|
Disclaimer: This is a personal page that reflects my own experiences, and not necessarily those of my current employer.
I work in marine resource management by blending ecological principles with mathematical modeling. I am interested in being a part of innovative conservation strategies for fishery management. I am currently employed with NOAA's National Marine Fisheries Service, Southwest Fisheries Science Center. I completed my master's degree at the School of Aquatic and Fishery Sciences at University of Washington, and am working on my PhD research in the same department. For my PhD project, I am working toward improving fisheries management through novel methods developments in machine learning applied to messy ecological data. I am also working toward multi-species management strategy evaluation for US West Coast forage fish fisheries. My published research includes developing improved machine learning methods for predicting messy fisheries data, shown through predictions of market squid (Doryteuthis opalescens) abundance. My current manuscript in progress takes these methods a step further by assessing the impact to forecasting given different forms of time series (e.g. regimes, environmental trends, cycles, random walks, ARIMA, black swan events, etc.), and finding that not all time series (and not all methods) are created equal. This work serves as a wake-up call to the fragility of seemingly robust machine learning methods when faced with messy data, showing that model-centric modelling is a trap without a data-centric understanding of how variable interact and perform in forecasts. My masters research developed an age-size-structured fishery assessment method, and applied it to a mix of data-rich and data-poor Alaskan populations, focusing on Eastern Bering Sea Tanner crab (Chionoecetes bairdi), Pribilof Island Blue King Crab (Paralithodes platypus), and Eastern Bering Sea Pacific cod (Gadus macrocephalus). We also explored the impact of uncertainty in these stock assessment methods. Other academic work includes bio-economic analyses of intersections between fisheries and fishers, as well as triple-bottom-line analysis of various fisheries management methods (ecology, economics, and communities). My earlier work includes responses of California market squid (Doryteuthis opalescens) to changes in the California Current Ecosystem, including this incredible opportunity to work with Smithsonian Magazine to show what amazing creatures squid are. Disclaimer: below is a description of my personal experience of the stock assessment process and not an official NOAA statement: In my role as a stock assessment scientist for the US, I use statistical models to develop estimates of fish populations and abundance for sustainable management. This is a rigorously reviewed process, which includes several processes that are overseen by the Fishery Management Councils. There are methods reviews for any methods used-- from data collection and processing to model estimation. There is a 3-day stock assessment review (STAR) process for any benchmark assessment, in which the stock assessment scientists (STAT) defend their assessment to a Council-appointed Scientific and Statistical Committee (SSC) sub-panel, as well as representatives from the Center for Independent Experts (CIE). These STAR panel reviews are open to the public. Update assessments are reviewed by an SSC sub-committee, and are often accompanied by strict rules regarding what updates are allowed. For the US West Coast, this means that in a catch-only update, only catch data can be added to the model and no other data and no model changes are allowed; in a regular update, new data only from data sources already in the most recent benchmark are allowed, and no model changes are allowed. There is typically a Terms of Reference (TOR) document that outlines specific rules for the Council processes and stock assessors. All finalized stock assessments should be accessible through the NOAA library. The stock assessment biomass estimates are then used by Councils and NOAA Regions to set catch limits and other metrics. Official NOAA statements regarding stock assessment can be accessed here. |