Research Project: The focus of this research training opportunity is remote sensing, specifically:
- development and application of mixture density networks (MDN), neural network approaches to predict/map chlorophyll concentrations; and
- development of predictive models for cyanobacterial blooms, both within estuaries and freshwater tidal rivers of the United States, using remote sensing images (e.g., Sentinel 2) as inputs.
The tools to be used include an extensive EPA database of paired chlorophyll/remote sensing reflectances with ancillary explanatory variables, USGS Earth Explorer, Google Earth Engine, toolboxes for atmospheric correction (e.g., ACCOLITE, POLYMER, SIAC), R and Python programming languages, an existing MDN Python toolbox for neural network analysis, and supercomputer access if needed.
The project integrates remote sensing, programming, and applied statistics.
The research participant may be involved in the following research activities:
- developing and applying approaches for bulk download of remote sensing images
- application of toolboxes for atmospheric corrections
- application of MDN methods for chlorophyll prediction, and
- development of logistic models to predict cyanobacteria metrics.
Learning Objectives: The research participant may learn about atmospheric corrections for Sentinel 2 data, managing and analyzing large imagery datasets within Google Earth Engine, machine learning approaches for predicting chlorophyll a, and development of predictive logistic models for cyanobacteria. The research participant may author or co-author on peer-reviewed publications, and may present at local and national meetings (possibly virtually). The participant will be a member of a multi-disciplinary research team.
https://www.zintellect.com/Opportunity/Details/EPA-ORD-CEMM-ACESD-2021-04