Oceanogr., 3 l(2), 1986, 266-276
0 1986, by the American Society of Limnology and Oceanography, Inc.
Limnol.
The effects and implication of tides and rainfall on the
circulation of water within salt marsh sediments’
Alan P. Carr and Michael
W. L. Blackley2
Institute of Oceanographic Sciences, Crossway, Taunton, Somerset, U.K.
Abstract
The response of the water table of an intertidal salt marsh in NW England has been investigated.
During neap tides, water seeps upward from the underlying sand-silt-clay interface. This effect
continues as the marsh is overtopped during spring tides but is then dwarfed by penetration of
water downward from the surface.
Under “typical conditions” the delay between high water in the river and maximum apparent
water level for the marsh becomes less as the spring-tide cycle maximum is approached. Lags also
appear smaller during summer, probably mainly due to desiccation cracks. Precipitation effects
are only important when rainfall is significant in quantity and coincides with low water during
neap tides.
This paper is one of a series (Carr and
Blackley 1985, 1986) examining aspect of
sedimentation and hydrology in the lower
zone of a salt marsh in Cumbria, northwest
England, and the relevance of such processesto the vertical distribution of radionuclides within the soil profile. Early research (e.g. Hetherington 1976; Aston and
Stanners 1979) suggestedthat radionuclides
were simply deposited on the marsh surface.
Later work has indicated that there is more
variability of accretion and erosion than was
initially proposed (e.g. Hamilton and Clarke
1984) but the possibility of radionuclide
profiles being related, at least in part, to
other factors has tended to be neglected. This
paper describes one such process.
We are especially grateful to P. Hardcastle
who was responsible for both the instrumentation and the recording system used
during the study. H. Hemond and W. Nuttle
provided comments on the original version
of this paper which went well beyond the
line of duty of referees and we are thankful
for their suggestions.
rivers Irt, Mite, and Esk (Fig. la, c). The
immediate coastline is composed largely of
sand dunes, but the estuary itself is more
complex. Although the bulk of the sediment
consists of sands and gravels underlain by
a coherent till, there are areas of silt and
clay, particularly along the banks of the estuary. Upon these marginal areas salt
marshes have developed, at least partially
as a consequence of the railway viaducts
constructed during the mid- 19th century.
The marsh discussed here, immediately
northeast of the Eskmeals viaduct (Fig. 1b),
is an example of such postrailway development. This ungrazed marsh extends over
an area of some 300 x 175 m and contains
a full range of plant communities distributed broadly in accordance with surface
height. Near the transducer site vegetation
consists principally of Puccinellia
maritima, Suaeda maritima, and Salicornia
spp. Detailed analysis of a core nearby
showed that down to 40 cm an average of
54% of the inorganic sediment consisted of
silt and clay-sized particles (> 6 $; < 15.6
pm) but from 40 to 60 cm this figure dropped
Description of the area
to 24%. Below 1 m only 2% of the material
The Ravenglass estuary, some 15 km was < 15.6 pm. All the rest of the mineral
south of the Sellafield (formerly Windscale) sediment was sand. The percentage of orreprocessing plant of British Nuclear Fuels ganic material was 9.9% between the surface
PLC, consists of the lower courses of the and 5 cm and fell progressively to 1.2% between 60 and 65 cm, with virtually none
LThis research has been carried out under contract
thereafter. At the time of sampling (Septemfor the U.K. Department of the Environment as part
ber 1983) water content, as a proportion of
ofits radioactive waste management research program.
dry weight, ranged between 54.7% at 5-10
2 Present address: IOS, Bidston Observatory, Birkenhead, Merseyside L43 7RA, U.K.
cm and 19.5% at 105-l 10 cm; the high near266
Salt marsh water circulation
267
Fig. 1. The experimental site in the context of the wider geographical setting.
surface figures reflect the higher proportion
of organic content and fine sediment there.
Although most of the clay in the estuary
consists of illite, some montmorillonite is
present (Kelly et al. 1982). The latter is reflected in the expansive nature of the clay
and the presence of desiccation cracks from
time to time at the research site.
Offshore the tidal range varies from about
6.5 m on springs to 3.5 m on neap tides and
each tide is essentially sinusoidal in form.
However, at the viaduct site in the river Esk
the pattern becomes highly asymmetric. The
inflowing tide is restricted to between 2 and
2.5 h and, while it reaches about 3.5-4.0-m
OD (Ordnance Datum, which is approximately Mean Sea Level) on springs and
2.5-m OD on neap tides, the water level
does not drop much below + l-m OD so
that the ebb flows for about 10 h per tide
(Fig. 2). The asymmetry is also reflected in
the salinity values. Hamilton and Clarke
(1984) quote values for high tide of N 3 1%~
falling to -4% for much of the low water
period. Assinder et al. (1984) give values of
1%~and Eakins et al. (in press) of 0.2%~as
a minimum. Even on a representative neap
high water we found a maximum of 32.5Y&
very close to typical seawater values, at
Ravenglass. Both water and salinity levels
may be modified by precipitation in the immediate area or within the Esk basin.
At low water most of the riverbed is exposed, showing the complex relation of ebb
and flood channels and local bedforms. The
situation at high water varies between conditions where the ground level of the lower
salt marsh is not quite reached to one where,
on equinoctial springs, even the highest
marsh is inundated.
268
Cat-r and Bkackley
OD
Cm
5.0
A
0
*
A
30 Jan
13 July a.m.
12 June p.m.
27 May p.m.
20 June a.m.
l
Tr =Transducer
-Tr #l
-Tr R5
1983
20 May p.m.
19 June p.m.
5 July a.m.
4 June p.m.
8 March
1.0 ‘,
-4
I
1
-2
I
I
0
I
I
2
I
I
4
I
I
6
I
I
8 Hours
Fig. 2. Examples of specific tides as recorded at Eskmeals viaduct. The upper diagram shows a range of
spring tides together with the surge of 30 January 1983. Note the variability of level during the low water period.
The lower diagram gives examples of neap tides. Ordnance Datum (OD) is about mean sea level.
Methods and results
Methods -During
the overall research
program various methods have been used
to assess the significance of different pro-
cesseson the vertical distribution of radionuclides in the salt marsh. We concentrate
here on the results obtained from six pressure transducers; other data collected during
Salt marsh water circulation
the study are discussed only when relevant
to apparent water movement within the
marsh sediments. In the present context
desiccation cracks and soil polygons developed at and below the mud surface may be
of particular importance.
One of the six Druck PTX gauge transducers used recorded water level in the river
Esk at Eskmeals viaduct; the others were
located between 25 and 85 cm below the
ground surface of a site in the lower salt
marsh (Fig. lb), where they recorded porewater pressure. At this site the marsh surface is at about +3-m OD and the junction
between the underlying sands and the silt
and clay of the marsh at about +2-m OD.
The marsh transducers were some 4 m away
from the nearest creek and about 1 m apart.
They were installed on 23 September 1982
and the river sensor on 21 January 1983.
The specification of the transducers for nonlinearity and hysteresis was < 20.1% max;
this gave short term repeatability of better
than f 0.35 cm of water, assuming constant
temperature. For temperature effects, the
specification was < *0.3-cm total error from
-2°C to + 30°C. The Druck transducers
converted gauge pressures (O-3 50 mbars) to
currents (4-20 mA). The specified measurable maximum pressure was equivalent to
a depth of - 350 cm of water. Since the river
transducer was at about + 1-m OD this
means that response to water levels over
-4.5-m OD could be attenuated. However,
in subsequent calibration tests, transducer
response proved essentially linear to 27 mA
(=5.0-m OD seawater), i.e. to 6.0-m OD.
The gauge outlets were vented to the atmosphere via smallbore tubes integral to the
cable construction. This provided good resolution and accuracy and obviated the need
for correction for atmospheric pressure. The
transducers had small porous pots, de-aired
before installation, mounted on the pressure
input to keep the pressure diaphragm uncontaminated by sediment. The porous pots
of the marsh transducers were surrounded
by sand, the rear sealed with bentonite and
the rest of the hole backfilled with the indigenous material. This should have prevented any direct seepage from above; in
any event, according to Vaughan (1974),
backfill material is unlikely to present a
269
problem since it can be some ten times more
permeable than the surrounding soil before
the influence on pore pressure is significant.
The data recording system converted the
transducer currents to frequency and totaled
this frequency on counters over a 40-s period. The final count was logged on EPROMs
(erasable programmable read only memory) every 12 min. When values for a transducer were identical on two successive 12min sampling intervals we assumed that high
water fell midway between, giving an effective 6-min sampling period. During quiescent periods no power was supplied to the
transducers. The resolution of the logging
system was -0.43 cm ofwater. The EPROM
capacity was 4,096 records; this gave a maximum interval of 34 days between EPROM
changes. Car-type batteries were changed at
the same interval. For test purposes the logging system also displayed the currents
drawn by each transducer. EPROMs were
returned to 10s (Taunton) for reading, their
data recorded onto magnetic tape, and a
quick-look printout was made.
Throughout the experiment periodic calibrations were made of river water level as
recorded by transducer 6 and as surveyed
by topographic leveling. Agreement was
consistently acceptable irrespective of the
level of the tide, and hence of the percentage
salinity; this was achieved by calibrating the
transducer in terms of seawater so that percentage error was minimum at high water.
Low salinities corresponded with low water.
As a result, when percentage errors would
be greater the transducer was only just covered, and absolute errors were small. The
same relationship between milliamps and
centimeters was used for all the marsh transducers (l-5) to produce an apparent water
level height for each sensor. Because of the
possibility of initial equilization problems
due to the drilling of the boreholes and because of the evident necessity of a river
transducer to clarify interpretation problems, we used no data before 21 January
1983.
There may be some conflict between the
measurement of pore pressures varying due
to drainage (falling tides), and changes attributable to undrained loading (rising tides).
Although drainage effects are likely to be
270
Carr and Blackley
Tide No. 10
HIGH TIDE LEVEL
River transducer
KEY
-
1 top
-....-
3 mtermedlate
---
5 bottom
0 Negatwe
pore
A Negatwe
threshold
pressure
‘, surge
value
a
1;
Highest
sprmg
hde
z
Lowest
neap
tide Cm m above
Ordnance
Datum)
10
HIGH TIDE LEVEL
River transducer
HIGH TIDE LEVEL
Ftwer Transducer
Fig. 3. Apparent water level of the top, middle, and lowest marsh transducers (1, 3, and 5) relative to the
observed high tide level in the river Esk estuary. The figure shows the greater variability in winter (a-21
January-l February 1983); the tendency toward closer agreement between marsh and river transducers over
spring tides and as summer progresses (b- 18 May-l 6 June; c- 16 June-18 July 1983); and the effect of the
intermittent diurnal tidal asymmetry (see text).
valid, pore pressure changes due to undrained loading may be rapid. The deformability of the piezometer may be different
from that of the surrounding soil so that
total pressure, and pressure changes, could
be atypical (Vaughan 1974). Despite these
qualifications,
the calculated apparent
heights seem plausible.
Evidence from pressure transducer records - Complete records are available for
408 tides between 21 January and 15 December 1983. Missing records are due to
storm damage, operational error, or, at the
end of the period, progressive instrument
failure attributable to condensation entering
the vented cables.
The data may be looked at in two ways:
as the actual or apparent height of the water
level for each high tide as given by trans-
ducer 6 and transducers 1-5, or as the lag
time between the peak water level in the
river and the corresponding peak porewater
pressure recorded by the transducers in the
marsh sediments. Although this relationship is primarily tidal, sometimes it may be
influenced by such factors as rainfall.
Figure 2 shows the changes in river water
level as recorded by transducer 6 for a representative range of spring and neap tides
and includes one of the two prominent surges
that affected the northwest coast of England
between 30 January and 1 February. It shows
for how short a period a tide can affect the
surface of even the lowest marsh areas: typically this time varies between barely 3 h at
springs to almost zero at extreme neaps.
Over much of the tidal cycle the water level
is constant at about + 1-m OD and the river
271
Salt marsh water circulation
Table 1. Mean delay times in minutes for transducers l-5 relative to river high water. Positive pore pressure
on transducer 1 and water level at transducer 6 > 3-m OD (i.e. overtopping marsh surface). Differences reflect
seasons, nature of spring tides, rainfall, and sampling periods. Lags for transducer 1 (top) are markedly less in
summer. Transducer 5 (bottom) is always Gust) in advance of transducer 4.
Transducer No.
2
1983
21 Jan-2 Feb
18 May-16 Jun
16 Jun-18 Jul
18 Jul-15 Aug
15 Aug-13 Sep
13 SeplO Ott
11 Ott-9 Nov
10 Nov-1 1 Dee
Total
13
37
38
33
38
35
44
44
282
15.2
10.9
5.2
z-2
7:o
11.3
16.1
9.9
flow continues to ebb. A similar tidal asymmetry has been described by Dankers et al.
(1984) for Delfzijl, Netherlands. The river
Esk “low water” level is partly the effect of
the viaduct which impedes drainage at the
recording site.
Figure 3 shows the apparent water level
in the marsh during winter and late springearly summer. All these values are plotted
relative to the respective high tide level in
the river nearby and thus eliminate the effect of the undrained response of the soil
when loaded by the tide. A number of features are apparent. Figure 3a gives some
indication of the variability possible when
substantial rainfall and tidal surges are present. Panels b and c are more representative
of typical conditions, although minor aberrations in, especially, the top transducer
records may reflect such factors as rainfall.
The records between 18 May and 18 July
1983 cover four periods of spring tides. Each
successive series shows increasing convergence of the top and intermediate marsh
transducers toward the river high-tide level.
This may be partly a response to the decrease in viscosity of water during summer
but is most likely due to an increase in vertical cracking at that time. This enables water
from those tides that overtop the marsh surface to penetrate rapidly below ground level.
Lag times (Table 1) are least during summer
months.
The diurnal asymmetry of the tides varies
between 0 and 0.6 m between successive
tides, depending on the time in the biweekly
tidal cycle and during the calendar year. At
17.5
19.3
15.9
10.2
16.4
16.3
22.9
26.1
18.6
16.6
13.9
14.4
10.5
15.3
21.2
17.6
19.2
16.2
15.2
15.4
14.9
13.8
16.7
14.4
19.5
17.6
16.2
14.3
14.4
14.4
13.7
16.4
14.1
19.0
16.2
15.6
spring tides it is most marked on the lowest
transducer (that which feels the tidal effects
for the longest time) and is well displayed
between tides 17 and 30 of Fig. 3c. At neap
tides all the marsh transducers may be affected in a similar fashion, whether or not
they show negative pore pressure, because
at neaps the marsh transducers are less able
to respond to the fluctuations of the tidal
curve in the river which then operates at a
lower level relative to the marsh sensors.
Tides 27-39 in Fig. 3b and 30-38 in Fig. 3c
demonstrate this. In the former example the
top transducer merely exhibits draining
throughout the period, as is also the case
between tides 1 and 7, and therefore cannot
be plotted. There are a few occasions, shown
as “negative pore pressure threshold values,” where the actual pore pressure is not
discernible because it is at or below the
threshold current consumption of the transducer. Although these records may exaggerate the difference between apparent water
levels at that height in the marsh and the
corresponding river high water, they clearly
do not do so to any great extent since the
changes are mostly in close agreement with
“real” data from the other, lower, marsh
transducers.
The closer height agreement on a seasonal
basis of transducer 6 with, especially, transducer 1 is probably linked with the shorter
lag time between about July and September.
This feature is well shown in Table 1. For
the period 16 June-l 8 July, there were 38
tides where the river water level exceeded
3.0-m OD (i.e. the ground height above the
272
Carr and Blackley
Table 2. Mean delay time in minutes for those occasions when transducer 1 had negative pore pressure
and river transducer high water % 2.75-m OD. Four
tides were indeterminable for transducers 3 and 4, 30
for transducer 2.
Table 3. Mean delay times in minutes for all available data between 21 January and 11 December 1983
for transducers 3-5. Indeterminate events have been
omitted.
Transducer No.
Transducer No.
Total
n
2
129
-
3
~47.8
4
5
~42.5
40.9
marsh transducers) and pore pressure was
positive on transducer 1. The average delay
time was 5.2 min. Comparable figures for
2 1 January-2 February were 15.2 min and
for 10 November-l 1 December 16.1 min.
Further data on lag times are given in Tables
2 and 3. Table 2 summarizes the occasions
when the river level at high water was
12.75-m OD (i.e. the approximate height
of the top sensor) and transducer 1 recorded
negative pore pressure. This condition is
typical of neap tides where there had been
no substantial rainfall to complicate the picture; as a result progressive drying-out of
the marsh surface zone occurred. Under
these conditions, delay times for transducer
1 (and often transducer 2) could not be resolved, but values for transducer 5, and
nearly always transducers 3 and 4, could
be calculated. Although the records from all
three bottom marsh transducers show long
delays, they were least for the lowest sensor.
Table 3 gives mean delay duration for transducers 3-5 for the total number of tides.
Differences may not be great, but the lowest transducer (No. 5) invariably has a
shorter lag than transducer 4 and there iswith one marginal exception-a similar relationship between transducers 4 and 3.
While the effective sampling interval was 6
min, there are a considerable number of
events even over a single “monthly” sampling period. It is therefore statistically acceptable to compare differences in mean
values for less than the effective 6-min measurement interval.
The representative event: discussion-The
shorter lag times of summer and early autumn may be attributed either to the cracking ofthe marsh surface, providing a rapid
water pathway downward, to the reduced
viscosity of water, or both. Viscosity of fresh
Total
n
3
4
5
557
26.0
24.1
22.5
(and sea)water is reduced by - 20% between
8” and 16°Cand -44% between 0” and 20°C.
To investigate the depth of vertical fissures in the marsh clays and silts, we selected a well developed group of soil polygons in an unvegetated area of the upper
salt marsh at the northwest corner of the
site. Cracks were seeded with barium sulfate. After the seasonal moisture deficit had
been eliminated, the area was cored and the
cores X-rayed. Barium sulfate was found
down to at least 23 cm, and potential pathways were apparent to about twice this
depth. These figures approach those of Reeve
et al. (1980) for cracking at inland sites with
clayey soils during dry summers. It is clear
that water (and by inference radionuclides
contained within it) from any tide level exceeding that of the fissured marsh surface
would immediately penetrate down into the
soil matrix. Such bypassing effects have been
described from other environments (e.g.
Kneale and White 1984, for a clay grassland
topsoil).
Vertical cracking of the vegetated marsh,
like that where the transducers were installed, is never very conspicuous. Nevertheless, the fact that, over spring tides in
summer, lag time, vis-a-vis the river high
water, is proportionately less between the
air-marsh interface and the top transducers
than between the underlying sand-marsh
interface and the bottom transducers points
to the predominance of this factor. Ravina
(1983, p. 154) cites Ritchie et al. who “have
shown that wet field clay soils, where water
movement may occur in invisible cracks and
slickensides, have hydraulic conductivities
25 times greater than the same soil which
has been repacked in the laboratory.” The
observation of Ravina (1984) that cracking
is dependent on the rate of moisture change
as well as its magnitude is particularly pertinent in the context of an intertidal area.
273
Salt marsh water circulation
External
NEAP TIDES
I
External
W.L.
-- 0.4
o(4);
4-5
i
I
I
l (3)
-- 0.6
(al
.‘.
.
-
*
’ ’
5 02
mm-l
:,
*.
.
* . *.
B
’
*7
.
..
. *a -,
Lag times appear
to be less during
summer months
l
4.3
-- 0.8
01
-4
3
Silt f Clay
02
-+2m--
water period)
2.
:
: Sand & gravel
,
l
**.c . 1 . l . . , .
---
l
-.
. . .:* - . . * *- . I .
Chl
*
Diagram of basic flow pattern : Ravenglass
Fig. 4. Summary diagram based on porewater pressure data, river Esk salt marsh. a. Although in example
water level reaches transducer at 42 cm, lag effects prevent water from reaching this height before river level
falls. Some capillary effects above. b. Fate of intermediate transducers depends on external water levels, duration,
etc. Marsh may or may not become fully saturated in time available.
Bouma (1980) found that hydraulic conductivities varied dramatically on a seasonal basis where swelling clay soils, like that
here, were present.
Water transport in the immediate surface
zone could also be enhanced to some extent
by the pathways provided by plant roots
(Bouma and Dekker 1978). This effect is
likely to be more widespread than vertical
cracking. Although much of the vegetation
at the transducer site is perennial, both root
growth and the tendency to surface desiccation would probably be maximal during
summer. Most roots at the site are within
the top 30 cm.
We referred above to the lowest marsh
transducer having the least mean delay time
overall, as well as at neap tides when the
top transducers were often merely subject
to prolonged drying. This, coupled with the
fact that transducer 5 always responded to
tidal fluctuations in the river, lends support
to the argument that water can circulate upward from the underlying silt-sand interface. Hamilton and Clarke (1984) also believed that for most areas of the estuary of
the river Esk there was upward and lateral
seepage, but they thought that this was
freshwater from land drainage. The importance of vertical water movement will vary
between spring and neap tides and in relation to the incidence of precipitation; the
situation is summarized in Fig. 4. At a site
adjacent to the river Severn, groundwater
levels responded rapidly to drawdown in the
neighboring channel; Thorne (1978) attributed this to the high permeability of the
underlying deposits, although even in the
silt and the clay above permeabilities of 2
m h-l were measured. Thorne believed that
vertical flow was dominant over lateral at
the site.
The rainfall eflect -Figures 5 and 6 record the effects of two periods of rain between 3 1 August and 18 September 1983,
both corresponding to neap tides. In neither
instance did the high water level of the river
quite reach the surface of the marsh (“3 m)
although it sometimes exceeded the level of
the uppermost transducer (transducer 1 =
2.73-m OD). Rainfall figures were observed
hourly at the meteorological station at Eskmeals.
Figure 5 shows, first, that the apparent
water level of transducer 1 is strongly influenced by rainfall, with the trace being al-
274
Rainfall
Carr and Blackley
Water level (ml
h -’
d
6-
Apparent
$1 ILl)
(mm
I[ Ralnfall
(hourly records
at Eskmeals)
- 2.9
/
River level
/
(Trans # 6)/
t
I
2.5
-2.8
-2.7
-2.6
- 2.5
17 Sep
GMT
18 Sep
Fig. 5. The effect of rainfall occurring at low water during neap tides. Rainfall figures are those from the
nearby Meteorological Office station at Eskmeals, 17-18 September 1983. Heights are relative to Ordnance
Datum (OD).
most the mirror image of the water level in
the river. Second, the rainfall was sufficient
to raise the water level in the river by up to
28 cm immediately before the morning high
tide of 18 September. This probably reflected the earlier precipitation in the basin of
the river Esk, although a similar effect could
also be generated during periods of onshore
winds. Figure 6 depicts data for the rather
longer period between 3 1 August and 3 September and includes transducers 2 and 3 in
addition to those shown in Fig. 5. There
were two substantial rainfalls early on 1
September and again around noon on 2 September, with smaller quantities thereafter.
Until the second rain the water level of the
river was little affected, but on that occasion
there was some elevation at low water, of
the same magnitude as that of 18 September
(Fig. 5). Not only did the much smaller rain
at midday 3 September produce inflections
in the water level curve, but low water for
the whole of that period was atypically high.
The uppermost marsh transducer (No. 1)
showed a complex picture. It responded
substantially to the first rain and, except for
very minor “kicks” related to the tidal cycle,
proceeded to drain from that time until the
next major event. There was then again a
substantial response with rapidly increasing
apparent water level. Thereafter, even small
quantities of rain caused conspicuous
changes in porewater pressure. The small
fingerlike peak in the afternoon of 2 September may be a reflection of the near-simultaneous high tide in the river; this high
tide is apparent on the record from the underlying transducer 2. Transducer 2 is interesting because although it was only 27
cm below transducer 1, the form of the profile is more closely comparable to that of
the river even if the magnitude is different.
Salt marsh water circulation
Rainfall
275
h -’
(mm> I
(: Ralnlall
(hourly
1;>-
records
at Eskmealsl
Water
level
Observed
(6)
2h
-
1c)-
EI-
.
6
Cm)
Apparent
(1.2.3)
\
\
...\I
2400
31 Aug
1200
2.5
2.0
1.5
2400
1200
2400
1200
2
1.0
3
3 Sep
GMT
Fig. 6. As Fig. 5, but for 31 August-3 September 1983. Transducer 2 data are a compromise between the
rainfall effects shown clearly from transducer 1 and the river water level recorded by transducer 6. Transducer
3 is almost entirely tidally dominated. The marsh sensor heights are at 2.73, 2.56, and 2.42 m.
Transducer 3 is hardly affected by rainfall
events at any time during the August-September example described, although it is
only 14 cm below transducer 2.
The rainfall data described above show
that the effect of precipitation on porewater
pressures, water circulation, and water
levels-both apparent and real - depends
largely on the time of precipitation. Rain
falling at high water spring tides would have
only an indirect effect, such as marginal
changesin pH and salinity and perhaps some
eventual increase in water flow and water
level at the subsequent river low waters.
However, rainfall at low water neap tides
can produce substantial modifications to the
normal neap-spring tidal picture, as Figs. 5
and 6 indicate. Such effects appear related
to both the amount of precipitation and the
duration of the storm. It is not surprising
therefore that there is some variability in
the apparent/actual water level and lag relationships both between individual marsh
transducers and between the marsh trans-
ducers and the transducer recording river
water level.
Trudgill et al. (1983) have shown that, in
a site having rendzina and brown earth soils,
the output of labeled soil-water was related
to rains exceeding 3 mm h-l and lasting 2
h or more. They emphasize the importance
of preferential flow paths as well as the hydraulic gradient. Gillham (1984) has commented on the highly disproportionate
manner in which shallow water tables may
be affected by precipitation; he noted the
large and rapid response to the incidence of
rainfall and observed that the legacy of past
events could be significant.
Conclusions
The work described here has shown that
the porewater pressure response both to the
regular changes in tidal water level and to
the intermittent
precipitation-induced
changes appears to be rapid. Although some
of the changes may be an artifact of the
instrumentation (Vaughan 1974) the re-
276
Carr and Blackley
sults, taken overall, produce a coherent picture. Water circulation, as reflected in apparent levels within the marsh, is manifestly
different between spring and neap tides, and
somewhat different between high and low
spring ranges. Lag times vary both between
particular cycles of spring tides and on a
seasonal basis. This picture is modified by
rainfall effects insofar as precipitation is
coincident with low water on neap tides.
The preferred water pathway via vertical
cracks and the lag/tidal cycle relationship
suggest mechanisms by which radionuelides, primarily of a conservative type (i.e.
in the water phase), may be circulated within the salt marsh sediment.
References
ASSINDER, D. J., M. KELLY, AND S. R. ASTON. 1984.
Conservative and non-conservative behaviour of
radionuclides in an estuarine environment, with
particular reference to the behaviour of plutonium
isotopes. Environ. Technol. Lett. 5: 23-30.
ASTON, S. T., AND D. A. STANNERS. 1979. The determination of estuarine sedimentation rates by
*34Cs/137Cs
and other artificial radionuclide profiles. Estuarine Coastal Mar. Sci. 9: 529-54 1.
BOUMA, J. 1980. Field measurements of hydraulic
properties characterizing water movement through
swelling clay soils. J. Hydrol. 45: 149-l 58.
AND L. W. DEKKER. 1978. A case study on
imiltration into dry clay soil: 1. Morphological
observations. Geoderma 20: 27-40.
CARR, A. P., AND M. W. BLACKLEY. 1985. Implications of sedimentological and hydrological processeson the distribution of radionuclides in a salt
marsh near Sellafield, Cumbria. Inst. Oceanogr.
Sci. Rep. 197. 88 p.
AND -.
1986. Implications of sedimentological and hydrological processes on the distribution of radionuclides: The example of a salt
marsh near Ravenglass, Cumbria. Estuarine
Coastal Shelf Sci. 22: in press.
DANKERS, N., M. BINSBERGEN, K. ZEGERS, R. LAANE,
AND M. RUTGERS VAN DER LOEFF. 1984. Transportation of water, particulate and dissolved organic and inorganic matter between a salt marsh
and the Ems-Dollard estuary, The Netherlands.
Estuarine Coastal Shelf Sci. 19: 143-l 65.
EAKINS, J., P. BURTON, D. C. HUMPHREYS, AND A. E.
LALLY. In press. The remobilisation of actinides
from contaminated intertidal sediments in the
Ravenglass estuary. Zn The behaviour of radionuclides in estuaries. Proc. Int. Seminar, Renesse,
The Netherlands, 1984.
GILLHAM, R. W. 1984. The capillary fringe and its
effect on water-table response. J. Hydrol. 67: 307324.
HAMILTON, E. T., AND K. R. CLARKE. 1984. The recent sedimentation history of the Esk estuary,
Cumbria, UK The application of radiochronology. Sci. Total Environ. 35: 325-386.
HETHERINGTON, J. A. 1976. The behaviour of plu. tonium nuclides in the Irish Sea, p. 8 l-l 06. Zn M.
W. Miller and J. N. Stannard [eds.], Environmental toxicity of aquatic radionuclides: Models and
mechanisms. Ann Arbor Sci.
KELLY, W. R., S. R. ASTON, AND D. ASSINDER. 1982.
The factors affecting the distribution of radionuelides in Irish Sea sediment profiles: A review, and
an analysis of 10s 1979 cores. Univ. Lancaster,
Dep. Environ. Sci. Rep. TR/5. 57 p.
KNEALE, W. R., AND R. E. WHITE. 1984. The movement of water through cores of a dry (cracked)
clay-loam grassland topsoil. J. Hydrol. 67: 361365.
RAvIN,~, I. 1983. The influence of vegetation on
moisture and volume changes. Geotechnique 33:
151-157.
-.
1984. The influence of vegetation on the
swelling and shrinkage of clays: Discussion. Geotechnique 34: 171.
REEVE, M. J., D. G. HALL, AND P. BULLOCK. 1980.
The effect of soil composition and environmental
factors on the shrinkage of some clayey British
soils. J. Soil. Sci. 31: 429-442.
THORNE, C. D. 1978. Processes of bank erosion in
river channels. Ph.D. thesis, Univ. East Anglia.
447 p.
TRUDGILL, S. T., A. M. PICKLES, K. R. SMETTEM, AND
R. W. CRABTREE. 1983. Soil-water residence time
and solute uptake: 1. Dye tracing and rainfall
events. J. Hydrol. 60: 257-279.
VAUGHAN, P. R. 1974. The measurement of pore
pressures with piezometers, p. 41 l-422. Zn Field
instrumentation in geotechnical engineering. Proc.
Symp. Brit. Geotech. Sot. Butterworths.
Submitted: 24 January 1985
Accepted: 13 September 1985