Ocean Solutions, Earth Solutions
()
About this ebook
Related to Ocean Solutions, Earth Solutions
Related ebooks
GIS for Science, Volume 3: Maps for Saving the Planet Rating: 0 out of 5 stars0 ratingsProspects for Resilience: Insights from New York City's Jamaica Bay Rating: 0 out of 5 stars0 ratingsPreserving Our Planet: GIS for Conservation Rating: 0 out of 5 stars0 ratingsRestoring Neighborhood Streams: Planning, Design, and Construction Rating: 0 out of 5 stars0 ratingsRestoring Streams in Cities: A Guide for Planners, Policymakers, and Citizens Rating: 0 out of 5 stars0 ratingsThe International Geodesign Collaboration: Changing Geography by Design Rating: 0 out of 5 stars0 ratingsGIS for Science, Volume 1: Applying Mapping and Spatial Analytics Rating: 0 out of 5 stars0 ratingsRemote Sensing and GIS in Ecosystem Management Rating: 0 out of 5 stars0 ratingsProtecting the Places We Love: Conservation Strategies for Entrusted Lands and Parks Rating: 0 out of 5 stars0 ratingsEnvironmental Restoration: Science And Strategies For Restoring The Earth Rating: 0 out of 5 stars0 ratingsPlanning for Biodiversity: Issues And Examples Rating: 0 out of 5 stars0 ratingsGIS for Surface Water: Using the National Hydrography Dataset Rating: 0 out of 5 stars0 ratingsValuing Place and Purpose: GIS for Land Administration Rating: 0 out of 5 stars0 ratingsPrepare, Respond, Renew: GIS for Wildland Fire Rating: 0 out of 5 stars0 ratingsMap Use: Reading, Analysis, Interpretation Rating: 0 out of 5 stars0 ratingsFinding a Sustainable Balance: GIS for Environmental Management Rating: 0 out of 5 stars0 ratingsGreen Infrastructure: Linking Landscapes and Communities Rating: 0 out of 5 stars0 ratingsStrategic Green Infrastructure Planning: A Multi-Scale Approach Rating: 0 out of 5 stars0 ratingsRoad Ecology: Science and Solutions Rating: 5 out of 5 stars5/5Fieldwork Handbook: A Practical Guide on the Go Rating: 0 out of 5 stars0 ratingsDealing with Disasters: GIS for Emergency Management Rating: 4 out of 5 stars4/5Resilient Communities across Geographies Rating: 0 out of 5 stars0 ratingsKnowing Nature: Conversations at the Intersection of Political Ecology and Science Studies Rating: 0 out of 5 stars0 ratingsSoil Physical Measurement and Interpretation for Land Evaluation Rating: 0 out of 5 stars0 ratingsWildlife-Habitat Relationships: Concepts and Applications Rating: 0 out of 5 stars0 ratingsImagery and GIS: Best Practices for Extracting Information from Imagery Rating: 4 out of 5 stars4/5The Historical Ecology Handbook: A Restorationist's Guide to Reference Ecosystems Rating: 0 out of 5 stars0 ratingsClimate Change and U.S. Cities: Urban Systems, Sectors, and Prospects for Action Rating: 0 out of 5 stars0 ratingsMarine Conservation Biology: The Science of Maintaining the Sea's Biodiversity Rating: 5 out of 5 stars5/5Understanding GIS: An ArcGIS® Pro Project Workbook Rating: 0 out of 5 stars0 ratings
Technology & Engineering For You
The Big Book of Maker Skills: Tools & Techniques for Building Great Tech Projects Rating: 4 out of 5 stars4/5The Art of War Rating: 4 out of 5 stars4/5The Art of War Rating: 4 out of 5 stars4/5Longitude: The True Story of a Lone Genius Who Solved the Greatest Scientific Problem of His Time Rating: 4 out of 5 stars4/5A Night to Remember: The Sinking of the Titanic Rating: 4 out of 5 stars4/5Logic Pro X For Dummies Rating: 0 out of 5 stars0 ratingsArtificial Intelligence: A Guide for Thinking Humans Rating: 4 out of 5 stars4/5The Big Book of Hacks: 264 Amazing DIY Tech Projects Rating: 4 out of 5 stars4/5Ultralearning: Master Hard Skills, Outsmart the Competition, and Accelerate Your Career Rating: 4 out of 5 stars4/580/20 Principle: The Secret to Working Less and Making More Rating: 5 out of 5 stars5/5Vanderbilt: The Rise and Fall of an American Dynasty Rating: 4 out of 5 stars4/5The Invisible Rainbow: A History of Electricity and Life Rating: 5 out of 5 stars5/5The Fast Track to Your Technician Class Ham Radio License: For Exams July 1, 2022 - June 30, 2026 Rating: 5 out of 5 stars5/5Death in Mud Lick: A Coal Country Fight against the Drug Companies That Delivered the Opioid Epidemic Rating: 4 out of 5 stars4/5The Systems Thinker: Essential Thinking Skills For Solving Problems, Managing Chaos, Rating: 4 out of 5 stars4/5The Wuhan Cover-Up: And the Terrifying Bioweapons Arms Race Rating: 5 out of 5 stars5/5Broken Money: Why Our Financial System is Failing Us and How We Can Make it Better Rating: 5 out of 5 stars5/5Selfie: How We Became So Self-Obsessed and What It's Doing to Us Rating: 4 out of 5 stars4/5The 48 Laws of Power in Practice: The 3 Most Powerful Laws & The 4 Indispensable Power Principles Rating: 5 out of 5 stars5/5Mother of God: An Extraordinary Journey into the Uncharted Tributaries of the Western Amazon Rating: 4 out of 5 stars4/5How to Disappear and Live Off the Grid: A CIA Insider's Guide Rating: 0 out of 5 stars0 ratingsA History of the American People Rating: 4 out of 5 stars4/5Seeing Further: The Story of Science and the Royal Society Rating: 4 out of 5 stars4/5After the Affair, Third Edition: Healing the Pain and Rebuilding Trust When a Partner Has Been Unfaithful Rating: 4 out of 5 stars4/5Smart Phone Dumb Phone: Free Yourself from Digital Addiction Rating: 0 out of 5 stars0 ratingsThe Homeowner's DIY Guide to Electrical Wiring Rating: 5 out of 5 stars5/5
Reviews for Ocean Solutions, Earth Solutions
0 ratings0 reviews
Book preview
Ocean Solutions, Earth Solutions - Dawn J. Wright
Introduction
Dawn J. Wright
The ocean dominates the surface of the earth and greatly affects our daily lives. It regulates Earth’s climate, plays a critical role in the hydrologic cycle, sustains a large portion of Earth’s biodiversity, supplies food and mineral resources, constitutes an important medium of national defense, provides an inexpensive means of transportation, is the final destination of many waste products, is a major location of human recreation, and inspires our aesthetic nature. However, it also carries with it the threat of deadly tsunamis and hurricanes, industrial accidents, and outbreaks of waterborne pathogens. (National Academy of Sciences National Research Council 2011)
Our ability to measure change in the ocean is increasing, not only because of improved measuring devices and scientific techniques but also because new GIS technology aids us in better understanding this dynamic environment. The domain of GIS has progressed from applications that merely collect and display data to complex simulation and modeling and the development of new research methods and concepts.
GIS has traditionally provided effective technological solutions to the integration, visualization, and analysis of heterogeneous, georeferenced data on land. As early as the late 1980s, Esri began to make significant progress in the application of ocean GIS solutions to nautical charting, commercial shipping, oil spill response, and defense and intelligence. Recent years have seen GIS increasingly used in ocean science, conservation, and resource management communities, in large part because of the Esri Ocean GIS initiative launched in 2012 (Wright 2012). Many challenges remain, especially regarding inconsistencies in ocean data models, formats, standards, tools, services, and terminology.
Toward addressing these challenges, Esri held a one-time-only Oceans Summit at its headquarters in Redlands, California, in November 2012 (Wright 2013). Intermediate to advanced ocean GIS analysts and developers attended this invitation-only, high-level strategic workshop. Their goal was to help Esri move forward in its approaches to ocean-centric software, associated data formats, tools, workflows, and computing platforms. The success of that summit led to the annual Esri Ocean GIS Forum, open to all and first held in November 2013 (Pratt 2013). Presentations at the forum cover a wide range of topics, including ocean exploration and science, coastal management and marine spatial planning, coastal resilience and conservation, hydrographic surveying, commercial ship tracking, and closing technological gaps in multidimensional data handling and analysis.
Attendees have lauded these earnest efforts to apply GIS to the ocean (the open ocean as well as on or near the coast) and to resolve a range of technological gaps. Going forward, they also understand that the health and sustainability of the planet still lies in the balance. With water covering 71% of the planet’s surface, it is clear that the ocean is critical to our lives, our energy, and our economy. This vital role is increasingly evident in the latest warnings from the UN Intergovernmental Panel on Climate Change (2013), the Third US National Climate Assessment (Melillo et al. 2014), the launch of the US White House Climate Data Initiative, various calls to action from the academic community (e.g., Liu et al. 2015; Lubchenco and Grorud-Colvert 2015), and the search for Malaysia Airlines Flight MH370 (Smith and Marks 2014; still ongoing at the time of this writing). Not surprisingly, the nongovernmental organization (NGO) sector has also recognized the need for action. The Clinton Foundation’s Blue Guardians, the XPRIZE Foundation (and its awarding of the Wendy Schmidt Ocean Health XPRIZE) (XPRIZE Foundation 2015), the Ocean Health Index initiative (Halpern et al. 2012; Halpern, Frazier, et al. 2015; Halpern, Longo, et al. 2015), the Mission Blue Sylvia Earle Alliance, the World Ocean Council, and the Schmidt Ocean Institute are just a few of the global-scale efforts emerging in recent years to focus on ocean sustainability.
Human-caused pressures have put the ocean in a state of deep crisis. And if the ocean is in crisis, the Earth is in crisis. At this pivotal juncture in history, we need good, digestible science to underpin solutions for
• protecting the ocean and ensuring our safety,
• managing and mitigating conflict among multiple and simultaneous uses of the ocean,
• geodesigning the ocean, and
• discovering and exploring a part of our planet that remains less understood than the moon, Mars, and Venus.
Ocean Solutions, Earth Solutions is about use-inspired science and realistic solutions for the ocean and thus the earth. With chapters drawn from among the best science presented at the 2013 Esri Ocean GIS Forum (first edition) and 2014 Forum (second edition), this book seeks to put that science into the hands of government decision-makers and ocean/coastal science researchers, state and local coastal zone managers, and ocean/coastal GIS practitioners. It also seeks to preserve good scholarship, including the emerging scholarship of students working on theses and dissertations. The editor and a sizable cadre of experts conducted a standard academic peer review of all chapters. The book encourages GIS best practices. Toward this end, it features an extensive digital supplement, including datasets with accompanying digital object identifiers (DOIs, in keeping with data publication trends; e.g., The Royal Society 2012; Leadbetter et al. 2013; Parsons and Fox 2013), geoprocessing workflows, GIS tools packaged as desktop extensions or web services, mobile apps, Python scripts, and Esri Story Map apps. Digital content for this book, described under Supplemental Resources
at the end of pertinent chapters, can be accessed by QR code or URL or in the book’s online resources (esri.com/esripress-resources/ocean-solutions-2e).
The solutions developed in these chapters lend credence to the hope that we can fix the problems of the ocean before it is too late—to build what Knowlton (2014) and others have tagged as a new #OceanOptimism hashtag on Twitter. In a similar vein, the Mission Blue (2014) initiative of the Sylvia Earle Alliance has established hope spots.
Indeed, the efforts of scientists, resource managers, coastal engineers, and local communities have shown many promising results around the world. As Knowlton urges, we need to trumpet and seed these solutions.
The chapters of Ocean Solutions, Earth Solutions are divided into four main themes:
1. Server/cloud GIS
2. Coastal and marine spatial planning
3. Analytical and mapping tools
4. Visualization
During the 1980s and 1990s, GIS architectures were largely file-oriented, with Esri datasets structured as coverages and later as shapefiles. In the late 1990s, geographic datasets were placed into relational databases to allow for multiuser access. The web has evolved more recently to embrace a service model powered by on-premise servers and/or the cloud, extending the old desktop data models with layers and web maps that travel with the data.
Chapter 1 represents the critical role that server/cloud GIS now plays in fisheries oceanography, particularly as local desktop resources are less able to handle the huge streams of data from ocean observatories, the complexity of numerical models, and the massive output files that these models generate. The chapter describes the on-premise, server-based Ecosystems Fisheries Oceanography Coordinated Investigations (EcoFOCI) Data Access and Analysis Tool (EcoDAAT) and the cloud-based LarvaMap, both developed for data integration and particle modeling to support fisheries oceanography and fisheries management. And because server/cloud GIS has become so important to data discovery and collaborative geodesign as part of coastal and marine spatial planning, it segues to the next theme of coastal and marine spatial planning.
Chapter 2 focuses on best practices for decision support guided by the popular Marxan approach when used in concert with GIS to inform multiple-objective ocean conservation and spatial planning. Chapter 3 (updated in the second edition) describes coastal and marine spatial planning programs under way in Martin County, Florida, and provides specific exemplars for data consolidation and organization, a workable data model and geodatabase schema, and GIS applications for managing artificial reefs and beaches.
The next theme highlights chapters that present specific analytic and mapping tools. Chapter 4 introduces tools for use in conjunction with the Coastal and Marine Ecological Classification Standard (CMECS), the first consistent US national classification for mapping and integrating ecosystem observations. The NatureServe observation toolkit and the CMECS crosswalking tool of the National Oceanic and Atmospheric Administration (NOAA) are examples of effective solutions for standardizing the collection of data in the field and uploading, validating, aggregating, and integrating that data with existing archives. Chapter 5 provides important guidelines and a desktop solution for accessing, mapping, and interpreting the output of numerical models used to project the effects induced by changes in atmospheric greenhouse gases. It describes the methodology used to extract the relevant ocean variables and considers the consequences of decision-making based on climate change scenarios. Policy makers and resource managers alike need such insights to create and manage climate adaptation and mitigation plans. A method and an accompanying ArcGIS tool for calculating a pollutant exposure index (PEI) for the Southern California Bight are the subject of chapter 6, which provides Python scripts for geoprocessing the exposure data and producing georeferenced PEI rasters for use with other spatial data to examine relative pollution risk for any area of interest within the mapped region.
Chapter 7 emphasizes mapping tools, in this case a map viewer to facilitate the management and monitoring of the Archipelago de Cabrera National Park in the Balearic Islands of Spain. The chapter includes useful guidelines for designing an intuitive interface with the Adobe Flash Integrated Development Environment along with the ArcGIS Viewer for Flex and creating the Open Geospatial Consortium Web Map Service. Chapter 8 (updated in the second edition) describes design principles and development of both a web app and a mobile app for mapping marine mammal distributions, drawing in large part from data contributed by citizen scientists. Chapter 9 (updated in the second edition) discusses the benefits of providing multiple computer pathways within the powerful web tool of the West Coast Ocean Data Portal to better enable search and connection to the services and datasets published by members of this network. These pathways include Catalog Services for the Web, the Esri Geoportal representational state transfer (REST) application programming interface (API), the OpenSearch API, and the Apache Solr REST API.
Chapter 10 analyzes land-cover data from the East End Marine Park of Saint Croix, US Virgin Islands, applying a landscape development intensity index for adjacent watersheds along with an analysis of benthic habitat data from both benthic habitat maps and in-water surveys. The resulting maps are a powerful tool for understanding the connections between actions on land and impacts at sea, specifically the impacts of land-based sources of pollution on benthic habitats and species composition on fringing coral reefs. Chapter 11 employs a suite of both desktop and ArcGIS Online mapping and spatial analysis tools to develop a collaborative proposal to the Pacific Fishery Management Council. The proposal, which represents nearly a year’s worth of discussion and trust building among commercial fishermen, conservationists, and local government, is for the design of spatial modifications to the current Pacific Coast groundfish essential fish habitat (EFH) conservation areas. The development of a Story Map app, in particular, helped this process. Chapter 12 focuses on Story Map apps again as one of many tools that the Aquarium of the Pacific uses to engage the public in lifelong learning experiences. The chapter provides best-practice examples in employing these tools to create a rich and informal science education experience that inspires learners to take action to protect and conserve ocean resources.
Chapter 13 begins the next major theme on practical solutions for improved visualization of the ocean. It reviews principles for integrating GIS and landscape visualization software to produce immersive and interactive four-dimensional models of the ocean. Planners and scientists provide feedback in a case study that tests these approaches within a marine protected area. Chapter 14 describes advances in visualization methods based on the Worldwide Telescope virtual globe and applies these methods to the management and analysis of data from marine microbial ecology and physical transport processes in the ocean. The discussion has implications on how to more effectively analyze and understand new harvests of large, complex datasets in three- and four-dimensional software environments that combine virtual globes with web GIS. Chapter 15 presents a framework adopted by the State of Oregon for managing the visual landscape of the territorial sea. The framework is part of the state’s effort to plan for marine renewable energy development. We could have included this chapter in the coastal and marine spatial management section of the book, but we put it here because of the emphasis on a visual resource scenic-quality inventory of publicly accessible viewpoints of the territorial sea, the adoption of visual class standards, and the modeling of viewsheds and viewshed class values. As such, the use of visualization tools described in this chapter may provide an important exemplar for other coastal states that are seeking to bring public participation and community values into the spatial framework for decision-making, thereby helping communities to define and conceptualize management, sustainability, and resilience.
Chapter 16 (updated in the second edition) returns the discussion to server/cloud GIS while drawing also on the theme of visualization. The chapter presents workflows using ArcGIS 10.2 GeoEvent Extension for Server as well as a series of Python scripts to map, in near real time, the observations of wave and profile gliders monitoring the shallow water depths of the Gulf of Mexico. Researchers will apply this process to visualize future observations from these platforms, which are critical for understanding and reducing ocean acidification and oxygen-poor conditions in the Gulf.
Chapter 17 ushers in brand-new content for the second edition with an overview of seamounts: how they are officially defined as features on the ocean floor, how they are mapped and classified, and why they are an important foundation for marine habitat protection. The chapter presents a new seamount classification scheme based on multivariate analysis of differences in area, height, peak depth, proximity to nearby features, and the area of steep slopes. A focus on seamount types within the Clarion-Clipperton Zone in the eastern Pacific Ocean and Kiribati’s marine jurisdictions evaluates the usefulness of the classification for improving existing management regimes and setting future management priorities.
Chapter 18 provides an overview of the benefits of using satellite image analysis and web services for a range of coastal applications, from mapping the shallow water bathymetry of the Great Barrier Reef, to exploring the seafloor habitats of Abu Dhabi, to recording various inland water quality sites, to monitoring dredging sites. Chapter 19 introduces the reader to The Nature Conservancy’s Coastal Resilience suite of web-based tools for examining coastal hazards as well as social, ecological, and economic assets. It also includes one of the more extensive sets of supplemental materials, featuring interactive learning modules, sample datasets, a full GitHub repository of computer code, and digital stories
in video mode.
Chapter 20 continues the thread of coastal resilience by describing collaborations among county regional planners and scientific and policy experts to shape climate adaptation planning at various scales throughout coastal California. The team conducted a spatially explicit coastal vulnerability assessment using the Natural Capital Project’s InVEST (Integrated Valuation of Ecosystem Services and Trade-offs) tool and ArcGIS with an eye toward evaluating the role of natural habitats in reducing exposure to rising sea levels and increasingly intense storms.
The author team for chapter 21 included several undergraduate students at the University of Redlands and chronicles their geospatial adventures while participating in a travel course to French Polynesia to study various applications of environmental science on the atoll known as Tetiaroa. This adventure was more than just a field trip. The students and faculty studied one of the last ground-nesting bird populations on Earth as well as ongoing changes in shoreline position to evaluate overall island geomorphology, all using kite aerial photography and GIS.
Chapter 22 returns to the theme of seafloor classification, but by way of bathymetric features that are probable modern-day paths for sediment drainage (i.e., canyons, valleys, channels, funnels, chutes, and hollows carved out by downslope sediment gravity movement, wave or current traction, and liquefaction). Identifying these pathways is important for understanding the supply of nutrients to benthic habitats and related controls on benthic community structure and biomass.
Finally, chapter 23 introduces the Navy Shore Geospatial Energy Module, a platform to help the US Navy become more resilient to climate change through reduced energy consumption, the integration of alternative energy sources, and ultimately the enhancement of energy security through sustainable energy usage. This platform, unique among the US Armed Forces, enables navy installation managers to track and map energy usage while geodesigning energy conservation projects and facilities management scenarios.
May this book not only inspire #OceanOptimism but drive real action with GIS. Such action needs a continued infusion of ideas and best practices. Hence, we envision Ocean Solutions, Earth Solutions as evolving into a series of research monographs, always based on Esri Ocean GIS Forum events as they continue to evolve.
References
Halpern, B. S., C. Longo, D. Hardy, K. L. McLeod, J. F. Samhouri, S. K. Katona, K. Kleisner, S. E. Lester, J. O’Leary, M. Ranelletti, A. A. Rosenberg, C. Scarborough, E. R. Selig, B. D. Best, D. R. Brumbaugh, F. S. Chapin, L. B. Crowder, K. L. Daly, S. C. Doney, C. Elfes, M. J. Fogarty, S. D. Gaines, K. I. Jacobsen, L. B. Karrer, H. M. Leslie, E. Neeley, D. Pauly, S. Polasky, B. Ris, K. St. Martin, G. S. Stone, U. R. Sumaila, and D. Zeller. 2012. An Index to Assess the Health and Benefits of the Global Ocean.
Nature 488:615–20.
Halpern, B. S., C. Longo, J. S. S. Lowndes, B. D. Best, M. Frazier, S. K. Katona, K. M. Kleisner, A. A. Rosenberg, C. Scarborough, and E. R. Selig. 2015. Patterns and Emerging Trends in Global Ocean Health.
PLOS ONE 10:e0117863.
Halpern, B. S., M. Frazier, J. Potapenko, K. S. Casey, K. Koenig, C. Longo, J. S. Lowndes, R. C. Rockwood, E. R. Selig, K. A. Selkoe, and S. Walbridge. 2015. Spatial and Temporal Changes in Cumulative Human Impacts on the World’s Ocean.
Nature Communications. doi:10.1038/ncomms8615.
Knowlton, N. 2014. Why Do We Have Trouble Talking about Success in Ocean Conservation?
http://po.st/IyKMYe. Last accessed October 18, 2015.
Leadbetter, A., L. Raymond, C. Chandler, L. Pikula, P. Pissierssens, and E. R. Urban. 2013. Ocean Data Publication Cookbook. IOC Manuals and Guides No. 64. Oostende, Belgium: UNESCO IOC International Oceanographic Data and Information Exchange. http://bit.ly/1lcHa6I.
Liu, J., H. Mooney, V. Hull, S. J. Davis, J. Gaskell, T. Hertel, J. Lubchenco, K. C. Seto, P. Gleick, C. Kremen, and S. Li. 2015. Systems Integration for Global Sustainability.
Science. doi:10.1126/science.1258832.
Lubchenco, J., and K. Grorud-Colvert. 2015. Making Waves: The Science and Politics of Ocean Protection.
Science. doi:10.1126/science.aad5443.
Melillo, J. M., T. C. Richmond, and G. W. Yohe, eds. 2014. Climate Change Impacts in the United States: The Third National Climate Assessment. Washington, DC: US Global Change Research Program. doi:10.7930/J0Z31WJ2.
Mission Blue. 2014. Hope Spots. http://mission-blue.org/hope-spots. Last accessed October 18, 2015.
National Academy of Sciences National Research Council. 2011. An Ocean Infrastructure Strategy for US Ocean Research in 2030. Washington, DC: National Academies Press.
Parsons, M. A., and P. A. Fox. 2013. Is Data Publication the Right Metaphor?
Data Science Journal 12:WDS32-WDS46.
Pratt, M. 2013. Saving the Blue Stuff: The First Ocean GIS Forum.
ArcUser 17 (1): 58–63.
The Royal Society. 2012. Science as an Open Enterprise: Open Data for Open Science. The Royal Society Science Policy Centre Report 02/12. London: The Royal Society. https://royalsociety.org/policy/projects/science-public-enterprise.
Smith, W. H. F., and K. M. Marks. 2014. Seafloor in the Malaysia Airlines Flight MH370 Search Area.
Eos, Transactions of the American Geophysical Union 95 (21): 173–74.
UN Intergovernmental Panel on Climate Change. 2013. Summary for Policy Makers.
In Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, edited by T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley. Cambridge: Cambridge University Press.
Wright, D. 2013. How Is GIS Meeting the Needs of Ocean (and Other) Sciences? Plus, Minus, Interesting.
http://blogs.esri.com/esri/arcgis/?p=28054. Last accessed October 18, 2015.
Wright, D. J. 2012. The Ocean GIS Initiative: Esri’s Commitment to Understanding Our Oceans.
Esri White Paper/e-Book J10129. Redlands, CA: Esri.
XPRIZE Foundation. 2015. Wendy Schmidt Ocean Health XPRIZE.
http://oceanhealth.xprize.org. Last accessed October 18, 2015.
Dawn J. Wright, ed.; 2016; Ocean Solutions, Earth Solutions, second edition; http://dx.doi.org/10.17128/9781589484603_int
CHAPTER 1
Cloudy with a Chance of Fish: ArcGIS for Server and Cloud-Based Fisheries Oceanography Applications
Tiffany C. Vance, Stephen Sontag, and Kyle Wilcox
Abstract
Wright et al. (2013) defined cyber-GIS as GIS detached from the desktop and deployed on the web.
This chapter describes two cyber-GIS applications—the Ecosystems Fisheries Oceanography Coordinated Investigations Data Access and Analysis Tool and LarvaMap—developed for data integration and particle modeling in support of fisheries oceanography research and fisheries management. Both were developed to make it easier for researchers to locate, link, analyze, and display in situ and model data. Because geospatial/temporal relationships are critical in understanding the early life history of commercially valuable fish species such as walleye pollock (Gadus chalcogrammus), integrating the generation of model results and in situ environmental data using GIS provides critically needed tools.
The Ecosystems Fisheries Oceanography Coordinated Investigations Data Access and Analysis Tool uses an ArcGIS for Server front end for selection, display, and analysis of data stored in an Oracle database. LarvaMap is a cloud-based tool for running particle dispersion models for fish and invertebrate larvae. It supports scenario testing and makes it easy for nonmodelers to configure and run models. The front end is a web-based map interface, and running the tracking model in a cloud resource allows scaling of computer resources to meet computational needs. Both applications provide basic display and analysis of data and can be integrated with desktop GIS for advanced analyses and two-, three-, and four-dimensional visualizations. The results from these projects can be translated to other types of particle tracking, other data management needs for environmental studies, and used as case studies of deployment of GIS tools to cyber-GIS computing resources.
Problem and Challenge
Knowledge of the complex relationships between the targeted fish species and its food sources, predators, and environment will aid the new emphasis on ecosystem-based approaches to fisheries management (Link 2002). Understanding these relationships can guide the management of a fishery and the setting of quotas, timing of fishing periods, and management of related fisheries or species that depend on the target species. Investigating these relationships requires a combination of field research and computer modeling. Field research gathers large amounts of disparate data that needs to be managed, integrated, analyzed, and visualized. Many times, the only common characteristic of this data is the fact that it can be geolocated. Spatial analyses and geospatial measures are critical in understanding this data. Models of ocean circulation, fish life histories, or complicated ecological interactions can require huge amounts of computing resources, generate massive output files, and be best understood using advanced visualization and analysis tools.
Providing the computational and analytical environment to support these studies is challenging. Local desktop resources are frequently insufficient. Tightly focused databases can make integrating information almost impossible, and visualization tools may be hard to use. New developments in cyber-GIS, specifically the expansion of ArcGIS for Server and increased access to cloud computing resources, have enabled us to develop and test a number of applications to address solutions for integrating disparate data for easy retrieval and analysis using ArcGIS for Server, either locally or in the cloud, and explore the possibilities of running models in the cloud and storing and displaying their results in a geospatially enabled application.
Introduction and Background
Parallel technological advances in computing and marine fisheries research have contributed greatly to truly implementing ecosystem approaches to fisheries management. Technologies first developed during World War II have been used both to improve sampling in marine systems and support modeling and geospatial analyses. Military surplus vessels formed the core of the oceanographic and fisheries research fleet in the 1950s and onward (Rainger 2000; Rozwadowski and Van Keuren 2004; Hamblin 2005), sound navigation and ranging (SONAR) techniques were adapted to become the core of fisheries hydro acoustics (Fornshell and Tesei 2013), and biological sampling became increasingly electronic (Dickey 2003; Wiebe and Benfield 2003). Computing resources evolved from large mainframe computers to minicomputers and personal computers, to computers on a chip embedded in deployed instruments, and finally back to large centralized, elastic computing resources in the cloud. Geospatial analysis has gone from the early days of GIS in SyMAP and the Canada Geographic Information System (Coppock and Rhind 1991) to ArcInfo to open-source GIS, cyber-GIS, and ArcGIS for Server providing GIS in the cloud.
GIS and other spatial analysis tools provide important tools to understand natural systems. The advent of cyber-GIS has made these tools even more powerful by removing constraints on the size and complexity of analyses through elastic computing resources and cloud-based storage for terabytes of data. Of course, what has been called the grand challenges
remain, of data intensity, computing intensity, concurrent intensity, and spatiotemporal intensity (Yang et al. 2011). The data integration and complex system modeling tools we have created with the Ecosystems Fisheries Oceanography Coordinated Investigations (EcoFOCI) Data Access and Analysis Tool (EcoDAAT) and LarvaMap take advantage of early answers to these challenges.
Wright et al. (2013) have written of cyber-GIS as GIS removed from the desktop to centralized, usually cloud-based resources. Wang et al. (2013) describe cyber-GIS as cyberinfrastructure-based GIS
and a subset of the broader realm of spatial cyberinfrastructure
as defined in Wright and Wang (2011). Yang et al. (2013) have identified a number of disciplines in which cyber-GIS can support critical advances. Topics related to fisheries include climate science, ecology, environmental health, and disaster management. They call for advancements in system architectures, which can directly support modeling and visualization and are key to understanding the outputs of complicated ecosystem models; data storage, which will support both the storage of field data and outputs of models; and data and process colocation, which may be less critical because of developments such as the grid caching we developed for LarvaMap described later in this chapter (Yang et al. 2013).
Nyerges (2010) specifically addresses the application of cyber-GIS to marine systems by looking at uses of cyber-GIS for modeling coastal marine policy and governance. Although his focus is on human communities, many of the elements he cites are equally applicable to marine ecosystems that also have a large number of interacting components,
creating systems that are computationally complex and intense
with spatiotemporal elements that require tremendous amounts of data and computational support.
EcoDAAT uses the cyber-GIS concepts of moving processing and data storage from the desktop to a centralized resource with a web-based thin client providing the user interface. The results can be used either for analyses via ArcGIS for Server or for local GIS or statistical analyses using ArcGIS or tools such as R. LarvaMap takes advantage of cloud-based computing resources by running its models on a cloud computing resource. The ability to launch additional model runs in the cloud provides easily scalable computer resources, and data can be stored either in the cloud or in local Open-Source Project for a Network Data Access Protocol (OPeNDAP)/Thematic Real-Time Environmental Distributed Data Services (THREDDS) servers. Both of these tools help meet the ecosystem-based management needs of integrating field data with models, integrating physical and biological data, and providing information on the early life history of commercially important species.
History and Status of the Walleye Pollock Fishery in Alaska
EcoDAAT and LarvaMap are first being applied to studies of the early life of walleye pollock (Gadus chalcogrammus), hereafter simply pollock (figure 1.1). Pollock are a groundfish found primarily in the North Pacific and Bering Sea that supports the largest commercial fishery in the United States by weight (Kendall et al. 1996; AFSC 2013). The primary human uses of pollock are in making fish sticks and fish sandwiches, and as surimi to make imitation crab. Juvenile pollock are a critical food source for flatfish (Lang et al. 2000), and adult seabirds serve as food for Pacific cod, Pacific halibut, sea lions, and other marine mammals.
Figure 1.1. Walleye pollock larvae, adults, the products made from pollock, and major predators of pollock. By Ingrid Spies and other photographers from the Alaska Fisheries Science Center, NOAA Fisheries Service; additional photograph by Nazila Merati.
In the Gulf of Alaska, pollock start as eggs spawned in March and April. The eggs are spawned at a depth of 100–250 m and hatch in two to three weeks. Larvae occur at an average depth of 40 m and develop over the next few months. At this stage, they are still at the mercy of oceanic currents and transported as though they were inanimate particles. Once the larvae are about 60 days old, they develop rudimentary swimming skills. Though not able to swim long distances, they are able to swim in bursts to consume prey and avoid potential predators. After 100 days, they are competent swimmers and can move up and down through the water column. At the age of three or four years, they are large enough to be commercially harvested and reach sexual maturity at age four or five. They can live for more than 20 years, but fish that old are unusual (AFSC 2010).
The Gulf of Alaska and Bering Sea pollock fishery started with a small trawl fishery in the 1950s, and catches increased rapidly in the 1970s and 1980s as the foreign fishing fleet targeted the population. The discovery of a spawning population of the fish in Shelikof Strait in 1981 led to the development of a roe fishery. Intense fishing in the international waters of the Donut Hole area of the western Bering Sea in the 1980s led to the collapse of that pollock population (Bailey 2013). By 1988, changes in fisheries regulations excluded the foreign fleet from American waters and transformed the pollock fishery into a domestic US fishery with the primary consumer products being frozen filets and surimi. In the 1990s, the rise of ecosystem-based management led to consideration of related species in managing the fishery; and in 1999, the Aleutian Islands portion of the fishery was closed in response to declines in Stellar sea lion populations, with the thought that fishing was overcompeting with sea lions for fish (National Research Council 2003). This fishery was reopened in 2005 (AFSC 2010). A complete history of the fishery, both domestic and international, can be found in Bailey (2013).
Management of the fishery depends on, among other things, an understanding of the early life history of pollock. Success in the early stages of life is one factor determining how many fish are eventually available to be harvested while preserving a sustainable population. The EcoFOCI program at the National Oceanic and Atmospheric Administration (NOAA) studies the factors that affect the variability in recruitment of pollock and other commercially harvested species in the Gulf of Alaska, North Pacific, and Bering Sea. Oceanographic and biological factors are studied using a combination of field and model studies. Field research includes physical oceanographic surveys and biological sampling to look at life stages from eggs to juveniles and post-age-1 fish. A particular emphasis of the sampling is studying pollock early life history. The zooplankton that supports the fish populations is also sampled and described. The samples collected also support age and growth, genetic, and food habit studies. The field data supports an understanding of the biological and physical factors affecting recruitment. The data also provides initial conditions and validation for model studies.
Understanding the Early Life History of Pollock Using Individual-Based Models
One of the grand challenges
in fisheries research is understanding the fluctuations in the size and structure of fish stocks. The classic work by Hjort (1914) described two hypotheses to explain the variability: The first, the Critical Period hypothesis, posits that the size and strength of a year class of fish is determined very early in life, when the fish have exhausted their yolk sack and started feeding, and is a function of the availability and quality of the plankton they are feeding on. The second, the Aberrant Drift hypothesis, examines the role of transport in taking fish larvae so far from place of birth that they are unable to return to support future populations. A variety of hypotheses has developed in the century since Hjort’s work, including the Match-Mismatch hypothesis of Cushing (1974), which relates the timing of larvae hatching and bloom of plankton to feed them; the effect of vertical stratification in the Stable Ocean hypothesis of Lasker (1978); and a number of other hypotheses described and related in Houde (2008) and Hare (2014).
In studies of marine fish recruitment, early studies to test hypotheses were primarily descriptive and considered aspects such as the scale of distributions of plankton (Haury et al. 1978). More recently, advances in the collection of quasi-synoptic in situ and satellite data, the rise of ecosystem approaches to studying marine systems (Levin 2006; Vance and Doel 2010), and rapid and striking advances in computers and data storage (Gentleman 2002; Yang et al. 2011) have radically changed research on recruitment. Hydrodynamic models are now able to resolve details at the scales that affect recruitment (tens of kilometers), and they can be run for weeks to months of model time to cover the early life history of slowly developing species. Biological and physical processes can now be coupled and modeled, and advances in visualization tools have made visualizing and interpreting large model outputs easier (Megrey et al. 2002; Vance et al. 2006; Vance 2008).
One requirement that these various hypotheses share is a need to know where fish larvae are located at various stages of life. This need has led to the development of techniques to model the dispersion of larvae based on hydrodynamic models of the ocean. At their simplest, these models distribute larvae as though they were inanimate particles drifting in ocean currents. More detailed models include larval behaviors such as the ability to swim, motion toward or away from light, and avoidance of warm or cool water (Hinckley et al. 1996; Hermann et al. 2001; Miller 2007; Hinckley et al. 2009). These models are the simplest of what are called individual-based models (IBM), which are called agent-based models in other disciplines. An example is the DisMELS model (Dispersal Model for Early Life History Stages), which incorporates behavior in early life stages to study the dispersion of eggs and larvae (Stockhausen 2006). More complex models describe both larval distributions and the distributions of prey for the larvae. These tools, often called nutrient-phytoplankton-zooplankton (NPZ) models, describe spatial and temporal interactions (Franks 2002). Upper-trophic level models such as FEAST (Forage and Euphausiid Abundance in Space and Time) add predator and bioenergetic parameters (Aydin et al. 2010; Ortiz 2014).
Integrating and Analyzing Field Data using EcoDAAT
Web-based GIS applications for data discovery and marine planning encompass a wide range of tools. SeaSketch (http://mcclintock.msi.ucsb.edu/projects/seasketch) supports collaborative design but does not support data storage or analysis. SIMoN (http://sanctuarysimon.org/regional_sections/maps/mpaviewer/) and similar tools provide mapping of the location of data collection sites and links to project websites but do not allow for analysis of the data. GIS-based data discovery/download-only tools, such as many of those listed for ecosystem-based management (EBM) on the EBM Tools Network (http://www.ebmtools.org/?q=ebm_tools_taxonomy#), acquire and manage data but do not focus on geoanalysis. The strength of EcoDAAT is that it allows location and downloading of raw data as well as simple geoanalysis.
Because the field sampling for EcoFOCI work on pollock and other forage species is multidisciplinary, the datasets gathered are in disparate formats. The datasets for ichthyoplankton, zooplankton, and environmental measurements were originally stored in separate Microsoft Access databases. This made it hard to answer questions such as What are the water temperatures where we find larger than average pollock larvae?
in a timely manner. Although the datasets were not enormous—the largest had about 250,000 records—network response time made it impractical to store the data in a single shared database on a network drive. Rather, it was necessary to have the complete databases on the desktop system of each user. This maintenance challenge caused databases to get out of sync if updates failed or were delayed. The physical oceanographic datasets of EcoFOCI collaborators contain up to six million conductivity-temperature-depth (CTD) records, making it impossible to deploy them as part of the existing databases.
The project to create EcoDAAT had a number of goals: to integrate the databases so that cross-data-type queries were possible, move the integrated database to a larger and more robust database application, and make the data selection and display more geospatial by providing a map-based interface and ArcGIS for Server-based analytical tools. To meet these goals, an ArcGIS for Server/Flex front end connected to an Oracle database was chosen. The system layout is shown in figure 1.2. The datasets are stored in a single Oracle database designed to contain, and link, the various types of data being collected. The data cascade is started at sea, where a stand-alone Java/SQLite application called the Cruise Operations Database (MasterCOD) is used to input metadata and collection information about every operation. The information includes the latitude and longitude of the operation, time and date, type(s) of equipment deployed, types of sample(s) gathered, and any notes about the operations. These records are used to track samples as they are analyzed and also to geolocate and cross-reference operations and samples. These datasets are used to populate the Haul tables of the Oracle database with core information and also to populate tables in two associated applications, ZooPPSI and IchPPSI. These two Java/SQLite applications are used by researchers at the Polish Plankton Sorting and Identification Center to record identifications and measurements of zooplankton and ichthyoplankton in samples sent to them for processing. Data from completed ZooPPSI and IchPPSI tables is loaded into the EcoDAAT Oracle tables using a series of Standard Query Language (SQL) scripts. These datasets are used to populate sample and specimen tables and associated tables to hold species and diet information. Information on larger age-0 and older fish collected during trawling operations is stored in Microsoft Excel spreadsheets at sea. The datasets are loaded into the Oracle database in a similar process to the loading of ichthyoplankton data. Laboratory analyses of stomach contents and other diet and condition-related information are loaded from text files and spreadsheets. Physical oceanographic datasets, including CTD and SeaCAT measurements, are processed using Sea-Bird Electronics files, reformatted into comma-separated value (CSV) files, and loaded using SQL scripts. Both CTD and bottle datasets are stored and can be linked to sample data.
The EcoDAAT front end uses the Flex viewer available with ArcGIS for Server (figure 1.2). This was chosen for ease of implementation and enhanced functionality. The front end allows for data selection by type of data (e.g., ichthyoplankton or oceanographic data), geographic location using graphic- or text-based searches, time of collection, and combined criteria. The interfaces for data selection are dynamic, and the available selection criteria vary based on the type of selection being made. Because some of the data in EcoDAAT is still undergoing quality assurance/quality control (QA/QC), the front end uses Lightweight Directory Access Protocol (LDAP) authentication for secure access. In the future, a tiered access scheme—in which data can be made fully public, shared with collaborators only, or limited to program scientists—will be implemented.
Figure 1.2. EcoDAAT system architecture showing integration of Flex, Oracle, and ArcGIS for Server. Esri.
The results of a search are available both in summary form on the map portion of the front end (with data locations highlighted and the standard GIS Identify tool implemented to query data) and as a table that can be output as a shapefile or .csv file for further analysis (figure 1.3). The results of a search for haul information can also be linked to searches for associated data to create complex queries.
Figure 1.3. Example of the map-based selection tool in the Ecosystems Data Access and Analysis tool (EcoDAAT) and a table of data. By NOAA Alaska Fisheries Science Center; data sources: NOAA Alaska Fisheries Science Center, Esri, DeLorme, GEBCO, NOAA NGDC, National Geographic Society, HERE, Geonames.org, and other contributors.
Simple geoanalyses such as creating and plotting a bathymetric cross section are available (figure 1.4). These types of tools will be expanded in the future by taking advantage of the geoprocessing tools available in ArcGIS for Server.
Figure 1.4. Example of the cross section tool in EcoDAAT. By NOAA Alaska Fisheries Science Center; data sources: NOAA Alaska Fisheries Science Center, Esri, DeLorme, GEBCO, NOAA NGDC, National Geographic Society, HERE, Geonames org, and other contributors.
Tracking the Transport and Dispersion of Larvae Using LarvaMap
Particle-tracking models enable the study of the motion of particles in the atmosphere or the ocean, from oil globules from an oil spill to plumes of smoke or sediment to even the ash from a volcanic eruption. Many models run locally and require setting up complicated parameter files for a model run. Examples include the Ichthyop tool for Lagrangian modeling of the Bay of Biscay and Mediterranean Sea (Lett et al. 2008) and the Lagrangian Transport Model (LTRANS), which was originally created to model the dispersion of oyster larvae in Chesapeake Bay (http://northweb.hpl.umces.edu/LTRANS.htm). Model runs can take hours to days to complete, depending on the number of particles released, duration of the run, and power of the underlying computer(s). There are an increasing number of web-based particle-tracking models, including the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Volcanic Ash Model (Draxler and Rolph 2003), which allows users to model ash transport from a list of volcanoes using one of five common atmospheric models. Connie2 (http://www.csiro.au/connie2/) allows the user to set up and run particle-tracking models for the seas around Australia and the Mediterranean Sea. Connie2 allows for either passive/inanimate particles or particles that show limited behaviors such as the ability to swim.
Particle tracking is a critical element of modeling the early life history of pollock and other commercially important species. The models can be as simple as one that tracks inanimate particles and as complex as one that describes the interactions at many trophic levels in an ecosystem. Lagrangian models follow an individual through space and time by calculating a path line and tracking the motion of the particle. Eulerian models use positions on a fixed spatial grid and streamlines to describe transport through the location. LarvaMap is a cloud-hosted implementation of a Lagrangian particle-tracking model that supports a variety of behaviors in a target species. It is an individual-based model capable of using output from a variety of ocean circulation models to drive the dispersion and movement of particles. The front end is a series of web pages for setting up larval behaviors, choosing the circulation or hydrodynamic model, configuring the model run, and launching the run. The model itself is run on a cloud computing resource (currently the Amazon cloud), but it could also be configured to run on a local resource or private cloud. The results are visualized as a track line in the web client, and results are available as an Esri shapefile and network Common Data Format (netCDF) file for further analysis.
LarvaMap is a combination of four pieces of technology working together (figure 1.5), consisting of
• A behavior library to create, catalog, and share larval behaviors: http://behavior.larvamap.axiomdatascience.com
• A fate and transport model written as a Python library: http://bit.ly/16dFk9x
• A model service implementing the fate and transport model in a cloud architecture
• A web client for interacting with the Model Service through a representational state transfer application programming interface (REST API): http://bit.ly/14aZinD
Figure 1.5. LarvaMap architecture that shows the interconnections between the various elements and the process of creating a model run. RPS/ASA.
Setting Larval Behaviors using the Behavior Library
The first step in running LarvaMap is discovering or creating a set of larval behaviors for an organism or species of interest using the Larval Behavior Library (figure 1.6). The library allows the creation of new behavior files, searching for existing files, and sharing files. Allowing behaviors to be exchanged between researchers, modelers, and students supports the creation of behaviors as a social process, with the ability to see how others have defined behaviors, to run models using both your own and community behavior files, and to work collaboratively on building behaviors. With the ease of running the models, it is possible to quickly and easily compare the results of different behaviors.
Figure 1.6. The interface to the Larval Behavior Library is used to define the behaviors of a specific species. By NOAA Alaska Fisheries Science Center; data sources: NOAA Alaska Fisheries Science Center, Esri, DeLorme, GEBCO, NOAA NGDC, National Geographic Society, HERE, Geonames.org, and other contributors.
Behaviors that can be defined include
• The geographic range of the organism and any keywords to be used for regional searches
• Adding life stages (e.g., egg, larval, juvenile) with associated duration in days or by temperature range
• Adding behaviors to life stages, including capabilities such as swimming speed and vertical migration, reactions such as avoidance behaviors, and sensory reactions such as avoiding salinities or temperatures
• The timing of the transition out of the plankton stage and where the organism will end up—settling on the bottom for benthic species and where in the water column pelagic species will reside
Configuring the Model Run Using the Web Client: Setting Parameters and Choosing Circulation Models
Configuring the model requires setting the location, date, time, and depth of the release and specifying the number of particles to be released. The behavior model is specified, usually by species. Vertical and horizontal dispersion coefficients can be set, and the underlying hydrodynamic model is specified (figure 1.7). Finally, the time step for the model runs is specified. Model durations of months and years can be supported, but the time step for the model must be selected to be reasonably small to provide the level of detail needed without requiring the calculation of an unnecessary number of time steps.
Figure 1.7. The interface for configuring and running a LarvaMap model. By NOAA Alaska Fisheries Science Center; data sources: NOAA Alaska Fisheries Science Center, Esri, DeLorme, GEBCO, NOAA NGDC, National Geographic Society, HERE, Geonames.org, and other contributors.
Hydrodynamic models include those provided via THREDDS/OPeNDAP server or available as a local file. OPeNDAP and file-based datasets can be added to the LarvaMap system as they are discovered. An administrator of the LarvaMap system can add a dataset through an interface, and datasets must be registered in LarvaMap before they can be used for particle forcing. Both local and OPeNDAP datasets can be registered in LarvaMap. A Python library called paegan (https://github.com/kwilcox/paegan) was developed as part of the LarvaMap project to facilitate consistent access to many formats of hydrodynamic model data. If the dataset opens in paegan as a CommonDataset object, it is usable in LarvaMap. Currently, paegan supports regular grids (delta y == delta x), curvilinear grids (sometimes called i/j grids), and static triangular meshes. Dynamically changing grids and nontriangular meshes are not yet supported. LarvaMap does not support uploading of any ocean circulation data files for internal storage. File-based datasets must be made accessible to all model runs in the LarvaMap architecture before being used. Installations of LarvaMap that use the distributed file system GlusterFS can achieve this.
The Regional Ocean Modeling System (ROMS) and US Navy Coastal Ocean Model (NCOM) models have been used to look at transport of pollock and herring in the North Pacific. Lowry (2013) used LarvaMap linked to an ROMS model for Prince William Sound to model drift and recruitment of herring. The NCOM region 7 model has been used to look at transport of sea urchin larvae off the California coast. The NCOM AmSeas model has been implemented for use in the Gulf of Mexico.
The Heart of LarvaMap: Fate and Transport Models
During a model run, the transport process is broken into smaller fragments called models. Each model implements a move method that moves a particle based on environmental and behavioral properties. Each model has access to environmental conditions from the hydrodynamic model (u, v, w; temperature; and salinity) at the current location and time of the particle. In a few specific scenarios, environmental data for a particle’s location and time may be unknown: particle has left the domain of the hydrodynamic dataset, or there is a data gap in the hydrodynamic dataset. Under these conditions, the last known environmental conditions that the particle experienced are used in the model.
LarvaMap implements three main models. The models are run sequentially, and the results are checked against the shoreline and bathymetry (figure 1.8). A typical time step in a model run is as follows:
• Run transport (moves particle).
• Check shoreline.
• Check bathymetry.
• Run dispersion (moves particle).
• Check shoreline.
• Check bathymetry.
• Run behavior (moves particle).
• Check shoreline.
• Check bathymetry.
Figure 1.8. Elements of the transport process during a single time step in a model run. Variables u, v, and w are the circulation model velocity components. RPS/ASA.
Transport
The transport model for LarvaMap is a parallelized Lagrangian transport model that uses hydrodynamic circulation data stored as netCDF/OPeNDAP files. The model is written in Python and uses the paegan library, which provides a data model for ocean and meteorological data stored in netCDF files and the paegan-transport libraries for the model itself. It moves a particle based on the current velocity components (u [zonal or east–west], v [meridional or north–south], and w [vertical]) from the hydrodynamic model. The vertical component (w) is optional. If the transport model receives no u, v, or w data for the place and time of the particle, the last known conditions are used.
Dispersion
Each model run is configurable using a horizontal and vertical dispersion coefficient (in m/s). These coefficients are used to compute the u, v, and w dispersion.
Behavior
The behavior model is set up using a configuration from the Larval Behavior Library. A configuration consists of many life stages, each with its own behaviors. The behavior model is responsible for calling the correct life stage model based on such factors as a particle’s development, age, and mass, and adding a dead
life stage to every behavior configuration. This model does not actually move the particle—the life stage model does. It can be thought of as a container model.
Life Stage
The life stage model grows each particle based on time or as a function of temperature. It also calls the correct behavioral models associated with the particle’s life stage. For example, if there are four diel behaviors configured for a single life stage, this life stage model determines which diel behavior to call based on the particle’s time. The behavioral models described as follows work a bit differently from the transport and life stage models. They do not return a new location for the particle; they only return the u, v, and w vectors. The vectors are summed from all behaviors by this life stage model; and the particle is only moved once, after all behaviors are completed. If the life stage model receives no temperature or salinity data for the place and time of the particle, the last known conditions are used.
Life Stage: Diel
The diel model moves a particle up and down in the water column based on the sun’s position in the sky. Each life stage of a larva has a configured static vertical swimming speed that is used here to determine how fast the particle can move to its desired depth range. If the particle’s vertical swimming speed puts the particle beyond the desired depth range, the swimming speed is recomputed for a single model time step to land the particle in the middle of the desired depth range. If the desired depth range will not be reached, or the particle will end up inside the depth range, the vertical swimming speed configured in the behavior is used. A particle below desired depth will swim up, a particle below desired depth will swim down, and a particle at the desired depth will do nothing.
Life Stage: Settlement
The settlement model decides whether a particle should settle under its current conditions. As described in the Larval Behavior Library section, there are two types of settlement:
• Benthic: settles particle at 1 m above the bathymetry if the bathymetry is within the configured settlement depth range
• Pelagic: settles particle at its current depth if it is within the configured depth range
The settlement model does not return u, v, or w vectors. It directly settles or moves the particle to the desired settlement location.
Life Stage: Dead
The dead model’s purpose is to kill a particle that has not settled and has aged beyond its settlement life stage. It continues to track environmental conditions around the dead particle, and the particle will continue to be forced by the other models (transport/dispersion), but the dead model always returns 0 m/s u, v, and w.
Under the Hood: Circulation, Shoreline, and Bathymetric Data for Model Runs
Hydrodynamic Caching Strategy
One of the challenges in running a particle tracking model is that the particles are sparsely distributed. The transport model supports long-running models on the order of months and years, with no upper time limit imposed by the trajectory model. Although the circulation model data must cover a large area to support any possible particle locations, at any given time step the particles occupy a small subset of this total area. Downloading the entire circulation/hydrodynamic model data for each time step is prohibitive. But because LarvaMap allows for remote data access over the data access protocol (DAP), the entire hydrodynamic dataset for the area of interest never has to be downloaded and stored locally. The data can be accessed on the fly, and the model downloads only the data it needs to run the model (figure 1.9). Data that is downloaded by the model is reused from particle to particle, so the same data is accessed and downloaded only once. Particles use the locally cached data in forcing algorithms.
Figure 1.9. The data caching scheme used to reduce the size of downloads from the underlying circulation model. RPS/ASA.
Shoreline and Bathymetry
The fate and transport model uses land polygon objects to test for shoreline collisions. The land polygons can be stored in an Esri shapefile or accessed from an Open Geospatial Consortium (OGC) web feature service (WFS) server. Every particle loads the shoreline upon creation and indexes the shoreline around its current location. Every time a particle tests its location against the shoreline file, it first makes sure that it has the appropriate data from the source shoreline dataset to do its calculation (figure 1.10). If it does not, it accesses the raw shoreline data, whether it is a shapefile or WFS server, and pulls a chunk of the data around its current location. This acts as a caching mechanism so that the particles are not reading from the shoreline data at every detection check. A similar process is used with bathymetric data to constrain particles to the water column.
Figure 1.10. Shoreline indexing scheme used to locate the appropriate shoreline segments in the vicinity of a particle. Avoidance behavior when a particle nears a shoreline is shown in the last two parts of the figure. RPS/ASA.
The Model Service: Implementing LarvaMap in the Cloud
The model service queues model run requests received from the web client and distributes them within a network of worker
servers. Worker servers can be running in many different places, including the Amazon cloud and local data centers. Each central processing unit (CPU) core on a worker server can force a single particle for the entire duration of its model. If 100 worker cores are available, 100 particles can be run in parallel. This can be a single model run with 100 particles, or 10 model runs with 10 particles each. The model service will automatically create new worker servers when the load demands it. Exactly when to spawn a new worker server is determined by how many particles are in the queue waiting to be run. If the number of particles in the queue is three times more than the number of CPU cores available on all worker servers, another worker server is initialized. Worker servers are shut down as the number of particles in the queue decreases. When no particles are in the queue, there is a minimum of one worker server running. Cloud computing technology allows LarvaMap to scale the number of worker servers dynamically as they are needed. In the near future, the model service will also be able to distribute a single particle’s model run onto multiple servers, which can reduce model run times to one-tenth of the current run time.
The model service collects the results from all the worker servers and pieces them together into the final results for each model run. It creates the final output formats and uploads the result files to a common location for the web client to access. Utilities are available to monitor the status of the model service currently running jobs and details about any failed jobs.
Outputs from Model Runs
Outputs from LarvaMap are available in a number of formats. A GeoJSON track line is created for display in the web client. A shapefile, a netCDF file, and a Hierarchical Data Format 5 (HDF5) file are produced for display and analysis in ArcGIS and other tools (figure 1.11). The output files contain the date, latitude, longitude, and depth of the particle; the temperature and salinity values at the point if they are available from the hydrodynamic model; and the u, v, and w speed components at the grid point. A log file is also produced for troubleshooting and system analysis. TopoJSON is being considered in the future to allow a web client to show the track lines for each individual particle