Isotope Hydrogeochemistry of Urban-Zone Groundwater, Central Africa Mohamed Fahmy Hussein Cairo University, Egypt Abstract The isotope hydrogeochemistry of Bangui City, Central Africa, is discussed. This small city (~0.65 million people on ~20km 2 ) is situated on Oubangui tributary of the Congo River. The urban- zone is highly despoiled and most of its population depends on water supply from the shallow the porous aquifer and/or pumping the deep fractured aquifer. The purpose is to define groundwaterchemical and isotope composition in relation to recharge, rock formation and the impact of the awful practices of the indigenous population on groundwater quality. The hydrochemical data demonstrated reaction with biogenic CO 2 gas, weathering, cation- exchange and NO 3 - pollution of anthropogenic origin. The conjunctive use of the chemical and isotope compositions ( 18 O and 2 H) showed the alteration of silicate and carbonate rocks in the zones of dilute and relatively charged groundwater. The isotopes illustrated the role of evaporation and transpiration in the water-balance, with a fraction of the transpired-vapor recycled. The regional “inverse continental isotope-effect” is attributed to differences in air-temperature, amount and altitude of precipitation, rather than to a claimed movement of humid air masses from Central Africa westward to the Atlantic. Isotope data showed that recharge during the humid-season may not exceed that taking place during the dry-season, and that the overall aquifer-recharge is mediocre. The data helped to distinguish rapid circulation in deep fractured formations northward versus slow one in porous sections southward, and showed that the system is void of evaporites and its carbonates are made-up of Calcite, not Dolomite. Pollution, by nitrate, is noticed downtown, and would extend everywhere in that urban-zone considered as one of the worst worldwide. Figure 1A Hydrographic network and sampling location map ________________________________________________________________________________________________Isotope Hydrochemistry of Urban Zone Groundwater, Central Africa. (Mohamed Fahmy Hussein) 1
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
8/14/2019 1 Isotope Geochemistry Central Africa 2009 WORD97
IntroductionAfrica, south of the Sahara, has abundant surface and groundwater basins that are seldom
studied. Oubangui River is the northern tributary of Congo River. It flows in a SW direction,
making part of the border between Congo and Central African Rep., where Bangui, is constructed
one century ago. Despite abundant river discharge and plentiful precipitation (1500mm.yr -1), the
city depends on groundwater supply. This study gives a first account on the hydrochemical and
isotope compositions of local groundwater. The objective is to conjugate geochemical and isotopedata to understand groundwater composition and aquifer dynamics.
Hydrogeological SettingThe groundwater of Bangui city (Figures 1A and B) is the subject of this study. The local
water-balance, aquifer recharge and dynamics are not well-known (Doyemet, 2006). The city is
underlain by two formations. The upper is an unconfined porous aquifer (100mdeep Quaternary
Oxisols) and the lower (~200m thick) is a semi-confined one made-up of fractured carbonate and/or
silicate rocks (underlain by the Precambrian schist and quartzite basement) where we sampled some
drilled wells. The two aquifers are hydraulically connected via a clay aquitard. The fractures run in
a SE direction formed 550±100 Ma ago by orogenesis tectonics of the Pan-African Oubanguide
Mountains (Kennedy, 1964). The local precipitation is falling in two seasons; wet (April/May toOctober/November) and dry (November/December to March/April) and 85% of rainfall takes place
during the wet-season. The mean air-temperature exceeds 26°C and the relative humidity is
1. Sampling and AnalysisThirteen groundwater samples (#1 to 13, Table 1 and Figure 1B) were collected, April, 2007
from pumping-wells drilled in the semi-confined aquifer. Seven heavy-storm events (#14 to 20)
were also sampled mostly during the 2007dry-season, as well as one run-off aliquot, and one
Oubangui River-water (#21). The samples were shipped for chemical and stable-isotope analysis
(18O and 2H) at the Atomic Energy Authority, Cairo, Egypt, using the standard methods. Theisotopic composition was measured on mass-spectrometer after CO2 equilibrium with water or after
water reduction on hot zinc for δ 18O and δ 2H, respectively.
2. Hydrogeochemical dataThe hydrochemical and isotope data is plotted in several ways (Figures 2 to 12). Groundwater
temperature is high (24 to 28.6°C) and enhances water reaction with aquifer solid-phases. It seems
that a karstic porosity allows rapid circulation showing initial high temperature of ~28 °C with low
EC of 50 µ moh/cm, whereas in the deep porous sections the EC is variable, up to 600 µ moh/cm
(relatively high EC where recharge warm water, T>24 °C, is circulating.)
The pH is the master piece of most hydrochemical reactions. The local air-born dust is made-
up of fine reddish oxide- and silicate-particles [e.g. Gibbsite, Al2O3.3H2O, Goethite, α−FeOOH and
Kaolinite, Al2Si2O5(OH)4]. Reaction of rainwater with such dust may result in significantly
increasing its pH value before reaching the ground. The pH of rainwater might become as high as
8.26 if it becomes saturated in air-born Calcite-dust during rainfall (Harte, 1985, page 108.)
through biogenic CO2 diffusion in the unsaturated zone. Finally, the deeply-percolated water reacts
with rock formations and shows the observed acidic pH values (Table 1) with high pCO2 values
(200 to 400 times -3.50) reflecting its high weathering capacity.
The aquifer is recharged through precipitation, so the chemistry of rainwater and groundwater
is closely related. The sum of ions (Σ ++Σ -) of rainwater is in the range 0.87-2.51 meq.l-1, whereas
that of groundwater is in the range 0.35-13.36 meq.l-1. Three groundwater samples have less ionic
strength than the most dilute rainwater sample. This would suggest that a section of the aquifer receives highly dilute precipitation and/or it has karst-type porosity.
2. 3. Hydrogeochemical DataCa and HCO3 are dominant in rainwater, whereas Mg and SO4 have the lowest levels. Na and
Cl have intermediate concentrations. A similar, but accentuated, trend is observed for groundwater
that significantly shows higher Mg contents. Piper and Schoeler diagrams (Figure 2 and 3) show
that rainwater has a Ca-Na-HCO3 composition that finally ends-up, in the aquifer, in a Ca-
Mg.HCO3 composition. The Mg increase (accompanied by Ca diminution) may be due to the
presence of olivine, attenuation of feldspar and/or dissolution of magnesian calcite.
K concentration is relatively high in two dilute GW samples (#10 and 4, Figure 6), and then it
stabilizes at low values in the rest of the samples. Two rainwater samples have higher K than inGW. It seems that K becomes rapidly fixed, by cation exchange, in the unsaturated zone and the
aquifer. Na also shows an early linear-increase as a direct result of non-carbonate rock weathering.
There is no reason for the later diminution of Na other than cation exchange.
Low Cl contents are noticed in most GW samples (Figure 6) but it is relatively high in the
most-charged sample #6 due to rock alteration (and like sample #4, it has twice Cl content than in
Oubangui river). This late increase is not comfortable since not accompanied by equivalent increase
in Na. Weathering of non-carbonate rocks is assumed for the release of Cl (and SO4). Erickson’s Cl-
Frequency Distribution can reveal two scenarios. When a “lognormal distribution” is obtained, it
indicates Cl gain through rainfall and evaporation. A “normal distribution” indicates Cl gain
through dissolution/mixing mechanism. We expect the 2nd scenario for the studied groundwater.
It is obvious that SO4 in the dilute GW is slightly higher than in rainwater. Dilute GW mayhave sulfate of air-born origin. Stumm and Morgan (1980) stated that the 40% of sulfate in river
water comes from maritime aerosols, 30% from gypsum dissolution and 30% is by sulfide
oxidation. It seems that Bangui aquifer is free of gypsum and anhydrite (see saturation indices in
section3). Sulfate decreased in two sites (#12 and 13), Figure 6. In the relatively-charged GW, SO4
stabilizes at low content due to reduction conditions (the dissolved-oxygen data - not shown -
indicate that the aquifer has moderate reduction conditions, at least in two sites (#1 and 9):
H2SO4 + 2 CH2O → H2S + 2 H2O + 2 CO2
HCO3 starts at a considerable level in rainwater, Table 1, and demonstrates a significant linear
increase in GW (up to 6 times its content in rainwater), Figure 6, indicates weathering of non-
carbonate rocks and dissolution of carbonate-rocks. Ca accounts for more than half of HCO3. Therest of HCO3 is covered by other cations (mainly Mg). On the composition diagrams (Figure 6) a
linear trend (with a slope of 0.48 and 0.31 for HCO3 and Ca, respectively) is obtained. Such
linearity shows dissolution-mixing mode of solute acquisition (Mazor, 2004). Linearity applies also
to Mg versus total ions (with a lower slope of 0.17). The (Ca+Mg) versus total ions line runs
exactly on that of HCO3 versus total ions.
NO3 is, in general, under the worrying limit (Figure 5) but in three sites downtown (#10, 1
and 4) high NO3 concentrations (16, 21 and 40 mg l-1, respectively) of anthropogenic origin
significantly exceed the admissible limit (15 mg l-1). As natural vegetation is cleared out NO3 would
build-up (Mazor, 2004) through nitrification (Stumm and Morgan, 1981):
Nitrification takes place in two steps (first the Nitrosomonas oxides NH4+ into NO2
- and then
the Nitrobacter transforms NO2- into NO3
-, Harte, 1985). Also, denitrification will be damped when
forests are cleared:
4 NO3- + 5 CH2O + 4H+
→ 5 CO2 + 2 N2 + 7 H2O
Since the composition diagrams (Figure 6) shows no disruption in the ionic relationships this
would confirm hydraulic connectivity between the unconfined and the semi-confined aquifers.
D’Amore diagrams (Figure 4) show six parameters for each sample. Samples #11, 9 and 10reflect compositions close to that of rainwater with early slight enrichment in HCO 3 through
weathering of non-carbonate rocks. The trend is accentuated in samples #8, 7 and 5 due to late
carbonate dissolution. Samples #3 and 13 show weathering of non-carbonate rocks whereas
samples #2, 4 and 1 show passage on Mg-Calcite while samples #12 and 6 show reaction with
limestone.
20
40
60
80
#19start
#21 riverend
0
100
0 100
PIPER CATIONSBangui Rainwater
% Mg
% Na
+K
% Ca
80
60
40
20
#19
start
#21 riverend
0
100
0 100
PIPER ANIONSBangui Rainwater
% Cl
% CO3+
HCO3
% SO4
80
60
40
20#11start
#6end
0
100
0 100
PIPER ANIONSBangui GW
% Cl
% CO3+
HCO3
% SO4
20
40
60
80
#11
start
#6end
0
100
0 100
PIPER CATIONSBangui GW
% Mg
% Na
+K
% Ca
0.01
0.1
1
10
1 2 3 4 5 6
meq l-1
Schoeler Diagram Bangui GW
6
5
7
13
8
12
1
4
2
3
10
9
11
CaMg Na Cl SO4HCO3
0.01
0.1
1
10
1 2 3 4 5 6
meq l-1
Schoeler Diagram Bangui Rain Water
21
14
17
16
15
18
20
19
CaMg SO4HCO3Mg Na Cl SO4
Figure 2 (Left) Piper triangles; the upper set is for GW and the lower set is for rainwater
Figure 3 (Right) Schoeler diagram. Samples are arranged according the total concentration (lowest at legend bottom)
The hydrochemical data show that the aquifer sections with dilute GW keep water
compositions close to that of rainwater. However, solute content becomes accentuated elsewhere;
early by weathering of non-carbonate rocks (#3 and 13) and later-on further by carbonate
dissolution (#1, 2, 4, 6 and 12). Dilute GW samples (in the eastern and northern sections of the city)
may reflect karstic conductivity (rapid flow and high response to recharge) whereas the southern
and central sections of the city have relatively charged GW and reflect porous carbonate system.
3. Processed hydrogeochemical Data NETPATH (Plummer et al, 1994) was used for the calculation of the saturation indices (SI)
with respect to certain solid-phases. Reversible reactions (equilibrium, cation exchange, and
precipitation) or irreversible reactions (dissolution, weathering, diagenesis, diffusion and
evaporation) may take place in groundwater. The reaction of GW with aquifer minerals is governed
by the change of Gibbs free energy (free enthalpy) that defines the equilibrium constant K:
lnK =- ∆G
°
RT, logK =
- ∆G°
2.3RT, pK =
∆G°
2.3RT
Where K equilibrium constant for the reaction a A + b B c C + d D
Figure 5 (Right) Nitrate (mg l-1) in GW and rainwater. The black diamond indicate polluted sites (#10, 4 and 1)
For Dolomite-SI only two samples (#5 and 6) are close to the saturation locus. This would say
that the aquifer might have Dolomite rocks. However, the predominance plot (Figure 10) clearly
illustrates that the carbonate-rock is not Dolomite but it is made-up of Calcite and/or Mg-Calcite(Stumm and Morgan, 1981). Mg-Calcite weathering may release Mg accompanied by Calcite
and/or Aragonite precipitation through incongruent dissolution.
For Halite and Gypsum, no GW sample, Figure 9, shows significant dissolution, i.e. the
aquifer in void of evaporites. Consequently, Na, Cl and SO4 are provided by alteration of non-
carbonate rocks and/or by maritime aerosols rather than being acquired from evaporites.
4. Isotope DataThe isotope hydrology approach (Kendall and McDonnell, 1998) and the isotope geochemical
methods (Allègre, 2005) have introduced sets of tools applicable to groundwater geochemistry. The
collected rainwater and GW samples were measured for 18O and 2H contents, Table 1, and plotted in
Figure 12 with reference to the Meteoric Water Line (MWL):
δ 2H%o = 8.00 δ 18O%o + 10.0 MWL (Craig line)
δ 2H%o = 8.17 δ 18O%o + 11.3 MWL (Rosanky line)
The “per mil” isotopic ratio is given with respect to the Standard Mean Ocean Water, SMOW:
δ%o/SMOW=R sample - R standard
R standard* 1000
Where R is the isotope ratio
The MWL gives an expression of the dependence of the rainwater isotope composition on the
“local mean annual air-temperature”, Ta °C, (Mazor, 2004). It has been shown (Siegenthaler and
Oeschager, 1980) that isotopically depleted rainwater is observed in seasons with low temperatures.
The equations of that temperature-dependency (Dansgaard, 1964) are:
δ 18O%o = (0.7 Ta °C) - 13.75
δ 2H%o = (5.6 Ta °C) - 100
The isotope composition of Bangui GW is (-2.31 to -0.67%o) and (-8.62 to +5.59%o) for 18O
and 2H, respectively (Table 1). The mean values are about -2%o and -4%o for δ 18O and δ 2H,
respectively. The isotope composition of the rainwater samples collected during the dry season
(with mean values of +1.4%o and +25.2%o for δ 18O and δ 2H, respectively) correspond to mean
air-temperatures of 21.6 and 22.4°C, respectively (mean value = 22°C). The unique rainwater sample
collected by the start of the humid season of 2007 (on April 28) is significantly depleted in 18O and2H (-4.46 and -39.5%o, respectively). It corresponds to air-temperature of 12 °C.
The significant difference of the isotopic composition of the dry and the humid seasons mayallow us to consider a mixture of two precipitation poles for calculating the mixing fraction. This
fraction shows that the aquifer is equally recharged during the two seasons. This may be
astonishing. However, it seems that the heavy showers of the humid season directly go to run-off
(due to rapid saturation of the low permeable Oxisols) rather than contributing to aquifer recharge.
The well-known “continental effect” makes the inland precipitation isotopically lighter than
that taking place near the ocean. If the same air mass is moving inland from Cameron to Bangui, the
isotope content of precipitation (and that of GW) should be more depleted in Bangui than in
Cameron. However, we observe a reversed situation, Figure 12.
It could be said that the reason of that inversion is the higher temperatures in Bangui. Higher
air-temperatures in Bangui than in Cameroun would comfortably account for the enriched isotope
contents in Bangui rain and GW compared to those of Cameron. Sigha-Nkamdjou, 1999, mentioned“inverse continental effect ” when moving eastward inside Cameron and interpreted that isotope
effect because of an “inverse movement of humid air-masses westward.” The last interpretation may
look good but needs support of relevant meteorological data. However, other factors may diminish
the isotope content of rainwater in Cameroun. Heavy storm events have isotopic signatures lighter
than for fair precipitation. The “amount effect” is interpreted because of: 1) heavy rainfall events
follow the formation of thick clouds that are naturally more depleted in 18O and 2H, and 2) the
saturated air during heavy raining-out prevents evaporation of raindrops. It may be possible that the
“amount effect” is more pronounced, and give rise to lighter isotope content, in Cameron than inBangui. Furthermore, the “altitude effect” (Bortolami et al, 1978) would result in depletion by
1.5%o for δ 18O and 12.5%o for δ 2H for each 500 meters of higher altitude. The altitude of
Bangui is only about 400amsl whereas topography in Cameron is much higher. Consequently,
rainwater in Cameron will be isotopically lighter than in Bangui.
So, the previous interpretation of the “inverse continental effect ” is replaced by three simple
explanations: 1) higher temperatures, and 2) less efficiency of “amount effect” and 3) lower
altitude, for Bangui than in Cameron.
For the river-water samples, from the south-east of Cameron, the isotope observations (Figure
11) demonstrate that the increase of the total dissolved solids could mainly be attributed to mixing
of different water-masses. The diagram is showing that river-water mineralization is only slightly
affected by evaporation and, to a much less extent, by dissolution. This means that these rivers donot significantly discharge groundwater but, rather, they reflect rapid response to runoff. Figure 11
shows a considerable range of δ 18O and δ 2H values for Cameron, both for precipitation and river-
water but with slight evaporation effect. The observed range for the isotope contents for Cameron
rainwater could also be mainly attributed to different air-temperatures, altitudes and amount effect.
The lack of evaporation is due to rapid runoff and high air-humidity.
Many isotope compositions (Figure 12) are to the left of the MWL for both Bangui and
Cameron data points. This strikingly indicates significant “recycling” of the intensive local
transpiration, as expected for the equatorial and subtropical forest regions. Such recycling, together
with the little evaporation, means that: 1) the evapotranspiration term in Bangui aquifer water-
balance is mostly dominated by the transpiration component (but the drainage-basin water-balance
should have an important evaporation term) whereas the water-pools formed after heavy rainfall is
lost to the atmosphere without contributing to aquifer recharge, 2) the moisture responsible of
subtropical rainfall has two sources; a) from external air-masses (traveling long distances across the
Atlantic Ocean) and, b) from local origin (recycled transpiration). This implies that the linear
elongation of the isotope data points along Craig line would reflect mixing of these two moisture-
poles. If, moreover, the assumed “recycled-transpiration” was more enriched in the heavy isotopes
in Bangui than in Cameron, this would give a 4th interpretation for the relative enrichment of
Bangui GW compared to precipitation and GW in Cameron and reinforces the postulation that
water-losses in the subtropical regions are mostly consumed in transpiration which is a “biological-
pump” almost free of isotope fractionation.
The isotopic composition of Oubangui river is midway between the mean isotopic of BanguiGW and the mean isotopic composition of the non-evaporated rainwater samples (#14, 17, 18 and
19) collected during the dry-season (but the river is isotopically slightly-shifted toward the
composition of rainwater). This means that >0.50 of Oubangui river water is fed by runoff, even
during the dry season, whereas the rest is fed by the base-flow. This is in conformity with
hydrodynamics of river head-reaches during its low-stage. During the wet-season, the isotopic
composition of the river-water will rapidly get depleted (toward -4 and -25%o for δ 18O and δ 2H,
respectively) through higher contribution of runoff to river-discharge at its high-stage, and the river
will recharge the aquifer. The “mean” isotopic composition of GW during the two seasons would
not significantly change; what will change is the isotopic composition of rainwater and river-water.
Bangui GW mineralization mechanisms are shown in Figures 6, 7 and 8. As illustrated by the
diagram of δ 18O versus individual ions the process seems being controlled (in the dilute GW) bythe initial composition of rainwater, but slightly modified by the contribution of silicate-rock
Figure 10 (Left) Predominance diagram for carbonate minerals in Bangui GW
Figure 11 (Right) Isotope compositions versus electric conductivity in river water (Cameron)
The dilute GW samples show more scatter for the isotope contents than the charged GW
samples. This scatter may correspond to sites where the aquifer is dynamic (GW with short turn-
over time represented by wells positioned on a fractured formation). When the aquifer is highly
dynamic the GW isotopic content will rapidly reflect any change taking place in the isotope and
chemical input-functions resulting in the variation of GW isotopic content and keeping the GW
dilute. On the contrary, a less dynamic aquifer will enjoy more time for homogenization and
damping of the input signal toward a “mean” isotopic composition and acquiring higher ionic
strengths. It can be seen (Figures 6, 7 and 8) that the stabilization of the isotope composition of the
less dynamic sites is approaching a “mean” isotopic content of -2%o and -4%o for 18O and 2H,
respectively.
The “stagnant” GW samples were collected from wells in the south of the city (Figure 1) and
have the highest HCO3, Ca and Mg contents. We believe that this section has porous carbonate-
rocks. Its relatively high ionic strength confirms GW reaction with carbonates and validates the
interpretation of the longer residence-time in the southern section of the urban-zone.It is obvious that any pollution event that would take place in this relatively “stagnant” part of
the aquifer will be difficult to control and remove. Unfortunately, the highly populated parts of the
city are established in the south of the urban-zone, where we observe the said “stagnant” sections.
Moreover, GW flow in Bangui aquifer takes place from the north to the south (Doyemet, 2006).
This makes things more complicated in the southern residential areas of the city.
Figure 12 Stable isotope composition of Bangui and Cameron water samples (in the legend: S=slope, I=intercept)
Conclusions
The collected samples allowed exploring the hydrochemical and isotopic composition of
Bangui GW and revealed anthropogenic contamination by nitrate downtown. Fractured silicate
rocks regulate the dilute GW whereas carbonate rocks add more solutes in the porous sections.
Biogenic CO2 from the unsaturated zone buffers the effect of carbonates and shifts the pH to acidic.
Rock alteration is the main mineralization mechanism but cation exchange also takes place. 18O and2H show no evidence on evaporation whereas transpiration is dominant and recycled. The southern
and central sections of the city have relatively charged and polluted GW whereas the eastern and
the northern sections have dilute-GW. The aquifer less-dynamic parts are toward the southern
sector in porous calcite formations, whereas the other sections are dynamic (fractured non-
carbonate and carbonate rocks). The “inverse isotope continental effect ” is attributed to basic
concepts of temperature, altitude and amount effects. Aquifer recharge seems equally fed through
the dry and humid seasons due to the low permeability of the Oxisols giving rise to runoff
dominance under heavy showers.
ReferencesAllègre C., 2005
Géologie Isotopique. Editions Belin Sup, Paris, 495pBortolami, G. C., Ricci, B., Suzella, G. F., and Zuppi, G. M., 1978. Isotope hydrology of the Val Coraoglia, Martime
Alps, Piedmont, Italy. In: Isotope Hydrology, IAEA, Vienna, 327-350Kennedy W. Q. 1964. The structural differentiation of Africa in the Pan-African (±550 million years) tectonic episode.
8th ann. Rep. Res. Inst. Afr. Geol., Leeds Univ.,UK., p: 48-49Dansgaard, W., 1964 (in: Mazor E., 2004). Stable isotopes in precipitation. Tellus 16, 436-469Doyemet A. 2006. Le Système aquifère de la région de Bangui (RCA)
Thèse de docteur de l’université. Univ. des Sciences et Technologies de Lille, France, 108pHarte J. 1985. Consider a Spherical Cow. A course in Environmental Problem Solving. William Kaufmann, Inc. Los
Altos, California, 283pMazor E., 2004. Chemical and Isotopic Groundwater Hydrology. Marcel Dekker, Inc. New York - Basel, 453pPlummer N. L, Prestemon E. C. and Parkhurst D. L, 1994. An interactive code (NETPATH) for modeling net
geochemical reactions along a flow path Version 2.0. U.S. Geol Survey, Water Res. Invest. Report 94-4169, Reston,Virginia, 130p
Sigha-Nkamdjou L., 1999. Fonctionnement hydrologique d’un écosystème forestier de l’Afrique Centrale: La Ngoko a
Moloundou (Sud-est du Cameroun). Travaux Documents Microfichés (TDM) No 111-F5, ORSTOM Éditions, Paris(réf. local à l’Alliance Française a Bangui no. 05590)
Siegenthaler U. and Oeschager H. 1980 (in: Mazor, 2004). Correlation of 18O in precipitation with temperature and
altitude. Nature 285, 314-317
Stumm W. and Morgan J, 1981. Aquatic Chemistry: An introduction emphasizing chemical equilibria in natural waters.
John Wiley & Sons, New York, 780
White W. M., 2007. Geochemistry. e-book on the Internet, 15 chapters, 701 p