Robust chronologies for landform development

Abstract

Obtaining chronological control for geomorphological sequences can be problematic due to the fragmentary and non-sequential nature of sediment and landform archives. The robust analysis of 14 C ages is often critical for the interpretation of these complicated sequences. This paper demonstrates a robust methodology for the 14 C dating of geomorphological sequences using a case study from the lower Ribble valley, northwest England. The approach adopted incorporates using greater numbers of ages, targeting plant macrofossils, obtaining replicate measurements from single horizons to assess the extent of reworking and the use of Bayesian approaches to test models of the relative order of events. The extent of reworking of organic materials and space-time dynamics of fluvial change means that it is critical that chronological control is sufficiently resourced with 14 C measurements. As a result Bayesian approaches are increasingly important for the evaluation of large data sets. Assessing the conformability of relative order models informed by interpretation of the geomorphology can identify contexts or materials that are out of sequence, and focuses attention on problem materials (reworking) and errors in interpretation (outlier ages). These relative order models provide a framework for the interrogation of sequences and a means for securing probability-based age estimates for events that occur between dated contexts. This approach has potential value in constraining the sequence of geomorphological development at scales that vary from individual sites to a catchment or region, furthering understanding of forcing and change in geomorphological systems. Figure 4. Landform and depositional contexts used for 14 C dating of geomorphological changes recorded in alluvial units (after Lewin et al., 2005; Chiverrell et al., 2008). RADIOCARBON DATING LANDFORM DEVELOPMENT IN FLUVIAL ENVIRONMENTS 325 Figure 6. Three different chronological models, (a), (c) and (d) for Lower House are used to estimate the age of the 'Geomorphological change', which equates to the switch from terrace T3 to T4 The large square brackets on the left along with the OxCal keywords define the model. A question mark (?) indicates that a result has been excluded from the model. (b) A schematic representation of the structure for Sequences models, with Sequences comprise dating information arranged in an expected order (oldest first) and that one event precedes another is incorporated into the resultant probability distributions. Phases are groups of ages for which no information is assumed about their relative ages but they share similar relationships with other items in the sequence model. Probability distributions have been estimated for features (Geomorphological change) that have not been specifically 14 C dated by placing Date queries within the sequence of 14 C ages. Details of the Markov Chain Monte Carlo sampling and the algorithms employed in Oxcal for this are available from the on-line manual (http://c14.arch.ox.ac.uk/oxcal.html) and in Bronk ...

From sedimentary records to sediment budgets: Multiple approaches to catchment sediment flux

“…In general, either a total sediment depth to the pre-existing land surface is used as a mean erosion rate for the period since the dam was constructed or depths derived from sequential bathymetric surveys are used. Future investigations of sediment budgets from lake sediments urgently need to incorporate catchment-wide sediment dynamics to explore the complex and nonlinear system dynamics in sedimentary systems ( [Dearing and Zolitschka, 1999] and [Coulthard and von De Wiel, 2007] A major approach to improving the precision and accuracy of chronologies using both 14 C and other sources has been through the use of Bayesian modelling of age-depth profiles ( [Sahu, 2004] and [Chiverrell et al, 2008b]) and wiggle-matching both of which are now included in calibration packages such as OxCal (v4.0) and Calib5 ( [Buck et al, 1996] and [Bronk Ramsey et al, 2001]). Such modelling is proving particularly important in dating lake sediments which are notorious for age inversions, hiatuses and a spread of dates from different fractions.…”

Coupling relationships: Hillslope–fluvial linkages in the Hodder catchment, NW England

“…These problems can be reduced through the careful selection of short-lived and fragile plant macrofossils, and through the robust statistical (Bayesian) assessment of sequences of radiocarbon ages (Bronk Ramsey, 2008). Here, we have employed an approach, using Oxcal (Bronk Ramsey, 2001), that allows theoretical models of the likely relative order of events in the geomorphological record to be tested (Chiverrell et al, 2009). Bayesian assessment of these relative order models ( Fig.…”

“…The new radiocarbon ages were obtained for horizons based on the potential to constrain geomorphological changes (sensu Lewin et al, 2005;Chiverrell et al, 2009), thus targeting palaeochannel fills and, in particular, the switches from active-channel to back-channel sedimentation typically at the base of each palaeochannel fill. Sample pre-processing followed the procedures specific to the Oxford (Bronk Ramsey and Hedges, 1997) and SUERC (Scottish University Environmental Reactor Centre) (Slota et al, 1987;Xu et al, 2004) laboratories.…”

Reply to comment: Cumulative probability functions and their role in evaluating the chronology of geomorphological events during the Holocene. Richard C. Chiverrell, Varyl, R. Thorndycraft And Thomas, O. Hoffmann, Journal of Quaternary Science 26: 76–85

“…Multiple parallel sampling for single entities would help eliminate rogue datings, e.g. Chiverrell et al (2009;, but if there is insufficient funding for this type of approach rogue ages can be eliminated, and the probability distributions for individual measurements refined, through the Bayesian analysis of a series of radiocarbon measurements. In the absence of a stratigraphic order implicit in a lake or peat sediment core, geomorphological reasoning can be used to provide prior information for a relative order model that can be tested using Bayesian approaches.…”

“…Whilst for the time and spatially transgressive nature of depositional processes in fluvial geomorphology this is challenging, it is possible to use a geomorphologically reasoned order of ages to build deposition models that can be tested using Bayesian statistics (e.g. Chiverrell et al, 2009;Howard et al, 2009). The posterior density functions for individual ages generated by this modelling of a series of radiocarbon ages focus on the relevant probability in the probability function and present a more robust estimate of the age of the samples.…”