Center Pivot Irrigation Systems and Where to Find Them: A Deep Learning Approach to Provide Inputs to Hydrologic and Economic Models

Daniel Cooley,Reed M Maxwell,Steven M Smith

doi:10.3389/frwa.2021.786016

Abstract

Availability and quality of administrative data on irrigation technology varies greatly across jurisdictions. Technology choice, however, will influence the parameters of coupled human-hydrological systems. Equally, changing parameters in the coupled system may drive technology adoption. Here we develop and demonstrate a deep learning approach to locate a particularly important irrigation technology—center pivot irrigation systems—throughout the Ogallala Aquifer. The model does not rely on super computers and thus provides a model for an accessible baseline to train and deploy on other geographies. We further demonstrate that accounting for the technology can improve the insights in both economic and hydrological models.

Highlights

Groundwater is a critical resource globally and in the U.S While it can be a renewable resource, it is rapidly becoming overexploited (Bierkens and Wada, 2019) with total extractions increasing by 8.3% from 2010 to 2015 while surface water use has declined by 13.9% in that same period (Dieter et al, 2018). This overuse has led to declining groundwater levels, and the Ogallala Aquifer is a prime example of this (Haacker et al, 2016)
Much of the decline in the Ogallala Aquifer’s water level can be accounted for by the agricultural sector which has overwhelmingly adopted center pivot irrigation systems (CPIS) that allow farmers to pump groundwater onto their fields efficiently (Gowda et al, 2018; USDA NASS, 2018). This decline of the water level imposes a real cost on groundwater users in terms of electricity use and well upkeep as they must drill their wells deeper which increases the effort necessary to transport the water to the surface (USGS, 2013)
Identifying CPIS from aerial or satellite imagery is a time-consuming and tedious process that may be cheaply replaced by utilizing the open-source deep learning model presented in this paper

Summary

INTRODUCTION

Groundwater is a critical resource globally and in the U.S While it can be a renewable resource, it is rapidly becoming overexploited (Bierkens and Wada, 2019) with total extractions increasing by 8.3% from 2010 to 2015 while surface water use has declined by 13.9% in that same period (Dieter et al, 2018). Irrigation offsets losses in dry and significantly dry years, but CPIS does even better at maintaining production levels closer to that in normal years The upshot of these three exercises is that economic relationships are sensitive to the irrigation technology present and the deep learning model can provide improved inputs for the economic models. These models all require many input parameters to define the location and extent of irrigated area and pumping wells As these input data are often not directly available, they are often based on coarser resolution (e.g., county level) information or land cover classification from nation datasets The results of our deep learning model allow for direct input of CPIS into hydrology models, providing a new way to estimate the effects on groundwater pumping The results of this current work provide an unprecedented, high resolution input dataset for the location of irrigation and pumping wells in the Ogallala Aquifer. The methodology used here can be applied over time to understand the expansion of groundwater pumping and depletion

CONCLUSION

Findings

DATA AVAILABILITY STATEMENT