Project Overview

Summary:

    The multidisciplinary team of researchers, led by Dr.  Natalia Sidorovskaia at the University of Louisiana at Lafayette, was awarded an RFP-IV consortium grant of $5,918,726 to conduct the RFP-IV project titled, “Littoral Acoustic Demonstration Center - Gulf Ecological Monitoring and Modeling (LADC-GEMM)”. The project collaboration involved  three other academic institutions (Oregon State University, University of New Orleans, University of Southern Mississippi) and three Gulf of Mexico coastal state small private companies (ASV (Broussard, LA), R2Sonic LLC (Austin, TX), Proteus Technologies LLC (Slidel, LA); 1 Principal Investigator (PI) (Sidorovskaia); 8 co-PIs (Azmy Ackleh, Juliette Ioup, David Mellinger, Christopher Tiemann, Kenneth Barbor, Danielle Greenhow, George Ioup, Stan Kuczaj); 5 Postdoctoral Research Associates (Thomas Guilment, Kun Li, Baoling Ma, Amy Veprauskas, Aijun Zhang); 13 PhD-level students (Britt Aguda, George Drouant, Selene Fregosi, Michael Haas, Md Istiaq Hossain, Jack Lebien, Kendal Leftwich, Bradley Sciacca, Joshua Veillon, Audra Ames, Kirk Bienvenu, James Leclere, Tingting Tang); 9 M.S.-level students (Matthew Firneno, Selene Fregosi, Yingxue Gao, Bibek Khadka, Britt Aguda, Audra Ames, Kirk Bienvenu, Jack Lebien, Sakib Mahmud); 13 undergraduate students (SydniCherise Austin, Simeon Benit, Wesley Cognevitch, Amanda Demesia, Conner Lemaire, Charles Maillho, Danielle Maurin, Matthew Firneno, Jared Risbourg, Bradley Sciacc, Joshua Veillon, Evan Wellmeyer, Trevor Wood); 5 high school students (Sarah Anselmo, Allison Chevis, Charlotte Fowler, Alexandria Hahn, Corinne Martin); and several research scientists, engineers, technicians,  and administration personnel.

     The LADC-GEMM consortium research seeks to advance our scientific understanding of how changes in the marine mammal distribution and abundance, caused by environmental stresses or human activities, could have a major impact on the function of the entire deepwater ecosystem.  This multidisciplinary consortium effort, using expertise from marine acoustics, biology, physics, engineering, mathematics, and computational predictive modeling, will provide data-driven answers to  how the regional marine mammal population in the Northern Gulf of Mexico (GoM) has been affected by the 2010 Deepwater Horizon (DWH) oil spill. The research results will supply the datasets for regional stock populations for the key GoM species of deep-diving marine mammals (endangered sperm whales, beaked whales, and dolphins) and will assist to formulate recommendations for needed mitigation efforts to improve various stock recoveries. The advancements in passive acoustic data collection and processing will provide tools for exploring relationships of regional abundance variations to long-term and short-term environmental factors, such as environmental disasters, weather conditions, natural seasonal migration, regional anthropogenic noise soundscapes due to industrial operations in an area, food supply, etc. The outcomes of the research, based on previously collected acoustic data from both before and immediately after the oil spill, plus new data collections, will not only provide an assessment of the long-term environmental impact of the recent spill event on deep-diving marine mammals, but will also aid in improving regulations, monitoring, and mitigation efforts for everyday industrial operations in the northern GoM. The new data processing and predictive modeling tools developed over the duration of this project will not only apply for forecasting regional population dynamics of marine mammals in the GoM but will also be applicable for studying the effects of environmental stressors on  marine mammal populations in other parts of the world's ocean.

     The LADC-GEMM consortium is in a unique position among those conducting passive acoustic studies in the GoM given its access to data unavailable elsewhere. Prior to the 2010 oil spill, LADC had conducted six broadband passive acoustic surveys in the GoM. In 2007, LADC conducted a two-week visual and acoustic survey of marine mammal activity just 9 miles and 23 miles from the DWH spill site, giving LADC a unique pre-spill baseline dataset of marine mammal activity and anthropogenic soundscapes near the oil spill site.  Earlier surveys had also been conducted at sites 50 miles from the incident site. In September 2010, LADC returned to those same survey sites to repeat underwater acoustic recordings, gathering data to support the first and possibly only comparisons of pre- and post-spill estimates of the marine mammal abundances in the vicinity of the event based on their acoustic activity.

     The major LADC-GEMM consortium goals are to:

• Establish a precedent of long-term consistent passive acoustic monitoring (PAM) of the marine mammal recovery after the oil spill, based on previously collected baseline data, continued data collection utilizing advanced PAM technologies, and development of population dynamics prediction models;
• Design and test a new cost-effective PAM approach for near real-time detection, characterization, and monitoring of the impact of environmental changes of different magnitude and duration on deep diving GoM marine mammals by utilizing the integrated experimental capabilities of the consortium, which will include bottom-moored listening buoys, deep-diving Seagliders, and autonomous surface vehicles;
• Develop an integrated acoustic data processing technique, which will allow in-depth understanding of the relations between observed/predicted abundance variations and underlying reasons decipherable from collected acoustic data, such as anthropogenic noise soundscapes, food supply, stock composition (adults versus calves, females versus males, etc.), and seasonal migrations; and
• Communicate the methods used and the results obtained to the public via presentations, teacher-training events about the technologies used, scientist-in-residence programs, web sites, videos, etc.

Research Highlights

     The LADC-GEMM consortium research, which included 82 outreach products and activities, resulted in 12 peer-reviewed publications and 85 scientific conference presentations to date and 43 datasets submitted to the GoMRI Information and Data Cooperative (GRIIDC), which are available to the public.  Significant outcomes of their research (all related to GoMRI Research Themes 3 and 4) are highlighted below.

Highlights: 

• LADC-GEMM consortium collected over 100 Tb of high-quality broadband passive acoustic data in the Mississippi Valley/Mississippi Canyon region of the GoM. These publically available datasets will serve generations of ocean scientists providing valuable snapshots of the GoM soundscapes across 2015-2017. These data could be harvested to answer a broad range of current and future questions about the GoM deepwater ecosystem. The utilization of these data will grow as scientists develop more advanced algorithms to decode information hidden in the data.
• The LADC-GEMM team was the first team to test and improve the effectiveness of Autonomous Surface Vehicles (ASV) and Glider-based passive acoustic monitoring platforms in the GoM environment.  We also developed and tested real-time communication and data-reporting capabilities for autonomous deep-water moorings (EARS-buoys). LADC-GEMM was the first team to record GoM beaked whales in 2007 and the first team to record beaked whales and Kogia’s species on the PAM-array towed by the ASV in 2017.
• The combination of real-time acoustic observations by an ASV-towed array and visual observations provided new inputs into the acoustic library of sounds produced by different species of GoM dolphins.
• Tailoring  Unsupervised Machine Learning techniques to passive acoustic data, our researchers developed the first automatic species-level classification algorithms for beaked and sperm whales for robust detection of these animals in Big Data. 
• Species-level classification is vital to studying regional habitat use by different species of beaked whales. The data processing results suggest that Cuvier’s beaked whales are the dominant species in the Mississippi Canyon region of the Northern GoM. The results show that Cuvier’s beaked whales tend to be more acoustically active at the deeper sites of the continental slope. Gervais’ beaked whales vocalize mostly at the shallower waters  and the number of detected clicks is consistently smaller over the same temporal intervals.
• We developed the region- and instrument-specific approaches to estimate the probability of detection of beaked and sperm whales from acoustic data to increase the accuracy of regional density estimations for beaked and sperm whales.
• The maximum estimated densities of both Gervais’ and Cuvier’s beaked whales in the region over the 12-day period in September 2010 were not achieved during the 120-day observational periods in 2015 and 2017. Two distinct implications (not mutually exclusive) are: (1) animals moved to the area due to higher prey availability (all anthropogenic activities were closed) and consumed polluted prey with significant negative health implications in 2010 and (2) beaked whale abundanbce in the region declined after the oil spill. 
• Observations of the long-term (from 2001 to 2015) regional abundance trends relevant to the habitat change and oil spill impact using passive acoustic monitoring for sperm whales indicate that  (1) sperm whales relocated away from the closest to the spill site in Fall 2010 and (2) sperm whale’s abundance in the region decreased in 2015 in comparison with pre-spill years dating back to 2001 and did not returned back to pre-spill levels in 2017.
• We advanced our understanding of Cuvier’s beaked whale click structure that will have a significant impact of the development of future detection and classification algorithms. The results show that observed Cuvier’s beaked whale clicks are a superposition of upsweep and downsweep chirps unlike the currently accepted upsweep-only structure of beaked whale clicks in the bioacoustics literature. By extension, it suggests that downsweep clicks found in passive acoustic recording data in the same frequency bandwidth as Cuvier’s beaked whale clicks belong to Cuvier’s beaked whale.
• Inferences from recorded data allowed us to confirm the ecological niche hypothesis of habitat-division between Cuvier’s and Gervais’ beaked whales in the region which will have an impact on population density estimations from visual and acoustic observations and on the design of future monitoring surveys.
  • We developed a novel matrix population model based on the life history cycle of the individuals to describe the population dynamics of the GoM sperm whale population with demographic stochasticity which accounts for individual differences in vital rates including reproduction and survivorship.
  • Using the above matrix model, we developed an analytical formula for calculating the population mean-recovery-time as a function of the magnitude and duration of the disturbance. This new formula has the advantage of saving substantial time that is usually required to compute thousands of sample paths of the model to obtain mean and variance of recovery time.  This work received the Rollie Lamberson Research Award Medal from the Resource Modeling Association in 2019.
  • Motivated to understand the interaction between the whales and its prey in the Gulf of Mexico, we developed a new predator-prey model for food-chain connectivity. Using Darwinian evolutionary game theory methodology, we modeled the impact of evolution in the prey to resist toxicants on the overall system dynamics. The results show that fast evolution to toxicant resistance in prey can help both predator and prey populations to persist during a disturbance when, without evolution, both may continue to decline if no targeted management strategies are applied.

 Proposal Abstract - RFP-IV PI Natalia Sidorovskaia