5.)Results and Discussions
AMS 14C results from the standards demonstrate that the sample pretreatment protocol does not have a measurable affect on the radiocarbon ages and stable carbon isotopic values (Table 1). Furthermore, 14C ages and δ13CPDB values obtained from the calcium oxalate standard and four crust sample aliquots digested in either acetic acid or phosphoric acid are statistically indistinguishable, indicating the type of acid used does not affect age or isotopic determinations (Table 2). Finally, radiocarbon ages from sample 41VV129-1 that contained two independent crust strata and a layer of prehistoric paint at the substrate/crust interface are in the correct stratigraphic order. The paint layer, expected to have been produced between 2950 and 4200 rcyrs BP, was below an oxalate stratum with an uncalibrated radiocarbon age of 3220±60 rcyrs BP while the upper crust stratum produced an uncalibrated radiocarbon age of 2000±80 rcyrs BP.
5.2 14C age distribution
The temporal distribution of radiocarbon ages indicates that the production of calcium oxalate occurred episodically during the middle to late Holocene, and nearly simultaneously at several sites at various times. For example, three overlapping 14C ages (1970±100, 2070±90 and 2080±120 cal. yr BP) indicate that oxalate production occurred at three di!erent sites (41VV129, 41VV167 and 41VV83, respectively) at essentially the same time. We suggest that there were four primary periods of oxalate production based on clustering of the 14C data (Table 2, Fig. 4). The average age difference within the de”ned clusters is less than 170 cal. yr BP, while the average age between the clusters is 810 cal. yr BP. Each cluster consists of data from multiple sites, which further supports concurrent lichen activity within various sites in the region.
5.3 Inter-site comparisons
The four 14C ages from site 41VV89 range from 680±80 to 1240±30 cal. yr BP, all of which fall within the earliest cluster. The radiocarbon ages of the sample split (41VV89-6A and 41VV89-6A2) were 1240±60 and 990±60 cal. yr BP, respectively, whereas the ages of samples collected ~1 m (41VV89-5B) and ~36 m (41VV89-26) apart were 680±80 and 820±60 cal. yr BP, respectively. The scatter of data could indicate the rate at which the lichen spread in the site, different episodes of oxalate production, or inherent errors in the 14C ages from the oxalate.
Five radiocarbon ages were obtained from two samples from site Pressa4. The first sample, Pressa4-1, was divided into approximately three equal splits, and the results showed disparate 14C ages (Table 3). The second sample from this site, Pressa4-3, was split into halves, and yielded overlapping ages (6030±80 and 6080±80 cal. yr BP). Overall, the five 14C ages from two Pressa4 samples ranged from 4140±100 to 7320±70 cal. yr BP. This strongly suggests multiple episodes of oxalate production within this site and on a small area of rock surface, and points to a potential problem for obtaining 14C ages from single samples corresponding to one period of lichen productivity. We opted not to use the data from Pressa4-1 in our reconstruction because of the likelihood that the ages do not represent single episodes of oxalate production.
Multiple periods of lichen productivity within some sites is further evident by radiocarbon ages from stratified samples. The three samples where upper and lower oxalate strata were removed and analyzed separately yielded results which show that each strata within the individual sample was produced at a different time. From sample 41VV167-1 we obtained 14C ages of 1360±50 and 2070±90 cal. yr BP for top and bottom layers, respectively, whereas for 41VV128-7 ages of 3010±110 and 4670±100 cal. yr BP were obtained for upper and lower layers. As previously noted, the sample with a paint layer (41VV129-1) also contained a stratified crust that yielded radiocarbon ages of 1970±100 and 3460±70 cal. yr BP for the upper and lower strata, respectively.
5.4 Paleoclimate reconstruction
We suggest that periods of high lichen productivity (as established by clusters of oxalate 14C ages) reflect dry climate periods due to increased temperature and/or reduced precipitation (Fig. 3). Gaps in the oxalate 14C record are interpreted as relatively wetter conditions when the lichen was absent or dormant. Accordingly, five xeric episodes are postulated for the middle and late Holocene: from 6380 to 6030 cal. yr BP, from 4670 to 4500 cal. yr BP, from 3790 to 2840 cal. yr BP, from 2080 to 1760 cal. yr BP, and from 1360 to 680 cal. yr BP. More moist conditions appear to have occurred during intervening periods (Fig. 3). The paucity of 14C ages older than 6380 cal. yr BP limits our ability to evaluate earlier climate regimes.
This reconstruction is consistent with previous records from the southwestern Edwards Plateau (Bryant and Holloway, 1985; Toomey et al., 1993; Blum et al., 1994) in predicting the onset of dry climate at 7320 cal. yr BP (6420 rcyrs BP), and a dominant warm period from 5020 to 2840 cal. yr BP (4400-2730 rcyrs BP). We also agree with a return to wet-cool conditions &2500 rcyrs BP with a mesic period from 2840 to 2080 cal. yr BP (2730-2080 rcyrs BP); however, our data would suggest a brief xeric period between 2080 and 1760 cal. yr BP (2080-1830 rcyrs BP), followed by a return to more mesic conditions from 1760 to 1360 cal. yr BP (1830-1440 rcyrs BP). These ephemeral wet/dry climate episodes could account for the inconsistency between the previous reconstructions. It is possible that Bryant and Holloway (1985) recorded the rapid return to dry conditions after the brief mesic interlude at 2500 rcyrs BP, but failed to resolve the ensuing wet period that we indicate between 1830 and 1330 rcyrs BP. Toomey et al. (1993) and Blum et al. (1994) did not interpret a brief return to dry conditions after about 2070 rcyrs BP, resulting in the prediction of an extended wet period. Finally, our reconstruction does not match the dry conditions for the last 1000 years inferred by Bryant and Holloway (1985), Toomey et al. (1993) and Blum et al. (1994). Instead, we predict more moist conditions from about 730 rcyrs BP to the present.
5.5 Limitations and Future Directions
The results of this study indicate an episodic production of biogenic calcium oxalate in southwestern Texas, and thus presents a potential means for obtaining paleoclimate information based on radiocarbon ages from the residue. However, the reliability of oxalate radiocarbon ages representing wet-dry climate fluctuations is subject to three assumptions: (1) oxalate was produced only during dry climate conditions; (2) multiple episodes of oxalate production can be distinguished in samples selected for radiocarbon dating; and (3) bicarbonate from the substrate was not metabolized by the lichen. For assumption 1, similarities in the micromorphology and chemistry of the oxalate crust throughout the southern Edwards Plateau (Russ et al., 1996), as well as the characteristic microenvironment in which it occurs, suggests the biogenic substance was produced by either a single or similar lichen species. Further studies are needed to determine whether biomarker compounds occur in the crust, so specific organisms can be identified as the source of the oxalate. If the 2nd assumption is not true then measurements of oxalate 14C ages would produce dates that are average ages of two or more periods of oxalate deposition, and would not correspond to climate regimes. We analyzed samples using optical microscopy and SEM in an attempt to exclude such samples in our data base. Moreover, we suggest new methods using lasers to combust independent oxalate strata in microstratified crusts (Watchman et al., 1993) could possibly be employed in future experiments to provide the selectivity in stratified samples. Finally, the 3rd assumption would cause anomalously old 14C ages due to inclusion of “dead” carbon from the limestone. This does not occur to a substantial degree since the 14C ages are relatively recent. Much older dates would be obtained if significant amounts of bicarbonate were incorporated within the oxalate. Furthermore, radiocarbon ages of discrete oxalate layers from stratified crust samples were in stratigraphic order, including the paint layer in sample 41VV129-1, further demonstrating that little if any limestone carbon was metabolized by the lichen.
The oxalate δ13CPDB values (mean=-10‰ range=7.6‰) indicate an enrichment in 13C compared to previous reports of δ13CPDB for lichen thalli, which range from -35 to -14‰ (Lange et al., 1988). As we argue above, the primary enrichment in 13C cannot be due to an incorporation of limestone carbon. Instead, we suggest the 13C enrichment reflects a specific metabolic shift such as that shown by Rivera and Smith (1979) for the production of solid oxalate by cacti. We cannot,
however, rule out a minor bicarbonate component in the production of oxalate.
Finally, reconstructing wet-dry climate shifts based on lichen productivity requires sufficient clumping of the oxalate 14C ages to differentiate between periods of oxalate production and episodes when no oxalate was produced. For our current data set (calibrated radiocarbon
ages) we tested the uniformity of the temporal distribution of the 14C ages across the interval of 0-ca. 7500 cal. yr BP using a Chi-square test of the deviation of the frequency of radiocarbon ages from a Uniform distribution (using BestFit software, Palisade Corp.). For this purpose, this interval was subdivided into a series of shorter time intervals and the frequency of ages within those intervals were tallied. Results therefore depend on the length of the intervals chosen. Five tests were run with 8-20 categories (intervals of ca. 350 to 900 yr). In no case did the Chi-square value approach significance (range of > 0.53 to > 0.85); thus, the statistical test does not indicate any deviation from a uniform distribution of radiocarbon data. While more robust statistical treatments could be established to discriminate between data clusters in the present set, it is clear that additional radiocarbon data are required to obtain confidence in representing climate regimes based on clusters of 14C ages.