Forecasting hydrologic controls on juvenile anadromous fish out-migration with process-based modeling and machine...

Area of Science:

• Ecology
• Hydrology
• Fisheries Science

Background:

• River herring (Alosa sp.) are foundational species in various ecosystems.
• Juvenile herring out-migration is vulnerable to stream drying and loss of hydrologic connectivity.
• Current water management decisions lack reliable predictions for out-migration success.

Purpose of the Study:

• To develop a predictive model for short-term forecasting of river herring out-migration loss.
• To support water managers in making informed operational decisions regarding water use and reservoir management.
• To provide a framework for assessing ecological impacts of streamflow connectivity loss.

Main Methods:

• Monitored streamflow and river herring out-migration over two years at three Long Island Sound sites.
• Utilized calibrated Soil and Water Assessment Tool (SWAT) models to generate extensive synthetic streamflow data.
• Trained random forest models using spawning reservoir depth and 30-day precipitation to predict out-migration loss.

Main Results:

• Developed models with 60%-80% accuracy for out-migration loss prediction at a 1.5-month lead time.
• Achieved 70%-90% accuracy for predictions within a two-week timeframe.
• Identified current spawning reservoir depth and prior 30-day precipitation as key predictors.

Conclusions:

• The developed forecasting tool can aid regional water management decisions impacting river herring.
• The model provides timely predictions to mitigate negative effects on herring out-migration.
• The tool's framework can be extended to predict ecological consequences of altered streamflow in impacted watersheds.