National Land Cover Database (NLCD) [2024 INV#WO0000000107887]

Overview

Development stage

DeployedSource: c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

2025-10-22 (day precision)Source: 10/22/2025

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Science

Operational functions

No function tag matched Derived by keyword rules; see methodology

Problem the AI is intended to solve

Characterization of the physical land cover is critical for managing the lands, waters, and resources of the United States. The National Land Cover Database (NLCD) is easily one of the most widely used datasets produced within the Department of Interior, with applications including natural resource management, hydrology, biodiversity and habitat, energy and minerals, natural disasters, agricultural sustainability, and more. The criticality of these data demand 1) frequent updates, and 2) a characterization of how our landscapes are changing across time and space. Our linkage of three deep learning models, including a generative transformer-based AI model, has been vital for not only improving our characterization of the landscape with NLCD, but for enabling us to produce land cover products faster and more efficiently. The application of AI to this problem has saved the government valuable resources.

Expected benefits

Our revamping of the NLCD methodology with a series of 3 linked deep learning models has allowed us to improve the product, reduce latency of delivery, and to produce that product more cheaply.

System outputs

Maps of land cover change over time for the United States.

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

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

The National Land Cover Database is built on a foundation of Landsat data, with a mapping interval of 1985 to present. Evaluation of performance of NLCD algorithms depends upon a rigorous accuracy assessment that uses a stratified sampling approach, high-resolution geospatial data, and manual interpretation from land cover experts. High-resolution data is visualized in TimeSync software, sourced from Google Earth and associated commercial high-resolution data such as Maxar and Planet.

Federal Data Catalog

www.usgs.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