Chemical and Mineralogical Properties of Tephras and Mobility of Chemical Elements in Tephra-Derived Soils

Ichiro YAMADA

Department of Agro-Environmental Management, Kyushu National Agricultural Experiment Station (Nishigoshi, Kumamoto, 861-11 Japan)

Abstract

The SiO2 content of tephras was closely correlated with the contents of other major elements except for the K2O content in Japan tephras, which varied considerably with the volcanic zones where the volcanoes were located. Volcanic glass was abundant in rhyolitic and basaltic tephras and the contents of heavy minerals and were high in dacitic, andesitic and basaltic andesitic tephras. Although the heavy mineral assemblage was correlated with the rock types of tephras and volcanic zones in the Japan tephras, these correlations were not clear in the New Zealand tephras. The composition of volcanic glass in andesitic, dacitic and rhyolitic tephras was felsic and non-colored glass with refractive indices about 1.50 was observed. In the mafic tephras, the composition of glass was intermediate and mafic and the colored glass with indices greater than 1.52 was observed. The stability sequence of the primary minerals in tephra-derived soils under temperate and humid climate conditions was graded as follows: volcanic glass < plagioclase = olivine < augite < hyperthene common hornblende ferromagnetic minerals. The mobility sequence of the major elements in tephra-derived soils under these climate conditions was considered to be as follows: Cao, Na2O > SiO2 > MgO, K2O > Al2O3, Fe2O3.

Discipline: Soil, fertilizers and plant nutrition Additional key words: volcanic glass, primary minerals, volcanic ash, Andosols 82 JARQ 31 (2) 1997 provinces with a significant reflection on the chemi­ alumina and alkali basalt contents, while the cal and mineralogical properties of volcanic products tholeliite petrographic province in Japan is lacking are different between Japan and New Zealand t>. In in New Zealand 1>. It is important to investigate the Japan, there are petrographic provinces with high properties of tephras in these counlries because

20 14 19- 12 18 - OD 17 . 10 ~ ~ ~ 16 . § 8 "' 0 15 · ON"' 6 Q) -"' 14 u. < 13 ~ -~ 4- ~~i 12 : 2 · 11 . i_ 10+.-, -,.,,..,....,..,,..,.,..,.,.....,...,.,..,..,....,,...,...-.--,..,..., 0 45 so 55 60 65 70 75 80 45 50 55 60 65 70 75 80 10 12 9 · • 8 ~ D 10 oa 7 ~ ...... 8 § 6 • • 0 ~ ~ 5 ..: 0 6 ~ 4 - <..>"' 3 · 4 - .. \~ 2 ..: 1 ..: 2 .irnf 0 JI.: -,., .,.,.., -r-,r--~-,-,,-~~~ - 0 • 45 50 55 60 65 70 75 80 45 50 55 60 65 70 75 80 4 . 6 . 3.5 • 5 0 D 3 ...... 2.5 · 4 · ~ ~ 0 2 ·JI~~. ON 3 :,,,:;"" jl D 1.5 :z"' 2- D 1 ..: 0.5 1 -: 0 -t-.--'-1.,.,.,.-r,-'TT"!-r,--r--,-,..,...,...... , 0 45 50 55 60 65 70 75 80 45 50 55 60 65 70 75 80 1.4 0.2 1.2 D D~ O 0 0. 15 ~ § 0.8 ~ 0 o 0.1 ""0.6 ...: c:: "~C\:Jo £10D\ I- :s 6 0.4 0.05 A b 0.2 - 0-1---.--.--,.,-...,..,..~~....---. 0 45 so 55 60 65 70 75 80 45 50 55 60 65 70 75 80 SiO (%) SiO (%) 2 2 D Japan e N.Z. A USA

Fig. l. Contents of major clements in relation to silica in tephras I. Yamada : Chemical and Mineralogical Characterisrics of Tephras and Volcanic Soils 83 tephras easily undergo a weathering process and of Japan tephras including the 26 samples reported release chemical elements affecting the fertility of by Shoji et al. 8>, 9 New Zealand samples and 2 USA tephra-derived soils. samples 10>. Correlation equations between Si02 Shoji et al. 3> classified tephras into 5 rock types content (x) and contents of other elements (y) of on the basis of silica content, as follows; tephra samples of Japan and New Zealand are as (70 to I OOOJo Si02), dacite (62 to 700Jo Si02), follows: (58 to 620Jo Si02), basaltic andesite (53 .5 to 580/o Japan tephras (n = 33): AbOJOJo y = - 0.250x Si02) and basalt (45 to 53.50Jo Si02). + 31.52 r = - 0.882***, Fe20 30Jo y = - 0.385x + Fig. I shows the analytical data of 33 samples 30.34 r = - 0.924***, MgOOJo y = - 0.247x + 17.78

100 wt.% K?muinupuri-c Towada-a 80 (rhyolite) (rhyolite)

60 VG VG 40 HM lo"- 20 /0-...... ---0 -o ---o- Pl o\ o o Pl /0 • 0 o-- , -., ...... -H~ 100 ·-·-·-·

80 Towada-Cu Tarumae-a (dacite) (andesite)

60 VG VG

40

20

80 Zao Hoei (basaltic andesite) (basalt)

60 VG VG 40

20 o--o--- / 0 0-0- 0 Pl , _ ,...... _ , , - • --- HM -----..0,::-:.,.. s 10 20 so 100 200 500 1000 S 10 20 so 100 200 500 1000

Fig. 2. Primary mineral composilion of frac1ion of each size VG : Volcanic glass. Pl : Plagioclase, HM: Heavy minerals. 84 JARQ 31(2) 1997

Table 1. Relationships between rock types and heavy mineral assemblage •>

Basaltic Rhyolite Dacite Andesite Basalt Total andesite 7 10 17 (15%) 4 12 31 6 53 (470/o) Pyroxene· hornblende 17 II 28 (25%) Hornblende 13 14 (13%) Total 5 42 49 6 10 112 (4%) (380/o) (44%) (5%) (9%) (100%) a): Figure shows the number of samples.

r = - 0.916***, Ca0% y = - 0.324x + 25.80 r = plagioclase were higher in intermediate tephras com­ - 0.926***, Na20% y = 0.107x - 3.139 r = 0.862***, posed of dacite, andesite and basaltic andcsite. K20% y=0.030x-0.701 r=0.340, Ti02% Compared with the contents of heavy minerals in y= - 0.018x+ 1.790 r= -0.579***, Mn0% fractions with different sizes, the contents were higher y = - 0.0029x + 0.315 r = - 0.536*** . in the 0.1 to 0.5 mm fraction and lower in the frac­ New Zealand tephras (n = 9): Al203 % y = tion finer than 0.05 mm. -0.248x+31.16 r=-0.923***, Pe203% Table I shows the relationship between heavy y = - 0.34Sx + 26. 79 r = - 0.976***, Mg0% mineral assemblage and rock types of tephras and 161 y = - 0.220x + 16.20 r = - 0.844***, Ca0% tephra-derived soils in Tohoku, Japan • The an­ 0 y = - 0.364x + 27.66 r = - 0.959* , Na20% desitic and dacitic samples accounted for the larger y = 0.049x + 0.330 r = 0.812***, K20% part of 112 samples while the samples of other rock y = 0.096x - 3.981 r = 0.926***, Ti02% types were not conspicuous. y = - 0.029x + 2. 325 r = - 0.996***, Mn0% Based on the heavy mineral composition, tephras y = - 0.0045x + 0.397 r = - 0.950***. and tephra-derived soils were divided into 4 types: The Si02 content of the tephras was closely cor­ olivine type, pyroxene type, pyroxene-hornblende type related with the contents of all the major elements and hornblende type16>. As shown in Table I, the except for K20 in the Japan tephras and with the pyroxene type accounted for almost half of all the contents of all the elements in the New Zealand samples. Olivine type was not observed in felsic tephras. The K20 content in the Japan tephras was samples. Hornblende type was not observed in an­ closely correlated with the petrographic provinces desitic to basaltic samples and pyroxene-hornblende (volcanic zone), being lowest in the Nasu volcanic type was not detected in mafic samples. These results zone (tholeliite petrographic province), intermediate indicated that the heavy mineral composition could in the Chokai volcanic zone (high alumina basalt be used to estimate the rock types, namely the charac­ petrographic province) and highest in the Daisen teristics of chemical elements in tephra-derived soils volcanic zone (alkali basalt petrographic province) 141, of Japan. However, the correlation between the whose volcanic zones are located in parallel to the heavy mineral composition and the rock types was Japan trench. The potassium contents in the tephras 11ot clear in the New Zealand tephras 10>. belonging to the tholeliite petrographic province were As for the correlation between the heavy mineral low, even though the chemical composition of the assemblage of tephras and volcanic zone of Japan, tephras was felsic. hornblende and biotite were detected in the tephras of the Daisen volcanic zone, but not in the tephras 2) Mineralogy of tephras of the Nasu volcanic zone. Hornblende was detect­ Fig. 2 shows the contents of volcanic glass, ed, but biotite was absent in the tephras of the Chokai plagioclase, and heavy minerals in 6 tephras with volcanic zone which is located between the Nasu vol­ 111 different rock types • The contents of light miner­ canic zone and Daisen volcanic zone. als which were dominated by volcanic glass were much higher than those of heavy minerals in all the 3) Volcanic glass in tephras samples. Especially, volcanic glass was overwhelm­ Since tephras were produced under the rapid cool­ ingly abundant in rhyolitic and basaltic tephras. On ing of , volcanic glass is the dominant primary the other hand, the contents of heavy minerals and mineral in tephras. Volcanic glass is also considered I. Yomodo: Chemical and Mi11eralogical Clwracrerisrics of Tephros and Vo/ca11k Soils 85

80 22 75 20 Al 20 3 70 • 18 65 16 0 60 14 • Si02 ss 12 so 10 45 so ss 60 65 70 75 45 so 55 60 65 70 75 10 5 9 4.5 8 Fe 203 4 MgO 7 3.5 6 3 s • • 2.5 0 • <.!) 4 2 D > 3 1.5 C. -..,0 2 1 C 1 • 0.5 Cl) • E 0 0 Cl) ~ 45 so ss 60 65 70 75 45 so ss 60 65 70 75 .c u Cl)"' 14 5 -0 12 . 4.5 D ~ _9 ~· • i...... CaO 4 .., 10 3.5 • • ...-,6;0 • "'C ..,Cl) 8 3 D o • C 2.5 i:. 0 0 u 6 2 4 1.5 Na 20 2 1 • 0.5 0 0 45 so ss 60 65 70 75 45 50 ss 60 65 70 75

5 1.4 D 4.5 1.2 4 Ti02 K 20 • 3.5 1 D 3 en • 0.8 . 2.5 • ~o •• ~~ o• .... 0.6 -~ 2 D D _...... - 1.5 _·E{J- DO 0.4 o~ •s...... __ 1 • -· -- D D • 0.5 0.2 DCl) ~- 0 0 45 so 55 60 65 70 75 45 so 55 60 65 70 75 Bulk SiO (%) Bulk SiO (%) 2 2

D Japan e N.Z. .A. USA

Fig. 3. Relationship between bulk Si02(0/o) and the content of each elemem in volcanic glass 86 JARQ 31 (2) 1997

to be the least resistant mineral t.o weathering. These that the content of crystalline minerals such as heavy facts indicate that volcanic glass is the most impor­ minerals and plagioclase was largest in the andesitic tant primary mineral as pareni material of tephra­ tephras and lowest in the rhyolitic and basaltic tephras derived soils. as shown in Fig. 2. Fig. 3 shows the relationship between the silica Since the greater part of Na20 and K20 in the content of Japan tephras and New Zealand tephras tephras is contained in volcanic glasses, the devia­ and the content of each chemical element of the vol­ tions from the regression line were very small in the 10 14 canic glass • >. The straight lines in this figure in­ tephras of all rock types. dicate the correlation equations between the bulk silica Determination of the refractive indices of volcan­ content and the contents of other elments in the Japan ic glasses is simple and is very useful for estimating tephras, given by Shoji et al. 8>. their chemical composition. The values of the in­ The silica contents of volcanic glass in the rhyo­ dices of non-colored volcanic glass in the rhyolitic, litic, dacitic and andesitic tephras ranged from 73 dacitic and andesitic tephras ranged between 1.48 to 780/o and these volcanic glasses were non-colored, and 1.52, while those of colored volcanic glasses while the silica contents of volcanic glasses in the ranged between 1.51 and 1.57 in the basaltic andesit­ basaltic andesitic and basaltic tephras ranged from ic tephras and were higher than I .55 in the basaltic 60 to 70% and 50 to 550/o, respectively and these tephras 14>. volcanic glasses were colored. The deviations from The morphology of volcanic glasses in the tephras the regression line of the silica content of tephras was divided into 4 types, namely spongy, fibrous, on that of volcanic glasses were greatest in andesitic platy and berry types 14>. Non-colored volcanic glass tephras and small in rhyolitic and basaltic tephras. consisted of one or some of these types, whereas The chemical composition of non-colored volcanic colored volcanic glass consisted of only the berry type. glasses in the rhyolitic, dacitic and andesitic tephras was very similar to the chemical composition of the 4) Stability of primary minerals rhyolitic tephras shown in Fig. I. Therefore, the The release of chemical elements from the primary deviations from the regression Jines of Al20J, Fe203, minerals with the progression of weathering is close­ MgO, Cao and Ti02 were also greatest in the an­ ly related 10 the stability of primary minerals 10 chem­ desitic tephras and small in the rhyolitic and basaltic ical weathering. Fig. 4 shows the relationships tephras. These deviations may account for the fact between the rates of stages III and IV of etching,

100 90 • ... D ...... 80 0 D ._,~ 0 0 g'b • • :=: 70 0 tO • + • o D .t. - 60 66.'{j oc. Cl) • • c,. oP Q) 0) c.~o

10 20 30 40 so 60 70 Clay(%) o P lagioclase • Olivine t:,. Augite .6. Hyperthene a Homblende

Fig. 4. Correlation between the clay CO!IIClll and higher degree of etching of primary minerals I. Yo111ado: Che111icol and Mineralogical Clwrac·rerisrics of Tephras and Volcanic Soils 87

100 • • • • • • • • • •• •• 75 • • • ••• •• • •• • • • • • • • • • ~ • . 13' so .• • •• > • • •• •• • • • • • • • 25 • • • ••• •• • • • • 0 • 0 20 40 60 80 Clay(%}

Fig. 5. Correlation between clay contents and volcanic glass contents

which correspond to highly weathered stages, and ferromagnetic minerals. the clay contents of tephra-derived soils in Tohoku Source of supply of chemical elements; Si02: 6 9 and Kanto, Japan • >. volcanic glass>> plagioclase >> . Ali0.J : Plagioclase, which is the second most abundant volcanic glass» plagioclase. Fe203: volcanic glass, component of tephras next to volcanic glass and is ferromagnetic minerals and in rhyolitic to therefore an important primary mineral in tephras, basaltic andesitic tephras and volcanic glass in basal­ was most susceptible to weathering in the primary tic tephras. MgO: volcanic glass> pyroxenes in crystalline minerals. Olivine was as susceptible as rhyolitic to basaltic andesitic lephras and volcanic plagioclase to weathering. Augite underwent rapid glass in basaltic tephras. CaO: volcanic glass and weaihering over 20% of clay, while hyperthene was plagioclase. Na20: volcanic glass» plagioclase. much more resistant than augite to weathering. Com­ K20: volcanic glass. mon hornblende was the most resistant among these minerals. Ferromagnetic minerals such as ilmenite 5) Mobility of chemical elements and titanomagnetite were more resistant than com­ rt has been generally recognized that not only mon hornblende except for the easy weathering of soil formation, but also soil properties are markedly 12 3 titanomagnetite under poorly drained soil influenced by the rock types of tephras · 1 >. Fig. conditions 15>. 6 shows the relationships between the clay contents Fig. 5 shows the relationship between the volcan­ and the weight ratio of each element to Ali03 in ic glass contents in the light mineral fraction of 0.1 tephra-derivcd soils in Tohoku, Japan. Since AhOJ to 0.2 mm size and the clay content in tephra-derived is the most immobile element during the weathering 11 3 5 soils of Tohoku, Japan >. The volcanic glass con­ of parent materials • >, the ratios of each element tent rapidly decreased with the progression of can be used for determining accurately the mobility weathering. of the chemical elements. The rock types of soil Based on the results obtained, the stability se­ samples were determined by the V-Zn belt methodn. quence of primary minerals in tephras with weather­ It appeared that both Cao and Na20 were highly ing was determined and the major primary minerals mobile and leached out rapidly in tephra-derivcd soils concerned with the source of supply of chemical cle­ under temperate and humid climate conditions. Si02 ments in the tephra were arranged as follows: was leached out gradually. Though Fe20J is report­ 5 Stability sequence; volcanic glass< plagioclase = ed 10 be more mobile than AliOi· >, the mobility olivine< augite< hyperthene ~ common hornblende~ of these elements did not differ appreciably in tephra- 88 JARQ 31 (2) 1997 derived soils under these climate conditions as andesitic samples. y = - 0.003x + 0.26, r = 2 described by Kurashima et al. '. - 0. 774*** in basaltic andesitic and basaltic samples. The mobility of MgO and K20 was more compli­ The equation of the relation between the clay cated. The equation of the relation between the clay contents (x) and K20/ Al203 x I OO(y) in samples of contents(x) and MgO/Al203 x IOO(y) in samples of each rock type was derived as follows: each rock type was derived as follows : y = - 0.008x + 0.091, r = - 0.461 "'** in rhyolitic y = 0.002x + 0.094, r = 0.009 in rhyolitic and and dacitic samples. y = - O.OOOlx + 0.065, r = dacitic samples. y = - 0.002x + 0. 188, r = - 0.077 in - 0.095 in andesitic samples. y = - 0.004x + 0.035, .. . 6 0.6 ..• . • .•• . .• l :...... • •. . • .••• . .. ..:i .. . .:~; .. : ...... M 4 : • • • • .• . ..,0.4 ... . 0 \ . .. : ,: . _ N t •• • •:;.. • •• • ~N ... ~ - . '! .. :. . . . : ·:. ~ ...... : •'. QN '!: • : ,: : • • : '• • ~ • :• • o"' ...... o,N Vl 2 ·-!·~:: .. . :., .. .,...... u. 0.2 ..

O+----~------~---~ 0 20 40 60 80 20 40 60 80 Clay(%) Clay(%)

0.6 .. .. . o.~ 0.4 .. ... o"' ...... 0 : ...... t ...... '. , , ~ 0.2 . .. . r0.2 ::;: ...... ·: .. ..ti:::: . . : . .·: I\.:...... ,•::.. .: . • • ,... •!:, : • .... • •••ti: • • • '•. .. ., . • • • :•. • • •• •• 1 •• • • • • • • •• :. .: : ...... : ...... • •• • !.· :: • ••• 01+------"'I'•._.__...._._• ...... "-' "'----, 20 40 60 80 0 20 40 60 80 Clay(%) Clay(%) 0.3

0.25

0.2 OM ••• _N • '! • ~ 0.1 S •• •, •• 0 • • • 0.15 N • t ' • • ! . .. ·: ~ . ,' ... 0.1 ••• • : : · o"' 0.1 . •, ...... ':.: ,. ~N '• • • • ...... ,:- . O • • •)!. ·: .. o.os '!,(;.N 0.05 .,, 'f \, •' t • '•! ··· •• • : . •, ...... ! .•. . " ~ • • I ...... •. . • • ' \I' •'. 0 .••• ...... 00 20 40 60 80 0 20 40 60 80 CI ay(%) Clay(~)

Fig. 6. Correlation between clay contents and weight ratio of each element to AhOJ of soil samples I. Yamada : Chemical mu/ Mineralogical Characteristics of Tephras and Volcanic Soils 89

r = - 0.404*** in basaltic andesitic and basaltic 25, 104-112. samples. 7) Shoji, S. et al. (1975): Relationships between the MgO and K20 in mafic samples, in which al­ geochemistry of ferromagnetic component and the chemical properties of air-born pyroclastic materials. lophane was the predominant clay minera1 8>, and K20 J. Jpn. Assoc. Mineral. Petrol. Econ. Geo/., 70, in fclsic samples were gradually lost with the progres­ 12-24. sion of weathering, but there were no relationships 8) Shoji, S. et al. (1975) : Chemical and mineralogical between the clay contents and the mobility of these studies of volcanic ashes. 1. Chemical composition elements in other samples. K20 accumulated in the of volcanic ashes and their classification. Soil Sci. Plant Nzar. , 21 , 31 1-318. top layers of non-allophanic Andosols where 2: 1 layer 9) Yamada, I. (1978): Thesis for Ph. D degree submit­ silicate clay predominated and only a small amount 2 ted of Tohoku Univ. [In Japanese). of MgO was removed by leaching l. On the other 10) Yamada, I. (1988): Tephra as parent matetial. /11 hand, MgO was gradually leached out and there was Proc. of the 9th Int. Soil Classification Workshop. no relationship between the clay contents and K20 eds. Kinloch, D. I. et al., 509-519. content in allophanic Andosols where allophane 11) Yamada, I. & Shoji, S. (1975): Relationships between panicle size and mineral composition of volcanic predominated in the clay fraction 4l. ashes. Tohoku J. Agric. Res. , 26, 7 -10. The mobility sequence of the major elements in 12) Yamada, I. & Shoji, S. (1975): Soils formed from the tephra-derived soils under the temperate and hu­ andesitic and basaltic volcanic ashes. 2. Soil proper­ mid climate conditions was proposed as follows : ties and fertilities problems. Tohoku J. Agric. Res., Cao, Na20>Si02>MgO, K20>Ah03, Fe20 3. 26, 102 - 116. 13) Yamada, I. & Shoji, S. (1980): Fertility of Andosols from basaltic and dacitic volcanic ashes. Tohoku J. References Agric. Res., 31, 149-155. 14) Yamada, I. & Shoji, S. (1983): Parent materials of the important Ando soils of Tohoku district. 2. Chem­ I) Kuno, S. (1976) : Volcanic rocks. lwanami, Tokyo ical and morphological properties of volcanic glasses [In Japanese). and relationships between the chemical and minera­ 2) Kurashima, K. ct al. (1981) : Mobilities and related logical properties of tephras and volcanic zones. J. factors of chemical elements in the topsoils of An­ Soil Sci. Plant Nutr., Jpn. , 54, 311 - 318 (In Japanese dosols in Tohoku, Japan. I. Mobility sequence of with English summary]. major chemical elements. Soil Sci., 132, 300-307. 15) Yamada, I. & Shoji, S. (1987): Weathering of 3) Loughnan, F. C. (1969) : Chemica l weathering of the titanomagnetitc in a re,cent rhyolitic ash in Miyagi silicate minerals. American Elsevier, New York, Prefecture, Japan. Soil Sci. Plant Nutr., 33, 493-499. 27-66. 16) Yamada, I. et al. (1980) : Parent materials of the 4) Nanzyo, M. (1993) : Chemical characteristics of important Ando soils of Tohoku district. I. Rock volcanic ash soils. In Volcanic ash soils. Elsevier types and primary mineral composition. J. Soil Sci. Amsterdam, 145 -188. Manure, Jpn., 5 1, 193-202 [In Japanese with English 5) Polynov, B. B. (1937) : The cycle of weathering. summary). Thomas Murby, London (translated by A. Muir), 17) Yamada, I. et al. (1986): Chemical weathering of 160-164. Ando soils in Tohoku, Japan. Abst. 1986 Meeting 6) Shoji, S. et al. (1974): Soils formed from andesitic Jpn. Soc. Soil Sci. Plalll Nutr., 32, 23. and basaltic volcanic ashes. 1. The namre of the par­ ent ashes and soil formation. Tohoku J. Agric. Res., (Received for publication, May 31, 1996)