TER LEX SOUTH DEVON. UK. - Field Studies Council...Field Studies,8, (1996) 629-661 SEDIMENT YIELDS AND SEDIMENT DELIVERY INTHE CATCHMENTS OFSIAPTON LO\TER LEX SOUTH DEVON. UK. I. D.

Field Studies,8, (1996) 629-661

SEDIMENT YIELDS AND SEDIMENT DELIVERYINTHE CATCHMENTS OFSIAPTON LO\TER LEX

SOUTH DEVON. UK.

I. D. L. FOSTERCentre For Enaironmental Research and Consuhanca, Geography Dioision,

Coxenty Uniztersity, Priory St., Coaentry CVl sFB, U.K.

P. N. OWENS eNo D. E.\TALLINGGeographg Department, Exeter Uniztersity, Rennes Driae, Exeter EX4 4RJ, U.K.

ABSTR,\CT

Existing sediment yield data for the catchments draining to Slapton Lower Ley arecompared with published estimates for the catchrnent of the nearby Old Mill reservoir andare shown to be only about one third of what might be expected based on the data fromthat catchment. In order to explain this difference, field surveys, coupled with r37Cs,

mineral magnetic, geochemical and physical analysis of catchment soils and floodplainand lake sediments, have been used to first, examine the delivery of sediment fromthe hillslopes to the Lower Ley and construct a sediment budget for the main lStart)catchment draining into the lake; secondly, identifu the dominant sources of rhe sedimentdeposited on the Start floodplain and in the Lower Ley and ascertain if there have beenany changes in the main sources since the Second World War; and thirdly, evaluate theevidence provided by the lake-sediment record at Slapton for interpreting catchmentprocesses. For the Start catchment, it is shou'n that a large amount of the material erodedfrom the hillslopes does not reach the basin outlet but is stored at lntermediare locationssuch as upslope of hedge boundar ies and on the f loodplain. The amount of mater ia lstored in these locations since 1954 has been estimated and a tentative sediment budgetfor the Start catchment has been constructed. This budget suggests that some I 5% of soileroded from hillslopes is deposited behind hedgerows, whilst a further 58%, is transferredto storage on the Start floodplain. Only 27nk of the eroded soil reaches the Lower Ley.The most recent floodplain sediment is dominated by topsoil, primarily from pastureland, with increased contributions from subsoil sources at greater depths. \X/ith theexception ofa short period ofcatchment disturbance in the 1940s, sediment accumulatingin the lake is oftopsoil origin, probably from areas ofgrazed pasture rather than cultivatedfields. The physical characteristics of the Ley (i.e. water residence time and sediment trapefliciency), and the low sediment delivery to the lake, introduce significant limitationsregarding the potential of the lake-sediment record at this site for inferring catchmentD T O C E S S E S .

INrRoouc.rroN

Over the last three decades, a substantial amount of research has been undertaken onthe hydrology, water quality and sediment yields of the catchments draining to SlaptonLey in South Devon (Troake &Walling, 1973, 1974;Troake et al.,1976; vanVlymen,1979; Burt et al., 1983, 1988; Johnes & O'Sull ivan, 1989; Heathwaite et al., 1990;Owens, 1990; Heathwaite, 1993; Heathwaite & Burt, 1993; O'Sull ivan, 1993) and onthe palaeolimnology of Slapton Ley (Crabtree & Round,1967; O'Sull ivan, 1989,1992;O'Sull ivan et al., 1989, 1991; Heathwaite & O'Sull ivan, 1991; Foster et al., 1993;

629

630 I. D. L. FosrEn, P. N. OweNs nNo D. E.tWar-lxc

Heathwaite, 1993). To date, however, no attempt has been made to evaluate theestimates of sediment yield produced by these various studies, to set them in a widerregional context or to examine the overall sediment budget of the catchments drainingto the Ley. The most recent estimates of annual suspended sediment yield from theGara, Slapton rVood, Start and Stokeley Barton streams were reviewed by O'Sullivan eta l . ( 1989 ) anda reshown inThb le 1 ( fo rs i t e l oca t i onsseeF ig . 1 ) .Thees t ima tes fo r theindividual catchments range from <10 to >70 t km-2 year-r and reveal a number ofpotential inconsistencies. For example, the specific sediment yield estimated for theSlaptonWood catchment for 1987-88 is significantly higher than the yields for the Gara,Start and Stokeley Barton streams for the same period and is aiso substantially higherthan the two other estimates for the same catchment based on different periods. The totalsediment flux to the Lower Ley at Slapton has also been estimated by O'Sullivan er a/.(199i) by using unsupported lead-210 ("oPb) and caesium-137 (13?Cs) measurementsto establish a chronology for two lake sediment cores collected in 1987 (Heathwaite &O'Sull ivan, 1991) and thereby estimate rates of sediment accumulation. This worksuggested that since ca 1977 about 8 t ha-1 of dry sediment have been depositedannually on the lake bed, which is equivalent to a total mass accumulating in the LowerLey of 610 t year t. By dividing the mass by the total catchment area of both the Upperand Lower Leys (46 km2), O'Sullivan er al. (1991) derived a catchment sediment yieldo f l 3 .4 t km-2yea r l .A l t hough th i sappea rs tobecons i s ten tw i th theannua l sed imen t

yield estimates for the Start stream, it assumes that sediment delivered to the HigherLey from the Gara and Slapton \7ood catchments also reaches the Lower Ley. This isunlikely, since Burt (1994 pers. comm.) has estimated that only ca 118 t of sediment istransferred from the Upper Ley to the Lower Ley each year. Recomputing the sedimentyield, based on the total catchment area of the Lower Ley (which is larger than the areascontributing to the river measuring stations ofThble 1) and correcting for the transferbetween the Upper and Lower Ley, provides a sediment yield estimate of ca 32 t km-2year r. However, this value needs further adjustment to take account of the organicmatter content of the lake sediment and the trap efficiency of the lake. The bottomsediments in the Lower Ley have an organic matter content exceeding 30o/o of the dryweight. Expressing these data on a minerogenic basis produces a reduced sediment yieldestimate of ca 22 t km 2 year r. On the basis of vanVlymen's (1979) data, the capacity-inflow ratio (the ratio of lake volume to the annual runoff input) for the Lower Ley iscalculated to be 0.05. Using the trap efficiency curve published by Brune (1953), thetrap efficiency for the Lower Ley, based on a capacity inflow ratio of 0.05, is estimatedto be only 760/o.The trap efficiency-corrected estimate of minerogenic sediment yield tothe Lower Ley therefore becomes 29 t km 2 year 1.

These various studies suggest that, with the exception of the 1987 88 estimate forthe Slapton $7ood stream, maximum sediment yields from the catchments draining toSlapton Ley are ca30 t km-2 year 51. However, an analysis of sediment deposits in thenearby Old Mill reservoir (see Fig. 1 for location), reported by Foster &\Tall ing (1994),has shown that sediment yields over the last l5 years, in an area of similar climate,l ithology and land use, were ca 90 t km-2 yssl-r (Fig. 2).The contemporary sedimentyield from the Old Mill catchment is approximately three times greater than the lake-sediment based estimate of sediment yield to Slapton Lower Ley calculated above.Although sediment yields from the Old Mill catchment have increased over the past 50years, evidence based on radionuclide and mineral magnetic fingerprinting of reservoirsediments and catchment soils suggested that, with the exception of a single high

SedimentYields and Deliaery in Slapton Lower Ley Catchments

,14- i I' : k ) |O t + < t > , r l _ , . . .

" i q \ A If - J \ lr ) ) l' v { ^ l

H I ' ' '

B

Exeter

Site location l Dartmoor

;\ ./

)o

TorbayN

l

i i

i

The SlaptonCatchments

0 k m 1 0

Frc . ILocation of the study sites (A and B) showing the approximate area of the Slapton Ley (Box i) and Old Mill

(Box ii) catchments. Details of the Slapton catchmenrs are given in C.

631

632 I. D. L. FosreR, P. N. OwsNs eNo D. E.\Wnr-r-rNc

magnitude event, sediment sources had not changed since the reservoir was constructec.ln 1942 (Foster & \Walling, 1994).

The above evidence suggests that there are significant differences in suspendedsediment yield between the catchments of the Lower Ley at Slapton and those from thecatchment of the nearby Old Mill reservoir. These differences may simply reflectcontrasting rates of soil erosion and sediment transport between these two areas.Flowever, the close proximity of the two areas, and their similarity in terms of currentland use, land use history, lithology, relief and climate, might be expected to producemore similar sediment yields. The possibil i ty that the rates of sediment mobil isationoperating in the two catchments are similar, but that contrasts in sediment deliveryprocesses and in the overall sediment budget of the catchments draining to the Ley mayresult in different sediment outputs, must also be considered.

None of the studies undertaken in the Slapton catchments to date has adopted aholistic approach, based on the lake-catchment framework, in order to quantit/ therelationship between rates of soil erosion, sediment transport and deposition, ancthereby to establish a sediment budget, or to identify changes in sediment sourcesthrough time (cl Oldfield, 1977; Foster et al., 1988).The work presented in this paperattempts to address this problem and has three specific aims:

. To examine the transfer of sediment through the Start catchment and intoSlapton Lower Ley, and to reconstruct a sediment budget for the Start catchments ince 1954.

. To identify the main sediment sources and to ascertain if there have been anychanges in the dominant sources over the same time period.

. To assess the value of the lake-sediment record for interpreting catchmentprocesses within the study area.

Before considering these issues in detail, a brief description of the study area isprovided.

THs Sruoy AnEe

Slapton Ley and its contributing catchments are located in the South Hams region ofSouth Devon (Fig. 1).The area is underlain by slates, shales, siltstones and mudstonesof Devonian age (Dineley, 1961; Brunsden, 1965). The oldest sedimentary rocks areLower Devonian shales, slates and grits. At the base of this sequence are the Dartmouthslates, which consist of purple-green glossy slates with grit and qvartz bands and theLower Devonian Meadfoot beds, which comprise grey shales, grits and calcareous shalebands.

Soils of the Denbigh series are to be found throughout the area, and Manod seriessoils occupy the main valleys of the Start and Gara streams (see Fig" lC for locations).Denbigh soils are stony, well-drained and of moderate depth. They are fine and loamyand produce typical brown earths overlying solid or shattered rock. Manod series soilsconsist of free draining fine loamy soils developed over Palaeozoic mudstones, siltstonesor slates.The Manod series is a typical brown podzolic soil (Trudgil l, 1983; Findlay era l . ,1984).

Mean annual precipitation for Slapton vil lage (1961-1993) is 1049 mm. Precipi-tation is highly variable, with monthly coefficients of variation exceeding 40% (Ratsey,1975).Variability is greatest in spring and autumn. Most precipitation is associated with

q)

I

;N

I

6;c()=o

SedimentYields and Deliaery in Slapton Lower Ley Catchments 633

40 '1950 1960

F tc .2Sediment yields in the catchment of the Old Mill reservoir (corrected for the autochthonous content) derived

from a reservoir-sediment based reconstruction of sediment yield (after Foster &Walling, 1994).

TReI-s l. Estimates of sediment yield from the Slapton catchments

Suspended sediment -vieldArea (t km r

1'ear r)

Catchment (km')* ( t 1 'ear r )"

" + t

Ga ra 2 i . 62 218 .48 9 .25SlaptonWood 0.93 67.76 72.86 11.2# 8.4S ta r t 10 .79 10 .1 .36 9 .67S toke l y Ba r t on 1 .53 47 . 83 31 .26 11 .6 'Total f6.87 642.30 1.7.42

* From O'Sul l ivan et a l . ( i989) (data for 1987-88)

f From Park (pers. comm.)# is based on 4 years ofdata (1978 1981)n is based on 2 years ofdata (1980-1981)

f FromTroake &Walling (1973) (data fot 1971 72).

1 0 0 -

8 0 -

6 0 -

40-

20-

0 -1 9 1970 1980 1990 2000

YEAR

234

634 I. D. L. FosrER, P. N. OwsNs aNo D. E.\fanruc

south-westerly and north-westerly airstreams from the Atlantic which account for 58%of the total precipitation and 57"/o of the rain days.

Table 2. Characteristics of Slapton Ley and itscatchment

Basin area (km2)

Lake (Lower Ley) area (km2)

Lake to catchment area ratioN{.ean depth (m)

Maximum depth (m)

Volume (106 mr)Maximum altitude (m)

Minimum altitude (m)

46o . 7 7 1

59.71 . 5 52 . 51 . 7 7

1 8 3c a 5

The freshwater lagoon of Slapton Ley (Fig. lC) developed as a result of theshoreward movement of sediment from the Skerries bank, which l ies some 2-3 kmsouth-east of the Slapton barrier beach, as post-glacial sea levels rose (Robinson, l96l;Hails, 1975). From radiocarbon dating of basal sediments it has been shown thatSlapton Ley formed at around 2889 + 50 years BP (Morey, 1976). The physicalcharacteristics of Slapton Ley and its catchments are listed in Thble 2.

Ttre Appnoncn

A number of investigations involving both field surveys and laboratory analysis of soilsand sediments have been undertaken by the authors in order to shed further light on thesediment delivery dynamics and sediment budgets of the study area and to identifychanges in sediment sources that may have occurred over the last 40 years. As part ofthese investigations, soil samples and floodplain and lake sediment cores have beencollected from a range of sites in the Slapton catchment (Fig.3).These have been sub-jected to physical, geochemical, mineral magnetic and radionuclide measurements. Thefollowing sections provide some general background to the techniques involved, moredetailed information is contained in Owens (1990, 1994) and Foster &\Walling (1994).

Caesium- 1 3 7 nteasurementsSignificant quantities of the artificial radionuclide r37Cs were released into the environ-ment by the testing of thermo-nuclear weapons during the period extending from theearly 1950s through to the mid 1970s (Fig.4). Release into the stratosphere resulted inglobal dispersal, and fallout to the ground surface occurred mostly with precipitation.The most significant fallout occurred between 1958 and 1965 and inputs have generallydeclined since this latter date. Caesium-137 was also liberated by the Chernobyl nuclearaccident in 1986 (Fig.4), although south-west England did not receive measurablel37Cs fallout as a result of this event.

Caesium-137 reaching the land surface as fallout is strongly and rapidly adsorbed bythe surface soil, particularly by the clay and organic fractions, and the distribution ofr37cs in an undisturbed soil profi le usually declines exponentially with depth. Incultivated soils, 137Cs is generally well mixed within the plough layer (Walling & Bradley,

SedimentYields and Deliaery in SlaDton Lower Ley Catchments

i

' i l t

!

o := 'c

9 - .o

: . *T " '

z*l--

.=

o g

v a ! : Y! 6 0 -

O

N X E A

p l. ! 6 |e E 8 |; E a - l

; t pl?l . .+

o

o l

o c

= " ; !c c l l

3 : g s

!c

g ? 3

o o , " 6> o n o* . = R *> O u u

ooE

. o- &.s, oI O

!

o; F

d =

Frc. 3Details of the lorver Start catchment and Slapton Lower Le1',

shorving locations for sampling soils, floodplain and lake sediments.

636 I. D. L. Fostnn, P. N. OweNs aNo D. E.WanrNc

1988; \Tall ing & Quine, 1990). Caesium-137 measurements have been used for anumber of purposes in studies of soil erosion and sediment transport (l Ritchie &McHenry, 1990). For example, r37Cs has been used for dating lake and reservoirsediments (Pennington et al.,1973; Ritchie et al.,1973); for assessing rates and patternsof soil erosion (Valling & Quine, 1990); for quantif,iing rates of sediment accretion onfloodplains during periods of overbank flow (Walling & Bradleg 1990;\Tall ing et al.,1992); and for identifiiing dominant sediment sources by comparing the activities oflikely source materials with those of fluvial suspended sediment or recently depositedsediments in lakes and reservoirs (\Walling &Woodward, 1992; Foster &\7all ing, 1994;He & Owens, 1995).

Mineral magneic, phgsical and geochemical propertiesThere is a growing body of evidence to suggest that mineral magnetic analysis of fluvialsediment may provide a semi-quantitative indication of dominant sediment sourceswithin a drainage basin (Walling et al., 1979;Thompson & Oldfield, 1986; Foster &\falling, 1994). A range of properties are measured in order to determine the responseof minerals to an applied magnetic field of varying strengths and the interpretation isoften made in terms of dominant magnetic mineralogy and magnetic domains or ofmagnetic grain size (cl Thompson & Oldfield, 1986). It is not the purpose of this paperto provide a detailed mineral magnetic interpretation of the sediments analysed, but touse magnetic properties in order to fingerprint the most likely sediment sources.

Geochemical analysis is often used for similar purposes in lake sediment studies.Since the pioneering work of Mackereth (1966), numerous attempts have been made tointerpret the geochemical signatures of lake sediments in order to determine the relativesignificance of atmospheric fallout, internal productivity and diagenesis, and catchmentinputs to the lake sediment body (Engstrom &Wright, 1984; Bengtsson & Enell, 1986;O'Sull ivan et al., l99l1, Owens & Slaymaker, 1993). Similar approaches have beenadopted in order to determine sediment provenance from an analysis of actively trans-ported fluvial sediment (Peart &Walling, 1988;Walling, 1990).

The physical properties of sediment, particularly particle size, bulk density and losson ignition, provide further information concerning the nature of the deposit, the energyconditions under which it was transported and/or deposited and historical changes inorganic productivity at the site of deposition (H6kanson & Jansson, 1983). Further-more, in order to compare sediment properties with source materials, it is essential thatcomparisons account for physical sorting processesr since many of the geochemical andmineral magnetic properties are influenced by particle size composition (Peart &$7all ing, 1982;\fall ing, 1990). Corrections are also required for organic matter contentbecause, in high quantities, it can dilute the mineral magnetic signal, whereas its presencecan preferentially increase the concentrations of hear.y metals and phosphorus throughsorption processes (Horowitz, 1991).

The following sections present evidence which is used to address the three mainaims identified at the beginning of the paper.

SeorNrsNr DnLrvERy rN THE Sranr CRrcinreNr

In this section, we examine the delivery of sediment from the hillslopes to the lake andsubsequently construct a tentative catchment sediment budget. Attention has beenfocused on the main catchment draining into the Lower Ley, namely the Start.

6- 500E

c0> 400.zO

F- enn

@

200

SedimentYields and Deliaery in SlaDton Lower Ley Catchments 637

Frc. 4The atmospheric 137Cs fallout as recorded at Chilton, Oxfordshire.

(Data courtesy of AERE Harwell).

Soil erosion and within-field storageFrom a consideration of the differences in sediment yield evident between Old Millreservoir and the catchments draining to Slapton Ley, it seems reasonable to suggestthat a significant amount of eroded material might enter storage before reaching the flowmeasuring station at the catchment outlet or Slapton Ley itself. A preliminary recon-naissance survey of the sediment budget of the Start catchment undertaken by theauthors revealed extensive areas of sediment accumulation immediately upslope ofhedge boundaries in many of the steep pasture fields which border the main channelnetwork. Soil depths on the south facing slopes of the lower Start Valley were measuredby augering along hillslope transects in the area indicated on Fig. 3 and shown in detailon Fig. 5. Soil depths varied from as little as 9 cm on the upper steep convex slopes to asmuch as 94 cm behind transverse hedgerows. These initial results suggested that asignificant amount of material eroded from steep upslope locations may be deposited atthe base of the slope behind hedgerows. To investigate this further, r37Cs activities weremeasured in cores retrieved from locations upslope of f ield boundaries and Fig. 5illustrates typical r37Cs profiles and indicates the total 137Cs inventories for seven cores.The basic premise behind using l37Cs measurements for estimating rates of erosion anddeposition, is that the total inventory at a particular point where erosion or depositionmay have occurred is compared to that of a stable 'reference' location in order toestablish the magnitude of the erosion or deposition involved (Walling & Quine, 1990).Within the vicinity of the study area, l37Cs reference inventories in undisturbed soilprofi les average ca26l mBq cm-2 (corrected to 1992). From Fig.5 it can be seen thatboth the total inventorv and the shaoe of the orofiles indicate that a substantial amount

638 I. D. L. FosrER, P. N. Orrrnxs nxo D. E. Wer-r-rNc;

of deposition has occurred upslope of these hedge boundaries since 1954 (i.e. the startof 1l7Cs fallout). For example, the total inventory for core 7 is over 400 mBq cm 2

greater than the local reference inventory and this difference can be accounted forby deposition of sediment-associated r37Cs from upslope locations. Also, the peak in137Cs activity for core 7 at between 1l and 15 cm depth belorv the soil surface can betentatively ascribed to the peak in lr7cs fallout in 1963 (Fig.4), suggesring an accumu-lation of >10 cm of material at this location since 1963. In the case of cores I and 3, theextended depths to which high concentrations of lr7Cs are found are again indicative ofsubstantial amounts of deposition since the onset of 137Cs fallout in the late 1950s.

A similar picture emerges from a consideration of the downcore organic matter andmagnetic susceptibility profiles for cores taken from the transect marked on Fig. 5 (Fig.6). High organic matter contents are recorded to a depth of ca 30 cm in the profi lelocated immediately behind the transverse hedgerow, and these decline to backgroundlevels of under 4o/o tn subsoil. In the case of frequency dependent magnetic susceptibility(Xfd), high values are recorded to depths of ca 30 cm on the steepest slopes, whereasin the core immediately upslope of the hedgerow high values are recorded to a depthof over 50 cm, below which levels fall sharply. These results further support theinterpretation given above, based on thel3TCs data, and again indicate a significantaccumulation of sediment at the slope foot.

An estimate of the amount of soil deposited behind the hedgerows can be derivedfrom the excess r37cs inventories of cores l, 2 and 3 (Fig. 5). The average excess 137cs

inventorv of these three cores retrieved from behind the hedge boundary is 148.7 mBqcm 2 and this can be ascribed to the deposition of sediment derived trom erosion atupslope locations (cf owens, 1994). owens (1994) used a numerical mass balancemodel to calculate that the average IrTCs content (decay corrected to 1991) of sedimentdeposited at these footslope sites since 1954 was ca 20 mBq g 1. It is estimated thatdeposition has occurred over an area of ca 800 m2 in this field, which has a total area ofca 100 000 m2. If i t is further assumed that this excess inventory ascribed to sedimentdeposition is representative of the entire area of deposition, then the total amount oftopsoil that has accumulated upslope of the hedge boundary since 1954 is cc 60 t.Thisis equivalent to a mean annual soil erosion rate from the entire field of ca 16 t km 2

vear "

Floodpl ain s ediment atiortAnalyses of suspended sediment ransport and sedimentation rates and patterns in theStart catchment, presented by Owens (1990) and Foster et at. (1993) have shown thatsuspended sediment concentrations can decline downstream b_v as much as 800%between Battleford \ilflood and Deer Bridge (see Fig. 3 for locations), suggesting that asignificant quantity of sediment is being lost to channel and floodplain storage upstreamof Deer Bridge. Other evidence appears to suggest that the StartValley may have been azone of net sediment accumulation throughout the 20th century. Fig. 7 maps the presentday and historical extent of the wetland in the Start Valley as depicted on the OrdnanceSurvey maps of 1889, 1907, 1955 and 1965. It is clear from these maps that, whereaswetland was found only downstream of Deer Bridge in 1889, this currently extends up-valley almost as far as Battleford Wood. Historical records have also documented asignificant and sustained increase in water level in the Ley since 1920, which may havebeen responsible for the rapid expansion of the wetland area after this time.

Slapton Soi l Core 7' t tCs

lmBq g ' ' ;0 1 0 2 0 3 0 4 0

0-5

- 6,_1 no " "

f ro - rso-

8 1s-20

20-25

25-30Total Inventory673 mBq cm 2

N

tI

-1. Core locat ion

36s t tbs

inventory

Slapton Soi l Core 1t t t C s l m B q g t ;

0 5 1 0 1 5 2 0 2

0-5

E C - r U- 1 n - i (c*&

rs-zoo

20-25

25-30

Total Inventory459 mBq cm 2

Slapton Soi l Core 1' t tCs lmBq g ' ;

0 5 1 0 1 5 2 0 2 5

0-52 c r n- J - t v

i ro - rs.-o-

_e 15-20U

20-25

25-30Total Inventory451 mBq cm

SedimentYields and Deliztery in SlaDton Louer Ley Catchmerfts

Frc. 5Details of soil coring locations, representative r3tCs profiles and 137Cs inventories.

(For location of site, see inser 'A' of Fig. 3).

639

x fd (nmr kg ' )

50 1 0 0 1 5 0

640 I. D. L. Fosrrn, P. N.

Soi l Core 2(Behind Hedgerow)

Organic matter ('/") xfd (nm3 kg ')

5 1 0 1 5 0 5 0 1 0 0 1 5 (

OwsNs eNo D. E.Wer-r-rNc

Soil Core 24(Mid Slope)

xtd (nm3 kgl )5 0 1 0 0 1 5 0

Soi l Core 28(Top Slope)

xfd (nm3 kgl)50 100 150 200

Ellc6a)o

0 1 0 2 0 3 0 4 0 0

Percentage of part ic le size

Frc. 6Dow'ncore variations in mineral magnetic, organic matter and particle size characteristics of pasture soils.The

location of soil coring positions is identified on Fig. 5.

In order to examine further the magnitude of floodplain sedimentation) two cores(SVl and SV2; see Fig. 3 for sampling locations) have been analysed for r37Cs activity.Both cores contain l l7Cs to depths of over 50 cm and their total inventories aresubstantially greater than the local reference fallout inventory of 261 mBq cm-2 (Fig. 8).The high inventories of the floodplain cores reflect the excess r37Cs deposited inassociation with suspended sediment during periods of overbank flow. The depth atwhich l37Cs reaches a maximum activity in the floodplain sediment core provides achronological marker horizon which can be equated with the peak of atmospheric falloutin 1963 (Fig. ). In the case of core SV2, which was collected in 1991, the location ofthe r37Cs peak at a depth of ca 37 cm from the surface implies a deposition rate of ca1.3 cm year 1.The deposition rate for core SVl (collected in 1987) is almost the same(ca 1.4 cm year-r ) .

Surveys of the Start valley undertaken between 1986 and 1988 revealed extensivedeposition of sediment between Deer Bridge and Battleford. Seven valley cross sectionswere surveyed by Owens (1990) at locations indicated on Fig. 3.The survey showed thatuncompacted sediments were found to depths in excess of 1.5 m in manyplaces (Fig.9)and the mean depth of sediment in the sampled area was estimated to be ca 0.79 m.Documentary evidence (such as Fig. 7) was used to establish that this sediment had

' ., F "k<2vm

SedimentYields and Delivery in Slapton Lower Ley Catchments

Frc. 7Expansion ofthe wetland in the lower Start valley since 1889

identified from Ordnance Survey maps.

641

60 80

642 I. D. L. FosreR, P. N. On'eNs Rxo D. E,.WaluNc

E e no " -

;E-q)

U 4 U

701

I

r60x,-:

70 - l

Core SV1

1 3 7 C s ( m B q g - 1 )

40 60 8(

Total Invenlory740.0 mBq cm 2

Core SV2

1 3 7 C s ( m B q g - 1 )

40 B0

Total Inventory795.1 mBq cm'2

rr;Cs inventori". o..i prur,l".Tl. ,i..,. nnuapt"in scdiment cores.(Coring lcrcarions are iclenti l ied on Fis. l l .

been deposited r.vithin the last 30 to 100 -vears.This gives a sedimentation rate between0.79 and 2.6 cm ]'ear 1' which agrees i l 'el l with the values of 1.3 cm year t and 1.4 cmyear-l (since 1963) estimated from the rrTCs depth distribution of two floodplain cores(SVl and SV2, Fig. 8). The total volume of sediment stored in the valley was estimatedby owens (1990) to be 75,711 mr, which is equivalenr to 34,070 t of dry minerogenicsediment. Thus, between 3.11 and 1136 t of sediment has been deposited annually onthe floodplain during the recent historical past. This represents a sediment input fromthe upstream catchment area of bet\,veen 28 and 95 (mean 62) t km 2 year-\.

A tentcrtiae sediment budgetIt has been demonstrated above that, since 1954, sediment equivalent to an erosion ratefiom the upstream catchment of 62 t km 2 year I has accumulated on the floodplainof the Start valley and that the equivalent of ca 16 t km 2 year I of eroded topsoii hasbeen stored behind field hedgerows. Adding the two values together increases the

0

SedimentYields and Deliz:ery in Slapton Lower Ley Catchments 643

estimate of sediment mobilised rvithin the catchment by ca 78 t km 2 year r. Further-more, it has been estimated previously from the lake sediment record that the averagesediment input to the Lower Ley since 1977 was ca 29 t km-2 1,ear

t. Although thetime-base for this iake-based estimate of sediment yield and the estimate of equivalent'yield' associated with floodplain and hedgero'*' storage are different, if i t is assumedthat the mean annual gross soil erosion rate in the Slapton catchments since 1954 canbe represented b1, the sum of the two storage components (floodplain and upslope ofhedgerows) and the lake-based estimate of sediment yield to the Lower Ley, a grosserosion rate of ca 107 t km 2 year-r can be obtained.This estirnate is higher than theaverage sediment yield for the period since 1954 obtained for the Old Mill catchmenr ofca 63 t km-2 year r, but it is consistent with the value for the same catchment over thelast l5 years of ca 90 t km-2 I 's21 I (Foster &\Walling, I99D Gf.Fig.2).There is only avery l imited area of f loodplain in the catchment of rhe Old Mill reservoir and theamount of sediment stored upslope of hedge boundaries is also l ikel-v to be less than inthe Start catchment because of differences in field sizes between the two basins. Thegross erosion rate in the Old Mill catchment is thus likely to be closer to the estimate ofsediment yield obtained from the reservoir deposits.

From the calculations given above, a tentative sediment budget has been constructedfor the Start catchment (Fig. l0). No estimate is made of within-channel storage (cfSutherland, 1990;Warburton, 1990;Walling & Quine, 1993).The budget suggests thatl5ok of the eroded soil has been stored behind hedgerorvs with 58% transferred to flood-plain storage. On11'27oh ol the soil eroded from fields reaches Slapton Ley.The value ofsediment output cited abor,e is lou- compared to r-alues estimated for most other catch-ments (e.g. Costa, 1975; Lambert &\Val l ing, 1987; Suther land, 1990;Wal i ing & Quine,1993) and reflects the large amount of material stored on the hil lslopes and on thefloodplain in this catchment. Other studies (e.g. Meade, 1982;Trimble, 1983; Roberts &Church, 1986; Phil l ips, 1991) have also dernonstrated that in certain situations theoutput of sediment from the contributing catchment ma-v be low due to storage effects.

IlsNrrp'icRrroN oF SpnntgNr SouRcss

In the previous section it was assumed that the majority of sediment is derived from thehillslopes. Using a sediment budget approach it was demonstrated that ca 58% of thesediment eroded from the hilislopes in the Start catchment is stored on the floodplainand tha t theou tpu to f sed imen t to theLowerLey i s ca27o /o o f t heg rosse ros ion . I n th i ssection 137Cs, mineral magnetic and geochemical measurements of the sedimentdeposited on the floodplain and in the lake are examined in order to confirm that thehillslopes represent the dominant source of the sediment and to evaluate whether therehave been any changes in sources since 1954. As the floodplain is located upstream ofthe Ley it seems logical to discuss the results obtained from these sediments beforemoving on to a discussion of the results obtained from the Lower Lev.

Tlte Starr floodplainIt is possible to infer whether the sediment deposited on the floodplain (and in theLower Ley) is dominated by topsoil or subsoil sources by comparing the mineralmagnetic properties of the sediment with representative catchment soils. From trig. 6,for exampie, it is apparent that Xfd may be able to discriminate between topsoil andsubsoil. On the other hand, Xfd% (Xfd expressed as a percentage of the total low

N S

Transect 7

Transect 6

Transect 5

---=-*J/

V 1 0

Transect 4

Transect 2

Transect 1

1 0 0

NI

Sediment

Channel

wm

R a n k c / h a d n o c

Sediment over 1 .5m deep

I. D. L. FosrrR, P. N. OweNs eNo D. E.Walr-rNc

F r c . 9Floodplain sedimentation in the lower Start valley

based on surveys of the transects located on Fig. 3 (after Owens, 1990).

SedimentYields and Deliaery in SlaDton Lower Ley Catchments

Sediment Yield toLower Start Valley

645

Soil Erosion107

( 1 0 0 )

'j'HedgerowStorage

1 6( 1 5 )

FloodplainStorage

62(58)

A tentative sediment traget rinr.il!,,13 ro., catchment to Slapton Le-v.Percentage: are given in parentheses.

frequency susceptibility) would not be an effective indicator of the contributions of

topsoils and subsoils to the floodplain sediment. Although not shown in Fig. 6, thepattern for low frequency susceptibility (Xl| is similar to that of Xfd, with topsoil values

averaging ca 1.4 pm3 kg I whilst subsoils average 0.2 pmr kg r. Thus, both Xlf and Xfd

could be used to identify the dominant source type and any changes over time.

Similarly, isothermal remanent magnetisation (IRIvl) acquisition curves have been

measured for a range of topsoils and subsoils on the Denbigh and Manod series soils

and are represented as envelope curves in Fig. 1 1. It is evident from these curves that

catchment topsoils are dominated by a markedly different magnetic mineralogy(probably magnetite) from subsoils (probably haematite). Fig. l l also shows the

envelope IR-lVl acquisition curves for sixteen samples taken at various depths from afloodplain core collected near to core SV2 (Fig. 3). It is clear from these data that thefloodplain core includes the extreme end members of the potential source soils rangingfrom almost 100% subsoil to 100% topsoil dominance.The former is generally believed

to be related to channel bank sources as there is no evidence of gully erosion in this

catchment which would provide a similar mineral magnetic response in the floodplain

sediment. The surface samples of the floodplain sediment core illustrated in Fig. I I have

similar IR.lvl acquisition curves to topsoil, suggesting that topsoil is the dominant sourceof sediment recently deposited on the floodplain surface.

Temporal variations in sediment sources can be investigated by examining the depthdistribution of certain diagnostic mineral magnetic properties and Fig. l2 presents the

depth distributions of Xlf, Xfd, HIR-fvL (the loss of magnetisation at a backfield of 0.lT

after saturation when exoressed on a mass specific basis) and the HIRIvt:Xfd ratio in the

9 1(85)

646 I. D. L. FosrrR, P. N. OweNs nNo D. E.$TnluNc

c

F6

f,

6U)s

tIField (mT)

Catchment Topsoils E Catchment Subsoi ls

Floodplain SedimentsU Lake sediments I

IRlvt acquisition curves for subsoils,::;,j,1. ,"0 floodptain and take sedimenrs.

floodplain sediment core mentioned above. Horizons representing the peak in measuredr37cs fallout and the first occurrence of 'r;cs in 1954 are at ca 37-3g and 60 6l cmdepth in the profi le, respecrively. Xlf values decline from above 0.2 pm3 kg 1 at thesurface to around 0.06 pm3 kg I below 20 cm depth. Although more variable, a similarpattern is detected for Xfd which is generally higher in the upper 20 cm of the core.Individually high Xfd values are recorded at 39,51,59 and 65 cm depth and wouldseem to indicate a significant influx of topsoil to the floodplain at rhese levels. HIRlvtvalues show a reversal in the general pattern shorvn by the susceptibility parameters,with values below 0.4 mAm2 kg-1 (with the exception of the surface sample) to a depthof 20 cm. HIRIVI slowly increases from 20 cm to a depth of 49 cm and then increasesrapidly to values in excess of 0. 8 mAmt kg t below 5 0 cm in the sediment column. TheHIRlvt:Xfd ratio is close to zero in the upper 13 cm of the core, below which it increasesand becomes more variable.

Since susceptibil i tv parameters broadly increase with an increasing topsoilcontribution and HIRM increases with a greater subsoil contribution, there is clearevidence that topsoil dominates the most recent period of floodplain sedimentation, withslightly increased subsoil contributions from 13 to 36 cm depth (approximately 1963)and a substantial increase in the subsoil contribution, probably as a result of channelbank erosion, before this date. The upper 13 cm of sediment probably represent a periodof t ime from the mid 1970s to the late 1980s when this core was collected. Isolatedreductions in the HIRlvl:Xfd ratio at depths of 39,51,59 and 65, suggest a significantinflux of topsoil over limited periods of time, possibly associated with individual highmagnitude storm events.

Having demonstrated above that the dominant source of the sediment depositedon the f loodpla in s ince the mid-1970s is topsoi l , i t should be possib le to usel37Cs

A

0

1 0

20

?n

40

50

bU

70

c

0

1 0

20

30

40

50

OU

70

u.. t0.0

SedimentYields and Deliverg in Slapton Lower Ley Catchments

X l f 1 p m 3 k g ' ;

0 .1 0.2

B X fd (nm 'kg ' ' ;

Cs - 137Chronology

1 963-

1 954+--------------

HIRM lmAm'kg " ' ;

0 .3 0.6 0.9

647

D HIRM X fd ra t io (uA m' )

0.0 0 .5 1 .0 1 .5 2 .0

E

-co-

o

1.20 .0

E

-co-

o

I Y O J-

1 954€

Frc. 12Mineral magnetic properties of a floodplain sediment core collected in the StartValley near core SV2

(see Fig. 3 for location).

648 I. D. L. FosrrR, P. N. OwpNs nNo D. E.\War-r-rNc

measurements to identifii whether the sediment is derived from pasture or cultivatedfields. To examine this point further, the l37cs activities in representative samples ofpasture and cultivated surface (top I cm) soils were measured and compared with thoseof surface sediment on the floodplain. Since there has been no significant fallout ofr37Cs in this region since the mid 1980s (there was no measurable Chernobyl fallout),the surface sediment accumulating on the floodplain since this time will only contain137Cs associated with sediment deposited during overbank flood events. The average137Cs concentration of pasture soils (n: l0) is 27.5 mBq g I (range 23 to 37 mBq g-t)and, for two cultivated soil profiles, the average concentrations in the plough layer werefound to be 13 and 16 mBq g-r (Owens,1994).These data provide similar contrastsbetween l37Cs concentrations in cultivated and pasture fields to those encountered inthe catchment of the Old Mill reservoir, where the l37Cs concentrations in cultivatedtopsoils average less than l0 mBq g r, whilst activities in pasture topsoils nnge from 24to 36 mBq g I (Foster &Walling, 1994). On the Start f loodplain, r37Cs concentrationswere measuredin2T surface sediment samples collected within an area of 100 mX 50 m,adjacent to coring point SV2 (Fig. 3) (Owens, 1994). Correcting for the mass of organicmaterial produced in-situ on the floodplain surface (ca 30o/o of the mass of the sedimentsamples), the average r37cs concentration of the clastic sediment is ca 22 mBq g-r(range ca 15 to 33 mBq g r). Even allowing for the potential enrichment or depletion ofr37cs in sediment deposited on the floodplain due to particle size effects (cl Walling &\woodward, 1992;He & owens, 1gg5), it is evident that the averager3Tcs activity of thefloodplain surface sediments more closely reflects the activities of soil samples collectec.from the surface of undisturbed pasture soils, rather than cultivated soils. This suggeststhat the dominant source of mobilised sediment in the Start carchment since the mid-1970s has been grazed pasture.

The Lower LeyIn addition to the mineral magnetic properties described in the previous section, therelationship between the HIRlvl:Xfd ratio and the percentage saturarion acquired at aforward magnetic field of 300mT mav also be used to differentiate berween topsoil andsubsoil, and Fig. l3 illustrates this relationship for samples raken at various depths fromlake sediment core SLl. Although not l inear, the relationship indicates rhat theseparameters are closei-v related and that, as the HIRIvI:Xfd ratio decreases to below 0.1.samples are dominated by topsoil. Above a value of 0.3, samples are dominated bysubsoil. Fig. 13 i l lustrates the presence of both topsoil and subsoil material in the lakesediment record, although topsoil appears to be the dominant source of the lakesediment.

Fig. I I plots the envelope curves for IR-lvl acquisition measurements on sedimentsamples taken at set depth intervals from core SLI collected from the Lower Ley (Fig.3). Again' in the majority of cases, the lake sediment samples more closely mirror theequivalent properties of catchment topsoils than of subsoils, suggesting that the formeris the more likely source of the sediment in this lake core.

Temporal changes in the relative importance of topsoil sources to sediment deliveredto the Ley can be assessed by examining the vertical distribution of mineral magneticproperties measured in lake sediment core SLI as shown in Fig. 14. Summary statisticsfor all measured properties are given in Thble 3. The average value for Xfd% is 5.28.Although this parameter is unsuitable for determining source type (i.e. topsoil orsubsoil), this value is a strong indicator that eroded material from the surrounding

SedimentYields and Deliztery in Slapton Lower Ley Catchments 649

70 75 80 85 90

o6 saluration acquired @ 300 mT Field

Frc . 13The relationship between the HIRNI:Xfd ratio and the percentage saturation acquired at a field of 300 mT

for samples taken at various depths from lake sediment core SL1 (see Fig. 3 for location).

catchment is reaching the Lower Ley, as the high value for this parameter indicates thepresence of secondary viscous magnetic minerals usually associated with fermentationprocesses in catchment soils under oxidising conditions (Thompson & Oldfield, 1986).

The plot of the HIR-lvl:Xfd ratio versus depth in the lake sediment profile (Fig. 14vi)is perhaps the best indicator of changes in sediment source. Although most sampies l ie

within the natural variability of catchment topsoils, the sample at 42 cm depth evidencesan increased subsoil/channel bank contribution. This can be dated to approximately1940-45, and corresponds with the init ial period of major agricultural change andlandscape disturbance identif ied by Heathwaite (1993).

The other mineral magnetic properties shown in Fig. l4 ali dispiay a similar trend)increasing gradually upcore. In the case of HIRlvl and Xfd, values have been adjustedfor loss on ignition at 450"C and 850oC in order to eliminate the diluting effects of

organic matter and carbonates, respectively. The S-Ratio, which is calculated from IR-lvl(0.8T) and a backfield measurement of remanence at 0.1 tesla, ranges between 0.4 and0.8 and also shows a slight increase upcore. It is tempting to argue that the upcorechange in susceptibil i ty and remanence parameters indicate a significant change insediment source, with topsoil assuming increasing importance. Correction of these datafor the effects of organic matter and carbonate content serves to enhance the upcoresignal by removing the diluting effects of these two diamagnetic components. However,consideration of some of the physical and geochemical properties of the lake sediments,as shown in Fig. 15, complicates this simple interpretation.

The data presented in Fig. 15 are for a total chemical extraction) rather than afractionation procedure as described by Heathwaite & O'Sull ivan (1991) for SlaptonLey, and Foster & Walling (1994) for the Old Mill reservoir. The chronology wasobtained by cross correlation of physical and geochemical characteristics between the

Io

4 n t

.9Ecc 0 .3!

X

> o 2c c - -

i

650 I. D. L. FosrnR, P. N. OwENs RNn D. E.\fnnrNc

x l f (um3 ko '1 )0 .2 o .4 "o .b o .B

x fd (nm3 kg1) IRM 0 .8T 1mAm2 kg1;2 0 4 0 " 6 0 o 2 4 ' 6 " 8 ' 1 o

E

;r tooo

i i

- - 850 "c- 4 C U t

- No correction

o.2S Ratioo .4 0 .6 0 .8

HIRM:Xfd Ratio (prA m-r)0 0 .1 0 .2 0 .3

E- e nc " "6c)o

40

IncreastngSubsoi l / Channel Bank

vi

Increasing Topsoil

Frc. l4Mineral magnetic properties of a sediment core retrieved from Slapton Lower Le-v

(core SLl ; F ig. 3) .

cores described here and those described and dated using 21OPb analysis by Heathwaite& Burt (1993). Even though the core was over 70 cm long, only the post-I930 data(55cm) are displayed.

Although the study basin is located in a relatively remote rural location, there isevidence to suggest that both total Pb and Zn concentrations have increased sub-stantially in the upper 60 cm of sediment. Whilst Zn is usually more soluble than Pb(Horowitz, 1991) the two patterns are closely correlated, suggesting that l i tt le post-depositional remobilisation of Zn has occurred. There are no known catchment sourceswhich may have contributed heavy metals to the lake sedimenr column at Slapton Ley,other than road drainage from the numerous small lanes which cross the catchment. Theupcore trend is similar to that attributed to the fallout of atmospheric contaminants

H I R M ( m A m 2 k g l )0 . 4 0 . 8 1 . 2 1 . 6

SedimentYields and Deliaery in Slapton Lower Ley Catchments 65r

ury uensrry (g cm ")0 0 .1 0 .2 0 .3 0 .4

Total P (ag gr)

400 800 1200

Fe : Mn Rat io100 150 200

% Organic Matter andCarbonate Content

0 1 0 2 0 3 0Chronology

1 987

.10

20EIc c nooo

40

20

co qn-ooo

40

50

60

50

60

r 932

Chronology

I 987

HO ano zn ($g g )50 100 150 200

i 9 3 1

Chrono logr

I 9 S i

trO

c J U

E"oo

40

50

60

/ l n c nD10 D90

Percent i les in cumulat ive PSD

Frc . 15Physical and chemical properties of a sediment core retrieved from Slapton Lo'"ver Le-v (core SLl; Fig.3).

6 5 2 I. D. L. FosrER, P. N. OwaNs nNo D. E.Walr-rNc

reported elsewhere in the UK (Foster & Dearing, 1987; Battarbee, 1988; Foster &Charlesworth, in press). Whilst concentrations are considerably lower than thoseassociated with contaminated reservoir sediments in Midland England (Foster et al.)1991a) they are considerably higher than those found in lakes in more remote west coasrsites in the UK such as the Scil ly Isles (Foster et al., l99lb).

The strong atmospheric pollution signal, as indicated by the Pb and Zn profiles, islikely to be associated with the deposition of highly magnetic carbonaceous particles andthis negates the use of mineral magnetic measurements of lake sediments to inferchanging sediment sources in the catchment through time (Foster & Dearing , 1987;Foster et al., l990b).This problem is less relevant to the interpretation of the floodplainsediment record, since the floodplain magnetic signal is much less sensitive to atmos-pheric flux inputs than the lake record. This is because the sedimentation rate (whenexpressed in g cm-2 year l) on the floodplain is more than double that in the lake, andthe input of magnetic carbonaceous particles will be significantly diluted in the former.Since atmospheric contaminants do not contain secondary magnetic minerals, anaverage Xfdo% exceeding 5 (Table 3) appears to provide the strongest evidence for thepresence of catchment soil in the lake sediments.

It may also be possible to use 137Cs measurements to identif i i sediment source. Aswith the floodplain sediment, the r37Cs activity of the surface lake sediment should beassociated with catchment inputs since there has been no significant recent atmosphericinputs.The depth distribution of r37Cs in one of the sediment cores collected from theLower Ley (core a in Fig. 3) is shown in Fig. l6i.The shape and total 137Cs inventory ofthis profi le are discussed in greater detail in the next section, and here attention isrestricted to the r37Cs concentrations, which increase upcore from values <10 mBq g-lbelow ca 30 cm depth to between 50 and 70 mBq g-r near the sediment surface.Thel37Cs concentrations of the surface lake-bottom sediments are considerably higher thanany of the potential catchment sources and this precludes an interpretation of sedimentsource based on 137Cs analysis alone. The high 137Cs activit ies in the Slapton Leysediments may reflect enrichment caused by sorting during delivery and the depositionof only very fine sediment particles in the Ley. Indeed, Fig. l5vi suggesrs that 90% ofthe sediment in the Ley has a diameter of less than 60pm. To date, no arrempt has beenmade to model these enrichment effects, but the high activities in the Lower Leysediments would, nevertheless, suggest that pasture topsoil is the more likely source.

To summarise the analysis presented above, topsoil sources, primarily pasture land,appear to dominate the most recent f loodplain sediment, with contributions fromsubsoil sources (probably due to channel erosion) increasing with depth. The mineralmagnetic data suggest that there have been periods of increased topsoil influx overlimited periods of time, perhaps associated with individual storm evenrs. The mineraimagnetic data suggest that catchment topsoil has also reached Slapton Lower Ley.\Whilst considerable downcore variability in the magnetic signal is recorded in the lakesediments, the trends in some physical and geochemical properties (i.e. heavy metalconcentrations) complicate any interpretation of potential sediment source changesthrough time derived solely from the lake sediment record. Similarly, there are problemsassociated with the use of 137Cs measurements to identifi' dominant sources and sourcechanges (and some further problems are discussed below). Flowever, the high l3?cs

concentrations of the lake sediment suggest that pasture topsoil is the most likely source.The dominance of topsoil erosion from pasture fields as the main sediment source

in recent t imes compares well with the known historv of land use in the Slapton

SedimentYields and Delioery in Slapton Lower Ley Catchments

TReI-s 3. Summary data for a Slapton Ley sediment core (core SL1)

653

Variable

Standarddeviation Minimum

Number of

Maximum samples

Dry density (g cm-3)\(/et density (g cm-r)DW ratioxlfxfdxfd%IRIvI qo.ee

BKIR,VTSRATIOHIR ,I

HIRlvl:XfdP (pg g- ' )

Pb (pg g ')

Ln lpg g ')

Mn (pg g t)

fe urg g ")Fe:Mn

Lu ( t r g g ' l

Ni (pg g ')

Al (pe g ')r rTCs (mBq g r)

D10 (pm)

D50 (pm)

D90 (pm)

SPAN

0 . 2 11 . 0 70 . 1 9o.23

13 .375.285.73

- 4 . t 3

0 .690.800 . 1 8

5t0.4795.227 5 . 2 80.24

2 t . 3 795.7 460.532 . 5 2

29.98t 7 . 5 54.15

i 6 . 1 813.952.10

0 . 1 10 . 1 00.080 . 1 1

t t .222.892 . t 02 .000 . 1 00 . 1 60 . 3 1

255.9617.4533.370.076.05

33.5615.779 . 7 28 . 7 1

22.640.483 . t 2

1 0 . 0 80 . 2 7

0 . 10 . 90 .0520 . 10 . 70 . 52 . 9

- 8 . r l

0.50 . 50 .0

tt2.o68.0933.330.06346 . 1

39.?rt6.8249.09

t . 8 l0 .03 . 7

l r .228.72 . 1

0 . 5l . )

0.3800 . 5

4 3 . 11 0 . i1 0 . 9- L 4

0 . 81 . 2t . 4

1200.0l 38 .551 3 8 . 8 9

0.354839.2

1 6 2 . 4 889.729 8 . 1 850.696 5 . 8

5 . 220.75 5 . 82 . 9

3 53 53 5353 53 53 53 53 53 53 53 53 53 53 53 53 53 53 53 53 5I : l1 41 41.1

catchments. In general, there has been an intensification of land use following the

Second\WorldWar. Heathwaite (1993) and Heathwaite & Burt (1993) reported that thearea of pasture land (both temporary and permanent) increased significantly after the

war. In 1986, permanent pasture occupied ca 560/o of riparian land use in the Start

catchment, while cultivated land only occupied3% of the riparian zone.The proximity

of riparian pasture land to the stream will have increased its relative contribution to the

suspended sediment load. Also, stocking densities associated with pasture land increasedsignificantly in the post-war period. Heathwaite et al. (1990) have demonstrated, usingrainfall simulation experiments, that runoff and sediment loss from overgrazed pasture is

an order of magnitude greater than that from both lightly grazed and temporary pasture

and is also significantly greater than that from cultivated land

EvaI-uarINc rHE SI-AproN LEy Borrou-SeDIMENT RscoRD

In the light of the information presented above, it is important to assess the value of thelake sediment record from Slapton Lower Ley for estimating catchment erosion ratesand for reconstructing changes in catchment processes. The examination of sedimentdelivery dynamics in the Start catchment provides strong evidence that only a small(<30%) proportion of the material eroded from the slopes actually reaches the Lower

654 I. D. L. FosrER, P. N. OwpNs nNl D. E.Weu;Nc

Ley. While the use of the lake sediment record may provide a reliable means ofcalculating sediment yields to the iake, this method alone results in a significant under-estimate of the erosion rate within the upstream catchment (owens, lgg0). \when theeffect of intermediate sediment storage behind hedge boundaries and on the floodplainis taken into account, the revised value is more closely comparable to that for the nearbycatchment of the Old Mill reservoir. The problems associated with sediment storage arefurther complicated by the fact that storage is a dynamic process that changes overtime (l rrimble, 1983; church & Slaymaker, l9g9). Sediment remobil ised fromintermediate storage areas in the lorver part of the catchment may complicate anyinterpretation of downstream sediment fluxes in terms of geomorphic processes in theupper part of the basin.

The potentiai probiems associated w-ith the use of the lake sediment record in theLower Ley for estimating sediment yields are further highlighted by examining the r37Csrecord contained within the sediments of Slapton Lower Le1, and Old Mill reservoir.Fig. l6 presents rlTCs profi les measured in a sediment core from each lake (core a, Fig.3 in the case of the Lower Ley). Unlike the atmospheric fallout record depicted inFig.4' r.vhich shows a decline inrsTCs fallout over rhe last 30 years, the 137Cs profi lefrom Slapton Ley shows increasing activity up-core) which may represent an increase inthe irTCs content of recently deposited sediment. The profi le shape appears to begenerallv consistent with other sites of relatively low sediment accumulation (Foster e/al., 1990a; wall ing & He, lgg3).In contrasr, the r37cs activity in the sediments ofthe old Mill reservoir rise rapidly to ca 100 mBq g-1 ar a depth of 60-6 1.5 cm, which istaken to indicate the depth of the 1963 peak in bomb-derived fallout (see Fig.4). Afterthis well-defined 1963 peak, high activit ies are sustained upcore, since erosion of r37Csenriched topsoil. largely from heavily grazed pasture) has continued to supplyr3TCs tothe resen'oir (Foster &\Walling, l9g4).

The total inventories of the two cores range from less than 200 mBq cm-2 inSlapton Le1'to over 1000 mBq cm 2 in the Oid Mill reservoir. Although it is temptingto compare directl l ' ther37Cs inventories of these lakes with the background inventorvfor the region (261 mBq cm 2), some care is required for trvo reasons. First, not allr37Cs input to the lake from the atmosphere or the catchment wil l be deposited on thelake bed. The amount deposited wil l depend on the mechanisms controll ing lr7Cstransfer to particulate material in the water coiumn and the trap efficiency of the waterbody. Secondlv, l lTCs mav not be uniforrniy deposited over the lake bed, since someareas accumulate sediment more rapidl-v than others ("focusing"). Focusing is l ikely toproduce a lake sediment inventorv greater than the local fallout inventory! whilst a lowtrap efficiency would lead to a lou'er vaiue. Correcting the l37Cs inventory for the trapefficiency of the Lower Ley, as described earlier, raises the value to 258 mBq crn- 2,which approximates the locai fallout inventory. However, even with the correction fortrap efficiency, it is evident that the r37Cs inventory for the Slapton Ley sediment core ismuch less than that for the Oid Mill reservoir.Thel3?Cs inventory in the sediment of'the Old Mill reservoir is more than four times greater than the iocal fallout inventory.The high sediment yields, as indicated in Fig. 2, appear to have been associated with asignificant influx to the reservoir of r37Cs derived from the erosion of catchment topsoilswhich has sustained high I rTCs activit ies during a period of declining atmosphericinflux. The Lower Ley, on the other hand, does not show the same parrern.

The comparison of the 137Cs profi les and inventories for the two lakes clearlyindicates that while the bottom sediment record of Old Mill reservoir is suitable for

SedimentYields and Deliz-tery in Slaoton Lower Ley Ca.tchments 655

i

137Cs (mBq g-1)

0 1 0 2 0 3 0 4 0 5 0

137Cs (mBq g- , )

20 40 60 80 100 120

t l

o

-co-

o

1 21 824303642485460

1 0 . 5

1 9 . 5

28.5

37.5

46.5

55.5

64.5

o

-co-

o

73

82.5

F IG . 16llTCs inventories and profiles ofsediment cores retrieved from Slapton Lou'er Ler

(i) and Old Mill Reservoir (ii).

estimating rates of, and temporal trends in, catchment erosion, that of Slapton LowerLey is unsuitable. Not only does most of the eroded sediment not actually reach theLower Ley, but there is also considerable uncertainty over the behaviour of the sedimentonce in the iake.The trap efficienc-v of the Lorver Le1,'is not particularly high (76o/o) and,because of the ver.v shallorv nature of the '"vater body (mean depth 1.55 m, Table 2),wind generated currents are l ikelv to mix and disturb the bottom sediment.Thus, 137Cs

(and 210Pb) measurements are l ikel-v to be of l imited value for dating purposes in thislake. The absence of a reliable detailed chronology renders any attempt at estimatingsediment yields to the Ley and paiaeoenvironmental reconstruction difficult.

There are also probiems associated with using the lake sediment record to identifi'the dominant sediment sources and source changes over time. Some of these problemshave already been highlighted in the previous section, but further uncertainties can behighlighted by comparing the IRM acquisit ion curves for both the lake and thefloodplain sediments (Fig. l1). The range associated with the curves for the lakesediment is significantly lower than that for the floodplain sediments. This is, perhaps,somewhat surprising since the age of the bottom of the lake sediment sequence isconsiderably older than that of the floodplain core, and the lake sediment mighttherefore be expected to include a wider range of source contributions given the landuse history of the catchment (cf Heathwaite, 1993).The most likely explanation is that

Total lnventory^196.0 mBq cm- '

Total lnventory .1017 .4 mBq cm'

656 I. D. L. FosreR, P. N. OweNs aNo D. E. \il7ALLrNG

selective sorting of the fluvial sediment passing through the Start floodplain and themarshy area immediately upstream of the Lower Ley (Fig. 3) results in the preferentialdeposition of coarser subsoil (which is haematite rich) and the transport to the lake bodyof finer (magnetite rich) topsoils. This in turn results in a bias of magnetic mineralogytowards topsoil derived materials.

The problems associated with examining source changes is also illustrated by theIron:Manganese ratio (Fe:Mn), which is often used in palaeolimnological studies as anindication of changing redox conditions, since Mn is more mobile in sediment understrongly reducing conditions (Engstrom & Vright, 1984). Interpretation of the Fe:Mnratio (Fig. 15) alone suggests that redox conditions have become more reducing inthe lower 20 cm and the upper 15 cm of the sediment coiumn. The particle size datapresented in Fig. 15 indicate the diameter of particles at the 10th, 50th and 90thpercentiles of the cumulative frequency distribution (Drn, D;o and Dqo respectively).Decreases in all measures are identified at ca 25 and 50 cm depth in the core, suggestingan influx of finer sediment at this time. This pattern is inversely related to the carbonateprofile and is directly related to the Fe:Mn ratio.The use of the Fe:Mn ratio as a palaeo-redox indicator in this lake requires some further consideration, since the trend appearsto mirror the particle size distribution of the accumulating sediment. Thus, variations inthe Fe:Mn ratio could reflect either changing sediment sources or changes in the particlesize composition of the deposited sediment.

CoNcr-usroNs

It has been shown that sediment f ield estimates for the catchments draining to SlaptonLey, based on monitoring and interpretation of lake sediment records, are lower thanwould be expected from a comparison with the estimates obtained for the catchment ofthe nearby Old Mill reservoir. The evidence presented above suggests that the mainreason for the apparently low sediment yield to Slapton Lower Ley is the significantaccumulation of sediment in the floodplain of the lower Start catchment which traps58% of the sediment moving down the Start stream. The quantity of sediment reachingthe Start stream is itself reduced by deposition behind hedgerows and it is estimated thatca l5%o of the total sediment mobil ised within the catchment since 1954 has been storedin such locations.

By adding the sediment accumulation rate on the floodplain and behind hedgerowsin the Start catchment to the estimated sediment yield to the Ley from the samecatchment, an estimate of the gross erosion rate of 107 t km 2 year 1 is obtained.Thisvalue is of a similar magnitude to that estimated by Foster &Walling (1994) for the OldMill reservoir catchment over the last l5 years (90 t km-2 year 1). It seems likely fromhistorical evidence that significant quantities of sediment began to accumulate upstreamof Deer Bridge after the rise in water level in the Ley in 1920s.

Mineral magnetic analysis of a floodplain sediment core suggests that topsoil hasrepresented the dominant sediment source since the mid 1970s. Before this time, thecontribution from subsoil sources was dominant, probably representing channel bankerosion. The mineral magnetic data also provide evidence that prior to the mid 1970sthere were times of increased erosion from topsoil during l imited periods. The 13?Cs

activity of surface sediment samples on the floodplain more closely resembles that ofpasture topsoil, suggesting that this is the dominant source of the recently depositedsediment. This observation asrees well with the documented land use historv of the

SedimentYields and Deliztery in Slapton Lower Ley Catchments 6 5 7

Slapton catchments, which i l lustrate an increase in the area of permanent pasture(especially in the riparian zone) and in the stocking densities on this land. Previousstudies have demonstrated that the sediment mobilised from heavily grazed pasture canbe an order of magnitude greater than that from lightly grazed pasture or fromcultivated land.

The mineral magnetic properties of the lake sediments have been compared with thesame properties in catchment soils. Although the interpretation of the mineral magneticproperties of the lake sediments is complicated by the presence of metal contaminantsand allochthonous material in the sediment column, it has been demonstrated that thesediments contain material derived from catchment soils. The HIRIvI:Xfd ratio appearsto be one of the best discriminators between topsoil and subsoil sources in both theSlapton Ley and Old Mill catchments. In both bottom-sediment records, there is noevidence of a major and sustained change in sediment source through time, althoughsediment deposited in Slapton Ley dating from the 1940s appears to contain moresubsoil; a feature which is attributed to the init ial disturbance caused by post-waragricultural intensification. The Old Mill reservoir also contains evidence of subsoilerosion, but here it relates to a single high magnitude event in the 1980s. The r37cs

concentrations associated u'ith the lake sediment suggest that topsoil from pasture fieldsmay be the dominant sediment source, although the high radiocaesium activit ies ofrecent lake sediments may be a function of selective transport and deposition of finer37Cs-enriched material in the Ley.

A lake-sediment based interpretation of catchment processes derived from ananalysis of the Lower Ley record alone must therefore be treated with considerablecaution for several reasons. First, the sediment yield to the Ley is less than 30%, of thegross rate of soil erosion from the catchment surface and it is not accurately' knou,nwhen significant sediment storage began and how this has changed over time. Secondly',the shallow water depth of the Lor,ver Le1' means that the estimated trap efficiencf isonly 7 6oh and that wind-generated currents are likely to have disturbed and mixed thedeposited sediments. This places some doubt on the use of radionuclides for sedimentdating in this lake. Thirdly, there is evidence to suggest that sediments in the Ley aredominated by fine silts, which indicates selective enrichment caused by sorting duringdeliverv, a process which mav influence radionuclide and chemical signatures in theIake.

Clearly, any attempt to use lake sediments for examining sediment (and solute)dynamics in the contributing catchment must be done in association with anunderstanding of contemporary and historical catchment processes and with anappreciation of the fate of the sediment once it reaches the lake. As many researchershave stressed (cl Owens, 1990; Dearing & Foster, 1993; Butcher et a\.,1993; Owens &Slaymaker, 1993), not ali lakes are suitable for interpreting processes operating in theupstream catchment.

Acxxonr-EIcEMENTS

The authors are grateful to a number of individuals for providing data and fieldassistance for the numerous sampling exercises from which the data presented here arederived. Special thanks go to Steve Bradle.v, Heather Dalgleish, John Dearing, QingpingHe, David Job, Tim Quine, Rachael Row, Fiona Swan, Elizabeth Turner and SharonWhitfield. S7e are particularly grateful to Chris Park (Lancaster University) for making

658 I. D. L. FosrER, P. N. Orvexs aNo D. E. War-r-rNc

unpublished sediment yield data available. I(eith Chell (Slapton Ley Field Centre) isalso gratefully acknowledged for making Centre facilities available for much of the fieldlvork. Finally, our thanks go to I{ate Muir and Steve Turnbull for producing thediagrams.

RspeRl'xcEsBarr.rr<err, R. \[t, (1988). Lahe AcidiJicatiort itt the United Kingdotn, 1800 1986. Department of rhe

Enr.ironment. HMSO. London.Bnxc-tssor, L. and ENELL, M., (1986). Chemical anall'sis. In: BERGLUND, B. (ed.) Hdndboole of Holocene

Palaeoecolog,y and Palaeolrydrolog1L Wiley, Chichesrer. pp. 123-45 l.Bnuxl, G. M.r (1953).'fhe trap efficiencl' of reservoirs. T'ransactions o;f the Anrcrican Geophltsicdl Union, 31,

407-4 1 8.Bnursor.:r, D., (1965). Dart*toor. The Geographical Assocratron.Bunr, T. P., BL.rrctHER, D. P., C()I-ES, x. and THou,ts, A. D., (1983). The Natural History of Slapton l-ey

Nature Reserve XV: H-vdrological processes in the SlaptonV/ood catchment. Field Studies,5,73l 752.Bunr, 1-. P., Anr<rr-t-, B. P, Tnuoc;u-r-, S. T. and'il/.rLLrsc, D. E., (1988). Stream nitrate levels in a small

catchmcnt in south-rvest England over a period of 15 -vears (1970-1985). Hldrological Proce-tsc.s, 2,267 284.

BtrL'culn, D. P., LARADZ. J. C., Por"t r..t, A. W R. and \VHtrn, P, (1993). Reservoir sedimenrarion rates in thesouthern Pennine region, UK. In: Mc\[rxus,J. and Ducr<, R.\{1 (eds) Geontorphologl'attd Sedinentologt'of Lakes und Resert'oirs. W'iier', Chichester. pp. 73 92.

CuuncH, L{ . and Slsvrnr tn, O., (1989). Disequi l ibr ium of Holocene sediment -v ie ld in g laciated Br i t ishColumbia. r \a lare. 337, 152 151.

Cos'r'1, J., (1975). Eflccts of agriculture on crosion and sedimentation in thc Piedmont Province, Marl'land.Bulletin ol the Geological Socie4' oJ Anten ca, 86, 1281 1286.

CIr,{B'r'REE, K. and Rou:io, F. E., ( I 967). Analvsis of a core from Slapton Le-v. ,Nczr Phlrologisr, 66, 255 27 0.DEARTNG, J. A. and FosrER, l. D. L.r (1993). Lake sediments and geomorphological processes: some thoughrs.

In: N,lcA(csus, J. and Ducr, R. \Xi (eds) Geonorphology attd Sedimentologl, o.f Lakes azd Reserrorrs. \ffrile1',

Chichester . pp. 5 14.D txn r - r v ,D . t - . , ( i 961 ) .TheDevon iansvs tem insou thDevonsh t re .F i e l dSn rd re . r , 1 (3 ) , 121 1 .10 .Excstnort , D. R. and rWnrcH' t ' , H. E. , (1984). Chemical st rat igraphy of lake sediments as a record of

environmental change. In: Hau'onrn, E. Y and Luxo, J. W G. (eds.) Lake Sedinents ancl IititirtnmtentalHistort'. Leicester University Press, Leicester. pp. 1 1,69.

FI\DLA\', l). C., CoLBoRxr, G. J. N., ColE, D.V., Hannoo,T. R., Hoc,tr., D. \'. and Sr.rrxls, S. J., (1934).Soils artd tlnir ttse irt SouthVest England. Soil Survey of England and Wales. Builetin No. 1.1. Harpcnden

Fost ' t ' : t t , I . D. L. and Cnanr- tsu 'oR-tHr S. N{. , (19S6). Heavy metals in the h1'drologicai c1-c le: t rends andexplanation. Hltdrological Processes, 10, 227 261'.

Fos' rnn, L D. L. , CHARI-ES$'oRTH, S. M. and I (erx, D. H. , (1991a). A comparatr \ -c studv of hearry metalcontamination in four reservoirs in the English Midlands. Hydrobiologia,2ll, 151 162.

Fosl'lR, I. D. 1.., CH.lnrEsn'onrH, S. M., De,rnrxc, J. A., Ketx, D. H. and D.u-ct-Ftrsl{. H., (199lb). Lakesediments: a surrogate measure of sediment-associated hearl' metal transport in fluvial s-vstems? In:Sedinrcnt attd StreantWltter Quali4'in a Changing Envnironntent (ProceedingsVienna S,vmposium). IAHSPub l . 203 , ) 2 \ - - J28 .

Fosrpn I. D. t-. and DE.\RING, J. A., (1987). Quantification of long term trends in atmospheric pollution andagticultural eutrophication: a lake-watershed approach. 7n The In.fluente o.f Clintare Change and Climatict'aridbilil ' on the Hytlrologic Regine andWater Resozrces (Procecdings \hncouver S-n-mposium). IAHS Publ.1 6 8 , 1 7 3 - 1 8 9 .

FosrER, I. D. L.' Dlantxc,J. A., Gnrn', R. and Onexn, K., (1988). Lake-catchmentsl an evaluation of theircontribution to studies of scdiment yield and delivery processes. In: Sedintertt Brdgers (Proceedings PortoA leg re Svmpo : rum; . IAHS Pub l . l 7 -1 ,4 l ? , 424 .

FosrER, I. D. L., Delruxc, J. A., Grun', R. and Oruxo, K., (1990a). The sedimentary clata base: An appraisalof lake and reservoir sediment based studies of sediment yield. In: Erosiott, Transport and DepositiottProcesses (Proceedings Jerusalem\X,brkshop). IAHS Publ. 189, l9 43.

FosrER, I. D. L.' Gn-ev', R. and DtaRrxc;, J. A., (1990b). Magnitude and frequency of sediment transport inagricultural catchments. In: Bo.cRDuax, J., Fosrln, I. D. L. and Dn.lnrxc, J. A., (eds) Soil Erosion onAgricuhural Land.l&/r1e.v, Chichester. pp. I 53 171 .

SedimentYields and DeliaerT, in Slapton Lower Ley Catchments 659

FosrER, I.,Wellrxc, D. and Ov'nxs, P., (1993). Sediment -vields and budgets in the StartValley. In: BURT,

T. P, (cd.) A Ficld Guide to the Geomorphology of the Slapton Region. British Geornorphological

Research Groupr FSC, OL:casiorml Publicatiorts of the Field Studies Couttcil,27. Shrewsbury, UK. pp. 25 30.

Fosrln. I. D. L. and\il.1r-uxc, D.8., (1994). Using reservoir sedimcnts to reconstruct sediment yields and

sources in the catchment of thc Old Mill Reservoir, South Devon, UK over the last 50 years. HJ'drologicdl

Sciences Journal ,39, J47 368.HA ILS ,J .R . , ( 1975 ) .Someaspec t so f t heQua te rna r_vh i s t o r yo fS ta r tBa_ l ' ,Devon .F ie l dS t t t d i es , 1 ,201 222 .

Harussox, L. and Jaxssox, M., (1983). Prhciples oJ Lake Sedinrenrologjr,. SpringerVerlag, Berlin.

Hr, Q. and Ovrxs, P. , (1995). Determinat ion of suspended sediment provenance using caesium-137,

unsupported lead-210 and radium-226: a numcr ical mix ing model approach. In: Fostr . ,n, I . D. L. ,

GL:RNEI-r-, A. M. and\X/Ese, B.Wi (eds) Sedintertt andWdter Quality in River Cdtchnenrs.\{)'ile-v, Chichester.pp. 207"227.

Hnrlulirrll:, A. L., (iS93). Catchment controls on the recent sediment history of Slapton Le1', south-wcst

England. In: Tnou.ls, D. S. G. and ALr.IsoN, R. (eds) Landscape Sensirixiy. \Wiley, Chichester. pp.

24 t 259 .Hl ;erHunnr- , A. L. and Bun, T. P. , (1993). The evidence for past and present erosion in the s lapton

catchment, southwest Devon. In: BELL, M. and BOARDMAN J. (eds) Past and Present Soil Erosion.

Oxbor.v Monograph 22, Oxford. pp. 89-100.HrxtHvatrn, A. L. and O'SuI-I-tr,.qx, P. E., (1991). Sequential inorganic chemical analysis of a core liom

Slapton Le1', Devon, UK. Hydrobiologia,214, 125 135.HhdlH{Ann, A. L. , Bum,T. P andTt<uocrt - r - , S.T. , (1990). Land-use contro ls on sediment product ion in a

lorvland catchment, south-rr"'est England. In: Bo,rRl,rnx, J., Fosfl.:R, I. D. L. and Drlzuxc, J. A., (eds)

Soil Erosion on Agriarlrural Lazd. \filey, Chichester. pp. 69-86.HoRo\\'ITZ, A. J., (1991). A Priner on Sedintent-Trace Elentent Cltenistrl,. 2nd ed. Lcrl'is. Chelsea, Michigan.

JoHxus. P. J. and O'Sur.r .nrx, P. E. , (1989). The Natural Histor-v of Slapton Lcy Nature Reserve XVII I .

Nitrogen and phosphorus losses from the catchment an export coethcient approach. Fteld Sndies,7,

285 309 .Ll , r tn lxL, C. P andWaluxc, D. E. , (1987). Floodplain scdimentat ion: a prelminar-v invest igat ion of the

contemporarv deposition r'vithin the lorver reaches of the River Culm, Deron, UI{. Gccgrulisieu Annaler,69A. 393 ,10,1.

M.rcr lu lu, F J H., (1966). Some chemical observat ions on postglacral lake sediments. Pl t t l t ts t tphical

Ti'dnsactiotts o.f the Ro1,al Society o.f Lortdort,250, 167 213.MhAr)E, R. H. , (1982). Sources, s inks, and storage of r iver sediment in the At lant ic drainage of the Uni ted

States. . fo i r la l o. l Geologt , ,90, 235 252.MoRrY, C. R., (i 976). The Natural Historl' of Slapton Le1' \ature Reserve IX:-I-he morphologl' and histor-v of

the lake basins. I'ie1J Srrrdr'e-'. .1, 353 3b8.Ot-or tnr-o, F. , (1977), Lakcs and thci r drainagc basins as fundamental uni ts of sediment-based ecological

studv. Plogrcs.r ri P/n,.ir'ia1 (ieograpl4l',1, -160 501.O'Sur-Ltr ' lx , P 8. , (1989). The ccosvstcm-u'atcrshcd cr)nccpr in the cnvironmental sc iences a review.

Imerrnt iota l - fourtn l o. f Etn: t rotr t t t t t t ta l Studies. I3. 27I 28I .O'Sut-t.trrr, P. 8., (19S2). Thc cutrophication of shallorv coastal lakes in Southwest England understanding

and recommendations lbr restoration. based on palaeolimnologl', historical records, and the modelling of

changing phosphorus loads. .l '1-r'drobrb I og i a, 213 I 211, 12 I -13 4.O'Sut- t . r r ' ,1x, P. E. , (1993). - \ lodcl l ing thc c i ' fccts of a l ternat ivc nutr ient contro l pol ic ies- the examplc of

Slapton Le.v, Devon. H1,drobio logia,25l , )51 361.O'Suruvex, P E., Hr:rHwurE, A. L., Frn-n, K. M. and Surrn, J. P. (eds) (1989). Sotnha:est England and tlt,:

Sltropshire-Cheshire XIeres. Excursion Guide A. Vth International S1'mposium on Palaeolimnologv,

Cumbria UI( . September 1989.C)'Sur-r-rr,a:1, P. E., HllrnvArrE, A. L., Allr-rRr', P. G., BnoorrrElD, D., Crucr, NI. ri(/., Moscr<ol, C.,

MULDER,T. B. ,Venrox, N. J. and\X'r i . r . rHLRS'r ' , J . M., (1991). Palaeol imnolog-v of Slapton Le1' . p"uotr .

Hydrobiologia, 214, 1 li 124.Ov'nxs, P N., (1990).Valle-v sedimentation at Slapton, South Devon, and its implications fbr the estimation of

lake sediment-based erosion rates. In: Bolnort.lx, J., Fosrpn, I. D. L. and Du-rntxc, l. A., (eds) Soi/Erosion on Agricuhural Land. Wile,v, Chichester. pp. 193-200.

OtvnNs, P. N., (1994). Tbuords itnprot:ed interpretatiort of caesiunt-137 nedsLrentents in soil erosion studies.Unpublished PhD Thesis, Exeter Universitv.

660 I. D. L. FosreR, P. N. OwnNs eNo D. E. !7er-lrNc

OwEns, P N. and SLAYN,LA.KER, O., (1993). Lacustrine sediment budgets in the Coast Mountains of BritishColumbia, Canada. In: McMANUS J. and DUCK, R. \V. (eds) Geomorphology and Sedimentology of Lakesand Reservoirs.Wiley, Chichester. pp. 105 123.

Planr, M. R. andWal l lNG, D.8. , (1982). Part ic le s ize character is t ics of f luv ia l suspended sediment. In:Recent Deaelopments in the Explanation and Prediction of Erosion and SedimentYietd (Prcceedings ExeterSymposium). IAHS Publ . 137,397 108.

Pram', M. R. andWen-txc' D. 8., (1988). Techniques for establishing suspended sediment sources in twodrainage basins in Devon, UK; a comparative assessment. In: Sedimenr Budgets (Proceedings Porto AlegreSymposium). IAHS Publ . 174,269 280.

PrxNIscroN,rW. F., Cer'tnnw, R. S. and FISHER, E. M., (1973). Observations on lake sediments using falloutCs-137 as a f facer. Narure,242,324-326.

PHTLLIPS, J. D.' (1991). Fluvial sediment budgets in the North Carolina Piedmont. Geontorphology,4,23l-241,.Rrrsrl S., (1975). The climate at Slapton Ley. Field Studies, l, 1,91-206.RrrcHIE, J. C. and McHrNnv, J. R., (1990). Application of radioactive fallout cesium-137 for measuring soi^

erosion and sediment accumulation rates and patterns: a review. Journal of Enaironmentat Quality, 19,215 233 .

RircHIE, J. C., McHExnr', J. R. and Gn-r-, A. C., (1973). Dating recenr reservoir sediments. Limnologjt andOceanography, 18, 254 26f .

Roernts, R. G. and CHURCH, M., (1986). The sediment budget in severely disturbed watersheds, QueenCharlotte Ranges, British Columbia. Canadian Journal of Forest Research,16, 1092-1 106.

RoeIxsoN, A. H. !7., (1961). The hydrography of Start Bay and its relationship to beach changes at Hallsands.Geographical Journal, 127, 63,77.

SuTnlnraso, R. A., (1990). Variability of in-channel sediment storage within a small semiarid drainage basin.Physical Geographjt, 11, 7 5 93.

THor'rPsoN, R. and or-or.rpr-o, F., (1986). Enztironnental Magtretism. Allen and unwin, London.Tnrr'tnLE, S.$il.' (1983). A sediment budget for Coon Creek basin in the Driftless Area,Wisconsin,1853-1977.

American Jottrnal of Science,283, 451-171.TROAKE, R. P. and \fALLING, D. E., (1973). The Natural Histor,v of Slapton Ley Nature Reserve VII. The

hydrology of the SlaptonVood stream. Field Stttdies,f,TIg 710.TROAKE, R. P. and WALLING, D. E., (1S71). Some observations on stream nitrate levels and fertiliser

application at Slapton, South Devon. Reports and Transactions o;f the Dev-onshire Association Jor theAdvancement of Science, 107 , 77-90.

TROAKE, R. P,Tnonxr, L. E. andrVaLLrxc, D. E. , (1976). Ni t rate loads of south Devon streams. MAFFTbchnical Bullerin, 32,340 351.

Tnuocrr-r-, S. T., (1983). The Natural History of Slapton Le-v Nature Reserve, XVI:The soils of Slapton Wood.Field Studies, 5, 833-840.

veN VLvnnN, C. D., (1979). The Natural Historl, of Slapton Le-v Nature Resen.e XIII: The rvater balance ofSlapton Ley. Field Studies, 5, 59-84.

WALLING, D. E. ' (1990). L inking the f ie ld to the r iver ; sediment del iver l ' f rom agr icul tural land. In:BOARDMAN, J., Fosrpn, I. D. L. and DranIxc, J. A., (eds) Soil Erosion on Agricultural Land.'Wltey,Chichester. pp. 129-152.

War-uxc, D. E. and Bn-uln1 S. B. , (1938). The use of Cs-137 to invest igate sediment del ivery f romcultivated areas in Devon, UK. In: Sedimenr Bzdgers (Proceedings PortoAlegre Symposium).IAHS Publ.t 7 4 , 3 2 5 3 3 6 .

rVerr- txc, D. E. and Brulor- lv , S. B. , (1990). Some appl icat ions of caesium-137 in rhe study of erosion,transport and deposition. ln'. Erosion, Transport and Depositiott Processes (Proceedings JerusalemSymposium). IAHS Publ . 189,179 203.

VaLLIxc, D. E. and HE, Q., (1993).Tou'ards improved interpretation of r37Cs profiles in lake sediments. In:MCMANUS J. and Ducr, R. !(. (eds) Geomorphologjt and Sedimentology of Lakes and Reseraoirs.Wiley,Chichester . pp.31 53.

WlLr-rNc, D. E. and Qursr,T. A., (1S90). f]se of caesium-137 to investigate patterns and rares of soil erosionon arable fields. In: BOARDMAN, J., FosrER, I. D. L. and DEARING, J. A., (eds) soil Erosion onAgricuhural Land. Wiley, Chichester. pp. 33,53.

\Walnsc, D. E. and QurNn, T. A., (1993). Using Chernobyl-derived fallout radionuclides to investigate therole of doq'nstream conveyance losses in the suspended sediment budget of the River Severn, IJnitedKingdom. Physical Geography, 14, 239 253.

SedimentYields and Delivery in Slapton Lower Ley Catchments 661

rWalI-INc, D. E. andriTooov,eno, I. C., (1992). Use of radiometric fingerprints to derive information onsuspended sediment sources. In: Erosion dnd Sediment Tiansport Monitoring Programmes in Riaer Basins(Proceedings Oslo Symposium). IAHS Publ. 210, 143 152.

Wet-t-rxc, D. E., Pranr, M., OLDFIELD, F. and Tnor,llsoN, R., (1979). Identifuing suspended sedimentsources by magnetic measurements on filter paper residue. Nature, 281, 1 10 1 13.

\WRI-uxc, D. E., QulNE, T. A. and HE, Q., (1992). Investigating contemporary rates of floodplain sedimen-tation. In: CARI.ING, P A. and Ptrrs, G. E. (eds) Lowland Floodplain Riaers: Geomorphological Perspectiaes.Viley, Chichester. pp. 165- 184.

\Wannunrou, J., (1990). An alpine proglacial fluvial sediment budget. Geografsha Annaler, 72A,261 272.

662 I. D. L. Fosrsn. P. N. OvzBNs eNo D. E,.\7er-r-rNc