QUB | Archaeology and Palaeoecology | The 14Chrono Centre
N Plot confidence intervals (off) [off] Enter '0' for on, or '1' for offThis menu controls the calculation of confidence intervals for age estimates and sediment accumulation rates, and also allows the calculation of confidence intervals for data values (which, in certain cases, need the confidence intervals for sediment accumulation rates). Certain extra information must be available to obtain confidence intervals. For % data, pollen sums must be available (marked `*': see above), and must be positioned in the input file after all of the taxa included in those sums. If no sum is found among the pollen taxa after a particular pollen type in the file, no confidence intervals will be calculated for that pollen type. Note that separate sums must be given for taxa in the main sum, aquatics, Sphagnum, indeterminables etc, if separate sums were used to calculate their percentages. The method of calculation follows Maher (1972). For concentrations (`c' in input Item 4 of main input file), additional data are needed (input Items 7-13, as above), and for accumulation rates (`b' in input Item 4), a C14 file must be present, and additional data for Items 7-14. These Items are described in Data preparation.
Confidence intervals for sediment age estimates and sediment accumulation rates are obtained, mostly, by simulation from the available radiocarbon dates and their associated errors (Bennett 1994), or calibrated dates. Values are obtained using means and standard deviations of radiocarbon ages, from 100 simulations by default, or by modelling from the irregular distributions of probabilities of calibrated ages. These may take a few minutes to calculate. The output file (see E2) includes sample standard deviations of age and accumulation rates. These need to be divided by the square root of the number of simulations to obtain standard deviations of the mean, for comparison with radiocarbon age errors.
Confidence intervals are `exact' (i.e., obtained by calculation, rather than simulation) for age estimates from linear interpolation, and for the gradient with 2-term polynomial line-fitting.
Using the linear interpolation method (menu L1), you should find that you get wider confidence intervals where the radiocarbon dates are closest together. The line-fitting methods should not show this effect.
Confidence intervals are not available in combination with interpolation (menu Bf) are not available, because the value at any level is a combination of values from several levels.
If 0 is selected, the following submenu, given below in full, appears (option 1 only appears for accumulation rate data):
1 Fine-scale variation in sediment accumulation [0.30] Default = 0.3 per 5 years 2 Method for calculating confidence intervals [1] 3 Method for generating random numbers [1] 4 Seed for random number generator [TIMER] 5 Number of simulations [100] 6 Include negative accumulation rates in models [off] 7 Attempts to model dates without -ve acc. rates [1000000] 9 Leave this menu
Return to main menu.
Archaeology and Palaeoecology | 42 Fitzwilliam St | Belfast BT9 6AX | Northern Ireland | tel +44 28 90 97 5136