Estimating fifty-two years of groundwater levels in different aquifer layers in the southern San Joaquin Valley, California

Abstract

Over the past century, California's San Joaquin Valley has faced dramatic groundwater level depletion. Significant spatial and temporal gaps in the records, unreliable measurements, and unknown depths to which most wells are drilled, hinder effective groundwater monitoring in the region. This study presents a novel method, which integrates a time series technique called Small Baseline subset (SBAS) with kriging to estimate yearly changes in groundwater levels, as well as total absolute head from 1971 to 2023 in the Southern San Joaquin Valley across two distinct aquifer layers – one shallow (mostly unconfined) aquifer and one deep (confined) aquifer. Firstly, 1,197 wells with known depths based on the depth of an extensive confining layer, the Corcoran Clay, were classified. Further, 348 wells with unknown depths, which were within the clay boundary, were categorized by examining the correlations and differences in the average yearly fluctuations of groundwater levels with the 1197 wells. Out of these 348 wells, 215 wells belonged to confined and 133 were assigned to mostly unconfined aquifers. 3,039 wells were outside the Corcoran Clay Layer, which were classified as mostly unconfined because they exhibited a similar distribution of seasonal groundwater fluctuations to the mostly unconfined group. For each of the aquifers, we used ordinary kriging to estimate groundwater level change over specific intervals (every 1, 2, 3, 4, 5, 6, 7, and 8 years for all wells with available data) across the study area. Using a system of linear equations, we then solved for the yearly change in groundwater level from 1971 to 2023. The mean reduction in groundwater level from 1971 to 2023 was observed to be 16 m for mostly unconfined aquifers and 23 m for confined aquifers. The groundwater levels declined sharply for both aquifers during drought periods because of the increased reliance on groundwater for irrigation. In addition, we used our predicted groundwater level data to estimate that a total of 21 km3 (0.8 to 2% of the total freshwater groundwater storage in the Central Valley) water was lost from the study area between 2015-2023, with most of the storage loss (76%) coming from mostly unconfined aquifers, followed by compaction of aquifer matrix (24%).