Two Point-Counting Methods for Modal Analyses of Holocene Sands: Their Comparison and Evaluation

堆積学研究,48号,85-93,1998 J.Sed.Soc.Japan.No.48,85-93,1998

Two point-counting methods for modal analyses of Holocene sands: their comparison and evaluation

Mayumi Nirasawa*, Yoko Minato* and Koji Yagishita*

Two-different point-counting methods of thin-sectioned sands have been argued in evaluat- ing their grain size effects and aims. The traditional method that reflects the dependence of modal composition on grain-size quite differs from the Gazzi-Dickinson method that shows the total independence of grain-size in thin-section analyses. Sets of goals by the two methods do not seem to be the same; the traditional method aims to demonstrate the effect of depositional environments, climates and diagenesis of sediments, whereas the primary target of the Gazzi- Dickinson method is mainly to clarify tectonic settings in provenances. In an attempt to avoid some confusions raised by the two different point-counting methods , we reveal that there will be no significant difference of petrographic data between two methods if we restrict the sand or sandstone samples less than 2.0 phi of mean diameter.

Key words: modal analyses, Gazzi-Dickinson method, traditional method, grain size, thin- section

1978) . Believers of the latter method are the stu- INTRODUCTION dents of William R. Dickinson, who developed his It has been well known that there are two counting method at Department of Geology, methods of point-counting in determing detrital Stanford University. The students of his school modes of sand and sandstone thin-sections maintain that modal composition can be demon- (Ingersoll et al., 1984) . One is the so-called tradi- strated independently of grain size ( e.g., tional method in which polymineralic coarsely Dickinson, 1970, Graham et al., 1976; Dickinson crystalline grains are counted as rock (lithic) and Suczek,1979; Ingersoll et al.,1984) . fragments. The other is known as the Gazzi- In this short paper we attempt to evaluate how Dickinson method in which the assignment of a much grain size can affect on the modes of crystal or a grain (under the cross-hair) within Holocene sand composition in QFL diagrams. a larger lithic fragment to the category of the We compare the modes of composition in testify- crystal or grain, but not to the category of the ing both the traditional and Gazzi-Dickinson larger rock fragment. The former method has methods in terms of grain-size parameter. To been used and developed by many workers. investigate the subject, we use weathered Particularly, one of them is the school of Indiana granitic rocks of so-called "masa" that deposited University, and the students supervised by Lee J. on relatively short-running river beds. Suttner understand well that there is a funda- mental dependence of modal composition on EXPLANATION FOR TWO DIFFERENT grain size (Suttner, 1974; Basu, 1976, Mack, COUNTING METHODS AND SAMPLING PROCEDURE Received : June 12, 1998, Accepted : August 17, 1998 *Department of Geology , Faculty of Education, Iwate University, Morioka, Iwate 020-8550, Japan By the traditional method we generally recog- 86 Mayumi Nirasawa, Yoko Minato and Koji Yagishita 1998

nize individual detrital grains as quartz, plagioclase or lithic fragment. If the plutonic (granitic) rock fragment as shown in Figure 1 is intersected by the cross-hair, we should count as a plutonic rock fragment. Consequently the compositional change is produced as this original detrital grain is altered in size through physical breakdown processes. In contrast, by the Gazzi-Dickinson method the rock fragment generally cannot be assigned to be "the rock fragment" . We should focus our atten- tion only to an individual crystal or grain within the rock fragment under the cross-hair. For example, plutonic rock fragments like Figure 1 Fig.1 Example of a plutonic rock (lithic) frag- are counted as either quartz, K-feldspar, plagio- ment seen in a thin-section made of very clase or mica, but never counted as the rock frag- coarse-grained sand. The grain is counted as a ment. In the case of volcanic rock fragment (e.g., plutonic rock fragment by the traditional method, but it should be counted as a mono- basalt) in the thin-section phenocrysts should be crystalline quartz grain by the Gazzi-Dickinson counted as olivine, augite, plagioclase or whatev- method. Bio: biotite, Qz: Quartz and P1; plagio- er else comes to the cross-hair, but not be count- clase. ed as the volcanic rock fragment. Because of such

Fig.2 Sampling sites for modal analyses of Holocene sediments on stream floors of the Ukedo and Takase Rivers, central Japan. Modal analyses by the traditional method for various sand-size sediments were conducted from site No. 1 to No. 10. Cross mark (X) denotes the sampling site of sediments, for which modal analyses by both the traditional and Gazzi-Dickinson methods were conducted (Fig. 4). J. Sed. Soc. Japan, No. 48 Two point-counting methods for modal analyses 87 controversial two point-counting methods, some Plateau were sampled along two river valleys, the workers tried and acquired modal data using Takase and Ukedo Rivers in Fukushima Pre- both methods (e.g., Baker et al., 1993; Trop and fecture, Japan (Fig. 2) . And the sediments were Ridgway, 1997). sieved into very coarse-, coarse-, medium- and To testify and compare the traditional and fine-grained sand size and impregnated by epoxi Gazzi-Dickinson methods, a number of Holocene resin in film boxes. Thin-sections were made sands of known provenance were collected. from these cylindrical boxes, and all were stained Modern sand-sized sediments derived from plu- by sodium-cobaltinitrite to distinguish K-feldspar tonic igneous (granitic) rocks of the Abukuma from plagioclase and quartz. For each thin-sec-

Table 1. Grain parameters for the two methods 88 Mayumi Nirasawa, Yoko Minato and Koji Yagishita 1998 tion we counted 300 points, using an appropriate zone, and K-Ar ages of the eastern block (97.4 to grid spacing that resulted in coverage of the 126 Ma) are rather older than those of western entire slide. block (85 to 100 Ma, Kubo and Yamamoto, 1990). In an attempt to produce QFL diagrams and Although it is far beyond the scope of this study their daughter diagrams, such as a QmFLt dia- to embark on a systematic characterization of the gram, some grain parameters were set forth source rocks, both intrusive bodies consist main- (Table 1) . In both the traditional and Gazzi- ly of fine-to medium-grained biotite granite, Dickinson methods, we recognize monocrys- medium-grained hornblende-biotite granodiorite talline quartz grains (Qm) . However, we desig- and granodiorite porphyry. Average of some nated only aphanitic quartzose aggregate grains modal analyses from these plutonic rocks are (i.e., chert) as polycrystalline quartz grain (Qp) given in Table 2. The amount of total feldspars by the Gazzi-Dickinson method, whereas all poly- (i.e., K-feldspar plus plagioclase) is generally crystalline quartz grains including chert were more than, or close to 50% in each mode, where- assigned as (Qp) by the traditional method. In as the amount of quartz is ubiquitously much less this thin-section study, however, we have encountered only few chert grains. Magic minerals derived from weathered granitic rocks, such as biotite (Ab) , hornblende (Ah) together with some opaques (Apq) , were assigned as accessory minerals. Ortho- or clino-pyroxene (Apx) derived from small-scale basic intrusive rocks was also assigned as the accessory mineral. Needless to say, these were not counted as con- stitutional detrital grains, when we made QFL diagrams (i.e., excluded from the three major component).

RESULTS AND DISCUSSION

Modal analyses of parent rocks and weathered sands: Rocks making up the Abukuma Plateau Fig.3 Plots of QKPI diagram from parent rocks are granites and granodiorites (Fig. 3). These along river-side bluffs of the Ukedo and Takase plutonic rocks are divided into two major intru- Rivers. Data for this diagram are shown in sive blocks by the NNW-SSE running fracture Table 2.

Table 2. Modal analyses (percentage) of some exposed Branitic rocks along bluffs of the Takase and Ukedo Rivers

6 i ot.=biotite, Hrb.=Hornblende, Opq=opaque, and Mi sc=miscellaneous. 2,000 points per thin-section were counted for the analyses. J. Sed. Soc. Japan, No. 48 Two point-counting methods for modal analyses 89

Table 3. Point-counting data (percentage) for se i ved fractions from Holocene sand (Takase River beds) using traditional method

The sampling localities i n the Takase River are shown i n Figure 2. Some vacant colums in the table means that, because of insufficiently impregnated epoxi resins i n film boxes, the thin-sections were unable to be prepared. 300 points per thin-section were counted for the analyses. 90 Mayumi Nirasawa, Yoko Minato and Koji Yagishita 1998 than 50% and partly less than one-quater of the fragments, but not counted as feldspars. The total mode. Biotite and hornblende are very com- other reason for the modal gap is due to weath- mon maiic minerals, together with a minor ering processes of the parent rocks, in which amount of opaque minerals. destruction of feldspars was substantially caused In contrast to such modal analyses of the par- under the medium humid climate, as revealed by ent rocks, Holocene sands on the stream floor of Suttner (1974) and Basu (1976). the Takase and Ukedo Rivers show that the total Variations in QmFLt diagrams due to two count- amount of feldspars in modes is much less than ing methods: Variations of plots of diagrams due that of the parent rocks (Table 3). The reason to the different point-counting methods are clear- for producing such modal gaps between the par- ly shown in the QmFLt diagram of Figure 4. The ent rocks and the Holocene sands can be modal plots of coarse-grained sand, counting in ascribed to two factors; (1) the point-counting the manner of the traditional method, are plotted method and (2) actual weathering processes. in the central part of the diagram. Using the Because thin-sections of the Holocene sands were same sand-size sample (i.e., the same thin-sec- investigated using the traditional counting tion), however, the plots by the Gazzi-Dickinson method, many feldspars included in larger plu- method occupy the quite different domain, near tonic rock fragments were counted as the rock the edge of QmF line of the diagram. This is obvi-

Table 4. Modal analyses (percentage) by two methods for seived fractions of Holocene sand

The sampling site i s shown as "X" i n the Ukedo River i n Figure 2. Note that there i s no rock fragment and no po Iycrystal line quartz (Qp) by the Gazzi- Di ck i nson method. Op i n terms of the Gazzi -Dickinson method i n this paper suggests the occurrence of chert grain, but the grain does not actually exist in sands on the river beds. 300 points were counted in each thin-section for the analyses. J. Sed. Soc. Japan, No. 48 Two point-counting methods for modal analyses 91

ously because that there should not be many rock As described, the studied Holocene sands were fragments (Lt) by the Gazzi-Dickinson method derived from granitic rocks in the Abukuma (Table 4) . Particularly the absence of plutonic Plateau, and the plots of modal analyses using rock fragments in the method is the major cause the Gazzi-Dickinson method are expected to be by that the plots are shifted toward the QmF line. located in the "intracratonic stable domain" It should be also mentioned that there is no poly- (see figure 2 of Dickinson and Suczek, 1979). crystalline quartzose grain (Qp) in terms of the Thus Figure 5 indicates that, in the case of Gazzi-Dickinson method in this study, because restricted grain size samples (finer than fine- there is no chert grain in treated thin-sections. grained sand), Dickinson's QFL diagram can be Shifts of plots due to the decrease of grain size applied even when we conduct modal analyses using the traditional method: By the traditional employing the traditional method. This means point-counting method, variations of modal com- that in a limited grain-size range the QFL dia- position due to grain size are chiefly caused by gram aquired by the traditional method can be the breakage of large grains into individual crys- used for discussions on tectonic settings in prove- tals or grains. Results using the traditional nances (e.g., Yagishita, 1985). method in Figure 5 clearly show the shift of CONCLUSIONS modal plots from the central part of QFL diagram toward the edge of QF line of the diagram, as the (1) Thin-section studies of weathered Holocene grain-size is decreased from coarse- to fine- sands from granitic rocks in the Abukuma grained sediments. Plateau, central Japan, were conducted using two

Fig.4 QmFLt plots of modal analysis data Fig.5 Shifts of QFL plots due to grain-size vari- obtained by the Gazzi-Dickinson method (G-D) ations by the traditional method of point-count- and the traditional method (TR) . Note that ing. Note that plots in QFL diagram migrate even using the same thin-section, difference of from the center toward the QF line with data between the two methods is very remark- decreasing size of sands. Q: total quartz, F: total able. Sampling site for this Holocene sediment feldspar, and L: total rock fragment. Thin-sec- is the cross mark (X) in Figure 1. Qm: Qm: tions were prepared from ten sampling locali- monocrystalline quartz, F: total feldspar, and ties along the Takase River in Figure 2, and the Lt; total rock fragment including aphanitic poly- data are shown in Table 3. crystalline quartz (e.g., chert). Data for this figure are shown in Table 4. 92 Mayumi Nirasawa, Yoko Minato and Koji Yagishita 1998 different point-counting methods (i.e., the tradi- 2182. tional and Gazzi-Dickinson methods). Graham, S.A., Ingersoll, R.V. and Dickinson, (2) Even when we observe the same thin-sec- W.R., 1976: Common provenance for lithic tion, domains of plots of mode in QmFLt dia- grains in Carboniferous sandstones from grams of Holocene sands quite differ from each Ouachita Mountains and Black Warrior other between the two methods. Basin. Jour. Sediment. Petrol., 46, 620-332. (3) If we restrict the grain size finer than the Ingersoll, R.V., Bullard, T.F., Ford, R.L., Grimm, finer-grained sands, however, the QFL diagram J.P., Pickle, J.D, and Sares, SW. , 1984: The aiming to analyze tectonic settings can be used effect of grain size on detrital modes: A test no matter what point-counting method we of the Gazzi-Dickinson point-counting employ (i.e., either traditional or Gazzi-Dick- method. Jour. Sediment. Petrol., 54,103-116. inson methods). Kubo, K. and Yamamoto, T., 1990: Cretaceous intrusive rocks of the Haramachi district, ACKNOWLEDGEMENTS eastern margin of the Abukuma Mountains - The original manuscript was greatly improved by petrography and K-Ar age. Jour. Geol. Soc. critical comments of an anonymous reviewer. Japan. 96, 731-743 (in Japanese with English His stimulating suggestions are very helpful for abstract). our ongoing studies of sedimentary petrography Mack, G.H., 1978: The survivability of labile light- of Holocene sands. mineral grain in fluvial, aeolian and littoral marine environments: the Permian Cutler REFERENCES and Cedar Mesa Formations, Moab, Utah. Baker, J.C., Fielding, CR. , Caritat, PD. and Sedimetology, 25, 587-604. Wikinson, MM. , 1993: Permian evolution of Suttner, L.J.,1974: Sedimentary petrographic sandstone composition in a complex back- provinces: An evaluation. In Ross, C.A. ed., arc extentional to foreland basin: The Bowen Paleogeographic provinces and provinciality. Basin, eastern Australia. Jour. Sediment. SEPM Spec. Publ. no. 21, 75-84. Petrol., 63, 881-893. Trop, J.M. and Ridgway, K.D., 1997: Petrofacies Basu, A., 1976: Petrology of Holocene fluvial sand and provenance of a Late Cretaceous suture derived from plutonic source rocks: implica- zone thrust-top basin, Cantwell Basin, cen- tions to paleoclimatic interpretation. Jour. tral Alaska Range. Jour. Sediment. Res., 67, Sediment. Petrol., 46, 694-709. 469-485. Dickinson, W.R., 1970: Interpreting detrital Yagishita, K., 1985: Evolution of a provenance as modes of graywacke and arkose. Jour. revealed by petrographic analyses of Creta- Sediment. Petrol., 40, 695-707. ceous formations in the Queen Charlotte ickinson, W.R. and Suczek,D C.A.,1979: Plate tec- Islands, British Columbia, Canada. Sedimen- tonics and sandstone compositions. Am. tology, 32, 671-684. Assoc. Petroleum Geologists Bull., 63, 2164- J. Sed. Soc. Japan, No. 48 Two point-counting methods for modal analyses 93

砂堆積物のモード解析における二つの測定法一比較と評価

韮澤真弓・湊陽子・八木下晃司,1998,堆積学研究,No,48,85-93 Nirasawa, M., Minato, Y. and Yagishita, K., 1998: Two point-counting methods for modal analyses of Holocene sands: their comparison and evaluation. Jour. Sed. Soc. Japan, No. 48, 85-93

薄片のモード比を求める上で,砂堆積物の粒径が測定結果にどのような影響を与えるかを検討した.薄片のモード解析において,粒径によりモード組成が左右される伝統的方法と,粒径による影響を受けないとされるGazzi-Dickinsonの方法とでは測定値に大きな差が生じる. そこで二つの異なるカウント方法により引き起こされる混乱を取り除くには,堆積物の粒径を2.0φ 以下の細粒砂に限定すれば,両者の測定方法によるデータには重要な差異が生じないことを明らかにした.このことは地質時代の砂岩においても,平均粒径が細かいもの(細粒砂以下)であれば,両測定法によって得られるデータにも大差は生じないことを意味しよう.