component wave heights and peak-direction
measured by the AWSR on board the T-ESD-2 and by CDIP
wave-buoy #045 on Oct. 21st at 15:31 UTC. During this period
of time there was 12-16 second period swell coming from the
south-south-west and wind-waves from the west with periods of
10 seconds and less. The bar-chart in the figure compares the
component-wave heights, derived from the directional
spectrums of the AWSR and the CDIP buoy, over several
frequency bands ranging from periods of 22 seconds down to 6
seconds. The bottom plot of the figure compares the peak wave
direction in each frequency bands as measured by the AWSR
and the CDIP buoy. Since Fig. 12 only shows the component
wave heights and peak direction at one specific time, Fig. 13 is
included to demonstrate how the component wave heights
evolved over the course of the week-long sea-test as measured
Fig. 10. FutureWaves™ User-Interface by three wave-sensing radars installed both of the JG15 ships
(ESD and LMSR).

IV. PERFORMANCE RESULTS

A. Directional Wave-Energy Spectrum

Using shipboard wave-radar, the directional wave-energy
spectrum and measured sea heights can be continuously
measured and reported to the ship’s master and crew in real-
time. To assess the accuracy of the computed radar-based
directional wave energy spectrum and associated sea heights we
made comparisons to nearby Coastal Data Information Program
(CDIP) moored wave-buoys. CDIP is an extensive network for
monitoring waves and beaches along the coastlines of the United
States and employs moored directional wave-buoys in
cooperation with NOAA's National Weather Service and
National Data Buoy Center (NDBC). CDIP reports detailed
wave information at 30-minute intervals including directional
wave-energy spectrum and wave heights delineated by period Fig. 12. Comparison of the component waveheights and peak direction
bands. We compared radar-based measurements obtained as measured by the shipboard radar on Oct. 21st at 15:31 UTC and CDIP
during JG15 to those at corresponding observation times as buoy #045 located ~10 nautical miles to the southeast.
reported by the nearest CDIP wave-buoy. Two CDIP buoys,
#043 and #045, were located about 10 nautical miles to
 B. Time-specific Ship Motion Forecast
Fig. 14 and 15 show an example of a 30-second forecast
signal spectrum and time-series (that is, a time-series of ship
motions that is forecasted 30 seconds into the future from the
time the forecast was generated) for ESD’s heave, roll, and
pitch over a 20-minute period on Oct 22, 2015 at about 1800
UTC. Fig 14 shows the corresponding spectrums of the
ground-truth (in red) and “30s-forecast” (in blue) signals over
the same 20-miute interval, which are useful for analyzing
performance. During this point in time the ESD was
performing S2S operations with the LMSR, both of which had
<1.0 knots of forward speed and a heading ~030oT. The wave
energy was mostly 12-16 second period swell coming from the
south-south-west, essentially parallel to the ships’ heading. The
correlation between the ground-truth and forecasted heave
motion signals shown is about 0.80. However, closer
inspection of the plot reveals that during some periods of time,
the correlation between the forecasted motion and the
measured motion appears to be very high and at other times the
correlation appears to be lower. Furthermore, it is of interest to
determine how well the system can forecast at intervals other
than 30 seconds, such as 15 or 60 seconds.

Fig. 14. The spectrums of the ground-truth

Fig. 13. Evolution of the component wavehieght in
(red) and “30s-forecast” (blue) signals over a 20-
three frequency bands as measured by three wave-sensing
miute data-collection interval, on Oct. 22, 2015
radars on two different ships and compared to CDIP buoy
at 1800 UTC.
#045 during the week-long JG15 sea-test.

Fig. 15. Plots comparing the ground-truth ship-motion time-series as measured by the on-board ship motion sensor
(red) to the computed “30s-forecast” (blue) of that motion over a 20-minute data-collection, on Oct. 22, 2015 at 1800 UTC.
 Fig. 16 shows another more detailed way of visualizing how Numerous plots as shown in Figs 14, 15, and 16 for each of
the motion-forecast performance is changing over the elapsed the ship motion parameters have been produced for the myriad
time from the beginning of a data-recording run, and over the of data obtained during the ESMF program sea tests. In general,
desired forecast time. Displayed are two-minute running- computed motion-forecast performance results out to 60
average correlations between the computed forecast and the on- seconds show correlations are on the order of 0.80~0.90 for
board motion sensor (ground-truth) over the same time period as cases where the magnitude of the motions are of operational
Fig. 14 (heave in this case). The vertical axis, elapsed time in the concern.
run, is the same for both sub-plots. The horizontal axis of the
left-subplot is heave in meters, and the horizontal of the right V. SUMMARY
color plot is computed forecast time, starting at 0 seconds and
going out to 60 seconds. The left sub-plot shows a 2-minute After extensive at-sea testing through the ONR ESMF
running average of the RMS ship heave recorded by the on- program, a viable system, FutureWaves™, has emerged that can
board sensor (red) and computed forecast by the ship motion provide real-time wave characterization and predict vessel
forecast system (blue). The color plot shows a 2-minute running motions time-series by employing wide-area ocean-surface
average of the correlation between the ground-truth ship motion Doppler radar operational-observations. FutureWaves™ is a
signal and the computed forecast of ship motion (with the method and system for sensing, decomposing, and
location on the x-axis corresponding to the forecast time) with reconstructing a wide-area wave-field in the proximity of
dark red being a perfect (1.0) correlation and dark blue being no waterborne vessels and then using the wave components predict
(0.0) correlation. As one looks towards the right of the color specific wave events at or near the vessels in the very near future
correlation plot, the value tends to decrease with increasing (out to minutes) and the resultant time-domain motions of the
forecast time as one might expect with future predictions. vessels. The time and geographic (spatiotemporal) location of
Looking at a vertical “slice” at 30-second x-axis mark of the specific forecasted waves (phase-resolved ocean surface height)
color correlation plot shows how well the 30-second forecast is reconstructed and used as information for user operational
correlated to the ground-truth ship motions. Now, comparing guidance along with the spectral (direction, period and energy)
the left sub-plot of RMS values to the correlation color plot on information via a display or sent directly to an external device
the right, shows that when the RMS ship motions are small (e.g. or system where it can be utilized for further processing where
at elapsed times of ~600, ~1100, and ~1500 seconds), the spatiotemporal or spectral wave information is required.
correlation of computed forecast to measured motion decreases The at-sea demonstrations have borne out the relevance and
and correspondingly system performance is degraded. This is utility of FutureWaves™ in enhancing safety of critically
expected because predicting small ship motions, and sensitive operations and in expanding operational envelopes
correspondingly decreased wave forcing, is difficult. where the ocean environment can have significant impact.
Fortunately, if ship motions are small there is much less concern FutureWaves™ provides the ship master and crew with the
for operations. advanced warning information necessary to assess and safely
respond to a challenging ocean environment. Such capability is
key to the safe execution of crucial offshore operations
involving the transfer of personnel and equipment at sea.

ACKNOWLEDGMENT
This material is based upon work supported by the U.S.
Office of Naval Research (Dr. Paul Hess, Program Manager)
under Contract No. N00014-11-D-0341. Any opinions, findings
and conclusions or recommendations expressed in this material
are those of the author(s) and do not necessarily reflect the views
of the U.S. Office of Naval Research.

REFERENCES
[1] Office of Naval Research, “Environmental and Ship Motion Forecasting
(ESMF)”, ONR BAA Announcement #10-019, 2010.
[2] “U.S. Navy Program Guide”, 2015. Department of the Navy, Washington,
D.C., http://www.navy.mil/ navydata/policy/seapower/npg15/top-
Fig. 16. Two-minute running-average correlation between the computed npg15.pdf
motion forecast and the on-board motion sensor (ground-truth) heave over [3] Abbott, M., “MSC's USNS Stockham, USNS Montford Point perform
the same time period as Fig. 14. The left sub-plot shows a 2-minute running 'skin-to-skin' maneuver”, Story Number: NNS160316-04, navy.mil,
average of the RMS ship heave in meters (0.0-0.15) recorded by the on-board March 16, 2016.
motion sensor (red) and the computed forecast (blue). The color plot shows
[4] Connell, B.S.H., Rudzinsky, J.P., Brundick, C.S., Milewski, W.M.,
a 2-minute running average of the correlation (with the location on the x-axis
Kusters, J.G., and Farquharson, G.F., “Development of an Environmental
corresponding to the forecast interval) with dark red being a perfect and Ship Motion Forecasting System”, Proceedings of OMAE 2015, St.
correlation (1.0) and dark blue being no (0.0) correlation. John’s, Newfoundland, Canada, 2015.
[5] Milewski, W.M., Connell, B.S.H., Vinciullo, V.J., and Kirschner, I.N.,
“Reduced Order Model for Motion Forecasts of One or More Vessels”,
Proceedings of OMAE 2015, St. John’s, Newfoundland, Canada, 2015.
[6] Cummins, W. E. The impulse response function and ship motions. No.
DTMB-1661. David Taylor Model Basin Washington DC, 1962
[7] Farquharson, G.,,Mower, J., and Plant, W. (APL-UW), Rudzinsky, J.,
Kusters, J., Cockrell, K., Frazer, B. and Connell, B. (APS) “Wave
Sensing Radar and Wave Reconstruction”, SoMAR-3, University of
Washington, June 2015
[8] Plant, W. J., & Farquharson, G. (2012). Wave shadowing and
modulation of microwave backscatter from the ocean. Journal of
Geophysical Research: Oceans (1978–2012), 117(C8).
[9] Plant, W. J., & Keller, W. C. (1990). Evidence of Bragg scattering in
microwave Doppler spectra of sea return. Journal of Geophysical
Research: Oceans (1978–2012), 95(C9).
[10] U.S. Navy, Commander Naval Surface Force, U.S. Pacific Fleet
“CULEBRA_KOA 2015”,
http://www.public.navy.mil/surfor/Pages/CulebraKoa2015.aspx#.V4171
PkrJhF
[11] Kimbrel, O., Sgt. U.S. Marine Corps Forces, Pacific “PALS-16
COMMENCES TO BUILD INDO-ASIA PACIFIC PARTNERSHIPS”,
July 2016