Delivery of marine chloride in precipitation and removal by rivers in the Murray-Darling Basin, Australia

Delivery of marine chloride in precipitation and removal by rivers in the Murray-Darling Basin, Australia

Journal of Hydrology, 154 (1994) 323 350 0022-1694/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved 323 [2] Delivery of marine chloride i...

1MB Sizes 0 Downloads 23 Views

Journal of Hydrology, 154 (1994) 323 350 0022-1694/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved

323

[2]

Delivery of marine chloride in precipitation and removal by rivers in the Murray-Darling Basin, Australia H. J a m e s S i m p s o n *'a, A n d r e w L. H e r c z e g b aLamont-Doherty Earth Observatory and Department of Geological Sciences, Columbia University, Palisades, NY 10964, USA bCentre for Groundwater Studies and CSIRO, Division of Water Resources, PMB No. 2, Glen Osmond, S.A, 5064, Australia (Received 14 April 1993; accepted 30 May 1993)

Abstract Rising groundwater tables and increasing river salinities are major problems in the semi-arid MurrayDarling (M-D) Drainage Basin of SE Australia. Chemical data from precipitation, rivers and groundwaters are integrated here to estimate chloride fluxes in the context of dramatic alteration of the hydrologic balance in the basin initiated during European settlement approximately 150 years ago. Our analysis indicates that about half of C1 deposition from the atmosphere in the western Murray Basin is derived from resuspended regional continental dust, resulting in high Ca/Na ratios in bulk precipitation samples. Estimates of recent marine deposition of CI to the M D Basin were made by excluding rainfall data from sites having high Ca/Na ratios. Basin-wide budgets for CI delivery from the atmosphere and removal by rivers indicate that the Darling Basin retains about 90% of current annual input of marine CI whereas the River Murray exports two to three times current annual input to the Murray Basin (exclusive of the Darling). Some implications from this study are as follows: (1) Re-evaluation of net marine C1 input by rain implies that recent model calculations of 36C1/C1inputs from the atmosphere to the Murray Basin may differ significantly from actual values. In particular, net 36C1/ CI input ratios from the atmosphere to the western Murray Basin may be a factor of two higher than those derived assuming all CI in rain is of recent marine origin. (2) The very large inventories of salt within the two catchments, especially the saline groundwaters of the Murray Basin, accumulated over very long periods, probably of the order of a few million years. The current rearrangements of stored salinity as a result, at least in part, of human perturbations of the hydrologic balance are likely to continue for an extended period because of the massive quantities of salts involved, and thus present major limitations on options for long-term management of the M - D Basin.

* C o r r e s p o n d i n g author.

SSD1 0022-1694(93)02345-X

324

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

1. Introduction

1.1. Background Surface waters in large areas of southeastern and western Australia have become progressively more saline during this century, owing to extensive replacement of native vegetation with shallow-rooted plants for grazing and cultivated crops (Allison et al., 1990). After removal of native deep-rooted Eucalyptus trees, which generally permit less than 1% of annual precipitation volume to pass the root zone, groundwater recharge has increased in some areas by several orders of magnitude. Because unconfined groundwaters in many areas of Australia have salinities comparable with seawater, enhanced recharge and consequent higher water tables can lead to (1) mobilization of high-salinity unsaturated zone moisture and (2) increased discharge of saline groundwater into surface waters. The most generally accepted explanation for the origin of the large amounts of dissolved salts in these groundwaters is delivery of marine aerosols to the land via rain plus dry deposition over very long periods of time. Nearly all of the rain-water is returned to the atmosphere by transpiration and evaporation, leaving residual soil moisture which eventually reaches the saturated zone having chloride concentrations [C1] that are orders of magnitude greater than the original rain. Stable isotope and radiocarbon compositions of these groundwaters (Leaney and Allison, 1986; Arad and Evans, 1987; Herczeg et al., 1989) and of the overlying soil moisture in the unsaturated zone (Allison and Hughes, 1983) provide very strong evidence in support of this mode of salinity evolution for Murray-Darling (M-D) Basin groundwaters. Possible alternative modes of formation, such as retention of seawater from previous marine intrusions millions of years ago, or solution of massive evaporite mineral deposits from aquifer formations, are not consistent with observed carbon, hydrogen and oxygen isotope compositions, or with observed stratigraphy and the generally accepted geological history of the basin. Assuming that extreme concentration of precipitation-derived ions by loss of water to the atmosphere is the dominant mode for formation of saline groundwaters in much of Australia, chemical properties of rain in SE Australia can be examined to assess the current balance of flux terms related to salinity budgets in the M - D Basin. Most of the rain chemistry data to be discussed are from the catchment of the River Murray and that of its longest tributary, the Darling River, which together form the most extensive and economically important network of interior surface drainage in the country. The current role of riverine transport in removing dissolved salts from the M - D Basin will then be considered. Finally, the approximate minimum

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

325

length of time required to accumulate the huge quantities of dissolved salts in the subsurface of SE Australia via rainout of marine aerosols will be estimated.

1.2. Cyclic salts in surface runoff A significant fraction of some dissolved ions transported by rivers throughout the world are derived from the atmosphere (i.e. by precipitation plus dry deposition) rather than from weathering reactions involving soils and rocks. The 'cyclic salt' can dominate the total river supply of C1 in drainage basins which are largely free of evaporite minerals exposed to meteoric waters. The importance of these cyclic salts to ions carried by rivers has been recognized for many decades (e.g. Conway, 1943; Junge and Gustafson, 1957; Stallard and Edmond, 1981). The C1 in most continental precipitation, except in regions with very high acidic atmospheric pollutant emissions from coal combustion such as the NE USA and NW Europe (Wagner and Steele, 1989; Moller, 1990), originated as marine aerosols generated at the ocean surface within a few days to a few weeks before deposition. The other important natural source of atmospheric C1 in arid continental regions involves wind entrainment of fine soil particles that include evaporite minerals as well as silicate and carbonate phases. Chloride derived from local soil dust can result in overestimation of the 'new' marine component of local atmospheric deposition, and should be excluded from cyclic salts representing additional delivery of marine salts to the land, as it involves redistribution of C1 between continental sites. 2. Marine chloride in precipitation over continents

2.1. General trends of precipitation [Cl] for Australia, North America and South America Precipitation [C1] over continental interiors tends to be substantially lower than for marine and coastal areas less than a few hundred kilometres from the sea. This difference results from the combined effects of (a) deposition of marine aerosols to the ground, primarily by precipitation, and (b) more intense vertical mixing of the lowest few kilometres of the atmosphere over land compared with over oceans (Junge, 1963). This latter process effectively dilutes the marine aerosol component in the lower atmosphere as air moves inland because a much greater fraction of the total tropospheric depth affects the chemistry of precipitation over the interior of continents.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

326

The rate of decline of [C1] in continental precipitation as a function of distance inland varies greatly from one continent to another as does its mean value in the interior (Junge, 1963). Precipitation [C1] plotted as a function of distance inland for rain samples from three continents illustrates some of these variations (Fig. 1). The trend line for the Amazon Basin was derived by combining data for both precipitation and large tributary rivers draining lowland rainforest sub-basins free of evaporite mineral exposure (Stallard, 1980), as representative of the mean rain [C1] of that area (Stallard and Edmond, 1981). We can describe some characteristics of these groups of observations by fitting exponential curves to data immediately inland from the sea-coast and constant values for continental interiors (Table 1). Precipitation [C1] in Victoria, in SE Australia (Hutton and Leslie, 1958) decreased rapidly with distance from the coast, but had a relatively high mean value for the interior (31 #equiv 1-1). At the other extreme, the rate of decrease as a function of distance inland of precipitation and tributary river [C1] from the Amazon Basin was much less than in SE Australia, but reached low mean values in the basin interior (7 #equiv 1-1). Two transects inland from the coasts 1000 Victoria, AU

./ o

100 )o°

A

o

Amazon, BZ

O:o :O

O"

O

A

E USA oAoo •

10



SE USA 1

0

i

i

i

i

!

200

400

600

800

1000

1200

distance to ocean (km) Fig. 1. Precipitation [CI] vs. distance inland for Victoria, Australia (open circles), E (filled squares) and SE (open triangles) USA, and Amazon River lowland tributary rivers plus rain (filled diamond + line). Data from Junge and Gustafson (1957) (USA); Hutton and Leslie (1958) (Australia); Stallard and Edmond (1981) (Brazil).

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

327

Table 1 Flux-weighted precipitation chloride concentration (Cli) trends vs. distance to nearest sea-coast (x), assuming exponential decrease from the sea-coast (C10) until a constant value for the continental interior is reached (Cls) Location

No. of [CI] values in mean

x, where Cli = Cls (km)

x, where CI0 at CIs Cli/Cl0 = ½ x = 0 mean (km) (#equiv 1-1) (#equiv 1-1)

(a) (b) (c) (d)

Victoria, Australia Eastern USA SE USA Amazon, Brazil

16 4 6 6

100 330 300 1700

20 60 200 600

(e) (f) (g) (h) (i) (j)

Coastal Interior Coastal Interior Eastern Eastern

-

-

SE Australia 5 SE Australia 35 USA 16 USA 47 USA 6 USA 8

-

800 160 17 57

-

184 144 -

-

-

31 5 4 7 52 6 6 8

(a) Total of 32 sites (wet deposition only; Hutton and Leslie, 1958) (see Fig. 1). (b) Total of seven sites (wet deposition only; Junge and Gustafson, 1957) (see Fig. 1). (c) Total of eight sites (wet deposition only; Junge and Gustafson, 1957) (see Fig. 1). (d) Total of seven sites, precipitation (wet deposition only) and tributary river [CI] from the Amazon Basin (Stallard and Edmond, 1981) (see Fig. 1). (e) Total of 40 sites, annual mean [C1] of monthly composites of rain plus dry deposition from three coastal sites: 35-100 km to ocean; 1974-1977 (Blackburn and McLeod, 1983); data for each year included separately. (f) Total of 40 sites, annual mean [C1] of monthly composites of rain plus dry deposition from 25 interior sites: 150-680km to ocean (Blackburn and McLeod, 1983); data for each year included separately. (g) Total of 63 sites, annual means of monthly composites (wet deposition only) from 16 sites: 7/ 1955-7/1956 (Junge and Werby, 1958); 16 sites designated here as 'coastal'. (h) Total of 63 sites, annual means of monthly composites (wet deposition only) from 47 sites: 7/ 1955-7/1956 (Junge and Werby, 1958); 47 sites designated here as 'interior'. (i) Amount-weighted mean values for individual precipitation events (wet deposition only) from interior sites over a period of 11 months from MAP3S network with eight sites: 8/1978-6/ 1979 (Pack, 1980). (j) Amount-weighted mean values for individual precipitation events (wet deposition only) from interior sites over a period of 11 months from EPRI network with nine sites: 8/1978-6/ 1979 (Pack, 1980). o f e a s t e r n a n d S E U S A ( J u n g e a n d G u s t a f s o n , 1957) i n d i c a t e d r a t e s o f d e c l i n e in p r e c i p i t a t i o n [C1] as a f u n c t i o n o f d i s t a n c e i n l a n d i n t e r m e d i a t e b e t w e e n t h o s e f r o m S E A u s t r a l i a a n d t h e A m a z o n Basin. T h e m e a n [C1] r e a c h e d f o r these t w o sets o f d a t a f o r t h e i n t e r i o r o f the U S A w e r e 5 a n d 4 # e q u i v 1 - 1 , respectively, nearly an order of magnitude lower than for Victoria, Australia.

328

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

Data from other investigations of precipitation chemistry also indicate that mean annual [C1] in rain over the interior of SE Australia (150-680 km from the nearest sea-coast; Table 1, f) is considerably higher (52 #equiv 1-1) than for mean annual precipitation [C1] over the interior of the USA (6-8 #equiv 1-1; Table 1, h-j). 2.2. Estimation of the 'new' marine component of chloride in rain over SE Australia

Relatively high total deposition of C1 per unit of rainfall over the interior of SE Australia could be the result of resuspension of soils containing dispersed evaporite minerals (Hutton and Leslie, 1958), enhancing dry deposition between precipitation events and also elevating rain [C1]. This possibility can be explored by considering mean annual rain compositions for the most extensive network of collection sites for precipitation chemistry reported for SE Australia (Blackburn and McLeod, 1983). Data for total annual wet plus dry deposition of major ions, annual precipitation amount and mean chemical composition of rain, assuming all deposition occurred in rain, have been reported for 28 locations (Fig. 2). Seven of these sites were sampled for more than 1 year (1974-1977), and the remainder represent a single year of data (1974). We have treated each mean annual value as a single flux-weighted observation, resulting in a total of 41 sets of annual mean rain compositions derived from 28-day sampling intervals. Five of these were originally categorized (Blackburn and McLeod, 1983) as coastal values (Table 1,e), 35 as interior values (Table 1, f) and one was excluded as a result of evidence of extensive local contamination from road dust. We subdivided these interior values into two categories, based primarily on relative abundances of Ca and Na (Fig. 3(A)). High ratios of Ca/Na (greater than unity) were considered as indicative of an important regional dust component (Group B) because Ca salts (i.e. CaCO3, CaSO 4 • 2H20 ) tend to be among the first to precipitate out of solution at the soil surface. Subsequent resuspension as wind-borne dust and partial dissolution of aerosols during rain episodes would lead to higher Ca/Na ratios in rain. Any collection site having a single year of high mean annual Ca/Na was retained in Group B for all years of sampling. The mean Ca/Na ratio in Group B was 2.1, compared with 0.30 for Group A interior samples and 0.19 for coastal samples, Group C (Table 2). The Mg/Na ratio also indicated significant differences between populations of interior sites for SE Australia (Fig. 3(B)), although the range of variations was considerably smaller than for Ca/Na. In contrast, the SO4/C1 equivalents ratios in the two groups of interior site values (Table 2, Fig. 3(C)) are not clearly distinguished (0.39 and 0.48),

H.J. Simpson, A.L. Herczeg I Journal of Hydrology 154 (1994) 323-350

A uSIT

RA'

i

329

Q b. . . . . . . .

Queensland

® kust r alia

i

O~\~0

(9

N

~

o I

lOO I

2~

3~

4~

i

I

I

5~tm I

I [

t

Fig. 2. Location of rain collection sites in SE Australia (site numbers from Blackburn and McLeod (1983)). Circles, interior sites judged free of significant regional continental dust contamination (see text for criteria); squares, interior sites judged to have appreciable dust contamination; diamonds, coastal group; M - D Drainage Basin is outlined.

indicating that additional rain Ca was probably derived from calcite and other carbonate minerals rather than gypsum components in continental dust. Because of high solubilities of C1 salts, it is likely that much of the proportion of C1 in the bulk precipitation samples attributed here to dust contamination was delivered via dry fallout rather than higher [C1] during rain episodes. Mean annual rain [C1] in SE Australia as a function of distance inland for Group A sites was somewhat lower than the general trend for all sites (Fig. 4(A)). The [C1] values in Group A sites showed little variation as a function of annual precipitation a m o u n t (Fig. 4(B)). The relatively constant values of mean annual rain [C1] for Group A sites as a function distance inland and annual precipitation a m o u n t indicates that estimation of total delivery of

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

330

i

300

,

,

,

l

,

I

I

,

n

A)

o" 200 ,.,1 v

o • •

I

,

I

A sites B sites C sites

• • •

c

oD



100



Uw

• •



0 0.0

•U



,

,

,

,

u

u

u

i

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

,

i

300 B)'



I

,

4.5

(eq/eq)

Ca/Na







,

I

I

I



200

A sites B sites C sites

v

t,,o~





O~ll



100 U

0.0

•m



i

u

i

i

i

0.2

0.4

0.6

0.8

1.0

1.2

Mg/Na (eq/eq) 300

I

C) o" o

=

o

200

A sites



B sites



C sites

& •

0

100



! u m

o I

0 0.0

0.2

,

I

i

0.4

I

0.6

i

I

0.8

n

I

1.0

i

1.2

S04/Cl (eq/eq)

Fig. 3. (A) SE Australia flux-weighted annual mean rain [CI] vs. [Ca]/[Na] in equivalents (Blackburn and McLeod, 1983). Group A sites are judged to have minimal dust contamination; Group B sites are judged to have dust contamination during at least 1 year of sampling; Group C sites are categorized as coastal. (B) SE Australia annual mean rain [C1] vs. [Mg]/[Na] in equivalents. (C) SE Australia annual mean rain [CI] vs. [SO4]/[CI ] in equivalents.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

331

Table 2 Mean annual atmospheric deposition for SE Australia (Blackburn and McLeod, 1983), with data aggregated into three groups: A, interior; B, interior with regional continental dust contribution; C, coastal (see list of sites at end of table) Parameter

Group A

Group B

Group C

Mean distance to ocean (km) Mean annual P (mm) No. of annual site values C1 deposition (kgha 1 year-l) [Cl] m i n i m u m b (#equiv 1-1) [C1]c (#equivl ~) [CI]d (#equivl 1) [Na] d (#equiv 1- l ) [Ca] d (#equivl 1) [Mg] d (#equivl l) [SO4] d (#equivl l) Na/CI d Ca/Na d Mg/Na d 804/C1 d

343 4- 143a 496 4- 164 19 6.8 4- 2.2 16 4- 4 39 4- 13 40 4, 8 51 4- 12 15 4, 11 9± 3 15 4- 6 1.28 4- 0.13 0.30 ! 0.17 0.18 + 0.04 0.39 ± 0.12

336 i 154 373 4. 150 16 7.2 4- 2.3 35 4, 10 54 4, 17 66 4- 24 75 4. 28 148 4. 83 33 4. 23 31 4. 12 1.14 4. 0.15 2.1 2:1.2 0.45 + 0.29 0.48 + 0.19

69 ± 23 768 4, 122 5 50 4. 17 88 ± 31 183 ± 59 184 + 59 184 i 60 33 ± 14 38 ± 12 30 4- 8 1.00 4- 0.05 0.19 -t- 0.10 0.21 4, 0.04 0.17 4. 0.03

a Uncertainties (±) reported for parameters are sample standard deviations (an 1). b [C1] M i n i m u m mean value for periods with high percipitation a m o u n t (50-150 m m of rain); 15 values for Group A, two values for Group B, three values for Group C (Blackburn and McLeod, 1983). c [CI] Derived from mean C1 deposition amounts for all sites in each group (amount-weighted precipitation value). d [X] Derived from mean of individual annual [X] values, unweighted for annual precipitation amount. G r o u p A sites include Charley±lie (1), Bourke (2), Wee Waa (3), Inverell (4), Cobar SCS (6), G u n n e d a h (7), Trangie (8), Wellington (9), Condobolin (10), Griffith (12), Temora (13), Cowra (14), Deniliquin - - 3 years (15), Wagga SCS (16), Wagga ARI (17), Albury (18) and K y a b r a m (22). G r o u p B sites include Hay - - 3 years (11), Merbein - - 3 years (19), Walpeup - - 3 years (20), Charlton (21), Loxton - - 3 years (23), Wanbi (24) and Fowlers Gap - - 2 years (28). Group C sites include Verdun - - 3 years (25), Kybybolite (26) and Mt. Gambier (27).

marine C1 to interior drainage basins of SE Australia can be made with some confidence from a limited number of deposition collection sites. Uncertainties introduced by possible evaporite minerals in wind-borne dust can be explicitly considered by excluding [C1] values for sites which have relatively high Ca/Na ratios. The basic assumption of this approach is that the C1 component in rain associated with resuspension of dust does not represent recent introduction to the continental interior of C1 from marine aerosols, but instead represents redistribution of salts previously delivered to the land surface.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

332

1000

A)

• •

o"

o

&•

=_..

A sites B sites C sites

0

,,- 100-

.O:o

O0

0

00

o

0 |



0

10 0

0

0

i

r

i

i

i

i

100

200

300

400

500

600

700

distance to ocean (km) 300

B) .-. 250 •-i 200

¢

150

0

100

o



A sites B sites C sites

ml

5O 0 0

14 12

i

i

200

400

i

,

1000

600 800 precipitation ( m m / y r )

C)

o

o0

1o "" ol

o

o

I

8 qll~ o

6 4

o

i

0

o .011x R^2 = 0.72 t

200

t

400 600 Annual P r e c i p i t a t i o n



B sites

i

800 (ram)

1000

Fig. 4. (A) SE Australia mean annual rain [CI] vs. distance to nearest sea-coast (Blackburn and McLeod, 1983); see Fig. 3 caption for site category designations (Groups A-C). (B) M - D Basin rain [C1] vs. annual precipitation amount. (C) M - D Basin annual deposition of CI vs. annual amount of rain; line based on linear least-squares fit to y-axis deviations for Group A sites.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

333

2.3. Approximate composition of soluble components of resuspended dust Comparison of mean compositions from Group A and Group B sites provides some indication of perturbations of rain chemistry in the M - D Basin as the result of resuspended dust. Using the difference (Group B minus Group A) in unweighted mean values (Table 2), the dust increments of soluble ions in equivalents units relative to [Ca] are 0.20 for [C1], 0.18 for [Na], and 0.12 for [SO4]. Thus for one equivalent of halite added, approximately five equivalents of calcium carbonate and 0.6 equivalents of gypsum would also be contributed. For rain Ca, carbonate minerals in resuspended dust appear to contribute about an order of magnitude more than sulphate minerals. If these ratios of soluble dust components are representative for the entire basin, [Ca] should be a very sensitive indicator of the dust component of [C1] in M - D Basin rain. This approach indicates only the net increase in dissolved ions in composite precipitation samples returned from collection sites, and may not be representative of the bulk composition of atmospheric dust samples, owing to very large difference in solubilities of halite and calcium carbonate. It is possible that dust deposited in the collectors could have appreciably more carbonate mineral material which did not completely dissolve in the composite precipitation samples.

2.4. Annual delivery of 'new' marine chloride to the M - D Drainage Basin via the atmosphere The relationship between annual CI deposition and annual precipitation amount at 17 locations (19 mean annual values) having only a minor component of dust-derived C1 (Fig. 4(C)) can now be used to estimate total quantities of marine C1 delivered annually to the M - D Basin (Table 3). Using long-term mean annual precipitation amounts for geographical segments of the basin, we obtained estimates of marine C1 deposition per unit area (kg ha-1 year l). These C1 deposition rates were then multiplied by sub-basin surface areas. Using this approach, current mean annual marine CI deposition is 20 x 109 mol, with 12 x 109 mol reaching the Darling River catchment and 8 x 109 mol delivered to the River Murray Basin exclusive of the Darling (Table 3). Blackburn and McLeod (1983) estimated total annual delivery of C1 to the entire M - D Basin in 1974-1975 of 8 x 109 to 20 x 109mol. The lower limit was based on averaging minimum mean [C1] observed at 17 locations during seasons of highest precipitation amount. The upper limit was derived from annual means of [C1] for 1974-1975 at stations with annual precipitation

334

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

Table 3 Mean annual atmospheric deposition of marine chloride to the M-D Drainage Basin, using dependence of C1 deposition amount vs. annual amount of rain (Fig. 4(C)) for Group A mean annual values (Blackburn and McLeod, 1983) Parameter

Darling a

Murray b

Murray plus Darling c

Area (106 km2) Rain (mm year-l) d CI deposition (kg ha- J) C1 deposition (106tons) CI deposition (109mol) [C1] in rain (mgl -I) [Cl] in rain (#equivl 1)

0.632 469

0.428 491

1.06 478

6.55

6.80

6.65

0.41 11.5 1.40 39

0.29 8.2 1.38 39

0.70 19.7 1.39 39

a The relationship between marine C1 deposition and annual P is assumed to be the same for the Darling and Murray basins. b Total catchment of the River Murray, excluding the Darling River. c Total catchment of the River Murray, including the Darling River. d Precipitation amounts based on aggregation of long-term mean annual values reported for meteorological districts which approximately correspond to surface runoff catchment boundaries (Simpson and Herczeg, 1991a). amounts of 4 0 0 - 5 0 0 m m , which corresponds approximately to the mean annual precipitation for the entire basin. This latter choice of sites by Blackburn and M c L e o d (1983) effectively excluded data from nearly all collection sites we have assigned to G r o u p B (i.e. those sites where chemical compositions indicate significant dust contamination). In an earlier paper, we used the mean o f the lower and upper limits of atmospheric deposition of C1 reported by Blackburn and M c L e o d (1983) to obtain an estimate of the annual amount-weighted mean of [C1] in precipitation for comparison with salinities in the River M u r r a y (Simpson and Herczeg, 1991a). As outlined here, our current assessment is that the upper limit of C1 deposition reported by Blackburn and M c L e o d (1983) is the most appropriate value to use. 2.5. Comparison o f rain [CI] in the M - D Basin with other regions in Australia Several coastal sites (classified here as less than 100kin from the ocean) located along the SW, south and SE coasts o f the continent have mean annual rain [C1] of 100-300 #equiv1-1 (Blackburn and McLeod, 1983; Ayers and Manton, 1991; Farrington et al., 1993). Interior sites (200-700kin from the ocean) in the M - D Basin judged free of appreciable dust contamination had mean annual rain [C1] of a b o u t 40 #equiv 1-1 (Fig. 5), a b o u t 20% of coastal rain [C1]. Mean annual rain [C1] at four interior sites in Western Australia

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

1000 l

. laC tsao. . . . . . . . . . . .

t-, J •

W. Aust

M/D A sites /

L'" / / . %

E

=®:"

.

oo¢o

.,,

teo°

,-~ m

,,



"

"

N ~ . Alice

0

"

X

"

10

335

M/D B sites

Wagga ÷

+ +

+

+ +

Katherine 1

0.0

i

i

i

i

I

0.5

1.0

1.5

2.0

2.5

=

I

3.0

i

I

3.5

=

I

4.0

4.5

Ca/Na (eq/eq) Fig. 5. Comparison of mean annual rain [C1] from several regions of Australia, including the M - D Basin.

located along a transect from 150 to 600km inland averaged about 70#equivl -I (Farrington et al., 1993), appreciably greater than for M - D Basin interior sites. In contrast, rain [C1] at Katherine (wet deposition only) located about 250 km inland from the north-central coastline averaged only about 8#equiv1-1 (Likens et al., 1987) (Fig. 5). Data for two other interior sites, Alice Springs (Hutton, 1983) and Wagga Wagga (Ayers and Manton, 1991) had mean annual rain [CI] of 25#equiv1-1 and 18 #equiv 1-1, respectively. There were substantial differences between sample collection and analytical methods for the precipitation chemistry studies cited above. However, rain [C1] values for interior locations in SW Australia and the M - D Basin do appear to be significantly greater than for northern Australia (Katherine), and probably also for central Australia (Alice Springs). The higher rain [C1] for these more southerly regions of Australia appears not to result primarily from resuspended dust, the magnitude of which can be explicitly considered, and thus appears to represent significantly higher increments of 'new' C1 from marine aerosols per unit volume of rain than for northern Australia and for the interior of North America.

H.J. Simpson, ,4.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

336

3. Chloride transported by major rivers in SE Australia 3.1. Trends of [Cl] for the River Murray in response to river discharge rate changes Median annual [CI] along the axis of the River Murray (Fig. 6) increases by more than two orders of magnitude during drought years (e.g. 1982-1983) and by about a factor of 50 during years with discharge (Q) near the mean for the years 1971-1989 (e.g. 1988-1989). Influx of C1 from agricultural lands with a high water table in the middle reaches of the Murray appears to be more important during years of average runoff than during drought years (Fig. 6), as evidenced by the rapid increase of [C1] near kilometer point (KP) 1400 for 1988-1989. We can use chemical analyses compiled for water quality monitoring by the Murray-Darling Basin Commission (Mackay et al., 1988; Herczeg et al., 1993) to provide information on integrated amounts of C1 removed by the Murray and the Darling over periods of months to years. Monthly Q for the Murray at the downstream end of the main channel (KP 71) has been estimated from gauged flows at KP 696, adjusted for diversions and evaporation (Simpson and Herczeg, 1991a). Over the period of 18 years considered here, monthly Q ranged over more than two orders of magnitude (Fig. 7(A)), from 0.03 km 3 to 7 km 3. Concentrations of C1 were inversely related to Q (Fig. 7(B)), with minimum monthly mean values of less than 2mequiv1-1 and maximum monthly means of about 8 mequiv 1-1 . The total range of monthly mean [C1] was less than one order of magnitude, whereas that for monthly Q was more than two orders of magnitude (Fig. 8). This relationship probably l

i

.

.

.

.

i

.

.

.

.

i

.

.

.

.

i

.

.

.

.

i

.

.

.

.

8-] ~k 19a2.s3 (low o)

i

( 34 ~ 2t 0

0

~ ~ ~ L . . - ~

FLOW DIRECTION 198849 ~" (>medIan Q)

500 1000 1 5 0 0 2 0 0 0 distance upstream (kin)

2500

Fig. 6. Median annual River Murray [C1] vs. distance upstream. 1982-1983 - - Extreme low Q, 0.8km3year -I, at kilometre point (KP) 71; 1988-1989 - - approximately mean annual Q, 7.6km 3 year -] , at KP 71.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350 lO.O'fi

|.0

o

0.1

C~

0.0

337

j i

i

eq

e~

.

.

.

.

Year -1

.

.

.

.

,

Year Fig. 7. (A) Monthly water discharge of River Murray to the head of Lake Alexandrina (KP 71) over a period of 18 years. (B) Monthly mean [CI] near downstream end of the channel of the River Murray (KP 115) over a period of 18 years (note that the [CI] axis is inverted to simplify comparison with variations in a).

results primarily from enhanced saline groundwater discharge to the Murray during months of high surface runoff, leading to much larger total riverine C1 fluxes in those periods (Simpson and Herczeg, 1991a, b; Herczeg et al., 1993).

3.2. Estimates of annual Cl riverine transport by the Murray and Darling Annual quantities of C1 transported out of the M - D Basin can be estimated from river discharge and [C1] data. We have compiled estimates of median annual values of Q and river [C1] for an 18 year period (1971-1989), as representative of the riverine transport of C1 during recent decades (Tables 4 and 5). Natural Q estimates for the furthest downstream monitoring locations (KP 696 on the River Murray and KP 825 for the Darling at its point of influx to the Murray) are included to indicate some of the integrated effects of surface water management in the basin. Natural Q values have been calculated by the M - D Basin Commission from observed Q, altered by

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

338 100

.

.

.

.

.

.

.

.

,

.

.

.

.

.

.

.

.

r

.

.

.

.

.

.

.

.

i

0

E 10

...=....,-

E

....

o

..~.:~ . . . . ; . .



• .'.,~.'..¢. ..'.;'. y-



ell •

4#





,

, . dl, # I"



Q.

• *'

i, jlb

@



*



1

.1

i

.01

i

i

i

. . . .

i

.1

.

.

.

.

.

.

.

.

t

1

i

i

. . . .

10

Q @ KP 71 (krn3/month) Fig. 8. Monthly mean [C1] vs. monthly Q for River Murray near downstream end of main channel (1971 1989).

amounts of diversions for irrigation, evaporation losses in storage reservoirs, and other perturbations. Total Q for both river systems was reduced by about half from natural Q over this period of 18 years. Chloride concentrations are listed for KP 87, the location furthest downstream on the main channel of the Murray for which such data have been reported. No attempt has been made here to adjust for the difference in median annual [C1] between KP 87 and KP 71, which should be less than a few per cent of the values in Table 4. Estimates of median annual [C1] for the Murray at KP 87 for 1971-1978 were made from a linear regression between median annual values of [C1] and total dissolved solids (TDS) at the same location for 1978-1989: Cl(mg1-1) = (0.427 x TDS) - 31, where TDS --- 0.6 x EC (electrical conductivity in millisiemens at 25°C). Riverine C1 transport estimates have also been compiled for the Darling River (Table 5), which drains an extensive network of tributary catchments, representing more than half of the total area of the M - D Basin. The values listed for median annual [C1] are those for the gauging location at Burtundy, which accounts for a b o u t 90% of total Q (1971-1989) from the two anabranches of the Darling (Burtundy plus Bulpunga). Estimates of median annual [C1] at the downstream end of the Darling for 1971-1978 were also

339

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350 Table 4 Flux of chloride by the River Murray into the head of Lake Alexandrina (KP 71)

Year

74-75 73-74 75-76 83-84 81 82 78-79 88-89 84 85 76 77 71-72 86-87 87 88 79 80 77 78 72 73 85-86 80 81 82 83

Mean Median Total %ofM+D

KP 696a (km3year -1)

Actual Q at KP 71 b (km3year -1)

[Cl] at KP 87 (mequiv1-1)

C1 Flux at KP 71 (M + D) c (mol× 109)

C1 Flux at KP 71 (M) d (mol × 109)

33.1 29.4 30.9 26.6 23.4 17.6 18.4 14.2 11.0 14.0 16.1 10.0 9.4 10.4 8.0 9.4 7.8 2.5 16.2 14.1 292 100

27.52 22.08 20.15 9.63 8.72 8.48 7.62 6.42 5.81 5.70 5.67 3.20 3.21 2.86 2.27 1.50 1.37 0.79 7.9 5.8 143 100

1.63 2.17 2.25 2.00 4.51 2.99 3.01 3.58 3.83 2.73 3.24 5.35 4.34 3.61 6.00 5.52 6.42 8.11 3.96 3.60

45.0 47.9 45.4 19.3 39.3 25.3 23.0 23.0 22.3 15.6 18.4 17.1 13.9 10.3 13.6 8.3 8.8 6.3 22.3 18.8 403 100

43.9 45.4 43.4 16.7 38.8 22.6 21.8 21.3 20.6 14.0 18.2 16.3 13.3 8.6 12.8 8.1 8.0 5.5 21.1 17.5 380 94

Natural Q at

a Estimated natural Q assuming no human perturbation in the basin. b Gauged actual Q at KP 696, reduced by diversion and evaporation between gauge and KP 71. c Annual CI flux by combined M - D system at KP 71. d Annual C1 flux by the River Murray minus C1 input by the Darling River (see Table 5).

made, on the basis of a linear regression between median annual [C1] at Burtundy and TDS for 1978-1989. Median annual values for the River Murray (Table 4) are listed in decreasing rank order of Q measured at the furthest downstream wellcalibrated flow gauge at KP 696. Over this period of 18 years, the Darling, whose main branch joins the Murray at KP 825, contributed 24% of Q at KP 71. Although the Darling usually provides a minor fraction of annual Q (less than 20%), in some years it can contribute more than half of the total (e.g. 1976-1977, 1977-1978, 1982-1983 and 1983-1984). Median annual [C1] of the Murray at KP 87 averaged about a factor of three greater than that of the Darling for all years discussed

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

340

Table 5 Flux of chloride by the Darling River (main channel plus anabranch) to the River Murray at KP 825 Year

Natural Q at KP 825a (km3 year -1)

Actual Q at KP 825, b Darling total (km 3 year -1)

[C1] at Burtundy (mequiv 1-1)

Cl flux of Darling to Murray at KP 825 (mol × 109)

74-75 73 74 75 76 83-84 81 82 78 79 88-89 84-85 76 77 71-72 86-87 87-88 79-80 77 78 72 73 85 86 80 81 82-83 Mean Median Total % of M + D

3.1 8.6 9.3 11.2 2.9 4.8 4.4 3.5 3.8 1.9 1.0 1.7 0.6 3.7 1.7 1.1 0.23 0.17 3.54 3.0 64 22%

2.39 5.17 4.85 5.63 0.32 2.94 1.82 2.13 4.21 1.19 0.07 0.33 0.41 1.65 0.47 0.13 0.33 0.69 1.93 1.42 35 24%

0.45 0.48 0.42 0.45 1.63 0.93 0.65 0.76 0.39 1.30 2.45 2.39 1.55 1.04 1,77 1.41 2.54 1.21 1.21 1.13 -

1.1 2.5 2.0 2.5 0.5 2.7 1.2 1.6 1.7 1.5 0.2 0.8 0.6 1.7 0.8 0.2 0.8 0.8 1.3 1.2 23 6%

a Estimated natural Q assuming no human perturbation in the basin. b Total influx of water from the Darling River to the River Murray obtained by adding main channel Q (Burtundy) and anabranch Q (Bulpunga).

here (Table 5, Fig. 9(A)). Thus total transport of C1 by the River Murray over the period 1971-1989 (Fig. 9(B)) included only a minor contribution (about 6%) from the Darling River, even though the Darling drains more than half of the total M - D Basin area. Comparison of median annual quantities of C1 transported by these two rivers (Table 6) over a period of 18 years (22 × 109molyear -1) with annual delivery from the atmosphere to the entire catchment of the combined system (20 x 109molyear -1) indicates that approximately the same total amount of C1 was transported out of the basin by the River Murray at KP 71 as was delivered by precipitation. However, the Darling

341

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

l°liilll~llili~--+-~-18

g

A

~

Murray

10

E~

/

t~

3 rr

/

0 I

I

I

I I

50

I I

I I

I I

I

t

t

I

I

t I

I L

I I

I

I

I

~

I

t L

t ~ -

~

-

50

B

Murray- Darling

40

I

~

40

E

o

~

~ 30

x

30 o= =o 03 TM

lO T

D n

o

1o

, , , , .

I~

04 I~

03 I~.

,

'q" i~

u[3 I~

¢D I'~

I~ I~

CO 1~

O~ I~

O 0O

,

*GO

¢~t CO

o

¢O GO

~l' ~ GO 0O

¢,O I ~ GO 0O

CO 0D

Fig. 9. (A) Median annual [CI] at downstream ends of the Murray and Darling rivers, 1971 1989. (B) Annual downstream transport of CI to the mouths of the Murray and Darling rivers, 1971-1989. R i v e r e x p o r t e d an a m o u n t e q u a l to o n l y a b o u t 11% o f a t m o s p h e r i c delivery o f C1 to the D a r l i n g Basin, whereas the M u r r a y c a t c h m e n t (exclusive o f the D a r l i n g ) e x p o r t e d two to three times the a n n u a l a m o u n t delivered f r o m the a t m o s p h e r e . It is likely t h a t p e r t u r b a t i o n s o f the basin's h y d r o l o g i c cycle d u r i n g this c e n t u r y have c o n t r i b u t e d to large-scale r e a r r a n g e m e n t s o f C1 stored within the c o m b i n e d M - D c a t c h m e n t .

342

H.J. Simpson, A.L. Herczeg/Journal of Hydrology 154 (1994) 323-350

Table 6 Atmospheric delivery of marine chloride to and annual chloride removed by riverine discharge from catchments of the Murray and Darling rivers Parameter

Atmospheric deposition per year (tons x 106) : pa Atmospheric deposition per year (mol × 109) = P ' Riverine discharge per year (mol x 109) = R b Riverine discharge per year relative to atmospheric input = RIP'

Murray (M) 0.29

Darling (D) 0.41

Murray plus Darling 0.70

I 1.5

20

21

1.3

22

260%

11%

110%

8.2

a p: See Table 3; based on total precipitation plus dry fallout sample collection during 1974 1977, with most data from 1974. b R: See Tables 4 and 5; based on mean annual [C1] at the downstream ends of the Murray and Darling rivers from 1971 to 1989.

4. Chloride inventory in the subsurface of the M - D Basin

4.1. General description of aquifers and stable isotope compositions of groundwaters The Murray Groundwater Basin includes a series of Tertiary to Holocene sedimentary sequences that contain four aquifer systems with a maximum thickness of 600 m. In most of this basin the sediment sequence has a total thickness of less than 200m (Brown, 1989), representing a thin veneer of sediment over a large area (3 x 105 km2). The fluvial Renmark Group and overlying marine Murray Group were deposited between 60 and 10 million years ago. The last marine intrusion was more than 2 million years ago and resulted in estuarine conditions in the western third of the basin. Present groundwater hydraulic gradients are low (about 1 : 1000), and groundwater residence times in the various aquifers have been estimated by several methods to be of the order of 104-106 years (Leaney and Allison, 1986). Although groundwaters in many parts of the southern and western M - D catchment resemble seawater in relative major ion chemistry, their stable isotope compositions are very different from that of seawater. Deuterium and 180 compositions of groundwaters with a large range of salinities in the Campaspe River Basin, in the south-central portion of the Murray Basin (Arad and Evans, 1987) and for the very saline waters in the western portion of the Murray Basin (Leaney and Allison, 1986) all are much more depleted in

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

343

heavy isotopes (2H and 180) than seawater (Fig. 10), demonstrating that despite high salinities, they have been recharged by meteoric waters and are not fossil marine waters. Stable isotope compositions for unconfined aquifers in the western Murray Basin deviate to the right of the World Meteoric Water Line (WMWL), owing to evaporative losses (Simpson and Herczeg, 1991b) during the recharge process (Fig. 10).

4.2. Outline of method for estimation of subsurface storage of Cl We cannot estimate total subsurface storage of C1 in the M - D Basin with much confidence. To do so would require detailed examination of results previously reported for a large number of aquifer systems, plus new observations for those non-potable groundwaters which have received little or no study. However, we can compile a few approximations of volumes and representative [C1] values of component systems to indicate order of magnitude values of C1 storage in the saturated zone potentially receiving recharge over the past few million years. Estimation of total C1 in the unsaturated zone and surface evaporite deposits located in ancient and active saline groundwater discharge zones is even more problematic, and is included here only to indicate the potential importance of these categories in the total C1 inventory. We will consider the mean thickness and [C1] of several groundwater systems, on the basis of a simplified treatment of aquifer subdivisions (Evans and Kellett, 1989). Only a small portion of the Great Artesian Basin A

10

•O0

_~

-10 ~,~ -20.

SMOW

C a m p a s ~

E-30 m

-4o ao

Murray Group GW

~~~'/

-50 -7

i

t

I

i

i

-6

-5

-4

-3

-2

-1

0

Oxygen-18 (~',.,,SMOW) Fig. 10. Stable isotope compositions (6D, 6180) of groundwaters in two regions of the Murray Basin: data for western Murray Basin from Leaney and Allison (1986); data for Campaspe River catchment from Arad and Evans (1987). SMOW indicates Standard Mean Ocean Water, which serves as the reference value for D and 180 data and corresponds closely to the likely isotopic composition of water from any previous marine intrusion in the Murray Basin. WMWL indicates the World Meteoric Water Line (6D = (8 × <5180) + 10%).

344

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

( G A B ) in eastern Australia, which represents the largest p o t a b l e w a t e r a q u i f e r o n the c o n t i n e n t , underlies the surface r u n o f f n e t w o r k o f the M - D c a t c h m e n t . T h e [C1] levels in the Jurassic f o r m a t i o n s (J aquifer) o f the G A B are m u c h l o w e r t h a n for g r o u n d w a t e r s in the s o u t h e r n h a l f o f the M - D Basin ( T a b l e 7) a n d thus r e p r e s e n t m u c h less storage o f C1 t h a n in the M u r r a y Basin despite a greater a q u i f e r thickness. T h e C1 s t o r e d in the C r e t a c e o u s f o r m a t i o n s ( K aquifer) o f the G A B is c o n s i d e r a b l y greater t h a n for the Jurassic f o r m a t i o n s , b u t p r o b a b l y is m u c h less t h a n t h a t in the M u r r a y Basin. T h r e e o f the general categories o f g r o u n d w a t e r sub-basins e m p l o y e d here have not been studied nearly as extensively as the G A B or M u r r a y Basin ( M B ) - - the D a r l i n g R i v e r Basin ( D R B ) a n d the F r a c t u r e d R o c k basins in the c a t c h m e n t s o f the R i v e r M u r r a y ( F R m ) a n d the D a r l i n g R i v e r ( F R d ) . T h e u n c o n f i n e d Table 7 Storage of chloride at and below land surface of the M-D Drainage Basin: order of magnitude estimates Parameter

MB a

FRm b

Area (106 km 2)

0.30

Saturation thickness (km) Porosity fraction Volume of groundwater (103 km 3) Mean [C1] (mequiv 1-1) Total C1 (1012mol)

0.1

0.10 0.05

0.15 5

0.15 0.8

200 1000

10 8

GAB - - jc 0.30 0.4 0.2 25 3 75

GAB - - K d

0.30 0.05

0.15 1.5

0.15 2

100 150

20 40

1.5 × 1015mol 0.5 × 1015mol (???) 2.0 × 1015mol (70% of E C1)

Darling R. Basin (30% of E C1)

1.4 8

0.6 12

1.8

0.5

Murray Groundwater Basin (MB). b Fractured Rock aquifers in the River Murray catchment (FRm). c Great Artesian Basin (GAB) - - Jurassic aquifer. a Great Artesian Basin (GAB) - - Cretaceous aquifer. e Darling River Groundwater Basin (DRB). f Fractured Rock aquifers in the Darling River catchment (FRd). a

0.20 0.05

0.2 25

R. Murray Basin

C1 stored = TC1 (1015 mol) Annual delivery of marine CI by P = ACI (109 mol) Minimum time to supply C1 -- TC1/AC1 (105 years)

FRd r

0.30 0.4

10 250

E CI in saturated zone Surface evaporation deposits + unsaturated zone Total

DRB e

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

345

shallow groundwaters of the DRB are saline and overlie the much fresher confined GAB aquifers. The two Fractured Rock aquifer categories used here include an extremely large range of salinities and recharge patterns, and are included only as a grossly simplified representation of the subsurface CI inventory outside the major sedimentary basin aquifers (Table 7). We have not given any estimates of C1 amounts associated with sedimentary formations classified as aquitards, assuming that they are, to a first approximation, not accessible for groundwaters recharged over the past few million years. Storage of C1 in the unsaturated zone and surface evaporite deposits fed in the past by saline groundwater outflows are included here as a total quantity equal to one-third of the total C1 storage in the saturated zone. We have very few constraints on this estimate, and have only a general impression that total C1 amounts in these categories are considerably lower than for the groundwater inventory in the western half of the Murray Basin and that these components of the C1 inventory in the Murray catchment exceed those of the Darling. The total subsurface plus surface evaporite deposit C1 inventory obtained by these approximations is 2 × 1015mol, about 70% of which is estimated to lie in the catchment of the River Murray and 30% in that of the Darling River. For comparison, Evans and Kellett (1989) independently estimated storage of C1 in the Murray Basin to be more than 1015mol (including a substantial portion of salts associated with aquitard formations). A first-order approximation of a minimum time required to accumulate C1 from marine sources via the atmosphere can then be obtained by invoking our previous estimates of annual deposition by rain plus dry fallout. For the Darling River catchment this minimum time is 5 × 104 years and for the River Murray catchment it is 1.8 × 105 years (Table 7). These minimum times assume (1) no C1 removal via riverine transport during the period of accumulation and (2) a constant delivery rate of marine C1 via the atmosphere equal to that estimated here for the present. Discharge of saline groundwaters into the River Murray has probably been an important component of the C1 budget in the Murray Basin for much of the past several million years, resulting in substantial export of C1 from the basin. The first of these assumptions is clearly not reasonable for the entire period of build-up of C1 and would lead to underestimation of actual C1 accumulation times. As regards the second assumption, during the last glacial period, considerable evidence indicates lower precipitation rates than at present for the interior of Australia for extended periods (Bowler, 1988; Wasson and Donnelly, 1991), which probably meant lower delivery rates of marine C1 via the atmosphere. Thus it seems likely that accumulation of marine C1 delivered via the

346

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

atmosphere to the subsurface of the M - D Basin has required at least an order of magnitude longer than the minimum estimates listed in Table 7. 5. Atmospheric delivery of marine chloride and

36C1

5.1. Issues related to use o f 36CI as a tracer for groundwater recharge ages

During the past decade 36C1 measurements have been made in old groundwaters in a number of areas around the world. The long radioactive half-life (0.3 million years) and lack of likely mechanisms for removal of C1 from solution have led to extensive measurement of 36C1 as a groundwater tracer. Two issues which have contributed to difficulties of interpretation of groundwater 36C1 data are (1) uncertainties in the long-term mean 36C1 to C1 ratio in precipitation-delivered 36C1 as a function of geographic location and (2) uncertainties in rates of in situ formation of 36C1 within aquifers (Andrews and Fontes, 1992). 5.2. General observations on 36Cl budgets in the M - D Basin

The 36C1/C1 ratios in the Murray-Mallee groundwaters have been observed to increase along the inferred direction of flow based on contours of water level elevation (Kellett et al., 1992), opposite to the trend expected on the basis of probable relative ages since recharge. This increase has been attributed to continued influx of 'recent' (with respect to the half-life of 36C1) saline recharge water along the flow path (Davie et al., 1989). These 36C1 observations for the western Murray Basin reinforce the critical need to establish appropriate 36C1/C1 ratios for atmospheric input as a function of geographic location. One implication of our interpretation of rain [C1] is that recent model calculations of 36C1/C1 ratios as a function of distance inland in SE Australia (Davie et al., 1989) may need substantial modification. If the marine C1 component accounts for only about half of total C1 in rain in the western Murray Basin, with the remainder resulting from resuspension of dust (Fig. 4(C)), the 36C1/C1 ratio of precipitation would be about twice those previously calculated based on the assumption that atmospheric C1 deposition (Blackburn and McLeod, 1983) was entirely derived from recent marine aerosols. This difference would result if C1 in the dust component had a 36C1/C1 ratio comparable with recent precipitation rather than extremely low ratios typical of the surface ocean. In addition, the gradient in 36C1/C1 ratios in rain as a function of distance

347

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

inland would be different from that derived from recent model calculations. This suggestion is based on the approximately constant values of marine C1 in rain observed as a function of distance from the sea-coast for sampling locations categorized here as relatively free of dust contamination (Fig. 4(A), G r o u p A sites). To illustrate changes in geographic trend o f 36C1/C1 ratios which would be expected from the conclusions reached here, we can compare 36C1/C1 ratios based on data from Fig. 4(B) used to derive annual deposition of marine C1 (Table 8) with those obtained from published contours o f 36C1/C1 ratios (Davie et al., 1989). The 36C1/C1 ratios in rain calculated with our revised approach are considerably higher in the western M u r r a y Basin, but lower than those previously calculated for the SE M u r r a y Basin.

6. Conclusions F r o m data and discussions outlined here, the following conclusions are offered: (1) a considerable fraction (approximately half) of C1 in composite rain samples from the western M u r r a y Basin is probably from resuspended continental dust rather than marine aerosols formed within the previous few weeks. (2) Excluding precipitation network [C1] data which have high C a / N a ratios indicative of dust contamination, the marine C1 c o m p o n e n t of Table 8 Model calculations of expected 36C1/C1 ratios in precipitation in the southern M D Basin (35°S latitude) Annual P (mm)

Marine C1 deposition (kg ha -1 year-l)a

36C1 fallout

36C1/C1 (×10-15)c

36C1/C1

(atomsm-2 s-l) b

300 400 500 600

4.7 5.8 6.9 8.0

18 18 18 18

71 58 49 42

20 40 60 70

(xl0-15) d

a Derived from Fig. 4(C). b Based on a revised estimate of 36C1 atmospheric fallout vs. latitude from spallation of 4°Ar only (Andrews and Fontes, 1992). c Calculated from the data in the second and third columns: 36C1 fallout/C1 deposition ((atomsm-2 s-1)/(kgha -1 year-l)) x 1.86 x 10 -14 = 36C1/C1x 10-15. d Derived from a west-to-east transect at 35°S, beginning at the River Murray near KP 100, using the contours of Fig. 1 of Davie et al. (1989).

348

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323 350

continental interior rainfall can be estimated as a function of annual precipitation amount. (3) The approximate ratios (equivalents units) of calcite, halite and gypsum in resuspended dust that perturb the chemistry of bulk precipitation samples in the western Murray Basin are 5 : 1 : 0.6. (4) Current mean annual marine C1 deposition from the atmosphere in the M - D Basin is 20 x 109tool, which is equal to the amount previously suggested as the maximum annual deposition of C1 from the atmosphere (Blackburn and McLeod, 1983). (5) Riverine [C1] and water discharge data for the period 1971-1989 indicate that the Darling River Basin is storing an amount of C1 equivalent to about 90% of the annual atmospheric input of marine C1, whereas the River Murray is exporting two to three times the annual amount it receives from the atmosphere. These C1 flux estimates indicate that large-scale rearrangements of C1 are currently occurring in the M D Basin. (6) Order of magnitude estimates of quantities of C1 stored in the subsurface of the M - D Basin are consistent with minimum C1 accumulation times of 5 x 104 years for the Darling catchment and 2 x 105 years for the Murray. If export of C1 were taken into account via the Murray and Darling rivers during accumulation of salts within the M - D Basin, these time estimates probably should be increased by at least an order of magnitude. (7) Input ratios of 36C1/C1 from the atmosphere to the surface predicted from the analysis presented here deviate substantially from those based on recently reported model calculations for the western Murray Basin. (8) Given the very long history of salt accumulation within the M - D Drainage Basin from marine aerosols (of the order of at least a few million years), degraded surface water quality and dry land salinization as a result of perturbed saline groundwaters in the catchments of the Murray and Darling rivers represent a very difficult and essentially permanent challenge to the human uses of the basin which have evolved over the past century.

Acknowledgements We thank P. Cook and J. Kellett for review of an earlier manuscript. Work begun under grants from Australian-American Educational Foundation (Fulbright Program) and US National Science Foundation (No. INT-8814385) provided initial stimulus for this research. A number of students at Columbia University also provided considerable help. This is Contribution 5079 from the Lamont-Doherty Earth Observatory of Columbia University.

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

349

References Allison, G.B. and Hughes, M.W., 1983. The use of natural tracers as indicators of soil-water movement in a temperate semi-arid region. J. Hydrol., 60:157 173. Allison, G.B., Cook, P.G., Barnett, S.R., Walker, G.R., Jolly, I.D. and Hughes, M.W., 1990. Land clearance and fiver salinisation in the western Murray Basin, Australia. J. Hydrol., 119: 1-20. Andrews, J.N. and Fontes, J.Ch., 1992. Importance of in-situ production of chlorine-36, argon-36 and carbon-14 in hydrology and hydrogeochemistry. In: Isotope Techniques in Water Resources Development 1991, Proceedings Series, IAEA-SM-319. International Atomic Energy Agency, Vienna, pp. 245-269. Arad, A. and Evans, R., 1987. The hydrogeology, hydrochemistry and environmental isotopes of the Campaspe River aquifer system, north central Victoria, Australia. J. Hydrol., 95: 63-86. Ayers, G.P. and Manton, M.J., 1991. Rainwater composition at two BAPMoN regional stations in SE Australia. Tellus, 43B: 379-389. Blackburn, G. and McLeod, S., 1983. Salinity of atmospheric precipitation in the Murray Darling drainage division, Australia. Aust. J. Soil Res., 21:411-434. Bowler, J.M., 1988. Environmental and salinity history of the Murray basin in the last 500,000 years. In: C.M. Brown and W.R. Evans (Editors), Abstr. Murray Basin 88 Conf., Aust. Bur. Miner. Resour., Div. Cont. Geol. Groundwater Ser. 12. Canberra, A.C.T., pp. 19 20. Brown, C.M., 1989. Structural and stratigraphic framework of groundwater occurrence and surface discharge in the Murray Basin, southeastern Australia. Aust. Bur. Miner. Resour. J. Geol. Geophys., 11: 127-146. Conway, E.J., 1943. Mean geochemical data in relation to oceanic evolution. Proc. R. Ir. Acad., Sect. A, 48:119 159. Davie, R.F., Kellett, J.R., Fifield, L.K., Evans, W.R., Calf, G.E., Bird, J.R., Topham, S. and Ophel, T.R., 1989. Chlorine 36 measurements in the Murray Basin: preliminary results from the Victorian and South Australian Mallee region. Aust. Bur. Miner. Resour. J. Aust. Geol. Geophys., 11: 261-272. Evans, W.R. and Kellett, J.R., 1989. The hydrogeology of the Murray Basin, southeastern Australia. Aust. Bur. Miner. Resour. J. Aust. Geol. Geophys., 11:147 166. Farrington, P., Salama, R.B., Bartle, G.A. and Watson, G.D., 1993. Accession of major ions in rainfall in the south-western region of Western Australia. CSIRO Div. Water Res. Rep. (Draft), Wembley, W.A. Herczeg, A.L., Leaney, F.W., Ghomari, R., Davie, R. and Fifield, L.K., 1989. Hydrochemistry of groundwaters in the south-western Murray Basin with special reference to the Woolpunda groundwater mound. Centre for Groundwater Studies, Rep. 13, CSIRO Div. Water Resources, Adelaide, S.A. Herczeg, A.L., Simpson, H.J. and Mazor, E., 1993. Transport of soluble salts in a large semiarid basin: River Murray, Australia. J. Hydrol., 144: 59-84. Hutton, J.T., 1983. Soluble ions in rainwater collected near Alice Springs, N.T., and their relation to locally derived atmospheric dust. Trans. R. Soc. S. Aust., 107: 138. Hutton, J.T. and Leslie, T.I., 1958. Accession of non-nitrogenous ions dissolved in rainwater to soils in Victoria. Aust. J. Agric. Res., 9: 492-507. Junge, C.E., 1963. Air Chemistry and Radioactivity. Academic Press, New York, 382 pp. Junge, C.E. and Gustafson, P.E., 1957. On the distribution of sea salt over the United States and its removal by precipitation. Tellus, 9:164 173.

350

H.J. Simpson, A.L. Herczeg / Journal of Hydrology 154 (1994) 323-350

Junge, C.E. and Werby, R.T., 1958. The concentration of chloride, sodium, potassium, calcium and sulfate in rain water over the United States. J. Meteorol., 15: 417-425. Kellett, J.R., Allan, G.L., Evans, W.R. and Fifield, L.K., 1992. Pre bomb-pulse chlorine-36 input to the regional unconfined aquifers of the Murray Basin. Third Murray Darling Basin Groundwater Workshop, Ext. Abstr., Renmark, S.A., 27 29 October, Aust. Bur. Miner. Resour., Canberra, A.C.T. Leaney, F.W. and Allison, G.B., 1986. Carbon-14 and stable isotope data for an area in the Murray Basin: its use in estimating recharge. J. Hydrol., 88: 129-145. Likens, G.E., Keene, W.C., Miller, J.M. and Galloway, J.N., 1987. Chemistry of precipitation from a remote, terrestrial site in Australia. J. Geophys. Res., 92:13299 13314. Mackay, N., Hillman, T. and Rolls, J., 1988. Water quality of the River Murray: review of monitoring 1978 1986. Murray Darling Basin Commission, Canberra, Water Qual. Rep. 1, 62 pp. Moller, D., 1990. The Na/C1 ratio in rainwater and the seasalt chloride cycle. Tellus, 42B: 254 262. Pack, D.H., 1980. Precipitation chemistry patterns: a two-network data set. Science, 208: 1143 1145. Simpson, H.J. and Herczeg, A.L., 1991a. Salinity and evaporation in the River Murray basin, Australia. J. Hydrol., 124:1 27. Simpson, H.J. and Herczeg, A.L., 1991b. Stable isotopes as an indicator of evaporation in the River Murray, Australia. Water Resour. Res., 27: 1925-1935. Stallard, R.F., 1980. Major element geochemistry of the Amazon River system. Ph.D. Thesis, Department of Oceanography, MIT WHOI, Cambridge, MA. Stallard, R.F. and Edmond, J.M., 1981. Geochemistry of the Amazon 1. Precipitation chemistry and the marine contribution to the dissolved load at the time of peak discharge. J. Geophys. Res., 86:9844 9858. Wagner, G.H. and Steele, K.F., 1989. Na/C1 ratios in rain across the USA, 1982-1986. Tellus, 41B: 444-451. Wasson, R.J. and Donnelly, T.H., 1991. Palaeoclimatic reconstructions for the last 30,000 years in Australia - - a contribution to predict future climate. CSIRO, Div. Water Res. Tech. Mem. 91/3, Canberrra, A.C.T.