Improving Prediction Capabilities for Barrier Island Landscape Change

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

2023-06-01 (day precision)Source: 6/1/2023

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Science

Operational functions

Image and video analysis Derived by keyword rules; see methodology

Problem the AI is intended to solve

This research uses several AI/ML tools to observe and analyze coastal landscape change at critical habitats along barrier islands. The work employs traditional ML methods (random forest model), and more cutting-edge methods, like foundation models (e.g. Segment Anything). AI tools that leverage human-in-the-loop methods (e.g. Doodler) enable efficient and reliable landcover map creation from aerial and satellite imagery and enhanced historical imagery that provides important long-term perspectives of landcover change. This work aids land management decision-making and risk assessments and provides data for model validation.

Expected benefits

More accurate and longer-term information leads to more efficient and effective resource management, protecting property and national interests.

System outputs

Data products include landcover maps and change maps; Publications describe model validation for coastal natural hazards and landscape evolution; Continued development of AI/ML methods and workflows to inform landscape change assessments.

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

Not reported

Sourcing

How it was built

Contract and in-houseSource: c) Developed with both contracting and in-house resources

Vendor (as reported)

Marda Science, LLC; ESRI; Python libraries (open-source)

Vendors (standardized)

Esri

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

Coastal aerial imagery and ground-reference data (example data release with attributed information is located at https://www.sciencebase.gov/catalog/item/67927873d34e88f5864c49b0)

Federal Data Catalog

Not reported

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