The late Pleistocene was a climatically dynamic period, with abrupt shifts between cool-wet and warm-dry conditions. Increased effective precipitation supported large pluvial lakes and long-lived spring ecosystems in valleys and basins throughout the western and southwestern U.S., but the source and seasonality of the increased precipitation are debated. Increases in the proportions of C₄/(C₄+ C₃) grasses in the diets of large grazers have been ascribed both to increases in summer precipitation and lower atmospheric CO₂ levels. Here we present stable carbon and oxygen isotope data from tooth enamel of late Pleistocene herbivores recovered from paleowetland deposits at Tule Spring Fossil Beds National Monument in the Las Vegas Valley of southern Nevada, as well as modern herbivores from the surrounding area. We use these data to investigate whether winter or summer precipitation was responsible for driving the relatively wet hydroclimate conditions that prevailed in the region during the late Pleistocene. We also evaluate whether late Pleistocene grass C₄/(C₄+ C₃) was higher than today, and potential drivers of any changes.

Tooth enamel δ¹⁸O values for Pleistocene Equus, Bison, and Mammuthus are generally low (average 22.0 ± 0.7‰, 2 s.e., VSMOW) compared to modern equids (27.8 ± 1.5‰), and imply lower water δ¹⁸O values (−16.1 ± 0.8‰) than modern precipitation (−10.5‰) or in waters present in active springs and wells in the Las Vegas Valley (−12.9‰), an area dominated by winter precipitation. In contrast, tooth enamel of Camelops (a browser) generally yielded higher δ¹⁸O values (23.9 ± 1.1‰), possibly suggesting drought tolerance. Mean δ¹³C values for the Pleistocene grazers (−6.6 ± 0.7‰, 2 s.e., VPDB) are considerably higher than for modern equids (−9.6 ± 0.4‰) and indicate more consumption of C₄ grass (17 ± 5%) than today (4 ± 4%). However, calculated C₄ grass consumption in the late Pleistocene is strikingly lower than the proportion of C₄ grass taxa currently present in the valley (55–60%). δ¹³C values in Camelops tooth enamel (−7.7 ± 1.0‰) are interpreted as reflecting moderate consumption (14 ± 8%) of Atriplex (saltbush), a C₄ shrub that flourishes in regions with hot, dry summers.

Lower water δ¹⁸O values, lower abundance of C₄ grasses, and the inferred presence of Atriplex are all consistent with general circulation models for the late Pleistocene that show enhanced delivery of winter precipitation, sourced from the north Pacific, into the interior western U.S. but do not support alternative models that infer enhanced delivery of summer precipitation, sourced from the tropics. In addition, we hypothesize that dietary competition among the diverse and abundant Pleistocene fauna may have driven the grazers analyzed here to feed preferentially on C₄ grasses. Dietary partitioning, especially when combined with decreased p_CO2 levels during the late Pleistocene, can explain the relatively high δ¹³C values observed in late Pleistocene grazers in the Las Vegas Valley and elsewhere in the southwestern U.S. without requiring additional summer precipitation. Pleistocene hydroclimate parameters derived from dietary and floral records may need to be reevaluated in the context of the potential effects of dietary preferences and lower p_CO2 levels on the stability of C₃ vs. C₄ plants.