Skip to main content Skip to secondary navigation

Obtaining the Large-Scale Structure of the Kaweah Subbasin From AEM Data

Main content start

The Place

The Kaweah Subbasin is located in the southern part of the Central Valley, referred to as the San Joaquin Valley, with  the eastern edge defined as the interface between the valley floor and the foothills of the  Sierra  Nevada Mountains.

The Problem

A groundwater model is a fundamental tool for sustainable groundwater management. Generating a high-quality model requires an accurate large-scale structure of the aquifer system such as a regional confining unit and basement surface. While traditional well-based approaches provide accurate point information of the large-scale structure, a large data gap between the wells creates difficulties to capture the continuity of the structure. The airborne electromagnetic (AEM) method, which can rapidly map out the subsurface resistivity of a large water basin, can be an alternate approach for filling in this data gap. However, the AEM method suffers from the limited resolution arising from the diffusive underlying physics creating challenges to accurately image the large-scale structure. 

In the study area, there were two important large-scale features of the aquifer system: 1) regional confining unit referred to as the Corcoran Clay, which divides the upper and lower aquifers; 2) bedrock surface, which is the impermeable base of the aquifer system. These two features were considered to be our “targets”. 

Kaewah Model Cross section
Figure 1. Hydrogeologic section of the Kaweah Subbasin.

Our Approach

For the improved imaging of the large-scale structure, we developed a targeted inversion approach. The approach was designed for the scenario where AEM data are being acquired in an area with no existing groundwater model and limited high-quality well data. To overcome the limited resolution of the AEM data, the targeted approach incorporated the prior knowledge about the targets into the inversion process: a) bedrock surface - there is a large resistivity contrast between the resistivity of saturated sands and clays which makes up the aquifer and the resistivity of granitic bedrock b) Corcoran Clay - there is a large resistivity contrast between the resistivity of the coarser-grained sediments and clay. For choosing hyper-parameters of the targeted approach, high-quality wells were used.

Results to Date

The final outputs of the approach were: a) final resistivity model and b) locations of the targets. A three-dimensional view of the final resistivity model is shown below. In the west, the Corcoran Clay divides the upper and lower aquifers, thinning out towards the east, where the upper and lower aquifers merge. In the eastern part of the subbasin, the resistive bedrock underlies low-resistivity sediments. As expected, we see sharp resistivity contrasts delineating the targets.

Kaweah Resistivity Model
Figure 3. A three-dimensional view of the final resistivity model from the targeted inversion approach.

The estimated locations of the targets were compared with the existing groundwater model of the study area generated by a large amount of well data. The Corcoran Clay from our approach and the groundwater model was compatible (~15% difference) demonstrating the value of our approach for accurately imaging the large-scale structure from the AEM data and limited high-quality well data. The bedrock surface from our approach and the groundwater model was significantly different (~150% difference). We concluded the bedrock surface from our approach was much more accurate than that of the groundwater model. This was based upon the lack of well data covering the bedrock surface along the eastern edge of the subbasin and the presence of considerable topography in the bedrock surface from our approach, as would be expected given the erosional history of an exposed bedrock surface, in contrast to the flat bedrock surface from the groundwater model.

The developed targeted approach is transferable to other regions where there is no existing model but AEM data and limited high-quality well data are available. With an ongoing state-wide AEM project in California ($12M) covering the entire Central Valley, we expect that our approach can be used to generate high-quality groundwater models for many water basins in the Central Valley.

Project Sponsors

California Department of Water Resources

Ministry of Environment and Food of Denmark, the Ecoinnovation Programme

Gordon and Betty Moore Foundation

NASA Applied Sciences Water Resources Program

 

Project Publications and Presentations

Kang, S., Knight, R., & Goebel, M. (2022). Improved Imaging of the Large-Scale Structure of a Groundwater System With Airborne Electromagnetic Data. Water Resources Research, 58(4), e2021WR031439. https://doi.org/https://doi.org/10.1029/2021WR031439

Project Leads / Contacts

Seogi Kang

Meredith Goebel

Rosemary Knight