The transport and fluvial redistribution of Chernobyl-derived radiocaesium within the River Wye basin, UK

The Science of the Total Environment, 1211 (1992) 109-131 Elsevier Science Publishers B.V., Amsterdam

109

The transport and fluvial redistribution of Chernobyl-derived radiocaesium within the River Wye basin, UK J.S. R o w a n a and D.E. Walling b ~Environmental Science Division, Institute of Environmental and Biological Sciences, Lancaster University LAI 4YQ, UK bGeography Department, University of Exeter, Rennes Drive, Exeter EX4 4R J, UK (Received May 22nd, 1991; accepted August 18th, 1991)

ABSTRACT Relatively little attention has been given to the long term prospect of fluvial transport processes redistributing Chernobyi-derived radiocaesium within the UK. Work undertaken within the Wye basin, central Wales, demonstrates a complex distribution of fallout at the catchment scale, with the bulk of the deposition concentrated in a narrow north-south band, situated in the west central areas of the basin, which contained in excess of 1500 Bq m -2 of 134Cs. Fluvial transport and redistribution of this material was demonstrated by river sampiing during the winter of 1988/89, when the radiocaesium content of suspended sediment transported by the River Wye ( - 3 0 - 5 0 mBq g-i of 137Cs) remained 3-5-times higher than pre-Chernobyl levels. Floodplain reaches displayed variable levels of secondary contamination, dependent upon the upstream supply of radiocaesium and local morphological controls. Accordingly, the highest 134Csinventories within the basin (> 6000 Bq m -2) were associated with rapidly accreting floodplain sites. A number of these sites experienced only limited amounts of direct atmospheric fallout. The importance of fluvial redistribution as a secondary contamination mechanism is thus highlighted. Key words: Wye; Chernobyl; radiocaesium; sediment-associated; redistribution

INTRODUCTION T h e 1986 accident at the C h e r n o b y l nuclear p o w e r station released substantial quantities o f ~9TCs a n d z34Cs into the a t m o s p h e r e a n d resulted in extensive, t h o u g h locally u n e v e n fallout across m u c h o f E u r o p e (cf. B a r n a b y , 1986). It has been estimated t h a t the i m p a c t o f the C h e r n o b y l accident within 0048-9697/92/$05.00 © 1992 Elsevier Science Publishers B.V. All rights reserved

110

J.S. ROWAN AND D.E. WALLING

the UK was to increase the existing weapons-test 137Cs inventory by an average of 40% (Cambray et al., 1987). Moreover, the high spatial variability of radiocaesium fallout over the UK, which was primarily a function of rainfall patterns, resulted in the bulk of this contamination being confined to the western upland areas of north and central Wales, Cumbria and Galloway (Smith and Clark, 1989). As a consequence of the limited geographic impact of the accident within Britain, concern for the environmental hazards associated with Chernobyl fallout has rightly focussed on these upland areas where significant levels of radionuclides continue to be found in grazing animals (Desmet et al., 1991). In contrast, little attention has been given to the potential for longer term redistribution of Chernobyl-derived fallout by fluvial processes and the possibility of secondary contamination occurring within areas adjacent to the main centres of fallout (Walling and Bradley, 1988a). Two factors can partly account for this apparent oversight. Firstly, in most circumstances 137Cs, the major fallout constituent with a relatively long half-life (t0.5 = 30.2 years), is rapidly fixed and immobilised by clay minerals within the upper layers of soil (Frissel and Pennders, 1983; Cremers et al., 1988; Bonnett et al., 1989). As a result further down-profile translocation is limited, and the physical transport processes of wind and water erosion are the primary pathways for ~37Cs redistribution (Rogowski and Tamura, 1970; McCallan et al., 1980; Campbell et al., 1982). Secondly, although fluvial processes are likely to dominate, the relatively low rates of soil loss and low sediment yield commonly found within the UK environment (Walling and Webb, 1987), suggest that the proportion of the total atmospheric deposition which might be mobilised in this manner will be low. Whilst these limiting factors should be acknowledged, substantial transfers of radiocaesium may still accompany sediment transport due to the preferential adsorption of radiocaesium onto fine-grained sediment (e.g. Bonnett et al., 1988). Moreover, the complex nature of the sediment delivery process (Walling, 1983), indicates that much of this sediment-bound radiocaesium could be stored within transitory and long-term sediment sinks (Marron, 1987; Trimble, 1990). The potential therefore exists for the localised accumulation of redistributed Chernobyl derived radiocaesium within channel, floodplain and reservoir/lake sediment sinks. In relative terms, the levels of fallout experienced in the UK were small compared to those found within the Soviet Union and many other areas of Europe (IAEA, 1986; Dorr and Munnich, 1987; Persson et al., 1987). Nevertheless, even the peripheral impact of the accident upon Britain has caused significant environmental and economic problems, and the largely overlooked redistribution of this material from upland sources may be responsible for increasing the geographic impact of the event still further. In order to in-

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN. UK

111

vestigate the potential impact of such redistribution, the spatial pattern of Chernobyl-derived fallout was mapped within the catchment of the River Wye. A parallel programme of monitoring the radiocaesium content of suspended sediment and floodplain soils was also undertaken to assess the degree to which fluvial transport and redistribution processes were operating to cause secondary contamination within the river basin. CATCHMENT CHARACTERISTICS

The 4200 km e Wye basin is a highly diverse catchment in terms of its geology, topography, soils, land use and climate, and this is reflected in a similarly diverse range of landscapes and landforms. Rising in the Plynlimon Mountains at an elevation of 680 m, the River Wye flows over 250 km before entering the Severn estuary south of Chepstow. Much of the western area of the catchment lies above 200 m, and is characterised by undulating plateaux, dissected by deep, steep-sided valleys. The country rock is predominantly composed of Ordovician shales, with grit and slate (Caradoc and Ashgill units), and Silurian shales, mudstones also containing some sandstone and limestone (Llandovery units). The soils of the western Welsh Mountains and southern Black Mountains are characteristically high in organics, have low pH values, and typically support only low intensity livestock rearing. Fringing these western uplands, the rolling countryside of the Welsh Borders represents a transitional zone which gradually merges into the more subdued relief of the eastern and southern lowlands. In this region, fertile brown earth soils have formed on the widespread Devonian Lower Old Red Sandstones. Along the main river corridor, and within the eastern lowlands, arable farming and intensive livestock rearing become increasingly prevalent. The pattern of annual precipitation over the catchment displays a marked rain shadow. The western uplands receive in excess of 2300 mm, whereas much of the eastern lowlands receive annual totals of less than 800 mm. Runoff patterns exhibit correspondingly variability, with mean annual totals in the north west typically greater than 1300 mm year -1 whereas southeastern areas generally experience totals below 500 mm year -t (NERC, 1989). It will be demonstrated that these catchment characteristics play an important role in controlling the movement of radiocaesium from slopes into the channel system. FIELD SAMPLING A N D MEASUREMENT P R O G R A M M E

The analytical approach, and the collection of field data were conditioned by the size of the study area. Conceptually, the aim was to examine different facets of the transport and sediment delivery process at a range of temporal

112

J,S. ROWAN AND D.E. WALLING

and spatial scales. The work reported in the present paper was undertaken as part of a larger study encompassing the entire Severn/Wye basin (Rowan, 1990; Walling et al., 1989). Soil samples were collected from across the catchment area in order to assess the levels of fallout received. Bulk water samples were collected during periods of high discharge for the recovery of suspended sediment. Channel and floodplain sediment were also sampled at a variety of locations along the main river. Within the Wye basin, replicate shallow soil cores were collected from a total of 66 sites located within flat undisturbed areas, likely to yield representative estimates of local fallout values (Cawse and Horrill, 1986). These sites were located on or close to the intersections of a 10 km grid. At each sampling site 6-12 subsamples were collected using an 18-cm 2 corer inserted to a depth of 12-15 cm. This work was undertaken during the early part of 1988. Bulk samples of river water (50-100 1) were collected from two sites on the main river, at Rhayader and Hereford (Belmont gauging station), during periods of flood flow (Fig. 1). The high suspended sediment concentrations generally observed at these times (Table 3) confirmed that significant quantities of fine-grained particulates were being transported by the river. Recovery of the suspended sediment was carried out by continuous flow centrifugation, and the samples were freeze-dried prior to gamma assay. The storage of fluvially-dispersed radiocaesium within channel and floodplain sediment sinks was of primary interest in this study, although only the results of the floodplain sampling are reported in this paper. Two complementary sampling strategies were undertaken, which involved coring the floodplain surface to variable depths using the 18 cm 2 corer described above. The first was a reconnaissance survey, aimed at documenting broadscale spatial variations in the levels of secondary contamination along the length of the main river. The second focussed on micro-scale variations in contamination related toreach-specific topographic controls influencing the inundation process. LABORATORY TECHNIQUES

The 137Cs and 134Cs contents of soil and sediment samples collected in this study were measured by gamma spectrometry using Canberra and Ortec coaxial high purity Ge detectors (crystal dimensions 60 x 55 mm). Counting times were typically of the order of 10 000-25 000 s and provided precision terms of ± 5% and ± 8% for 137Csand 134Cs, respectively, at the 95% confidence level. Samples of suspended sediment were gently disaggregated and analysed without further processing. However, soil and floodplain sediment were ground and screened to < 2 mm in rotary sieves before being loaded into Marinelli beakers for counting.

113

CHERNOBYL-DER1VED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

\ t"~ ~'1 /

~"

0

25km ,.I

~ {~'"J~~ Rhayade~/.~_

\ ~,,/

L ~---/~ ,.,.'-~H e~e'fo e rd~

) I

(

k

/

-t

~llll Stu.dYc~') catcrlment

_

K

t" .3

J

I

I<'/'

Chepstow,ii~/ ,.,Y/

Fig. 1. Location of suspended sediment sampling stations, the River Wye basin, UK.

Prior to the input of radiocaesium from the Chernobyl accident, 137Cs was already widely dispersed in the environment due to atmospheric fallout from above ground weapons testing, indeed a survey of weapons-test 137Cs inventories undertaken throughout Wales during 1984 reported values in the range of 3290-11390 Bq m -2 (Cawse et al., 1988). In contrast, the other major component of Chernobyl fallout 134Cs (t0.5 = 2.1 years) is not found in weapons test debris. The presence of this radionuclide beyond the local influences of nuclear installations, can therefore be solely attributed to Chernobyl fallout, and accordingly decay-corrected 134Cs data are most commonly reported in this paper. Within Europe, the ratio of 134Cs/137Cs in Chernobyl-derived fallout ranged between 0.45 and 0.80 (Koster et al., 1987; Brydsten and Jansson, 1989). Within the UK, an average ratio of 0,6 has been widely cited (Cambray et al., 1987), and by application of this ratio it

] ]4

J.S. ROWAN AND D.E WALLING

is possible to distinguish between Chernobyl-derived ~37Cs and |37Cs already present from weapons-test fallout. In many respects the fate of IChernobyl-derived ~37Cs is of greater interest because it is more abundant, and by virtue of its longer half-life, will constitute a persistent environmental pollutant for many years. RESULTS AND INTERPRETATION

Spatial distribution of Chernobyl-derived fallout The areal pattern of Chernobyl 134Cs fallout determined from the input survey is illustrated in Fig. 2A. The interpolated map was produced by geostatistical analysis, and plotted using the UNIRAS graphics package. Several elongate features can be identified within the ~34Cs distribution, most obviously the continuous ribbon containing inventories in excess of 1500 Bq m -2, which stretches from the Black Mountains in the south of the region through to the Berwyn Mountains in the northwest of the adjacent Severn basin. Also contained within this north south trending band, several isolated pockets of relatively higher fallout were recorded, wherein the levels of U4Cs were greater than 4000 Bq m -2. Bounding the 1500 Bq m -2 fallout band, a much larger and less clearly defined zone is delimited by the 500 Bq m -2 contour. The variability of the atmospheric deposition is highlighted by the relatively low fallout totals received by the western and eastern margins of the Wye basin. In the Plynlimon headwater areas, values were typically below 250 Bq m -z, corresponding well with the findings of Bonnett et al. (1989). Similarly, fallout totals decline to the east, with much of the Lower Wye receiving totals under 500 Bq m -2. The unit mass concentration of radiocaesium (mBq g-l) found within the surface layers of a soil is governed by a number of inter-related variables, including the magnitude of the atmospheric input, bulk density characteristics and the capacity of the surface layers to retain and prevent down-profile translocation of radiocaesium. The retention capacity is influenced by such factors as clay content and mineralogy, pH, and the concentrations of other chemical solutions (Livens and Rimmer, 1988). The chemical behaviour of Chernobyl fallout within the upland areas of the UK has important implications for the redistribution potential. If, instead of being retained within the surface layers of the soil, fallout is dispersed within the profile, the effectiveness of surface transport processes to mobilise this material will be reduced. It is generally accepted that weapons-test fallout was in most locations rapidly and irreversibly adsorbed by the upper layers of the soil. However, because the bulk of the Chernobyl fallout occurred within upland areas,

115

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

A ~34CsBq m: 2500 2000 1500 1000 5OO 250 100

B

13~CsmBqg1 100

25 10

Fig. 2. Patterns of Chernobyl-derived ~Cs fallout across the Wye basin (0-15 cm). A) Total deposition (Bq m-2, B) Concentration in soils (mBq g-i). which are dominated, by acidic peaty soils deficient in clay minerals, a number of workers have suggested that Chernobyl-derived fallout may be characterised by a higher degree of availability, involving enhanced recycling of radiocaesium between soil and vegetation stores, and greater down-profile migration (Livens and Loveland, 1988; Kirk and Staunton, 1989). To ex-

116

J.S. ROWAN AND D.E. WALLING

amine this possibility, a number of depth-incremental soil profiles were collected throughout the Wye basin, and an example of such a profile is reproduced as Fig. 3. The radiocaesium distribution of this sample is consistent with most of the profiles obtained for the present study and indicates that the downward mobility was limited and that the bulk of the Chernobyl fallout was retained in the upper 10 cm of most soils. According to these direct measurements it would appear that physical processes of erosion, transport and redistribution will be important determinants of the long-term environmental fate of Chernobyl-derived radiocaesium, demonstrating a behaviour similar in many respects to weapons-derived radiocaesium. The spatial distribution of 134Cs unit mass activity is illustrated in Fig. 2B, and a relatively restricted area of elevated concentrations is again apparent. It should be emphasised that this distribution refers to average soil ]34Cs concentrations to a depth of 15 cm. Accordingly, these averages will underestimate the activity of material most likely to be mobilised by surficial erosion processes. Despite this qualification, it is clear that the major source of Chernobyl-derived radiocaesium to the channel network of the Wye basin

A o

o ~

soo

Activity (rnBq g.1) 1ooo 1,oo 2ooo

'

II

'

B 2,oo

o

l

°N 'I

I

12-

600

A E

Activity ( B q m 2) 1600 2000 I I

2400 I

2800

12' ~

L~ t-

e-

a

134CS

16-

[

18

Total Inventory 137Cs 7566 Bq m "2

[ '%s

134Cs 1548 Bq m "2

24 ¸

24-

30

30

Fig. 3. Verticaldistribution of radiocaesiumwithin undisturbedsoil profile, BlackMountains (SO 298300), July 1988.

117

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

will be associated with the mobilisation of particulates from the most contaminated west central upland sources. As a corollary, the southeastern areas of the basin are capable of providing only a limited supply of Chernobylderived fallout. Radiocaesium levels in suspended sediment

The radiocaesium levels in suspended sediment samples collected at the two monitoring stations on the River Wye are summarised by season in Table 1. Sampling commenced in October 1987 and continued until June 1989. No sediment was obtained prior to the Chernobyl accident, however, on the basis of the 134Cs/137Csratio previously described, estimates of the pre-Chernobyl levels were produced and these are also included in Table 1. Similar studies conducted on other rivers following the deposition of Chernobyl fallout indicate that substantial 'flushes' of radiocaesium occurred in both solution and sediment-associated form over periods ranging from days to weeks following atmospheric deposition (Peterson et al., 1986; Schoer, 1988; HaUdin et al., 1990; Desmet et al., 1990). For example, Walling and Bradley (1988a) reported a peak in the radiocaesium content of suspended sediment transported by the neighbouring River Severn within 3 days of the fallout episode. During this time the maximum 137Cscontent of 1450 ± 92 mBq g-~ exceeded pre-Chernobyl levels by more than two orders of magnitude. In most cases it is likely that some initial transfer of radiocaesium in solution would have occurred, and the subsequent scavenging of the dissolved phase onto particles would have been an important secondary

TABLE 1 Sampling distribution and the mean caesium-137 content of suspended sediment from the River Wye Sampling station

Oct-Mar 1985-86

Oct-Mar 1987-88

May-Sep 1988

Oct-Mar 1988-89

Total

Rhayader number 137Cs (mBq g-t)

-22.7 (5.2)

4 60.5 (4.9)

1 59.2

3 66.3 (1.4)

8 --

Hereford number 137Cs (mBq g-t)

-11.1 (4.1)

7 42.1 (11.8)

1 26.3

4 55.8 (3.6)

12 --

The Oct-Mar 1985-86 values of 137Cs content represent estimated pre-Chernobyl values. Bracketed figures refer to standard errors.

118

J,S. ROWAN AND D.E. WALLING

mechanisms contributing to these peak values (Hilton et al., 1989; Bonnett, 1990). Nevertheless, following the relatively short-lived 'flush', rapid fixing of the fallout in soils ensured that sediment-associated transport became the dominant pathway for radiocaesium redistribution (Waber et al., 1987). The average 137Cs content of suspended sediment transported by the River Wye during the winter of 1988/90 remained 3-5-times higher than the pre-Chernobyl estimates (Table 1). Confirmation of the role of Chernobyi fallout in accounting for the enhanced levels of 137Cs measured during the study period was provided by the presence of corresponding levels of 134Cs in these samples. As an example, the maximum 137Cs activity measured, at the Hereford collection station was 107.9 4- 6.2 mBq g-~, and this sample also contained 56.3 4- 3.4 mBq g-~ of 134Cs (site locations recorded in Fig. 1). In general, the radiocaesium content of suspended sediment samples collected from the upstream site at Rhayader were higher than those from the downstream Hereford site. Non-parametric Mann-Whitney U-tests confirmed that a statistically significant difference existed between the mean radiocaesium contents of the two stations (P < 0.05). These differences are graphically displayed in Fig. 4, where the 137Cs content of suspended sediment obtained during storm events are plotted against the ambient sediment concentrations measured at the time of collection. The Rhayader samples displayed a mean suspended sediment concentration of 75.0 mg l -l, which was found to be an order of magnitude lower than the mean value of 760.8 mg 1-l produced at Hereford. This pattern indicates an increased availability of sediment in the downstream reaches of the River Wye which is in agreement with previous findings (Oborne et al., 1980; Antoine and BensonEvans, 1988). Of additional note is the inter-storm variability in radiocaesium content observed at the two stations. Whereas the Rhayader site showed little variation throughout the study period, the radiocaesium content of sediment collected at Hereford evidenced a pronounced decline in 137Cs activity as the ambient suspended sediment concentration increased. Analysis of this negative relationship, using least-squares regression, indicated that log-transformed suspended sediment concentration data could account for up to 42% of the variability in the 137Cs activity measured at Hereford (P < 0.05). On the basis of the fallout distribution, which showed the bulk of the atmospheric fallout was deposited downstream of the Rhayader sampling station, a priori reasoning suggests that the radiocaesium content suspended sediment at Hereford would have been the greater of the two sites. The most likely explanation to account for the trends illustrated in Fig. 4 is that sediment derived from the relatively highly contaminated sub-catchments south of Rhayader, e.g. the River Ithon, is mixed with and diluted by proportional-

119

C H E R N O B Y L - D E R I V E D R A D I O C A E S I U M W I T H I N T H E R I V E R W Y E BASIN, U K

,--- 1000 "7, cn {:T rn

E EI1) .E m

v

~

~ej

[]

100 •

00

•

Rhayader samples

ffl

0 *-. ¢-

•

[~ []

[] []

[][] Hereford samples

[]

[]

10

i0

E

._~ ffl I

I

I

loo

looo

~oooo

Suspended sediment concentration (mg l 1) Fig. 4. Relationship between 137Cs content of suspended sediment and ambient sediment concentrations at two stations on the River Wye.

ly greater amounts of sediment derived from other less contaminated sources within the lower portion of the basin. These later sources are likely to consist of river bank and channel bed degradation and surface erosion within the cultivated lowlands upstream of Hereford where direct Chernobyl fallout totals were low. The dilution behaviour noted above is considered in more detail in Fig. 5. Three suspended sediment samples were collected from a series of floodwaves moving through the Belmont (Hereford) gauging station over the period 13 December 1987 to January 31, 1988. The times when sediment samples were collected are highlighted. The first sample, taken on the recession limb of an early peak contained 36.5 4- 2.3 mBq g-~ of 137Cs. The second, collected 3 days later on January 2, when the discharge was considerably higher, exhibited a 137Cs content of 44.3 4- 3.1 mBq g-l. The suspended sediment concentrations in these samples were !34 and 654 mg 1-1, respectively. Both samples also contained substantial quantities of 134Cs indicating that the bulk of the radiocaesium was Chernobyl-derived. The

120

J.S. ROWANAND D.E.WALLING

350-

137Cs44,28 + 3,07 mBq g.1

B

13~s16.64+ 0.66rnBqg'1

C

300 ~

250-1 ~3r [

-t

Cs 36.47 + 2.30 mBq g

2004

A

,7" >,,

0 December

January

Fig. 5. 1~TCscontent of suspended sediment during three successive floods at Hereford.

storm sampled on January 24 differed significantly from the previous events in two important ways. Firstly, no detectable 134Cs w a s measured, and the 137Cs content of the sample was considerably below the seasonal average for the site. Secondly, although comparable in discharge magnitude to the storm of 2 January, the 2900 mg 1-~ suspended sediment concentration associated with this sample was far higher than in the previous events. Because such high sediment concentrations are unlikely to originate from the stable upland areas of the catchment, the sediment concentration data logically point to the cultivated lowland areas upstream of Hereford zone as providing the main source of this sediment. The distribution of Chernobyl fallout illustrated in Fig. 2 indicated that the bulk of the fallout within the Wye basin was deposited to the north and south of the main river corridor. If the cultivated lowland regions of Hereford and Worcestershire become important sources of sediment during high flow periods, the combined effects of low Chernobyl fallout totals and locally high unit-area sediment yields could account for the comparatively low values of 137Cs and ~34Cs measured at Hereford. Furthermore, if as is likely, cultivated land represents a major sediment source in the middle

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN. UK

121

reaches of the Wye basin, radiocaesium activities in eroded soils will be further reduced due to tillage practices mixing the surface fallout with less contaminated soils within the plough layer. Another factor likely to account for these differences relates to a substantial proportion of the sediment load at Hereford being derived from bank sources. Much of the channel of the River Wye is deeply incised into the floodplain, and large volumes of loosely consolidated sediment, unlabelled with radiocaesium, are accordingly available for transport during a variety of flow conditions. As an illustration of the rates of bank erosion previously recorded in the region, Newson and Leeks (1987) have reported rates of bank retreat as high as 1.60 m year -1. To demonstrate how the radiocaesium content of sediment can be reduced as a flood-wave is routed downstream, the activities of samples collected close to the same turbidity maxima at both upstream and downstream sites are compared in Table 2. The storms of 12.11.87 and 23-24.11.87 both show substantial decreases in radiocaesium activity from the Rhayader to the Hereford sites. However, the pattern is not entirely consistent, as shown by the storm of 19-20.11.87 when the Hereford sample contained substantially higher activities. This event is noteworthy since it expresses the limitations of characterising large dynamic storm events by a single sample. Indeed this result may represent a pulse of high activity sediment derived from a more heavily contaminated sub-catchment south of Rhayader (e.g. the Ithon) contributing to the storm hydrograph sampled at Hereford. Estimates of the rate of radiocaesium removal over the study period were produced from mean annual sediment yields calculated for the two sites by Mitchell and Gerrard (1987). These data, a listing of fallout inventories, and the annual export rates are presented in Table 3. Despite having a lower catchment-wide fallout inventory and a lower mean annual suspended sediment yield, the Rhayader site exhibited a higher proportional annual export rate compared to Hereford, by virtue of the higher m e a n 137Cs content of sediment. These factors point to substantial differences in the mobility of radiocaesium between the two stations and again indicate contrasts in the TABLE 2 Comparison of the caesium-137 content (mBq g-l) of suspended sediment collected during the same flood event at two sites on the River Wye Sampling date

Rhayader

Hereford

12.11.87 19-20.11.87 23-24.03.89

70.3 -4- 7.9 56.6 + 3.2 65.4 ± 1.9

44.9 ± 1.8 107.9 ± 6.2 35.2 ± 0.8

Error terms expressed as 95% confidence intervals.

122

J.S.ROWANANDD.E.WALLING

TABLE 3 Sediment-associatedtransfer of Chernobyl-derived caesium-137for two sites within the Wye basin Rhayader (Ddol Farm) Catchment area Mean catchment 137Cs inventory Total deposition of Chernobyl 137Cs Suspended sediment yield Mean suspended concentration Mean 137Cscontent of suspended sediment Annual 137Csexport Annual percentage 137Cs lOSS1987--89

(km2) (Bq m -2) (TBq)

Hereford (Belmont)

174

1895

1179

2083

0.21

3.95

(t km-z year -l)

16.9

32.0

(mg l-l)

75.0

760.8

(Bq kg-1) (GBq year-I)

39.8 0.17

(% year-l)

0.056

31.5 1.91 0.048

sediment delivery systems of the two catchments. The estimated annual export rate of Chernobyl-derived 137Cs at Rhayader is in close agreement with the findings of Bonnett et al. (1989) working in the Plynlimon headwaters of the basin. In contrast, the relatively low annual 137Cs export proportion estimated for the Wye at Hereford reflects the importance of river banks and cultivated fields as sediment sources. Nevertheless, based on the respective annual 137Csexport totals of 0.17 and 1.91 GBq year -1, it is clear that considerable amounts of radiocaesium are being transported in association with suspended sediment in the Wye basin. In view of the fact that the volume of sediment delivered to the catchment outlet is only a small fraction of the gross erosion occurring within the basin (Walling, 1983), the implication of these findings are that even greater amounts of radiocaesium are being mobilised and redistributed into sediment sinks within the catchment system.

Redistribution of Chernobyl-derived radiocaesium with& the floodplain system Overbank flows and inundation caused by the retarded drainage of local streams may introduce sediment and sediment-bound pollutants onto floodplains (Macklin and DowseR, 1989). However, most existing work on floodplain contamination pertains to the redistribution of heavy metals and

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

123

deals with the hydrodynamic redistribution of mining wastes that originate either from a single source, or from a limited number of point sources within a mineralised region (e.g. Bradley and Cox, 1986). In contrast, 137Cs from atomic weapons testing and Chernobyl-derived radiocaesium must be viewed as a non-point source pollutant which will be contributed by the entire basin and whose behaviour is therefore complex. The River Wye is over 250 km long and is noted for well developed floodplain systems whose widths vary downstream depending on a variety of lithological, hydrological and human controls (Brown, 1987). A complementary sampling programme, involving both a broad-scale reconnaissance survey along the entire length of the main channel and a single case-study examination of local variations within a specific reach, was employed to investigate the secondary contamination of the floodplain by Chernobylderived radiocaesium. For the broad-scale study, samples were collected from a range of morphological units including: bank breaches, topographic depressions, cut-off channels, drainage ditches, flood channels, freely drained land, and locations close to potential over-bank flow obstacles such as fences and trees. Most samples were collected within 200 m of the main channel. The inherent weakness of the macro-scale approach is that only an approximate indication of micro-scale variations in sedimentation and radiocaesium loadings can be achieved. To compensate for this, and to provide detailed data on 134Cs and 137Cs inventories across particular floodplain reaches, an intensive coring programme was carried out over a single site at Moccas. A total of 53 floodplain cores were collected from various locations along the River Wye, and the sampling distribution is shown in Fig. 6A. A summary comparison of the mean ~34Cs inventories in the upper 15 cm of floodplain sediment with the equivalent local fallout inventories derived from the input sampling programme supplemented by additional samples collected adjacent to the floodplain, is presented in Table 4. A difference of means test (Student's t-test) indicated that the near-channel floodplain zone within the Wye basin contained significantly higher levels of ~34Cs than areas which received only direct atmospheric fallout inputs (P < 0.05). The capacity of floodplains to act as sinks for sediment-bound contaminants such as radiocaesium is thus highlighted. Downstream variations in floodplain activity levels are illustrated in Fig. 6B. The dashed lines represent the local direct fallout, and are presented as the 95% confidence interval, which takes account of the analytical precision (cf. Walling and Bradley, 1988b). Accordingly, where floodplain samples plot outside this reference envelope, a statistically significant enhancement or depletion of the local direct fallout component is indicated. Substantial downstream variation in local fallout totals is evident, and this

124

J.S. ROWANANDD.E.WALLING

A f~,~5 el k "1

25 km !

0 I

,;

/

-%

L?

i /J

) •

,

\.j

Floodplain sampling site J C~

N,

k

t

B A 0

mE t",r "¢D

- - [ - - 95% confidenceinterval of direct falloutestimates

4-

3~.~ 2-

E ._~

O

~

0

0

0

@ 0

1-

0 0

i

0

40

80

120

160

200

240

Distance downstream (kin) Fig. 6. Floodplain sampling on the River Wye (summer 1988). (A) Location of coring sites, (B) Downstream variations in 134Cs deposition relative to direct atmospheric fallout totals.

125

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN. UK

TABLE 4 A comparison of the mean floodplain caesium-134 inventory with local fallout inputs and the associated flux rates (1986-1988) River Wye

Inputs Bq m -2

Floodplain Bq m -2

Flux rates Bq m -2 year -1

I34Cs

794 (76)

1272 (171)

+ 239 (74)

Fallout inventories have been derived from the regional input sampling programme and supplementary data. Bracketed values refer to standard errors of the mean. A significant difference exists between the sample means at the 95% confidence interval.

is consistent with the uneven distribution of fallout previously discussed. A fallout m a x i m a occurred between 50 and 90 k m d o w n s t r e a m o f the headwaters o f the River Wye. A cluster o f elevated floodplain 134Cs inventories were also f o u n d within a n d a r o u n d a n d within this zone where peak levels ranged f r o m 2000 to over 6000 Bq m -2. The supply o f radiocaesium f r o m upstream sources represents a major control on the spatial variability o f floodplain c o n t a m i n a t i o n . The broadscale survey also identified the i m p o r t a n t gemorphological and hydrological factors which act as secondary mechanisms influencing the dispersion of the c o n t a m i n a n t (Table 5). The existence o f substantially elevated floodplain ISaCs inventories in the zone between 45 a n d 100 k m d o w n s t r e a m are unlikely to have been derived from sediment mobilised from the local floodplain surface, instead it is more likely to be related to a sediment input from the River Ithon. This was a c a t c h m e n t which received higher average fallout totals t h a n the m a i n River W y e network upstream. F u r t h e r downTABLE 5 Factors influencing floodplain inundation and associated deposition of sediment-associated radiocaesium 1. 2. 3. 4. 5.

Availability and concentration of upstream sources of radiocaesium. The conveyance capacity of the hydrological system to mobilise and transport sediment from geomorphological source areas. Fluvial transport parameters: particle size enrichment, chemical transformations, and suspended sediment concentrations. The geometric configuration and conductivity of the floodplain surface influencing inundation extent and floodwater residence times. The time-scale over which the deposition processes operate, and the magnitude/frequency of flood events which have occurred.

126

stream,

J.S. ROWAN AND D.E. WALLING

134Cs inventories generally declined, although minor peaks

associated with sediment inputs from other sub-catchments are a feature of the distribution. The sample yielding the highest 134Cs inventory was obtained at the furthest downstream location from a low terrace 0.5 m above the summer flow level, and was most probably the product of deposition of sediment derived from the highly contaminated Monnow sub-catchment. The general decline in the levels-of floodplain contamination which occur in the downstreaha direction is associated with dilution effects causing suspended sediment deposited on the floodplain to have a reduced radiocaesium content. Accordingly, floodplain contamination also tends to decline downstream of tributary junctions, although this may also be associated with a decrease in the depth of sedimentation (Petts, 1988). The highest rates of radiocaesium accumulation were found in low-lying, near-channel zones that were frequently flooded. In Fig. 6B the peak value of any group of samples was generally associated with relict point bars, scroll features, secondary channels or bank breaches where locally high sedimentation rates have caused the greatest contamination. Away from the channel, the most significant radiocaesium accumulations were associated with enclosed depressions where ponding effects were often important (Lewin and Hughes, 1980). The lowest rates of radiocaesium accumulation were associated with hummocky morphological units and in confined floodplain reaches where high velocity flood currents limited the opportunities for deposition. Few samples plotted below the reference envelope indicating that the floodplain was primarily acting as a sink, and not a source of radiocaesium. In a few areassome localised source may be associated with confined floodplain reaches where over-bank flow velocities are high and some scour is possible. To supplement the macro-scale approach, a reach specific examination of radiocaesium dispersal was undertaken at Moccas, a site 8 km upstream of Hereford. The 134Cs content of floodplain sediment (0-5 cm) is shown in Fig. 7A, and an immediate correspondence to the topography (Fig. 7B) is evident. This site was chosen not as being representative of the general floodplain systems, but because of the pronounced influence of bank breaches, and drainage ditches in controlling the flood water inundation sequence. The distribution of 134Cs is clearly related to the topography, and in Fig. 7C, the inverse relationship between contamination levels and relative height on the floodplain confirms the role of inundation frequency in controlling sedimentation rates. Other detailed reach investigations undertaken on the River Severn can be found in Rowan (1990) and Walling et al. (1991), and all confirm that fluvial transport and redistribution processes can cause substantial secondary contamination of areas receiving only relatively low levels of atmospheric input.

127

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

C

E o 3 0 0 "10

I,~ ~

O O

"~

"~ 2 0 0 0 ~

o

O

O 0

~ -~ 100

o

0

o

oO

0

8~o-~

0

0 0

0~

0 ' -1.1

'

.0~.9

'-o'.7'-o:~ '-o'.3 '-o'.1

o'.~ 'o'.3 ' o:s 'o:7

Relative Floodplain Elevation

Fig. 7. Reach specific variations in 1 3 4 C s accumulation (0-5 cm), Moccas floodplain, north of Hereford. (A) 134Cs inventories, (B) Topography, (C) Relationship between accumulated 134Cs inventory and floodplain elevation.

128

J.S. ROWAN AND D.E. WALLING

CONCLUSIONS Within the Wye basin, the fluvial transport and redistribution of Chernobyl-derived radiocaesium has caused substantial secondary contamination of floodplains with Chernobyl-derived radiocaesium due to deposition of suspended sediment. An event-based sampling programme carried out at two sites on the main river indicated that during the period October 1987-June 1989, suspended sediment contained 3-5-times the levels of radiocaesium than were encountered prior to the accident. Broad-scale and detailed surveys of this deposition indicate variable rates of secondary contamination depending upon the upstream supply of sediment and local morphological characteristics of the floodplain surface. In comparison with other areas of Europe, and the USSR in particular, the impact of the Chernobyl accident was relatively minor. However, evidence from the Wye basin study confirms that fluvial processes can be important vectors for radiocaesium mobilisation causing substantial secondary contamination in areas distant from the main centres of fallout. Further work is required to develop these initial findings into a modelling framework which can be generally applied to all river basins so that predictive scenarios can be formulated in the event of future radiation emissions. The results additionally offer scope to link the spatial patterns of sediment deposition identified by the radiocaesium data with hydrological modelling work (Knight, 1989; Gee et al., 1990). ACKNOWLEDGEMENTS The authors gratefully acknowledge the Natural Environmental Research Council for the provision of a studentship for JSR and a research grant to DEW. Thanks are extended to Milwyn and Owen Evans of Rhayader sewage works, and Elizabeth Payne of Broomy Hill water treatment works for collecting the bulk water samples for suspended sediment analysis, and to Welsh Water for supplying flow records for the Belmont gauging station. The cooperation of the many farmers who permitted access to input sampling sites and to the floodplains of the River Wye is also appreciated. Finally we would like to thank Dr. David Higgitt (Lancaster University) for his contribution to the field work and Mr. Jim Grapes, of Exeter University, for assistance with the gamma spectrometry. REFERENCES Antoine, S.E. and K. Benson-Evans, 1988. Environmentaland hydrologicalcharacteristicsof the River Wye system, Wales, UK. Acta Hydrochim. Hydrobiol., 16: 3-37.

CHERNOBYL-DERIVED RADIOCAESIUM WITHIN THE RIVER WYE BASIN, UK

129

Barnaby, F., 1986. Chernobyl: the consequences in Europe. Ambio, 15: 332-334. Bonnet, P.J.P., 1990. A review of the erosional behaviour of radionuclides in selected drainage basins. J. Environ. Radioactivity, 11: 251-266. Bonnet, P.J.P., P.G. Appleby and F. Oldfield, 1988. Radionuclides in coastal and estuarine sediments from Wirral and Lancashire. Sci. Total Environ. 70: 215-236. Bonnett, P.J.P., G.J.L. Leeks and R.S. Cambray, 1989. Transport processes for Chernobyllabelled sediments: preliminary evidence from upland Wales. Land Degrad. Rehab., 1: 39-50. Bradley, S.B. and J.J. Cox, 1986. Heavy metals in the Hamps and Manifold valleys, North Staffordshire, U.K.: Distribution in floodplain soils. Sci. Total Environ., 50: 103-128. Brown, A.G., 1987. Long-term sediment storage in the Severn and Wye catchments. In: K.J. Gregory, J. Lewin and J.B. Thornes (Eds), Palaeohydrology in Practice. John Wiley and Sons Ltd., Chichester, pp. 307-332. Brydsten, L. and M. Jansson, 1989. Studies of estuarine sediment dynamics using ~37Csfrom the Tjernobyl accident as a tracer. Est. Coast. Shelf Sci., 28: 249-259. Campbell, B.L., R.J. Loughran and G.L. Elliot, 1982. Caesium-137 as an indicator of geomorphic processes in a drainage basin system. Aust. Geog. Studies, 20, 1, 49-64. Cambray, R.S., P.A. Cawse, J.A. Garland, J.A.B. Gibon, P. Johnson, G.N.J. Lewis, D. Newton, L. Salmon and B.O. Wade, 1987. Observations on radioactivity from the Chernobyl accident. Nuclear Energy, 26: 77-101. Cawse, P.A. and A.D. Horrill, 1986. A suvey of caesium-137 and plutonium in British soils in 1977. AERE-R 10155, HMSO, London. Cawse, P.A., R.S. Cambray, S.J. Baker and P.J. Burton, 1988. A survey of background levels of environmental radioactivity in Wales, 1984-1986 (pre-Chernobyl), AERE-R 12535, HMSO, London. Cremers, A., A. Elsen, P. De Peter and A. Maes, 1988. Quantitative analysis of radiocaesium in soils. Nature, 335: 247-249. Desmet, G.M., L.R. Van Loon and B.J. Howard, 1991. Chemical speciation and bioavilability of elements in the environment and their relevance to radioecology. Sci. Total Environ., 100: 105-124. Dorr, H. and K.O. Munnich, 1987. Spatial distribution of soil 137Cs and 134Cs in West Germany. Naturwissenschaften, 74: 249-251. Frissel, M.J. and R. Pennders, 1983. Models for the accumulation and migration of 9°Sr, 137Cs, 239'24°pu and 241Amin the upper layers of soils. Spec. Pub. Br. Ecol. Soc., 3: 63-72. Gee, M.D., M.G. Anderson and L. Baird, 1990. Large-scale floodplain modelling. Earth Surf. Process. Landf., 15: 513-523. Halldin, S., A. Rodhe and B. Bjrman, 1990. Urban storm water transport and wash-off of caesium-137 after the Chernobyl accident. Water Air Soil Pollut., 49: 139-158. Hilton, J., W. Davidson, F. Livens, M. Kelly and J. Hamilton-Taylor, 1989. Transport mechanisms and rates for the long-lived Chernobyl deposits. DOE Report, PECD 7/9/385. IAEA, 1986. Summer report on the post accident review meeting on the Chernobyl accident. Vienna Safety Series 75-INSAG-1. Knight, D.W., 1989. Hydraulics of flood channels. In: K.J. Beven and P.A. Carling (Eds), Floods. John Wiley and Sons Ltd., Chichester, pp. 83-105. Koster, H.W., F.C.M. Mattern and R.M.J. Pennders, 1987. The radioactive contamination of the Dutch soil due to the accident of the nuclear power station at Chernobyl. (in Dutch), Working Document VI, RIVM, Bilthoven, Netherlands. Livens, F.R. and D.L. Rimmer, 1988. Physio-chemical controls on artificial radionuclides in soil. Soil Use Mgmt., 4: 69-75.

130

J,S. ROWAN AND D.E. WALLING

Lewin, J. and D.A. Hughes, 1980. Welsh floodplain studies 1I. Application of a quantitative inundation model. J. Hydrology, 46: 35-49. Marron, D.C., 1987. Floodplain storage of metal contaminated sediments downstream of a gold-mine at Lead, South Dakota. In: R.C. Averett and D.M. McKnight (Eds), Chemical Quality of Water and the Hydrological Cycle. Lewis, Michigan, pp. 193-209. Macklin, M.G. and R.B. Dowsett, 1989. The chemical and physical speciation of trace metals in fine-grained overbank flood sediments in the Tyne basin, northeast England. Catena, 16: 309-317. McCallan, M.E., B.M. O'Leary and C.W. Rose, 1980. Redistribution of caesium-137 by erosion and deposition on an Australian soil. Aust. J. Soil Res., 18: 119-128. Mitchell, D.J. and A.J. Gerrard, 1987. Morphological responses and sediment patterns. In: K.J. Gregory, J. Lewin and J.B. Thornes (Eds), Palaeohydrology in Practice. John Wiley and Sons Ltd., Chichester, pp. 177-200. NERC, 1989. Hydrological data UK1988 Yearbook. Institute of Hydrology and British Geological Survey, NERC, Wallingford. Oborne, A.C., M.P. Brooker and R.W. Edwards, 1980. The chemistry of the River Wye. J. Hydrology, 45: 233-252. Persson, C., H. Rodhe and L.E. De Geer, 1987. The Chernobyl accident - a meterorological analysis of how radionuclides reached and were deposited in Sweden. Ambio, 16: 20-30. Petersen, R.C., L. Landner and H. Blanck, 1986. Assessment of the impact of the Chernobyl accident on the biota of Swedish streams and lakes. Ambio, 15: 327-331. Peterson, W., H.-D. Knauth and R. Pepelnick, 1990. Vertical Distribution of Chernobyl isotopes and their correlation with heavy metals and organic carbon in sediment cores of the Elbe estuary. Sci. Total Environ., 97/98: 531-547. Petts, G.E., 1988. Accumulation of fine sediment within substrate gravels along two regulated river, UK. Reg. Rivers: Res. Mgmt., 2: 141-153. Rogowski, A.S. and T. Tamura, 1970. Erosional behaviour of cesium-137. Health Phys., 18: 467-477. Rowan, J.S., 1990. Sediment-associated transport and redistribution of Chernobyl-derived radiocaesium in fluvial systems. Ph.D. thesis, University of Exeter. Schoer, J., 1988. Investigation of transport processes along the Elbe River using Chernobyl radionuclides as tracers. Environ. Technol. Lett., 9: 317-324. Smith, F.B. and M.J. Clark, 1989. The transport and deposition of airborne debris from the Chernobyl nuclear power plant accident with special emphasis on the consequences to the United Kingdom. Met. Office Sci. Paper, No. 42, HMSO, London. Trimble, S.W., 1990. Geomorphic effects of vegetation cover and management: some time and space considerations in prediction of erosion and sediment yield. In: J.B. Thornes (Ed), Vegetation and Erosion. John Wiley and Sons Ltd., Chichester, pp. 55-65. Waber, U., H. Von Gutern and U. Krahenbuhl, 1987. The impact of the Chernobyl accident on a river/groundwater aquifer. Radiochim. Acta, 41: 191-199. Walling, D.E., 1983. The sediment delivery problem. J. Hydrology, 65: 209-237. Wailing, D.E. and S.B. Bradley, 1988a. Transport and redistribution of Chernobyl fallout radionuclides by fluvial processes: some preliminary evidence. Environ. Geochem. Health, 10: 35-39. Walling, D.E. and S.B. Bradley, 1988b. The use of caesium-137 measurements to investigate sediment delivery from cultivated areas in Devon, UK. In: M.P. Bordas and D.E. Walling (Eds), Sediment Budgets. Proc. Porte Alegre Symp., IAHS Publ. No. 174, pp. 325-335. Walling, D.E., T.A. Quine and J.S. Rowan, 1991. Fluvial transport and redistribution of Chernobyl fallout radionuclides. Hydrobiol. (in press).

CHERNOBYL-DER1VED RADIOCAESIUM WITHIN THE RIVER WYE BASIN. UK

131

Walling, D.E., J.S. Rowan and S.B. Bradley, 1989. Sediment-associated transport and redistribution of Chernobyl fallout radionuclides. In: R.F. Hadley and E.D. Ongley (Eds), Sediment in the Environment. Proc. Baltimore Symp., IAHS Publ. No. 184, pp. 37-46. Walling, D.E. and B.W. Webb, 1987. Suspended load in gravel-bed rivers: UK experience. In: C.R. Thorne, J.C. Bathurst and R.D. Hey (Eds), Sediment Transport in Gravel-bed Rivers. John Wiley and Sons, Chichester, pp. 691-734.