Using Video Monitoring to Assess Emigration Patterns for Juvenile Alewife.

Marjadi, Meghna N.*¹, Joel K. Llopiz², Meghan-Grace Slocombe³, John J. Shephard⁴, Sidney Batchelder², Ryan Govostes², Allison H. Roy⁵, ¹Massachusetts Cooperative Fish and Wildlife Research Unit, University of Massachusetts, Amherst, MA, ²Woods Hole Oceanographic Institution, Woods Hole, MA, ³University of Massachusetts, Department of Environmental Conservation, Amherst, MA, ⁴Massachusetts Division of Marine Fisheries, New Bedford, MA, ⁵USGS Massachusetts Cooperative Fish and Wildlife Research Unit, University of Massachusetts, Department of Environmental Conservation, Amherst, MA

Early life stage survival can greatly influence adult population size for anadromous species like river herring (Alosa pseudoharengus and A. aestivalis). Managers and scientists have limited capacity to assess the number of juvenile fishes leaving freshwater ecosystems and the environmental conditions associated with emigration trends. Trapping can only cover limited time periods and areas, while collecting video footage without computational processing is extremely time consuming. River herring juveniles hatch in early spring and stay in freshwater environments for several months before emigrating to the ocean between June and December. Proposed environmental and seasonal emigration triggers include lunar cycles, sharp decreases in water temperature, increased stream discharge, rainfall, water visibility, and zooplankton density. While some studies have identified relationships between proposed triggers and juvenile movement, these studies reported inconsistent emigration trends, covered only one study season, and did not monitor emigration continuously throughout the season. We developed a continuous video monitoring system and computational model to assess juvenile river herring emigration trends over three years (2017-2019). Each year, we deployed our camera system between June and December at the outlet of Benoit’s Pond in the Monument River in Bourne, MA. We developed, trained, and tested a neural network to process video footage. Environmental data (water temperature, zooplankton density) were collected in the field or from other sources alongside emigration footage, and emigration trends will be compared against proposed emigration triggers. We will present our model development and validation process as well as preliminary analyses of juvenile emigration trends.