Dante Horemans

About

Our research focuses on ecological forecasting and understanding broader ecological dynamics in estuarine and coastal systems. We study both primary producers and higher trophic species using laboratory, in situ, and remote sensing data, coupled with state-of-the-art hybrid numerical modeling approaches. Specifically, we integrate process-based models to identify driving factors and feedback mechanisms within dynamic systems, and combine them with machine learning and satellite remote sensing to improve applicability and predictive capacity.

We have focused on forecasting harmful algal blooms (HABs) through the development of hybrid numerical models that predict bloom occurrence in the Chesapeake Bay based on physical-biogeochemical environmental conditions. We examined how machine learning model complexity (Horemans et al., 2023) and coupling of processed-based models with machine learning (Horemans et al., 2024, 2025) affect the accuracy of our forecasts. Using these insights, we generate daily forecasts of the HAB species Prorocentrum minimum, and more recently including multiple HAB species that are of stakeholders’ concern. These forecasts are accessible through the operational Chesapeake Bay Environmental Forecasting System (https://www.vims.edu/cbefs.