Skip to main content
Quantifying Impacts of Agricultural Land Use and Irrigation Practice on High Resolution Regional Weather Prediction
Project Start Date
05/05/2025
Project End Date
05/05/2028
Grant Number
24-LCLUC24_2-0012

Team Members:

Person Name Person role on project Affiliation
Cenlin He Principal Investigator National Center for Atmospheric Research (NCAR) , Boulder   , USA    
Zhe Zhang Co-Investigator National Center for Atmospheric Research (NCAR) , Boulder    ,  USA
Tzu-Shun Lin Co-Investigator National Center for Atmospheric Research (NCAR),  Boulder    ,   USA    
Liping Di Co-Investigator George Mason University, Fairfax, US
Abstract

The US Corn Belt is one of the most important food baskets in the world where agricultural management practices have substantially altered the landscape and regional weather and climate through land-atmosphere interactions. Evidence showed that the expansion and intensification of croplands and irrigation practices in this area have resulted in notable changes in regional weather patterns (e.g., temperature cooling and precipitation increase). However, the intricate interplay among crop dynamics, irrigation, groundwater, and the atmosphere in this region and associated impacts on short-term weather prediction have not been fully understood or quantified particularly at high-resolution. Moreover, current weather prediction models suffer from errors caused by inaccurate, static, outdated, and/or coarse resolution input data of agricultural land use and irrigation activities, as well as the lack of model representation of crop and irrigation dynamics during prediction periods. Our proposal team members have recently developed a series of machine learning (ML) based techniques to create field-scale accurate historical and in-season crop and irrigation maps based on satellite observations. This provides a promising opportunity and foundation for this proposed project to enhance agricultural land use and irrigation input maps for weather prediction models based on NASA satellite data.

This project aims to leverage a suite of remote sensing data and ML techniques to develop a suite of dynamic high-resolution annual crop and irrigation data over the continental US (CONUS) during 2000-present and incorporate them into the state-of-the-art community WRF/NoahMP-Crop model to improve convection-permitting weather prediction and the knowledge of agriculture-weather interactions. We will use the US Corn Belt as a testbed and address two science questions:

          Could we improve weather prediction by developing and implementing dynamic high-resolution annual crop and irrigation maps in WRF?

          What are the key factors and mechanisms controlling agricultural impacts on weather prediction and associated uncertainties, particularly for extreme events?

Accordingly, we will conduct three tasks:

Task 1. We will replace the current static and outdated crop and irrigation input maps in WRF by developing a suite of annual maps (2000-present) for field-scale (30-m) historical and in-season crop types with rotation patterns, state-level crop planting and harvesting dates, and 4-km crop growing degree days (GDDs), as well as 5-yearly field-scale (30-m) irrigation area maps.

Task 2. We will implement the new data (Task 1) into the widely-used community WRF/NoahMP-Crop model to conduct short-term (1~7 days) convection-permitting (4-km) numerical weather predictions/hindcasts and detailed model evaluation over the US Corn Belt during major crop seasons (May-September) for representative years (e.g., wet, dry, normal years). 

Task 3. We will conduct a series of model sensitivity simulations to quantify key factors in affecting weather prediction, associated mechanisms, and uncertainty, during both non-extreme and extreme conditions (i.e., extreme precipitation, heat waves, and droughts).

Project Research Area