Estimating Unmetered Ground Water Irrigation Demand with High-Resolution Remote Sensing Data
Project ID: 1545
Principal Investigator: Eve Halper
Research Topic: Water Resource Data Analysis
Priority Area Assignments: 2012 (Climate Adaptation), 2013 (Climate Adaptation)
Funded Fiscal Years: 2012
Keywords: unmetered wells, domestic wells, exempt wells, remote sensing, vegetation indices, lidar,
In many watersheds, a substantial amount of ground water pumping is attributed to unmetered domestic wells. Well owners are not required to meter or report their water use, and the sheer number of permitted wells makes it impractical to install measuring devices. The lack of information on domestic well pumping introduces substantial uncertainty into the regional water balances that form the basis of Bureau of Reclamation (Reclamation) supply augmentation studies. As a result, projects may be undersized and, therefore, lack effectiveness or be oversized and be more costly than necessary.
We propose the development of a method that uses Geographic Information System (GIS) data and high-resolution remote sensing data, which is increasingly available in many communities, to identify vegetation irrigated with domestic wells. The same data can be combined with widely available meteorological data to estimate evapo-transpiration (ET) rates. Together, these techniques can generate a direct estimate of consumptive water use by domestic wells. Our area of focus is the Sierra Vista Subwatershed in southeastern Arizona, but methods should be generalizable to other areas, especially arid and semiarid regions.
Specifically we ask:
Can irrigated residential landscapes be effectively separated from vegetation supported by natural water sources using high-resolution pattern analysis techniques and Light Detection and Ranging (LIDAR) derived hydrography?
Does the combination of multispectral imagery and LIDAR data provide enough information to categorize high and low water demanding plant species?
Can a reasonable estimate of ET of this vegetation be generated using a combination of a remote-sensing derived vegetation index, commonly available meteorological data, and plant-specific ET coefficients?
How do these estimates compare with the values currently being used for domestic well irrigation demand?
Need and Benefit
Reclamation water supply studies typically begin by establishing a water budget to gauge the impact of different types of water demands (agricultural, domestic, etc). Large surface diversions and ground water wells are generally metered, and the quantification of these uses is straightforward. However, many watersheds contain substantial numbers of small, unmetered wells that are legally exempt from reporting their pumping.
The volume of consumptive use from these wells is a great source of uncertainty in water balance analyses, and the use of estimates from other watersheds to approximate this demand is the focus of much controversy among stakeholder groups. This situation has been documented in Reclamation studies of the southeast Arizona's Sierra Vista Subwatershed, northern Arizona's Mogollon Rim, and the Washington State's Yakima River Basin. A cost-effective method of directly estimating the consumptive water demand of unmetered wells could resolve this controversy and expedite regional decisionmaking for supply augmentation projects.
A direct method for estimating domestic well demand would also reduce the likelihood of either underestimating and overestimating the water supply augmentation needs. An underestimate of demand could result in the under-sizing of an augmentation project, compromising a solution's effectiveness and leaving important needs unmet. Overestimating a project's size could lead to the waste of natural and financial resources by providing more water than is actually necessary. Both of these outcomes are detrimental to Reclamation's goals of providing water to communities and preserving the environment. The development of a verifiable method for estimating irrigation by unmetered wells is therefore critical to the appropriate design of Reclamation augmentation projects in areas with large numbers of unmetered domestic wells.
In order to be widely useful, a process for estimating irrigation demand from unmetered domestic wells should require only commonly available data. This proposal employs datasets that are typically acquired for other purposes (GIS, multispectral imagery, and LIDAR) and re-purposes them for an additional use. In addition, Nagler's method for estimating ET using vegetation indices has already undergone extensive field testing in several Western watersheds in recent Reclamation-funded studies. Hence, this proposal represents further development of Reclamation-sponsored research.
The following documents were not reviewed. Statements made in these documents are those of the authors. The findings have not been verified.
Final Report for Estimating Unmetered Ground Water Irrigation Demand with High-Resolution Remote Sensing Data (final, PDF,
By Eve Halper
Report completed on May 29, 2014
It may be possible to answer this question with better resolution data sets, in a different study area, or where irrigated plant types are distinct from background vegetation. However, with the data available in southestern Arizona, such an analysis was not possible.