Linked fire activity and climate whiplash in California during the early Holocene

The study investigates how hydroclimate variability (“climate whiplash”) relates to wildfire activity and vegetation change in California during the early Holocene, focusing on the 8.2 kyr event. Observations and models indicate California’s rainy season is shortening, with declining autumn precipitation, enhanced moisture stress, increased flammability, and projected increases in precipitation volatility. Such oscillations between wet and dry conditions can promote fine-fuel growth in wet winters and elevate fire risk in dry autumns. Speleothems provide multi-proxy isotopic and geochemical records sensitive to precipitation amount, source, and seasonality and have documented dynamic California hydroclimate over the last several millennia. Prior work at White Moon Cave (WMC) using δ13C, δ44Ca, and P/Ca in stalagmite WMC1 indicated enhanced precipitation volatility spanning the 8.2 kyr event and a precursor event near 8.3 ka. This study aims to test whether wildfire signals are captured in speleothems and how they relate to hydroclimate variability and vegetation shifts in a pre-anthropogenic context.

The paper situates its work within evidence that California’s hydroclimate is sharpening toward a shorter winter rainy season with increasing interannual precipitation variability, both linked to anthropogenic warming. Tree-ring and lake sediment records in western North America document associations between fire activity, summer temperature, and hydrological drought, with enhanced fire during periods of high interannual variability. Previous stalagmite studies in California have recorded dynamic hydroclimate and identified the 8.2 kyr event and a precursor event as periods of increased volatility. Levoglucosan has been established as a tracer of biomass burning in aerosols and ice cores, while lignin oxidation products (LOPs) are used to infer vegetation types (angiosperm vs gymnosperm; woody vs non-woody). However, applications of levoglucosan and its isomers in speleothems are rare, motivating development of new proxies in cave archives to reconstruct paleo-fire and vegetation and link them to hydroclimate variability.

Study site: White Moon Cave (WMC), Santa Cruz Mountains, California (N37°00’, W122°11’) in a warm-summer Mediterranean climate; overlying vegetation is mixed evergreen forest (coastal Redwood, Douglas Fir, Coyote Bush, grasses). The site burned during the 2020 CZU Lightning Complex Fire. Paleo samples: Analyzed a section of stalagmite WMC1 spanning ~8.6–6.9 ka BP, bracketing the 8.2 kyr event. Levoglucosan (biomass burning marker) and lignin oxidation products (LOPs; vanillyl [V], syringyl [S], cinnamyl [C] groups) were measured on 20 speleothem samples to reconstruct fire activity and vegetation composition. LOP ratios (S/V and C/V) were used to distinguish angiosperm vs gymnosperm and woody vs non-woody inputs. Modern cave monitoring: Dripwater and glass plates (3 years of calcite growth) were collected at multiple drip sites (WMC1–WMC6) in 2019, 2020, and 2021 to evaluate transport and incorporation of levoglucosan and LOPs into calcite and to contextualize paleo signals. Sample preparation: Speleothem segments (0.5–0.9 cm; 0.5–1.3 g) were cleaned (DCM/MeOH sonication, Milli-Q rinses), acid-rinsed (0.6% HCl), dried (150 °C), pulverized, and stored at −20 °C. Levoglucosan extraction used MeOH with ultrasound-assisted extraction and filtration; concentrates were prepared in ACN/H2O and filtered (0.2 μm). LOPs were obtained by oxidative digestion: dissolution in HCl, SPE using HLB cartridges, CuSO4 oxidation with L-ascorbic acid under N2 in microwave digestion vessels, re-acidification, SPE cleanup, and elution with ACN/NH4OH, followed by concentration and filtration. Instrumental analysis: Levoglucosan and LOPs were quantified using LC-HRMS (Dionex Ultimate 3000 coupled to Q Exactive Orbitrap with heated ESI, negative mode). Separation employed phenyl-type and PFP columns under H2O/ACN gradients with ammonium acetate; targeted MS2 (HCD, 35% NCE) complemented full scans. Dripwater solid-phase extractions were processed similarly. Chronology and statistics: U–Th dates from WMC1 were used to construct age models for sample segments (modeled upper and lower age bounds) using R (Stalagmite aging function). Changepoint analysis on the continuous δ18O time series (PELT algorithm with MBIC penalty) identified shifts in mean and variance to characterize hydroclimate volatility. Additional proxies (δ13C, δ44Ca, P/Ca) and 87Sr/86Sr were assessed; Sr isotopes tested for potential changes in seepage routing that could artifactually alter organic inputs. Data archiving: Levoglucosan and LOP data archived at NOAA NCEI (study/37018); previously published proxies at study/32012 and /22270.

- Elevated fire activity during the 8.2 kyr event: Speleothem levoglucosan concentrations were elevated between 8217 ± 23 and 7847 ± 20 years BP, peaking between 8167 ± 20 and 8019 ± 15 years BP, indicating increased local fire activity relative to 8.5–6.9 ka BP background levels. - Vegetation shifts: LOP ratios (S/V vs C/V) separate into three zones, showing a shift from non-woody to woody vegetation between 8300–8200 BP, followed by a return toward more non-woody vegetation between 8000–7000 BP. The woody interval (8217 ± 23 to 7793 ± 25 BP) overlaps the levoglucosan peak period, linking more woody vegetation with heightened fire activity. - Hydroclimate volatility: Carbon and calcium isotope proxies and changepoint analysis on δ18O indicate increased interannual precipitation variability during the precursor (~8.3 ka) and 8.2 kyr events. Estimated rainfall oscillations during the 8.2 kyr event were comparable to the instrumental record; during the precursor they exceeded recent multi-year drought and post-drought variability. P/Ca also increased near the onset of elevated levoglucosan and vegetation change. - Transport pathway control unlikely: No shift in stalagmite 87Sr/86Sr occurred during the levoglucosan peak, arguing against sustained changes in seepage routing as the cause of organic variability; thus, levoglucosan and LOP changes likely reflect surface fire and vegetation signals. - Modern cave context: Dripwater levoglucosan (ng/mL) across sites: WMC1=0; WMC2=5; WMC3=10; WMC4=15; WMC5=20; WMC6=25. Modern calcite levoglucosan concentrations (3.6 ± 0.3 and 6.6 ± 0.3 ng g−1) exceed all early Holocene speleothem samples; younger WMC1 cores (<6900 BP to ~240 BP, imprecisely dated) show up to 4.7 ± 0.3 ng g−1, suggesting increased fire inputs in the mid-to-late Holocene, consistent with regional records. Modern calcite LOP concentrations are higher than early Holocene values and C/V and S/V ratios match the current mixed evergreen forest and resemble the woody interval in the speleothem. - Duration: Peak levoglucosan spans four consecutive samples, suggesting elevated fire activity persisted for ~150 years during the 8.2 kyr event. - Regional coherence: Early–mid Holocene charcoal and pollen records from the Pacific Northwest and coastal California indicate increased fire activity and shifts toward more woody vegetation broadly consistent with WMC1 findings.

The results demonstrate a coupling among hydroclimate variability (climate whiplash), vegetation shifts, and wildfire activity during the early Holocene in coastal California. Tree-ring and paleorecords elsewhere in western North America show that fire activity responds to high interannual variability in moisture and higher temperatures. At WMC1, winter-dominated precipitation proxies captured enhanced interannual variability during the 8.3 ka precursor and 8.2 kyr event, while levoglucosan reflects summer fire activity from the preceding dry season. Enhanced seasonality and higher early Holocene summer insolation likely raised summer temperatures. Despite small mean precipitation changes modeled for western North America during the 8.2 kyr event, the stalagmite proxies indicate heightened interannual volatility. The timing of elevated levoglucosan and a shift toward woody vegetation immediately following the precursor event suggests that abrupt hydroclimatic variability promoted vegetation change that in turn supported increased fire activity. Elevated P/Ca contemporaneous with levoglucosan increases supports enhanced surface/soil inputs, potentially linked to fire. These findings parallel modern relationships between PDSI variability, temperature, vegetation, and fire, implying that projected increases in precipitation volatility with warming may similarly intensify fire regimes by altering fuels and seasonal moisture patterns.

This study provides the first detailed application of levoglucosan and lignin oxidation products in a speleothem to reconstruct paleo-fire activity and vegetation composition, demonstrating that during the early Holocene 8.2 kyr event in coastal California, elevated fire activity coincided with a shift to more woody vegetation and heightened hydroclimate volatility. The combined multi-proxy speleothem record indicates tight coupling between climate whiplash and fire. Methodologically, quantifying levoglucosan and LOPs in small speleothem samples and dripwater opens new avenues for high-resolution paleo-fire and paleo-vegetation reconstructions beyond tree-ring timescales. Future work should prioritize cave systems with well-constrained modern and historical fire records and complementary archives (e.g., tree-ring fire scars, lake sediments with charcoal/pollen/PAHs), and develop joint levoglucosan–PAH analyses to refine interpretations of fire sources and transport. Fast-growing temperate-region speleothems offer strong potential to elucidate drivers of precipitation volatility and climate–vegetation–fire linkages under different climate states.

- Temporal resolution: The speleothem record cannot resolve interannual fire variability; peak levoglucosan spans ~150 years, limiting inference on year-to-year dynamics. - Chronological uncertainty: Portions of the stalagmite younger than 6900 BP lack precise ages (estimated to ~240 BP), constraining interpretation of trends toward the present. - Proxy transport and preservation: The longevity and transport pathways of levoglucosan and LOPs in soils and the epikarst, and their embedding in calcite, require further monitoring to assess signal fidelity and potential diagenetic or hydrologic biases. - Site specificity: Findings reflect conditions at a single cave site; broader spatial generalization requires additional records. - Analytical caveats: Although 14C-levoglucosan spike recoveries support data quality and Sr isotopes rule out routing changes, potential unrecognized factors (e.g., episodic inputs, heterogeneity in drip contributions) cannot be completely excluded.