Building Coupled Storm Surge and Wave Operational Forecasting Capacity for Western Alaska

  • Participants

    This project is funded by the Integrated Ocean Observing System, Ocean Technology Transition Project Grant NOAA # NA18NOS0120164.

    Project lead institution: University of Notre Dame

    Principal investigator: Joannes Westerink
    Department of Civil and Environmental Engineering and Earth Sciences

    Collaborating institutions:

    • The University of Texas at Austin
    • NOAA’s Great Lakes Environmental Research Laboratory (GLERL)
    • The Cooperative Institute for Great Lakes Research (CIGLR)
    • Alaska Ocean Observing System (AOOS) and Axiom Data Science, LLC.

    In-kind modeling collaborators:

    • NOAA’s National Centers for Environmental Prediction (NCEP)
    • NOAA’s National Ocean Service (NOS) Coast Survey and Development Laboratory (CSDL)

    Other partners:

    • Alaska Division of Geological & Geophysical Surveys ( DGGS)
    • Western Alaska Land Conservation Cooperative (WALCC)
    • Alaska’s NOAA National Weather Service (NWS) Weather Forecast Offices (WFO).

    See Project Descriptions and Team Bios drop down menu above for participant information.

  • Background

    Alaska’s coast is a unique and irreplaceable natural, social and economic system. The region has a complex geography and highly energetic atmospheric and ocean circulation and wind-wave conditions. This combined with the extensive continental shelf and coastal floodplain leaves many western Alaska communities vulnerable to storm surge and flooding events. Compounding the situation are strong winter storms under varying ocean ice cover that make this a uniquely challenging location for predicting and responding to flood-related hazards. Ice coverage varies dramatically both intra- and inter-annually and ice drift speeds have doubled over the past two decades due to reduced coverage, effectively increasing the atmosphere’s transfer of momentum to the water. Air-sea momentum transfer formulae in the presence of ice utilized by wave and storm surge models have improved over the years, but remain a significant source of uncertainty in models forecasting storm surge in this environment.

    Models rely on in situ observations for validation to ensure the models are reporting accurately. They also depend on accurate inputs for bathymetry, shoreline topography and boundary conditions necessary to initiate the models. Unfortunately, Alaska coasts have historically received less attention than the rest of the continental U.S. in terms of real-world observations, and as a result suffer from a higher degree of uncertainty in terms of understanding coastal water level, current and wind-wave simulation capacity. At present, there are only four year-round verified NOAA National Water Level Network (NWLON) water level stations along the thousands of miles of coastline in western Alaska and the Alaska Arctic.

    NWLON network stations in Alaska. The most recent Unalakleet NWLON Station was installed in 2016 and is shown in pink.

    Western Alaska is also limited in weather observations and replete of nearshore current and circulation information. Regional forecasters and the many communities they serve are severely limited in their assessment of the threat from a specific storm event and have no basis to determine impact risk or evaluate safe evacuation routes and locations. These problems become even more vexing in light of continued diminishing ice conditions in and around fall through winter months, when the most intense storm events occur.

    The situation in Shishmaref is not helped by the fact that in addition to the encroaching ocean swells, the melting of Permafrost (the frozen layer of earth and rock that exists in perennial cold climates) destabilizes the shoreline and makes the earth still more vulnerable to erosion.
    Photo provided for use in the GRID-Arendal resources library by: Lawrence Hislop

  • Filling the Observing Gaps and Motivation for Models

    Observations and models go hand in hand. For one, accurate water level and ice observations are fundamental for storm-surge forecasting, informed emergency response, ecosystem management, safe navigation, and efficient mapping/charting. Portions of Alaska’s remote coastline are among the nation’s most vulnerable to geohazards such as coastal sea level rise, tsunami, extra-tropical storm surge, and erosion. Unfortunately, gaps in active, real-world observations for water level variations and ice conditions are limiting our state’s ability to provide useful marine forecasts and this lack of information is endangering coastal populations and infrastructure.

    In this photo you can see a collapsed block of ice-rich permafrost along Drew Point, Alaska. Coastal bluffs in this region can erode 20 meters/year (~65 feet). USGS scientists continually research the causes of major permafrost thaw and bluff retreat along the Arctic coast of Alaska. In addition, with the loss of sea ice to protect the beaches from ocean waves, salt water inundation the coastal habitats. You can learn more at: USGS Alaska Science Center:
    Photo courtesy of USGS Pacific Coastal and Marine Science Center
    Photo by: Benjamin Jones, USGS

    Western and northern Alaska are difficult and expensive to provide adequate in situ observational capacity because of limited access, seasonal ice, lack of coastal infrastructure and rapid coastal erosion. These conditions can render coastal observing systems using conventional water level sensing technologies unaffordable in much of these areas.

  • The Alaska Water Level Watch

    The Alaska Water Level Watch (AWLW) is a collaborative group working to improve the quality, coverage, and accessibility to water level observations in Alaska’s coastal zone. The AWLW maintains an observing map showing where water level and storm inundation information is available. The most recent updated 2018 map is available on the AWLW website.

    Alaska coastal water level stations map, 2018. Most water level observations come from shorter term tide stations (yellow circles) that were deployed only long enough to establish datums and provide core tidal harmonics for deterministic tide prediction.  They were also deployed during summer, and  do not provide data on fall and winter wind-driven surge events.

    Efforts are progressing across this region to increase water level and weather observing, including working with local tribes and community members to install and maintain equipment used for pre-storm and post-storm water level and shoreline erosion information. However, the need for forecast, nowcast and emergency response information is immediate. Groups working to increase observing capacity are not able to afford the necessary installations to adequately observe the Alaska coastline. Models offer a solution for not only that, but also, models provide us with the ability to predict what to expect with given preconditions. So though community post-storm assessments cannot advise on the nowcast needs, they offer validation and support data for developing forecast models. Increasing observations while improving forecast model capabilities is therefore, mutually beneficial.

  • We already have storm surge models...Why can’t we use those and what else do we need? - Model Motivation

    Though progress has been made on our ability to produce meaningful forecasts nationwide, existing models do not contain the complexity necessary to reproduce observed conditions where they exist in much of the Alaska region, and existing models need to be modified to more accurately predict storm surge conditions, especially for the less well-understood western Alaska coastline. In order to improve hindcasting and forecasting skill for coastal water levels and circulation as well as wind-wave environments, this modeling project seeks to integrate multi-physics high resolution models for all of Alaska that meld coastal long-wave hydrodynamics, wind waves, ice physics, and ocean circulation with the driving meteorology. We will especially focus high mesh resolution on Alaska’s inner shelves, adjacent estuaries and bays and the coastal floodplains, regions where this is crucial in order to obtain accurate model results. Integrating all processes that impact coastal water levels into a single framework combined with a high-resolution representation of the coast will drive more robust and reliable computations.

  • Goal

    The ultimate goal for this project is to deliver an improved coupled surge, wave and ice forecasting capacity to NOAA National Centers for Environmental Protection (NCEP) and NOAA National Ocean Service (NOS) Coastal Survey Development Lab. The wave-surge modeling system will be utilized for the purpose of forecasting real time risk assessment for western Alaska and its communities.

    An integrated storm wave and surge coastal ocean model will be developed that focuses on refining and transitioning a recently completed high-fidelity research model to a real time operational forecasting capacity for the coasts of Western Alaska. The proposed system will be based on recent advances in coupled surge, wave, ocean circulation, and sea ice models including ADCIRC, WAVEWATCH III, Global RTOFS, and CICE in the northern Pacific Ocean, Bering, Chukchi and Arctic Seas.

  • Technical Details on Our Modeling Approach

    The integration of multiple physical processes spanning across the energy spectrum of the ocean and the application of high localized mesh resolution to correctly resolve these processes are at the heart of this project. We will couple the ADCIRC, WAVEWATCH III, Global RTOFS, and CICE models through ESMF/NUOPC to compute surge and tides, wind waves, ocean currents and sea ice properties. Each model will compute select processes/information and the linkages will inform the other models so that the combined total energy of the ocean can be much better accounted for. The high-resolution unstructured mesh ADCIRC model will cover all Alaskan waters, including the Gulf of Alaska and the Bering, Chukchi and Beaufort Seas. The resulting ALaska Coastal Ocean Forecast System (ALCOFS) is illustrated in the following figure showing linkages and interactions between model components.

    The integrated ALCOFS (ALaska Coastal Ocean Forecast System) showing linkages and interactions between model components.

    The ADCIRC models will evolve from an existing high-fidelity storm surge model for Alaska and will compute tides and storm surge. We are focusing high-resolution on steep topographic continental shelf breaks, the Aleutian Islands, the inner shelf, estuaries and bays, and Alaska’s coastal floodplain, where geometric and topographic details are vital for solution accuracy. The ADCIRC models will be tested within a global shell with global coverage with the high-resolution Alaska model as an integrated inset, as illustrated in the diagram below. This global strategy is advantageous since it eliminates open water boundary conditions from the solution and has robustness and accuracy advantages.

    Seamlessly merging the Alaskan regional mesh into a global mesh.

    ADCIRC will be implemented in two dimensional barotropic form appropriate for computing wind, atmospheric pressure as well as tidally driven circulation on continental shelves and inland regions. The two-dimensional implementation reduces computational expense and allows for high mesh resolution to be applied on the inner shelf and coastal regions. To account for baroclinicity and its flow features at minimal expense, the baroclinic pressure gradient terms in ADCIRC will be driven by temperature and salinity fields from the much coarser grid three dimensional Global RTOFS model. The baroclinic field data from Global RTOFS will also be used to drive internal tide dissipation terms in ADCIRC. It is computationally advantageous to use the coarser three-dimensional Global RTOFS model as the internal model driver for ADCIRC instead of using ADCIRC’s own internal baroclinic module since the high-resolution inner shelf mesh would add prohibitive computational expense when resolving on the order of a hundred vertical layers needed to compute the temperature and salinity fields. Thus, the large-scale ocean baroclinic ocean physics, which can drive coastal setups and setdowns, can be directly incorporated into the ADCIRC hydrodynamics at very little expense since Global RTOFS is already running operationally.

    ADCIRC will be two way coupled to WAVEWATCH III, forcing wave radiation stress induced setup and currents in ADCIRC and incorporating the effects of time varying water column depth and currents in WAVEWATCH III. This integrates the wind wave and long wave spectra although the infra-gravity wave portion of the spectrum is still missing. Nonetheless, coastal water levels, currents, and wind waves are much more accurately computed than without the coupling.

    Coupling of a regional CICE model to ADCIRC, WAVEWATCH III, and Global RTOFS will inform ADCIRC and WAVEWATCH III of the ice coverage impacting generation and dissipation terms in both models. Specifically, this coupling will compute wave dissipation as waves penetrate into ice covered regions which will then drive additional wave radiation stress terms in ADCIRC, forcing additional currents and set-up. Furthermore, ice coverage data such as concentration, thickness and ice characteristics will allow improved wind to ice to ocean stresses to be computed either directly or parametrically. To reduce uncertainty in computing the effect of sea ice in the circulation model, we will blend and assess a parameterization-based approach and direct simulation and the interaction of the sea ice with air and water.

  • Model Outcomes and Transition to Operational Capacity

    We will develop products that evolve in progressive refinements to the existing regional model and will include a lighter operational model and a high-resolution version used to quantify the errors associated with the mesh resolution of the operational model. We will prototype end to end models with forecasts, hindcasts, and uncertainty quantification and provide an assessment of model component interactions and impact on solution fidelity and runtime. All products will be publicly available through AOOS.

  • Complementary Efforts - Other Regional Activities and Observations

    Western and Northern Alaska Regional Data and Observations Manifest: There are many sources of information in Western Alaska, which this data manifest attempts to list. Many data sources are not federal or producing real time data, making some data not as easily discovered. This document is being maintained throughout the project (updated quarterly or as needed), to include all data sources, but primarily to provide links to sources of data relevant to this project but that are less well known. One can visualize most of what is available in terms of water level on the AWLW Alaska Coastal Water Level Sensors Map and other data resources directly on the AOOS Ocean Data Explorer’s real-time sensor map and historical sensor catalog.

    Sample map showing both real time sensors – wave observations on 19 August 2019, and historical sensors – water level data heatmaps that can be used to identify valuable data resources for model validation.

    Central Beaufort Sea Wave and Hydrodynamic Modeling Study – Bureau of Ocean Energy Management sponsored project: Renewed interest in nearshore oil exploration and production in the central Beaufort Sea has created a general need to advance understanding of the existing wind, wave, current, storm, and sediment dynamics conditions in the central and eastern Beaufort Sea coastal region. The particular study was prompted by the 2015 submission of a Development and Production Plan (DPP) by Hilcorp Alaska, LLC (HAK) to develop the North Slope Liberty prospect in the eastern Beaufort Sea, east of Prudhoe Bay in the Stefansson Sound region. Historical and new observations (waves, currents, pressure, meteorology and water levels) and model information can be leveraged for relevant applications during this project. For more project information and links to data, visit the project webpage, which has a link to a data portal (under development):

    The US Geological Survey: USGS is currently developing a coupled wave/hydrodynamic-sediment transport model for application to their ongoing work on coastal erosion in the Kaktovik region. For more details, please visit:

    Long-term Ecological Research (LTER) Beaufort Sea Lagoons: This 5+ year NSF funded LTER effort is aimed at understanding how exchange between barrier island fringed lagoons (including Stefansson Sound) and areas offshore of the barrier islands controls ecosystems in the sheltered lagoon systems. Sensors deployed as part of the LTER can be leveraged for this project. For more information on the LTER Project, refer to: