Lead PI: Young-Oh Kwon (Woods Hole Oceanographic Institution)

Co-PI: Ke Chen (Woods Hole Oceanographic Institution), Glen Gawarkiewicz (Woods Hole Oceanographic Institution), Terry Joyce (Woods Hole Oceanographic Institution), Janet Nye (Stony Brook University), Jon Hare (NOAA/NMFS), Paula Fratantoni (NOAA/NMFS), Vincent Saba (NOAA/NMFS), Tim Miller (NOAA/NMFS)

The Northeast U.S. Shelf (NES) Large Marine Ecosystem (LME) supports some of the most commercially valuable fisheries in the world and has experienced dramatic ecosystem change in response to fishing pressure, climate variability and climate change, the combined effects of which create a huge challenge for the fisheries stock assessment in this region. Fisheries stock assessments describe the past and current population abundance of a fish population and importantly, develop a forecast for future population growth. Most stock assessment forecasts are used to set Annual Catch Limits over a 6–36 month scale. This forecast has been typically based on the biology of the fish and is highly uncertain. Incorporating physical environmental variables into the stock assessment population model and subsequent forecast could improve model performance and reduce uncertainty in future population size.

Therefore, a reliable prediction of the NES environmental variables, such as the ocean temperature, could lead to a significant improvement of the fisheries stock assessment. However, the current generation climate model-based seasonal-to-interannual predictions exhibit a limited prediction skill in the coastal environment. On the other hand, recent studies by the PIs reported statistically significant correlations between the NES temperature and the multiple large-scale climate features, such as the Gulf Stream path variability, with some lead time up to a few years, which indicates predictability.

Here, we propose to develop a seasonal-to-interannual statistical prediction system for ocean temperatures on the NES, which will be tailored to the needs of the National Marine Fisheries Service (NMFS) Northeast Fisheries Science Center (NEFSC) for fisheries stock assessment. The measures of uncertainty and predictability skill of the prediction product will be rigorously evaluated against three independent long-term regional hindcast simulations based on probabilistic skill metrics. Our first goal is to use previously described statistical relationships linking shelf ocean temperature to atmospheric circulation, Gulf Stream path, and coastal sea-level to develop a prediction system at the 3–36 month time scale. Our second goal is to evaluate this system in the context of selected stock assessments performed by NOAA Fisheries. Our third goal is to clarify the dynamical basis for the statistical relationships using ocean hindcast models and coupled ocean-atmosphere models.

This proposal is targeting the FY 2017 NOAA Modeling, Analysis, Predictions, and Projections (MAPP) Program solicitation Competition 2: Research to explore seasonal prediction of coastal high water levels and changing living marine resources and its third sub-element: Develop and evaluate experimental probabilistic-based prediction products tailored to the needs of the NOS and/or NMFS, as appropriate, by proposing to develop and test a new statistical seasonal-to-interannual prediction system of NES temperature specifically tailored to the needs of NMFS stock assessment, by the team of PIs including the scientists from NMFS. Our proposed work addresses every element of the NOAA’s long-term climate goal of advancing scientific understanding, monitoring, and prediction of climate and its impacts, to enable effective decisions.

Lead PI: Michael Jacox (University of California, Santa Cruz)

Co-PI: Michael Alexander (NOAA/ESRL), Steven Bograd (NOAA/SFSC, Christopher Edwards (University of California, Santa Cruz), Jerome Fiechter (University of California, Santa Cruz), Elliott Hazen (NOAA/SFSC), Samantha Siedlecki (University of Washington)

One of the greatest challenges in fisheries management is balancing environmental and economic interests by maintaining productive fisheries while limiting bycatch. Static closure rules are often inefficient in this regard, and momentum is building to employ management strategies informed by environmental conditions and predicted species distributions. One such example off the US west coast is the California Current System (CCS) drift gillnet fishery (DGN), whose regulations are enacted by NOAA/NMFS. The DGN targets sustainable stocks of commercially valuable swordfish, but accidental take of non-targeted species (e.g., turtles, cetaceans, sea lions) has led to large-scale fishery closures and declining productivity of the DGN fleet. Recently, dynamic management strategies for the DGN have been explored using observations and statistical models to predict target- by-catch species distributions in near real time, allowing managers to reexamine specific closure rules. Extending these efforts to seasonal forecasts would enhance their utility by giving managers additional time for adaptive responses, but no forecast system currently exists that offers sufficient spatial resolution to capture key physical processes in the CCS and the broad spatial coverage needed to manage wide-ranging fish, turtles, and marine mammals. We propose to produce and validate downscaled seasonal reforecasts for ~3 decades of CCS physical conditions as well as species distributions for target- and by-catch species of interest to US west coast fisheries. Key elements of the proposed work plan are (1) extract and bias-correct global NMME fields, use them to force downscaled reforecasts (i.e., retrospective forecast experiments predicting what happened in the past) of CCS physics, and validate CCS reforecasts with observations, (2) run and validate species distribution reforecasts for target- and by-catch species in the CCS, and (3) determine the added value of an ensemble approach to forecasting living marine resources. Environmental and fisheries data for validating 30+ years of reforecasts is already in hand. The proposed project will provide (i) a set of downscaled seasonal climate reforecasts that can be applied to diverse science and management questions, (ii) target-and by-catch species distribution reforecasts that can be used to reexamine DGN closure rules in collaboration with NOAA/NMFS partners, and (iii) a seasonal forecasting framework that can be applied in fisheries management off the US west coast and elsewhere.

Lead PI: Samantha Siedlecki (University of Connecticut)

Co-PI: Isaac Kaplan (NOAA/NWFSC), Nicholas Bond (University of Washington), Al Hermann (University of Washington), Jan Newton (University of Washington), Mike Alexander (NOAA/ESRL), Simone Alin (NOAA/PMEL) Advisory Council: Joe Schumacker (Quinault Department of Fisheries), Dan Ayers (Washington Department of Fish and Wildlife), Kelly Corbett (Oregon Dept Fish and Wildlife)

The Dungeness crab (Metacarcinus magister) fishery is the most valuable on the US West Coast, with landed values ranging from $100 million to $250 million dollars per year for 2013-2015. In the Pacific Northwest the states and tribes co-manage this fishery and must make critical decisions on seasonal timescales. There is strong interest from crab fishery managers to inform decisions with seasonal forecasts of ocean conditions that are known to affect Dungeness crabs: temperature, oxygen concentrations, aragonite saturation states, and lower-trophic level production rates. We propose to augment an existing forecast system to address the needs of the Dungeness crab fishery managers and stakeholders. This forecast system, JISAO’s Seasonal Coastal Ocean Prediction of the Ecosystem (J-SCOPE, http://www.nanoos.org/products/j-scope/) features dynamical downscaling of regional ocean conditions in Washington and Oregon waters using a combination of a high-resolution regional model with biogeochemistry and forecasts from NOAA’s Climate Forecast System (CFS). We will tailor and test the ability of J-SCOPE to forecast the autumn period critical to crab fishery managers, and determine whether forcing with a multi-model ensemble (North American Multi-Model Ensemble or NMME) reduces bias and better conveys uncertainty, relative to our previous efforts using CFS alone. Using the J-SCOPE  forecasts, we will quantify the relationship between local ocean conditions and three decision- oriented metrics from the crab fishery: (1) meat quality index, (2) spatial and interannual variability in the spatial distribution of crab catch and abundance, and (3) likelihood of summertime hypoxic events. Forecasts for the crab fishery will be delivered via the existing J- SCOPE app on the NANOOS portal (Northwest Association of Networked Ocean Observing Systems) in partnership with state and tribal managers on our Advisory Council. This project addresses directly the second of MAPP’s competitions on advancing the prediction of subseasonal to seasonal phenomena. This project will address objectives 1 and 3 of the MAPP Program.

Lead PI: Mark Merrifield (University of Hawai’i at Mānoa (UH)), Arun Kumar (NOAA/CPC), Gary Mitchum (University of South Florida)

Co-PI: Matthew Widlansky (UH Sea Level Center), Philip Thompson (UH Sea Level Center), H. Annamalai (International Pacific Research Center), William Sweet (NOAA/NOS/CO-OPS), Eric Leuliette (NOAA/STAR), John Marra (NOAA/NCEI)

Coastal high water events are increasing in frequency and severity as global ocean levels rise. With higher relative sea levels, minor coastal flooding is occurring more often during periods of higher astronomical tides. If combined with above-normal seasonal sea levels, often associated with climate-driven variability in the ocean, coastal flooding becomes more severe. Such total high water events expose coastlines to potentially damaging storm-related flooding, yet no seasonal prediction of coastal high water exists on a national scale.

Regional sea levels are affected by the winds, as well as ocean circulation changes. In the Pacific, sea level variations (±30 cm) associated with the El Niño-Southern Oscillation (ENSO), as well as more local processes such as eddies and upwelling or downwelling, impact Hawaii, U.S.-affiliated Pacific Islands (USAPI), and the West Coast. In the Atlantic, sea level anomalies also emerge during El Niño from seasonal changes in storm tracks and winds, impacting the East Coast. On more frequent timescales, fluctuations in the Gulf Stream can produce sea level anomalies (±15 cm) along the U.S. mid- and south-Atlantic Coasts. With recent advancements in forecasting seasonal climate variability using state-of-the-art coupled ocean-atmosphere models, which have the ability to assimilate and predict sea level, come the opportunity to predict the potential for future high water events many months in advance for the U.S. Coast.

Our goal is to construct multi-model forecasts based on seasonal prediction systems and evaluate their skill across the NOAA tide gauge network in the continental U.S., as well as the NOAA and University of Hawaii Sea Level Center (UHSLC) network in Hawaii, the USAPI, the Gulf of Mexico, and Caribbean. Preliminary work suggests that forecast skill can also be improved in some regions by incorporating statistical relationships between observed and predicted sea level variability or by connecting it with the modes of atmospheric variability. From the multi-model forecasts we will provide regional seasonal sea level anomaly outlooks nationwide. A geospatial web portal will be developed to deliver the outlooks, such as high-water alert calendars, which can be combined or incorporated into new or existing NOAA coastal-flood products.

Our team will focus on three objectives by performing the following tasks to deliver a prototype seasonal prediction system:

i. We will explore the processes responsible for sea level variability on monthly to interannual timescales in the Pacific, Atlantic, Gulf of Mexico, and Caribbean coastal regions.

ii. We will process sea level forecasts from operational as well as experimental modeling frameworks to develop a prototype ensemble seasonal prediction system for coastal sea level anomalies.

iii. We will use the multi-model prediction system to provide monthly outlooks for seasonal sea level anomalies across the Nation.

We aim to deliver a framework for seasonal sea level forecasting which strengthens the production of climate data and information that informs the management of climate-related risks. Our forecast framework seeks to reduce the residual between predicted tides and observed water levels by predicting relative sea level changes.

Lead PI: Lesley Thorne (Stony Brook University)

Co-PI: Janet Nye (Stony Brook University), Hyemi Kim (Stony Brook University)

Distributional shifts related to climate variability and oceanographic conditions have been observed across a broad range of marine taxa, and these shifts have both negative and positive socioeconomic consequences. The Northeast United States Large Marine Ecosystem (NEUS LME), a highly productive marine ecosystem that supports important commercial and recreational fisheries, has experienced some of the highest rates of warming in the last few decades. Climate-driven shifts in fish distribution have been observed in this region for most fish species and for whole fish communities. In addition to distributional changes, climate-driven environmental change can influence the phenology of marine processes such as the timing of breeding or spawning, seasonal movements and migrations. Marine organisms show species-specific responses to changing thermal regimes, and thus distributional overlap between species can be strongly impacted by climate-driven change. Therefore, quantifying the biophysical links driving species distributions and understanding how seasonality and climate-driven change together impact living marine resources is imperative to predicting the impacts of future change.

Bycatch, the incidental capture of non-target species in fisheries, is an important source of mortality for several marine mammal and fish species in the NEUS LME, and is strongly impacted by both species overlap and overlap between fishermen and non-target species. In addition to impacts on non-target species, bycatch is a concern for commercial fishermen as it can increase costs and decrease yield. Bycatch could be reduced by incorporating dynamic environmental variables to more precisely estimate spatio-temporal limits on the distribution of marine mammals and commercially important fish. Accurately predicting the distribution of living marine resources on seasonal timescales would be particularly beneficial since it would allow fishing and management approaches to be adjusted based on environmental conditions. Recently developed state-of-the-art seasonal climate models (e.g NMME, S2S) now provide the unprecedented opportunity to make significant strides in the seasonal predictions of living marine resources. However, despite the advantage of the hybrid model in seasonal prediction, this method has been applied only to limited properties of climate phenomena such as hurricane activity. Here we propose to use output from these climate models to generate probabilistic predictions of forage fish and marine mammal distribution in order to inform dynamic management of protected species in the NEUS LME. Specifically, this work will develop bycatch reduction tools to inform decision making for fishermen and managers by highlighting regions that fishermen could avoid in order to decrease the likelihood of bycatch.

The proposed research directly addresses priorities of the MAPP program. A major goal of the MAPP program is to increase the resilience and intelligence of coastal communities through improved products and services relevant to NOAA. The proposed work will contribute to this goal by elucidating the impacts of climate-driven environmental variability on fish and protected marine mammal species, and by generating probabilistic predictions regarding the biological impacts of forecasted changes in the NEUS LME. These products can then be used to inform the management of commercially and ecologically important species, and to provide information required for dynamic management. The proposed studies are thus consistent with the mission of NOAA  MAPP “to enhance the Nation’s capability to predict variability and changes in the Earth’s climate system” and directly contribute to NOAA’s long-term goals, especially for “(1) improved scientific understanding of the changing climate system and its impacts” and “(3) mitigation and adaptation choices supported by sustained, reliable, and timely climate services”.

Lead PI: Kerim Aydin (NOAA/AFSC), Albert Hermann (University of Washington), Michael Alexander (NOAA/ESRL), Phyllis Stabeno (NOAA/PMEL)

Co-PI: Wei Cheng (JISAO, UW/PMEL, OAR, NOAA Affiliate), Kelly Kearney (JISAO, UW/AFSC, NMFS, NOAA Affiliate), Ivonne Ortiz (JISAP, UW/AFSC, NMFS, NOAA Affiliate)

As part of the NOAA Integrated Ecosystem Assessment (IEA) Program, the NMFS Alaska Fisheries Science Center, the OAR Pacific Marine Environmental Laboratory, and the Joint Institute for the Study of the Atmosphere and Ocean at the University of Washington have collaborated to produce seasonal forecasts of the eastern Bering Sea as part of the North Pacific Fishery Management Council’s Ecosystem Status Report. The forecasts incorporate dynamically-downscaling climate information into a regional ocean model coupled to a nutrient- phytoplankton-zooplankton model for the Bering Sea (called Bering10K hereafter). Bering10K model has been tested for the past four years with promising results for a 9-month lead-time forecast of the Bering Sea Cold Pool, a major habitat feature of bottom temperature that determines fish and crab recruitment and distribution. As fish and crab from the eastern Bering Sea account for over 40% of the annual catches in the United States, we envision seasonal forecasts of the Cold Pool, sea-ice cover, ocean temperature, and other outputs to be of interest to numerous stakeholders, from resource management agencies and coastal communities to research institutions and industry service providers. To date, our forecasts have provided direct guidance, in the management council setting, for determining the outlook for key stocks such as Bering Sea walleye pollock and snow crab.

This project would further develop our seasonal forecasting ability, focusing on both technical improvement and applicability. Our goals are to: 1) conduct systematic re-forecasting experiments, driving our regional model with an ensemble of global re-forecasts to assess regional skill and sources of predictability; 2) expand from our present use of NOAA’s Climate Forecast System (CFS) to include other members of the North American Multi-Model Ensemble (NMME); 3) fine tune our regional models to improve accuracy in 9-month forecasts of key environmental features such as sea-ice cover, cold pool, and water column temperature; 4) develop seasonal outlooks of environmental indices as well as fish and crab distribution and recruitment which are management relevant and consistent with ecosystem based management; and 5) conduct a workshop in collaboration with stakeholders from management, fishing industry and Alaskan Native communities to identify additional type, format, extent and frequency of information that would be most useful to those stakeholders for forecast delivery, and design and deliver final products accordingly.

This proposal targets the MAPP competition, specifically on the priority to predict seasonal impacts on the distribution and abundance of fish stocks or other living marine resources. This project will directly increase the production, delivery, and use of climate-related information in fisheries management by producing and applying seasonal environmental predictions directly to the management of living marine resources as per the MAPP topical area. In particular, we have established a direct link between the Bering10K model and fisheries management: seasonal forecasts relevant to managed stocks are provided each year to the North Pacific Fishery Management Council in the direct context of groundfish and crab quota setting, informing final decisions on modifying recommended quotas. Thus, it responds directly to the NOAA long-term climate goals of applying modeling and prediction to maintaining the sustainability of marine ecosystems.

Lead PI: Arthur Miller (Scripps Institution of Oceanography)

Co-PI: Antonietta Capotondi (NOAA/ESRL/PSD), Emanuele Di Lorenzo (Georgia Institute of Technology)

The California Current upwelling system (CCS) supports one of the most productive marine ecosystems in the world and is a primary source of ecosystem services for the U.S. including fishing, shipping, and recreation. Despite the empirical evidence of ENSO influence upon the California Current marine ecosystems, the detailed influence of different ENSO events is unclear, and the degree of predictability of the various ecosystem drivers for specific tropical Pacific conditions has never been quantified. The goal of this proposal is to: 1) Use high-resolution ocean reanalysis of the CCS to link the physical drivers of the CCS ecosystem (temperature, upwelling velocity, alongshore & cross-shore transport) to local climate forcing functions (e.g. alongshore winds, wind stress curl, heat fluxes, precipitation and river runoff) at seasonal timescale; 2) Use long reanalysis products (e.g. SODAsi.3, 20CRv2c, CERA-20C) in combination with multiple linear regression and Singular Value Decomposition to objectively link the climate forcing functions variations in the CCS region with conditions (e.g. sea surface temperature, thermocline depth, sea surface height, tropical wind stresses) in the tropical Pacific that can optimally force them at seasonal timescales; and 3) Use Linear Inverse Modeling (LIM) and the North American Multi Model Ensemble (NMME) to determine the predictability and uncertainty of the forcing functions along the CCS region, compare the LIM and NMME forecast skills, and explore possible sources of error in the NMME models.

The proposed research will directly address the first objective of the call, "Explore how selected modes of climate or ocean variability relate to seasonal variations in fields such as sea level height and ocean temperature that are of primary relevance to predictability for the topical areas of the call, and evaluate the seasonal prediction skill of these modes”, in that it explores how the leading mode of tropical Pacific climate variability (e.g. ENSO) will affect a set of variables, including sea level height and ocean temperature, that are fundamental to ecosystem dynamics in the California Current System. The proposed study also sets the foundation for the development of a probabilistic prediction system, as described in objective 3 of the call, for use in predicting components of the ecosystem that are relevant to the marine resources managed by NOAA NMFS. This study will also utilize and evaluate the latest versions of the century-long SODAsi and 20CR reanalysis products, and also provide an evaluation of a NOAA-sponsored operational forecast system, the North American Multi-Model Ensemble (NMME) in the area of ecosystem predictions.

Lead PI: Scott Sheridan (Kent State University)

Co-PI: Cameron Lee (Kent State University) Key People: Doug Pirhalla (NOAA/CSC), Varis Ransibrahmanakul (NOAA/NOS)

Introduction to the problem: Changes in sea levels have been studied on many spatiotemporal levels, from the local to the global, and short-term to long-term, as well as secular trends. One of the key drivers in seasonal fluctuations in coastal sea levels are ambient atmospheric conditions. Thus, the ability to predict anomalous sea levels should be viewed within the context of the ability to predict the modes of atmospheric variability that affect these seasonal anomalies. The improvement of mid-range to seasonal forecasts of atmospheric conditions has long been a priority of the weather/climate modeling world, and the North American Multi-Model Ensemble (NMME) experiment has been designed to help overcome a number of uncertainties in climate predictions. The ability of models to forecast anomalous sea levels can thus be examined in light of their ability to predict atmospheric circulation. Rationale and objectives: We focus on two main objectives. First, we will assess the relationship between short-term to seasonal-term atmospheric circulation patterns and anomalous coastal sea-level values for all oceanic tidal gauges in the conterminous United States from 1982-2016. Our hypothesis is that the occurrence of extreme atmospheric circulation patterns, as well as the anomalous frequency of these patterns, can be associated with anomalous sea levels locally and regionally on multiple timescales. Second, we will assess the ability of the NMME to successfully simulate both the arrays of atmospheric circulation patterns that are identified, in terms of their overall frequency, persistence, and seasonality, as well as anomalous sea levels using the relationships that were developed. Summary of the work to be completed: We will obtain tidal gauge data for the conterminous US, and classify circulation patterns (CP) using multiple variables, via self-organizing maps. The relationship between CPs and anomalous sea-levels will  be analyzed by examining the short-term and seasonal-term relationships between anomalous sea-level values and individual CPs, and then modeling the time series with non-linear auto regressive models with exogenous input (NARX models). The output of the NARX model will not evaluate the relationship between atmospheric circulation and anomalous sea-level, but also the role of individual drivers. Once this is complete, forecast data from the NMME will be used to evaluate the ability of the model to reproduce observed synoptic circulation patterns, as well as modeled sea-level anomalies.