This shows you the differences between two versions of the page.

Both sides previous revision Previous revision | |||

age-depth_chronologies [2007/07/04 11:53] mblaauw |
age-depth_chronologies [2015/06/23 17:22] (current) |
||
---|---|---|---|

Line 1: | Line 1: | ||

+ | ===== Age-depth modelling in the EPD ===== | ||

+ | |||

+ | By far most of the sites in the EPD have some some of age control. Age models are vitally important for palaeoecological/palaeoclimatological studies, especially when multiple proxy archives need to be compared (e.g., an EPD search for mid-Holocene sites). Basically we can distinguish two types of age-controls in EPD sites: pollen-stratigraphic and radiocarbon dates. Whereas pollen-stratigraphy can give us relative and often rather imprecise information (e.g., a site with a mid-Holocene pollen spectrum), radiocarbon (<sup>14</sup>C) can potentially provide more precise and absolute ages to pollen cores. On the other hand, cores often contain only a handful of <sup>14</sup>C dates owing to constraints on budget and abundance of datable organic material. From this limited amount of information, and through assumptions about the accumulation history (sedimentation rate, hiatuses), the ages of all depths of a core need to be estimated somehow (age-depth modelling). | ||

+ | |||

+ | === Radiocarbon === | ||

+ | |||

+ | Atmospheric CO<sub>2</sub> contains a tiny proportion (only 1 in every 10<sup>12</sup> atoms) of the <sup>14</sup>C isotope. As the CO<sub>2</sub> is incorporated into the carbon cycle, living plants in equilibrium with the atmosphere (and their consumers) will obtain the same concentration of <sup>14</sup>C atoms (the concentration will be lower in systems with old carbon present, such as oceans and hard-water lakes). Because radiocarbon is radioactive, as time goes on the <sup>14</sup>C atoms will fall apart (click [[http://www.chrono.qub.ac.uk/Members/MBlaauw/C14decay.html]] for an animation). When exactly an individual <sup>14</sup>C atom will disappear cannot be predicted; it could happen within a few seconds, but it could also survive several tens of thousands of years. In organic material, after waiting 5568 years (the half-life of <sup>14</sup>C) only half of the original concentration of <sup>14</sup>C atoms will remain. By measuring the remaining concentration of radiocarbon, the <sup>14</sup>C age of fossil material can thus be determined. Currently the limit for <sup>14</sup>C dating lies around c. 10 <sup>14</sup>C half-lifes; thus <sup>14</sup>C dates older than c. 40-60,000 years obtain infinite ages when measured (//Is Infinite// column in EPD Dating table). | ||

+ | |||

+ | To-be-dated fossil material (//Material Dated// in the EPD Dating table) can either be bulk sediment/lake mud, or a selection of material from the sediment (e.g. macrofossils such as seeds, twigs and leaves). The <sup>14</sup>C/C ratio of the fossil material can be measured in two ways: through counting the number of <sup>14</sup>C disintegrations over a time-interval (//conventional radiometric dating//, either using //liquid scintillation// or //proportional gas counting//, see [[http://www.c14dating.com]]), or through counting the <sup>14</sup>C/C mass ratio directly (accelerator mass spectrometry or AMS). For the former method, large samples are needed, often spanning several cm of sediment (//Thickness// in the EPD Dating tables), and consisting of bulk sediments. The AMS technique can handle much smaller samples, and can thus do with thinner sediment slices (more precise/accurate age estimates). | ||

+ | |||

+ | Owing to laboratory treatments, machine drift, and statistical scatter, <sup>14</sup>C ages are always reported with an uncertainty or error estimate, e.g., 2450 ± 50 years (//Age//, //Limit Younger// and //Limit Older// in the EPD Dating tables). The dates always have a unique label given by the <sup>14</sup>C laboratory (//Lab Number// in the EPD Dating tables) so that its details can be retrieved if needed. | ||

+ | |||

+ | Most laboratories measure the d<sup>13</sup>C value of the dates (//delta13C// in EPD Dating tables). This is done to correct for fractionation; during assimilation of CO2, organisms will slightly favour the lighter isotopes <sup>12</sup>C and <sup>13</sup>C over the heavier <sup>14</sup>C. This discrimination against <sup>14</sup>C is corrected for by measuring the ratio of the stable isotopes <sup>13</sup>C/<sup>12</sup>C, and by assuming that the <sup>14</sup>C/12C fractionation in the sample is twice that of the <sup>13</sup>C/<sup>12</sup>C. | ||

+ | |||

+ | From measuring the <sup>14</sup>C ages of material of exactly known calendar age (dendrochronologically dated tree-rings), it has become clear that the concentration of atmospheric <sup>14</sup>C has fluctuated in the past. Thus there is no one-to-one relationship between real age and <sup>14</sup>C age; <sup>14</sup>C ages need to be calibrated. For most northern hemisphere terrestrial samples with <sup>14</sup>C ages up to c. 21,000 years, the IntCal04 calibration curve (Reimer et al. 2004) is recommended; no accepted calibration curve exists yet for samples older than 21,000 years, and for the most recent, “post-bomb” <sup>14</sup>C dates, dedicated calibration curves should be used (Hua and Barbetti 2004). | ||

+ | |||

+ | Calibration of <sup>14</sup>C ages (e.g. on CALIB, [[http://calib.qub.ac.uk]]) and subsequent interpretation of the calibrated ages is not a straightforward task. Owing to the wiggly nature of the calibration curve, calibrated ages often possess multiple and asymmetric local maxima in their probability range (see Fig 1). It is not recommended to obtain point calendar age estimates using the intercepts with the calibration curve, because they do not provide a robust indicator of sample calendar age. Instead the weighted average or the median probability of the probability distribution are preferred (Telford et al 2004). | ||

+ | |||

+ | {{calibrate.png|}}\\ | ||

+ | Fig. 1. Example of a calibrated <sup>14</sup>C date | ||

+ | |||

+ | === Age-depth modelling === | ||

+ | |||

+ | As mentioned above, constraints in budget and datable material means that not every depth of a pollen profile can be dated individually. Most cores only have a handful of dates spanning several millennia and meters of sediment. Thus, some sort of interpolation and/or extrapolation is needed to estimate the ages of non-dated levels (Bennett, 1994). Most age-depth models in the EPD were constructed using interpolation between the dated levels (often including the core top and its year of sampling as extra dating point), while some EPD age-depth models use linear, polynomial or power regression to estimate the ages of the non-dated depths. Sometimes the age-depth model is extrapolated beyond the dated levels; extrapolated ages should be treated with much caution, as should core sections with low dating resolution. | ||

+ | |||

+ | The choice of age-depth model (see //Default Chronology// in EPD site information) is somewhat subjective. Often a range of models is tested on the pollen core, and then the model which seems the most “realistic” is used. The settings of the site could help in deciding about which model to use; sites which can be assumed to have accumulated in a relatively stable manner over the studied period, could for example be modelled using linear regression, while sites with many likely changes in sedimentation rate (e.g. sites which changed from a lake to a mire) could perhaps be better modelled using linear interpolation between the dated levels. Indeed, it is recommended to check how different modelling assumptions would impact the age-depth models of individual sites. | ||

+ | |||

+ | Sometimes some dates appear to disagree with the other dates and the model used. If this is the case, such outlying dates are removed from the analysis, and a remark about this is made in the EPD Default Chronology or Dating Samples table. | ||

+ | |||

+ | Most of the age-depth models currently provided in the EPD are based on uncalibrated <sup>14</sup>C ages. As <sup>14</sup>C dates need to be calibrated for reliable age-depth models, we are currently working on updating our age-depth modelling routines. We are also working on ways to estimate the age uncertainties for the cores (e.g. Heegaard et al. 2005, Blaauw et al. 2007). | ||

+ | |||

+ | === References === | ||

+ | |||

+ | Bennett KD, 1994. Confidence intervals for age estimates and deposition times in late-Quaternary sediment sequences. //The Holocene// **14**: 337-348 | ||

+ | |||

+ | Bennett KD, 2007. Psimpoll and pscomb programs for plotting and analysis. http://www.chrono.qub.ac.uk/psimpoll/psimpoll.html | ||

+ | |||

+ | Blaauw M, JA Christen, 2005. Radiocarbon peat chronologies and environmental change. //Applied Statistics// **54**: 805-816 | ||

+ | |||

+ | Blaauw M, Christen JA, Mauquoy D, van der Plicht J, Bennett KD, 2007. Testing the timing of radiocarbon-dated events between proxy archives. //The Holocene// **17**: 283-288 | ||

+ | |||

+ | Heegaard E, Birks HJB, Telford RJ, 2005. Relationships between calibrated ages and depth in stratigraphical sequences: an estimation procedure by mixed-effect regression. //The Holocene// **15**: 612-618 | ||

+ | |||

+ | Hua Q, Barbetti M, 2004. Review of tropospheric bomb C-14 data for carbon cycle modeling and age calibration purposes. //Radiocarbon// **46**: 1273-1298 | ||

+ | |||

+ | Reimer PJ, Baillie MGL, Bard E, Bayliss A, Beck JW, Bertrand CJH, Blackwell PG, Buck CE, Burr GS, Cutler KB, Damon PE, Edwards RL, Fairbanks RG, Friedrich M, Guilderson TP, Hogg AG, Hughen KA, Kromer B, McCormac FG, Manning SW, Ramsey CB, Reimer RW, Remmele S, Southon JR, Stuiver M, Talamo S, Taylor FW, van der Plicht J, Weyhenmeyer CE, 2004. IntCal04 terrestrial radiocarbon age calibration, 26 - 0 ka BP. //Radiocarbon// **46**: 1029-1058. | ||

+ | |||

+ | Telford RJ, Heegaard E, Birks HJB, 2004. The intercept is a poor estimate of a calibrated radiocarbon age. //The Holocene// **14**: 296-298. | ||