Flolocene isostasy and relative sea-level changes on the east...

212

Flolocene isostasy and relative sea-level changes on theeast coast of England

I. SHENNAN,I K. LAMBECK,2 B. HORTON,I J. INNES,I J. LLOYD,IJ. MCARTHUR' & M. RUTHERFORD'

I Environmental Research centre, Department of Geography, (Jniversity of Durham,Durham DHI 3LE, UK (e-mail: [email protected])

2 Research schoot of Earth sciences, The Australian Nationar university,Canberra ACT 0200. Australia

Abst¡act: Analysis of sealevel data from the east coast of England identifies local-scale andregional scale factors to explain spatÌal and temporal variations in the altitude of Holocenesea-level index points. The isostatic effect of the glacial rebound process, including boththe ice (glacio-isostatic) and water (hydro-isostatic) load contributions, explains refional-

. .. scale differences between eight areas: c.20mrange at 8cal.ka¡p and,by cai..kasÞ.rela-tive seaJevel in Northumberland was above present, whereas in areas to the south relativesea level has been below present throughout the Holocene. Estimates for pre-industrial rela-tive sea-level change range from I .04 * 0. I 2tnrna-t in the Fenland to - i.30 + 0.6g mm a

-r(i-e. sealevel fall) in north Northumberland, although this may overestimate the current rateof sea-level fall. Isostatic effects will produce similar relative differences in rates of sea-levelchange through the twenty-first century. The data agree closely with the patterns predictedby glacio- and hydro-isostatic models, but small sysiematic differences alõng the east coastawait testing against new ice models. Local scale processes identified include differentialisostatic efects within the Humber Estuary and the Fenland, tide range changes during theHolocene, and the effects of sediment consolidation- These processes help explain the vária-tion in altitude between sea-level reconstructions derived from index poinis ta-ken from basalpeats and those from peats intercalated within thick sequences of Éiolocene sediments.

Analysis of data from estuaries and coastallowlands around Great Britain reveals differentrelative seaJevel changes owing primarily to thespatially variable consequences of glacio- andhydro-isostasy (e.g. Shennan 1983, 1989; Lam-beck1993a,b, 1995; Peltier 1998). These isosraticfactors combine with eustatic changes in oceanvolume to produce regional-scale rel¿tive sea-level changes. Great Britain lies beyond the limitof Scandinavi an ice at the last glacial maximum,but close enough to result in isostatic movementcaused by that ice-sheet. Although small incomparison, the British ice-sheet is the primarycause of the spatial variation of regional-scalerelative sealevel changes along the east coastof England. Both observations and model pre-dictions (e.g. Shennan 1989; Shennan et al. li95;Lambeck 1993a,b, 1995) in areas under thickice-cover from relatively small ice-sheets suchas the British ice-sheet typically record a fall-rise-fall relative sea-level from the Late Deven-sian to the present (Fig. 1). The age and alti-

tude of the early Holocene minimum and mid-Holocene maximum varying according to geo-graphic location relative to the centre ofthickest-ice. In contrast, areas close to the limit of the ice,and beyond, record a general trend ofrising rela-tive sealevel (RSL) throughout the Holocene.(Fig. 1).

Although the interaction of eustasy andisostatic factors produces the general pattern ofrelative sea-level changes various factors operateat the coast and within an estuary that influencethe registration of relative sea-level changes inthe sedimentary record. These local-scale factorsinclude modifications in the tidal regime alongthe estuary and the relationship between thefreshwater table and tide levels. Furthermore,changes in elevation of the sediment recording apast sea-level since the time of deposition mustbe taken into account. Such changes in elevationmay include consolidation due to the accumula-tion of overlying sediments and consolidationdue to land drainage.

l:f-î

it:È

From: S¡rP¡nqeN, I. & Auon¡ws, J. (eds) Holocene Land-Ocean Interaction and Environmental Chønge around.île Yorlh.Sea. Geological Society, London, Special Publications, 166, 215-298. I-86239-054-1100/$15.õ0 @ TheGeological Society of London 2000.

ät

Fig. 1. Site location map with inset of schematicrelative sea-level curves from northwest Scotland andsoutheast England based on glacio- and hydro-isostatic rebound models (after Lambeck 1995).

For each site the changetime r, and location g canmatically as

in RSL (AÇ,) atbe expressed sche-

A(.'r(", p) :AÇ*(r) * A(¡*(r,9)

* A€ro""r(z, p) (1)

where AÇ,,(r) is the time-dependent eustaticfunction; A{i"o(r, rp) is the total isostatic effect ofthe glacial rebound process, including both theice (glacio-isostatic) and water (hydro-isostatic)load contributions; and A{ro"ul(2, p) is the totaleffect of local processes within the estuary. Inorder to use observations from the sedimentaryrecord to reconstruct sea-level change the localfactors can be expressed schematically

Â6o""r(¡, ç): L€r¡a.G,ç) * LÊ""a?,ç) Q)

where Af,¡¿"G,ò is the total effect of tidal-regime changes and the elevation of the sedimentwith reference to tide levels at the time of deposi-tion, and A€""¿(r, tp) is the total effect of sedi-ment consolidation since the time of deposition.

This paper is one of a series arising mainlyfrom two projects (Project 316, ModellingHolocene depositional regimes in the westernNorth Sea at 1000-year time intervals, and Pro-

I. SHENNAN ET AL.

ject 313, Differential crustal movements withinthe river-atmosphere-coast study (RACS) site,Berwick-upon-Tweed to North Norfolk) withinthe Land-Ocean Interaction Study (LOIS).These two projects address different aspects ofreconstructing relative sealevel change andcrustal movements since the last glacial max-imum (e.g. Shennan et al. this volume, in pressa,b). Other LOIS projects contributed data tothis paper (see Andrews et al. 1999; Brew et al.1999; Metcalfe et al. 1999; Orford et al. 1999;Plater et al. 1999). Previous modelling studies(e.g. Lambeck 1993a,b, 1995, 1996: Lambecket al. 1998) demonstrated the general trends inHolocene RSLs for different parts of the eastcoast and indeed the previously available datafrom eastern England were used to develop andvalidate the earth and ice models. These studiesprovide robust parameters for the earth model(e.g. Lambeck 1996) and the eustatic function(Fleming et al. 1998). Lambeck et al. (1998)updated the model of the Scandinavian ice sheetcompared to the earlier papers and a separatepaper arising from the current LOIS projectsdiscusses a modified model of the British ice-sheet (Shennar et al. this volume). Therefore,the present paper does not discuss the detail ofthe ongoing glacio- and hydro-isostatic model-ling. The major discrepancies between observa-tions and predictions occur mostly in those areasunder the thickest ice or where ice limits andvolumes are less well known, and for the LateDevensian and early Holocene times, i.e. beforeScal.kanp, for which there are few data pointsfrom the east coast of England.

The aims of this paper are threefold. Firstly todistinguish and quantify the local-scale factors,A(,ia"(r, rp) and AÇ"a(r,p), for separate sectionsof the east coast of England between Berwick-upon-Tweed, north Northumberland, and northNorfolþ secondly, to quantify the regional-scaledifferences (i.e. between the separate sectionsof coast) in RSL caused by the interaction ofeustasy and isostasy; and finally, to quantifythe pre-twentieth century rate of sealevel risefor the different coastline sections and comparethem to the Intergovernmental Panel on Cli-mate Change (IPCC) predictions of future sea-level change.

Data

A number of recent projects in eastern England(e.g. Waller 1994; van de Noort & Ellis, 1995,1997; Long et al. 1998) and the LOIS pro-gramme have greatly enhanced the databaseavailable for reconstructine Holocene sea-level

I

;

Tif:,

iìl:1.

iíÍii:.

itx

ISOSTASY AND RELATIVE SEA-LEVEL CHANGE 27'/IFË

change since the analyses ofShennan (1989) andLambeck (1995). Collectively, they provide alarge dafabase of reliable index points forreconstructing Holocene sealevel changes. Theprocedures for evaluating individual samples asrelative sea-level index points are routinelyexplained in many publications (e.g. Shennanl986a,b; van de Plassche 1986; Long et al.1998). In brief, Iithostratigraphic and biostrati-graphic data are used to quantify, with an errorterm, the waterlevel at which the sample formsin relation to tidal regime and identify thetendency of sea-level movement represented bythe sample. A positive tendency is deflned asan increase in marine influence and a negativetendency is a decrease in marine influence. Theage of the indicator comes from calibrating the'' radiocarboh age of the sample, expressed asthe mean calibrated age and the 95o/o probabilityrange (from 'method A' of Stuiver & Reimer1993). The pollen data also provide a coarse-scale chronological check on the accuracy of theradiocarbon ages.

Much of the sealevel data used in thisanalysis existed before the LOIS project began.The ¡adiocarbon database held at the Environ-mental Research Centre (ERC), University ofDurham, consisted of 790 validated sea-levelindex points of which 225 came from the eastcoast between Berwick-upon-Tweed and northNorfolk. The temporal and spatial distributionsof the index points were uneven and limited thevalue of earlier analyses. Most of the recordedindex points were of mid-late Holocene age, aîimbalance reflecting the field distribution ofsurviving sediments suitable for field investiga-tion. Sediments younger than about 2cal. ka¿phave been destroyed by recent agricultural andindustrial development in the coastal zone. Indexpoints of earlier Holocene age often lie at verygreat depth or seaward of the current coastline.In addition, other important factors such assediment compaction and tidal-range variationswere not quantified in the existing database.Therefore the sampling strategy within the LOISprojects aimed, where possible, to fill gaps in thegeographical distribution (Fig. 2) of sea-levelindex points and to extend the temporal range ofthe data set (Fig. 3). The scatter shown in theage-altitude distribution of sea-level index points(Fig. 3b) indicates the magnitude of the sum ofthe spatially dependent components, A€i*(", p),A6,i¿"(", p) and A€."a(r, p), that are to beexplained.

In general the spatial improvement of the dataset and its extension towards the earlier Holo-cene was very successful. The database nowavailable from the east coast of Eneland

Fig. 2. Location of sea-level index points from the eastcoast of England and western No¡th Sea. The dotsrepresent study sites; each site may contain several sea-level index ooints.

between Berwick-upon-Tweed and north Nor-folk comprises 388 sealevel index points quan-titatively related to a past tide level together withan error estimate, and a further 7l data pointsthat provide limits on the maximum altitùde ofthe contemporary local sea-level. Less successfulwas the recovery of samples more recent than2cal. kanp, due mainly to the great scarcity ofsurviving deposits of that age. New data

ililrli;1,

S.i

:tr

iT'

fi#il

278 I. SHENNAN E7: AL.

+ Pre-LOIS

O LOIS

80

r

IPre-LolS

trLOIS

0 3 6 9 1 2ka cal BP

1 0

-10

g -roJ

fl. -zo

-50

-60

40

1 41 21 0

Thousand calibrated years BP

Fig. 3. Age-altitude plot, and frequency distribution (inset), of sea-level index points f¡om the east coast ofEngland and western North Sea.

collection also targeted basal peats, which, beingless susceptible to compaction, should providemore reliable altitude data and allow an assess-ment of compaction effects on intercalated peatsof comparable age. Where possible, sampledcores were placed within a sedimentary contextby the addition of transects of cores, prior to theselection ofthe core to be analysed. The originalsampling design identified six geographicalregions, each with different pre-existing dataand thus sampling requirements. The finalsampling programme also had to conform tothe requirements of the other LOIS projects (seeother papers in this volume).

Examples of data collected

In the following section we present litho-,chrono- and biostratigraphic records from twosites that illustrate the types of sea-level indexpoint and the limiting data used to reconstructrelative sea-level changes. Pollen and spores areshown as percentages of a total land pollen sumof at least 300, which does not include aquaticsand spores. Foraminifera and diatom data areshown as numbers counted, as at some levels onlysmall though significant totals were achieveddue to poor preservation. Stratigraphic symbolsfollow a simplified Troels-Smith (1955) scheme.

Bridge Mill (BM95l7A), northNorthumberland

This site in northern Northumberland (Fig. 1)lies in a small area of coastal plain belowl0mOD to the west of Holy Island. Coringtransects across this area show the Holocenesediments to consist almost everywhere of silt,clay and sand, but localized organic intercala-tions occur within the clastic sequence in asheltered area in the lee of the upland to thewest. A thin upper peat occurs in several cores,and impersistent lower organic units are alsopresent, although usually with a high clayfraction. Two distinct peat layers are bestpreserved at the hillslope foot in core BM95/7A fNU 0407 45671, where the lower peat restsupon sand, which could not be penetrated. Thetwo organic units in BM95l7é- are overlain andseparated by slightly organic blue-grey silt clay,the intercalated clay layer including a highlyorganic horizon. The lithostratigraphy is shownas part of Fig. 4 which presents detailed radio-carbon and biostratigraphic data from the core.

Pollen levels were selected near lithostrati-graphic unit boundaries, which representchanges in depositional environment, as well asat intervals throughout the units. Representativelevels from each of the lithostratigraphic unitswere analysed for foraminifera and diatoms toprove their depositional origin and, ifintertidal,

3462.3802 ' 80

90

4426.4827.1 0 0

+4.50m OD -

1 1 0

120

1 S 06002-6266 .

140

1 5 0

+4.00m OD .

1 6 0

7014-7281 .

170

Shtubs / - Herbs

Fig' 4' Bridge Mill, Northumbe¡land: microfossil diagram, with interpretative log at right edge. Pollen expressed_ as per cent total land pollen; spores as per cent E(landpolien * spores); foraminifera as raw counts; f,iatoms as ¡arv counts and grorrped ãccording toiali"ity ;la;;;;ñlyhutouour, mesohalobous and oligohalobous. Calibratedradiocarbon ages, altitudes (metres above ordnance Datum (oD)) ana ãepttr (centimetrei) down-córe shown to the left of the lithology column. The sediment legend isdrawn according to Troels-Smith (1955).

rrrTTl I-rl T|TTTI nTrTTl nnFlTrTTTt rrT'Tl nnFTTTTTTrl nFTT'-Ta nrrrl rrl FTì rïrl |-rt Ti-Trl20 40 20 40 20 40 20 40 20 40 20 40 60 15oO 30 30 150 20i - -= ì= i F . - u . . - : ¡ , - r

lrf rlil cray a sin t':{:l-l-J Hum¡fled FiìlÏ'!ÌETil Herbaceous ll*il s"n¿Lr.l.t L_n:::! peat tiiiiiiiiiiiiil p""t r.:.:,:.:.:.r

[:i'] :i;:l 'r:ä1¡1ikriij'1.;i:l'1i1lF-;;li

F luv ia l

Reedswampupper

Sa l tmarsh

l,lii[tr:',,o=r;:t:,]:,*,r,:il:,.t,j|,,

aU)ÊÞv)

zF 1

ñ--l

?f¡JU)rrJ

-lrI

F]-o

zG]

LOWerSa l tmarsh

UpperSal tmarsh

N!\o

Tt =-#.7

280 I. SHENNAN ET AL

to assess their reference waterlevel within the changing conditions of water-level and salinity.tidal cycle. The lower peat is shown to be mid- Table I illustrates how these data are used toHolocene in age by a rise in Alnus pollen. reconstruct past RSLs. All four contacts are'îhe

Alnus pollen and a fall in Ulmus above the validated sea-level index points with a knownupper contact supports the radiocarbon date of altitude, date and indicative meaning, which is5290+608P (6002-6266cal.nn) at *4.25mOD. linked to a past tide level. The indicativeThe date of 6285 +658P (7014-7281cal.nn) at meaning, of a coastal sample is the relationship+ 3.89 m OD on the base of this peat is also sup- of the local environment in which it accumulatedported by the pollen data. Gramineae dominates to a contemporaneous reference tide level (vanlocal wetland taxa through this peat and into the de Plassche 1986). The indicative meaning canoverlying clay. Saltmarsh pollen such as Cheno- vary according to the type of evidence and it ispodiaceae and Plantago maritima are present comrnonly expressed in terms of an indicativethroughout, but increase in frequency and range range and a reference water-level. The former isin the clay. Foraminifera test linings are pres- a vertical range within which the coastal sampleent in the pollen preparations and in the sep- can occur and the latter a waterlevel to which .a ra te |ypreparedsamplessa l tmarsh foramin i fe ra theassemblage isass igned, fo rexample ,meanJadammina mauescens, Miliammina fusca and high water spring tide (MHWST), mean tide " .:

' T r o c h a m m i n a i n f l a t a a r c e o m m o n i n . b o t h p e a t ' . l e v e l ( M T L ) ; e t c . . ' ( v a n d e P l a s s c h e . 1 9 8 6 ) . , T h e - . - , : , , ' i ' \ . . . . ' . : , : iand clay. Diatom preservation was poor except base of the lower peat records saltmarsh peatfor a single sample, from the upper part of the formation at MHWST and the upper-threepeat, with a varied saltmarsh assemblage domi- dated contacts record changes between clasticnated by Diploneis interrupta. This Iower peat- and organic sedimentation occurring aroundclay couplet represents intertidal deposition, MHWST. The lower peat records positive sea-passing up-core from upper to lower saltmarsh. level tendencies and the upper peat records

The upper peat and surface clay are less negative then positive tendency.clearly saltmarsh, containing no foraminiferaand a few freshwater diatoms. Although the fullrange of saltmarsh pollen appears, increased South Farm, Sunk Island (HMB9),Cyperaceae and Gramineae frequencies sug- Humberside :,gest deposition in a reedswamp with regular -.,tidal input. The lower and upper contacts of the This site exemplifies stratigraphic data that ";'upper peat are dated 4105+558P (4426-4827 cannot be referred to a past tidal level and so i.cal.nr) at +4.59mOD and 3360È60nv (3462- are not valid sea-level index points per se, but li.3802caI.BP) at +4.77 mOD. which nevertheless provide valuable limiting +

The sediments at Bridge Mill represent data that constrain the position of past sea- :''

deposition in the upper intertidal zone under levels by providing maximum altitudes for

Table 1. Sea-level index points from Bridge Mill

Laboratory code AA24223 é^A24224 ¡^424225 ^424226

raC age* 1o (a)Calibrated age (cal. a nr)

Max.MedianMin.

Altitude (mOD)Reference water-level*,tIndicative rangefTendencyChange in RSL from present$

3360 + 60 4105 + 55 5290+60 628s X65

3802 4827 6266 728r3606 4s65 6101 12003462 4426 6002 70t4+4.77 +4.59 +4.25 +3.89MHWST -0.20 MHWST +0.09 MHWST -0.20 MHWST+0.20 +0.20 +0.20 +0.20+ - + ++2.57 +0.20 +2.10+0.20 +2.05 +.0.20 +1.49 +0.20

Change in relative sealevel (RSL) is calculated as altitude minus the reference water level.*The reference water-level is given as a mathematical expression of tidal parameters plus/minus an indicativedifference. This is the distance from the mid-point of the indicative range to the reference water-level.tLocal mean high water spring tide (MHWST) is 2.40mOD.

f The indicative range (given as a maximum) is the most probable vertical range in which the sample occurs.$ The RSL error range is calculated as the square root of the sum of squares of altitudinal error, sample thickness,tidelevel error and indicative ranee.

H M B S

.òø- ò\'

^ o '

. ! t ^ot ¡9. ,

1525

1 530

't 535

Iriril[iiiilL'rtrI

tirirl

lf$l

ffil trrrt ll trtrt ll trtrt I

z806-8069 r I s4o

1 545

1 1 . 7 5 m O D '

1 5 5 0

FTn rrfll]

.ù\oQ,ø' Q\

ITTFI

9397-9646 |

1 s70- 1 2 m O D -

1 5 8 0

t 5 8 s

Fig' 5' south Fa¡m' sunk Island (HMB8), Flumberside: microfossil diagram with interpretative log at right edge. pollen expressed as per cent total land pollen; spores as percent Ð(land pollen ¡ spores); foraminifera as raw counts; diatoms as rai counts and grouped ".roi¿ing

t-o rãiirity classes polyhalobous, mesohalobous and oligohalobous.Calibratedradiocarbonages,alt i tudes(metresbelowordnancedatum(oD))anddeptñ(centimetres)¿oî, ' -co."shôwntodrawn according to Troels-Smith (1955).

1 5 9 0 il-T.l f'fl'F Tl-Fl FFI lll-TTl flnnnniîfrlT'l'n nTTrIrì nnfln Tn nnnnrn fn nTT.in Tn |-,Fl20 20 40 20 20 20 20 40 20 20 40 60 80 20 20 20 20 40 60 20 20ffiiÌc'"ras'n EFI ¡""¡n'"0 llÏl."no ffirr.u"

f ; r . r . i r , f , : . : . .G . , i

Re€dswamp

Lacust¡¡ne

Fluv¡al

282 I. SHENNAN ET AL.i

highest tidal levels. HMBS L'1A257017601 lieson Sunk Island, a reclaimed area near thenorthern shore of the outer Humber Estuary(Fig. l). More than l0 m of marine sands cover alayer of organic clay which contains Phragmitesremains. This clay overlies a peat, organic clay,limnic mud and silt clay sequence, which rests onthe pre-Holocene sand diamicton. The contactbetween the peat and the Phragmites-richorganic clay is very sharp and represents anerosive break in the sediment pile. The lowerpart of the succession is shown in Fig. 5, whichpresents the biostratigraphic analyses.

The limnic mud and silt clay units contain apollen record that represents transitional LateDevensian and early Holocene vegetation com-munities. The lowest level near the base of thelimnió'mud is characterized by Betúla and Juni-perus, with some Gramíneae and Cyperaceae.Pediastrum algal colonies support the interpre-tation of a lacustrine origin. At the start of claydeposition, above the limnic unit, many tundra-type open habitat herbs like Helianthemum,Rumex and, Thalictrum appear. Within the clayand in the peat above, the open ground herbflora is replaced firstly by Juniperus and Betulaagain, followed by Corylus and then deciduoustrees Ulmus, Quercus and Alnus. This successionis typical of the sequence from Late DevensianInterstadial to Loch Lomond (Younger Dryas)Stadial and then to Holocene. The changefrom clay to peat formation is dated 8555 +65ny (9397-9646cat. ne), which agrees with pine,birch and hazel pollen assemblages from otherregions. The peat below the sharp upper contact

Table 2. Limiting dates from Sunk Island

with the overlying clay is dated 7145 +60 ne (7806-8063 cal. nr) at - I 1.67 m OD, whichagain supports the mid-Holocene pollen datawith high Alnus. A rich marine diatom and salt-marsh foraminiferal assemblage occurs withinthis upper clay, suggesting a marine origin, butthe peat and other sediments below this marineclay arc barren of diatoms and foraminifera.There are no saltmarsh pollen indicators in theupper level of the peat, which apparently formedunder completely freshwater conditions. Thesebiostratigraphic data confirm an erosive hiatusbetween the peat and marine clay, of unknownduration. This hiatus means that the datedpeat-upper clay contact cannot be accepted as asealevel index point related directly to a pasttidal level. Nevertheless, the age and altitude(Table 2\ give á limitiñg valuè becàúse'at thattime freshwater peat was forming and tidalinfluence must have operated at or below thataltitude. Such peats may form in a wide verti-cal range depending on local palaeogeography,especially their relationship to active tidal chan-nels and the groundwater table. Therefore, theindicative range for these dates is from aboveMHWST, where freshwatel peat formation iscontrolled by tidal influence, to just below MTL,where groundwater level is the controlling fac-tor (Godwin 1940; Shennan 1982). Subsequentmarine inundation truncated the peat profile,and then laid down marine clays unconformablyover the peat.

The analyses described in the following sec-tion use both sea-level index points of the typesdescribed from Bridge Mill, with quantified

Laboratory code AA2558t é^425582

taC age|Lo (a)Calibrated age (cal. a rr)

Max.MedianMin.

Altitude (mOD)Refe¡ence water-level *,t

Indicative rangefTendencyChange in RSL from present$

7145 +60

806379237806-tt.67>[MHWST+MTL|2+1.58Limiting< -14.61 + 1.58

8555 + 65

964694819397-i l .95>IMHWST+MTL]/2+ 1 . 5 8Limiting< -14 .89 + 1 .58

.F.':

ä

Ii:i.¡:..l

Change in relative sea-level (RSL) is calculated as altitude minus the reference water-level.* The reference water-level is given as a mathematical expression of tidal parameters plus/minus an indicative difference. This is the distance from the mid-ooint of the indicative ranseto the reference water-level.

tLocal mean high water spring tide (MHWST) is 3.32mOD.{The indicative range is the most probable vertical range in which the sample occurs,although the sample could occur above that range.

$ The RSL error range is calculated as the square root of the sum of squares of altitudinalerror, sample thickness, tide level erro¡ and indicative range.

tiii

ISOSTASY AND RELATIVE SEA-LEVEL CHANGE

flr

relationships to tide levels (e.g. Table l), and thelimiting type of data points illustrated fromSunk Island (e.9. Table 2).

Analysis

Differences in RSL along the east coast due toA€"",(") and A(¡"o(r, g) produce a continuumwithin the range shown in Fig. 3b. In order tosummarize the information in graphical formthe data are grouped into geographical units andplotted against a single curve for that region(Fig. 6). The curve is the prediction fromLambeck (1995) for one location in the regionand does not indicate any predicted within-region variations in RSl.. .

The ice model and earth models of Lambeck(1995) predict RSL slightly below the observa-

283

tions from Northumberland, the Tees Estuary,the Lincolnshi¡e Marshes and, to a lesser extent,north Norfolk and the Humber. For the Fen-land the predictions lie within the upper range ofthe observations. Lambeck et al. (1998), Flem-ing et al. (1998) and Shennan et al. (this volume,in press a, ó) describe improvements in the iceand earth models and the eustatic factor andthese are not discussed in detail in this paper.Lambeck et al. (1998) discuss changes to theScandinavia ice model and Shennan et al. fthisvolume, in press a, b) evaluate earth-modelparameters and revisions to the ice model forGreat Britain, but at this stage they cannotidentify a unique solution for the ice model.Furthermore, modelling of tidal ranges in theNo¡th Sea.during the Holoce¡e (Shennan et al.this volume) predicts changes of the oppositesign needed to produce a better fit between the

50

-5

-20_25-30

l.lorthum berland

+f¡þ.

\ *"!F

+ observat¡ons

- pred¡cted

2 4 6 8

Thousand calibrated yE BP

0

-10

-20-25-30-J5

-40

Tees Estuary

--t

i,liì1 ;

tiijì

¡t;Lincolnsh¡re Marshes

North Norfolk

2 4 6 8 1 0

Thousand cal¡brated yrs BP

Fig. 6. Relative seaJevel observations (validated index points) and predicted sea-level curves for the initialsix regions.

d -tu-20

: - l n

_ 1 5

-10

.15

-10

Hum ber Estuary

F

+

284 I. SHENNAN ET AL.

H:-F

Lambeck (1995) predictions and the observa-tions. Also, all of the previous analyses do notconsider the possible effects of sediment con-solidation and tidal-range changes through time.Paul & Barras (1998) illustrate how sedimentconsolidation can be estimated, but this requiresquantitative information on the lithology of thesequences above and below the dated sample,which is not available for a large majority of thesea-level index points.

Therefore, we have adopted the followingexploratory approach based on analysis ofthe residuals between the observations and thesummary relative sea-level curve. For each geo-graphical region in the following section the sea-level predictions from the model of Lambeck(1995) are summarized with a best-frt poly.nomial. The linear term is then modified toproduce a best-fit solution to the index pointsfrom basal peats only, subject to the conditionthat the same solution does not conflict withthe set of limiting dates. The reasoning for start-ing from a solution for the basal peats is thatthey are probably less influenced by sedimentconsolidation than index points from peatsintercalated between thick Holocene clastic sedi-ments (e.g. Jelgersma, 1966; van de Plassche1986; Törnqvist et al. 1998). Where seaJevelindex points on basal peats are taken from thebottom of the peat, compaction is not a factor.If sea-level index points are taken from the topof the basal peat, however, these too will besubject to some compaction.

Modification of the linear term has a similarnet effect for the areas and time periods underconsideration to changing an earth- or ice-modelparameter (e.g. Lambeck l993a,b) or the effectof a change in tidal range through the Holo-cene. The distribution of residuals against timeevaluates the validity of changing the linearterm. Comparable analyses were undertaken foreach region. To illustrate the approach we pre-sent details of the analysis of the Humber dataand then provide summary illustrations from theother areas.

Detailed example: the Humber Estuary

In relation to its size very little is known aboutthe Holocene evolution of one of Britain's mostimportant estuaries (Gaunt & Tooley 19741' vande Noort & Ell is 1995, 1997;Longet a/. 1998). Asa result almost 40 cores were allocated to theHumber, distributed along the estuary and in theseveral major river lowlands that drain into it.Thirty-six new index points were added to thedatabase and the record now extends to almost

+

- RSL mdel

^ Liriting dates

o Basal ¡ndex points

I htercalated ¡ndex [þints

+l , {' - 1

-5

0 2 4 6 8 1 0 1 2thousand €librated years BP

Fig. 7. Humber Estuary relative sea-level index points.Error te¡ms are calculated as described in the text foreach index point, but only appear when they extendbWlnd the size of the symbol

,: .

8cal.asp. Further details are given in otherpapers in this volume (Andrews et al.;Rees et aI.;Ridgway et a.l.;Ìl|l[efcalfe et al. this volume).

Regional factors.The plot of all validated index points from theHumber Estuary conflrms the upward trend ofHolocene RSL typical of an area at or beyondthe margins of the last British ice-sheet (Fig. 7).The seaJevel predictions of Lambeck (1995) forthe Humber Estuary are modifled to producea revised regional RSL curve (compare Figs 6and 7). Departures from the revised regionalRSL curve (Fig. 7) suggest the importance oflocal processes, such as sediment consolidation,changes in palaeotidal range or within-regiondifferences in the isostatic effects. Determiningthe relative importance of each of these isdifficult at this stage in the analysis, but carefulexamination of the residuals (i.e. the differencebetween the RSL summary curve and observa-tions) provides an indication of their potentialimportance.

Local factors

Palaeotidal changes. Major changes in coastalconfiguration have occurred during the Holo-cene as changes in the rate of relative sealevelrise, sediment supply and catchment inputs ofsediment and water have varied. Accompanyingthese will have been changes in tidal range,caused by variations in tidal prism as well asestuary configuration. As a first analysis, the RSLsummary assumes no such changes and is basedon present tidal variations within the Humberestuary. This hypothesis is now tested, initially

ti:,

ii,',

li rl

i::l l

ü:a-l

l:¡Í¡tä¡.:ì::i:ri

il

o 2 4 6 8 1 0

fhousand callbntdl yeaÉ BP

Fig. 8. Humber Estuary relative sea-level index pointsf¡om basal peats: inner and outer estuary.

by dividing the sealevel index points into thosefrom the inner estuary, west of the HumberGap and outer estuary, east of the Humber Gap(Fig. 8).

Index points younger than 6cal.kanp fromthe outer estuary plot above the regional rela-

285

tive sealevel curve, whereas those from theinner estuary plot below (Fig. 8). This contrastmust be qualified by the clustered distributionof the data through time, and more data arerequired to test this hypothesis further. Never-theless these differences suggest that palaeotidalchanges or differential isostatic movements mayhave occurred within the Humber Estuary.

The separation into inner and outer estuarydata is an oversimpliflcation because as the tidalwave penetrates an estuary it is modified bychanges in width and depth, river flow andincreased friction, which together cause non-linear tidal distortions of the higher tidalharmonics. In the Humber Estuary this resultsin an increase in the altitude of tide levels upestuary. This relationship is used in the subse-quent analyses by using the up-estuary distance,from an arbitary reference point, Grimsby,based on our reconstructions of palaeogeogra-phy at the time of sediment deposition. Thisstill does not take into account possible changes

4

Ê 2

i oE - 2ê ¿

€&

4

0

4

2

0

O 2 ¿ l 6 8r.iit:.i :

iiÈ;,'l

d b E P

Depth ot overturden

o 2 4 6 8 1 0d b B P

Depth of overburden

-50 0 50 100dlrhB uÞ sûÐ{h}

Depth to båsc4

o

4

o 5 t o 1 5dlmt(ñ'

Tot¡l lh¡ckneEg

0 5 't0 15dtut(ú)

Total thlcknrss

0 5 1 0 1 5dlffit(ñ) 3

irl

4

: ñE - 2E

o 5 l o t 5rd¡ff i i(m)

0 1 0 2 0.d¡ff i |fñ)

Fig. 9. Humber Estuary: scatter plots for sealevelindex points from basal peats showing therelationships between residuals (observed versusmodelled relative sea level) and five parameters(age, distance up estuary, depth of overburden,depth to base of Holocene sequence and totalthickness of Holocene sequence) that may indicatelocal-scale processes. Table 3 gives the correlationcoefficients.

Fig. 10. Humber Estuary: scatter plots for sealevelindex points from intercalated peats showing therelationships between residuals (observed versusmodelled ¡elative sealevel) and five parameters(age, distance up estuary, depth of overburden,depth to base of Holocene sequence and totalthickness of Holocene sequence) that may indicatelocal-scale processes. Table 3 gives the correlationcoefficients.


0 50 100d.¡EqPáÙryûn)

4

¡E - 2

Ocptñ to b¡se

FF -*lo

*,-1,r-r 2

rE


in tidal prism through time due to the chang-ing sea-level, bathymetry and palaeogeographywithin the Humber Estuary.

Vy'hen the residuals are plotted against dis-tance there is a strong relationship for basal(Fig. 9) and intercalated (Fig. l0) peats. The cor-relation coefficients (ruasar : -0.61; r¡r¡erç¿1¿¡s¿ :-0.43) are above the critical value at both the5 and lVo significance levels (Table 3). Theresiduals for both basal and intercalated sea-level index points have weaker positive relation-ships with age (Figs 9 and l0). One hypothesis toexplain the trend with age is that tidal rangewithin the whole estuary was less in the pastthan at present. The trends seen with both ageand distance suggest that in the past the tidalrange in the inner estuary had decreased to a

. greater extent .-than the outer ,estuary com-pared with the present day (i.e. the increase intidal amplitude up estuary seen at the presentday was reduced in the past). This suggestionis not unexpected, given the progressive inflllingof the Humber Estuary during the Holocene.An additional consideration is differential iso-static movement within the estuary. The isolinesof predicted mean sea level for 10, 8 and7 r4Ckasp lie almost north-south across theHumber Estuary (Lambeck 1995) and an elementof the trend of residuals with distance up estu-ary could result from such isostatic movement.

In order to discriminate between differentialisostatic movement and tidal range changesthrough time within an area that lies across theisolines of isostatic model predictions (e.g. Lam-beck 1995) it is necessary to analyse residualsfrom RSL predictions based on the geographicposition of each index point, rather than theRSL summary for the large area used here. Untilthe revised glacio- and hydro-isostatic modellingdescribed earlier is completed this preferredapproach is not possible.

Sediment consolidation. Following deposition,sediment consolidation will lower index pointsfrom their original elevation and, unless cor-rected for, will lead to an over-estimate of therate and magnitude of relative sea-level rise inthe Humber Estuary. The effects of consolida-tion can be especially severe for index points thatare found above and below considerable thick-nesses of Holocene sediment (such as those fromintercalated peats). This problem is reduced, butnot entirely removed, by using index points frombasal peats, which rest at, or close to, the baseof the Holocene sediment column. Subdivisionof the index points into basal and intercalatedpeats provides an initial assessment (Fig. 7), andsuggests that the influence of sediment consoli-dation may indeed be signiûcant. Most of thesamples from intercalated peats lie below those

j

Table 3. Pearson's product-moment correlation cofficients

Area Typet No Critical Critical Agevalue value(0.05) (0.01)

Distance Depth of Depth to Totaloverburden base of thickness

Holocene ofsequence Holocene

sequence

NorthumberlandNorthCentral

South

Tees

Humber

Lincolnshire Marshes

Fenland

North No¡folk

6 0 .716 0 .711

6 0 .716 0 .71

L6 0.479 0.60

2t 0.4223 0.4240 0.3011 0 .55l7 0.4685 0.21

r08 0.2017 0.46t4 0.50

-0.s8 0.62-0.79 -0.56

0.01 -0.93-0.70 0.07

0.2s -0.66-0.68 -0.4s-0.14 -0.48-0.48 0.20-0.52 -0.51-0.09 0.27-0.65 -0.41

0.25 _-O.rt-0.48 -0.87

BIBIBIBIBIBIBIBI

0.830.83

0.830.830.590.740.540.540.390.680.s80.270.250.580.62

-0.51 -0.90 0.34-0.15 0.60 0.s7

0.16 -0.87 -0.970.76 0.36 0.26

-0.19 -0.48 -0.54-0.03 -0.34 -0.28

0.01 -0.s3 -0.s90.42 -0.61 0.270.08 -0.43 -0.230.37 -0.60 0.280.54 -0.12 -0.23

-0.19 -0.06 -0.35-0.04 0.09 -0.37-0.09 0.1 I -0.28-0.30 -0.01 -0.81

?1

l;,+!i

iii'

*Values in bold exceed the critical value at the 0.05 significance level (two-tailed).

t B, basal peat sealevel index points; I, intercalated peat sealevel index points. Basal index points include samplesfrom the top and bottom of Deats.

ISOSTASY AND RELATIVE SEA-LEVEL CHANGE 287

from basal peats of the same age. This is notwholly unexpected and invites further analysis.

In the absence of the detailed lithological dataneeded for quantitative assessment of consoli-dation (e.g. Paul & Barras 1998) its effects onindex points from peats may be considered to bedependent upon three more easily availableparameters. This approach does not model theconsolidation process, but gives an indicationof the net effects. The parameters available foreach index point are: (a) the thickness of sedi-ment overburden; (b) the depth of sedimentbelow to the base of the Holocene; (c) thethickness of the whole Holocene sequence(Figs 9 and 10). For the basal peat index pointsonly the depth of sediment to the base of theHolocenç. is signiûcant (Table 3),, correlationcoemcient rbasal: -0.48. The positive correla-tions with the other two variables are neitherstatistically significant nor of the correct sign tosuggest consolidation effects.

For samples from intercalated peats, depth ofsediment to the base of the Holocene and thethickness of the Holocene sequence suggeststrong empirical relationships with the magni-tude of residuals (Fig. 10 and Table 3). Thecorrelation coefficients for both parameters aregteater than the critical value at the 5 and l%osignificance level suggesting further research inthis area is justified. For example, at this stage ofanalysis no account has been taken of thevaria-tion in sediment types, including the proportionofdifferent grain size distributions, organic con-tent, water content, or of drainage histories.

Summary: regional sea level, sedimentconsolidation and tidal change within theHumber Estuary

The regional sea-level curve (Fig. 7) provides asummary of the Holocene sea-level data fromthe Humber Estuary. It is suggested that sedi-ment consolidation, tidal changes and differen-tial isostatic movements have played importantroles in producing the scatter of relative sea-leveldata around the regional trend. Establishing therelative importance of these factors is not pos-sible at this stage and requires further modeliingof each process.

This procedure of data analysis is nowrepeated for the data collected from the fiveother study areas with the following amend-ment. Graphs of the relationships betweenresiduals and local processes are only shownfor those exceeding the critical value at the 5o/osignificance level (Table 3). These relationships,

therefore, reject the null hypothesis suggestingthere is no empirical relationship between theparameters.

Northumberland

This is a critical region, as it is the mostnortherly in the study area and so is likely toshow most clearly the effects of differentialcrustal movement. Only 14 pre-LOIS validatedindex points were available (Plater & Shennan1992; Shennan 1992). Twenty-eight new indexpoints have been produced with the databasenow extending to pre-8 cal. ka nr (Shennan et al.in press å). These index points. come fromindividual sites, with almost 60 km between themost southerly and the most northerly. Geo-graphically the sites fall into three clusters,labelled here north, central and south (Fig. 1,although no index points come from the mostsoutherly parts of the Northumberland coast).Previous investigations illustrate a systematicincrease in the altitude with distance northwardsfor seaJevel index points of the same age (e.g.Plater & Shennan 1992; Shennan 1992; alsoShennan et al. in press ó) and the threefolddivision is used in the following analyses. Noneof the data come from large distances upestuaries, so the distance parameter used isdistance from the northernmost point. Sincethe Northumberland coast lies along the axis ofdifferential uplift any trend against this distanceparameter will in part result from within-areadifferential movement.

Despite the lack of index points for the last3 ka in all three Northumberland areas, thepresence of intertidal clastic sediments abovepresent high tide level demonstrates a fall inRSL since the time of the youngest index point.The distributions of the data from the wholeof Northumberland constrain the form of thesummary RSL curves (Figs ll, 13, 15) andthe three differ only in the value of the linearterms in the polynomials.

The summary curve for NorthumberlandNorth (Fig. ll) indicates a mid-late Holocenemaximum around 2.5m above present. Thebasal peat data come from only two sites, sowhile the significant correlation with distancenorth-south may indicate differential crustalmovement any local effects changing tide levelsat either site will also contribute to the distribu-tion in Fig. 12. The significant correlation withdepth to base of the Holocene for index pointsfrom intercalated peats indicates a sedimentconsolidation effect (Fig. 12 and Table 3).

;:ri

i,:i

ii: t

a:rl

f!::::.,ni:l

lirirf

a+-RSL model

A Limiting dates

O Basal index points

I lntercalated index points


Fig. 11.'Northrimberland, North: rêlative sea-levèl iidex'points. Error'teims aie'calculated as desciibed in thetext for each index point, but only appear when they extend beyond the size of the symbol.

Basal index points: dístance Intercalated index points:depth to base

2 4d¡stance N€ (km)

Fig. 12. Northumberland, North: scatter plots showing the statistically significant relationships from Table 3between residuals (observed versus modelled relative sea-level) for basal and intercalated index points andparameters that may indicate local-scale processes (distance north-south and depth to base of Holocenesequence).

-RSL model

A Limiting dates


I Intercalated index points


Fig. 13. Northumberland, Central: relative sealevel index points. Error terms are calculated as described in thetext fo¡ each index point, but only appear when they extend beyond the size ofthe symbol.

1 41 21 0

4

2

0-2

4-6

4

^ 2t

ã oõoE - 'ao 4

6

1 2sediment (ml

e: - rU'É,

-2

-3

41 41 21 0


The very small data set for NorthumberlandCentral shows a maximu.m around lm a.bovepresent (Fig. 13) and a within-area trend withdistance (Fig. 16), which again should be quali-ûed by the fact that the data are from few sites.Sediment consolidation effects are evident fromthe significant correlations with both depth of

lntercalated index points:distance

sediment overburden and total thickness of theHolocene sequence.

Northumberland South reveals a mid-lateHolocene maximum less than 1m above present(Fig. l5). The age distribution of index pointsfrom basal peats is very restricted, and given theerror estimates for both the age and altitudeof residuals, no inference is made from the cor-relation (Fig. 16 and Table 3). The significantcorrelation with distance, in this case for datafrom intercalated peats, must be qualified withthe observation that the largest negative resi-duals of those data younger than 4cal.kaspcome from peat beds recently exposed on thebeach at Druridge Bay, south Northumberland.These peats have been exposed following thelandward migration of sand dunes, so they wereonce covered by many metres of dune sànd,which would have caused consolidation of thepeat and underlying Holocene sediment. Sedi-ment consolidation of intercalated index pointsis further shown by the significant correlationwith depth of overburden and sediment thick-ness (Table 3).

Tees Estuary

Nineteen index points were previously availablefrom within the Tees Estuary (Fig. 1) (Tooley1978; Shennan 1992). Eleven new index pointshave been gained under LOIS, which extendthe Tees record to most of the period 3-l Ical. kanr; further details in Plater et a/. (thisvolume). The data indicate that sea-level did notrise above its present level (Fig. 17). Many of thesamples come from thick sediment sequencesand the scatter seen on the RSL plot appears toresult at least in part from sediment consolida-tion. Significant correlations for altitude residuals against depth to base for the basal peatsand depth of overburden and total Holocenethickness for the intercalated peat index pointssupport this view (Fig. 18). The distance param-eter is, like the Humber, distance up estuary.The correlation between residuals and distanceis only significant at the 5o/olevel for intercalatedpeats, but this may indicate a change throughtime of the tidal prism; although the basalpeat data do not show a significant correlationcoefficient.

Lincolnshire Marshes

The area between the Humber Estuary and theFenland south of Gibraltar Point (Fig. 1) hadnot been studied in detail, and few index pointswere available despite the importance of the

4

E 2

ã 0ç - 2o o 4

-6

4

È 2€ - 0oÇ - 2E 4

-6

:rii:t:,il:tiii

È

4

e z€ - 0E - z6

€ 4-Þ

0 1 0 2 0 3 0

d¡stance t{-S (km,

Intercalated index points:depth of overburden

2sed¡menl (ml

Intercalated index points:total thickness

0 2 4 6sed¡ment (ml

Fig. 14. Northumberland, Central: scatter plotsshowing the statistically significant relationships fromTable 3 between residuals (observed versus modelledrelative sea-level) for intercalated index points andparameters that may indicate local-scale processes(distance north-south, depth of overburden and totalthickness of Holocene sequence).

Limiting dates

Basal index points

Intercalated index points

Thousand cal ibrated years BP: , . , - . r ' i i . . ; -

Fig. 15. Northumberland, South: relative sea-level index points. Error terms are calculated as described in thetext for each index point, but only appear when they extend beyond the size of the symbol.

Basal index points: age Intercalated index points :distance

ø4

40distânce N-S (kml

Fig. 16. Northumberland, South: scatter plots showing the statistically significant relationships from Table 3between residuals (observed versus modelled relative sea-level) for basal and intercalated index points andparameters that may indicate local-scale processes (age and distance north-south).

-RSL model


¡ Intercalated index points

Thousand cal¡brated yearc¡ BP

Fig. 17. Tees Estuary relative sealevel index points. Error te¡ms are calculated as desc¡ibed in the text for eachindex point, but only appear when they extend beyond the size of the symbol.

4

3

2

1

^ n

-t -1al,ú - 2

-3

-4

-6

1 0

4

^ 2Éã o!oF - 26oo 4

-6

4

2Eã o!o

F - 26ao 4

-6

ii

tr:.i-

FÍili:

60207 8cal ka BP

^ €El - ag,n -1o

-12

-14

-16

+{

ffi

ISOSTASY AND RELATIVE

Basal index points: depth tobase

SEA-LEVEL CHANGE

Intercalated index points:distance

2sediment (m)

lntercalated index points:depth of overburden

2 4 6sedíment (ml

Lincolnshire coast as a geological and geody-namic link between the major estuaries of theHumber and Wash (Brew 1997). There are now28 index points for this region, including for the

d¡stânc€ up ætuary Gml

Intercalated index points:total thickness

0 5 1 0 1 5sed¡ment (ml

first time, data from calcareous material withinclastic sediment units (further details in Hortonet al. this volume). The sealevel index pointsrecord a rising curve towards present (Fig. t9),

-2

4-6

E 2g 0Ê - zt 4

4

2

0-2

4-6

4

= 2

! - ooE - 2û€ ¿

-o

201 0-10

Fig. 18. Tees Estuary: scatter plots showing the statistically significant relationships from Table 3 betweenresiduals (observed versus modelled relative sealevel) for basal and intercalated index points and parameters thatmay indicate local-scale processes (distance up estuâ¡y, depth of overburden, depth to base of Hoiocene sequenceand total thickness of Holocene sequence). 'l:l'

:Ì::t7,,

iì:,:':;. t.

Ítì;f

L

+ Jr-ì

+- RSL rmdel

^ Limiting dates

o Basal index points

t Intercalated index points

Thousand cal ibrated years BP

Fig. 19. Lincolnshire Marshes relative sealevel index points. Error terms are calculated as described in the textfor each index point, but only appear when they extend beyond the size of the symbol.

-6

g - 8JI -10E

- 1 2

-2

À

-14

-16

-18

1 0

oo

o ?

¿ > L I. SHENNAN ET AL.

with the majority of the samples from inter-calated peats falling on or below the RSL curve.Horton èt al. (this volume) discuss the data olderthan 7.5cal. kasp, which includes dated samplesof foraminifera from clastic sediments. Differ-ential crustal movement within the area issuggested by the significant correlation withdistance, north-south along the coast as with the

Basal index points: distance

0 2 0 4 0 6 0d¡stance N-S (km)

Intercalated index points:age

0 2 4 6 8 1 0cal ka BP

lntercalated index points:depth to base

0 5 t 0sediment (m)

fig.20. Lincolnshire Marshes: scatter plots showingthe statistically significant relationships from Table 3between residuals (observed versus modelled relativesea-level) for basal and intercalated index points andparameters that may indicate local-scale processes(age, distance north-south, depth to base of Holocenesequence).

Northumberland analysis, for the basal peatindex points (Fig. 20 and Table 3). As discussedin the Northumberland section, this could alsoinclude a tidal change effect. The cluster ofsamples from foraminifera discussed by Hortonet al. (this volume) heavily influences the correla-tion with age (Fig. 20), and without which thereis no significant trend. Sediment consolidation issuggested in the signif,cant correlation for sedi-ment thickness to the base of the Holocene.

Fenland

Since a substantial mid Holocene Fenland data-base existed already (158 index points), LOISsampling concentrated on deeper sedimentscloser to the coast where earlier index pointscould be recovered. In this way the Fenlanddatabase has been extended by 35, includingseveral more index points of about 7 cal. ka np orolder. This region (Fig. l) is large enough to holdout the possibility of recording differences incrustal history within it. Furthermore, the spreadof the basal peat index points around the RSLcurve (Fig. 21) is no greater than other regionswith considerably smaller data sets (e.g. com-pare Fig. 22with Fig. 9). The spatial distributionof samples is large, over 90 km, and previousinvestigations indicate separate areas within theFenland basin showing contrasting sedimentchronolo gies (e. g. Shennan 1 98 6 b ; W aller 199 4).The limitations of differentiating between iso-static and palaeotidal effects as outlined inthe Humber analysis also occur here. The lack ofa signif,cant correlation between residuals anddistance for either basal (r6""u1: -0.06) or inter-calated (rintercalated :0.09) peat index points(Table 3) should not at this stage be used tosay there is no differential crustal movementwithin the Fenland basin. It is possible that achange in tidal range through time could have aneffect of similar magnitude and opposite sign.Model results indicate that tidal range changeswill have occurred in the Fenland area duringthe Holocene, but the predictions are poorlyconstrained by the resolution of the tidal modelsavailable or by the palaeogeography and palaeo-bathymetry models used (e.g. Hinton 1992;Shennan et al. this volume).

The only sediment parameter available for themajority of Fenland samples currently is thedepth of overburden. For index points fromboth basal and intercalated peats the correlationwith residual altitude is significant (Fig.22 andTable 3), indicating the effects o[ sedimentconsolidation.

E6,î,oEU'

l¡o

IÈã oEoç - 26 'Io 4

-6

4

2

0-2

4

2Èã oEo

F - 2øoo 4

-6

:Ì::i : l

-6

T¡iÅ

brr

ê4tta

o

ISOSTASY AND RELATIVE SEA.LEVEL CHANGE

-5

çIro(nÉ,

- 1 5

-20

. Thousand calibrated years BP

Fig. 21. Fenland relative seaJevel index points. Error terms are calculated as described in the text for each indexpoint, but only appear when they extend beyond the size of the symbol.

zv5

Fig.22. Fenland: scatter plots showing the statistically significant relationships from Table 3 between residuals(observed versus modelled relative sealevel) for basal and intercalated index points and parameters that mayindicate local-scale processes (depth of overburden).

Basal index points: depth ofoverburden

o ""4¡J9tt-¡

20

North Norfolk

Locating deeper sequences and recovering sev-eral basal peats previously known but notsystematically sampled enhanced the existingdistribution of data points for north Norfolk(Fig. l). Some extension along the north Norfolkcoast was also achieved and 23 new index pointsadded (Andrews et al. this volume). Index þointsfrom basal peats and the limiting dates fromfreshwater peats give a well-constrained seaJevelcurve, with most of the dates from intercalatedpeats lying below the curve (Fig. 23). Cur-rent tidal ranges change significantly east-westalong the north Norfolk coast, with MFIWSTfor Cromer, in the east, at 2.55mOD, rising to3.55mOD to the west at Hunstanton. The dis-tance parameter used in the analysis of residuals


.t

f - ;

0 5 1 0sed¡m€nt (m,

(Table 3) is distance west-east, but it shows nosigniflcant correlation with either basal or inter-calated peat index points, which may have indi-cated an important spatially dependent change intide levels between sites. Significant correlationswith depth of overburden and total thicknessof the Holocene sequence indicate the likely neteffects of sediment consolidation (Fig. 2q.

Regional summary: relative sea-leyelchanges between north Northumberland andnorth Norfolk

The eight regional relative sea-level curvesdescribed separately in the sections above (Figs7, l l , 13, 15,17, 19,21,23) show a consistent

ilrtl,,r..:,:;4,it:;-t:

;lì:''Él'

Llr{i:,J¿

4

2

0

-2

4

-6

4

^ 2

ã oEoF - 2âo 4

-6

{-o Basal index points

Limiting dates

r Intercalated index points

294 T. SHENNAN ET AL.

-5 +Ii+r ++dr-ì ü +-RSL model

A Limiting


t Intercelated index points+

-10

El - rsatú,

-20

-25

-301 2


Fig. 23. North Norfolk relative sea-level index points. Error terms are calculated as described in the text for eachindex point, but only appear when they extend beyond the size of the symbol.


Intercalated index points:total th¡ckness

the linear term may overestimate the current rateof sealevel fall because the polynomialsare fltted to data for ¿.3.5-Scal.kasp andshow the rate of fall increasing from the sea-level maximum to the present (Figs 11, 13, l5).Linear solutions from the youngest index pointto the present give rates of c. =0.66mma-' forNorthumberland North and c. -0.l2mma-lfor Northumberland Central and South. Atpresent there are no observations to test thisfurther. Nevertheless, the observations of mid-late Holocene sea-level index points abovepresent in Northumberland and below presentin the Tees Estuary support the change in signfor current rates of sea-level change betweenthese regions (Figs 25 and 26).

4

2

0-2

4

€

4

^ 2Eã oEoF - 2aao 4

€0 1 0 2 0

sed¡ment (ml

ßig. ?A. North Norfolk: scatter plots showing the statistically significant relationships from Table 3 betweenresiduals (observed versus modelled relative sealevel) for intercalated index points and parameters that mayindicate local-scale processes (depth of overburden and total thickness of Holocene sequence).

5 1 0 1 5 2 08od¡mont (m)

sequence of increasingly higher sea level to thenorth at any point in time (Fig. 25). Thissequence is consistent with model predictions(e.g. Lambeck 1995), but explanation of theoffsets between predictions and observations(Fig. 6) awaits the results of modelling usingrecent analyses of the earth- and ice-modelparameters (Lambeck et al. 1998; Shennan etal. this volume). Current rates of sea-level riserange from 1.04t0.lr.mma-l in the Fenlandto -1.30t0.68mma-' (i.e. sealevel fall) inNorthumberland North (Fig. 26). The scarcityof data for the last 3 ka from Northumber-land, where the data sets are the smallest, givesIess well-constrained statistical summaries thanother regions. The approach of changing only

o

\|\\t



Fig. 25. Holocene relative sealevel curves for the east coast ofEngland, based on the summary curves for eacharea shown in the preceding figures.

Fig.26. Late Holocene-present rates of relative sealevel change for the east coast of England based on theextrapolation of the summary curve for each area shown in the preceding figures. The error term is the 95o/oregression confidence interval based on the sum of squ¿res of the ¡esiduals from the regression. For theNorthumberland areas the open square shows the rate based on the linear ût through the youngest index pointonly (see text for discussion).

^ - 5gJat,É. _10

E OoÐGr - 1

Jvtú,

-2

-3

Ë Ë Ë H E C E Ëí q 3 * È E F 9F E Ë Èz : > =

z -

Comparison with IPCC scenarios of futuresea-level rise

The report of the Intergovernmental Panel onClimate Change 1995 (IPCC 95) summarizes arange of global mean sealevel rise scenarios(Houghton et al. 1996). The'best-guess' or'mid-range' predictions described in the technical

summary and summary for policy makers inIPCC 95 refer to sea-level predictions from theIS92a prediction, which is discussed in detail byWarrick et al. (1996). The rise in sea-level forIS92a predicts a rise of 20 cm from I 990 to 2050,within a range of uncertainty of 7-39 cm; and49 cm for 1990¿100, range 20-86 cm. The meanprediction translates into a rate of change in

--+- Northumberland-North--¡- Northum berland-Centra,-x- Northumberland-South---r- Tees--+- Humber--¡- Lincolnshire Marshes--r- Fenland--.-- North Norfolk

f +r T rï


ôFenland

*global mean with lower &upper limits

I Northumberland-North

)+I

+1850 1900 1950 2000 2050 2100 2150

Yearsfig.27. Rates of sealevel change 1890-2100. Global mean data with error limits from Warrick et al. (1996),assuming zero eustatic sea-level rise in the pre-industrial era, shown as 1890. The late-Holocene rates forNorthumberland North and Fenland (mean values from Fig. 26), shown fo¡ 1890 and added to the global meandata to give the values for 2050 and 2100.

1 0

Ê 6

Eë 40,|EÉ , 2

-2

excess of 5 mm a-r by 2100 (Fig. 27). Assumingmaintenance of the late Holocene relative differ-ences between the Fenland and Northumber-land North (Fig. 26) these data in combinationindicate a rate of sea-level rise bv 2100 in excessof 6mma-l for the Fenland and 4mma-lfor Northumberland North. Such predictionshave potentially very important implications forcoastline changes and management. Detailedconsideration of these is beyond the scope of thispaper except to point out that NorthumberlandNorth has not experienced such rates of seaJevelrise since before 4 cal. ka Bp (Fig. I l), though thismust be qualified by the different time-scales ofthe comparison, and for much of the period thecoastline has developed under falling RSL. Thetransition sometime in the next century fromrelative fall to relative rise may initiate signifi-cant coastline changes. Similarly, an increase inthe rate of seaJevel rise in the Fenland may besufficient to initiate a change from a regressive totransgressive phase (e.g. Shennan 1987; AllenL990, r99s, 1997).

The uncertainties of the IPCC projections areIarge relative to the mean predicted sea-levelrise (Fig. 2T. If sealevel rise is towards thelower end of the range, the differential crustalmovement or isostatic factor, i.e. the differencebetween Fenland and Northumberland North, isereater than the 1990-2100 increase. For sea-

level predictions at the higher end of the range,the differential isostatic factor becomes propor-tionately less important, around 25o/o.

Conclusions

This analysis ofsealevel data from the east coastof England has identifled local-scale and regio-nal-scale factors that explain spatial and tem-poral variations in the altitude of sealevel indexpoints. The isostatic effect ofthe glacial reboundprocess, including both the ice (glacio-isostatic)and water (hydro-isostatic) load contributions,explains the differences at the regional scale,which are manifested in approximately a 20-mrange for features formed 8kasp. By 4kanr,RSL in Northumberland was above present;whereas in areas to the south RSL has beenbelow present throughout the Holocene. Thesedifferences will produce different rates of sea-level change through the twenty-ûrst century.The data agree closely with the patterns pre-dicted by glacio- and hydro-isostatic models, butsmall systematic differences along the east coastawait testing against new ice models. New pre-dictions based on the revised models will providethe basis of testing further the effects of local-scale processes identified here. These includedifferential isostatic effects within the Fenland,tide-range changes during the Holocene, and the

t:.1: :i,.tìi

a

effects of sediment consolidation. The paucity ofdata for the last 3 cal. ka Bp remains a constrainton the analyses from most areas and is anotherpriority for future research.

This is publication number 590 of the Land-OceanInteraction Study (LOIS) Community Research pro-g¡amme and the work was supported by NERC grantsGST I 02 I 07 60, cST I 02 I 07 61 and GST/02I4766 underLOIS special topics 313, 316 and 348, respectively.Additional data were supplied with the collaborationof,principal investigators and research stafffrom otherLOIS projects and core programmes (details in prelaceof this volume). We are very grateful to the NERCradiocarbon laboratory at East Kildb¡ide for theradiocarbon dates and to landowners for access tofield sites. We thank T. Törnqvist and C. Vita-Finzifor their excellent reviews of the original version andsuggestions for improvements-

References

AI-I¡N, J. R. L. 1990. Salt-marsh growth andstratification: a numerical model with soecialreference to the Severn Estuary, soutirwestBritain. Marine Geology, 95, 77-96.

-1995. Salt-marsh growth and fluctuating sea level:implications of a simulation model for Flandriancoastal stratigraphy and peat based sealevelcurves. Sedimentary Geology, 100, 2l-45.

-1997 . On the minimum amplitude of regional sea-level fluctuations during the Flandrian. Journal ofQuaternary Science, 12, 501-505.

AuoRrws, J. E., BooMER, I., Berrrp, I., Bal-so¡r, p.,Bnrsrow, C., CHnosror.¡, P. N., FuNNELL, B. M.,HÂRwooD, G. M., JoN¡s, R., MAHER, B. A. &Sulrrlrrlrrro, G. B. 2000. Sedimentary evolution ofthe north Norfolk barrier coastline in the contextof Holocene sea-level change. This volume.

Bnnw, D. S. 1997. Holocene lithostratigraphy andbroad scale evolution of the Lincolnshire Out-marsh, eastern England. East Mídland Geogra-pher,20,20-32.

-, Horr, T., Pvr, K. & N¡wsseM, R. 2000.Holocene sedimenta¡y evolution and palaeocoas-tlines of the Fenland embayment, eastern Eng-land. Thís volume.

FrevrrNc, K., JoH¡.rsrox, P., Zw¡p-rz, D., yoroy-AMA, Y-, LnvrEcr, K. & CHAPPELI, J. 1998.Defining the eustatic sea-level curve since the lastglacial maximum using far and intermediate-fieldsites. Earth and Planetary Science Letters, 163,t ¿ t - 3 + z -

GeuNr, G. D. & TooLEy, M. J. 1974. Evidence forFlandrian sea-level changes in the HumberEstuary and adjacent areas. Bulletin of theInstitute of Geological Sciences, 48,25-41.

Goowrx, H. 1940. Studies in the post-glacial history ofBritish vegetation. III: Fenland pollen diagrams.IV: Post-glacial changes of relative land and sealevel in the English Fenland.. Philosophical Trans-actions of the Royal Society of London 8,230,239-303.

297

Hr¡.rrou, A. C. 1992. Palaeotidal changes within thearea of the Wash during the Holocene. Proceedingsof t he G e o lo g ís t s' As s o ciat ion, 103(3), 259 -27 Z.

HoRroN, B. P., Eowanos, R. J. & Lrovo, J. M. 2000.Implications and applications of a microfossiltransfer function in Holocene sea-level studies.Thís volume.

HoucHroN, J. J., M¡rno Flrso, L. G., CerrnNoen,B. 4., Hannrs, N., K¿,rrBN¡Bnc, A. & MASKELL,K. (eds) 1996. Climate Change 1995: The Scienceof Climate Change. Cambridge University Press,Cambridge.

Jercensrraa, S. 1966. SeaJevel changes during the last10000 years. In: Se,wvrn, J. S. (ed.) World Cli-mate 8000-0 ¡c. Royal Meteorological Society,London, 54-69.

LeMs¡cr, K. 1993a. Glacial rebound of the BritishIsles. I. Preliminary model results. GeophysicalJoumal Internationa!, 715, 941-9 59. .

-1993b. Glacial rebound of the British Isles. II.A high-resolution, high-precision model. Geophy-sical fournal International, f15, 960-990.

-1995. Late Devensian and Holocene sho¡elines ofthe British Isles and North Sea from models ofglacio-hydro-isostatic rebound. Journal of theGeological Society of London, 152, 437-448.

-1996. Limits on the areal extent of the Barents Seaice sheet in Late \Veischelian time. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology, 12, 4l-5L.

-, Stutrs¡n, C. & JoHNsroN. P. 1998. Sealevelchange, glacial rebound and mantle viscosity fornorthern Europe. Geophysical Journal Interna-tional,134, 102-144.

LoNc, A. J., Irwes, J. B., Krnev, J. R., LLoyÞ, J. M.,RUTHERFoRD, M. M., SnenNnN, I. & TooLEy,M. J. 1998. Holocene sea-level change and coastalevolution in the Humber Estuary, eastern Eng-land: an assessment of rapid coastal change. TheHolocene,8, 229¿47.

Mrrcer-rE, S. E., Elus, S., HoRToN, B. P., Ivr.¡es,J. 8., McAnr¡run, J. J., Mnr-rHNen,4., PARKEs,4., Per¡rrcr, J. S., REES, J. G., Rlncwev, J.,RUTHERFoRD, M. M., SHENNAN, I. & Tool¡y,M. J. 2000. The Holocene evolution of theHumber Estuary: reconstructing change in adynamic environment. This volume-

ORFoRD, J. D., Wnsox, P., WTNTLE, A. G., KNrcHT,J. & Bnel¡y, S. 2000. Coastal dune initiation inNorthumberland and Norfolk, eastern UK: cli-mate and sea-level changes as possible forcingagents for dune initiation. This volume.

Paur, M. A. & Bernas, B. F. 1998. A geo-technical correction for post-depositional sedi-ment compression: examples from the Fo¡thValley, Scotland. Journal of Quaternary Science,13,171-r76.

Pslrtex', W. R. 1998. Postglacial variations in the levelof the sea: implications for climate dynamics andsolid-earth geophysics. Reviews of Geophysics, 36,603-689.

PLATER, A. J. & Sr¡r¡¡NeN. l. 1992. Evidence ofHolocene sealevel change from the Northumber-land coast, eastern England. Proceedings of theGeologísts' Association, 103, 201-216.


: ,i:l

È

:

ñ

"fi

æiil¡¡iìi.å

lli*i

298 I. SHENNAN ET AL.

-, Rrocwev, J., RavN¡n, 8., Ss¡r.¡NnN, I.,. . . HoRToN, B. P., H¡.wonrs, E. Y., Wnrcnr, M. R.

& WtNrre, A. G. 2000. Sediment provenance andflux in the Tees estuary: the record f¡om the LateDevensian to the present. This volume.

REEs, J. G., Rrocwev, J., Elus, S., KNox, R. W. O'8.,NEwSHAM, R. & PARKES, A. 2000. Holocene sedi-ment storage in the Humber Estuary. Thís volume.

Rrocwe,v, J., ANonEws, J. E., Erl¡s, S., HoRroN,B. P., INN¡s, J. 8., KNox, R. W. O'8.,McAnrsun, J. J., Mesrn, B. A., Mercelre,S. E., MrrrpsNsn, 4., Panrgs, 4., R¡rs, J. G.,SAMwAys, G, M. & SHENNAN., I. 2000. Analysisand interpretation of Holocene sedimentârysequences in the Humber Estuary. This volume-

SHTNNaN, l. 1982. Interpretation of Flandrian sea-level d¿ta from the Fenland, England. Proceedingsof the Geologtsts' Associatìon,93, 53-63.

-1983. Flandrian and Late Devensian sea-levelchanges and crustal movements in England'andWales. 1z: Svtru, D. E. & DAwsoN, A. G. (eds)Shorelines and Isostasy. Academic Press, London,255-283.

-1986a. Flandrian sea-level changes in the Fenland.I. The geographical setting and evidence ofrelative sealevel changes. Journal of QuaternaryScience, l, ll9-154.

-1986b. Flandrian sealevel changes in the Fenland.IL Tendencies of sea-level movement, altitudinalchanges and local and regional factors. Journal of

Quate rnary S cience, l, 1 55 -17 9.-1987. Impacts on the Wash of sea-level rise.

Research and Survey in Nature Conservation, T,77-90.

-1989. Holocene crustal movements and sea-levelchanges in Great BriTain. Journal of QuaternarySctence, 4,77-89.

-1992. Late Quaternary sea-level changes andcrustal movements in eastern England and easternScotland: an assessment ofmodels ofcoastal evo-lution. Quat er nar y Internat ional, 15 | 16, 1 6 1- I 73.

-, HoRToN, B. P., Ix¡¡Es, J. 8., GEHRELS, W. R.,Llovo, J. M., McAnrHUR, J. J., Rrrrnrnrono,M. M. & W¡Ncnr¡ro, R. in press å- Late

Quaternary sea-level changes, crustal movementsand coastal evolution in Northumberland. Journalof Quaternary Science.

-, Irwrs, J. 8., LoNc, A. J. & ZoNc, Y. 1995.Late Devensian and Holocene relative sealevelchanges in northwestern Scotland: new data totest existing models. Quaternary International, 26,97-123.

-, T.AMBEcK, K., FLATHER, R., HoRToN, B. P.,McArrHUR, J. J., INNES, J. 8., Llovo, J. M. &Ruruennono, M. M. 2000. Modelling westernNorth Sea palaeogeographies and tidal changesduring the Holocene. This volume.

Honro¡¡, B. P., IUN¡s, J. 8., LLoYD,J. M., McAnrHUR, J. J., Puncslr, A. &Rurnrnnono, M. M. in press ¿. Late Devensianand Holocene records of relative sea-level changesin northwest Scotland and their implicationsfor glacio-hydro-isostatic modelling. QuaternaryScience Reviews-

SrurvER, M. & RETMER, P. J. 1993. Extended rac database and revised CALIB 3.0 laC age calib¡ationprogram. Radiocarbon, 35, 21 5-230.

TooLEy, M. J. 1978. The history of Hartlepool Bay.International Journal of Nautical Archaeology andUnderwater Exp loratíon, 7, 7 l-:7 5 -

TÕRNevrsr, T. 8., veN Re¡, M. H. M., vAN'T VEER,R. & ve¡¡ Gspr, B. 1998. Improvirig metliodologyfor high-resolution reconstruction of sealevel riseand neotectonics by palaeocological analysis andAMS r4C dating of basal peats. QudternaryResearch, 49,72-85.

Tno¡rs-SvIrs, J. 1955. Characferization of unconso-lidated sedimenfs. Danmarks Geologiske Under-sogelse, Series IV, 3, 38-73.

VANDE NooRT, R. & ELLIS, S. 1995. Wetland Heritageof Holderness: an Archaeological Survey. HumberWetlands Project, University of Huil, Hull.

- 8L -1997 . Wetland Heritage of the HumberheadLevels: an Archaeologícal Szrrey. Humber Wet-lands Project, University of Hull, Hull.

vAN DE Presscnr, O. (ed.) 1986. Sea-level research:a manual for the collection and evaluatíon of data.Geobooks, Norwich.

WALLER, M. 1994. The Fenland Project, Number 9:Flandrian Environmental Change ín Fenland, EastAnglia. Archaeology Report Number 70, Fen-land Project Committee, Cambridgeshire CountyCouncil, Cambridge.'Wennrcr,

R. 4., L¡ Pnovosr, C., MEIER, M. F.,OERLEMANs, J. & WooowoRrg, P. L. 1996.Changes in sea level. 1¿: Houcsro¡¡, J. T.,M¡rne Frrno, L. G., CALLANDER, B..{., HARRrs,N., KATTENBERG, A. & MASKELL, K. (eds)The Science of Climate Change: Contributíon ofl[/orking Group I to the Second AssessmentReport of the Intergovernmental Panel on CIí-mate Change. Cambridge University Press, Cam-bridge,572.

l ¡ . ' , '

:,:.:::l-: iir .

i':i.,,i i l l : f . .

:rir,::

ì.i'.r-

, l¡i.i

" ,fÍillir

'{¿

0352_001.pdf0352_022.pdf

Flolocene isostasy and relative sea-level changes on the east...

Documents

Transcript of Flolocene isostasy and relative sea-level changes on the east...