Random forest models for predicting water quality of inland waters from remotely sensed imagery

Overview

Development stage

PilotSource: b) Pilot – The use case has been deployed in a limited test or pilot capacity.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

2025-05-01 (day precision)Source: 5/1/2025

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Energy and the Environment

Operational functions

Image and video analysis Derived by keyword rules; see methodology

Problem the AI is intended to solve

Inland waterbodies (i.e. rivers, lakes, reservoirs, ponds, etc.) can face issues of poor water quality which may pose issues to water users, water infrastructure, and ecosystems. While it is important to monitor the water quality of these waterbodies, it can be costly and labor intensive to continuously monitor sites across the US. This project aims to improve the efficiency of monitoring by using Machine Learning (Random Forest) models to predict two water quality parameters of waterbodies seen in remotely sensed images. These remotely sensed images span the entire Conterminous US (CONUS) and are collected approximately every 5 days, allowing for water quality monitoring at high temporal and spatial resolutions with relatively low effort and cost.

Expected benefits

cost savings by estimating water quality parameters across the entire country on a frequent (5 day) schedule with minimal labor. By making these modeled water quality products publicly available, it may aid in managing water resources

System outputs

Prediction of chlorophyll concentrations and turbidity values in remotely sensed images of water bodies.

Technology

AI classification

Classical / predictive MLSource: Classical/Predictive Machine Learning: Models trained on data to make predictions or classifications based on identified patterns or relatio…

System name(s)

Not reported

Custom-developed code

No

Public source code

www.sciencebase.gov ↗

Sourcing

How it was built

Developed in-houseSource: b) Developed in-house

Vendor (as reported)

Not reported

Vendors (standardized)

None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII

No

Privacy Impact Assessment

Not reported

Authorization to Operate

No

Demographic features

k) None of the above

Training and evaluation data

Data used to train the models are remotely sensed images from the European Space Agency's Sentinel-2 satellite program and in-situ USGS water quality data

Federal Data Catalog

www.sciencebase.gov ↗

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Potential impacts and how they were identified

Not reported