Analysis of the surface microtextures and morphologies of beach quartz grains in Japan and implications for provenance research

Area I

Area I is located in Shimokita Peninsula, Aomori Prefecture (Fig. 1b). The eastern coastline of area I faces the Pacific Ocean, and the western line faces toward Mutsu Bay. Towada volcano, which is located to the southwest of the sampling sites, erupted at the rate of once every 1000 and 2000 years over the past 10,000 years (Hayakawa 1985) and spread some volcanic ash to the northeast, east, and southeast (Machida and Arai 2003). Towada volcano and the surrounding area are covered with basaltic andesite to rhyodacite with phenocrysts of plagioclase, augite, hypersthene, magnetite, and occasional olivine (Hayakawa 1985; Machida and Arai 2003). This area also includes moderately sized hills covered with volcanic and pyroclastic rocks (Pleistocene to Holocene), along with a few rivers with lengths up to 140 km near the sampling sites. The geological setting of the sampling site is sedimentary rock (Pleistocene to Holocene) (Tsushima 1964; Kamada et al. 1991).

Area II

Area II is located in Niigata and Toyama Prefectures (Fig. 1c). The coastline of this area faces the Sea of Japan, and the southern part of area II contains a large mountainous area with an elevation of over 3000 m, known as the “Japanese Alps.” Steep rivers flow over plains from the Japanese Alps to the Sea of Japan [e.g., the average river slope of the Hime River is 1/80 (Handa et al. 2013)]. The water levels of these rivers are increased by meltwater from the Japanese Alps in the snowmelt season (Fukuda 1939; Fukuda and Sekiguchi 1940). The geological settings of the river basins are complex, and various sediment types such as igneous, sedimentary, and metamorphic rocks cover the area (Sakamoto et al. 1967; Yamada et al. 1989; Takeuchi et al. 1994; Harayama et al. 1996). The Itoigawa–Shizuoka Tectonic Line (ISTL) runs from north to south along the Hime River valley. The eastern side of the ISTL is dominated by sedimentary and volcanic rocks, whereas plutonic and metamorphic rocks cover the western side (Sakamoto et al. 1967; Yamada et al. 1989; Harayama et al. 1996).

Area III

Area III is located in Ibaraki and Chiba Prefectures (Fig. 1d) within the Kanto plain. The eastern coastline is a sandy beach that faces the Pacific Ocean. The Naka River has a length of 150 km and is covered with Quaternary volcanic at upper reach, Paleozoic sediments at middle reach, and Quaternary strata at lower reach (Yamamoto et al. 2000; Yoshioka et al. 2001). The Tone River, which is the second longest river in Japan, has a length of 322 km. The Tone River flows from the mountainous area, which is primarily covered with Cenozoic volcanic and Mesozoic sedimentary rocks, and then through the alluvial plain (Sudo et al. 1991). The Kanto plain is filled with Neogene–Quaternary sediments (Unosawa et al. 1983; Yoshioka et al. 2001).

Area IV

Area IV is located in Izu Peninsula, Shizuoka Prefecture (Fig. 1e). The elevation of the peninsula is less than 1500 m. This area contains no long rivers and is widely covered with Neogene to Quaternary volcanic rocks (e.g., andesite and basalt) (Sugiyama et al. 2010; Takeuchi et al. 2015). Narrow beaches with widths of several tens of meters are developed along the peninsula, as indicated by the GSI Map (Geospatial Information Authority of Japan 2013).

Area V

Area V is located on the island of Shikoku in Tokushima Prefecture (Fig. 1f). The Yoshino River, which has a length of 194 km and flows near the Median Tectonic Line, supplied the beach sediments around the sampling site (Mitsui et al. 1979). The northern side of the river course is covered with sedimentary rock, whereas the southern side is covered with regional metamorphic rock (e.g., green schist) (Isomi 1959; Makimoto et al. 1995; Matsuura et al. 2002).

Analysis of grain size and mineral composition

Particle size distribution was measured using a laser diffraction particle size analyzer (Shimadzu, SALD-3000S), and the mineral composition of each sample was determined using a stereomicroscope. For each sample except fluvial sediment III-4, over 500 sand particles with sizes in ranging from 0.1 to 1 mm were observed and classified into four groups with reference to the method of Ishii (Ishii et al. 2009): “colored minerals,” “magnetic minerals,” “colorless minerals,” and “rock fragments and others.”

Analysis of quartz shape and surface

The dried sand samples were treated chemically to remove carbonates, iron oxides, and organic matter (Vos et al. 2014), and clay, silt, and other small particles attached to the sand grain surfaces were removed by wet sieving using an aperture size of 0.1 mm without ultrasonic vibration (Porter 1962). Quartz grains with sizes in the range of 0.1 to 1 mm were randomly handpicked under a stereomicroscope and placed on SEM stubs. After coating with a layer of platinum (JEOL, JFC-1600), 20 quartz grains for each sample were observed by SEM (JEOL, JSM-6610LV).

Each quartz grain was captured in a secondary electron image, and the quartz grain shape was evaluated based on the SEM images using the free software ImageJ v1.49 (Abràmoff et al. 2004; Schneider et al. 2012). ImageJ was used to determine the shape parameters, including circularity, aspect ratio, and roundness, followed by the calculation of a roundness parameter, R (Takashimizu and Iiyoshi 2016). R parameter denotes circularity corrected by aspect ratio and shows the smoothness of grain outline.

Area I, Aomori

Moving along the coastline, the median grain diameter decreased from Φ0.8 to 2.4 from I-1 to I-6 (Table 2). I-3 contained coarse particles and exhibited poor sorting (Additional file 1: Figure S1). The mineral composition differed greatly between samples; the proportion of colored minerals and magnetic minerals increased in the southern area from 16% (I-1) to 86% (I-6) and from 10% (I-7) to 91% (I-9) (Fig. 2). Most observed colored minerals such as augite and hypersthene exhibited euhedral crystal shapes. The proportions of colorless minerals and rock fragments increased in the northern area. All nine samples in area I contained rounded grain outlines (R > 0.75, Table 2). The surfaces of all samples in this area exhibited a high frequency of v-shaped percussion cracks (> 60%), medium conchoidal fractures (> 45%), and straight steps (> 45%), whereas small holes were rare (< 20%) (Fig. 4).

Area II, Niigata and Toyama

Moving along the coastline, the median grain diameter increased from Φ1.7 to − 0.2 from II-1 to II-3 (Table 2). On the basis of the mineral composition analysis, colorless minerals, rock fragments, and others were dominant (Fig. 2, Table 2). The proportion of colorless minerals increased in the western part of area II from 45% (II-1) to 77% (II-4). Quartz roundness was less than 0.75, and the grain outlines were more angular than those in the other areas (Table 2). A high frequency of v-shaped percussion cracks (70%) was observed in sample II-4, and all samples, except II-1, exhibited high frequencies of small holes (> 70%) (Table 2).

Area III, Ibaraki and Chiba

Although III-2 (Ibaraki) and III-3 (Chiba) were collected from sites separated by only ~ 6 km, their median diameters were markedly different (Table 2). Large amounts of colorless minerals, rock fragments, and others were observed in all three samples from area III (Fig. 2, Table 2). III-3 contained highly rounded quartz grains (Table 2). V-shaped percussion cracks were observed on the grain surfaces in all samples, whereas small holes were rare (< 30%) (Table 2). The number of large conchoidal fractures exceeded the number of small conchoidal fractures in III-3 (Fig. 4).

Area IV, Shizuoka

The median grain diameter of IV-1 was Φ1.3 (Table 2). As in area I, colored and magnetic mineral groups were common (47%) in area IV (Fig. 2). The grains exhibited rounded outlines (Table 2). Medium conchoidal fractures and v-shaped percussion cracks were abundant on the grain surfaces (80%), whereas features related to chemical reactions were sparse. Small holes on the grain surfaces were also rare (Table 2).

Area V, Tokushima

V-1 had a median diameter of Φ1.7 (Table 2) and a high proportion of rock fragments (> 70%), most of which were green schist (Table 2). The quartz grains exhibited spatulate shapes and had the lowest roundness among all areas (Table 2). The grain surfaces in V-1 were characterized by sparse v-shaped percussion cracks (30%), abundant small holes (80%), and abundant small and medium conchoidal fractures (> 70%) (Fig. 4, Table 2). The existing v-shaped percussion cracks were only partially observed and existed mainly at the grain edges, and the area of surface coverage was low.

Grain size and mineral composition

Grain size and mineral composition differed markedly between sampling areas. Thus, the analyses can be used to estimate sediment provenance in some areas. In area I (Aomori), grain size and mineral composition showed distinct trends with location along the coastline (Fig. 2, Table 2). Mineral composition trend is considered to be attributed by the presence of volcano. This area is covered with terrace deposit and volcanic rocks (Yoshida 2017). Grains with β-quartz shapes were contained in samples except for I-4 and I-6. Quartz occurred as euhedral phenocrysts in such as rapidly cooled extrusive and hypabyssal rocks (Gribble and Hall 1992). By microscopic observation, samples taken from the southern area contained abundant magnetic minerals and colored minerals such as augite and hypersthene, which is also contained in tephra of Towada volcano (Hayakawa 1985; Machida and Arai 2003). Although a more detailed investigation is needed to clearly identify the provenance of colored and magnetic minerals in area I, coastal sand in the southern part of area I were more strongly influenced by pyroclastic deposits of volcano. The change in the proportion of heavy minerals along the coastline may indicate the direction of coastal drift. The direction of sand drift in this area is north near site I-3 to I-6, no direction near site I-2 and south near site I-1 (Sato et al. 1972). However, in this study, the median diameter of the samples was larger in the northern part of area I than in the southern part. Thus, sand drift does not likely explain the observed differences in mineral composition in area I. So what caused variation in grain size distributions? It appears that many factors relate to grain-size trend such as offshore bathymetry, wave energy, sediment composition, and sand grain size in parent rock. It is hard to decide the main cause by the result of mineral composition and quartz morphology, so other data is needed to explain the grain-size trend in this area.

In area III (Ibaraki and Chiba), although the sampling locations of III-2 (Ibaraki) and III-3 (Chiba) were close to one another (6 km apart), these samples had completely different grain sizes (Table 2). For comparison, the median diameter of the fluvial sediments III-4, 8 km upstream of the mouth of the Tone River, was larger than III-2 (Table 2). Sato et al. (2000) and Uda et al. (2007a) reported that beach sand at site III-2 was supplied from the Tone River. Conversely, grain size at III-3 was larger than fluvial sediments III-4 and the coastal line shows a natural pocket beach. In consideration of the grain size and geomorphological setting, III-2 was likely derived from a fluvial supply, whereas the origin of III-3 was not fluvial. That is, III-2 and III-3 originated from different sources.

In area V (Tokushima), the types of rock fragments found in the samples also provide clues as to the origins of the quartz grains. V-1 contained a large number of rock fragments, most of which were identified as green schist (Table 2). Green schist covers a wide area of the drainage basin of Yoshino River, which flows near site V-1 (Isomi 1959; Makimoto et al. 1995; Matsuura et al. 2002). This suggests that sediments, including fragments of green schist, were transported from the upstream area of Yoshino River and deposited on the beach.

Quartz shape

In the classification table of Vos (Vos et al. 2014), grain outline is divided into three groups based on appearance: angular, subangular, and rounded. We attempted to quantify grain outline in this study. Quartz outline was evaluated based on a new roundness parameter calculated in ImageJ (Takashimizu and Iiyoshi 2016). This parameter accurately reflects the quartz outline observed in SEM images and provides a consistent interpretation of sediment supply and parent rock, as detailed below.

The samples from areas II (Niigata and Toyama) and V (Tokushima) exhibited roundness values under 0.75 and relatively angular grain outlines (Fig. 3d, e, i) compared to the samples from the other areas. In area II (Niigata and Toyama), where steep rivers flow, grains are thought to be transported in high-energy subaqueous environments without edge rounding (Helland and Holmes 1997), resulting in angular outlines. V-1 (Tokushima) was also influenced by river supply (Yoshino River) and showed the lowest roundness among all samples. Pyökäri (1997) also stated that roundness of sediment grains at beach discharged from river was poor. Another reason for V-1 showing angular outline was the parent rock. V-1 is likely derived from green schist, containing irregular shaped oblate quartz (Nishikawa et al. 1994). Thus, the angular quartz outline of V-1 may have resulted from regional metamorphism.

On the other hand, I-1 through I-9 (Aomori), III-3 (Chiba), and IV-1 (Shizuoka) contained relatively rounded quartz grains (R > 0.75) partly because the corresponding sediments were not influenced by a large river and were subjected to long periods in coastal environments. Thus, the rounded grain outlines likely resulted from long transport distance and time or large amount of recycled sediments (Madhavaraju et al. 2009). Alternatively, the rounded grain outline may reflect the original grain shape generated by the parent rock. As indicated in previous studies, grain outline is related to both the transportation process and the original grain shape of the source rock (Goudie and Watson 1981; Kleesment 2009; Costa et al. 2013). In this study, the geological settings of sampling area containing rounded quartz grain are volcanic rock or ash. The samples in area I (Aomori) were influenced by the volcano, and most of area IV (Shizuoka) is covered with volcanic sediment (Hayakawa 1985; Machida and Arai 2003; Sugiyama et al. 2010; Takeuchi et al. 2015). III-3 (Chiba) contains grains with β-quartz shapes which are thought to be influenced by volcanic rock (Gribble and Hall 1992). The rounded quartz grains observed in this study may be attributed to the original shapes of the volcanic rocks, along with the depositional environment.

Quartz surface features

Quartz grains from coastal environments are characterized by subangular to rounded outlines, silica precipitation, oriented etch pits, and v-shaped percussion cracks (Krinsley and Doornkamp 1973; Higgs 1979). In this study, features produced by pressure or powerful impacts on grains such as flat cleavage face, small to large conchoidal fractures, arcuate steps, and straight steps were common in all samples, whereas features produced by chemical reaction were sparse. This microtextural trend corresponded to the neither nearshore nor dune but beach sample in Costa’s model, namely fresh surfaces and percussion marks are dominant features (Costa et al. 2012). Moreover, Margolis (1968) reported that sand grains from beaches with low wave activity (mean wave height, 0–0.1 m) exhibit features produced by chemical reaction, whereas grains from high-energy beaches (mean wave height, 0.5 m<) show mechanical features. Mean wave heights of studied areas in 2016 were approximately 0.7 m for area I Aomori; 0.3–0.6 m for area II, Niigata and Toyama; 0.8 m for area III, Ibaraki and Chiba; 0.6 m for area IV, Shizuoka; and 0.3 m for area V, Tokushima (Nationwide Ocean Wave information network for Ports and HArbourS, NOWPHAS 2016). This suggests that all 18 samples in this study were subjected to moderate to high energy beach conditions.

In the course of this study, 15 types of surface microtextures were observed, with large differences in microtexture frequency observed between two of them: v-shaped percussion cracks and small holes (Fig. 4, Table 2). Conchoidal fractures (Nos.11–13) and steps (Nos.14–15) could not be used because they were very common among all the samples. The size of conchoidal fractures may not be related to the sedimentary history but were related to the grain size (Fig. 4, Table 2). Other microtextures except v-shaped percussion cracks and small holes could not be used to interpret the sedimentary history clearly because they were rare frequency or no definitive differences in all samples (Fig. 4). Therefore, two surface features are considered to be characteristic features that reflect the diversity of the beach sand samples, as discussed in detail below.

V-shaped percussion cracks

In the subaqueous environment, v-shaped percussion cracks are produced by grain-to-grain collisions (Krinsley and Doornkamp 1973; Margolis and Krinsley 1974; Mahaney and Kalm 2000). In this study, the observed differences in the frequency of this feature are thought to show the transportation process. No major river mouth is located near the sampling sites in areas I (Aomori) and IV (Shizuoka), where the grains have many percussion cracks; by contrast, sites II-1 through II-3 (Niigata and Toyama) and V-1 (Tokushima), in which the grains exhibited few percussion cracks, were located near the mouths of rivers. This correlation is discussed below in terms of the geomorphological settings of Japan. Compared to rivers in other parts of the world, most rivers in Japan are short and steep, and the sediment yields from steep watersheds are much larger than those from less steep watersheds (Ohmori 1983; Oguchi et al. 2001). Fluvial sediments in Japan are thought to have short transport times in the subaqueous environment before being deposited on the coast. II-1 to II-3 and V-1 did not have quartz grains with marine characteristics (i.e., rounded outlines, medium relief, silica precipitation, and v-shaped percussion cracks) (Krinsley and Doornkamp 1973; Higgs 1979; Vos et al. 2014). These samples, which were collected near a river mouth, are thought to have been influenced by river supply. The sediments corresponding to these samples were only located on the coast for a short time as a result of the large amount of sediment supply from rivers (0.9~6.2 × 10⁶ m³/year, Udo et al. 2016); thus, there was no sufficient time to produce percussion cracks on their grain surfaces in the coastal environment. This interpretation is consistent with the findings of Margolis and Kennett (1971), who found that v-shaped percussion cracks generally occur on less than 50% of grains of fluvial origin and on more than 50% of grains from high-energy beaches.

However, grains from some sampling sites near a river mouth had abundant percussion cracks. For example, III-3 (Chiba), which was collected from approximately 4 km away from the mouth of the Tone River, had abundant percussion cracks (95%; Table 2) and exhibited marine characteristics. III-3 was not supplied from the Tone River as described in discussion of grain size and mineral composition section. The coastline at site III-3 shows a natural pocket beach, and the grains may have been subjected to extended periods in the coastal environment, resulting in the percussion cracks. The environment of site III-1 resembles to that of site III-3. Site III-1 (Ibaraki) is approximately 6 km away from the mouth of the Naka River. However, considering the southward sand drift at river mouth (Uda et al. 2007b) and coastline of pocket beach, III-1 was not influenced by river supply and showed abundant percussion cracks (95%: Table 2) like III-3.

In conclusion, in this study, the differences in the frequency of v-shaped percussion cracks on quartz grain surfaces observed between the samples are related to transportation process (i.e., whether the sediments at the sampling sites were primarily influenced by fluvial supply or not).

Small and large holes

In this study, the small holes on the quartz grain surfaces, which were sometimes aligned on a fresh plane with a diameter of approximately 1 μm, were derived from the remnants of small inclusions in the grain (Le Ribault 1977; Higgs 1979). Small volumes of liquid or gas inclusions within a crystal were frequently observed by optical microscopy in grains with small holes on the surface. The sizes of the small holes observed by SEM agreed with those determined under optical microscope (~ 5 μm).

The small holes observed in this study were derived from the remnants of small inclusions rather than by mechanical forces, like the v-shaped percussion cracks. Inclusions are trapped during the growth of the host crystal or during the healing of fractures in the host crystal after crystallization. The presence of inclusions indicates the presence of fluid during the initial crystal growth or subsequent fracture healing (Roedder 2003). Many small inclusions are generated when the crystal grows rapidly (Sorby 1858; Takenouchi 1975). Inclusion was used to get information of geology and geochemistry, for example, the search for ore deposits (Roedder 2003). Considering the mechanisms of inclusion trapping, the frequency of small holes on the quartz grain surface may provide information about the geologic source of the quartz. All nine samples collected in area I (Aomori) had low frequencies of small holes (< 20%), indicating that the quartz grains in these samples may be derived from similar geological provenances. In area II (Niigata and Toyama), II-1 showed a lower frequency of small holes (Table 2) than the other samples (II-2 through II-4). This may indicate that the sediments at site II-1, which is located on the eastern side of the ISTL, have a different geological provenance from the sediments at the other sampling sites.

Large holes with diameter > 10 μm also seemed to be derived from large inclusions (Fig. 3p). Large inclusions are generated when the crystal grows slowly under high temperature (Sorby 1858; Takenouchi 1975). On the basis of the stereomicroscopic observations, the samples collected in areas I (Aomori), III (Ibaraki and Chiba), and IV (Shizuoka) contained some grains with large clear or gray fluid inclusions. Thus, the frequencies of the large holes also have the potential to provide information about geological provenance; however, the differences in the frequencies of large holes observed in this study (0% to 40%) were lower than those of small holes. Inclusion size is thought to be related to inclusion abundance; inclusions with size < 10 μm are one or two orders of magnitude more abundant than those with size > 10 μm (Roedder 2003). Thus, in MQS analysis, the more frequently observed small holes may be more helpful for the study of provenance compared to large holes because of the number of inclusions remnant on surface.