Reducing Uncertainty in Estimates of Irrigation Water-Use

Author(s): Westerman, D.; Wilson, J.; Painter, J.; Torak, L.

The Mississippi Alluvial Plain (MAP) is one of the most important agricultural regions in the United States, and the MAP region has seen substantial declines in groundwater levels and reductions in stream base flow that have led to concerns about sustainability and future availability of the water resources. One of the tools used to understand groundwater responses in the MAP region to current and future water-use demands is groundwater modeling; however, one of the largest sources of uncertainty in groundwater modeling of the MAP region is irrigation water-use estimates. The U.S. Geological Survey (USGS) is working closely with local and state cooperators to help improve estimates of water-use demand within the MAP region. A publicly available water-use network is being established that includes 46 real-time monitoring flow meters installed on irrigation wells. In addition to the real-time data, existing State programs will be leveraged to obtain water-use measurements from hundreds of additional metered sites within the MAP region. The metered water-use data will be essential in providing authoritative datasets for estimating water-use demands based on crop types, climatic variables, and the variety of soil types present within the MAP region. Using the water-use metered data as the main driver, the USGS is developing a national water-use model with the goal of estimating monthly groundwater use for irrigation at a spatial resolution of 1 kilometer. The initial version of the water-use model will be aimed at quantifying irrigated acres, estimating irrigation rates developed from current water-use metered data, and developing estimates of water use for the entire MAP region. Future, more sophisticated versions of the water-use model will aim to incorporate additional site-specific water-use data, develop irrigation rates as a function of climate variation based on the metered data, use remote-sensing data to estimate irrigated acres, and implement geostatistical and machine-learning approaches to spatially and temporally estimate groundwater use for irrigation. The real-time flow meter data collected as part of this project, when coupled with real-time remote sensing data, will allow for real-time prediction of water use. Estimates from the water-use model will be used directly as input into the current groundwater-flow model, which will help guide future refinement of both the water-use and groundwater models, capture uncertainties in the data, and identify data gaps.

Go back


Past Conference Archive