AWA: Academic Writing at Auckland

A Fine Line: Secondary Ion Mass Spectroscopy Measurements of Obsidian Hydration Bands as an Alternative to Radiocarbon for Dating New Zealand’s Prehistory

Describe your research question/hypothesis or research objective. That is, what will be the focus of your investigation?

Dating archaeological sites in New Zealand is hindered by issues of calibration in radiocarbon dating. The presence of a large wiggle across the period of time in which the prehistory of the islands occurs causes margins of error that significantly increase the age ranges for sites. Obsidian Hydration Dating (OHD) had been proposed as an alternative to overcome this problem due to the frequency of obsidian artefacts in prehistoric sites. Unfortunately, traditional methods for OHD relied on visual analysis meaning that there was still a significant amount of error. Recent developments in technology have led to advances in this method of dating with many scholars in other parts of the world reassessing its potential. Secondary Ion Mass Spectrometry (SIMS) has proven successful in many of these studies by reducing the associated error ranges as well as being fast and relatively inexpensive when compared to radiocarbon dating. SIMS therefore has the potential in a New Zealand context to not only overcome the issues of radiocarbon calibration but to offer reduced costs and increased efficiency for archaeological research. This project assesses the viability of using SIMS in New Zealand archaeology by comparison of its accuracy with established radiocarbon dates.

How does your research build on existing scholarship in anthropology and closely related disciplines? Give specific examples of this scholarship and its findings.

Liritzis and Laskaris (2011) reviewed the history of OHD from its conception by Friedman and Smith in 1960 through to the most recent developments. They noted that during this time the initial excitement of a cheap and accurate dating method had waned due to the inaccuracies caused by theoretical and methodological short comings (Liritzis and Laskaris, 2011). Initially assumed to be a linear relationship between climate, time and chemical composition of the material investigations found this was not the case and inaccuracies were compounded by errors in optically measuring hydration bands (Liritzis and Laskaris, 2011; Anovitz et al., 1999). Rogers’ investigations concluded that removal of systematic errors caused by the optical method along with better understanding of composition, accurate effective hydration temperatures (EHT), knowledge of intrinsic hydroxyl levels and larger sample sizes would greatly increase accuracy of the method (Rogers, 2008, 2010). Anovitz et al. (1999) advocated for advancing the mathematical theory to increase accuracy when modelling the hydration process using a combination of experimental and archaeological samples. Both identified SIMS as valuable in improving measurements of hydration bands for both experimental and archaeological material as well as analysing chemical composition (Anovitz et al., 1999; Rogers, 2006, 2010).

SIMS has seen wide use in analyses and preservation of archaeological material yet its application in OHD has been relatively recent (Dowsett and Adriaens, 2004). In that time archaeologists have used the method alongside independent data from Thermo Luminescence and Radiocarbon dating with results showing good correlation to the established chronologies (Riciputi et al., 2002; Stevenson et al., 2004; Liritzis and Laskaris, 2009; Liritzis, 2010; Laskaris et al., 2011). While margins of error were acceptable Anovitz et al. (1999) identified that analysis of artefacts older than 2000 years increased inaccuracy due to the complicated interactions of many variables over long time frames which poorly fit standard mathematical models commonly used. Discovery that atomic and isotopic analyses of water and minerals contained in hydration bands could offer palaeoclimatic information to aid in calculating values for mathematical constants was useful in overcoming this (Anovitz et al., 1999).

New Zealand’s short prehistory is well within the aforementioned 2000 year timeframe allowing reasonably standardised mathematical formulas to be applied based upon the variables of time, climate and composition with minimal increases in error ranges (Anovitz et al., 1999). Previous investigations into OHD in New Zealand done by Stevenson et al. (1996) were conducted using computer assisted imaging providing dates which compared well to those obtained using radiocarbon dating. However, there still remained a noticeable margin of error. Given the establishment of a difference between hydration and optical boundaries following this research and the inherently small hydration bands of artefacts of such a young age it would seem that this investigation should be repeated to see if the margin of error can be reduced using SIMS (Anovitz et al., 1999). If established as a reliable method of dating SIMS would also allow for a reduced emphasis on radiocarbon dating while supplying sourcing information and aiding in palaeoclimatic reconstructions providing a large amount of information for minimal expense.

How will you go about collecting and analysing this evidence? What will your variables and how are they operationally defined?

In order to be comparative to the original study conducted by Stevenson et al. (1996) it will be necessary to analyse the samples that they worked with. The sites involved were Puriri (T12/318), Waiwhau (T13/756), Pukehue Paa (T12/229), Hurumoimoi Paa (T12/347), Twilight Beach (N1 & 2/976) and Sunde site (N38/24) (Stevenson et al., 1996). As OHD is a destructive method, with optical assessment requiring a thin section of the artefact, it may not be possible to work with the same artefacts used by these scholars. As such a new selection of artefacts would need to be drawn from the material of each site. In order to accurately replicate the sample the artefacts would need to come from the same source areas since chemical composition affects the outcomes as described above.

The variables for this part of the study would be the sites, operationally defined as the physical areas of archaeological excavation and study named above, obsidian artefacts, obsidian flakes gathered within the archaeological matrix of the excavations that show minimal post-depositional disturbance or damage, and geochemical source, the volcanic area of origin which the obsidian can be geochemically traced to. Another variable required for this study is radiocarbon dates. These are operationally defined as the age ranges within two standard deviations of the mean obtained from radiocarbon dating techniques. This information would be located within the published articles and grey literature for the sites. A review of the sources for this from Stevenson et al.’s (1996) original study would be necessary. The final variable required would be Stevenson et al.’s (1996) age determinations from their OHD results, operationally defined as the published dates found using the enhanced optical technique in their article.

In order for the statistical testing to be as accurate as possible it would be preferable to extend the study beyond the six sites originally examined by Stevenson et al. (1996). The sites added to the study would require the same variables to be present, obsidian artefacts obtained from archaeological excavations and calibrated radiocarbon dates. These sites would be identified through a review of the published and grey literature on New Zealand archaeological investigations and the variables obtained from these writings and the collected material.

All obsidian artefacts will then need to be dated using the hydration bands. To obtain measurements of the hydration bands a SIMS spectrometer would be used with a measurement scale of micrometres (µm). The hydration band is defined as the depth of externally derived water from the exposed surface of the obsidian. Also required for the calculations are density, EHT, activation energy, percentage of intrinsic water and high temperature hydration rate. Density, weight per cubic centimetre, will be measured in grams per cubic centimetre (g cm³). The EHT, average temperature during exposure, will be measured in degrees Celsius as will the high temperature hydration rate which is an experimentally derived value for the temperature required for obsidian of a certain composition to absorb a pre-determined amount of external water in a set time frame. Activation energy, a mathematical constant derived from the same experiments, is measured in kilo joules per mole (kj mol⁻¹). The intrinsic water content is determined as the amount of water included in the natural composition of the obsidian, usually as hydroxyl, and measured in parts per million then converted into a percentage of weight (weight %). These scales are the same as those used by Stevenson et al. (1996) to measure the hydration bands allowing for accurate comparison. Age ranges from both OHD and radiocarbon, measured in years AD, will be converted to a mean for statistical testing.

Sourcing of all obsidian artefacts in the samples is also required for an accurate analysis of the hydration band and evaluating SIMS use in geochemical sourcing. This will be achieved using X-ray Fluorescence (XRF) measuring the elements in the obsidian as parts per million (ppm). It will then be repeated with the SIMS analysis as part of measuring the hydration band and intrinsic water content again using ppm of the elements.

Potential biases for this research will be associated with the selection of obsidian to be analysed. Obsidian that has been damaged after its original deposition will cause modification of the original hydration band and new surfaces to be exposed altering the final dates. It is therefore important to select artefacts that appear undamaged or have edges that can be identified as exposed at the time of manufacture.

What statistical tests are appropriate to answer your research question? How will you interpret your results?

All samples used in the statistical analyses will first undergo evaluations to ensure they meet the assumptions for parametric tests. This will include the Shapiro-Wilk test for normal distribution, with a result of greater than 0.05 representing normally distributed data, and Levene’s test for homoscedascity, a value above 0.05 indicating that variances are equal. If these tests suggest there is a violation of the assumptions log, natural and square root transformations will be applied and the tests repeated to see if the issues are rectified. Once these tests are completed each individual site will have the radiocarbon and SIMS OHD dates compared using Student’s Independent Two Sample T-tests if the assumptions are met or Mann – Whitney U tests if they are violated. Results with a p-value >0.05 will fail to reject the null hypothesis, being that there is no significant difference between the C¹â´ and SIMS OHD dates. Student’s Paired T-tests, or Wilcoxon tests if assumptions are violated, will be used to determine if the optical OHD and SIMS OHD dates differ. A p-value of <0.05 will cause the null to be rejected and the alternative hypothesis, that there is a statistically significant difference between the dates obtained by the methods, to be accepted.

Sourcing analyses will be carried out also. Firstly, Discriminant Function Analysis will be used to classify samples from the University of Auckland reference collection using XRF results. The graph will show clusters of material based upon the differences in elemental composition. The classified samples will then be added to the unknown samples from the archaeological sites. XRF values from the unknown samples will then be added to determine their geological sources. The same analysis will then be run using values obtained by SIMS for both the known and unknown samples. Plots from the XRF and SIMS can then be compared to see if there is a difference in the groupings.

The goal of the Wenner-Gren Foundation is to support original and innovative research in anthropology. What contribution does your project make to anthropological theory and to the discipline?

The potential of SIMS to improve OHD has been investigated elsewhere in the world with promising outcomes. New Zealand’s current dependence on C¹â´ dating results in dates with wide error ranges for high costs. SIMS offers a solution which has the possibility to not only provide more accurate dates for archaeological sites but return geochemical sourcing and palaeoclimatic information as well for a fraction of the cost. This would reduce the expenses of archaeological investigations while giving dates with less error so that the prehistory of New Zealand may be better understood possibly altering the chronologies that currently exist. If it proves effective here it may offer a model that is useable in other areas of the Pacific and the world as well as an independent means of verifying radiocarbon dates.

Bibliography

Anovitz, Lawrence M., J. Michael Elam, Lee R. Riciputi and David R. Cole. 1999. The Failure of Obsidian Hydration Dating: Sources, Implications, and New Directions. Journal of Archaeological Science. (1999) 26: 735–752.

Dowsett, Mark and Annemie Adriaens. 2004. The Role of SIMS in Cultural Heritage Studies. Nuclear Instruments and Methods in Physics Research B. 226 (2004): 38–52.

Laskaris, N., A. Sampson, F. Mavridis and I. Liritzis. Late Pleistocene/Early Holocene Seafaring in the Aegean: New Obsidian Hydration Dates with the SIMS-SS Method. Journal of Archaeological Science. 38 (2011): 2475-2479.

2009. Liritzis and N. Laskaris. 2009. Advances in Obsidian Hydration Dating by Secondary Ion Mass Spectrometry: World Examples. Nuclear Instruments and Methods in Physics Research B. 267 (2009): 144–150.

Liritzis, Ioannis. 2010. Strofilas (Andros Island, Greece): New Evidence for the Cycladic Final Neolithic Period through Novel Dating Methods using Luminescence and Obsidian Hydration. Journal of Archaeological Science. 37 (2010): 1367–1377.

Liritzis, Ioannis and Nickolaos Laskaris. 2011. Fifty Years of Obsidian Hydration Dating in Archaeology. Journal of Non-Crystalline Solids. 357 (2011) 2011–2023.

Riciputi, Lee R., J. Michael Elam, Lawrence M. Anovitz and David R. Cole. 2002. Obsidian Diffusion Dating by Secondary Ion Mass Spectrometry: A Test using Results from Mound 65, Chalco, Mexico. Journal of Archaeological Science. 29 (2002): 1055–1075.

Rogers, Alexander K. 2006. Induced Hydration of Obsidian: A Simulation Study of Accuracy Requirements. Journal of Archaeological Science. 33 (2006): 1696-1705.

Rogers, Alexander K. 2008. Obsidian Hydration Dating: Accuracy and Resolution Limitations Imposed by Intrinsic Water Variability. Journal of Archaeological Science. 35 (2008): 2009-2016.

Rogers, Alexander K. 2010. Accuracy of Obsidian Hydration Dating based on Obsidian-Radiocarbon Association and Optical Microscopy. Journal of Archaeological Science. 37 (2010) 3239-3246.

Stevenson, Christopher M., Peter J. Sheppard, Douglas G. Sutton and Wallace Ambrose. 1996. Advances in the Hydration Dating of New Zealand Obsidian. Journal of Archaeological Science. 23 (1996): 233–242.

Stevenson, Christopher M., Ihab Abdelrehim and Steven W Novak. 2004. High Precision Measurement of Obsidian Hydration Layers on Artefacts from the Hopewell Site using Secondary Ion Mass Spectrometry. American Antiquity. 69 (3): 555-568.