Lake sediments record past hydrologic variability, but natural lakes are often sparse in semiarid and arid regions, making the calibration of paleohydrologic models a challenge. At Lake Elsinore, the largest of the few natural lakes in Southern California, we explore and develop a novel transfer function approach for reconstructing lake depth. Using 32 modern surface sediment samples spanning Lake Elsinore’s littoral to profundal zones, we establish a statistical relationship between lake depth and sediment elemental geochemistry composition analyzed via X-ray fluorescence (XRF). We develop lake depth transfer functions using weighted averaging-partial least squares (WA-PLS) and modern analog technique (MAT). Application of the WA-PLS C5 elemental geochemistry-based transfer function to Lake Elsinore sediment cores reveals a climatically sensitive and variable lake hydrology over the past 32,000 years. The reconstruction suggests a prolonged shallowing during an early Marine Isotope Stage 2 (MIS 2) mega-drought between 28,000 and 25,000 cal yr BP, a deep lake spanning the last glacial maximum, a wet–dry response to the Younger Dryas, and a highly dynamic MIS 1/Holocene lake. This single-lake elemental geochemistry technique may be useful in similar settings for reconstructing lake depth and inferring past hydrologic changes.