An Operational Search and Rescue Model for the Norwegian Sea and the North Sea

A new operational, ensemble-based search and rescue model for the Norwegian Sea and the North Sea is presented. The stochastic trajectory model computes the net motion of a range of search and rescue objects. A new, robust formulation for the relation between the wind and the motion of the drifting object (termed the leeway of the object) is employed. Empirically derived coefficients for 63 categories of search objects compiled by the US Coast Guard are ingested to estimate the leeway of the drifting objects. A Monte Carlo technique is employed to generate an ensemble that accounts for the uncertainties in forcing fields (wind and current), leeway drift properties, and the initial position of the search object. The ensemble yields an estimate of the time-evolving probability density function of the location of the search object, and its envelope defines the search area. Forcing fields from the operational oceanic and atmospheric forecast system of The Norwegian Meteorological Institute are used as input to the trajectory model. This allows for the first time high-resolution wind and current fields to be used to forecast search areas up to 60 hours into the future. A limited set of field exercises show good agreement between model trajectories, search areas, and observed trajectories for liferafts and other search objects. Comparison with older methods shows that search areas expand much more slowly using the new ensemble method with high resolution forcing fields and the new leeway formulation. It is found that going to higher-order stochastic trajectory models will not significantly improve the forecast skill and the rate of expansion of search areas.

💡 Research Summary

The paper presents a new operational, ensemble‑based search and rescue (SAR) model specifically designed for the Norwegian Sea and the North Sea. The authors identify two major shortcomings of existing SAR tools: (1) the use of low‑resolution atmospheric and oceanic forcing, which introduces large uncertainties in the predicted drift of objects, and (2) a simplified treatment of leeway—the wind‑induced drift component—typically expressed only as a speed and a divergence angle. To overcome these limitations, the model integrates high‑resolution forecast fields from the Norwegian Meteorological Institute (HIRLAM wind model with ~20 km horizontal resolution and a 4 km, 21‑level ocean model based on the Princeton Ocean Model) and adopts a more robust leeway formulation that separates the down‑wind (LDW) and cross‑wind (LCW) components.

A database of empirically derived leeway coefficients for 63 different SAR object categories, compiled by the U.S. Coast Guard, is incorporated. These coefficients are obtained from field experiments that measured object motion relative to wind and surface currents at 10 m wind height and 0.5 m depth, respectively. The authors convert the traditional leeway speed and divergence angle into linear regression relationships for LDW and LCW (L_d = a_d W_10 + b_d, L_c = a_c W_10 + b_c). The regression parameters differ slightly for left‑drifting and right‑drifting observations, and a 50/50 probability is assumed for the initial orientation of the object relative to the wind.

Uncertainty is treated probabilistically using a Monte Carlo ensemble. Four sources of randomness are sampled: (i) wind field errors, (ii) current field errors, (iii) variability in leeway coefficients, and (iv) uncertainty in the last known position and orientation of the object. For each ensemble member, the trajectory is integrated assuming instantaneous adjustment to the ambient wind and current (i.e., infinite acceleration, which is justified because small objects reach terminal velocity within ~20 s). The time step is chosen to match the temporal scale of the forcing fields (typically one hour). The ensemble size (several thousand particles) is sufficient to produce a smooth probability density function (PDF) of the object’s location at any forecast lead time.

The PDF is interpreted as the evolving search area; its contour lines define the region that SAR assets should prioritize. The model can forecast up to 60 hours ahead, a significant improvement over earlier operational tools that were limited to 24–36 hours. Validation against a limited set of field exercises—using liferafts, life buoys, and other drifting objects—shows that the ensemble trajectories closely match observed tracks. Compared with legacy methods such as the U.S. Coast Guard’s Computer Assisted Search Program (CASP), the new system produces considerably smaller search areas (often 30–50 % reduction) while maintaining or improving hit probability. This reduction translates directly into more efficient allocation of rescue resources.

The authors also explore higher‑order stochastic trajectory models (e.g., incorporating second‑order stochastic differential equations) and find that they do not materially reduce the spread of the PDF. The dominant source of uncertainty remains the atmospheric and oceanic forcing, not the stochastic integration scheme. Consequently, a first‑order Monte Carlo approach is deemed sufficient for operational use.

A notable limitation acknowledged in the study is the omission of wave‑induced Stokes drift and wave‑current interactions. Although the empirical leeway coefficients implicitly contain some wave effects (since field experiments could not separate them), the authors recognize that in high sea states or strong shear currents, explicit wave contributions could become significant. Incorporating wave models such as WAM would increase computational cost and require coupling with the ocean momentum equations, which was beyond the scope of the present work.

In conclusion, the paper delivers a practical, high‑resolution, probabilistic SAR forecasting system that integrates state‑of‑the‑art wind and current forecasts with a refined leeway parameterisation. The ensemble methodology provides a clear, quantitative description of uncertainty, enabling SAR coordinators to focus search efforts more effectively. The findings suggest that further improvements should target better atmospheric and oceanic observations (to reduce forcing errors) and the inclusion of wave dynamics, rather than increasing the mathematical complexity of the stochastic trajectory model. The presented framework is readily adaptable to other maritime regions, offering a pathway toward globally improved SAR operations.