Forecasting contrasting coastal and estuarine hydrodynamics with OPENCoastS

https://doi.org/10.1016/j.envsoft.2021.105132Get rights and content

Highlights

  • Web platform to generate on-demand forecast systems for multiple circulation options.

  • Open source software at https://opencoasts.ncg.ingrid.pt/, using EOSC resources.

  • Forced by GFS, WRF or ARPEGE (atmosphere), CMEMS or FES2014 (ocean) and WW3 (waves).

  • Data/model comparison with Copernicus Sentinel data and in-situ EMODnet field data.

Abstract

Robust and accurate coastal forecasts require models to represent the relevant processes, prediction computational tools and reliable computational resources. OPENCoastS is a free, open-source WebGIS platform to develop on-demand hydrodynamic forecast systems that started as a simple 2D engine. OPENCoastS provides a visualization and download interface with in-situ and Sentinel satellite data comparison. 2D tidal, 2D wave & current interaction and 3D baroclinic flows are now included, forced by several atmospheric, oceanic and riverine forcings.

Four applications demonstrate OPENCoastS’ capacity. The prediction of the 2020 typhoon season in Taiwan illustrates the use of the service using only large-scale public data. An application to the Bay of Biscay shows the importance of waves on extreme water levels during storms. A nearshore deployment in Figueira da Foz harbor assesses the impact of bathymetry on coupled wave and current circulation. 3D baroclinic circulation forecasts in Tagus estuary are validated by independent data.

Introduction

Coastal forecast systems provide predictions of environmental variables at time scales of a few days. Environmental variables include water levels, velocities, wave parameters, pollutant concentrations and sediment fluxes. These forecast systems have a wide range of applications in coastal and harbor management (Viegas et al., 2009; Bedri et al., 2014; Oliveira et al., 2015), civil protection (Breivik and Allen, 2008; Fortunato et al., 2017a; Ferrarin et al., 2019; Stokes et al., 2020), navigation (Orseau et al., 2021), military operations and recreation (e.g. windguru.cz, magicseaweed.com). Some of these forecast systems cover spatial scales from oceans and regional seas to coastal regions, using downscaling techniques over structured and unstructured grids (Trotta et al., 2016, 2021). They are developed and operated by research centers, meteorological and hydrographic organizations, harbor administrations and private companies.

In spite of the growing development of coastal forecast systems, their dissemination remains limited by their implementation and maintenance costs. These costs are mostly associated with very specialized human resources, with backgrounds in both numerical modeling and information technologies, and also with dedicated computational resources to guarantee a timely delivery of predictions.

However, several evolutions are paving the way for a drastic increase in the development and adoption of coastal forecast systems. First, higher resolutions, more stable numerical schemes and better parameterizations reduce the need for calibration and the effort required to optimize the numerical parameters. As a result, the skills required from modelers decrease and forecasts become more robust. Second, the growing availability of online near-real time data (e.g., GEBCO, EMODNET), atmospheric forecasts (e.g., GFS, WRF, ARPEGE) and large-scale ocean models (e.g., FES2014, CMEMS, HYCOM) provide free access to the information required to force local forecasts worldwide. Third, large computational infrastructures, both public and commercial, can now provide the computational power to perform demanding simulations without the need to acquire and operate these infrastructures. The European Open Science Cloud (EOSC) and the Partnership for Advanced Computing in Europe (PRACE) are examples of such public infrastructures.

A fourth evolution that can drastically reduce the cost of generating and operating coastal forecast systems is automation. The recent development of Web-based platforms that can simultaneously generate and operate coastal modeling systems with minimal human intervention will reduce the cost of forecast systems, thereby fostering their dissemination. Examples of these tools remain scarce in the coastal and ocean communities. WebMARVL (the Virtual Marine Laboratory, Oke et al., 2016), for setting up ocean circulation and wave models, Delft-FEWS, dedicated to hydrological and coastal flood forecasting (Werner et al., 2013), and OPENCoastS, to generate coastal forecast systems for any location in a few minutes (Oliveira et al., 2020) are the most comprehensive platforms available. OPENCoastS is a user-friendly platform supported by EOSC computational services and resources. It is freely available to all users whereas, for instance, WebMARVL is dedicated to the Australian communities. The original version of the platform described in Oliveira et al. (2020) was however limited to simple physics (i.e., 2D depth-averaged shallow water flows). Now it has matured and addresses more complex flows, including wave and currents interactions and 3D baroclinic flows. The only inputs requested to the users to set up a new deployment are the horizontal grid file and, for the 3D runs, also the vertical grid. The platform is maintained in operation through the use of European Open Science cloud (EOSC) resources and forecasts still take only a few minutes to generate.

This paper aims at demonstrating how forecast systems built using the OPENCoastS service can provide accurate prediction of complex flows in estuarine and coastal environments. “Complex flows” refer here to flows associated with extreme atmospheric events, breaking waves, and strong density gradients, and at scales ranging from tens of meters to thousands of kilometers. Four demonstration examples are presented herein that cover various spatial scales (from basin-wide to estuarine scales), different forcing agents (tides, waves, river flow, wind and atmospheric pressure), applied in distinct geographies (European and Asian coasts). These examples address different scientific questions (e.g., coastal inundation, salinity dynamics in estuaries) and the forcing agents include tides, waves, river flow, wind and atmospheric pressure. The criteria behind the selection of the applications are summarized in Table 1. The evolution of the platform, from its original version to its present capabilities, is also detailed to promote the usage of the service software by other teams. It is now freely available under licence Apache License Version 2.0 at https://gitlab.com/opencoasts/eosc-hub/webportal.

The paper is organized as follows. First, the OPENCoastS platform is briefly described, with an emphasis on the most recent features. Then, the capabilities of the platform to support operational management in coastal systems are demonstrated through four examples of application. In section 3, these examples are used to illustrate and discuss the lessons learned from the first three years of development of OPENCoastS. Finally, the potential and the present limitations of OPENCoastS are discussed and its evolution is anticipated.

Section snippets

Overview of the OPENCoastS service

The OPENCoastS service provides accurate circulation forecasts in any coastal system of choice (Oliveira et al., 2020). This is achieved through the use of the process-comprehensive suite of numerical models provided by SCHISM (Zhang et al., 2016), and of a complex computational web platform. SCHISM was chosen because it encompasses all relevant processes, and the web platform was built to run it seamlessly and automates the whole prediction workflow. This combination provides the users the

Extreme water levels in the coast of Taiwan

The northwestern Pacific Ocean is the most active tropical cyclone basin on Earth (Elsner and Liu, 2003). The most severe of these cyclones, locally known as typhoons, can generate extreme storm surges that can have devastating effects on the shores of the Philippines, China, Taiwan and Japan. Here, we illustrate the generation of a forecast system for the coast of Taiwan with OPENCoastS and its validation using only publicly available data.

Typhoon tracks can be divided into three groups (

Discussion, conclusions and future perspectives

Over the past three years, OPENCoastS has grown from an innovative on-demand platform that addressed simple 2D barotropic forecasts to a powerful tool that solves all circulation options, used by over 400 users and applied on all continents. Most past applications are scientific ones, to understand the importance of processes at a site or to explore the influence of numerical and physical parameters or forcing sources on forecasts, among other goals. Several deployments were also built to

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was funded by the European Commission through the H2020 projects EOSC-hub (Grant Agreement No 777536) and EOSC-Synergy (Grant Agreement No 857647), by Lisboa2020 Operational Program through the INCD project (LISBOA-01-0145-FEDER-022153) and by the Fundação para a Ciência e a Tecnologia through projects MOSAIC.pt (PTDC/CTA-AMB/28909/2017) and AQUAMON (PTDC/CCI-COM/30142/2017). This work made use of results produced with the support of the Portuguese National Grid Initiative; more

References (58)

  • T. Guérin et al.

    A numerical scheme for coastal morphodynamic modelling on unstructured grids

    Ocean Model.

    (2016)
  • W. Huang et al.

    Compounding factors for extreme flooding around galveston Bay during hurricane harvey

    Ocean Model.

    (2021)
  • W.-C. Liu et al.

    Investigating typhoon-induced storm surge and waves in the coast of Taiwan using an integrally-coupled tide-surge-wave model

    Ocean Eng.

    (2020)
  • M.S. Longuet-Higgins et al.

    Radiation stresses in water waves; a physical discussion, with applications

    Deep Sea Res. Oceanogr. Abstr.

    (1964)
  • P.J. Lynett et al.

    Inter-model analysis of tsunami-induced coastal currents

    Ocean Model.

    (2017)
  • K.J. Roberts et al.

    On the automatic and a priori design of unstructured mesh resolution for coastal ocean circulation models

    Ocean Model.

    (2019)
  • K. Stokes et al.

    Forecasting coastal overtopping at engineered and naturally defended coastlines

    Coast. Eng.

    (2021)
  • F. Trotta et al.

    A structured and unstructured grid relocatable ocean platform for forecasting (SURF)

    Deep Sea Res. Part II Top. Stud. Oceanogr.

    (2016)
  • A.S. Valente et al.

    On the observability of the fortnightly cycle of the Tagus estuary turbid plume using MODIS ocean colour images

    J. Mar. Syst.

    (2009)
  • M. Werner et al.

    The Delft-FEWS flow forecasting system

    Environ. Model. Software

    (2013)
  • B. Wolf et al.

    Using the salinity preferences of benthic macroinvertebrates to classify running waters in brackish marshes in Germany

    Ecol. Indicat.

    (2009)
  • Y. Zhang et al.

    A new vertical coordinate system for a 3D unstructured-grid model

    Ocean Model.

    (2015)
  • Y.J. Zhang et al.

    Seamless cross-scale modeling with schism

    Ocean Model.

    (2016)
  • J. Allan et al.

    Maritime tsunami evacuation guidelines for the Pacific Northwest coast of Oregon

    Nat. Hazards

    (2018)
  • Plano de Gestão da Região Hidrográfica do Tejo, Relatório Técnico – Síntese. Ministério da Agricultura, do Mar, do Ambiente e do Ordenamento do Território

    (2012)
  • V. Brotas et al.

    Padrões de variabilidade sazonal e interanual de nutrientes e fitoplâncton no estuário do Tejo. Plano de Ordenamento do Estuário do Tejo, Saberes e Reflexões. Gabinete de Ordenamento do Território, Administração da Região Hidrográfica do Tejo, Ministério do Ambiente, do Ordenamento do Território e do Desenvolvimento Regional

    (2009)
  • J.M. Castanheiro

    Distribution, transport and sedimentation of suspended matter in the Tejo Estuary

  • C. Chen et al.

    Extratropical storm inundation testbed: intermodel comparisons in Scituate, Massachusetts

    J. Geophys. Res. Oceans

    (2013)
  • D.L. Codiga

    Unified Tidal Analysis and Prediction Using the Utide Matlab Functions

    (2011)
  • Cited by (12)

    • Wave-induced mean currents and setup over barred and steep sandy beaches

      2022, Ocean Modelling
      Citation Excerpt :

      Identifying this discrepancy not only reveals the difficulty in measuring the wave setup close the shoreline on steep beaches, but it underlines the need to further develop the capacity of phase-averaged modelling approaches to predict extreme water levels at the shoreline. Indeed, phase-averaged models fully-coupled to oceanic circulation models play a critical role in operational applications or in early-warning systems worldwide (e.g., Gillibrand et al., 2011; Ferrarin et al., 2013; Sembiring et al., 2015; Khan et al., 2021; Oliveira et al., 2021). In this context, the present findings suggest that modelling approaches relying on the Vortex-Force formalism (either 2DH or 3D) should be preferred over the radiation stress-based approach for improved predictions of mean water levels along wave-exposed coastlines.

    • Modelling the contribution of wind waves to Cap Ferret's updrift erosion

      2022, Coastal Engineering
      Citation Excerpt :

      This value falls within the range of values found in the literature for tidal inlets (Bruneau et al., 2011; Orescanin et al., 2016) and was set after calibration tests performed with tidal forcing only. For the tests, the model's open boundary was forced with 16 tidal components from the regional tide model of Bertin et al. (2012), and with wave spectra issued from OPENCoastS′ unstructured WaveWatchIII model (Oliveira et al., 2021; WWIII Development Group, 2016), forced with wind fields from the Climate Forecast System Reanalysis CFSR (Saha et al., 2010). The model's performances were evaluated in terms of modelled elevation at three locations across the tidal inlet (Fig. 4A and B and C), compared with observations from June 2014.

    View all citing articles on Scopus
    View full text