High exchangeable calcium concentrations in soils on Barro Colorado Island, Panama

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

High exchangeable calcium concentrations in soils on Barro Colorado Island, Panama

Tobias Messmer a, Helmut Elsenbeer b,c, Wolfgang Wilcke a,⁎a University of Berne, Geographic Institute, Hallerstrasse 12, 3012 Bern, Switzerlandb University of Potsdam, Institute of Earth and Environmental Sciences, Karl-Liebknecht-Strasse 24-25, 14476 Potsdam, Germanyc Smithsonian Tropical Research Institute, Balboa, Panama

a b s t r a c ta r t i c l e i n f o

Article history:Received 19 February 2013Received in revised form 28 October 2013Accepted 29 October 2013Available online 22 November 2013

Keywords:PanamaCation-exchange capacityExchangeable CaClay mineralogyWeathering indices

The soils on four lithologies (basaltic conglomerates, Bohio; Andesite; volcanoclastic sediments with basalticagglomerates, Caimito volcanic; foraminiferal limestone, Caimito marine) on Barro Colorado Island (BCI) havehigh exchangeable Ca concentrations and cation-exchange capacities (CEC) compared to other tropical soils onsimilar parent material. In the 0–10 cm layer of 24 mineral soils, pH values ranged from 5.7 (Caimito volcanic andAndesite) to 6.5 (Caimito marine), concentrations of exchangeable Ca from 134 mmolc kg

−1 (Caimito volcanic)to 585 mmolc kg−1 (Caimito marine), and cation exchange capacities from 317 mmolc kg−1 (Caimito volcanic)to 933 mmolc kg−1 (Caimitomarine). X-ray diffractometry of the fractionb2 μmrevealed that smectites dominatedthe clay mineral assemblage in soil except on Caimito volcanic, where kaolinite was the dominant clay mineral.Exchangeable Ca concentrations decreasedwith increasing soil depth except on Caimitomarine. Theweathering in-dices Chemical Index of Alteration (CIA), Plagioclase Index of Alteration (PIA) andWeathering Index of Parker (WIP)determined forfive soils on all geological formations, suggested that in contrast to expectation the topsoil (0–10 cm)appeared to be the least and the subsoil (50–70 cm) and saprolite (isomorphically weathered rock in the soilmatrix) the most weathered. Additionally, the weathering indices indicated depletion of base cations and enrich-ment of Al-(hydr)oxides throughout the soil profile. Tree species did not have an effect on soil properties. Impededleaching and the related occurrence of overland flow seem to be important in determining claymineralogy. Our re-sults suggest that (i) edaphic conditions favor the formation of smectites on most lithologies resulting in high CECand thus high retention capacity for Ca and (ii) that there is an external source such as dust or sea spray depositionsupplying Ca to the soils.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

It has long been accepted by soil scientists that the soils of the tropicsare as diverse as the landscapes in which they developed (Sanchez andLogan, 1992; Sollins et al., 1994). One of the misconceptions that persistis the presumed general infertility of tropical lowland soils (Sollins et al.,1994) on account of their advanced degree of weathering. Strongweathering results in the depletion of cationic nutrients, such as Ca, andin the prevalence of typical suites of low-activity clay minerals. Thismisconception, which Richter and Babbar (1991) traced back to theearly 19th century, tacitly but wrongly assumes that tropical lowlandsoils neither receive any atmospheric input from volcanic, oceanic,anthropogenic or dryland sources (Boy and Wilcke, 2008; Kurtz et al.,2001; Likens et al., 1998;Wiegand et al., 2005), nor experience any denu-dation. While the former assumption was safely put to rest (Muhs et al.,1990; Pett-Ridge et al., 2009; Prospero et al., 1981), the latter receivedoverdue scrutiny only recently (Bern et al., 2005; Porder and Chadwick,2009; Vitousek et al., 2003; Zimmermann et al., 2012).

Themacronutrient Ca in soil can have two sources: (i) minerals of theparent material like calcite or plagioclase (Plg) and (ii) deposition fromthe atmosphere resulting from external sources such as sea spray, volca-nic exhalations, desert dust, or anthropogenic emissions (Boy andWilcke, 2008; Kurtz et al., 2001; Likens et al., 1998; Wiegand et al.,2005). Central America is influenced by winds transporting Sahara dustover the Atlantic Ocean and depositing dust on Caribbean islands (Muhset al., 1990; Pett-Ridge et al., 2009; Prospero et al., 1981). To our knowl-edge, there is no reported evidence for dust input in Panama yet, althoughGolley et al. (1976) reported a high flux of 37 kg ha−1 yr−1 of Ca inthroughfall in the Darién region.

Studies of toposequences of tropical soils on basaltic parent materialrevealed that the occurrence of smectite and kaolinite is related totopographic positions and drainage conditions. Smectites were identifiedon footslope positions with poor drainage and kaolinite on topslopepositions with good drainage. On positions with intermediate drainage,interstratified kaolinite/smectite clay minerals occur with CEC of250–1100 mmolc kg−1, depending on the proportion of kaolinite in theclay mineral (Bühmann and Grubb, 1991; Herbillon et al., 1981; Kantorand Schwertmann, 1974; Righi et al., 1998, 1999; Vingiani et al., 2004).In many mature tropical soils, pH-depending variable charge of organicmatter is the major contributor to CEC (Sollins, 1987). However, Yavitt

Geoderma 217–218 (2014) 212–224

⁎ Corresponding author. Tel.: +41 316313896; fax: +41 316318512.E-mail address: [email protected] (W. Wilcke).

0016-7061/$ – see front matter © 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.geoderma.2013.10.021

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma


andWright (2002) measured a point of zero net charge (PZNC) of pH b2and showed that soils with a permanent charge dominate on BCI. This iscorroborated by Baillie et al. (2007) who concluded from field evidencethat smectites dominate in soils on Caimito marine and Bohio, and thatkaolinitic clays dominate in soils on Andesite and Caimito volcanic. Thereport of Baillie et al. (2007) is based on 484 auger observation and 30 de-tailed soil profile descriptions. Johnsson and Stallard (1989) identifiedsmectite as dominant clay mineral in stream sediments on Caimitomarine and Bohio and to minor extent on Andesite and Caimito volcanicby XRD measurements of the fine-earth fraction.

Individual tree species affect nutrient cycling and thus soil proper-ties in mixed forests in temperate climates because of different nutrientdemands, rooting depths and litterfall characteristics as shown byFujinuma et al. (2005), Dijkstra and Smits (2002) and Dijkstra et al.(2003). The fig species Ficus insipidaWilld. and the cashew Anacardiumexcelsum (Bertero & Balb. ex Kunth) Skeels on BCI show contrasting Caconcentrations in fruits. Fruits of the figs had an average concentrationof 11.9 ± 3.5 mg g−1 Ca in drymatter of pulp. On the other hand, fruitsof cashew had an average concentration of 0.80 ± 0.25 mg g−1 Ca indrymatter of pulp (L. Albrecht, M. Tschapka and E. K. V. Kalko, personalcommunication). Because these data suggest a different Ca demand, weexpected contrasting impacts of fig and cashew trees on the nutrientstatus in soils near these tree species.

On BCI, exchangeable Ca concentrations of around 300 mmolc kg−1

were reported and attributed to themarine facies of the Caimito forma-tion (Yavitt and Wright, 2002; Yavitt et al., 1992) and to frequentrejuvenation of soil material (Johnsson and Stallard, 1989). These rejuve-nation processes are likely either induced by tectonic uplift associatedwith high erosion rates (Nichols et al., 2005) or by a combination ofnear-surfaceflow inducedby shallow impermeable soil layers and unpro-tected soil surface because of rapid decomposition of nutrient-rich leaflitter and a sparse understory caused by closed canopies (Zimmermannet al., 2012).

To investigate the reasons for the high CEC and base metal concen-trations, our objectives were i) to evaluate two competing hypotheses

regarding the origin of soil Ca in this lowland humid tropical environ-ment — i.e. lithogenic vs. eolian, ii) to explain the mechanism of Caretention, and iii) to assess the role of recycling in maintaining highsoil Ca levels.

2. Material and methods

2.1. Study site

The study sites are located at latitude 9°08′–9°11′ N and longitude79°49′–79°52′ W on Barro Colorado Island (BCI) in the Gatún Lakewhich was created in 1914 when the Panama Canal was dammed(Fig. 1). The island has a size of 15 km2. The highest point of the islandis 168 m above sea level and 138 m above lake level. The climate is clas-sified as Tropical Monsoon [Am] in the Koeppen Climate Classificationwith amean annual precipitation (MAP) of 2600 mmand a pronounceddry season between January andApril with an average rainfall of around90 mm (Windsor, 1990). Mean annual temperature is about 27 °Cwitha diurnal range of 8–10 °C and a monthly deviation of 2 °C (Leigh,1999). Vegetation is classified as a TropicalMoist Forest in theHoldridgesystem (Holdridge and Budowski, 1956). Barro Colorado Island consistsof three geological formations of Oligocene to Miocene age. Two are ofsedimentary origin (the Bohio and Caimito formations), while thethird is an andesitic lavaflow capping the center of the island. The oldestformation is the Bohio formation dating back to the early Oligocene,which forms the northwestern part and stretches from north to souththrough the center of the island. Of the 300 m-thick formation, onlythe uppermost 125 mare exposed on BCI. This formation consistsmain-ly of coarse basaltic conglomerates, set in a sandy matrix ofvolcaniclastic origin. Clasts are spherically weathered and attain diame-ters of cm to meters; most clasts range from 5 to 10 cm (Johnsson andStallard, 1989; Woodring, 1958). The andesite lava flow is of earlyMiocene age and caps the central part of the island. It is a resistant,non-vesicular, phenocrystic andesite reaching a thickness of 85 m(Woodring, 1958). Themineral assemblagemainly contains plagioclase,

Fig. 1. Location of the study sites onBarro Colorado Island, Panama. F andAdenotefig (Ficus insipida) and cashew trees (Anacardium excelsum), respectively. The sites are numbered from1-12.

213T. Messmer et al. / Geoderma 217–218 (2014) 212–224


pyroxenes and magnetite (Johnsson and Stallard, 1989). The Caimitomarine facies is of Oligocene age and overlies the Bohio formation inthe southwestern portion and a small outcrop in the northeast. OnBCI, the lowermost 100 m of a total of 300 m are exposed (Woodring,1958). This formation consists mainly of foraminiferal limestone withabundant pelecypods (Johnsson and Stallard, 1989). It also contains alarge detrital component consisting of vitric volcaniclastic debris,plagioclase, and quartz (Johnsson and Stallard, 1989). However,Woodring (1958) describes this formation as consisting of tuffaceoussandstone and tuffaceous siltstone. According to Woodring (1958),minor constituents are foraminiferal limestone, conglomerates, tuff,and agglomerates. The Caimito volcanic formation resides in the south-eastern part of the island. It consists mainly of volcaniclastic sandstonesof altered volcanic lithic fragments and basaltic agglomerates (Johnssonand Stallard, 1989; Woodring, 1958). Large boulder fields are found onthis formation, where boulders are mostly between 5 and 20 cm in di-ameter and spherically weathered. Intrusive dikes and sills are commonon the Bohio and the Caimito volcanic formation. Their compositions arebasaltic to basaltic andesite but are difficult to identify because of theirlithological similarity to the host rock (Baillie et al., 2007). The mainstructural feature is the Lutz-Drayton fault system striking NNE–SSW.

The topography of BCI is mainly related to geology (Johnsson andStallard, 1989). The andesite cap forms a broad plateau in the centerof the island dipping from NE down to SW. The Caimito marine in thesouthwestern part and the Caimito volcanic to the east are character-ized by gentle slopes and broad drainage divides. In contrast, theBohio formation is deeply dissected with slopes of up to 30°.

Relying on field data without laboratory support, soils were tenta-tively classified in the WRB classification system (IUSS Working GroupWRB, 2006) by Baillie et al. (2007). Cambisols were described on eachof the geologic formations while Ferralsols were classified on andesite,Caimito volcanic, and Bohio, respectively. To a minor extent, the fieldsurvey suggested the presence of Alisols and Luvisols on Andesite andCaimito marine, respectively, and Alisols on Caimito volcanic.

2.2. Field sampling

To include the influence of different tree species on the Ca concen-trations and distributions in soil, we selected three individuals of figtrees (F. insipida, a Ca-rich tree species) and cashew trees (A. excelsum,a Ca-poor tree species), respectively, on each of the four geologic forma-tions. In the dry season 2011 (April toMay), soil profiles were hand-dugat b3 m distance from the stems, and sampled by depth increments(0–10, 10–20, 20–30, 50–70 cm). Texture was estimated in the field(FAO, 2006). Samples of unweathered parent material were collectedfor each of the soils at the surface assuming that chemical compositionof these surface boulders represents that of the parentmaterial. Soils ad-jacent to fig trees are designated with an F, soils near cashew trees withan A. The soils F1–F3 and A1–A3 were on Bohio, F4–F6 and A4–A6 onandesite, F7–F9 and A7–A9 on Caimito marine, and F10–F12 and A10–A12 on Caimito volcanic (Fig. 1). Bulk soil samples were dried at 40 °Cfor 48 h, visible roots hand-picked and samples sieved to b2 mm. Bulkdensity and water content were determined with 5 replicates of250 ml cores per depth collected from the walls of the soil pits next tothe fig trees and thought to be representative for the adjacent cashewtrees. After determination of bulk density, the five replicates per depthincrement were bulked and wet sieved with a mesh size of N2 mm todetermine the stone content gravimetrically. The stone content wasalso estimated volumetrically in the field.

2.3. Chemical analyses

Soil pH was measured in deionized water and 1 M KCl with a soil tosolution ratio of 1:2.5, respectively, using a glass electrode (WTW3310,and SenTix 41electrode). Concentrations of C andNwere determined bydry combustion using a vario El III (Elementar Analysensysteme GmbH,

Hanau, Germany).We did not expect the soils to contain carbonates be-cause of the low pHvalues (Baillie et al., 2007). Even on the Caimitoma-rine formation which originally contained carbonates, we did notobserve effervescence by addition of 10% HCl. Therefore, we assumedthat total carbon equals organic carbon.

Exchangeable cations (K+, Na+, Mg2+, Ca2+, and Al3+) were ex-tracted with an unbuffered 1 M NH4NO3 solution (soil to solutionratio 1:20) by shaking for 2 h at room temperature on an orbital shaker(modified DIN ISO 19730 [2008] method). Samples were centrifugedand the supernatant solutions were filtered through ashless filters(Munktell Grade 392). Thereafter, the remaining samples were saturat-ed with 1 M NH4OAc, buffered with acetic acid to pH 7 (soil:solutionratio 1:20) and shaken for 2 h. After centrifugation, supernatant solu-tion was discarded and the samples were washed three times with96% ethanol to remove excess NH4

+. Ammonium on the exchange siteswas extracted with 1 M KCl (soil:solution ratio 1:10) shaken for 2 h andmeasuredwith a continuous flow analyzer (Seal Analytica AA3HR) to de-termine potential CEC. Cations were analyzed with flame atomic absorp-tion spectrometry (ZEEnit 650 P, Analytik Jena). For each soil, onerandomly selected sample was analyzed in replicate. Sum of the chargeequivalents of K+, Na+, Mg2+, Ca2+, and Al3+ gave the effective cation-exchange capacity (ECEC), the NH4

+ concentration after KCl extractionthe potential cation-exchange capacity, the sumof the charge equivalentsof K+, Na+, Mg2+, Ca2+ divided by ECEC the base saturation (BS).

Total concentrations of major elements were determined withwavelength dispersive X-ray fluorescence spectrometry (XRF) using aPhillips PW 2400 on pressed powder pellets and using the UNIQUANTinterpretation software. The sum was normalized to 100%. Major ele-ments are given as oxideswith an error of 1%. The analysis was conduct-ed at the Department of Geosciences at the University of Fribourg,Switzerland. Loss on ignition (LOI) was determined by ignition of soiland rock samples in ceramic crucibles at 1050 °C for 2 h.

2.4. XRD-analysis

Five soils (F2, F6, F9, A9, and F11, Fig. 1), their corresponding isomor-phically weathered rocks (in the following designated as saprolite) andparent materials were selected for XRF-analysis. They were selected asrepresentatives for their respective lithology. On Caimito marine we se-lected two profiles (F9, A9) because of the obvious influence of colluvialmaterial beneath the fig trees. F9 is strongly influenced by erosion pro-cesses. This is not the case for A9, representing an almost undisturbedsoil development in the Lutz catchment. XRD analysis was performedon soil samples for identification and semi-quantification of secondaryminerals in the clay fraction. The fraction b2 μmwas separated and ori-ented samples on glass slides were prepared and analyzed on a PhilipsPW 3710. The fraction b2 μm was obtained after sedimentation ofbulk soil sample b2 mm in an Atterberg cylinder for 16 h at room tem-perature. The clay suspension was saturated with 1 M CaCl2 and twospecimens per sample were mounted with a pipette on glass slides.One sample was air dried, measured and subsequently ethylene-glycolsolvated for 24 h at 50 °C (EG). The second was heated to 550 °C for1 h. The samples were measured with a Philips PW 3710 operatingwith Cu Kα-radiation at 40 kV and 30 mA. Diffraction patternswere re-corded by step scanning from 2° to 40° 2θ, with a step size of 0.02° andcounting time for 1 s per step. Smectite was identified by its variablespacing of the 001 peak in air-dried state (1.2–1.5 nm), expansion to1.69–1.78 nm with EG solvation and its collapse to around 1.0 nmwhen heated to 550 °C for 1 h (Moore and Reynolds, 1997). Interstrat-ified kaolinite–smectite clay minerals were identified by the broad dif-fraction peaks in EG-solvated samples between 0.82 and 0.72 nm,according to Righi et al. (1999) and Vingiani et al. (2004). Another indi-cation of interstratified kaolinite/smectite is the shifting pattern of thepeaks from air-dried to EG-solvated samples from 0.72 nm of the 001peak and 0.357 nm of the 002 peak to 0.81 nm and 0.351 nm, respec-tively (Moore and Reynolds, 1997). Kaolinite was identified by its 001

214 T. Messmer et al. / Geoderma 217–218 (2014) 212–224


peak at 0.72 nm and its 002 peak at 0.35 in the air-dried samples, bytheir disappearance when heated to 550 °C, and by the lack of anyshift upon EG-solvation. Zeolites were identified with the help of 020peaks at 0.896 nm in the air-dried samples and Cristobalite by its 101peak at 0.404 nm in all samples (air dried, EG-solvated, and heated,Moore and Reynolds, 1997). We did not quantify clay mineralconcentrations because the method of sedimentation onto glass slidesis appropriate for the identification of clay minerals but not for theirquantification (Środoń, 2006). Therefore, our results are expressedin a semi-quantitative way. We measured the peak height abovebackground and calculated the relative abundance of the clay mineralsin the sample.

2.5. Weathering indices

In order to determine the depth distribution of weathering intensityin the selected soils (F2, F6, F9, A9, and F11), we applied three common-ly used weathering indices which were proven to be well applicable tolithologies with intermediate to basic chemistry (Duzgoren-Aydinet al., 2002; Price and Velbel, 2003), the Chemical Index of Alteration(CIA) (Nesbitt and Young, 1982), the Plagioclase Index of Alteration(PIA) (Fedo et al., 1995), and the Weathering Index of Parker (WIP)(Parker, 1970). To facilitate comparisons with other published resultsand because most Ca, and large parts of K and Na concentrationswere in exchangeable form, which is highly unusual, we calculated allweathering indices both for total metal concentrations and totalminus exchangeable metal concentrations, respectively.

2.5.1. Chemical Index of Alteration (CIA)According to Nesbitt and Young (1982), the degree of weathering

can be estimated by calculation of the Chemical Index of Alteration(CIA), based on molecular proportions (i.e. mass% of the oxide of anelement divided by molar weight of the oxide, Eq. (2))

CIA ¼ Al2O3= Al2O3 þ CaOþ K2OþNa2Oð Þ½ � � 100: ð2Þ

Assuming that Al is immobile, changes in CIA reflect changing pro-portions of feldspar and aluminum-rich secondary minerals in the soilprofile. During weathering, feldspars are dissolved by acid hydrolysisand their constituting cations Na, Mg, Ca, and K are leached. The moreimmobile elements Si andAl remain in the soil solution and recrystallizeto oxidic minerals. However, Si may be removed in acidic environmentswith high leaching rates which are common in the tropics duringdesilication, thus Al is the residual element. Therefore, low values ofCIA indicate little chemical alteration while high values point to anintensive alteration and leaching of themobile cations relative to the re-sidual Al during weathering and soil formation. CIA values for primaryminerals are usually lower than 50, between 70 and 85 for smectite,and around 100 for kaolinite, gibbsite and boehmite (Nesbitt andYoung, 1982).

2.5.2. Plagioclase Index of Alteration (PIA)The PIAwas introduced by Fedo et al. (1995). Plagioclase is themost

abundant mineral in the earth's crust and is highly susceptible to alter-ation and weathering. In basaltic to andesitic rocks, the plagioclasegroup ranging from albite (sodium feldspar) to anorthite (calcium feld-spar) is one of the major constituents. If minerals of the anorthite-endmember dominate in the basaltic–andesitic parentmaterial, plagioclaseis suggested to be themain supplier of Ca in environments inwhich pri-mary minerals are preserved in the solum. The PIA takes in account theelements in a molecular proportion occurring in plagioclase (Eq. (3))

PIA ¼ Al2O3= Al2O3 þ CaOþNa2Oð Þ½ � � 100: ð3Þ

As for the CIA, values for unweathered parentmaterial are around50or less, and the weathered substrates have values between 80 and 100

indicating minerals ranging from smectites to oxidic and kandicminerals.

2.5.3. Weathering Index of Parker (WIP)TheWIPwas introduced by Parker (1970) for silicate rocks and is ap-

plicable for rocks ranging from acid to basic chemistry (Duzgoren-Aydinet al., 2002; Price and Velbel, 2003). The index is based on the atomicproportions (i.e. mass% of the element divided by atomic weight ofthe element) of the base elements Ca, K, Mg and Na. These are assumedto be the most mobile major elements in humid and acidic environ-ments, in which silicate weathering is driven by hydrolysis and cationsare removed by leaching through the profile. The index is calculatedwith the individual mobilities of the elements, based on their bondstrengths with oxygen (Eq. (4))

WIP ¼ Na=0:35þMg=0:9þ K=0:25þ Ca=0:7½ � � 100: ð4Þ

The values for unweathered parent material depending on themineral assemblage are ~100, and for intensively weathered materialnear 0. The advantage of the WIP over the other indices is that it takesonlymobile cations into account and it does not assume immobility of Al.

2.5.4. Open-system mass balance of CaWe calculated gains and losses of Ca in soils compared to parent ma-

terial in the respective depth increments using the open-systemmass bal-ance function (Eq. (5)).

τ j;w ¼ ρw � CCa;w

� �.ρp � CCa;p

� �� εCa;w þ 1� �

–1: ð5Þ

Variables are bulk density of soils (ρw), parentmaterial (ρp) and totalCa concentrations in soil (CCa,w) and in parent material (CCa,p), respec-tively. A positive value of τCa,w indicates a gain of Ca compared to theparent material and a negative value of τCa,w indicates a loss of Ca, theτ value represents the mass fraction gained or lost. The open-systemmass balance also considers the change in density and volume duringweathering as the strain (εi,w, Brimhall and Dietrich, 1987; Chadwickand Brimhall, 1990, Eq. (6)).

εi;w ¼ ρp � CTi;p

� �.ρw � CTi;w

� �–1: ð6Þ

Additional variables to Eq. (5) are concentration of Ti considered asthe immobile reference element in parent material (CTi,p) and in soils(CTi,w).

2.6. Statistical evaluation

To achieve normal distribution of all the data sets we log-transformedall data prior to statistical analysis. This resulted in normality for all datasets according to the Lilliefors test. After confirming variance homogene-ity with the Levené test, we used one-way Analysis of Variance (ANOVA)followed by Tukey–Kramer honestly significant difference (HSD) Post-hoc test of the means to determine effects of geology, tree species, andsoil depth on ECEC, BS, exchangeable Ca concentrations and total elementconcentrations. For the analysis of ECEC, BS and exchangeable Ca concen-trationswe excluded soils F3 andA6because of their topographic positionon backslope and footslope influencing the vertical distribution ofchemical properties which in turn differed too strongly from the othertrees. For the analysis of total Ca concentrations we excluded soils F9and A9 because of an extraordinarily high Ca concentration. Significancewas set at p b 0.05. Statistical analyses were performed using JMP10(SAS Institute Inc.).



3. Results

3.1. Exchangeable Ca concentrations and CEC on BCI in a pan-tropicalcontext

The concentrations of exchangeable Ca and (E)CEC in surface soils onBCI are among the highest evermeasured in tropical soils on volcanic par-ent material (Table 1, Fig. 2). Similar to BCI, the lithology of Caribbean lo-cations ranges from volcanic origin with an intermediate compositionresulting from the subduction of the South America Plate under theCarribean Plate tomassive coral limestone. An exception isHawaii emerg-ing through a hot spot and whose material mainly originates from theastenosphere. The range of MAP in the Caribbean and Pacific region isbetween 1400 mm on Barbados (Lathwell, 1974) and 4000 mm inCosta Rica (Sollins et al., 1994), and between 1220 and 2500 mm onHawaii (Porder and Chadwick, 2009) to 5000 mm in Malaysia (Proctor,1983), respectively. In regions with MAP N1400 mm, most mature soilsare deeplyweathered, have lowbase cation concentrations, lowCa reten-tion ability because of the lack of smectites and the dominance ofAl-(hydr)oxides like gibbsite in the soils. Mineral assemblages in soilson Pacific Islands and the Caribbean depend on content of volcanic ash

and glasses which transform to poorly crystalline secondary mineralssuch as allophane (Sollins et al., 1994). We also included soils of the Am-azon region into our comparison because they still dominate our percep-tion of soil fertility in the tropics. The average concentration ofexchangeable Ca in surface soils (0–10 cm) of BCI is 4.6-fold of the Carib-bean soils, 1.7-fold the concentration of Hawaiian soils, and 4.9-fold theconcentration in the other soils. When taking the average of all the soilsbut those on BCI, the average concentration is 3.2-fold higher in surfacesoils on BCI than in the other soils, of which all but the Amazonian soilsdeveloped fromsimilar lithologies and are located near oceans. The differ-ences between soils on BCI and the soils used for comparison are evenmore pronounced for the subsoils. The average exchangeable Ca concen-tration on BCI at the 50–70 cm soil depth is 7-fold higher than that in theCaribbean, 1.4-fold higher than that on Hawaii, and 5.2-fold higher thanthat in the other soils. The BCI subsoils have 3.4-fold higher Ca concentra-tions than the mean of all other subsoils.

The average CEC in the surface soils on BCI (641 mmolc kg−1) is 1.7-fold the CEC found in the Caribbean, 2.2-fold that on Hawaii, and 2.2-fold that in other soils. The differences in CEC between BCI and othersoils are again more pronounced for the subsoil. The average on BCI(606 mmolc kg−1) is 3.3-fold that in the Caribbean, 2.7-fold the

Table 1Results of the literature survey of (effective) cation exchange capacities in tropical soils. MAP = mean annual precipitation, ECEC = effective cation exchange capacity, CEC = cationexchange capacity.

Sites MAP[mm]

Parent material pH [H2O] ECEC [mmolc kg−1] CEC [mmolc kg−1] References

Surface Subsurface Surface Subsurface Surface Subsurface

Barro Colorado Island This studyBohio 2600 Basaltic conglomerates 6.1 4.9 500 420 791 734Andesite 2600 Andesitic lava flow 5.7 5.0 337 296 522 517Caimito volcanic 2600 Volcanoclastic sandstone 5.7 5.2 176 57 317 229Caimito marine 2600 Foraminiferal limestone 6.5 6.2 666 656 933 943

La Selva, Costa Rica Sollins et al. (1994)Matabuey 4000 Lava flow 5.0 4.8 45.5 27.8 280 201Helechal 4000 Lava flow/alluvial deposits 4.3 4.4 36.8 22.0 293 193

Puerto Rico Beinroth (1982)Catalina 1800 Basaltic andesite lava 5.1 5.0 – – 162 77Bayamon 1800 Limestone, blanket deposits 5.6 5.9 – – 151 86Carreras 1800 Siltstone, volcanic sandstone 5.0 4.3 – – 190 139Picacho 4500 Quartzdiorite 4.8 5.2 – – 100 66

Luquillo, Puerto Rico Silver et al. (1994)Ridge 3500 Volcanoclastic 4.7 4.8 85 53 – –

Slope 3500 Sandstone 4.2 5.0 100 70 – –

Jamaica Tanner (1977)Mor Ridge 2500 Volcanics 3.0 4.5 113 2 1850 510Mull Ridge 2500 Granodiorite 3.7 4.5 141 2.0 190 10Wet Slope 2500 Granodiorite 4.0 4.5 110 70 290 200John Crow 2500 Limestone 4.8 4.8 212 181 830 640 Grubb and Tanner (1976)

Barbados Lathwell (1974)St. Johns 1400 Limestone 7.3 7.7 – – 353 179

AmazoniaSan Carlos 3560 Sandstone 4.5 5.6 – – 147 54 Jordan (1982)Site no.14, Pará 1750 Sediment 4.9 26.5 – – – Buschbacher et al. (1988)Site no.15, Pará 1750 Sediment 4.4 19.1 – – –

Hawaii Porder and Chadwick (2009)10 kyr-flow 1220 Basalt lava flow 5.7 6.9 492 586 – –

10-kyr flow 2400 Basalt lava flow 5.9 5.6 150 41 – –





Thailand, Khon Buria 1300 Basalt 6.2 6.2 159 106 188 123 Thanachit et al. (2006)Semporna Peninsula, Sabahb Burnham (1987)Andesitic agglomerate 2250 Andesitic agglomerates – 331 – 331 – –

Basalt 2250 Basalt – 146 – 146 – –

Sarawak, Malaysia Proctor (1983)Alluvial forest 5000 Alluvial deposits 4.4 4.8 – – 380 150Dipterocarp forest 5000 Colluvial deposits 4.1 4.7 – – 370 120

North Queensland, Australia Brasell et al. (1980)Site no.1 1400 Basalt flow and pyroclasts 6.6 6.1 343 174 431 313Site no.2 2500 5.1 4.8 85 16 276 181

a Sampling depth: 11-20 cm.b Sampling depth: 10-40 cm.



Table 2Vertical distribution of selectedmean properties of the study soils (standard errors in brackets). The upper case letters indicate differences among geological formations, lower case lettersamong depth increments on the corresponding formation (Tukey–Kramer HSD).

Geology n Deptha pH BD Corg Exchangeable Ca ECEC CEC BS

[cm] [KCl] [g cm−3] [%] mmolc kg−1 [%]

Bohio 6 0–10 5.2 0.82 5.21 366 A,a 500 A,a 791 A,a 100 A,a(0.3) (0.05) (1.98) (147) (160) (203) (0)

6 10–20 4.4 0.95 1.83 275 A,ab 415 A,a 646 A,a 98 A,a(0.7) (0.09) (0.51) (110) (134) (210) (2)

6 20–30 4.2 0.92 1.38 278 A,ab 440 A,a 728 A,a 94 A,a(0.7) (0.06) (0.69) (153) (161) (200) (6)

5b 50–70 3.7 0.99 0.67 191 A,b 352 A,a 641 A,a 78 A,a(0.7) (0.07) (0.18) (177) (163) (210) (17)

Andesite 6 0–10 4.9 1.02 3.72 204 B,a 273 B,a 421 B,a 100 B,a(0.6) (0.10) (0.97) (103) (113) (132) (0)

6 10–20 3.9 1.12 1.51 151 B,ab 249 B,a 421 B,a 90 B,a(0.6) (0.06) (0.24) (98) (117) (118) (5)

6 20–30 3.7 1.17 1.06 125 B,ab 269 B,a 486 B,a 66 B,ab(0.3) (0.02) (0.16) (106) (99) (111) (11)

6 50–70 3.6 1.17 0.45 102 B,b 296 B,a 517 B,a 48 B,b(0.1) (0.08) (0.16) (137) (135) (179) (17)

Caimito marine 6 0–10 5.6 0.69 4.88 585 C,a 666 C,a 933 C,a 100 A,a(0.3) (1.57) (60) (57) (49) (0)

6 10–20 5.3 0.86 2.33 573 C,a 650 C,a 946 C,a 100 A,a(0.3) (0.96) (65) (57) (113) (0)

6 20–30 5.1 0.82 1.40 566 C,a 641 C,a 909 C,a 100 A,a(0.3) (0.58) (58) (50) (59) (0)

6 50–70 4.5 0.90 0.73 575 C,a 656 C,a 943 C,a 100 A,a(0.6) (0.29) (58) (42) (81) (0)

Caimito volcanic 6 0–10 5.1 0.96 5.32 134 D,a 176 D,a 317 D,a 100 A,a(0.3) (0.05) (1.1) (37) (43) (52) (0)

6 10–20 5.0 1.10 3.06 79 D,ab 108 D,b 248 D,ab 100 A,a(0.4) (0.09) (0.68) (26) (28) (44) (0)

6 20–30 4.7 1.17 1.88 46 D,b 71 D,bc 258 D,ab 98 A,a(0.5) (0.09) (0.47) (12) (10) (35) (2)

6 50–70 4.6 1.19 0.58 27 D,c 57 D,c 229 D,b 91 A,a(0.4) (0.03) (0.18) (15) (26) (43) (6)

a : Bohio: n = 1 (10–20 and 20–30 cm); n = 2 (50–70 cm); Andesite: n = 3 (10–20 cm); n = 4 (20–30 and 50–70 cm); Caimito volcanic: n = 1 (20–30 cm); n = 2 (50–70 cm);n.d. = not detected.

b n = 5 because the soil A3 is shallow and was only sampled to the 20–30 cm increment. Bulk density is of the fine earth fraction (b2 mm).

Fig. 2. Relationship ofmean annual precipitationwith exchangeable Ca concentrations. Trend lines were calculatedwithout samples from BCI, John Crow in Jamaica and the Sabah sampleof andesitic agglomerate because we classified them as outliers.Data from: La Selva, Costa Rica (Sollins et al., 1994); Puerto Rico (Beinroth, 1982); Luquillo, Puerto Rico (Silver et al., 1994); Jamaica (Grubb and Tanner, 1976; Tanner, 1977); Barbados(Lathwell, 1974); San Carlos (Jordan, 1982); Pará, Brasilia (Buschbacher et al., 1988); Hawaii (Porder and Chadwick, 2009); Thailand (Thanachit et al., 2006); Semporna Peninsula, Sabah(Burnham, 1987); Sarawak, Malaysia (Proctor, 1983); North Queensland, Australia (Brasell et al., 1980).



concentration onHawaii, and 3.9-fold that in the other soils. Overall, theCEC in the subsurface on BCI is 3.2-fold higher than on average of allother soils.

3.2. Vertical distribution of CEC and exchangeable Ca

There were no significant differences in concentrations of mean ex-changeable Ca, ECEC, and CEC between the soils under figs and undercashews. Therefore we pooled all six soils on each geological formationfor statistical analysis (Table 2). The partly high standard errors inTable 2 represent the spatial heterogeneity of the soils (likelymainly driv-en by their topographic positions) and were not tree species-specific. Ex-changeable Ca concentrations, ECEC and CEC differed partly significantlyamong the geological formations. The highest concentrations of Ca oc-curred in soils on Caimitomarine. ECEC and CECwere highest on Caimitomarine and decreased in the order Bohio N Andesite N Caimito volcanic.The lowest Ca concentrations, ECEC and CEC consistently occurred onCaimito volcanic.

Exchangeable Ca concentrationsdecreased significantlywithdepthonBohio, Andesite and Caimito volcanic, while on Caimito marine Ca con-centrations did not vary significantlywith depth (Table 2).Mass balances,

using Ti as immobile index element, revealed τCa,Ti values, describing theloss of Ca relative to Ti because of weathering (Brimhall and Dietrich,1987; Chadwick and Brimhall, 1990), of −0.73 to −0.98 on Bohio (F2),−0.90 to −0.99 on andesite (F6), −0.95 to −0.99 on Caimito marine(F9 and A9) and−0.96 to−0.97 on Caimito volcanic (F11).With respectto ECEC andCEC, the only significant depth trendwasdetected onCaimitovolcanic where ECEC and CEC were highest in the topsoil. Topsoils (0–10 cm) on all geological formationswere saturated to 100%with base cat-ions and this was also true for all depth sections on Caimito marine. OnBohio and Caimito volcanic, BS decreased slightly but not significantlywith depth, on Caimito marine, BS remained constant and on AndesiteBS decreased significantly. Consistent with the decrease in exchangeableCa concentrations and partly BS with increasing depth, pH values de-creased with depth by 1.2 pH units on Bohio, by 0.8 on andesite, by 0.3on Caimito marine, and by 0.5 on Caimito volcanic, respectively. In soils,Ca is mainly distributed between primary minerals and the cation ex-change complex. On all geologic formations, more than half up to over80% of Ca was stored in exchangeable form (Fig. 3). The contribution ofexchangeableK andNa to total concentrationwas smaller, but still around40 and 30%, respectively. The largest part of Mg was not exchangeableand thus must form structural part of minerals.

Fig. 3.Contributions of exchangeable basemetals to total element concentrations in %. Datawere aggregated over the four sampled depths. Error bars are standarddeviations. The soil F2 isdeveloped from the Bohio geological formation, F6 from Andesite, F9 and A9 from Caimito marine, and F11 from Caimito volcanic.

Table 3Elemental composition (as determined with X-ray Fluorescence analysis) and loss on ignition (LOI) of the selected soils.

Soil SiO2 Al2O3 TiO2 Fe2O3 MnO MgOwt. [%]

CaO Na2O K2O P2O5 LOI

BohioF2 0–10 36.0 19.9 0.46 13.7 0.28 1.09 0.99 0.03 0.11 0.16 26.7

Saprolite 27.7 36.4 0.69 10.8 0.82 0.56 0.03 0.04 0.11 0.30 21.8Rock 44.5 23.5 0.43 12.0 0.19 3.1 3.44 0.90 0.98 0.17 10.3

AndesiteF6 0–10 44.2 13.4 0.55 18.7 0.09 0.82 0.84 0.03 0.14 0.13 20.4

Saprolite 64.4 9.8 0.27 19.8 0.09 0.18 0.11 0.04 0.06 0.12 4.8Rock 50.0 18.4 0.54 11.6 0.18 3.91 8.6 2.56 0.94 0.44 2.2

Caimito marineMean of F9 and A9 0–10 44 15 0.39 9.2 0.25 2.1 2.4 0.05 0.21 0.20 25

Saprolite 54 17 0.44 8.1 0.21 2.4 2.0 0.07 0.15 0.21 15Rock 25 7.4 0.22 5.6 0.08 1.8 31 0.09 0.23 0.15 30

Caimito volcanicF11 0–10 25.7 20.7 0.67 17.2 0.29 0.33 0.81 0.03 0.08 0.20 33.2

50–70a 30.0 25.0 0.77 20.2 0.22 0.26 0.09 0.02 0.07 0.09 28.1Rock 46.4 21.8 0.59 12.1 0.21 3.09 7.97 2.11 0.90 0.46 3.8

a Saprolite was not accessible.



3.3. Weathering indices

On all lithologies, the concentrations of Ca, K, andMgwere higher inthe topsoil than in the saprolite or subsoil (except forMg in Caimitoma-rine, Table 3). For the other elements, the relationship of topsoil tosaprolite/subsoil concentrations was inconsistent with a tendencytowards amore frequent depletion than accumulation in the topsoil rel-ative to the saprolite/subsoil. Among the base metals, Ca was outstand-ing in that its accumulation in the topsoil was the highest of all elementsin three of the four lithologies ranging 760–3300% of the saprolite/subsoil concentration. The only exception was Caimito marine wherethe topsoil concentration of Ca was 120% while the topsoil concentra-tion of K was 140% of that of the saprolite. The weathering indices CIA,PIA, and WIP showed all similar vertical trends with a sharp differenti-ation between parent material on one hand and saprolite and soil onthe other hand (Fig. 4). The parent materials of andesite and Caimitovolcanic had CIA and PIA values around 50, respectively, which confirmthat primaryminerals were themain rock constituents. The CIA and PIAvalues of the parent material on Bohio were slightly N70 indicating thepresence of smectite which is likely inherited from its sedimentary ori-gin. This assumption is supported by the fact that there were broadpeaks at 1.45 and 4.48 nm in an own unpublished XRD analysis ofunoriented powdered rock material from Bohio. The parent materialof the Caimito marine formation revealed CIA and PIA values of slightlyN11 pointing to the dominance of calcite and the occurrence of Al-bearing minerals as only minor constituents.

The saprolites on Bohio and andesite had CIA and PIA values N96 in-dicating the dominance of Al-(hydr)oxides in line with their advancedstate of weathering. The saprolite on Caimito marine with CIA and PIAvalues between 78 and 84 was dominated by base cations and thusstill contained weatherable minerals.

The CIA and PIA index values indicate consistently that the topsoil(0–10 cm depth) is apparently less weathered than subsoils on allfour lithologies. On Bohio and andesite, the values of CIA and PIA consis-tently increased fromuppermineral soil (0–10 cm) to the deepest layer(50–70 cm). On Caimito volcanic, the values of CIA and PIA in 0–10 cmincrements were slightly higher than on Bohio and Andesite indicatingthat oxidic minerals weremore common. The 50–70 cm increments onBohio, Andesite and Caimito volcanic had CIA and PIA values around 99indicating almost complete depletion of base cations and enrichment ofAl-bearing minerals. The least weathered soil occurred on Caimito ma-rine where CIA and PIA values ranged between 70 and 84, indicatingthat smectites dominated the mineralogy.

The WIP values, which considered the mobile cations, decreasedwith depth in soils on Bohio, Andesite and Caimito volcanic. Caimito vol-canic had the lowest WIP, suggesting that the soil F11 was the mostweathered and had lost most of the primary base cations. Again, thevalues indicated that the saprolitewasmostweathered. On Caimitoma-rine, soil F9 seemed to be less weathered in the 0–20 cm incrementsthan soil A9, supporting the assumption that it contained some colluvialmaterial which rejuvenated soil F9.

The difference in weathering indices calculated for non-exchangeablemetals only and for total concentrations (including exchangeablemetals)decreased with increasing soil depth on Bohio, andesite and Caimitovolcanic (Inlet graph of Fig. 4). On Bohio, the differences were the largestbecause of the particularly high contribution of exchangeable metals tototal metal concentrations.

3.4. Clay mineralogy

The mineral composition of the clay fraction (Table 4) of the five se-lected soils (F2, F6, F9, F11, and A9) ranged from smectitic on Caimitomarine (F9, A9) over mixed smectitic–kaolinitic on Bohio and andesite(F2 and F6), to kaolinitic on Caimito volcanic (F11). The soil F2 onBohio was dominated by smectite decreasing slightly from 0–10 cm to20–30 cm and increasing sharply in the 50–60 cm increment,

respectively. In the depth increments of 0–10 and 10–20 cm, addition-ally interstratified kaolinite–smectite clay minerals were identified.Applying the method proposed by Moore and Reynolds (1997), the ka-olinite accounted for 30 to 40% in the randomly interstratified kaolinite/smectite. Table 5 summarizes further field observations concerningmineralogical properties of the study soils.

The soil F6 on andesite was dominated by smectite, with a ratio ofsmectite to kaolinite of approx. 3:1. The smectite content was similar

00 20

Chemical Index of Alteration

So

il d

epth

[cm

]S

oil

dep

th [

cm]

So

il d

epth

[cm

]

Plagioclase Index of Alteration

40 60 80 100

0 20 40 60 80 100

Weathering Index of Parker

0 20 40 60

5

15

25

soil

dept

h [c

m]

35

45

60

5

15

25

soil

dept

h [c

m]

35

45

60

5

15

25

soil

dept

h [c

m]

35

45

60

01020

01020

0 -5 -10

80 100

10

20

30

40

50

60

70

SAPPM

010203040506070

SAP PM

010203040506070

SAP PM

F2 F6 F9 A9 F11

Bohio Andesite C. marine C. volcanic

A)

B)

C)

Fig. 4. Vertical distribution of (A.) Chemical Index of Alteration (Nesbitt and Young, 1982)(B.) Plagioclase Index of Alteration (Fedo et al., 1995), and (C.) theWeathering Index of Parker(1970) in five soils under Ficus insipida Willd. (F) and Anacardium excelsum L. (A) calculatedfor non-exchangeablemetal concentrations only. The soil F2 is developed from the Bohio geo-logical formation, F6 from Andesite, F9 and A9 from Caimito marine, and F11 from Caimitovolcanic. The inlet graphs show the mean differences in weathering indices calculated fornon-exchangeable metals and total metal concentrations (including exchangeable metals).White bars refer to the soils F2, F6, and F11, and the gray bars to the soils F9 andA9onCaimitomarine. Error bars are standard deviations.



in all depth increments. On Caimito marine, smectite either clearlydominated (in soil F9) or was the sole clay-size constituent (in soilA9). In soil F9 we identified zeolites, which were abundant in the0–10 cm depth but only occurred in traces in the 20–30 cm depthincrement. The presence of zeolites and cristobalite indicated the in-fluence of volcanic ash material and its hydrothermal alterationproducts (Brindley and Brown, 1980; Yerima et al., 1985). The claymineral assemblage in the soil F11 on Caimito volcanic consisted ofkaolinite. Only in the 50–60 cm increments, small amounts of

smectite were present. In 50–60 cm increment, surprisingly, the001 peak at ca. 0.99 nm which is typical for illite appeared althoughthe presence of illite in the parent material as indicated by the XRDanalysis of unoriented bulk samples (own unpublished data) wasunlikely. Illite is supposed to occur in themuscovite–kaolinite trajec-tory as intermediate mineral because of transformation of thephyllosilicate muscovite by dissolution of potassium (Righi et al.,1998). However, the occurrence of muscovite is unlikely in basalticto andesitic rocks.

Table 5Field morphological observations of the five selected soils F2, F6, F9, A9, and F11.

Soil Depth[cm]

Topographicposition

Texturea Stone contentb

[wt %]Colluvial materialc Hydromorphyd Drainagee

BohioF2 0–10 Ridge SL 4 intermediate

10–20 CL 420–30 CL 5 impeded50–70 C 6

AndesiteF6 0–10 Plateau SL 7 free

10–20 SL 620–30 SiCL 29 x impeded50–70 C 26 x

Caimito marineF9 0–10 Valley bottom CL 4 x free

10–20 CL 4 x20–30 CL 4 x50–70 C 22 x impeded

A9 0–10 Ridge SL 4 free10–20 SiCL 1020–30 SiCL 1050–70 SC 10

Caimito volcanicF11 0–10 Slope Toe L 40 free

10–20 L 3020–30 CL 2050–70 CL 20

a Texture notation follows the Guidelines for Soil Description (FAO, 2006).b Stone content was determined from the bulked samples of five cores, each with a volume of 250 ml.c Deposition of colluvial material is indicated by x.d Occurrence of hydromorphic features such as mottling is indicated by x.e Assessment by observation in the soil profiles.

Table 4Mineralogical composition of the clay fraction (b2 μm) of the soils determined by XRD. Sm = smectite, Kaol = kaolinite, K/S = interstratified kaolinite/smectite clay mineral,Zeo = zeolite, Cri = cristobalite, Qz = quartz, Ana = anatase, Gib = gibbsite, Goe = goethite, Lepid = lepidocrocite, brackets = traces were determined, n.d. = not determined.

Soil Depth Sm Kaol K/S Zeo Illite Cri Qz Ana Gib Goe Lepid[cm] - % -

Bohio F2 0–10 72 n.d. 28 +++ +F2 10–20 71 n.d. 29 +++ +

20–30 70 30 n.d. +++ +50–60 77 23 n.d. + + +

Andesite 0–10 73 27 n.d. + + + +F6 10–20 75 25 n.d. + + + +

20–30 76 24 n.d. + + + +50–60 74 26 n.d. + + +

Caimito marine 0–10 95 6 n.d. ++F9 10–20 94 6 n.d. + (+)

20–30 94 6 n.d. (+) (+)50–60 83 17 n.d. (+) (+) +

A9 0–10 100 n.d. n.d.10–20 100 n.d. n.d.20–30 100 n.d. n.d.50–60 100 n.d. n.d.

Caimito volcanic 0–10 n.d. 100 n.d. + +F11 10–20 n.d. 100 n.d. + +

20–30 n.d. 100 n.d. + +50–60 5 95 n.d. + + + +



4. Discussion

4.1. Exchangeable Ca concentrations and cation exchange capacities in apan-tropical context

Exchangeable Ca concentrations and CEC in soils of BCI are at theupper end of the range reported for tropical soils (Table 1, Fig. 2).These high exchangeable Ca concentrations and CEC values likely pre-clude the classification of Ferralsols and Alisols as previously done basedon a field survey (Baillie et al., 2007), because Ferralsols and Alisols arecharacterized by low base saturations and the dominance of low-activity clays. Exchangeable Ca concentrations and CEC in soils on BCIare comparable to those on Hawaii, North Queensland and Barbadoswith a MAP between 1200 and 1400 mm, respectively, despite a muchhigher MAP (2600 mm), which might suggest pronounced leaching ofbase cations. The concentration of exchangeable Ca on BCI is comparablewith the concentrations in soils at John Crow Mountains in Jamaica(Tanner, 1977) where dust inputs were already shown by Muhs et al.(2007). The expectation is that at a MAP between 1400 and 2000 mm,the soil solution containsmostly Si, Al, and Fe cations, favoring the precip-itation of kaolinite and hence a low exchange capacity (Chadwick et al.,2003; Fisher and Ryan, 2006; Porder and Chadwick, 2009).

On the other hand, the Ca deposition rate via throughfall of 30–37 kg ha−1 yr−1 in Panama (Cavelier et al., 1997; Golley et al., 1976) isconsiderably higher than that in Puerto Rico (20 kg ha−1 yr1, McDowell,1998; Pett-Ridge et al., 2009), Jamaica (11 kg ha−1 yr−1, Hafkenscheid,2000) and Costa Rica (12 and 24 kg ha−1 yr−1, respectively, Clarket al., 1998; Hölscher et al., 2003). However, high Ca fluxes are not onlythe result of external inputs but may also reflect a pronounced recyclingon the Ca-rich soils of BCI (Vitousek and Sanford, 1986). The assessmentof net Ca inputs from the atmosphere to soils on BCI would require themeasurement of bulk deposition and an estimate of dry deposition.

4.2. Clay minerals in the mineral assemblage — a reason for high CEC

Our observation of high CEC and the dominance of permanent-chargecomponents identified by a PZNC of pH b2 in the soils on BCI (Yavitt andWright, 2002) led to the assumption that clayminerals are themain con-tributors to the charge characteristics. Therefore, we explored the claymineralogical composition of the study soils. Smectites with their highpermanent charge belong to the main minerals contributing to CEC(Yavitt andWright, 2002). This finding is confirmed by the identificationof smectites in sediments of streams draining the deeply dissected Bohioand Caimito marine formations (Johnsson and Stallard, 1989). On thegentle terrain of andesite and Caimito volcanic, claymineralogy in streamsediments ismore dominated by kaolinite although smectite is present athigh levels (Johnsson and Stallard, 1989).

The occurrence of smectites inmoist tropical environments is attrib-uted to pedological conditions characterized by pH values N5 and highconcentrations of base cations in soil solution, or to poor drainage(Fisher and Ryan, 2006). These conditions are produced either byrapid decomposition of primary minerals in young soils not yetdesilicated and ferralitized or by the permanent supply of weatherableminerals by either erosion or by aerosol or dust input (Muhs, 2001). Al-though not identified, the type of smectite is indicated by the molarratio of Al:Fe of between 2:1 and 3:1 and the high CEC of around 500–700 mmolc kg−1 in the soils, thus suggesting that ferruginous beidellitewhich is also constrained to weathering of basic igneous rocks is thedominant smectite on BCI (Fisher and Ryan, 2006; Kantor andSchwertmann, 1974; Righi et al., 1998; Ryan and Huertas, 2009).

The comparatively low degree of weathering of some soils on BCImay be attributed to several factors, all of which involve hydrology inoneway or another. Mean annual precipitation by itself is not a good in-dicator of leaching intensity, unless it is known or can be assumed thatthe rainfall total expressed asMAP is actually available for leaching. Thisis true wherever all rainfall infiltrates into the soil and then percolates

through it, but not where overland flow is frequent, as it is on BCI onBohio and Caimito marine (Godsey et al., 2004; Zimmermann et al.,2012). Any rainfall that does not enter the soil cannot possibly contrib-ute to leaching, and this portion of rainfall was estimated as one-thirdfor one catchment on BCI (Dietrich et al., 1982; A. Zimmermann, person-al communication). That is, the leaching-effective precipitation is closerto 1500 mm. Further, the frequent occurrence of overland flowon someparts of BCI affects the degree ofweathering in yet anotherway. The rateof erosion attributable to overland flow (Zimmermann et al., 2012)maywell match or even exceed the rate of weathering and hence maintainsoils in a ‘youthful’ state. Topography as an additional factor results inpoor drainage in some parts, such as at the bottom of the anti-dipslope of the andesite cap, the location of soil profile F6. Poor drainageimplies restricted leaching, and hence conditions favorable for theformation of smectite.

The occurrence of interstratified kaolinite/smectite (K/S, see soil F2,Table 4) in BCI-like environments has yet to be confirmed. Theseminerals occur in red-black soil sequences in Mediterranean climateswith MAP of 800–1000 mm (Righi et al., 1999; Vingiani et al., 2004),in the subtropical climate of Barbados with MAP of 1100 mm (Muhs,2001), in El Salvador with MAP of 1800 mm (Yerima et al., 1985) andin semiarid regions (Bühmann and Grubb, 1991; Herbillon et al.,1981). In contrast, no K/S minerals were found where MAP exceeds3000 mm (Nieuwnhuyse and van Breemen, 1997). However, Ryanand Huertas (2009) proposed that K/S minerals are common in moisttropical environment, but went undetected because of the difficulty oftheir detection. Ryan and Huertas (2009) identified a smectite–K/S–kaolinite transformation trajectory in moist tropical Costa Rica withMAP of 3000 mm and a dry season of 3 month. BCI receives a mean an-nual precipitation of 2600 mm intermediate between the semiarid toarid regions where K/S minerals occur and the (per)humid oneswhere no K/Sminerals occur. The occurrence of K/Sminerals is attribut-ed to (i) parent material rich in aluminum and iron, (ii) a humid tosemiarid climate, (iii) plants extracting SiO2, (iv) sufficient time ofweathering and (v) intermediate drainage (Hughes et al., 1993). Al-though the formation of K/S minerals is still under debate, Hugheset al. (1993) suggest that the transition of smectite to kaolinite is bestexplained by dissolution–recrystallization processes or by dissolutionof some layers of the smectite and subsequent crystallization of kaolin-ite in the interlayer of the remaining smectite (Środoń, 1980).

The occurrence of K/S minerals in the 0–10 and 10–20 cm depths insoil F2 on Bohio points to intermediate drainage conditions on thisplateau-like site (Bühmann and Grubb, 1991; Herbillon et al., 1981;Righi et al., 1998, 1999; Vingiani et al., 2004). In deeper soil (50–70 cm) of F2 on Bohio, the drainage is more impeded and soil pHdrops to b5. However, smectite dominates over kaolinite because ofthe high nutrient status, and K/S minerals are absent. In this acidic andwaterlogged condition, a direct transition from smectite to disorderedkaolinite is expected because of the dissolution of smectite and precip-itation of kaolinite under acidic conditions (Fisher and Ryan, 2006;Środoń, 2006).

The soil F6 on Andesite shows a uniform distribution of smectite tokaolinite of 3:1 throughout the profile. This seems exceptional becauseof the acidic conditions of pH 5.7 in the 0–10 cm depths to pH 4.9 inthe 50–70 cm depths. However, the impeded drainage indicated bystrong hydromorphic patterns in addition with the nutrient-rich condi-tions (Tables 2 and 5), favors the precipitation of smectite. The kaoliniteis most likely the product of the dissolution of smectites and the directrecrystallization of the remaining Al- and Si-ions.

The soils F9 and A9 in the Lutz catchment on Caimito marine differfrom the soils on Bohio and Andesite in that they are dominated by smec-tites, which account for over 90% of all minerals in the clay fraction in F9and for almost 100% in A9 (Table 4). As pointed out above, this catchmentis subject to high erosion rates (Zimmermann et al., 2012), which resultsin a continuous rejuvenation of soils incompatible with the formation ofweathering products such as kaolinite. This interpretation is also borne



out by the concentration of exchangeable cations in soils on this forma-tion (Table 2).

Because of their three-dimensional structure with large structuralchannels and cavities, zeolites possess a high CEC between 1000and 3000 mmolc kg−1 which likely attributes to the CEC on Caimitomarine formation. Zeolites in soils are reported in a variety of climatesas product of alteration of volcanic glass in open hydrologic systems(e.g., Bhattacharyya et al., 1999). According to Woodring (1958) andJohnsson and Stallard (1989), the parent material of the Caimito marineformation is on one hand foraminiferal limestone (as the source of basecations as Ca) and on the other hand tuffaceous silt- and sandstone withvitric volcaniclastic debris. Groundwater percolating through the columnof volcanic ash or glass is enriched in Si and Al as well as alkali and earthalkali cations. Hence, zeolites will precipitate from this solution in areaswhere volcanic glass is being dissolved. The zeolites are incorporatedinto the soil as lithogenicmineral phasewhen the zeolite-rich sedimenta-ry deposits are exposed to the land surface and soil formation occurs(Ming and Boettinger, 2001). The presence of volcanic glass or its formerpresence was indicated by cristobalite on Caimito marine (F9) as well ason Bohio (F2), adjacent to Caimito marine (Table 4). However, the soilA9 differs from the soil F9 most likely because there was no depositionof soil material from the slopes and neither zeolites nor cristobalitewere detected.

The soil on Caimito volcanic, F11, with the lowest CEC of all studysoils differs in the mineral assemblage of the b2 μm fraction from theother soils (Tables 2 and 4). The dominating secondary mineralsthroughout the soil profile are kaolinite, gibbsite, and goethite. Howev-er, the absence of smectite is surprising because in soil F11 abundanthardly weathered blocks were removed while excavating the profile,indicating that base cations are supplied continuously by weathering.This leads to the assumption that drainage conditions on this foot-slope position are much better and thus removal of dissolved basecations is much higher than on other comparable sites on BCI. Indeed,in soil F11, CEC and exchangeable Ca concentrations were still higherthan those in other kaolinite-dominated soils of the moist tropics(Table 1, Fig. 2). Ryan and Huertas (2009) reported a kaolinite in CostaRica with a CEC of 180 mmolc kg−1 soil and suggested that elevatedCECwas attributable to the substitution of Mg2+ for Al3+ in the octahe-dral sheets. Because Mg is abundantly present in the parent materialand soils of Caimito volcanic including F11 and is mainly mineral-bound (Table 3), it can be assumed that isomorphic substitution of Alby Mg might be a reason for elevated CEC in soil F11.

Our mineralogical analyses demonstrated that the high CEC in allsoils on BCI is mainly attributable to high concentration of smectites inmost and the presence of zeolites in some soils. On the only lithologyon which no smectites occurred in soils, the dominating kaolinite hadan unusually high permanent surface charge likely because of the highdegree of isomorphic substitution of Al by Mg (Moore and Reynolds,1997). Thus, the study soils have a high capacity to retain cationsreleased from parent material or deposited from the atmosphere.

4.3. Weathering trends

The increasing trends of CIA and PIA, and the decreasing gradient ofWIP from topsoil to subsoil (regardless of whether exchangeable metalswere included in their calculation or not) are uncommon in moisttropical environments, where the vertical distribution of weatheringindices usually decrease from top to bottom (Price and Velbel, 2003).Weathering agents are assumed to reach the soil via rainfall or becauseof the concentration of roots in the topsoil leading to a higher concentra-tion of root exudates and organic acids acting as source of H+. Thus, theindex values of CIA, PIA, andWIP suggested soil rejuvenation because oferosion which exposes little weathered subsoil rock material near thesurface (Zimmermann et al., 2012). In the nearby river Chagres basin,Nichols et al. (2005) reported a denudation rate of up to 100 mm kyr−1

and an erosion rate of 2.7 Mg ha−1 yr−1 caused by tectonic uplift

which is proposed to be in steady-state, indicated by cosmogenic 10Beconcentrations in the sediments in the catchment. This is consistentwith the results of Zimmermann et al. (2012) who report an erosionrate of 2 Mg ha−1 yr−1 in the Lutz catchment. However, the latterhypothesize that the combination of regular near-surface flow, thenutrient- rich soils and sparse understory triggers erosion under closedforests. Because of these high erosion rates fresh unweathered materialis regularly exposed at the surface, weathers and releases dissolved basecations supplying the soil with lithogenic nutrients.

It was shown by various authors that parent material even in highlyweathered soils is the main source of base cations (Bern et al., 2005;Pett-Ridge et al., 2009; Porder et al., 2006; Vitousek et al., 2003). ThisCa source was also proposed by Johnsson and Stallard (1989) andYavitt et al. (1992) for BCI, although the latter authors based their con-clusion on studies on Caimitomarinewhichmight not be generalized tothewhole BCI. Among the candidate minerals for Ca release, plagioclasemight be the most relevant because, Sak et al. (2004) showed thatplagioclase is the mineral most susceptible for rapid weathering byacidic weathering agents.

Furthermore, the influence of plants on the redistribution of lithogenicbase cations (Ca, K, Mg, and P) by uptake and recycling to the soil surfacevia litterfall, throughfall and stemflow, partially compensating forleaching losses, is well documented (Dijkstra and Smits, 2002; Jobbágyand Jackson, 2001; Porder and Chadwick, 2009) and generally appliesfor limiting nutrients. On BCI, biological redistribution of P and Kwas sug-gested by Barthold et al. (2008) and Dieter et al. (2010). At the nearbyGigante peninsula it was shown that P and K are co-limiting nutrients(Kaspari et al., 2008; Wright et al., 2011). However, it is unlikely that Cais a limiting nutrient onBCI because of the high bioavailable Ca concentra-tion in soil. The latter assumption is supported by our finding that therewere no differences in Ca concentrations of soil below fig trees with ahigh and cashew treeswith a lowCa requirement. Thus, wewould expecta differential vertical distribution of K and Mg versus Ca which is not thecase (data not shown).

Finally, external inputs might be an important reason for high Caconcentrations in soil and accumulation in topsoil. This suggestion iscorroborated by the fact that the base saturation is higher than expectedfrom the acid pH values of the soils (Young, 1976) indicative of non-equilibrium conditions in which base metal leaching is continuouslycompensated by deposition. Additionally, the Ca mass balances of thefive profiles, which in all cases show a smaller loss of Ca in the topsoilthan in the subsoil as indicated by the respective τ values, suggest anexternal input of Ca on BCI. This assumption is further supported bythe fact that theweathering indices CIA, PIA andWIP showanunexpect-ed vertical distribution suggesting the smallest degree of weathering inthe topsoil both if calculated with or without exchangeable metals(Fig. 4). When the weathering indices calculated with and withoutexchangeable metals are compared, it becomes obvious that the effectof the exchangeable metals on the weathering indices is strongest inthe topsoil, which points at a base metal input to the topsoil. OnHawaii, several authors have shown along a chronosequence that withincreasing age of the soils the external supply with nutrientstransported from the Asian deserts become more and more importantto maintain the nutrient stock supporting the local vegetation (Kurtzet al., 2001; Vitousek et al., 2003; Wiegand et al., 2005). Refreshmentof soil material with Sahara dust is known for Caribbean islands suchas Barbados, the Lesser Antilles (Muhs, 2001; Muhs et al., 1990), andPuerto Rico (Pett-Ridge et al., 2009). For Caribbean islands, Muhs et al.(1990) showed that Saharan dust is the major parent material for soilsin the last 900,000 yr and that bauxite deposits on different Caribbeanisland on Tertiary carbonate rocks are related to this trans-Atlantictransport mechanism of trade winds. According to Prospero et al.(1981), the main constituents of Saharan dust deposited on Barbadosare mica (60%), quartz (10%), kaolinite (6%), plagioclase (6%), calcite(5%), chlorite (4%), microcline and gypsum (each 3%), and goethite(2%). Saharan dust is supposed to be relatively uniform in composition



across the Atlantic to the Caribbean region, implying that it is reasonableto assume a similar composition for the dust reaching Panama. Mica arenot found in the soil matrix on BCI by XRD (own unpublished data), butmica are susceptible to dissolution because of its phyllosilicate structureunder warm and humid conditions (Muhs, 2001). The deposition ofmica with Sahara dust might therefore explain the possible presenceof illite in soil F11 on Caimito volcanic. Although we do not know ofany report on Sahara dust deposition in Panama, the high Ca depositionof between 30 and 37 kg ha−1 yr−1 with throughfall in Panama report-ed by several authors (Cavelier et al., 1997; Golley et al., 1976)might in-dicate Sahara dust inputs. This interpretation is supported by the factthat the Ca fluxes exceed the fluxes in other tropical sites by morethan 10 kg Ca ha−1 even when compared with sites for which dust in-puts were already shown (Puerto Rico, McDowell, 1998; Pett-Ridge etal., 2009; Jamaica, Hafkenscheid, 2000).

5. Conclusion

Concentrations of exchangeable Ca in soils on BCI are higher thanthose reported from many other locations in the tropics. An importantreason for high Ca concentrations is the high CEC which is mainlyattributable to the abundance of smectites in soils on three of the fourlithologies and the occurrence of some zeolite. Smectites likely precipitatebecause the soil solution has high concentrations of basemetals and a pHvalue N5 combined with poor to intermediate drainage. InterstratifiedK/S minerals are present as intermediate secondary mineral betweensmectite and kaolinite and contribute to CEC.

The various lithologies on BCI consist, with the exception of Caimitomarine, of intermediate basaltic to andesitic rocks in which the rapidlydecomposing plagioclase is most likely the only native source of Ca forthe soils. There are indications that pronounced soil erosion exposes lit-tleweathered rockmaterial near the soil surface acting as a source of Ca.The fact that the highest exchangeable Ca concentrations occurred inthe topsoil and that soils on BCI were apparently more weathered inthe subsoil than in the topsoil as implied by three weathering indices,and a reported deposition rate of more than 30 kg Ca ha−1 yr−1 withthroughfall in Panama, strongly suggest an atmospheric source of basemetals rather than a lithogenic one. Because Ca is unlikely to be agrowth-limiting element, the Ca input from above can also hardly beexplained by the base pumping effect of plants. Therefore, we proposethat an external source, such as Saharan dust input, is the main driversupplying BCI with Ca and K.

Acknowledgments

We thank Oris Acevedo and Belkys Jimenez for their organizationaland logistical support on BCI, the late Elisabeth K.V. Kalko, LarissaAlbrecht and Inga Geipel for their helpful advices about and introduc-tion to BCI, the Smithsonian Tropical Research Institute (STRI) forgranting access to its research station, the Panamanian authorities(ANAM) for the export permit, the German Research Foundation(DFG,Wi1601/14-1) for financial support, and one reviewer for insight-ful comments on an earlier version of this manuscript. H. E. acknowl-edges support by the Smithsonian Tropical Research Institute duringhis sabbatical in 2012. We also thank the head of the XRD lab, UrsEggenberger and the technical assistance Christine Lemp at the Instituteof Geology in Bern for their instruction in XRD-measurements andDaniela Fischer for XRF-measurements.

References

Baillie, I., Elsenbeer, H., Barthold, F.K., Grimm, R., Stallard, R., 2007. Semi-detailed Soil Surveyof Barro Colorado Island, Panama. online: http://biogeodb.stri.si.edu/bioinformatics/bci_soil_map/.

Barthold, F.K., Stallard, R.F., Elsenbeer, H., 2008. Soil nutrient–landscape relationships in alowland tropical rainforest in Panama. For. Ecol. Manag. 255, 1135–1148.

Beinroth, F., 1982. Some highly weathered soils of Puerto Rico, 1. Morphology, formationand classification. Geoderma 27, 1–73.

Bern, C.R., Townsend, A.R., Farmer, G.L., 2005. Unexpected dominance of parent-materialstrontium in a tropical forest on highly weathered soils. Ecology 86, 626–632.

Bhattacharyya, T., Pal, D., Srivastava, P., 1999. Role of zeolites in persistence of high alti-tude ferruginous Alfisols of the humid tropical Western Ghats, India. Geoderma 90,263–276.

Boy, J., Wilcke, W., 2008. Tropical Andean forest derives calcium and magnesium fromSaharan dust. Glob. Biogeochem. Cycles 22, GB1027.

Brasell, H.M., Unwin, G.L., Stocker, G.C., 1980. The quantity, temporal distribution andmineral-element content of litterfall in two forest types at two sites in tropicalAustralia. J. Ecol. 68, 123–139.

Brimhall, G.H., Dietrich, W.E., 1987. Constitutivemass balance relations between chemicalcomposition, volume, density, porosity, and strain in metasomatic hydrochemicalsystems: results on weathering and pedogenesis. Geochim. Cosmochim. Acta 51,567–587.

Brindley, G.W., Brown, G., 1980. Crystal Structures of Clay Minerals and Their X-rayIdentification. Mineralogical Society, London, UK.

Bühmann, C., Grubb, P.L.C., 1991. A kaolin-smectite interstratification sequence from a redand black complex. Clay Miner. 26, 343–358.

Burnham, C.P., 1987. Pedological processes and nutrient supply from parent material intropical soils. In: Proctor, J. (Ed.), Mineral Nutrients in Tropical Forest and SavannaEcosystems. Special Publication No. 9 of The British Ecological Society. BlackwellScientific Publ., pp. 27–41.

Buschbacher, A.R., Uhl, C., Serrao, E.A.S., 1988. Abandoned pastures in Eastern Amazonia.II. Nutrient stocks in the soil and vegetation. J. Ecol. 76, 682–699.

Cavelier, J., Jaramillo, M., Solis, D., de Leon, D., 1997. Water balance and nutrient inputs inbulk precipitation in tropical montane cloud forest in Panama. J. Hydrol. 193, 83–96.

Chadwick, O.A., Brimhall, G., 1990. From a black to a gray box—a mass balance interpre-tation of pedogenesis. Geomorphology 3, 369–390.

Chadwick, O.A., Gavenda, R.T., Kelly, E.F., Ziegler, K., Olson, C.G., Elliott, W.C., Hendricks,D.M., 2003. The impact of climate on biogeochemical functioning of volcanic soils.Chem. Geol. 202, 195–223.

Clark, K.L., Nadkarni, N.M., Schaefer, D., Gholz, H.L., 1998. Atmospheric deposition and netretention of ions by the canopy in a tropical montane forest, Monteverde, Costa Rica.J. Trop. Ecol. 14, 27–45.

Dieter, D., Elsenbeer, H., Turner, B.L., 2010. Phosphorus fractionation in lowland tropicalrainforest soils in central Panama. Catena 82, 118–125.

Dietrich, W.E., Windsor, D.M., Dunne, T., 1982. Geology, climate, and hydrology of BarroColorado Island. In: Leigh, E.G., Rand, S.A., Windsor, D.M. (Eds.), The Ecology of aTropical Forest: Seasonal Rhythms and Long Term Changes. Smithsonian Institution,Washington, DC, pp. 21–46.

Dijkstra, F.A., Smits, M.M., 2002. Tree species effects on calcium cycling: the role of calci-um uptake in deep soils. Ecosystems 5, 385–398.

Dijkstra, F.A., van Breemen, N., Jongmans, A.G., Davies, G.R., Likens, G.E., 2003. Calciumweathering in forested soils and the effect of different tree species. Biogeochemistry1995, 253–275.

DIN ISO 19730, 2008. Soil Quality — Extraction of Trace Elements From Soil UsingAmmonium Nitrate Solution.

Duzgoren-Aydin, N., Aydin, A., Malpas, J., 2002. Re-assessment of chemical weatheringindices: case study on pyroclastic rocks of Hong Kong. Eng. Geol. 63, 99–119.

FAO, 2006. Guidelines for Soil Description. Food and Agriculture Organization of theUnited Nations, Rome, Italy.

Fedo, C.M., Nesbitt, H.W., Young, G.M., 1995. Unraveling the effects of potassiummetaso-matism in sedimentary rocks and paleosols, with implications for paleoweatheringconditions and provenance. Geology 23, 921–924.

Fisher, G.B., Ryan, P.C., 2006. The smectite-to-disordered kaolinite transition in a tropicalsoil chronosequence, Pacific coast, Costa Rica. Clay Clay Miner. 54, 571–586.

Fujinuma, R., Bockheim, J., Balster, N., 2005. Base-cation cycling by individual tree speciesin old growth forests of Upper Michigan, USA. Biogeochemistry 74, 357–376.

Godsey, S., Elsenbeer, H., Stallard, R.F., 2004. Overland flow generation in two lithologicallydistinct rainforest catchments. J. Hydrol. 295, 276–290.

Golley, F., McGinnis, J.T., Clements, R.G., Child, G.I., Duever, M.J., 1976. Mineral Cycling in aTropical Moist Forest Ecosystem. University of Georgia Press, Athens, USA.

Grubb, P.J., Tanner, E.V.J., 1976. The montane forests and soils of Jamaica: a re-assessment.J. Arnold Arbor. 57, 313–368.

Hafkenscheid, R., 2000. Hydrology and Biogeochemistry of Tropical Montane Rain Forestsof Contrasting Stature in the Blue Mountains, Jamaica. (PhD. thesis) VU Amsterdam.

Herbillon, A.J., Frankart, R., Vielvoye, L., 1981. An occurrence of interstratified kaolinite–smectite minerals in a red-black soil toposequence. Clay Miner. 16, 195–201.

Holdridge, L.R., Budowski, G., 1956. Report of an ecological survey of the Republic ofPanama. Caribb. For. 17, 92–110.

Hölscher, D., Köhler, L., Leuschner, C., Kappelle, M., 2003. Nutrient fluxes in stemflow andthroughfall in three successional stages of an upper montane rain forest in Costa Rica.J. Trop. Ecol. 19, 557–565.

Hughes, R.E., Moore, D.M., Reynolds Jr., R.C., 1993. The nature, detection and occurrence,and origin of kaolinite/smectite. In: Murray, H.H., Bundy, W.M., Harvey, C.C. (Eds.),Kaolin Genesis and Utilization. Special Publication No. 1. Clay Minerals Society,Boulder, Colorado, pp. 291–323.

IUSSWorking GroupWRB, 2006.World Soil Resource Reports No. 103. Food and AgricultureOrganization of the United Nations, Rome, Italy.

Jobbágy, E.G., Jackson, R.B., 2001. The distribution of soil nutrients with depth: globalpatterns and the imprint of plants. Biogeochemistry 53, 51–77.

Johnsson, M.J., Stallard, R.F., 1989. Physiographic controls on the composition of sedimentsderived from volcanic and sedimentary terrains on Barro Colorado Island, Panama.J. Sediment. Res. 59, 768–781.



Jordan, C.F., 1982. Nutrient Cycling in Tropical Forest Ecosystems: Principles and TheirApplication in Management and Conservation. John Wiley & Sons, Chichester, UK.

Kantor, W., Schwertmann, U., 1974. Mineralogy and genesis of clays in red-black soiltoposequences on basic igneous rocks in Kenya. J. Soil Sci. 25, 67–78.

Kaspari, M., Garcia, M.N., Harms, K.E., Santana, M., Wright, S.J., Yavitt, J.B., 2008. Multiplenutrients limit litterfall and decomposition in a tropical forest. Ecol. Lett. 11, 35–43.

Kurtz, A.C., Derry, L.A., Chadwick, O.A., 2001. Accretion of Asian dust to Hawaiian soils:isotopic, elemental, and mineral mass balances. Geochim. Cosmochim. Acta 65,1971–1983.

Lathwell, D.J., 1974. Report of Caribbean and tropical America Soils Conference, Held atUniversity of West Indies. Cornell University, Ithaca, NY, USA.

Leigh, E.G., 1999. Tropical Forest Ecology: A View From Barro Colorado Island. OxfordUniversity Press, New York, USA.

Likens, G.E., Driscoll, C.T., Buso, D.C., Siccama, T.G., Johnson, C.E., Lovett, G.M., Fahey, T.J.,Reiners, W.A., Ryan, D.F., Martin, C.W., Bailey, S.W., 1998. The biogeochemistry of cal-cium at Hubbard Brook. Biogeochemistry 41, 89–173.

McDowell, W.H., 1998. Internal fluxes in a Puerto Rican rain forest. J. Trop. Ecol. 14,521–536.

Ming, D.W., Boettinger, J.L., 2001. Zeolites in soil environments. In: Bish, D.L., Ming,D.W. (Eds.), Geology, mineralogy, properties, and utilization of natural zeolites.Reviews in Mineralogy and Geochemistry. Mineralogical Society of America,Washington D.C.

Moore, D.M., Reynolds, R.C., 1997. X-ray Diffraction and the Identification and Analysis ofClay Minerals. Oxford University Press, UK.

Muhs, D.R., 2001. Evolution of soils on Quaternary reef terraces of Barbados, West Indies.Quat. Res. 56, 66–78.

Muhs, D.R., Bush, C.A., Stewart, K.C., Rowland, T.R., Crittenden, R.C., 1990. Geochemical ev-idence of Saharan dust parent material for soils developed on quaternary limestonesof Caribbean and Western Atlantic Islands. Quat. Res. 33, 157–177.

Muhs, D.R., Budahn, J.R., Muhs, D.R., Budahn, J.R., 2007. Geochemical evidence for Africandust and volcanic ash inputs to Terra Rossa Soils on carbonate reef terraces, NorthernJamaica, West Indies. USGS Staff — Published Research. Paper 172.

Nesbitt, H., Young, G.M., 1982. Early Proterozoic climates and plate motions inferred frommajor element chemistry of lutites. Nature 299, 715–717.

Nichols, K.K., Bierman, P.R., Finkel, R., Larsen, J., 2005. Long-term sediment generationrates for the upper Rio Chagres basin: implications for Panama Canal watershedman-agement. In: Harmon, R.S. (Ed.), The Río Chagres: A multidisciplinary profile of aTropical Watershed. Water, Science and Technology Library. Springer, Dodrecht,Netherlands.

Nieuwnhuyse, A., van Breemen, N., 1997. Quantitative aspects of weathering andneoformation in selected Costa Rican volcanic soils. Soil Sci. Soc. Am. J. 61, 1450–1458.

Parker, A., 1970. An index of weathering for silicate rocks. Geol. Mag. 107, 501–504.Pett-Ridge, J.C., Derry, L.A., Barrows, J.K., 2009. Ca/Sr and 87Sr/86Sr ratios as tracers of Ca

and Sr cycling in the Rio Icacos watershed, Luquillo Mountains, Puerto Rico. Chem.Geol. 267, 32–45.

Porder, S., Chadwick, O.A., 2009. Climate and soil-age constraints on nutrient uplift andretention by plants. Ecology 90, 623–636.

Porder, S., Clark, D.A., Vitousek, P.M., 2006. Persistence of rock-derived nutrients in thewet tropical forests of La Selva, Costa Rica. Ecology 87, 594–602.

Price, J.R., Velbel, M.A., 2003. Chemical weathering indices applied toweathering profiles de-veloped on heterogeneous felsic metamorphic parent rocks. Chem. Geol. 202, 397–416.

Proctor, J., 1983. Mineral nutrients in tropical forests. Prog. Phys. Geogr. 7, 422–431.Prospero, J., Glaccum, R., Nees, R., 1981. Atmospheric transport of soil dust from Africa to

South America. Nature 289, 570–572.Richter, D.D., Babbar, L.I., 1991. Soil diversity in the tropics. Adv. Ecol. Res. 21, 315–389.Righi, D., Terribile, F., Petit, S., 1998. Pedogenic formation of high-charge beidellite in a

vertisol of Sardinia (Italy). Clay Clay Miner. 46, 167–177.

Righi, D., Terribile, F., Petit, S., 1999. Pedogenic formation of kaolinite–smectite mixedlayers in a soil toposequence developed from basaltic parent material in Sardinia(Italy). Clay Clay Miner. 47, 505–514.

Ryan, P.C., Huertas, F.J., 2009. The temporal evolution of pedogenic Fe–smectite to Fe–kaolin via interstratified kaolin–smectite in a moist tropical soil chronosequence.Geoderma 151, 1–15.

Sak, P.B., Fisher, D.M., Gardner, T.W., Murphy, K., Brantley, S.L., 2004. Rates of weatheringrind formation on Costa Rican basalt. Geochim. Cosmochim. Acta 68, 1453–1472.

Sanchez, P.A., Logan, T.J., 1992.Myths and science about the chemistry and fertility of soilsin the tropics. Soil Sci. Soc. Am. Spec. Publ. 29, 35–46.

Silver, W., Sactena, F.N., Johnson, A.H., Siccama, T.G., Sanchez, M.J., 1994. Nutrientavailability in a montane wet tropical forest: spatial patterns and methodologicalconsiderations. Plant Soil 164, 129–145.

Sollins, P., 1987. Factors affecting nutrient cycling in tropical soils. In: Proctor, J. (Ed.),Mineral nutrients in tropical forest and savanna ecosystems. Special Publication No.9 of the British Ecological Society. Blackwell Scientific Publ., pp. 27–42.

Sollins, P., Sancho, M.F., Mata, Ch.R., Sanford, R.L., 1994. Soils and soil process research. In:McDade, L.A., Bawa, K.S., Hespenheide, H.A., Hartshorn, G.S. (Eds.), La Selva: Ecologyand Natural History of a Neotropical Rain Forest. University of Chicago Press, Chicago,pp. 34–53.

Środoń, J., 1980. Synthesis of mixed-layer kaolinite/smectite. Clay Clay Miner. 28,419–424.

Środoń, J., 2006. Identification and quantitative analysis of clay minerals. In: Bergaya, F.,Theng, B.K.G., Lagaly, G. (Eds.), Handbook of Clay Science. Elsevier.

Tanner, E.V.J., 1977. Four montane rain forests of Jamaica: a quantitative characterization ofthe floristics, the soils and the foliarmineral levels, and a discussion of the interrelations.J. Ecol. 65, 883–918.

Thanachit, S., Suddhiprakarn, A., Kheoruenromne, I., Gilkes, R.J., 2006. The geochemistryof soils on a catena on basalt at Khon Buri, northeast Thailand. Geoderma 135, 81–96.

Vingiani, S., Righi, D., Petit, S., Terribile, F., 2004. Mixed-layer kaolinite–smectiteminerals in ared-black soil sequence from basalt in Sardinia (Italy). Clay Clay Miner. 52, 473–483.

Vitousek, P.M., Sanford, R.L., 1986. Nutrient cycling in moist tropical forest. Annu. Rev.Ecol. Syst. 17, 137–167.

Vitousek, P.M., Chadwick, O.A., Matson, P., Allison, S., Derry, L., Kettley, L., Luers, A., Mecking,E., Monastra, V., Porder, S., 2003. Erosion and the rejuvenation of weathering-derivednutrient supply in an old tropical landscape. Ecosystems 6, 762–772.

Wiegand, B.A., Chadwick, O.A., Vitousek, P.M., Wooden, J.L., 2005. Ca cycling and isotopicfluxes in forested ecosystem. Geophys. Res. Lett. 32, 2–5.

Windsor, D.M., 1990. Climate and moisture availability in a tropical forest: long-term en-vironmental records from Barro Colorado Island, Panama. Smithsonian Contributionsto Earth Sciences.Smithsonian Institution Press, Washington D.C.

Woodring, W.P., 1958. Geology of Barro Colorado Island, Canal Zone. Smithson. Misc.Collect. 135, 1–39.

Wright, S.J., Yavitt, J.B., Wurzburger, N., Turner, B.L., Tanner, E.V.J., Sayer, E.J., Santiago, L.S.,Kaspari, M., Hedin, L.O., Harms, K.E., Garcia, M.N., Corre, M.D., 2011. Potassium,phosphorus, or nitrogen limit root allocation, tree growth, or litter production in alowland tropical forest. Ecology 92, 1616–1625.

Yavitt, J.B., Wright, S.J., 2002. Charge characteristics of soil in a lowland tropical moistforest in Panama in response to dry-season irrigation. Soil Res. 40, 269–281.

Yavitt, J.B., Wieder, R.K., Wright, S.J., 1992. Soil nutrient dynamics in response to irrigationof a Panamanian tropical moist forest. Biogeochemistry 19, 1–25.

Yerima, B.P.K., Calhoun, F.G., Senkayi, A.L., Dixon, J.B., 1985. Occurrence of interstratifiedkaolinite–smectite in El Salvador Vertisols. Soil Sci. Soc. Am. J. 49, 462–466.

Young, A., 1976. Tropical Soils and Soil Survey. Cambridge University Press, Cambridge.Zimmermann, A., Francke, T., Elsenbeer, H., 2012. Forests and erosion: Insights from a study

of suspended-sediment dynamics in an overland flow-prone rainforest catchment.J. Hydrol. 428–429, 170–181.


High exchangeable calcium concentrations in soils on Barro Colorado Island, Panama

Documents

Transcript of High exchangeable calcium concentrations in soils on Barro Colorado Island, Panama