1154-1161Vol. 78 No.5ACTA GEOLOGICA SINICAOct. 2004Water Self-Softening Processes at Waterfall SitesCHEN Jing’an', David Dian ZHANG', WANG Shijie' and XIAO Tangfu'1 State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry,Chinese Academy of Sciences, Guiyang, Guizhou 550002;E-mail: jinganchen @vip.163.com2 Department of Geography and Geology, University of Hong Kong,Pokfulam Road, Hong KongAbstract Many rivers in tropical and subtropical karst regions are supersaturated with respect to CaCO3 and have highwater hardness. After flowing through waterfall sites, river water is usually softened, accompanied by tufa formation,which is simply described as a result of water turbulence in fast-flowing water. In this paper, a series of laboratoryexperiments are designed to simulate the hydrological conditions at waterfall sites. The influences of air-water interface,water flow velocity, aeration and solid-water interface on water softening are compared and evaluated on a quantitativebasis. The results show that the enhanced inorganic CO2 outgassing due to sudden hydrological changes occurring atwaterfall sites is the principal cause of water softening at waterfall sites. Both air-water interface area and water flowvelocity increase as a result of the“aeration effect",“l(fā)ow pressure effect" and“jet-flow effect" at waterfall sites, whichgreatly accelerates CO2 outgassing and therefore makes natural waters become highly supersaturated with respect toCaCO3, consequently resulting in much CaCO3 deposition and reduction of water hardness. Aeration, rapidly increasingair- water interface area and water flow velocity, proves to be efctive in reducing water hardness. This study may providea cheap, safe and effective way to soften water.Key words: water softening. CO2 outgassing, CaCO3 precipitation, waterfall site1 Introductionbecause of the lack of free energy to create new surfaceareas, unavailability of reactive calcite to act as nucleationThe total hardness of natural waters, as an importantites, and inhibition of PO上, Mg2+ and organic ligandsindicator of water quality, is defined as the sum of calcium(Bermer, 1975; Reddy 1977; Dandurand et al. 1982;and magnesium concentrations. It is often expressed as theBuhmann and Dreybrodt 1987; Lorah and Herman 1988;equivalent amount of calcium carbonate that could beLebron and Suarez 1996). This is why river water canformed from calcium and magnesium in the solution (Yangmaintain high hardness while no calcite precipitationet al, 1999). High hardness is harmful at domestic andappears in many karst rivers. In the Colorado River systemindustrial scale because substantive calcite may form andof the southwest United States, calcite precipitation is notbe accumulated when being heated in pipes, boilers anddetectable although the water reaches 4 -6 timescooking utensils (Saurina et al, 2002). Moreover, highsupersaturation with respect to calcite (Suarez 1983).hardness of drinking water may increase the risk of someIt is well known that high supersaturation of water withdiseases such as atopic dermatitis (Miyake et al., 2004).respect to calcite usually results from the removal of CO2Many rivers in tropical and subtropical karst regions havefrom the water. Several studies have attributed the COrhigh water hardness. However, it has been found that waterremoval to turbulence, mixing of different waters andhardness is reduced when the river water flows throughmetabolic uptake of CO2 by photosynthetic plantswaterfall sites, accompanied by calcite precipitation and(Jacobson and Usdowski, 1975; Chafetz and Folk, 1984;tufa formation. This is referred to as self- softeningHerman and Lorah, 1987; Viles and Goudie 1990). Studiesprocesses. Study on the water self-softening processes mayof waterfall tufa have shown that a great amount ofbe helpful for finding effective water softening measures.inorganic CO outrassing occhrred at waterfall sitesPrecipitation of calcite in natural waters can be simply(Cha中國煤化工d Hernan 1988; Ford,described by the following reaction:1989;YHC N M H Get al, 2002). ThroughCa2*+2HCO;女CaCO3↓+ CO2↑+H2O(1)detailea tela measurements ot temperature, pH, calciumconcentration and alkalinity of water samples from twoGenerally, precipitation of calcite requires the waters tosmall streams in Southwest Germany, Merz-Preiβ andbe 5 to 10 times supersaturation with respect to calcite.Vol.78 No.5ACTA GEOLOGICA SINICAOct.20041155Riding (1999) found that the principal cause ofresulting from detraining of the dissolved gases from thesupersaturation in fast-flowing streams is inorganic carbonwater flow. The“jet-flow effect" usually results in muchdioxide outgassing while photosynthetic uptake of carbon larger air-water interface area because the jet flows consistdioxide and temperature effects are negligible. This isof sprays, droplets and broken streams (Chanson andtestified by the field observation that organisms exist alongCummings, 1996; Zhang et al, 2001). Thus it can be seenthe entire river channel while hardness reduction and tufathat the main hydrological changes, occurring at waterfalldeposition occur mainly at waterfall sites. However,sites as a result of these effects, are the enlargement of air-hardness reduction or calcite precipitation at waterfall siteswater interface area and the increasc of water flow velocity.has been simply described as a result of water turbulence.We believe these two hydrological changes are the majorKnowledge about how environmental conditions change atcause of water softening or calcite precipitation. In order towaterfall sites and how these changes affect water hardnesstestify our hypothesis, a series of laboratory experimentsis rather scarce. We believe the principal cause of hardnesswere designed to simulate changes in air-water interfacereduction is the sudden hydrological change occurring atarea and flow conditions at waterfall sites, and to evaluatewaterfall sites. In this paper, a series of experiments arequantitatively their influences on water softening.designed to simulate the hydrological conditions atIn order to acquire quantative data on water softeningwaterfall sites, and various factors affecting water softeningrates, we need to continuously monitor Cat+ concentrationare evaluated quantitatively.changes. In fact, it is impossible to determine Ca2+concentrations continuously and accurately during the2 Experiments and Methodssimulating processes because the simulating conditionswould be changed if a part of solution were taken out forUnderstanding the hydrological changes occurring aCa2concentrationdetermination.Fortunately,waterfall sites is the key to design simulating experimentsconductivity, which along carbonate depositing systemsand thus to investigate the factors controlling waterfluctuates primarily as a result of carbonate depositionsoftening. Three major changes occur when river channel(Drysdale et al, 2002), is often a good tracer of Ca2+flow approaches to a waterfall. Firstly, sudden changes inconcentration variations in natural karst waters (Groleau etflow conditions at waterfall sites can lead to air entrainment,al., 2000; Drysdale et al., 2002). In order to testify whetherwhich is termed“natural aeration". In rivers, aeration canconductivity variability can represent Ca2+ concentrationappear in high-velocity flow, plunging free-jet flow, thevariations, Ca'+ concentrations and conductivities ofwakes of topographic iregularities on channel beds, and inlaboratory solutions with different Ca2+ concentrationshydraulic-jump configurations, which suck and trap airwere measured respectively on an atomic absorptioninside the water body and create many air bubblesspectrometer (PE5100) and a portable multi-parameter(Chanson and Qiao, 1994; Chanson and Cummings, 1996;instrument (pIONneer 65). The good correlation betweenChanson and Toombes, 2003). These conditions are mostthem (Fig. 1) suggests that conductivity can be used as aobvious at waterfall sites, where“white-water phenomena”surrogate measureof Ca2+ concentrations. Ca2*+occur. Here we call the phenomena induced by airconcentrations can be calculated according to the initialentrainment as the“aeration effect".Ca+ concentrations and the conductivity in the sampleSecondly, water pressure at high velocity at waterfallsolutions. The water softening rates can be deduced fromsites is reduced according to the Bermoulli effect. At lowerthe Ca+ concentration variations.pressures, dissolved gases can be released from water asCaCO3 solution for simulating experiments was preparedtiny air bubbles according to the Henry' s law. This resultsin air detrainment at waterfall sites and is called “l(fā)ow1500ppressure effct".y= 5.37x+ 17.52耳1200-Finally, the fast-flowing and falling water at waterfallR = 09995sites is broken into many water droplets, small streams andsprays due to the initial jet-flow turbulence and to the shearforces of the surrounding air. We call this phenomenon“jet-flow effectr.中國煤化工The“aeration effect creates air bubbles at waterfall sites二150 200 250that can greatly increase the size of the air-water interfaceTHCNMHGon(mgy)area. The“l(fā)ow pressure effect" at high flow velocity notonly creates bubbles and thus enlarges the air-waterFig. 1. Correlation between Ca2+ concentrations andconductivity in sample solutions.interface area, but also reduces CO2 content in the water1156Water Self-Softening Processes at Waterfall SitesChen et al.by adding pure CaCO3 grains in a 10-liter glass containerwas put into a cylindrical glass vessel with a water-surfacefull of ditilled water. Pure CO2 gas was then continuouslyarea of 55.65 cm'. The glass vessel was put on a magneticpumped into the water for 72 hours. The CaCO3 solutionbeater and a round magnetic rod with a length of 4 cm waswas passed through a 0.45-μm filter before used as sampleput in the middle of the bottom of the vessel in order tosolutions for experimental studies. The initialmake the solution move. When the magnetic beaterconductivities of the sample solutions range from 1050 toworked, the sample solution had a rotation rate of 301210 us/cm.revolutions per second (i.e., an average flow velocity of 4All the following simulating experiments were carriedm/s). Conductivity variations were measured under flowingout in a cultivator box set at 20°C in order to eliminate thecondition at time intervals of 10 -20 minutes at theinfluence of temperature on calcite precipitation. Changesbeginning and at time intervals of 30- 60 minutes when theof conductivity and pH in sample solutions were measuredsample solution showed only small conductivityrespectively on a portable multiparameter instrument andfluctuations. As a comparison, 550 ml of sample solutionan Orion 818 pH meter. The pH electrode was calibratedwith an initial conductivity of 1049 μs/cm was put into awith pH 4.00 and 6.86 buffers every two hours.cylindrical glass vessel with the same water-surface area of55.65 cm, and conductivity variations under stationary1.1 Experiments on water softening processescondition were monitored every 10- 20 minutes at theThe first aim of this experiment is to systematicallybeginning and every 30- -60 minutes when the solutioncomprehend the water softening processes and the secondgradually approached to equilibrium.aim is to examine the dependence of water softening onCO2 outgassing. 550 ml of sample solution was put into a1.4 Aeration experiments11.21 cm-diameter cylindrical glass vessel with a water-The aim of this experiment is to evaluate quantitativelysurface area of 98.70 cm2. The pH and conductivity of thethe influence of aeration on water softening. 550 ml ofsample solution were measured under stationary conditionsample solution with an initial conductivity of 1047 μs/cmat time intervals of 10- -20 minutes at the beginning in orderto understand the fast-changing process, and at timearea of 55.65 cm'. A bubble producer was designed tointervals of 30- 60 minutes when the sample solutionsimulate the aeration effect. Airflow with normalshowed only small fluctuations in pH and conductivity andatmnospheric CO2 partial pressure passed through a tube tofinally appeared to reach equilibrium. The“equilibrated"the bubble producer in the sample solution to produce aboutsolution was filtered through a 0.45-um filter, and its Ca2+4000 air bubbles in 0.2 second, which was the period forconcentrations were measured on an atomic absorptioneach bubble to rise from the small holes of the producer tospectrometerthe water surface. Thus a total water- air interface area ofabout 602 cm2 was added to the original air-water interface.1.2 Air-water interface experimentsConductivity variations were measured under acrationIn order to evaluate quantitatively the influence of air-condition at time intervals of 1-5 minutes at the beginningwater interface on water softening and calcite precipitation,and at time intervals of 5- -20 minutes when the sample550 ml of sample solution was put into three cylindricalsolution gradually approached to equilibrium.glass vessels respectively with different water surface areasof 55.65, 98.70 and 155.49 cm2. These sample solutions1.5 Solid-water interface experimentshad the same initial conductivity of 1052 us/cm.In order to evaluate quantitatively the influence of solid-Conductivity changes with time in the three samplewater interface on water softening and examine the relativesolutions were measured on a portable multi-parameterimportance of air-water interface and solid-water interface,instrument under stationary condition at time intervals of550 ml of sample solution was respectively put into twoabout 10 hours until the solutions appeared to stopcylindrical glass vessels with the same surface area of 55.65precipitating. Ca2+ concentrations of the initial solutionscm*, one of which a calcite tablet with a surface area ofand“equilibrated”solutions were measured on an atomic65.31 cm2 was put in the middle of the bottom. The sampleabsorption spectrometer after they were filtered through asolutions had the same initial conductivity of 1180 us/cm.0.45-μ m filter.Condtwo solutions weremonit中國煤化Iner istrument under1.3 Flowing water experimentsstatiolYHC N M H Gof about 10 hours untilThe main aim of this experiment is to compare the waterthe solutions appeared to stop precipitating. Ca2+softening rates in standing and flowing waters. 550 ml ofconcentrations of the initial solutions and "equilibrated"sample solution with an initial conductivity of 1031 us/cmsolutions were measured on an atomic absorptionVol.78 No.5ACTA GEOLOGICA SINICAOct, 20041157spectrometer after they were filtered through a 0.45-umstage, the supersaturation degree of water with respect tofilter in order to compare the water softening rates.calcite rose to such a high level that the nucleation barriercan be overcome and the fast calcite precipitation started.2 ResultsMuch H* was released into the solution during theprocesses of calcite precipitation, leading to the decrease of2.1Water softening processes observed underpH, which can be shown by equation (4). On the otherstationary conditionhand, the released H+ and HCO5 were gradually convertedVariaions in pH, conductivity and calcite precipitationto CO2 (Dreybrodt et al, 1997), leading to the reduction ofrate with time under stationary condition are shown in Fig.H* contents, which can be shown by equation (5). When a2, from which the following three stages were identified.balance was reached between the consumption ofH+ and itsThe water softening rates can be reflected by the calciteproduction, the pH value would remain stable. When theprecipitation rates because the reduction of water hardnessconsumption rate of H* outpaced its producion rate, the pHresulted mainly from the precipitation of calcite.value would rise. When the conversion rate slowed down,he pH value would again decline. These processes2.1.1 The fast outgassing stagealtermately appeared during the whole stage and led to theThis stage lasted about 25 hours and was characterizedfrequent fluctuations of pH (Fig. 2). The calciteby the constant conductivity and a rapid increase in pHprecipitation rates ranged from 8.0 to 92.0 μg/(:min)(Fig. 2). The Pco, difference between the ambientduring this stage with an average of 27.6 ug(-min) (Fig.atmosphere and the sample solution resulted in quick2).diffusion of CO2 from the water to the atmosphere, thusCa2*+2HCO5女CaCO3↓+ HCO3 +H+(4)leading to the reduction of H* contents. This process can beH*+HCO3臺CO2↑+H2O(5)shown by the reaction (2).H*+HCO3白H2CO3白CO2↑+H2O(2)2.1.3 The equilibrium stageThe supersaturation degree of water with respect toThis stage was characterized by the relatively stablecalcite is usually expressed by IAP/Ke, where IAPconductivity and pH value (Fig. 2). The rates of CO2represents the ionic activity product and Ke represents theoutgassing and water softening slowed down because theequilibrium constant of CaCO3. One convenient method forPco2 difference between the atmosphere and the samplethe computation of IAP is:solution decreased and the solution gradually approachedto“equilibrium". The calcite precipitation rates were belowIAP=(Cat)(CO3 )=8.0 g(-min) with an average of 4.2 μg/(l min) Fig. 2).{rcz*X[Ca*]XTHco' -xK2XAlk}( H)(3)Although the conductivity of the sample solutionwhere rca2+ and THco- are respectively the activitydecreased continuously with time, the calcite precipitationcefficients of Ca+t and HCO5, [Ca2] is the Ca2+rates fluctuated from time to time (Fig. 2). Two reasonsconcentrations, K2 is the second dissociation constant formay explain this phenomenon. Firstly, the sample solutionH2CO3, Alk is the alkalinity of the water, (H) is the Htwas supersaturated with respect to calcite, so theactivity defined through measuring pH values (pH =-logconductivity showed a decreasing trend during the calcite(H). From equation (3), it can be easily seen that thesupersaturation degree of water with respect to calcite8.00increases with the reduction of H+ contents. Detailed. Conductivity一precipitation rate . pH200-calculations showed the supersaturation degree of the7.50sample solution with respect to calcite increases from 0.821000to 6.97 when pH rises from 6.22 at the beginning to 7.15 at3800 A7.00告the end of this stage. However, the constant conductivity of享音600the sample solution indicated calcite precipitation did not400commence. The calcite precipitation rate is almost equal to6.50zero during this stage (Fig 2).200f"NMwmrmas」中國煤化工000 120000 1500002.1.2 The fast softening stageThis stage lasted about 170 hours and was characterizedTYHCNMH G'by a sharp drop in conductivity (Fig. 2), which reflected theRig. 2. Variations in conductivity， pH and calciterapid reduction of Ca2+ concentrations or water hardness.precipitation rate in sample solutions under stationaryBecause of the great loss of CO2 during the fast outgassingcondition.1158Water Self-Softening Processes at Waterfall Sies .Chen et al.pecipitation processes. Secondly, the supersaturationtransfer rate may be expressed as:degree of water with respect to calcite varied as a result ofSC = K xax(Cum-Cm)(6)frequent pH fuctuations (Fig. 2) according to equation (3),causing the calcite precipitation rates to vary from time towhere Cps is the dissolved gas concentrations, K is thetime.mass transfer cofficient, a is the specific surface area andCa is the concentrations of the dissolved gas in water at2.2 The infuence of the air-water interface on watereqilibrium (Chanson, 1995; Chanson and Toombes,softening2003). The mass transfer coefficient (K) is almost constantVariations in Ca2+ concentrations under stationary(Kawase and Moo Yong, 1992). The specifc surface areacondition in the three sample solutins respectively with(x) is defined as the air water surface area per unit volumewater-surface areas of 55.65, 98.70 and 155.49 cm' areof air and water (Chanson, 1995; Chanson and Toombes,shown in Fig. 3.2003). In our experiments, the three sample soluions haveObviously, an enlargement of air-water iterface area notthe same volume, the same Csu and the same initial Cas, sOonly shortened the period for the sample solution to reachthe variations of the specific surface area can beeqilibriumn (Table 1), but also aclerated the waterrepresented by those of the air-water interface area, andsoftening rates indicated by a faster decrease in Ca^*equation (6) can be simplified as:concentations (Fig. 3). The high crrelation coffcientbetween the air-water interface area and the water sofening二Cu =Kx$x(Csu -Cm)(7)rates (Fig. 4) suggests that the water softening rates, to adgreat extent, depend on the air water interface area. Higherwhere K is a constant and s is the air-water interface area.ItPooz in the water than in the ambient atmosphere drivesis obvious from equation (7) that the air water interfaceCO2 to dffuse quickly from the water to the air. All otherarea decides the CO2 outasing rate, thus contolling thethings being equal, the larger the air-water interface area is,water softening rates.the faster the CO2 will diffuse, the earlier the calciteMer-Preiβ and Riding (1999) found that a greaterpecipitation will commence. In fac, this phenomenon canbe well explained by the Table 1 Comparison of water softening rates in solutions with dferent air-water interfacediffusion theory. The mass areastransfer rate of a chemicalSurface. areaInitial Ca2*Equilibrated Ca2* Bquilibrium timeWater softeningspecies across an interface isconcentration (mg/) concentration (mg/)b)rate (mg/-H)a function of the moleculardiffusion cofficient, the55.665254.638.83120.69negative gradient of gas98.702470.88concentrationandheinterface area. If the155.4937.91731.25chemicalf interest ivolatile (e.g. CO2), the gas1.50r55.65 cnm'.98.7 cmf 155.49 cm?y= 0.0057x+ 0.3559240fR' = 0.98922001.20-160f40|; ili i; ...中國煤化工140 170fYHC N M H Grea(em)Time ()Fig. 3. Vriations in Ca2+ concenrations in sample solutionsFig. 4. Crrelation between the air water interface area and thewith different air-water interface areas.Vol.78No.5.ACTA GEOLOGICA SINICAOct. 20041159channel width (i.e, larger air-water interface area) allowedis the same for the same temperature, gas and solvent.more rapid CO2 outgassing while a narrower channelTherefore, dissolved gases can diffuse more quickly fromlimited CO2 outgassing and calcite precipitation. Thisflowing water than from stationary water. On the otherprovided direct field evidence for our experiment results.hand, fast flowing water not only induces turbulence whichcan cause more effective collision among dissolved ions2.3 The infuence of flow velocity on water softeningand, therefore, accelerates chemical reactions, but alsoVariations of Ca2+ concentrations in the sample solutionsreduces the thickness of diffusion boundary layers at bothunder stationary and flowing conditions are shown in Fig.solid-water and air-water interfaces, which acceleratesmass transfer through the two interfaces and thus speeds upFrom Fig. 5 and Table 2, it can be seen that the watercalcite precipitation or water softening (Liu et al, 1995;softening rate in flowing water is more than four timesZhang et al, 2001). All these factors would jointly causelarger than in stationary water, while the eqilbrium time ismuch earlier and faster hardness reduction in flowing wateronly one-fourth of that in stationary water. This gives clearthan in stationary water. This phenomenon also occurs inevidence for the influence of flowing conditions on waterthe field. Liu et al. carried out in-situ experiments tosoftening.measure calcite deposition rates at the dam sites with fastThe behavior of a fluid under varying flowing conditionswater flow (0.5- 2 m/s), as well as inside pools with stillis described quantitatively by Bernoulli's law:water in the Huanglong Ravine, China. The result showedthat the depositional rates in fast flowing water were fourP+zpv2 + pgh =[constant]8)times high those in still water although there was nodifference in hydrochemistry (Liu et al., 1995). The goodwhere P is the static pressure (Nm), ρ is the fluid densityagreement between the field measurements and ourlaboratory observations provides convincing evidence that(kg/m), v is the velocity of fluid flow (m/s) and h is thethe flowing conditions can exert a large influence on waterheight above a reference surface (m). Obviously, ansoftening rates in natural waters.increase in flow velocity will result in a decrease in staticpressure from equation (8), so the water pressure is lower in2.4 The influence of aeration on water softeninga moving fluid than in a stationary fluid.Variations of Ca2+ concentrations with time in sampleIt is well known that the solubility of a gas in waterdecreases with decreasingpressure if the temperature Table 2 Comparison of water softening rates in solutions under stationary and flowingstays constant according to conditionsthe Henry' s law, i.e,Initial Ca2*Equilibrated Ca*Equilibrium timeWater softeningFlow conditionP=k.C(9)concentration (mg/)concentation (mg/)(hr)rate (mg/.h)where P is the static pressureStationary254.638.8120.69(N/m2)，C is the gasconcentration and k is theFlowing251.738.1703.05Henry's law constant, which300astationary。flowing250。aeration● stationary200160-12010中國煤化工300060009000 12000 15000YHCNMHG12000 15000Time (min)1HN wwun)Fig.5. Variations in Ca2+ concentrations in sample solutionsFig.6. Variations in Ca2* concentrations in sample solutionsunder stationary and aeration conditions.under stationary and flowing conditions.1160Water Self- Softening Processes at Waterfall SitesChen et al.solutions under acration and stationary conditions arresulting in much CaCO3 deposition and fast reduction ofshown in Fig. 6.water hardness. This is the major cause of water self-It is obvious from Fig. 6 that the Ca2+ concentrations ofsoftening at waterfall sites and can be simply explained bythe sample solution decline very rapidly under aerationthe model shown in Fig. 8.condition, dropping to 15% of the original level in 25From the preceding discussions, we can see that CO2hours. In comparison, the Ca2+ concentrations of the sampleoutgassing plays the most important role in water softeningsolution under stationary condition show a much flatterand an air-water interface is necessary in order to completedecreasing trend. Detailed calculations show that the watersoftening rate under aeration condition is more than 12350times larger than in stationary water, while the equilibrium300▲with tablet● without tablettime is only one-twelfth of that in stationary water (Table250On one hand, the bubbles from the holes increased the150fwater-air interface area by about 602 cm2. On the otherhand, the rising bubbles induced strong water movement.These together greatly promoted CO2 outgassing and50 100 150 200 250 300 350 400accelerated water softening.Time (1)2.5 The influence of solid-water interface on waterFig. 7. Variations in Ca* concentrations in sample solutionssofteningwith/without calcite tablet.Although the increase of solid-water interface areaaccelerated water softening Fig. 7), it didn't affect thesoftening rate so much as the increase of air-water interfaceHydrological changes at waterall sitesarea (Fig. 3). Moreover, when river water flows over awaterfall, there will be a much greater increase in air-waterinterface area than in solid-water interface area. Therefore,Acraon effect| Low pressure efectJet-flow effectthe air-water interface is much more important incontrolling water softening than the solid-water interface atwaterfall sites. CO2 outgassing plays a major role inIncrease of air-water interfacc area|[Increase of water flow velocity|reducing water hardness.3 Discussion and ConclusionsFast CO2 outgassingMany rivers in tropical and subtropical karst regions areWater becoming highly supersaturated with respect to CaCO,supersaturated with respect to CaCO3 and have high waterhardness because of the high contents of dissolved Ca2*+↓derived from groundwater. After river water flows throughPrecipitation ofCaCO,waterfall sites, CaCO3 can be deposited and river water willbe usually softened.Our experiment results showed that the enlargement ofReduction of water hardnessair-water interface, the increase of water flow velocity andaeration condition can greatly accelerate CO2 outgassingand lead to fast reduction of water hardness. At waterfallFigure 8 A simple model for water self softening processes atwaterfall sites.sites, both air-water interface area and water flow velocitytend to increase as a result ofTable 3 Comparison of water softening rates in solutions under stationary aerationthe "aeration effect", “l(fā)owconditionspressure effect" and“jet- floweffect", which will greatlyConditionconcentration (mg/1)Initial Ca2+Eq中國煤化工Precipitation rate(mg/-h)accelerate CO2 outgassingand, therefore, make natural:MYHCNMHGwaterbecomehighlyStationary254.638.310.69supersaturated with respect toAeration37.1258.70CaCO3,consequentlyVol. 78 No. 5ACTA GEOLOGICA SINICAOct, 20041161the softening processes. However, there is no air-waterCO2: The conversion to CO2 by the slow process H*+HCO3~→interface in the tap water pipe, which is a closed system, soCO2+H2O as a rate limiting step. Geochimi. Cosmochim. Acta,361: 3897-3904.CO2 outgassing does not occur and the water hardnessDrysdale, R.N,, Taylor, M.P.. and Ihlenfeld, C, 2002. Factorsremains unchanged from the pipe entrance to the exit. Incontrolling the chemical evolution of travertine-depositingorder to reduce the hardness of tap water, necessaryrivers of the Barkly karst, northern Australia. Hydrologicalmeasures should be adopted before the water enters theProcesses, 16: 2941-2962.pipe system. Aeration, which can greatly increase air-waterFord, T.D., 1989. Tufa: a freshwater limestone. Geology Today, 5:60-63.interface and water flow velocity, has proved itself to be anGroleau, A., Sarazin, G, Vincon-Leite, B., Tassin, B., andeffective way to reduce water hardness in our experiments.Quiblier-Lloberas, C, 2000. Tracing calcite precipitation withIn fact, acration processes have been successfully adoptedspecific conductance in a hard water alpine lake. Waterin water treatment to remove harmful gases, but not yetResearch, 34: 4151-4160.been applied to softening water until now. What we haveHerman, J.S., and Lorah, M.M., 1987. CO2 outgassing and calciteprecipitation in Falling Spring Creek, Virginia, USA. Chemi.learned from this research may provide a cheap, safe andGeol, 62: 251-262.effective way for softening water.Jacobson, R.L, and Usdowski, E, 1975. Geochemical controls ona calcite precipitating spring. Contributions to Mineralogy andAcknowledgementsPetrology, 51: 65-74.Kawase, Y, and Moo-Yong, M, 1992. Correlations for liquid-phase mass transfer coefficients in bubble column reactors withThis research was supported jointly by the CRCG SeedNewtonian and non-Newtonian fluids. Canadian J. Chemi,Grant of the University of Hong Kong and the NationalEng., 70: 48-54.Natural Science Foundation of China (Nos. 90202003 andLebron, I, and Suarez, D.L., 1996. Calcite nucleation and40303014). We thank anonymous reviewers for theirprecipitation kinetics as affected by dissolved organic matter athelpful comments and Professor Xu Zhongluen for25C and pH > 7.5. Geochim. Cosmochim. Acta, 60:2765- 2276.correcting the English manuscript.Liu, Z, Svensson, U, Dreybrodt, W., Yuan, D.X., and Buhmann,D.，1995. Hydrodynamic control of inorganic calciteManuscript received June 25, 2004precipitation in Huanglong Ravine, China: field measurementsaccepted July 26, 2004and theoretical prediction of deposition rates. Geochim.Edited by Liu XinzhuCosmochim. Acta, 59: 3087 -3097.Lorah, M.M., and Herman, JS, 1988. The chemical evolution ofa travertine-depositing stream: Geochemical processes andReferencemass transfer reactions. Water Resources Res, 24: 1541-1552.Bemer, R.A., 1975. The role of magnesium in the crystal growthMerz-Preiβ，M., and Riding. R., 1999. Cyanobacterial tufaof calcite and aragonite from seawater. Geochim Cosmochimcalcification in two freshwater streams: ambient environment,Acta, 39: 489- -504.chemical thresholds and biological processes. SedimentaryBuhmann, D, and Dreybrodt, W., 1987. Calcite dissolutionGeol, 126: 103-124.kinetics in the system H2O-CO2-CaCO3 with participation ofMiyake, Y., Yokoyama, T, Yura, A., Iki, M., and Shimizu, T,foreign ions. Chemi. Geol, 64: 89 -102.2004. Ecological association of water hardness with prevalenceChafetz, H.S., and Folk, R.L, 1984. Travertine: depositionalf childhood atopic dermatitis in a Japanese urban area.morphology and the bacterially constructed constituents. J.Environmental Res., 94: 33- -37.Sedimentary Petrology, 54: 289- -316.Reddy, M.M., 1977. Crystallization of calcium carbonate inChanson, H, and Cummings, P.D., 1996. Air-water interface areapresence of trace concentration of phosphorus-containedin supercritical flows down small slope chutes. Department ofanions, I, inhibition of phosphate and glycerophorus ions at pHCivil Engineering Research Series No.CE15I, University of8.8 and 25°C.J. Crystal Growth, 41: 287- -295.Queensland.Saurina, J, Avies, EL, Moal, A.L, and Cassou, S.H, 2002.Chanson, H, and Qiao, G.L, 1994, Air bubble entrainment andDetermination of calcium and total hardness in natural watersgas transfer at bydraulic jump. Deparment of Civilusing a potentiometric sensor array. Analytica Caca, 464:Engineering Research Series No.CE149， University of89- -98.Suarez, D.L， 1983. Calcite supersaturation and precipitationChanson, H, and Toombes, L, 2003. Strong interactions betweenkinetics in the lower Colorado River, All-American Canal, andfree -surface aeration and turbulence in an open channel flow.East Highland Canal. Water Resource Res, 19: 653- -661.Experimental Thermal and Fluid Science, 27: 525- -535.Viles, HA., and Goudie, A.S, 1990. Tufas, Travertines and aliedChanson, H, 1995. Air-water gas transfer at hydraulic jump withcarbonate deposits. Progress Phys. Geogr, 14: 19- 41.parially developed inflow. Water Research, 29: 2247- 2254.Yang CY Chin HF Chenc MF _Tsai, S.S., Hung, C.F., andDandurand, JL, Gout, R., Hoefs, J, Menschel, G. Schott, J, and中國煤化工ncer morality and totalUsdowski, E, 1982. Kinetically controlled variations of majoring water. Environmentalcomponents and carbon isotopes in a calcite-precipitatingYHCNMHGspring. Chemi. Geol, 36: 299 -315.Zhang, D. Zhang, Y, Zhu, A., and Chen, X, 2001. PhysicalDreybrodt, W, Elsenlohr, L, Madry, B., and Ringer, S., 1997.mechanisms of river waterfall tufa (travertine) formation. J.Precipitation kinetics of calcite in the system CaCO3-H2O-Sedimentary Res., 71: 205- -216.