Denudation Rates of the Hawaiian Islands by Rivers and Groundwaters1

YUAN-HUI LIZ

ABSTRACT: The carbonic acid produced by the microbial oxidation oforganic matter is the most important chemical weathering agent ofthe Hawaiian basaltic rocks. The total denudation rate of the Hawaiian Islands ranges from 11 to 50 mg/cmz /yr (or 0.04-0.19 mm /yr), The island of Hawaii has the lowest denuda tion rate among the Hawaiian Islands.

BASED ON THE REMOVAL RATE of dissolved River at Hilo on the island of Hawaii (USGS calcium by rivers, Moberly (1963) estimated Station No. 16213000, monthly data from the total denudation rate (including both October 1979 to September 1981). One excep chemical and physical) of the Kaneohe Bay tion is sulfate concentration data, which are watershed on the island of Oahu, Hawaii, to quite variable and show no direct correlation be about 34 mg/cm"/yr or 0.13 mm /yr (based with runoff. Another interesting observation on a density ofbasaltic rocks in the area of2.6 is the near constancy or only slight decrease 3 g/cm ). Since the denudation rate ofany river of chloride ion concentration as the run drainage area is a complex function ofrainfall off increases. The chloride ion in rain and rate, river runoff, vegetation coverage, tex river waters is supplied mainly by the sea-salt ture and composition of parental rocks, etc. aerosols from the ocean. The [SOi-]/[Cn (Garrels and Mackenzie 1971, Sherman and weight ratio in rainwater along Saddle Road, Ikawa 1968), this paper addresses the ques which is almost parallel to Wailuku River, tion of how comparable Moberly's estimated ranges from 0.20 near seashore to 0.86 inland value is to other areas in the Hawaiian island at an elevation of1625 m (Eriksson 1957). The . chain. The specific purposes of the present average is about 0.4 as compared to 0.14 paperare to demonstrate (1) what factors con for seawater, probably indicating an excess trol the chemical compositions of rivers and source ofdimethylsulfide (DMS), which is re groundwaters in the Hawaiian Islands, and leased to the atmosphere at the air-sea inter (2) how the chemical and physical denudation face and oxidized to SOz and sulfate (Bates rates in the Hawaiian Islands vary as func and Cline 1985). Other possible sources ofthe tions of various environmental factors. The excess sulfate are the anthropogenic inputs extensive water resources data base collected from the continents and local volcanic gases. by the U.S. Geological Survey annual report The volcanic input is, however, relatively from 1975 to 1984 is the basis for this work. small (Harding and Miller 1982). With an average chloride concentration of about 5-6 mg/liter in the Wailuku River water, the sulfate CHEMICAL COMPOSITION OF RIVERS AND contribution from rainwater should be about GROUNDWATERS 2-2.4 mg/liter in river water. The observed In general, the dissolved concentrations of [SOi-] of most samples of Wailuku River major ions, [M], in rivers decrease and the water are, however, much less than 2-2.4 mg/ suspended particle concentrations increase liter, indicating a high degree of sorption of

when the river runoff (Fw , in cubic feet per sulfate by soils (Hasan et al. 1970). second) increases. Figure 1 shows these rela The concentrations of dissolved organic tionships for water samples from the Wailuku carbon, organic nitrogen, inorganic nitrogen, and phosphate in the same water samples are also quite variable, and again show no rela 1 Manuscript accepted 23 July 1987. 2 Hawaii Institute of Geophysics, University of Hawaii, tionship to the runoff (Figure 2). Apparently, Department ofOceanography, Honolulu, Hawaii 96822. the direct source of these dissolved nutrients 253 254 PACIFIC SCIENCE, Volume 42, July/October 1988

Ill. I Ll. Suspended tr:.Ll.tA Ll. Ll.Ll.Ll. )( matter Ll. Ll.Ll. 2 - A 10 • Ll. If" - x A • ~ • x! .., • x Ll. Si 02 • 0 • o 000 • • A Q) ~o ~ • 0 • • 0 0 • x J( 00 • 0 0 T. • • Ca++ 0 00 0 ++ 10 x x lO( >c 0 Mg - ... Jt. • • ~. • .. • A M A !::All. • K+ · A • • • <, Ll.A AA. O'l A ~A E A Na+xO,l 1 I I I C ...0 • ...,.,.0 • • • c .. .' Q) --. • u • • c • • 0 0 u 10 HCO- - .. • 3 0 • 0 ~ 0 0 • a 0 0 0 0 0 0 o dJ 0 CI- 0 0 o 0 0 .9.. • • • ~ • l- • •• • • • • • • • 50=4 • I I 10 102 103 104 Runoff (ft3/ sec)

FIGURE 3. The concentrations (after subtracting sea-salt contributions) of major cations and dissolved Si02 plotted against bicarbonate ion concentration in rivers and groundwaters of Hawaii and Kauai islands. Denudation Rates in Hawaii-e-Li 255 is main ly the organic matter accumulated in sea-salt contributions were estimated by the soils and some from fertilizers used on the ([M]/[Cn)seaw ater x [Cn sample, where [M] is sugarcane in the watershed, rather than the the concentration of any particular ion. The weathering of basaltic rocks. concentrations of Mg?", Ca2+, Na", and K + The relationships between the concentra are more or less linearly related to that of tions of bicarbonate and major cations as bicarbonate (slope of near 1 in Figure 3; the well as dissolved Si02 after subtracting the greater scatter for Na+ and K + is in part sea-salt contributions are shown in Figure 3 caused by large sea-salt corrections), even for water samples from the Wailuku River at though river and groundwater samples were Hilo (USGS Station No. 16213000), Waimea taken in different years and localities. The River at Waimea on the island of Kauai Si02 concentration does not increase as fast (USGS Station No. 16031000, October 1979 as [HC03"] (slope of less than 1 in Figure 3). September 1981), and groundwater samples Also, [Si02 ] in groundwaters tends to be around the islands of Hawaii and Kauai higher than in rivers of similar [HC03"] (October I972-September 1975). The lo (Figure 3). This difference may be partly ex cations where these water samples were plained by the uptake ofsilicon by the surface taken are shown in Figures 4 and 8. The vegetation (Fox et al. 1967).

r----~-----,.----_r_------, 10 I I

1 ++ + T C org. <, t 0> 1 0 0 E 0 + 00 0 0 0 0 0 0 c 0 0 0 0 0 °i • 0 0 Norg. 0 0 •• • • -~ •• •• ••• -c 1()' • •N inorg. Q) - .. • u C • 0 A A U A • A P AA A A A

A A A -2 10 10 102 103 104 Runoff ( ft3/ sec)

FIGURE 2. The concentrations of dissolved nutrients in the Wailuku River at Hilo as a function of runoff (October I979-September 1981). 256 PACIFIC SCIENCE, Volume 42, July/October 1988

100 I I I o Wailuku River at Hilo ~ l:;. Ground water in Hawaii / Waimea River near / ~ Ground water in Kauai /' • / .../ Waimea in Kauai / ~... / ...... / A A /~ ...... / /~ / / .z/£/ ... 10 - A Ao' ...... ~j+.'" ~/.1'"A • .II. A.JY ~~ ...... \~~. ... ./ 0 0 0/:· A .~~~ o 0,0· A A 9- /,~ A 8 ,. . (a++ /~o A Mg++ / • o / / 00 E o / o/QP Q. / d" Q. , r :~1Jf ~?O I I p. .--£~A ... '" c A ... 0 "AfSA ~ ... ~ A

0 ~;;t: A -~ Si02 -c .,J" • A ... Q) 10 - ,,;:s .0 t u 9-a- ... c A .-- 0 rt 0 "" A .. U ... 0 A ... A ... AA 0'1 ~A • 0/./e• " AA ~ ...... oC\:)~ /• .. ~ ~o . ... •• <- ...... 0 o •• 1 - 0/• - 0 • 0 0 0 0 / . 00. .. 10 80Qlo 0 / 0 .~'eo 0/ • 0 - o. • A ... e 0 • 'It.: /0 • 0 0 / 0 • 0• + K+ Na 0.1 I I I I 10 100 ... 10 100 HC03 (ppm)

Oahu AVE . 74,1JB.

Hawaii AVE, 63.'16.

Kaual AvE. 128,!Cl

FIGURE 4 . The concentrations of bicarbonate and silica (underlined) in the groundwaters of Hawaii, Oahu , and Kauai islands. 258 PACIFIC SCIENCE, Volume 42, July/October 1988

Except for coral reefs and cap rocks along The relative mobility ofelements during the the shorelines, carbonate rocks are rare on weathering ofparental rocks can be expressed the Hawaiian Islands. Therefore, the main by the partition ratio K d = concentration of source of bicarbonate in Hawaiian rivers an element in solid divided by that in solution and groundwaters must be carbonic acid after sea-salt correction. The higher the Kd , produced by microbial oxidation of organic the lower the mobility of an element. One matter in the soil columns and root respira example is shown in Table 1 for the average tion. The high partial pressure of CO 2 , and alkaline series basalts on the island of Hawaii thus carbonic acid, in soils is well docu (Basaltic Volcanism Study Project 1981) and mented (Buyanovsky and Wagner 1983, long the sea-salt corrected chemical composition of and Schappert 1972, Keller 1986, Yamaguchi Wailuku River water at Hilo (USGS Station et al. 1967). The direct incorporation ofatmo No . 16713000, March 1981, when the run spheric CO 2 into soil via rainfall at pH less off rate is similar to the long-term average). than 5 is negligible. The abundance oforganic According to the relative magnitude of Kd's matter, root system, and microbial activities in Table 1, the sequence of relative mobil in soils are, in turn, closely related to vegeta ity is Ca> K > Na ~ Mg > Ba ~ Si ~ P .» tion coverage and rainfall rate. The high Mn ~ Fe, which is similar to that given by [Mg2+] and the linear correlation between Colman (1982) based on the composition [Mg2+] and [HC03"] (Figure 3) also do not changes in weathering rings of basalts and indicate the dissolution of carbonate rocks andesites. One exception is that the mobility [no dolomite on islands; the atmospheric dust ofK obtained by Colman falls between Si and inputs ofcalcite and dolomite from the Asian Fe. As mentioned earlier, the concentration of continent are negligibly small (Ferguson silica in the Hawaiian groundwaters is always 1970)] as the main source of bicarbonate, higher than in rivers at a comparable [HC03"] Mg2+ , and Ca2+. The light c5 1 3 C values of (Figure 3). Therefore, the mobility of Si in dissolved inorganic carbon in both rivers and groundwater becomes very similar to that of groundwaters on Oahu [mostly within the Mg. -18 1974)] (Pco , range of ± 2%0 (Hufen further The partial pressure of CO 2 2 in at support this conclusion. mospheres), pH, and activity of bicarbonate

After the sea-salt correction, the positive (aHCO, ) are related by the following equation: charge of major cations in river and ground pH = -log K x K X P water samples is mostly balanced by bi C0 2 1 C0 2 carbonate anion (Figure 3). Therefore, the + 10gaHco, (1) hydrogen ion of carbonic acid should be the major chemical weathering agent that in TABLE I teracts with basaltic rocks to produce second THE PARTITION RATIOS (Kd)OF ELEMENTS BETWEEN ary minerals, dissolved cations, bicarbonate, ALKALINE SERIES BASALTS (BASALT VOLCANISM and silica in natural waters (e.g., Garrels and STUDY PROJECT 1981) AND THE WAILUK U RIVER Mackenzie 1971, Li 1972). The hydrogen W ATERS ONTHE I SLAND OF HAWAil ion contribution from rainwater with pH of BASALTS RIVER K d around 4.5 -5 in the Hawaiian Islands (Dugan (%) (mg /l) ( x 104 ml/g) and Ekern 1984, Eriksson 1957, Miller and Yoshinaga 1981) is relatively minor except in Si 23.7 4.1 5.8 active volcanic areas on the island of Hawaii Fe 8.8 0.12 73 where the pH of rainwater can be locally Mg 3.9 1.4 2.8 Ca 5.6 3.0 1.9 as low as 3-4 (Harding and Miller 1982). Na 3.1 l.l 2.8 Organic acids in soil do contribute some K 1.4 0.7 2.1 hydrogen ion, but they are eventually oxidized Mn 0.16 0.002 80 to carbonic acid by microorganisms during P 0.20 0.Q3 6.7 the percolation ofwater down the soil column Ba 0.04 0.007 5.7 AI 8.24 (Cronan and Aiken 1985). Denudation Rates in Hawaii -e-Li 259 where K = a co /(P x a 0) = 10- 1.47 island of Oahu consist mainly of gibbsite coz Hz ,co zH z - 6 .3 5 and K 1 = aH+ x aHCO , /aHzCO , = 10 at and hallo ysite/kaolinite in the top 25 ft and 25°C. The relationship is shown as solid lines smectite with variable amounts of pedogenic in Figure 5A [the values for Kcoz and K 1 are feldspar below 30 ft (Jones 1986). Sherman taken from Drever (1982)]. The observed pH (1952) also demonstrated that the composi and [HC0:3], which is approximately equal to tion of minerals in Hawaiian soils changes aHCO" for the water samples given in Figure 3 systematically from gibbsite and goethite to as well as samples from an upstream sta kaolinites, then smectites, as the average rain tion of the Waimea River (USGS Station No. fall decreases. I 60I 9000, October I 979-September 1981) are also plotted in Figure 5A . Except for a few CHEMICAL AND PHYSI CAL DENUDATION RATES data points, the Pcoz in rivers and ground waters is always much higher than the atmo The daily discharge rates of dissolved (Fd ) and (~ 3 5 spheric Pcoz 10- . atm) , again suggesting particulate (Fp ) matter by the river at any the continuous input of carbonic acid by sampling station are estimated by microbial oxidation of organic matter even in F = F; x ([Ca 2+ ] + [Mg2+] + [Na +] the subsurface groundwater reservoirs . The d + [K+] + [SiO ]) microbial activity in groundwater aquifers z F = x C is well documented (e.g., Hirsch and Rades p r; p Rohkohl 1983).The sources oforganic matter where F; = daily runoff of the river at the in the Hawaiian groundwater aquifers should station (usually given in cubic feet per second, be the dissolved and fine particulate or col but can be converted to liters per day by multi 3) loidal organic matter carried down by the per plying by 2447 x 10 ; Cp = river suspended colating waters. particle concentration (in milligrams per li If a parcel of natural water is in equilibrium ter); and [Mr s = concentrations (after sea with calcite, then salt corrections) of major dissolved species (in milligrams per liter). For the estimate of log (acaz+/aw ) = 10g(Ksp / K z) F , the [HC0:3] is excluded , since the source - log aHCO (2) d i for bicarbonate is the organic matter in soils 8 34 where Ksp = aCa2+ x aco z- = 10- . and rather than basaltic rocks , as mentioned ear K z = aw x aCO z- /aHCO i = 310- 10.33. The reia lier. The Fd and Fp data as a function of F; tionship given ih equation (2) is shown as a for the Wailuku River at Hilo are shown in solid line in Figure 5B. [The values for Ksp Figure 7 as an example. The values for Fp were and K z are from Drever (1982).]The observed obtained from daily E; and Cp data for the aCa2+ /a w and aHCO- for various samples (no period between October 1980 and September sea-salt correction;) are also plotted in the 198I. The Fd values were obtained from bi same figure, where it can be seen that the weekly chemical data for the period between waters are steadily approaching saturation October 1979 and September 198I. The Fd with respect to calcite from upstream to rate increases more or less linearly with F; up down stream and in groundwater reservoirs. to about 103ft3/sec. In other word s, the disso The plots of chemica l data from the same lution rate of basalts is directly proportional set of samples (no sea-salt corrections) on to runoff or rainfall in a watershed. At the various activity-activity phase diagrams higher Fw , however , the rate of change of Fd (taken from Drever 1982), as shown in Figure decreases gradually. In contrast, the Fp in 6, indicate that the stable secondary minerals crease is a power function ofFw , as also shown after weathering of basalts should be gibbs by Ekern (1976) and Doty et al. (1981). At ite in upstream, kaolinite in downstream, lower Fw , the Fd is dominant over Fp • and various smectites and/or feldspar in the The average daily discharge rates of dis groundwaters. This prediction is in agree solved matter (Ed) as obtained by calculating ment with the observation that the weathered (I /n) x L s; for 3 observations and by read materials from several 60-ft wells on the ing off the Fd at F; = (I /n) x LF; in the cali- 9 .' 5 / o Wailu ku River at Hila in Hawaii A Groundwater in Hawaii )( up stream } Waimea River ... Groundwater in KO ... down stream in Kauai • A • 0 • Calcite A ... saturationr ..,. 8 • .... • 4 • • ce fl.· .. ... • .~ AA • . ••• d>.o. • • • ... 6' ...~ .0.•• /j. /j. ~ \f • •• + o ~ ..; J: oellll. /j. .... ~ .o./j. 0 A./j. 0 0 .0. • <, • /j.A I • 0 0 + + 0 7 • 0 o· /j. 3 0- 0 ~ 0 o u • 0 0 /j. 0 , 0 n",' 0> ~C) 0 0 " • -oJ X x 0 x 6 0 x 2

x (A) (8)

x )C 5 )( -5 -4 -3 -2 )( -4 -3 -2 Log a HC0 Log 0HC03 3

F IGURE 5. (A ) Log aileD, plotted again st pH and (D) log aileD, plotted against log ae. >+lall + for rivers and groundwatcrs in Hawaii and Kauai. Denudation Rates in Hawaii-e-Li 261

16 + NJ: o 14 <, et. (J 12 o Ol ..3 10 Gibbsite

x 6 '---- -l...--'--'------'-- --'---...1..-----'-----'---' -5

8 I ~I 7 Albite 6 IN -" 10u 5 I t: ~I~ I g Ej 6 Ie <{I + +J: 4 J: I 0 I ~ <, I + + 0 5 Gibbsite I Z o"" 3 0 I I OJ Ol 4 o 0 I -' 2 -' I I 3 9 "01 Pyrophyll ite "~ Kaolinit e 2 *,,1 I " " O L..-_ --'-_-'----1-__'------L----'-_ --L--1-----' -5 -5

FIGURE 6. The log activity-log activity ratio plots of rivers and groundwaters from Hawaii and Kauai islands (the symbols are the same as in Figure 5), and the stability fields of various secondary minerals given by Drever (1982). bration curve (dashed line in Figure 7) are drainage area (A) upstream can be calculated very similar (as one would have expected, be by multiplying 365 x Ed /A . The CDR calcu cause Fd is more or less linearly correlated to lation results for three Hawaiian islands are F; except at high Fw ) . Therefore, if one can shown in Figure 8. establish the calibration curve for Fd versus F; In general , the CDR is higher on the wind at any hydrographic station, then the Ed can ward (northeast) than on the leeward side be roughly estimated by using Ew obtained for of each island. This observation is consistent long periods of time (at least more than 10 with the fact that the windward side of each years). Once Ed is known, the average chem island has higher rainfall, thus also generally ical denudation rate (CDR) of any river denser vegetation and higher carbonic acid 262 PACIFIC SCIENCE, Volume 42, July/October 1988

I I I -

Wailuku River at Hilo

-;::: 10 2 - - o -0 <,

...... Q) o - 0:::

Q) 0> L-o -C U

Ave rage water discharge 000 I ClCISl I 10-2 I I ' I I 10 10 2 103 10 4 Runoff ( ft 3 / sec)

FIGURE 7. The daily discharg e rates of dissolved chemical species and suspended particles as a function of river runoff in the Wailuku River at Hilo. CHEMICAL RIVER DENUDATION RATE : (1'lf!Sl.CAl.) P'S/al/VR

FIGURE 8. The chemical and physical (in parentheses) denudation rates of river drainages on three Hawaiian islands. 264 PACIFIC SCIENCE, Volume 42, July/October 1988

production than the leeward side. Compar recharge and discharge rates in the Hawaiian ison of the CDR on the windward sides of Islands should be equal. Therefore, the aver these three islands shows that it is lowest on age chemical denudation rate of an island by the island of Hawaii, probably reflecting the groundwaters can be calculated by the sum high surface area of relatively young basaltic mation of average concentrations of major rocks free of or low in vegetation and humus dissolved species in groundwater (with proper coverage on this island. sea-salt correction) multiplied by the island Only a few river stations provide daily rec wide groundwater recharge rate and divided ords on Cp for more than several years since by the total area of the island. The island wide 1972. For those stations, I chose the F; and groundwater recharge rate and the total area Cp data in a year when the annual water dis of each island are given by Macdonald et at. charge rate was close to its long-term average (1983). The chemical compositions ofground (more than 10 years) to calculate the average waters between 1972 and 1975 are summa daily discharge rate of suspended particles, rized in the USGS Water Resources Data (l j365)~)Cp Fp = x Fw ). The average phys 1975. Interestingly, the islandwide average ical denudation rate (PDR) of a watershed is CORs by groundwaters among the three is then defined by 365 x FpjA. The PDR results lands are very similar, and their magnitudes are shown in Figure 8 (numbers in paren are comparable to those for rivers (Table 2). theses). For the Waikele River on Oahu and Table 2 summarizes the CDR by ground the Waimea River on Kauai, however , there waters and the possible ranges of CDR and are no extensive daily data on Cpo Therefore, POR by rivers, along with the total denuda all the Fpdata from several years were plotted tion rates (chemical and physical) in three against F; in a diagram similar to Figure 7 to Hawaiian islands . A CDR of 1 mgjcm2 jyr obtain the calibration curve first. Then I again corresponds to a leaching rate of about chose the daily F; data from a year when the 0.2 mg Cajcm 2 j yr for river water and 0.13 mg annual water discharge rate was close to its Ca jcm 2 jyr for groundwater. On the other long-term average to obtain daily .Fp values hand, the Ca content in average Hawaiian from the calibration curve and then to calcu basaltic rocks is about 6.4 ± 0.5% (or 9 ± 1% late Fp and the POR. This procedure may CaO; Basaltic Volcanism Study Project 1981). have an uncertainty of50% because , as shown Therefore, from the total CDR (rivers + in Figure 7, the Fp is very sensitive to a few groundwater), one can estimate the mass of ~ large Fp values at the high end Fw • Doty et fresh basaltic rocks that would interact annu al. (1981) also summarized the Fp values over ally with waters to lose all their Ca and most a 5-year period on the island of Oahu, based of their Mg, Na, K, and Si02 contents in on the USGS data. The PDR data are too few dissolved form and their residual solid, prob to make any generalization except that the ably as river suspended particles. These results PDR of the Wailuku River watershed is again are also shown in Table 2. Interestingly, these the lowest among the three islands. estimates fall nicely within the ranges of the Assuming a steady state, the groundwater total denudation rates obtained from the "ob-

TABLE 2 SUMMARY OF DENUDATION RATES (mg/cm2/yr) INTHREE HAWAIIA NISLANDS

Hawaii Oahu Kauai

Chemical (C = R + G) 4.7-6.8 5.1- 11.9 4.9- 9.4 River (R) 0.5- 2.4 0.7-7.5 1.2-5.7 Groundwater (G) 4.2 4.4 3.7 Physical (P) 6 6-30 ?- 40 ± 20 - -- Tolal (P + C) 11-13 11-42 ?- 49 ± 20 Tot al calculated from Ca leaching rate ll-16 11-32 11- 26 Denudation Rates in Hawaii-LI 265 served " PDR and CDR. The finding is con 5. The relative mobility of elements during sistent with the suggestion that the formation the weathering of Hawaiian basaltic rocks processes of Hawaiian amphitheater-headed is roughly in the order ofCa > K > Na ~ valleys consist of repeated cycles of a soil Mg > Ba ~ Si ~ P » Mn ~ Fe. formation period through chemical weather 6. The discharge rate of dissolved matter ing and subsequent soil avalanching or mass is always higher than that of particulate wasting (Scott and Street 1976). matter in the Hawaiian rivers when the The Waimea Canyon on the island of river runoff is relatively low . The former Kauai is as deep as 1000 m. The average ero increases more or less linearly and the sion depth is about 500 m in the Waimea River latter exponentially with increasing river watershed since the eruption of Makaweli runoff. lava (3.5-4.05 x 106 yr ago ; Macdonald et al. 7. The chemical denudation rate by rivers is 1983). Therefore, the average total denuda always higher on the wet windward side tion rate of the Waimea watershed is about than the dry leeward side of the Hawaiian 0.13 ± 0.02 mm/yr or 34 ± 5 mg/cm2 /yr, Islands. which is in good agreement with that esti 8. The chemical denudation rates by ground mated from the present CDR and PDR (49 ± waters among the Hawaiian Islands are 20 mg/cm 2 /yr, Table 2), and is by chance very similar, and their magnitudes are equal to that of the Kaneohe Bay watershed comparable to those by rivers . given by Moberly (1963). 9. The total denudation rate ofthe Hawaiian The average chemical denudation rates Islands ranges from 11 to 50 mg/cm 2 /yr. of the North and South American, Asian, Moberly's estimate for Kaneohe Bay falls African, and European continents is 3 ± 1 nicely within the range. The island of mg/cm 2 /yr (Garrels and Mackenzie 1971), as Hawaii tends to ha ve the lowest rate. compared to 5-12 mg/cm2/yrin the Hawaiian Islands. The latter values are comparable to those for the carbonaceous rocks in southern ACKNOWLEDGMENTS China (8-12 mg/cm 2/yr; Li et al. 1984), where the karst topography is world famous. Critical reviews and helpful suggestions by K . Chave, P. Ekern, F. Mackenzie, R. Moberly, and S. Smith are acknowledged. CONCLUSIONS 1. The hydrogen ion of carbonic acid , which LITERATURE CITED is produced by microbial oxidation of organic matter, is the major chemical BASALTIC VOLCANISM STUDY PROJECT. 1981. weathering agent of basaltic rocks in the Basaltic volcanism on the terrestrial planets. Hawaiian Islands. Pergamon Press , New York. 2. After correcting for the sea-salt contribu BATES, T. S., and J. D . CUNE. 1985. The role tion, the charge of major cations in the of the ocean in a regional sulfur cycle. Hawaiian rivers and groundwater is bal J. Geophys. Res . 90: 9168-9172. anced mainly by bicarbonate anion. BUYANOVSKY, G. A., and G . H . WAGNER. 3. The majority of Hawaiian waters are 1983. Annual cycles ofcarbon dioxide level undersaturated with respect to calcite but in soil and air. Soil Sci. Soc. Amer. J. 47: steadily approaching saturation from up 1139-1145. stream to downstream and groundwater COLMAN, S. M. 1982. Chemical weathering reservoirs. of basalts and andesites: Evidence from 4. The stable or metastable alteration prod weathering rinds. USGS Prof. Pap. 1246. ucts of primary rock weathering are Washington, D. C. gibbsite upstream, kaolinite downstream, CRONAN, C. S., and G. R . AIKEN. 1985. and smectites and K-feldspar in ground Chemistry and transport of soluble humic water. substances in forested watersheds of the 266 PACIFIC SCIENCE, Volume 42, July/October 1988

Adirondack Park, New York, Geochim. water samples. Ph.D. Thesis. University of Cosmochim. Acta 49: 1697-1705. Hawaii, Honolulu. DoTY, R. D., H. B. WOOD, and R. A. JONES, R. C. 1986. Mineralogical analysi s MERRIAM. 1981 . Suspended sediment pro of exploratory well soil samples in the duction from forested watersheds on Oahu, Wahiawa Plateau. Dept. ofAgronomy and Hawaii. Water Res. Bull. 17: 399-405. Soil Science, University ofHawaii (internal DREYER, J. I. 1982. The geochemistry of report), Honolulu. natural waters . Prentice-Hall, Englewood JONG, E., and H. J. V. SCHAPPERT. 1972. Cliffs, N. J. Calculation of soil respiration and activity DUGAN, G. L., and P. C. EKERN. 1984. from CO 2 profiles in the soil. Soil Sci. Chemical constituents of rainfall at dif 113: 328-333. ferent locations on Oahu, Hawaii. Water KELLER, M. 1986. Emissions of N 2 0 , CH 4 , Resources Res. Center Techn. Rept. No . and CO 2 from tropical forest soils. J. 160. Honolulu. Geophys. Res. 91 : 11791-11802. EKERN, P. C. 1976. Turbidity and sediment LI, Y. H. 1972. Geochemical mass balance rating curves for streams on Oahu, Hawaii. among lithosphere, hydrosphere, and at Soil erosion: Prediction and control. Soil mosphere. Amer. J. Sci. 272: 119-137. Conserv. Soc. Amer. Ankeny, Iowa. LI, Y. H., H. TERAOKA, T. S. YANG, and J. S. ERIKSSON, E. 1957.The chemical composition CHEN. 1984. The elemental composition of of Hawaiian rainfall. Tellus 9: 509-520. suspended particles from the Yellow and FERGUSON, W. S. 1970. Atmospheric dusts Yantze rivers. Geochim. Cosmochim. Acta from the North Pacific-A short note on a 48: 1561-1564. long-range eolian transport. J. Geophys. MACDONALD, G. A., A. T. ABBOTT, and F. L. Res. 75: 1137-1139. PETERSEN. 1983. Volcanoes in the sea. 2d ed. Fox, R. L., J. A. SILVA, 0 . R. YOUNGE, D . L. Univ. Hawaii Press, Honolulu. PLUCKNETT, and G. D. SHERMAN. 1967.Soil MILLER, J. M. and A. Yoshinaga. 1981. The and plant silicon and silicate response by pH of Hawaiian precipitation-A prelim sugar cane . Soil Sci. Soc. Amer. Proc. 31: inary report. Geophys. Res. Lett. 8: 779 775-779 782. GARRELS, R. M., and F. T. MACKENZIE. 1971. MOBERLY, R., JR. 1963. Rate ofdenudation in Evolution of sedimentary rocks . Norton, Hawaii. J. Geol. 71: 371-375. New York. SCOTT, G. A. J., and J. M. STREET. 1976. The HARDING, D., and J. M. MILLER. 1982. The role of chemical weathering in the forma influence on rain chemistry ofthe Hawaiian tion of Hawaiian amphitheater-headed volcano Kilauea. J. Geophys. Res. 87: valleys. Zeitschrift fiir Geomorph. 20 : 171 1125-1230. 189. HASAN, S. M., R. L. Fox, and C. C. BOYD . SHERMAN, G. D. 1952. The titanium content 1970. Solubility and availability of sorbed of Hawaiian soils and its significance. Soil sulfate in Hawaiian soils. Soil Sci. Soc. Sci. Proc. 16: 15-18. Amer. Proc. 35: 897-901. SHERMAN, G. D. , and H. IKAWA. 1968. Soil HIRSCH, P., and E. RADES-RoHKOHL. 1983. sequences in the Hawaiian Islands. Pac. Microbial diversity in a groundwater aq Sci. 22 :458-464. uifer in northern Germany. In C. Nash, II, U. S. GEOLOGICAL SURVEY. Water resources and L. Underkofler, eds. Developments in data: Hawaii and other Pacific areas. An industrial microbiology. Vol. 24. Society nual report from 1975to 1984. Washington, for Industrial Microbiology, Arlington, D.C. Virginia. YAMAGUCHI, M., W. J. FLOCKER, and F. D. HUFEN, T. H. 1974. A geohydrological in HOWARD. 1967. Soil atmosphere as in vestigation of Honolulu's basal waters fluenced by temperature and moisture. Soil based on isotopic and chemical analysis of Sci. Soc. Amer. Proc. 31 : 164-167.