Skip to main content
ARS Home » Pacific West Area » Kimberly, Idaho » Northwest Irrigation and Soils Research » Research » Research Project #444583

Research Project: Mapping Irrigation Methods using Artificial Intelligence

Location: Northwest Irrigation and Soils Research

Project Number: 2054-13000-010-010-S
Project Type: Non-Assistance Cooperative Agreement

Start Date: Jul 1, 2023
End Date: May 31, 2028

Develop a tool using Artificial Intelligence to map irrigation methods in the northwestern United States. This product will allow the identification and mapping irrigation methods classified as surface, sprinkler or other irrigation methods in irrigated areas of the northwest. Maps of irrigation methods of major agricultural areas in the northwestern region will be developed for years with available input data.

Publicly available earth observation satellite products such as Landsat will be used to develop this irrigation method mapping technique. A previously developed deep learning model based on the U-Net architecture will be evaluated for its generalization capability in irrigated areas of the northwestern U.S. Various improvement strategies to this initial deep learning model will be explored. Potential improvement strategies to explore include addition of vegetation and/or moisture indices as model inputs, alternative input image normalization, inclusion of additional spatial information, or higher resolution satellite products. Training data will be extracted from the Utah Water Related Land Use (WRLU) dataset. This dataset is collected by the Utah Division of Water Resources as part of the state water plan. This dataset is available from years 2003 to 2021. The dataset contains many important water-related attributes collected at the field scale across the state, including irrigation methods classified as sprinkler or flood irrigation, dry crop, sub-irrigated, drip, and none.