Title:

Data Science for Storm Events

Storms change the concentration of contaminants in watersheds, which in turn affects the concentration in all surrounding bodies of water. Clearing watersheds of such contaminants is crucial to preserving aquatic ecosystems. If we derive a log-log relationship between flow rate and water solute concentration, one could predict the concentration of solutes in watersheds, allowing for expedited contaminant cleanup. By providing a utility for researchers to predict solute concentrations in watersheds after storm events have passed, researchers will be able to help create state budgets for watershed cleanup. Data has been collected from ten different sites in New Hampshire consisting of solute measurements, flow rate and other various readings. Our project is focused on various machine learning techniques, such as logistic regression and data cleaning, to train and deploy models that can detect storm events by analyzing flow rate. Being able to detect storm events allows researchers to look at historical events and their associated solute concentrations. This project lays down the groundwork for researchers at the University of New Hampshire’s College of Life Sciences and Agriculture to build models that can predict solute concentrations after a storm event has passed.

Name

Comment*

Comments are viewable only by submitter

Captcha*