The development of the database presented here is described in Brooks & Edwards (2006). Users are advised to read this prior to using the database, as it outlines a number of important limitations associated with the data it contains.
Reference: Brooks, A.J., and Edwards, R.J. (2006) The Development of a Sea-Level Database for Ireland. Irish Journal of Earth Science v. 24 p 13-27.
DISCLAIMER: Whilst we have made every effort to ensure that the data contained within this database are correct, we cannot guarantee that all entries are error-free. The user is therefore advised to consult with the source publications associated with any of the data they choose to employ. A full set of references are available here. If you notice any errors, please email me.
This database was produced as part of project SC-2003-0215-Y entitled Holocene Sea Level Change and Glacio-Isostatic Adjustment in Ireland funded by Enterprise Ireland.
- What do we mean by 'sea-level data'?
- Database structure
- Limitations & things to be aware of
- What about new or missing data?
- References and Associated Reading
- Download Database (.xls file) Last Updated: November 2006
The delicate balance between post-glacial changes in land and ocean level has helped to shape Ireland ’s coastline. The pattern of relative sea-level (RSL) changes produced by this interplay contains information concerning both the timing and magnitude of eustatic sea-level change, and the spatial variability associated with glacio-isostatic adjustment (GIA). Consequently, a firm understanding of RSL changes around Ireland underpins a variety of research, ranging from studies examining ice sheet history, Earth rheology, climate and sea level, to those concerned with coastal evolution, palaeogeography and archaeology.
Several attempts have been made to model RSL change and GIA around the Irish coast and elsewhere (e.g. Lambeck 1996; Lambeck and Purcell 2001). However, some authors have argued for caution in the adoption of these models, highlighting contradictions with field data (e.g. McCabe 1997; Smith 2005). It is, therefore, vital that the collection of primary data continues in association with ongoing model development.
The database presented here is a first attempt at compiling a standardized sea-level database for Ireland, which is consistent with established protocols for the evaluation of sea-level information. The available data are of varied quality, yet limitations associated with this are mitigated by applying a rigorous appraisal of data errors and the consequent adoption of a quality tier system. We hope that this will prove a valuable tool to a range of user groups including coastal scientists, archaeologists and geophysicists interested in understanding patterns of GIA in Ireland.
Over thirty years of international research, co-ordinated under the auspices of a series of International Geoscience Programme (IGCP) projects, has produced a well-defined methodology for developing records of relative sea-level change from sedimentary coasts (Edwards 2005). Central to this is the use of sea-level index points which fix the past altitude of sea level in time and space (Tooley 1978; Preuss 1979; van de Plassche 1986). Detailed consideration of sea-level index points and their associated error terms are given in a series of publications (e.g. Kidson and Heyworth 1979; Devoy 1982; Heyworth and Kidson 1982; Shennan 1982; 1986). In brief, for a sample to be established as a sea-level index point, it must possess information regarding its location (latitude and longitude), its altitude (relative to a levelling datum), its age (commonly inferred from radiocarbon dating), and its vertical relationship to a contemporaneous tide level (termed the indicative meaning). The latter is important when accounting for the differing vertical distributions of coastal sub-environments and associated sea-level indicators. Sea-level index points are commonly derived from lithostratigraphic contacts between terrestrial and marine sediments, with supporting microfossil data (e.g. foraminifera, diatoms) being used to delimit the onset or removal of brackish/marine conditions.
Where a dated sample does not possess an indicative meaning (e.g. a freshwater peat deposit) it cannot be used to reconstruct the former altitude of RSL. However, given knowledge of its depositional environment it can be used as a limit of possible RSL, and is thus referred to as a limiting date. For example, in the case of a freshwater peat deposit, the sample must have formed above the upper limit of marine influence indicating that, at the time of accumulation, local RSL must have resided at a level somewhere below the sample altitude.
Sea-level index points are routinely plotted as points on age–altitude diagrams with associated vertical and temporal error terms. These points fix the former altitude of RSL in time, but do not provide any information on the nature of sea-level change between points. The resolution at which past RSL changes can be reconstructed is therefore a function of the number and distribution of data points and the relative magnitudes of their associated error terms. Since RSL records are influenced by vertical land movements, it is important that sea-level data are only combined from geographically restricted areas in order to avoid the influence of differential crustal movements (Tooley 1978). In turn, the extraction of coherent sets of sea-level data can provide information on differential crustal movements between regions that may be used to model the GIA process (e.g. Shennan 1989; Shennan and Horton 2002).
The database presents sea-level data with information spanning over 20 fields, covering fundamental variables such as location, altitude and age, coupled with important supporting data such as tidal information, type of dated material and the indicative meaning of the sample (see 'Fields' worksheet in the database file for details). These data have been extracted from published literature and a full set of source references is given in the database file.
We sub-divide the data into four classes based on the quality of sea-level information they contain.
Primary Index Points
These represent the highest quality sea-level data currently available from Ireland, and possess information on location, age, altitude and indicative meaning. Importantly, the relationship between the dated sea-level indicator and the environment in which it formed is quantified and error estimates for age and altitude are included.
Secondary Index Points
These offer lower-quality information on the former position of sea level than that offered by primary index points. They are derived from dated sea-level indicators, but one or more of the core variables is unquantified or associated with significant uncertainty. This may include: an absence of accurate levelling data; poor chronological control; limited accompanying microfossil analysis (ambiguous environment); an unclear relationship between the dated material and sea level; association with an erosive contact or stratigraphic unconformity. In addition, sea-level data sourced from archaeological evidence are included in this category since their indicative meanings are often poorly constrained.
***As a consequence of the uncertainties associated with these data, only tentative inferences on the position of former sea level can be made on the basis of secondary index points alone***
Limiting Dates (Type I)
These are derived from samples of known age and associated with a known environment that has no quantifiable relationship with sea level. As a consequence, it is only possible to infer whether sea level was above or below a certain altitude at a given time. These limiting dates are derived from material such as freshwater peat and in situ tree stumps.
***These data cannot be used to fix the former altitude of relative sea-level. They simple limit the upper or lower bounds of 'possible relative sea-level. NOTE: Many of the data presented in Carter (1982) are limiting in nature and should not be interpreted as sea-level index points.***
Limiting Dates (Type II)
These differ from Type I dates in that they are derived from material whose source environment is unclear or contested. Once again, these data only indicate whether sea level occupied a position above or below a certain altitude. However, the (possible) allochthonous nature of the dated material advises against attaching great significance to the age of the deposit, a point emphasized by numerous sea-level researchers (e.g. Kidson 1982). Instead, the date is seen as a ‘maximum age’ for the deposit since the dated material may have been deposited then reworked several times before finally occupying its current position in the stratigraphic column.
An example of Type II data are radiocarbon dates derived from wood that is not demonstrably in situ (e.g. recumbent trunks in estuarine silt). In addition, a number of accelerator mass spectrometry (AMS) dates derived from foraminifera contained with ‘glaciomarine muds’ are also included since, whilst potentially important sources of sea-level data, their origins are the source of much controversy.
***It is possible that, as new information becomes available, some of these (and indeed other) data may be re-assigned to a different class***
Key Database Fields
The database fields are closely related to the core attributes of a sea-level index point outlined above. Given the variable quality of the different data classes, not all entries possess data for the entire set of fields.
This information is provided in the form of site name, grid reference and latitude/longitude in decimal degrees. All primary index points have location information that is accurate to within 1km. For the purposes of displaying information, data are also assigned to a broad geographical area, where each area exhibits largely homogeneous RSL histories.
Most entries in the database have age information provided by radiocarbon dating. The database contains the sample laboratory code, radiocarbon age and calibrated calendar date calculated using CALIB 5.0.1 (Stuiver et al. 2005). Details are included in the database .xls file. Whilst some of the radiocarbon dates contained within the database are corroborated via pollen chonostratigraphic data, the vast majority have no such supporting evidence. Consequently, erroneous age estimates due to contamination may remain undetected. It is also the case that if erosion has occurred between a terrestrial–marine contact in the stratigraphic column, any age estimate for the date of transition between the two environments may overestimate the true age of the event. To guard against this, detailed microfossil analysis can be employed to demonstrate a continuous transition across the stratigraphic contact. This has been carried out for all samples contained within the primary index point tier. It should be noted that for the case of limiting dates, an erosive contact is not important since no attempt is being made to fix the time of a transition between a marine and terrestrial environment.
More than one levelling datum has been employed in research around the coastline of Ireland . In the database, all sample altitudes refer to Ordnance Datum ( Belfast ). Conversions were as follows: Irish Ordnance Datum (Poolbeg) lies 2.7m below Ordnance Datum ( Belfast ) which in turn lies approximately 0.03m below Ordnance Datum (Malin Head) (Admiralty Tide Tables 1997).
The indicative meaning of a sample describes its vertical position relative to the tidal frame at the time of its formation. Where available, the database provides information on the inferred indicative meaning, the nature of the evidence from which this is derived, and the local tidal parameters used in reconstruction. For primary index points, all indicative meanings are derived from a combination of lithostratigraphic and biostratigraphic data. For other data classes, the formation of freshwater peat is conservatively assigned to elevations above mean high water of spring tides (MHWST), whilst mean high water of neap tides (MHWNT) is used as the boundary between the formation of organic-rich intertidal deposits (e.g. salt-marsh ) from inter-tidal to sub-tidal minerogenic sediments containing shells.
Specific limitations are associated with the differing classes and fields of data, reflecting wider limitations associated with existing sea-level methodology. In many cases, the influence of these additional variables cannot be precisely quantified and, as a result, the uncertainties associated with age or altitude may be larger than those indicated by the error terms. Below, the key points to be aware of are highlighted, along with reference to publications which deal with these issues in more detail.
Location, data distribution & regions
The quality of a local record of relative sea-level change is dependent upon the accuracy, precision and distribution (in time and space) of sea-level data. This is highly variable around the coast of Ireland with the result that sea-level changes in certain parts of the country are still associated with considerable uncertainty (see Brooks & Edwards, 2006, for details).
The combination of sea-level index points from different sites into ‘regional curves’ assumes that inter-site differences in RSL change are not substantial. This assumption becomes increasingly prone to error as the areas become larger, or the individual records are analysed at higher resolutions. It will also be more suspect in areas with complex hydrographic regimes and coastal geometries since significant inter-site variability in tidal parameters (and hence indicative meanings) may be produced. The current 22 geographical areas reflect broadly similar modelled RSL histories derived from recent attempts to model the GIA process (Brooks et al., submitted). These regions are therefore suitable for input to geophysical models with similar, comparatively coarse, spatial resolutions. Conversely, care should be taken in extrapolating the general results from these broad regions to detailed, site specific studies.
The reliable interpretation of age data rests upon the assumption that the estimated age is correct, and the dated sea-level indicator reflects the timing of the sea level change of interest. This may not be the case if the sample has been contaminated by older or younger carbon, such as by rootlet penetration or inwashed material. In particular, it is important that the stratigraphic contact between terrestrial and marine sediments is not erosional or associated with a hiatus, since this will result in the overestimation of the change in marine influence. Recently, Waller et al. (2006) have questioned the reliability of SLIs based upon (apparently continuous) transgressive contacts. Their findings suggest that radiocarbon dates from the upper surface of peat layers should in most instances only be regarded as limiting ages for the deposition of the overlying clastic sediments, since peat growth often appears to slow down or cease well in advance of marine inundation. Other issues concerning the age of sea-level index points and dating material for sea-level reconstructions are given in van de Plassche (1986) and Edwards (2004).
Errors in altitude can be introduced during sample collection as a consequence of limitations in the precision of the surveying equipment used, uncertainties inherent within the benchmark network, and errors associated with the measurement of lithostratigraphic contacts in the field (Shennan 1982, 1986). An additional potential source of error comes from uncertainties associated with establishing the indicative meaning of the sample (see below). For further details on the calcuation of the vertical error term, see Brooks & Edwards (2006).
A further unquantified error may be introduced by post-depositional compaction of the sediment column, either under its own weight or as a consequence of subsequent loading by water or an overlying sediment burden. Whilst the potential influence of compaction on Holocene RSL records is long established and widely acknowledged (e.g. Jelgersma 1961 ; Terzaghi and Peck 1967; Greensmith and Tucker 1971a, 1971b, 1973 ; Tooley 1978; Heyworth and Kidson 1982; Shennan 1986 ; Allen 1996; Haslett et al. 1998; Shennan et al. 2000b; Edwards 2006), it has commonly been set aside due to the lack of a formal means for correcting for its influence (Allen 2000; Shennan and Horton 2002). Compaction serves to lower the reconstructed altitude of former RSLs and will be greatest where index points are established from thick intercalated sedimentary sequences. Index points should therefore be interpreted in light of the stratigraphic data supplied in the database.
Each primary index point contained within the database has an associated indicative meaning which has been constructed with reference to contemporary tidal characteristics. Since the majority of data points in the database are collected some distance from an established tide gauge, site-specific tidal parameters have to be extrapolated or interpolated from the closest available records. This will introduce an error term which may only be accurately corrected by on-site data-logging of tidal conditions.
It is our intention to update this database as new information becomes available. The basic standard for information is that it is published in a peer-reviewed work. Information from unpublished sources (e.g. PhD theses) offered by the authors will be considered if it can fulfill all the criteria required by the database fields.
Please let us know about any data that we've missed or that has recently become available, and we'll do our best to update the database as soon as possible. Email me.
For general references relating to sea-level reconstruction and included in the text above, click here (.pdf file).
A full list of references associated with the data presented in the database is included in the SL_Database.xls file.
The database file is standard microsoft excel spreadsheet called SL_Database.xls
This database was last updated November 2006, download the DATABASE here (XLS file 225KB)
Contact: email@example.com Last updated: March 30th 2006