Dating record
Based on the BACON dating calculations, the Cave Springs Cave guano core spanned the last ~12,000 cal yr BP and constitutes the oldest guano core environmental record from the southeastern United States. It is important to address only three radiocarbon dates were used to construct the model, however other studies reconstructing paleoclimate utilizing bat guano have also used three or less radiocarbon dates (Bird et al. 2007; Wurster et al. 2010; Onac et al. 2014; Cleary et al. 2016). The comparison of radiocarbon dates between guano core studies has been described in Tsalickis et al. (Accepted Manuscript). The calibrated radiocarbon ages contain multiple paleoclimatic events in this region. The Hypsithermal event occurred 5000-9000 cal yr BP and has been characterized by a warmer environment (Delcourt 1980; Delcourt and Delcourt 1980; Driese et al. 2008; Tanner et al. 2015). The HCO occurred at the same time (5000-9000 cal yr BP) and is characterized by increases in Quercus spp. transition from Pinus spp. and a moisture shift from dry to wet conditions, inferred from pollen (Frey 1953; Watts 1970; Delcourt and Delcourt 1980; Watts and Hansen 1994). Unlike other sources of paleo data (e.g., lake sediment cores, ice cores) guano cores contain few disturbances related to climate, bioturbation, or redox alteration (Onac et al. 2014; Cleary et al. 2016). The cavern where the guano pile occurred did not show signs of historic flooding or water disturbance as noted in other areas of the cave. As a result, the bat colonies are assumed to have maintained a consistent migratory pattern with summer roosting and did not interact with the guano piles below the roosting areas.
Elemental changes
Calcium has the highest concentrations ( >80 mg/g) in the early middle Holocene to lower Holocene while Phosphorus has higher concentrations ( > 40 mg/g) in the lower Holocene/upper Pleistocene. Bird et al. 2007 also documented an increase in phosphorus concentrations in the older portion of his bat guano core, calcium however, was not measured. Phosphate minerals are common in caves containing bat guano and form due to reactions between guano and clay minerals in acidic conditions (Giurgiu and Tămaş 2013). This could explain why phosphorus and calcium concentrations are highest in the oldest portion of the Cave Springs Cave core – more time allows for phosphate minerals such as brushite, hydroxylapatite, or variscite to form through digestion, dissolution, double replacement, and redox reactions (Onac and Forti 2011; Giurgiu and Tămaş 2013). Additionally, there are few studies that have examined minerals from bat guano deposits in caves (Fiore and Laviano 1991; Dumitraş et al. 2002; Dumitraş et al. 2004; Onac et al. 2004; Marincea et al. 2006; Frost and Palmer 2011).
Deuterium evidence
Hydrogen (H) isotopes can undergo fractionation through evaporation, condensation, and precipitation as well as through trophic interactions (Schimmelmann and DeNiro 1986; Gröcke et al. 2006; Wurster et al. 2008; Wurster et al. 2010; Peters et al. 2012). According to Peters et al. (2012), evaporation favors light isotopes therefore, the water in the bodies of animals, and in this case bats, tends to be deuterium enriched. This causes trophic enrichment of H2 as the H incorporated into animal tissues is from an enriched source. When considering this mechanism in insectivorous bat guano, there is a three-tier or four-tier trophic system possible (plant tissue, insect chitin, bat guano) or (plant tissue, plant-eating insect, carnivorous insect, bat guano) allowing H2 to become enriched by the time it is deposited in bat guano (Gröcke et al. 2006; Onac et al. 2014; Cleary et al. 2016). H2/H1 ratios have frequently been used as measures of paleoclimate (Birchall et al. 2005) despite a lack of understanding trophic discrimination of H isotopes between resources and consumer tissues (Peters et al. 2012). Many paleoclimate studies utilizing bat guano have also not incorporated H2 isotope fractionation through trophic levels into their interpretation (Bird et al. 2007; Onac et al. 2014; Forray et al. 2015; Choa et al. 2016; Cleary et al. 2016). Wurster et al. (2010) is the first guano study that has applied effects of H2 isotopic fractionation in their guano cores via labile hydrogen exchange with atmospheric H2O and reported H2 values from their study fall between -188‰ to -143‰.
Hydrogen isotopic fractionation is a process that was not considered in this study and is also not often considered in other bat guano studies reconstructing paleoclimate (Bird et al. 2007; Onac et al. 2014; Forray et al. 2015; Choa et al. 2016; Cleary et al. 2016). One reason δ2H is not often considered in bat guano studies is due to its high vulnerability to error. δ2H profiles can be highly influenced by degradation or contamination, or environmental effects (Wurster et al. 2010). It is accepted that bat guano δ2H reflects shifts in moisture source or precipitation amount (Wurster et al. 2008; Wurster et al. 2010), but interpretation is lacking. According to the interpretation of Wurster et al. 2010, lighter δ2H values reflect higher precipitation while heavier δ2H values indicate more aridity. Using the δ2H values from Cave Springs Cave and based on our statistical analysis, we are not interpreting between wet versus dry periods, but instead indicating that there are three distinct shifts in moisture for the region: 1) the upper Holocene (0 – 4,000 cal yr BP), 2) the middle Holocene (4,000 – 10,000 cal yr BP), and 3) the lower Holocene/upper Pleistocene (11,000 – 12,000 cal yr BP).
δ2H and pollen comparisons
Pollen records from locations in northwest Georgia, Alabama, Florida, South Carolina, and North Carolina (Frey 1953; Watts 1971; Watts 1975; Delcourt 1980) provide evidence for a hypsithermal event in the southeastern United States occurring between 9000 and 5000 cal yr BP (Delcourt 1980; Delcourt and Delcourt 1980; Whitehead and Sheehan 1985). The hypsithermal event is coeval with persisting negative guano δ2H values interpreted to reflect a wet period. Pollen records also provide evidence for the HCO which has been frequently observed in previous cores collected from lake sediments across the Gulf Coastal Plain (Frey 1953; Watts 1970; Watts 1971; Watts 1975; Delcourt 1979; Delcourt and Delcourt 1980; Whitehead and Sheehan 1985; Watts and Hansen 1994). Additionally, Quercus is at its maximum and Pinus at its minimum from 9000 to 5000 cal yr BP and this was apparent in lake sediment cores from Bartow County, Georgia, USA (241 km east of Priceville, Alabama) when pine began to dominate about 5000 cal yr BP and replaced oak forests during the upper Holocene (Figure 7). Thus, the Quercus to Pinus shift would also be expected to have occurred in northern Alabama with the most likely explanation for this shift to be a change in precipitation. Pinus are better adapted to drier environments (Watts 1971; Watts 1975; Delcourt 1980; LaMoreaux et al. 2009), allowing them to thrive for the last 4000 years since this is when an inferred decrease in moisture in the Cave Springs Cave δ2H record is observed. Quercus dominated from 10,000 – 4,000 cal yr BP due to wetter conditions (Delcourt 1980; Watts and Hansen 1994), which is also supported by the Cave Springs Cave δ2H record.
It is important to take into consideration that pollen may yield different results than bat guano. Pollen producing vegetation can respond to additional environmental factors such as temperature, topography, soil type, and reproductive patterns. δ2H stable isotopes provide a direct measure of moisture, which could cause stratigraphic differences from pollen alterations due to temperature. Furthermore, regional climate differences could cause variations in paleoclimate signals as has been shown from the studies previously mentioned as well as new charcoal records from Florida (Mendieta et al. 2018).
Paleoenvironmental studies based on stable isotope records from wetland and sediment cores disagree on the dominant climate regimes in the southeastern United States across the Holocene. Lake sediment cores from east Georgia, Florida, and the North Carolina and Georgia coastal plains provide evidence for a wet climate during the lower Holocene/upper Pleistocene to middle Holocene (Leigh 2008; Filley et al. 2001; Goman and Leigh 2004; and LaMoreaux et al. 2009), while other studies utilizing sediment and wetland cores interpret the upper to middle Holocene as being a dry period (Delcourt 1979; Driese et al. 2008; Tanner et al. 2015). Other pollen records also support a wet climate during the upper to middle Holocene time period (Watts and Hanson 1994; Goman and Leigh 2004; and LaMoreaux et al. 2009). Whitehead and Sheehan (1985) describe the middle to lower Holocene as being a dry period and attribute the cause to changes in lightning induced fire frequency and anthropogenic changes in land use to southern Florida. The δ2H guano record reported here is inferred as being linked to precipitation alone, thus differences between pollen and δ2H stratigraphies are not surprising given the lack of additional environmental factors recorded in guano. An accepted consensus has not been achieved and additional records are needed to provide a regional pattern.