Introduction

Coastal central California is characterized by limited water resources, relying heavily on water sourced elsewhere in California and/or Colorado. The region’s hydrologic budget is dominated by winter precipitation; therefore, it is highly sensitive to changes in the amount of this component (Cayan and Peterson 1989; Schonher and Nicholson 1989). As a result, it is essential to understand prior hydroclimatic changes and their forcings to better assess present and future water variability in the region. For example, will changing climate produce wetter or drier winters, or will the future reflect more extreme weather, in general (Dettinger 2011; Das et al. 2013; Neelin et al. 2013; Pierce et al. 2013; Seager et al. 2013; Griffin and Anchukaitis 2014; Williams et al. 2015)? Whatever the scenario, it is critical to characterize the past to understand the future. This is true even if the modern climatic boundary conditions are unlike those in the recent geologic past (e.g., CO2 concentrations). Periods of past extremes—e.g., multicentennial scale drought and pluvials—provide a glimpse of potential future extremes even in the context of unprecedented anthropogenic forcings (Cook et al. 2004, 2010; Kirby et al. 2012, 2014; Mensing et al. 2013; Heusser et al. 2015). Moreover, the strong role played by the Pacific region over the Holocene in coastal southern CA suggests that future climate change will remain tightly coupled to Pacific conditions (Kirby et al. 2010, 2012, 2014, 2015). Thus, it is equally important to both characterize and explain past hydroclimatic change.

There are several late Holocene paleoclimatic records from coastal central California, defined here as the region north of Santa Barbara/Zaca Lake and south of San Francisco (Fig. 1) (Patch and Jones 1984; Michaelsen et al. 1987; Haston and Michaelsen 1997; Jones and Kennett 1999; Byrne et al. 2001; Malamud-Roam et al. 2006; Plater et al. 2006; Cowart and Byrne 2013; Anderson et al. 2015; Ejarque et al. 2015). Here, a few sites are discussed as relevant to the Abbott Lake record. Malamud-Roam et al. (2006) sediment core research from the San Francisco Bay-Delta system show dry conditions between 1000 and 800 cal years BP and wet conditions between 550 and 200 cal years BP, the latter being contemporaneous to the Little Ice Age (LIA) and the former to the Medieval Climatic Anomaly (MCA). Byrne et al. (2001) sediment core research, also from the San Francisco Estuary, indicates dry conditions between 1700 and 730 cal years BP. Their results suggest dry conditions during the MCA. Sediments from Pinto Lake, located 80 km north of Abbott Lake, suggest two periods reduced precipitation and lower lake levels from 1300 to 1050 and 675 to 200 cal years BP (Plater et al. 2006). Moving south, evidence for changes in water availability during the MCA or LIA from Anderson et al. (2015) coastal site is equivocal. Notably, Anderson et al. site is located only 50 km west of Zaca Lake where a distinct dry MCA and variably wet LIA signal are inferred (Kirby et al. 2014). Finally, sediment core data from Dune pond, 50 km south of Zaca Lake, indicate greater local moisture between 700 and 150 cal years BP (Ejarque et al. 2015). In all, these coastal central California sites suggest geographically variable MCA and LIA signals. It is arguable, however, that the coastal sites are influenced more than inland sites by marine conditions such as coastal fog, which may modulate their response to regional versus local hydroclimatic change (Anderson et al. 2015; Ejarque et al. 2015). Similarly, the San Francisco Bay-Delta-Estuary sites receive a combination of water from distal (e.g., Sierra Nevada Mtns) and proximal regions, possibly obfuscating the predominant hydroclimatic signal as geographically disparate water sources change in relative dominance through time.

Fig. 1
figure 1

a Regional map with relevant comparison sites. b, c Box and whisker plots shown for relevant precipitation sites as discussed in text. Precipitation data from http://www.wrcc.dri.edu

Full size image

The highest-resolution, best dated records for the region are aggregate tree-ring studies (Cook et al. 2004, 2010; MacDonald and Case 2005; Griffin and Anchukaitis 2014). Plotting spatially relevant PDSI grid points 36 and 48 (http://www.ncdc.noaa.gov/paleo/newpdsi.html) reveal significant multi-decadal droughts from 1100–900 to 750–650 cal years BP. A trend towards wetter conditions (or less drought) begins ~500 cal years BP, peaking at 200 cal years BP. However, tree-ring records are spatially and temporally limited older than 1000 cal years BP and do not record the same signal as lakes such as discrete precipitation run-off events (e.g. floods), vegetation changes, source water variability, and integrated fire frequency. For this reason, other paleoclimatic archives, such as lacustrine sediments or cave deposits, are required to extend the record of analysis as well as to examine components of the climate system not preserved (recorded) by the tree ring records. For these reasons, a new record from Abbott Lake was acquired to characterize and explain past hydroclimatic change in coastal central California (Fig. 1).

The objectives of this paper are: (1) to characterize hydroclimatic variability of coastal central California using sediment from Abbott Lake over the past 1400 cal years BP; (2) to explain this hydroclimatic variability in the context of known forcings such as the Pacific Decadal Oscillation and tropical and north Pacific SSTs; and, (3) to evaluate the spatiotemporal variability along coastal central and southern California through comparison to Zaca Lake (Dingemans et al. 2014; Feakins et al. 2014; Kirby et al. 2014), Lake Elsinore (Kirby et al. 2010), and flood event layers in Santa Barbara Basin (Hendy et al. 2013) (Fig. 1).

Study site

Abbott Lake is located within the Santa Lucia Range in Monterey County approximately 20 km east of the Pacific Ocean (Figs. 1, 2). Abbott Lake formed as a meander cutoff and subsequent channel abandonment of the Arroyo Seco River during the transition from a meandering to a braided system (Finnegan and Balco 2013). The modern lake has a maximum length of approximately 0.46 km and a maximum width of 0.08 km. The lake’s western half is dominated by Mesozoic granitic rocks whereas the eastern half (the largest portion of the lake’s drainage basin) is predominantly friable Tertiary sandstone (Dibblee and Minch 2006; Fig. 2). The drainage basin (0.75 km2) is steep and evidence for mass movement is common. The primary source of water for the lake is direct run-off from its drainage basin. Notably, there are no obvious active tributaries into the lake; although, there is abundant evidence for past channels in the drainage basin that presently do not connect to the lake. An incised channel at the lake’s western edge suggests Abbott Lake overflows into a down-valley dry lake during periods of wetter climate.

Fig. 2
figure 2

a Abbott Lake site map with local geology and topography [Mesozoic granitic rocks (left side of figure) and Tertiary sandstone (right side of figure) (Dibblee and Minch 2006)]. b Image looking towards site 6 (as shown on a). The prominent tree line is approximately 4–5 m higher than lake level (AD 2014)

Full size image

Abbott Lake is relatively shallow with a maximum depth of 2–2.5 m in the central basin as of April 2014. The presence of precipitated calcium carbonate on rock surfaces (tufa), a prominent tree line on the southern side of the lake, and aquatic gastropods in raised lacustrine sediments on the north side of the lake indicates Abbott Lake was, at times, much deeper. Several measurements were taken by hand along the south side of the lake from the tree line to the modern lake surface in April 2014 to estimate the maximum potential depth (Fig. 2). In the central basin, the tree line is approximately 4–5 m above the modern lake surface therefore equaling a maximum modern depth of 7–7.5 m before spillover (Fig. 2). On the eastern end of the lake where the sediment cores were extracted for this study, the tree line extended only 2–3 m above the modern sediment surface as expected based on the lake’s modern geometry (Fig. 2). Based on these measurements, the core site for this study is located 2–3 m below water level during modern lake level highstands, or overflow. 2400 years ago, however, the same core site would have been located between 7.5 and 8.5 m (2–3 m plus sediment core length [5.5 m]) below water depth. Consequently, our core site has experienced a long term shoaling due to volumetric loss associated with rapid lake sedimentation (0.2–0.4 cm year−1). The point of this modern- and paleo-depth analysis is to highlight that although shallow today, the lake is occasionally much deeper and sustained for long enough periods of time to precipitate and accumulate tufa, deposit and preserve raised lacustrine sediments, and develop a well-defined tree-line (Fig. 2). Consequently, the core site used for this study—although a wetland environment today—was at times under considerable water depth during past pluvials.

Modern climate setting

The modern climate of coastal central California is characterized as Mediterranean with cool, wet winters and hot, dry summers. A cross-longitude comparison of total winter precipitation values (October–March) from Big Sur to Paloma to King City reveals a prominent rain shadow effect (Fig. 1). Abbott Lake sits in the coastal Santa Lucia Range (~30 km east of Big Sur) where orographically enhanced precipitation should impact the lake’s hydrology. Furthermore, a cross-latitude comparison of Big Sur, Lompoc, Oxnard, and Culver City (north to south) illustrates a strong latitudinal gradient in winter precipitation (Fig. 1). That is, annual precipitation is highest in coastal central California and decreases to the south. This latitudinal and longitudinal comparison highlights Abbott Lake’s important location to fill a geographical gap in understanding the spatiotemporal patterns of past hydroclimatic variablity.

The amount of winter precipitation coastal central California receives is determined by the position of the Pacific subtropical high. For example, wet winters are attributed generally to a weakened subtropical high, which allows storm tracks to shift southward more frequently (Cayan and Peterson 1989; Schonher and Nicholson 1989).

The position of the subtropical high is influenced predominantly by Pacific Ocean–atmosphere dynamics, specifically El Niño-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (Schonher and Nicholson 1989; Castello and Shelton 2004; Wise 2010, 2016). Importantly, the Pacific affects climate in coastal southern California on interannual to millennial timescales, as recognized in both annual-to-interannual resolution tree ring studies and centennial-to-millennial resolution lake studies (Cook et al. 2004, 2010; MacDonald and Case 2005; Cook et al. 2011; Seager et al. Kirby et al. 2010, 2012, 2014, 2015).

The interplay between ENSO and the PDO affects significantly the frequency and position of winter storms tracking across the study region, which in turn modulate the amount of precipitation Abbott Lake receives. Given the current position of Abbott Lake south of the modern western US precipitation dipole (Cayan et al. 1998; Dettinger et al. 1998; Wise 2010, 2016), a positive PDO phase in combination with El Niño conditions often generates a significant increase in winter precipitation due to a southerly displaced winter storm track (Cayan and Peterson 1989; Mantua and Hare 2002; MacDonald and Case 2005; Sung et al. 2014). However, the study region may experience a reverse response to ENSO and PDO if the western US precipitation dipole shifts southward of Abbott Lake (Cayan et al. 1998; Dettinger et al. 1998; Wise 2010, 2016; Sung et al. 2014). In addition, this region’s winter climatology is influenced by atmospheric river (AR) storms, which are responsible for transporting up to 90 % of all Pacific-sourced tropical/sub-tropical moisture to higher latitudes in the western United States (Ralph et al. 2006; Dettinger 2011; Dettinger et al. 2011). In particular, atmospheric rivers generate copious precipitation along the west coast of California. Orographic uplift enhances moisture loss particularly in the mountain ranges along California’s coast such as the Santa Lucia Mountains, in which Abbott Lake is located. In fact, atmospheric rivers account for nearly all of the historically large storms and their associated floods within the region (Dettinger 2011; Dettinger et al. 2011; Ralph and Dettinger 2011). Notably, the run-off generated by atmospheric river storms is often well above average and mobilizes otherwise intransigent sediment (Gray et al. 2015a, b). Abbott Lake’s steep drainage basin coupled with its location within the coastal Santa Lucia Range makes the study site especially sensitive to changes in the magnitude and/or frequency of atmospheric river storms. In coastal central California where Abbott Lake resides, ENSO and PDO are shown to modulate the contribution from atmospheric rivers, such that more precipitation is produced from atmospheric rivers during a positive PDO phase and during El Niño conditions (Dettinger et al. 2011). Given age control-resolution limitations and the varied climatic sensitivity of the analyses, it is unlikely that our Abbott Lake core captures distinct AR events; rather, it is more reasonable that Abbott Lake records general changes in wetness or dryness that reflect changes in the relative contribution of ARs.

Methods

Two sediment cores (AL13PC-1 and AL13PC-2) were collected from the lake’s eastern littoral zone using an Eijkelkamp lined pound core capable of extracting 33 cm drives (Fig. 2). The coring site was determined based on six reconnaissance soil gouge cores extracted down the axis of the lake’s eastern and western now-exposed surfaces (Fig. 2). Core AL13PC-1 (the focus of this paper) is comprised of fifteen sediment drives extending 495 cm below the surface with an 87 % recovery rate. Core AL13PC-2 was taken from the same location to extend the sediment record to 623 cm. It is comprised of six sediment drives extending from 425 to 623 cm with a 70 % recovery rate. This paper discusses only the upper 300 cm of sediment, or approximately 1400 cal years, in core AL13PC-1 where sediment recovery and age control is most robust. The cores were split, photographed, visually described, sampled, and stored at California State University, Fullerton.

Six charcoal samples (>125 μm) and one gastropod sample were collected from core AL13PC-1. Two charcoal samples (>125 μm) were collected from core AL13PC-2. In total, nine accelerator mass spectrometry (AMS) 14C dates were measured at the University of California, Irvine Keck-Carbon Cycle AMS Facility. The charcoal samples were pretreated using a standard acid–base–acid (ABA) method. The gastropod shells were leached of 50 % of their mass using HCl to remove secondary carbonate.

Magnetic susceptibility was measured at contiguous 1 cm intervals using a Bartington MS2 magnetic susceptibility meter at 0.465 kHz. Every sample was measured once, rotated 180° on its y-axis, and measured a second time.

Total organic matter (% TOM, 550 °C) and total carbonate (% TC, 950 °C) were measured at contiguous 1 cm intervals using the loss on ignition method (LOI) of Dean (1974). Total carbonate values less than 4–5 % may reflect clay lattice dewatering, which occurs during the 950 °C heating stage (Dean 1974). However, the sediments used for this research are predominantly silty sands, sandy silts, or clayey silts. As a result, we do not consider clay lattice dewatering a significant contributor to the LOI 950 °C final values.

Grain size was measured at every other centimeter using a Malvern Mastersizer 2000 grain size analysis machine coupled to a Hydro 2000G. Each sample was pretreated with ≥30 mL of 30 % H2O2 to remove organic matter, 10 mL of 1 M HCl to remove carbonate, and 10 mL of 1 M NaOH to remove biogenic silica. A tuff standard (TS2: n = 9525, \( {{\bar{\text{x}}}}= 4.61 \pm 0.18\)) was run twice at the beginning of each day, once every ten samples, and once at the end of every day to evaluate the equipment’s analytical stability over time.

Gastropods were sampled at 5-cm intervals and contiguous 1 cm intervals where visually abundant. All samples were cleaned using 74-μm and 125-μm sieves and isolated into different groups based on key visual differences (e.g. shell size and shape, number and direction of coils, aperture shape, etc.). Each gastropod group was photographed using a tabletop scanning electron microscope (SEM) at the University of Southern California and identified before additional observations to ensure subsequent correct identification.

Twenty sample depths were selected for stable oxygen and carbon isotopic analysis on an identified obligate aquatic species (Gyraulus parvus). A minimum of 3 gastropods per sample depth were washed, cleaned and sonicated with DI water and subsequently homogenized for whole shell analysis. The powdered gastropod samples were then analyzed in the Saskatchewan Isotope Laboratory. Stable isotope values are obtained using a Finnigan Kiel-IV carbonate preparation device directly coupled to the dual inlet of a Finnigan MAT 253 isotope ratio mass spectrometer. Twenty to fifty micrograms of carbonate were reacted at 70 °C with 3 drops of anhydrous phosphoric acid for 420 s. The CO2 evolved was then cryogenically purified before being passed to the mass spectrometer for analysis. Isotope ratios are corrected for acid fractionation and 17O contribution using the Craig correction and reported in per mil notation relative to the VPDB scale. Data is directly calibrated against the international standard NBS-19 that is by definition δ13C = 1.95 ‰ VPDB and δ18O = −2.20 ‰ VPDB. Precision/accuracy of data are monitored through routine analysis of NBS-19 and in-house check standards which have been stringently calibrated against NBS-19. Precision/accuracy of δ13C and δ18O are 0.05 and 0.11 ‰, respectively (n = 25).

Results

The nine AMS 14C dates measured from cores AL13PC-1 and AL13PC-2 were converted to calendar years before present using an IntCal13 calibration curve (Reimer et al. 2013) embedded in the Bacon version 2.2 software program (Table 1) (Blaauw and Christen 2011). A single age model was constructed in Bacon using Bayesian statistics (Blaauw and Christen 2011) (Fig. 3). However, only the upper 300 cm or 1400 cal years BP are discussed due to low confidence in the age model older than 1400 cal years BP (Fig. 3).

Table 1 Age control
Full size table
Fig. 3
figure 3

Core AL13PC-1 age model showing the 2-sigma probability distribution for each radiocarbon date (purple) and maximum/minimum calculated at and between each sample depth (gray shading). Colored box highlights the section (depth and age) of interest for this paper. (Color figure online)

Full size image

Core AL13PC-1 is characterized by brown, reddish-brown, and gray layers (Fig. 4). Gastropods are visually abundant throughout most of the core and tend to be fragmented in the sandy units. Magnetic susceptibility values range from 0.2 × 10−7 to 12.9 × 10−7 m3 kg−1 with an average of 1.4 × 10−7 m3 kg−1 (Fig. 4). Percent TOM values range from 1.0 to 33.3 % with an average of 10.1 %; percent TC values range from 0.7 to 50.5 % with an average of 11.6 % (Fig. 4). Core sediments consist predominantly of silty sands, sandy silts, and clayey silts. The sand values range from 0 to 84.9 % with an average of 47.6 %; the silt values range from 12.2 to 77.4 % with an average of 44.0 %. The clay values range from 0.4 to 33.3 % with an average of 8.4 % (Fig. 4). Sand data are standardized according to Kirby et al. (2010) in order to view the data in standard deviations above or below the mean. Four gastropod groups were identified including the Family Physidae, genus Helisoma, species Gyraulus parvus, and species Ferrissia fragilis. This study focuses on the obligate aquatic species, G. parvus, present from 490–390, 295–135, 120–80 to 20–0 cm (Fig. 4). δ18O values of the G. parvus shells range from −2.6 to 3.3 ‰VPDB with an average of 0.9 ‰VPDB (Fig. 5).

Fig. 4
figure 4

Core AL13PC-1 sedimentological results. From left to right: magnetic susceptibility (CHI), LOI 550 °C or percent total organic matter (TOM %), LOI 950 °C or percent total carbonate (TC %), grain size as clay-silt-sand, presence (P)/absence (A) of G. parvus, as well as δ18O(G. parvus) values and δ13C(G. parvus) values in per mil (‰) VPDB. Red arrows = AMS 14C sample depths. Color changes on stratigraphic column reflect subjective sediment color changes. Checkered boxes represent core gaps. (Color figure online)

Full size image
Fig. 5
figure 5

Abbott Lake depositional model as viewed from the southeast end of the lake (Fig. 2). Top diagram shows Deep Lake model (more run-off). Bottom diagram shows Shallow Lake scenario (less run-off). Gray column represents hypothetical coring site

Full size image

Discussion

Sediment analysis interpretations

Grain size

Changes in run-off duration and/or intensity influence strongly the distribution of grain size observed in lake sediment (Håkanson and Jansson 2002). Research on rivers in coastal central and southern California generally show that increased precipitation during intermittent, intense winter storms—such as during atmospheric rivers or exceptionally wet winters—generates higher river discharge and enhanced mobilization of sand (Inman and Jenkins 1999; Farnsworth and Milliman 2003; Gray et al. 2015a, b). This modern observation, in combination with twentieth century grain size/lake level/river discharge relationships (Kirby et al. 2010, 2014), is the basis for using changes in sand content in coastal southern (and now coastal central) CA lake sediments as an indicator of variable runoff and thus wet versus dry paleoclimates. Consistent with existing regional studies on Elsinore and Zaca (Kirby et al. 2010, 2014), changes in sand content in Abbott Lake are interpreted as variations in winter climate wetness and/or runoff intensity (i.e., more sand = wetter winters and/or more intense run-off).

An alternative interpretation associates an increase in sand to lake level regression and subsequent migration of the sandy littoral zone basinward (Heusser et al. 2015; Shuman et al. 2009, 2015). As discussed in our depositional model, the lake’s broad, low angle bathymetry generates large changes in surface area with small changes in lake depth. The result is the migration of dense macrophytes across the littoral zone during lake lowstands. This dense macrophytic zone inhibits the transport of sand while capturing the finer sediment that is more easily commuted into, and through, the vegetation. Consequently, we interpret changes in percent sand to reflect run-off variability (i.e. wetter conditions) rather than lake level regression (i.e. drier conditions).

Percent total organic matter (LOI 550 °C)

Total organic matter in lake sediment is a combined measure of autochthonous and allochthonous contributions. Variations in autochthonous organic matter often reflect changes in primary productivity in the lake and nutrient supply (Mullins 1998; Anderson 2001; Kirby et al. 2007). Variations in allochthonous organic matter may reflect changes in precipitation and the subsequent run-off of organic matter into the lake basin (Meyers and Ishiwatari 1993; Kirby et al. 2012). In Abbott Lake, changes in total organic matter are negatively correlated to percent total sand (r = −0.52, p < 0.0001). If changes in % TOM reflect an increase in run-off derived allochthonous organic matter, then sand and % TOM should be positively correlated because higher sand content is interpreted to reflect more runoff. The observed opposite relationship leads us to conclude that changes in % TOM reflect changes in the depositional environment over time through the expansion or contraction of autochthonous littoral macrophytes. Thus, lower total organic matter is interpreted as a reduction of macrophytes across a narrow littoral zone during highstands and vice versa for higher total organics.

Gastropod oxygen isotopes

Gastropods represent one of the largest classes of Mollusca and can serve as an indicator of environmental conditions (Strong et al. 2008). For this study, we focus on G. parvus because it is found consistently in shallow freshwater environments including the littoral zones of lakes and ponds (Miller 1964; McMullen and Zakrzewski 1972; Rymer et al. 1988; Miller et al. 2000). δ18O values of obligate aquatic gastropods are often interpreted as an indicator of lake water δ18O values (Stuiver 1970; Whitmore et al. 1996; Curtis et al. 1998; Abell and Hoelzmann 2000). As an obligate aquatic species, G. parvus was selected for isotopic analysis to minimize the complicating isotopic effects of environmental variability on species that inhabit both sub-aerial and sub-aqueous environments. Consequently, it is our contention that the δ18O values of the obligate aquatic species G. parvus are a first-order recorder of lake water δ18O, likely via changes in the precipitation:evaporation ratio. For example, higher δ18O(G. parvus) values may reflect greater evaporation and removal of 16O during dry conditions and vice versa for wet conditions. Alternatively, changes in δ18O(G. parvus) values may reflect changes in lake water δ18O as controlled by changing winter precipitation amount and/or sources. For example, dry winters should be characterized by less cold season low δ18O precipitation and its subsequent run-off and vice versa for wet winters. Whatever the predominant control—P:E or amount/source—the net change in the δ18O(lake water) is in the same direction and thus recorded similarly in the δ18O(G. parvus) values. In other words, lower δ18O(G. parvus) values are interpreted to reflect wet winters with more cold season low δ18O (precipitation) run-off and/or less evaporation and vice versa for higher δ18O(G. parvus) values.

Depositional model

A depositional model was developed to account for the drainage basin’s bedrock, its topography, the lake’s broad littoral zone along the eastern side (where the core was extracted), and the lake’s assumed response to winter precipitation run-off (Figs. 2, 5). Abbott Lake’s hydrology responds directly to winter precipitation variability. Consequently, wet winters generate higher-than-average precipitation run-off and mobilize more sand into the lake. The increase in total sand is amplified by Abbott Lake’s steep drainage basin coupled with the presence of friable sandstone bedrock. This sandstone bedrock is the most common surficial rock type in the drainage basin (Fig. 2). During dry winters, however, precipitation and run-off in the drainage basin is reduced. The decrease in transport energy reduces the movement of sand resulting in a lower proportion of sand in the lake sediment. As the lake’s surface elevation drops in response to reduced winter precipitation, the littoral zone expands across the broad and low angle eastern end of the lake, as was observed during the 2011–2015 AD drought (Fig. 2). Subsequently, macrophytes also expand across the littoral zone—also observed during the recent drought—both inhibiting the transport of sand and capturing the finer sediment that is more easily commuted into, and through, the dense macrophyte-rich littoral zone. Concurrently, the obligate gastropod G. parvus precipitates a shell with higher oxygen isotope values that reflect a more evaporative hydrologic regime or reduced cold season low δ18O (precipitation) runoff. In summary, dry winters are reflected by low sand content, high total organic matter, and high δ18O(G. parvus) values and vice versa for wet winters.

1400 years of hydroclimatic variability at Abbott Lake

Using changes in sand content, percent total organic matter, and δ18O(G. parvus) values, a 1400 year winter precipitation/run-off history is inferred for Abbott Lake, divided into wet (W), dry (D), and transitional (T) intervals (Fig. 6). Sand is used as the primary guide for the intervals above. We focus our discussion on the Medieval Climatic Anomaly (MCA: 1050–650 cal years BP) and the Little Ice Age (LIA: 500–100 cal years BP) using the Masson-Delmotte et al. (2013) age ranges (Fig. 6).

Fig. 6
figure 6

Core AL13PC-1 sediment analysis results showing percent total organic matter, standardized percent sand, and δ18O(G. parvus) values. Sediment analysis interpretations are shown along y-axes. LIA Little Ice Age (Masson-Delmotte et al. 2013), MCA Medieval Climatic Anomaly (Masson-Delmotte et al. 2013), W wet, D dry, T transitional climate

Full size image

At first glance, it is apparent that the MCA and LIA are different from one another in terms of run-off (i.e., precipitation) variability (Fig. 6). The MCA is generally dry, characterized by less run-off (low sand), higher total organic matter (expanded littoral zone), and higher δ18O(G. parvus) values [greater lake water evaporation and/or less cold season low δ18O (precipitation)]. Conversely, the LIA is characterized by more run-off (high sand), lower total organic matter (contracted littoral zone), and lower δ18O(G. parvus) values [less lake water evaporation and/or more cold season low δ18O (precipitation)]. Peak MCA conditions occur from 920 to 680 cal years BP during which run-off is reduced for ~240 years. The transition into the LIA begins ~680 cal years BP with a small increase in sand. This transition, however, is interrupted by a ~80 year return to dry winters between 590 and 510 cal years BP. Wetter conditions return at ~510 cal years BP. Peak LIA conditions occur between 510–380 and 240–175 cal years BP with two notable excursions to drier conditions from 380–300 to 175–130 cal years BP. In all, the timing of the MCA and LIA as defined by the Abbott Lake sediment analyses are consistent with the Masson-Delmotte et al. (2013) age ranges. Rather than serendipitous correlations, we contend that Abbott Lake sediments capture both the MCA and LIA as major hydroclimatic periods over the past 1400 years in coastal central CA.

Regional site comparisons

A comparison moving from south to north between Lake Elsinore (Kirby et al. 2010), Zaca Lake (Dingemans et al. 2014; Feakins et al. 2014; Kirby et al. 2014), and Abbott Lake shows that all three sites record a variably wet LIA and a less variable, generally dry MCA (Figs. 1, 7). All three records indicate also a gradational transition out of the MCA and into the LIA beginning ca. 600–700 cal years BP. Inferred peak LIA wet conditions are temporally congruent between the southern sites—Zaca and Elsinore (ca. 350–250 cal years BP)—but incongruent with Abbott Lake.

Fig. 7
figure 7

Regional site comparison. a Abbott Lake (this study) standardized percent sand, b Standardized Zaca Lake percent sand >125–2000 µm (Kirby et al. 2014), c standardized Lake Elsinore percent sand (Kirby et al. 2010). Small blue arrows at top represent Santa Barbara Basin (SBB) flood layers (FL). LIA Little Ice Age (Masson-Delmotte et al. 2013), MCA Medieval Climatic Anomaly (Masson-Delmotte et al. 2013). (Color figure online)

Full size image

Flood events from Santa Barbara Basin are compared to the sites above to serve as a more regional integration of past hydroclimates (Fig. 7) (Schimmelmann et al. 1998; Hendy et al. 2013). Notably, floods are virtually absent during the MCA and certainly during peak MCA conditions at all three sites. During the LIA, however, flood event layers are present suggesting more intense and/or more frequent runoff events. Regardless of various sites’ absolute timing, it is apparent from this comparison of the three lake and one marine sites that the LIA in coastal southern and central CA is wet with occasional large runoff events (in some cases imprinting on the marine environment) (Fig. 7). Conversely, the MCA is dry with evidence for reduced run-off both in the terrestrial and marine environments.

There are two additional lake sites from coastal southern CA—Mod Pond and Dune Pond—both located along the Pacific coast, west of Zaca Lake (Anderson et al. 2015; Ejarque et al. 2015) (Fig. 1). Both sites were analyzed for pollen and charcoal to reconstruct past vegetation change and fire frequency (Anderson et al. 2015; Ejarque et al. 2015). However, evidence for the MCA or the LIA is equivocal at Mod Pond, especially in the context of Holocene human occupation and disturbance of the landscape. Sediment core data from Dune pond, however, indicate greater local moisture between 700 and 150 cal years BP, contemporaneous to long term trends in increasing wetness at Elsinore, Zaca, and Abbott (Ejarque et al. 2015).

Forcing analysis

Finally, we examine the Abbott Lake record in the context of Pacific forcings using a tree-ring based Pacific Decadal Oscillation reconstruction (MacDonald and Case 2005), a N Pacific PDO region SST reconstruction (Mann et al. 2009), a foraminiferal-based sea surface temperature record from Santa Barbara Basin (Kennett et al. 2007), and a tropical Pacific zonal SST gradient reconstruction (Rustic et al. 2015) (Fig. 8). The rationale for this Pacific focused comparison is based on geologic (and modern) observations from coastal central and southern CA, including paleo-sites [e.g., Lake Elsinore, Zaca Lake, Silver Lake (Kirby et al. 2015)], where a strong relationship between Pacific sea surface temperatures and/or combined Pacific ocean–atmosphere conditions (e.g., ENSO, PDO) is observed (Namias et al. 1988; Hanson et al. 2006; Kirby et al. 2010, 2014; Wise 2010, 2016; Cook et al. 2011).

Fig. 8
figure 8

Forcing comparison. a Abbott Lake (this study) standardized percent sand, b Pacific Decadal Oscillation (PDO) index reconstruction (MacDonald and Case 2005), c North Pacific PDO region SST averaged over the central North Pacific region (22.5°N–57.5°N, 152.5°E–132.5°W) (Mann et al. 2009), d tropical Pacific E–W Zonal SST gradient (Rustic et al. 2015), e Santa Barbara Basin SSTs (Kennett et al. 2007). LIA Little Ice Age (Masson-Delmotte et al. 2013), MCA Medieval Climatic Anomaly (Masson-Delmotte et al. 2013), W wet, D dry, T transitional climate

Full size image

For the tree-ring based PDO reconstruction, we use the MacDonald and Case (2005) PDO reconstruction over other PDO reconstructions (Biondi et al. 2001; D’Arrigo et al. 2001). The MacDonald and Case (2005) reconstruction is preferred for this study because the incorporated tree-ring-data sites lie both north and south of Abbott Lake thus capturing the latitudinal breadth of the PDO signal. We also use the Mann et al. (2009) N Pacific PDO region SST reconstruction because it captures oceanic conditions from the north central Pacific region where the PDO’s SST signature is strongly developed. The comparison between the PDO reconstructions and the Abbott Lake sand data is intriguing (Fig. 8). In general, intervals of high PDO index values and cooler N Pacific PDO region SSTs correspond to intervals of more sand and vice versa. For example, the MCA is generally characterized by negative PDO index values, warmer N Pacific PDO region SSTs, and less sand (vice versa for the LIA). This relationship makes sense in the context of the modern PDO and its associated winter climatology (Hanson et al. 2006). Today, higher PDO indices and cooler N Pacific PDO region SSTs are associated with more frequent winter storm tracks across coastal southern and central CA. This relationship was also observed over most of the past 1000 years via a comparison between Lake Elsinore sand content and the MacDonald and Case (2005) PDO reconstruction (Kirby et al. 2010). The Elsinore record, however, is out-of-phase with the PDO index during peak LIA conditions (max. sand centered on 300 cal years BP at Elsinore) (Kirby et al. 2010) (Fig. 6). Interestingly, at the same time at Abbott Lake (minimum sand centered on 300 cal years BP) low PDO index values correspond to lower sand and thus less inferred runoff. This disparity between the two sites (Elsinore and Abbott) during the LIA may indicate latitudinal changes in the mean position of the winter storm track during the LIA.

We also compare the Abbott Lake sand record to a foraminiferal-based SST record from Santa Barbara Basin (Kennett et al. 2007) (Fig. 8). Today, warmer SSTs along coastal southern California occur in response to ocean–atmosphere conditions (e.g. El Niño) associated with enhanced winter precipitation (Thunell et al. 1999; Barron et al. 2015). Our comparison reveals generally above average SSTs during the wet LIA and below average SSTs during the dry MCA.

Finally, we compare the Abbott Lake record to a tropical Pacific zonal SST gradient reconstruction (Rustic et al. 2015). The rationale for this comparison is the strong modern relationship between tropical Pacific SSTs and winter climatology for our study region (Schonher and Nicholson 1989; Dettinger et al. 1998; Castello and Shelton 2004). In general, El Niño conditions are associated with enhanced winter precipitation in coastal central and southern California. In the paleo-record, a reduced E–W tropical Pacific SST gradient is interpreted to reflect El Niño or El Niño-like conditions (Rustic et al. 2015). Comparing Rustic et al. (2015) to the Abbott Lake sand record shows a reduction in the E–W tropical Pacific SST gradient during the LIA, conditions that should enhance winter precipitation in the study region. The relationship is less apparent during the MCA.

In summary, our comparison between Pacific-sourced forcings and the Abbott Lake run-off record yields confidence that late Holocene hydroclimatic variability in coastal central California is linked to Pacific conditions. This finding is consistent with that previously observed at Lake Elsinore, Zaca Lake, and Silver Lake (Kirby et al. 2010, 2014, 2015).

Conclusions

Here, we present a new 1400 year sediment record from Abbott Lake in coastal central California which includes the MCA and LIA. The results indicate that coastal central California was relatively dry (less runoff) during the MCA and variably wet (more runoff) during the LIA. These results are compared to Zaca Lake and Lake Elsinore, both located south of Abbott Lake, and to flood event layers from Santa Barbara Basin (SBB) (Hendy et al. 2013). All three sites, including SBB’s flood event history, indicate a variably wet LIA and dry MCA; however, the absolute timing of wet and dry episodes is more congruent between Zaca and Elsinore than to Abbott Lake suggesting spatiotemporal variability along the southern to central California coast. Finally, a comparison to various Pacific-sourced forcings links late Holocene hydroclimates along coastal central California to tropical and N Pacific conditions. Evidence for this connection at various timescales from coastal southern California sites—15000 years at Silver Lake (Kirby et al. 2015), 3000 years at Zaca Lake (Kirby et al. 2014), and 1400 years at Abbott Lake (this paper)—give reason to suspect that future hydroclimatic change in coastal central and southern California will remain coupled to changes in the Pacific.