1 Introduction

New Zealand (NZ) is an island country in the southwestern Pacific Ocean comprising two major land masses—the North Island and the South Island. The primary volcanic activity is concentrated in the North Island, which runs from Bay of Islands, Auckland (Auckland Volcanic Field AVF), Rotorua–Taupo (Taupo Volcanic Zone TVZ), and Taranaki/Egmont E (Lowe and Balks, 2019). While the high activity of TVZ is subduction-driven, the lesser activity of AVF is due to feeding from a magmatic region. The central North Island, particularly TVZ, has witnessed the biggest eruptions in this country—the most destructive NZ eruption of recent times being Mount Tarawera in 1886. On the other hand, the AVF, covering much of the metropolitan area of Auckland, includes existing 53 monogenetic volcanic centers but the field is still active (Lindsay et al. 2011).

Tephra refers to the solid constituent contained in volcanic gas in an explosive volcanic eruption. It can be broadly classified into two groups—‘tephra fall deposit’ and ‘pyroclastic flow deposit’. The tephra fall deposit is the depositing product of the volcanic ash and pumice carried on in the wind during the Plinian (explosive column rising over the volcano) eruption. As the volcanic deposits constitute a variety of grain sizes and densities, they fall at relative distances from the volcano. Hence, the tephra fall deposits usually have a uniform grain size distribution. In contrast, the pyroclastic flow deposit is provided by the volcanic column containing pumice, volcanic ash, and scoria yielding to gravity action and further flowing above the ground. These kinds of deposits—‘tephra airfall’ and ‘pyroclastic flow’ deposits—are abundantly found in Rotorua and Auckland regions due to the high volcanism observed in these regions. For a city such as Auckland, which is home to 1.5 million people, the management of such deposits becomes even more critical to restore the functionality of the area in case of an eruption (Hayes et al. 2017).

Many of today’s TVZ and AVF soils comprise weathered tephra materials that originated from previous airfall events; with future eruptions tending to produce newer identical deposits. These would range from fine-grained ashes to granular scoria or pumice. Because of abundance and economic considerations, these airfall deposits have been used in the past and will be used in the future in geotechnical engineering applications, for example as structural backfills or land reclamation fills. However, understating their response when subjected to extreme rain and flood events due to climate change, and associated potential accelerated weathering processes in combination with earthquake conditions, is of paramount importance before their use as structural fills can be further recommended. To do so, detailed laboratory investigations of their physical, chemical, compaction, and geotechnical engineering properties (strength, compressibility, collapsibility, liquefaction potential, etc.) are necessary.

As also these volcanic deposits are products of different times and eruptions, these might differ based on their physical and chemical properties (Pathiranagei et al. 2023). Accordingly, this paper provides useful insights from physical (grain size, specific gravity, and morphology), chemical (elemental and mineralogy using X-ray fluorescence and X-ray diffraction), and compaction tests (maximum dry density, optimum water content, and particle breakage) for selected volcanic tephra samples sourced from the North Island of New Zealand in the Rotorua, Taupo and Auckland regions. As a part of a detailed laboratory investigation, relationships between physio–chemical and compaction parameters were first established, in this paper, as they provide a preliminary direction for their selection as backfilling materials for geotechnical applications. The shear strength, one-dimensional compressibility, collapsibility due to water inundation, and liquefaction potential of such materials have been also assessed; their correlations with the physio–chemical and compaction properties will be presented in due course in companion papers by the authors focusing on detailed geotechnical aspects of compacted tephras.

2 Geological Background and Sampled Tephras

This study focuses on the physical, chemical, and compaction properties of ‘recent’, ‘young’, and ‘older’ tephras formed approximately 135, 705, and 140 ka years ago in the TVZ region along with ‘older’ depositions dating back to 200–260 ka and 35–180 ka years in the AVF. The TVZ is a volcanic zone in the North Island of New Zealand and has been highly active for the past 2 Ma. It runs offshore from the Bay of Plenty in the north to Mt. Ruapehu in the south (Fig. 1) including the most productive calderas in the world—Okataina and Taupo (Lowe 1990). The AVF, on the other hand, is a small-volume volcanic field within New Zealand’s largest city, Auckland. Considering different geological sites, eruption times and possible weathering effects, the differences between the sampled tephras, thus, need thorough investigation.

Fig. 1
figure 1

Map showing TVZ comprising calderas and AVF containing volcanic centers with collective locations of sampled tephras (derived from Lowe 1990)

Full size image

The tephras under consideration in this study were from the following different eruptions: Tarawera 1886 and Kaharoa 1315 (Okataina caldera in TVZ), Taupo 140 ka (Taupo caldera in TVZ), Pupuke 200 ka and Maungataketake 85–200 ka (AVF). In NZ’s geological history, the Tarawera and Kaharoa tephra samples come under very young (or recent) tephra and young tephra, respectively.

When a large quantity of tephra is deposited, the soil developed on it can be classified as a recent deposit such as Tarawera ash and lapilli, and Rotomahana mud from the 1886 ‘basaltic’ eruption (Rijkse 1974). The 1315 Kaharoa eruption, instead, from Mt. Tarawera is the youngest ‘rhyolitic’ eruption and the largest in terms of volumetric deposition in the last millennium in NZ (Lowe and Balks 2019). The eruption triggered approximately 5 km3 of rhyolitic deposits of block and ash flows to the north, southeast, and northwest of the volcano. The deposit namely, Kaharoa ash, was a coarse ash, with abundant blocks of pumiceous rhyolite, dacites, obsidian, and devitrified rhyolite. Blocks of granodiorite and granite were also found in them (Ewart and Cole 1967; Nairn et al. 2004). Nonetheless, the erupted products were mainly pyroclastic with a total volume of deposited materials of around 1 km3 (Nairn 1979). In addition to the basaltic eruption, the eruption along the surrounding areas was explosive enough to pulverize lake sediments and country rhyolitic rocks resulting in ‘fall’ and ‘surge’ deposits known as the Rotomahana Mud.

The Taupo, Pupuke and Maungataketake basaltic samples were comparatively ‘older’ samples with formation dating to around 140 ka, 200–260 ka, and 35–180 ka years ago (Brown et al. 1994; Lindsay et al. 2011). Rhyolite is the predominant material (~ 98%) at Taupo from the past eruptions originating 300 ka ago. Although basalt is rare in Taupo, it has erupted from several vents in the region close to Lake Taupo resulting in ‘red’ scoria deposits of two types—‘drier cones’ farther from the lake and ‘wetter tuff’ deposits nearer to the lake (Brown et al. 1994).

Pupuke, considered to be the first volcano marking the evolution of AVF, generated ash-lapilli (gravel) rich beds (overlaying a fine tuff bed) near the crater rim of the lake. The volume of volcanic deposits was estimated to be around 0.07 km3. Lastly, Maungataketake, a small-volume AVF volcano, showcased pheratomagmatism along with wet violent base surges. In the distal areas from the vent (as in our case), the beds were fine-grained with thin beds of angular and glassy lapilli beds (Flores et al. 2014). Eruptions like wet base surges are likely to occur in the future in Auckland, hence their knowledge of mechanical behavior can be of significant importance.

As such in this study, the areas of tephras’ collection were central and northern North Island, and included Rotorua and Taupo townships as well as Auckland city. Heterogeneous tephra samples were collected from the eruption zones, as shown in Fig. 1, and named accordingly: Kaharoa Ash (KA), Tarawera Basalt (TB), Rotomahana Mud (RM), Taupo Scoria (TS), Pupuke Basalt (PB) and Maungataketake Ash (MA). The tephras, thus, investigated included a wide array of soil types from clayey sands, and silty sands to gravelly sands as illustrated below through physical and chemical characterizations.

3 Physical and Chemical Characterization

To assess the suitability as backfilling materials of the NZ tephra samples collected, firstly conventional laboratory tests were carried out, including both physical (particle size distribution, specific gravity, Scanning Electron Microscope (SEM)) and chemical (X-ray Fluorescence (XRF) and X-ray Diffraction (XRD) analyses characterizations (Rendón et al. 2020). All the materials were oven-dried at 105 °C for 24 h before carrying out the tests. The results of these are presented henceforth.

3.1 Physical Properties and Morphology (SEM)

The gradation characteristics were examined by performing sieving and hydrometer analyses (ASTM 2017a, 2017b) on three samples for each tephra type. The deposits contained particle sizes up to 50 mm. For laboratory testing purposes, only the particles passing through the 8 mm sieve were used. The particle size distribution (PSD) curves were averaged out and representative PSD curves of each tephra type are presented in Fig. 2. Based on the Unified Soil Classification System (USCS), the White Kaharoa Ash (WKA), Black Pumiceous Ashes (BPA1 and BPA2), Golden Kaharoa Ash (GKA), and Tarawera Basalt with Kaharoa Ash (TBKA) samples classified as well-graded silty sands (SW-SM). The Taupo Scoria (TS) sample is classified as well-graded gravel (GW). The Tarawera Basalts (TB1 and TB2) are categorized into poorly-graded silty sand (SM) and silty gravels (GM) respectively. The Pupuke Basalt (PB) is designated as a poorly graded sand (SP) with an almost negligible amount of fines. The Rotomahana Mud (RM) tephra is classified as a low-plasticity clayey-sand (CL) with liquid limit LL = 36.8%, plastic limit PL = 23.8%, and plastic index PI = 13.0%; along with the Maungataketake Ash (MA) as sandy silt (ML) sample with LL = 30.1%, PL = 26.4% and PI = 3.7% (ASTM 2017c). Being base surge deposits, both RM and MA samples exhibited low plasticity. Except for RM and MA samples, the remaining samples contained non-plastic fines. The index properties are as such reported in Table 1.

Fig. 2
figure 2

Tested particle size distributions of collected tephra samples

Full size image
Table 1 Index properties of tephra samples
Full size table

Based on the SEM results carried out at 150× magnification and 100 μm scaling (Fig. 3) on 0.075 mm fraction size, the texture of the tephras categorized as vesicular type containing voids within and on the surface (Gobin et al. 2023). The rhyolitic WKA and BPA samples contained fibrous, glassy, and highly angular particles enclosing elongated voids (‘stretch marks’ or ‘layers’ type of voids). The TB and TS samples, though both basaltic in origin, also differed in terms of their shape and appearance. TB pertained a subangular blocky, sponge-like structure with larger voids on a rough surface (Plinian type deposit produced after the significant rise of pulverized gas and ashes); while the TS showcased angular, sharp-edged and jagged topography with uneven distribution of voids (Strombolian type deposit generated as short bursts of lava near vent). The RM sample had a dusky-chalky (softer) appearance due to hydrothermal alteration (Lowe and Balks 2019; Heap and Violay 2021) of previous country rocks and rhyolitic rocks with very small voids on the surface.

Fig. 3
figure 3

Morphology (shape, structure and voids) of selected tephras

Full size image

As indicated through the visual observations and SEM images, owing to the presence of internal voids and different types of tephra in the bulk sample, the specific gravity (Gs) might be different among individual particles of a particular size as well as different fraction sizes of a particular tephra. The estimation of Gs was, therefore, made for all fractions (Table 2) by water pycnometer and vacuum application (ASTM 2014). The pycnometer was continually agitated under a vacuum for at least 2 h to ensure suspension of tephra and removal of air from within.

Table 2 Gs values of investigated fractions of all tephras
Full size table

Over the investigated range of particle sizes for the tephras, the Gs values for WKA, BPA2, PB, and TS samples showed the greatest change with the variation in particle size—i.e., the trend followed a decreasing pattern with the increase in particle size (Fig. 4), with the highest percentage decrease for WKA (25.8%), then BPA2 (22.1%), followed by PB (15.2%), TS (13.9%) and MA (3.2%). The reason behind this could be attributed to the smaller amount of internal voids into which water could not penetrate upon the decrease in particle size to around 0.075 mm (Wesley 2001; Cecconi et al. 2010; Stringer 2019).

Fig. 4
figure 4

Specific gravity values of investigated fractions of tephras

Full size image

On the other hand, the rhyolitic GKA sample showed a marginal decrement of 1.0% in Gs in comparison to the other rhyolitic WKA and BPA2 samples. The basaltic PB, TS, and MA samples showed a lower decrement in Gs values with the increase in particle size than the rhyolitic WKA and BPA2 samples. This could be further attributed to the nature of the formation of different tephras with rhyolitic being more gaseous than basaltic, therefore, leading to more internal voids. On the contrary, the Gs values of the RM sample appeared to showcase an increasing trend from 0.15 to 2.00 mm due to intermixed basaltic fractions.

Some irregular trends were also observed for basaltic TB1, TB2, and intermixed BPA1, TBKA samples. For TB2, the Gs showed an increase from 6.30 mm to 0.60 mm particle size, and thereafter a decrease from 0.60 mm up to fines. There appeared to be a similar trend for TB1, though the increase in Gs was apparent up to 1.18 mm, below which the Gs seemed to decrease. Similarly, for TBKA and BPA1, the increase in Gs was apparent only from 6.30 mm to 2.00 mm particle size, below which Gs reduced considerably. Also, the Gs values of TBKA were comparatively higher than BPA1 (and similar to TB1 and TB2) in the range between 6.30 and 2.00 mm, indicating definite intermixing between rhyolitic and basaltic components. Contrary to BPA2, BPA1 also seemed to include basaltic fractions in it (as evident from higher Gs values in the range 6.30–0.60 mm).

Intermixing among different types of deposits is a common occurrence due to the flow and redeposition of tephra deposits. Due to intermixing, the Gs of the bulk sample is different from the Gs of the individual fractions of a certain tephra deposit. Therefore, the Gs of the bulk sample were taken into account and calculated as:

$${G}_{{\text{s}}}=\sum\limits_{{\text{i}}=1}^{9}{G}_{{\text{si}}} . {x}_{{\text{i}}}$$
(1)

where Gsi is the specific gravity value of a certain grain size and xi is the percentage of tephra particles that have Gsi.

To obtain the best possible estimate of Gs of tephra deposits, a comparison was made between original fines, fines produced by crushing bulk sample (below 8.00 mm), and bulk sample (Fig. 5). For WKA, BPA2, and TS samples, the bulk sample showed the lowest Gs values and nearly identical Gs values evaluated out for original and crushed bulk fines. The latter similarity indicated identical types of tephra deposits in these samples. The GKA sample, in comparison to WKA, showed lower Gs for original bulk fines due to reworked fines (below Rotomahana mud) and finer-grained composition.

Fig. 5
figure 5

A comparison of Gs of (i)—original bulk fines, (ii) bulk crushed into fines and (iii) bulk sample as a whole

Full size image

For BPA1 and TBKA, the Gs values seemed to be close-ranged, indicating the dominant role played by a higher percentage rhyolitic component present in particle size < 2.00 mm suppressing higher Gs of lower percentage > 2.00 mm basaltic fractions. Likewise, the RM and MA samples also seemed to have slight differences, although the presence of slightly plastic fines played a major role in deciding closer Gs values here. Lastly, the crushed fines of bulk samples TB1, TB2, and PB presented significantly higher Gs values (11.0, 7.5, and 7.8% respectively) than original bulk fines due to the predominance of gravelly basalt in their gradation.

Overall, based on all the Gs results, the majority of the tephras could be classified as light-weight materials except for TS, PB, and MA samples which showcased Gs values higher (Aldeeky and Al Hattamleh 2018; Francisca and Bogado 2019) than 2.682 (Gs of hard-grained quartz geomaterials such as New Brighton beach sand; Sood et al. 2022a). The lower Gs values could be attributed to the abundance of lighter minerals with a range of 70–95% (Yamada 1988). Within the minerals category, the relative presence can be arranged as non-colored volcanic glass >  > plagioclase feldspars >  > silica minerals (quartz, cristobalite, and tridymite) ≈ mica. These minerals can have a further influence on the mechanical characteristics and therefore need quantification. The relative quantification of these minerals in respective tephras is described next in the chemical properties section.

3.2 Chemical Properties: Elemental State and Degree of Weathering (XRF) and Mineralogical Composition (XRD)

3.2.1 Elemental State and Degree of Weathering by X-ray Fluorescence (XRF)

To establish chemical configurations, the X-ray fluorescence (XRF) analyses were performed firstly for major oxide elemental composition (Table 3) and loss of ignition (LOI) determinations respectively. The LOI tests were carried out by heating the samples to 1000 °C. The studied tephras constituted major element oxides ranging between 45.73 and 75.10% SiO2, 0.15–2.25% TiO2, 11.72–17.23% Al2O3, 1.30–13.82% Fe2O3, 0.06–0.17 MnO, 0.23–13.23% MgO, 1.09–12.09% CaO, 2.26–3.90% Na2O, 0.40–3.38% K2O, 0.04–0.50% P2O5 and 0.20–8.11% LOI values indicating a wide variation from basalts to rhyolites. Also, as evident from the major element oxides of TBKAm (1.18 mm—fines) and TBKAf (< 0.075mm) in comparison to TBKAc (6.30–2.00 mm), the TBKA sample showed signs of intermixing with Kaharoa ash as the major component (89.8%—TBKAm and TBKAf) and Tarawera basalt as the minor component (11.3%—TBKAc).

Table 3 Major oxide analyses (XRF) of tephra samples
Full size table

It is desired that the tephra sample to be used as land reclamation material can withstand the external atmospheric conditions for a long time and not decompose readily after a shorter period. After the release of tephra into the atmosphere, significant changes in mineral chemistry occur, which ultimately result in weathering. Depending on the nature of deposition and accompanied weathering changes (Rendón et al. 2020), elements such as Si, Al, Fe, Na, K, Ca, and Mg in tephras get depleted. To measure the amount of weathering induced, several weathering indices have been developed by different researchers using the major elemental composition (Table 4).

Table 4 List of weathering indices studied
Full size table

A particular weathering index measures the relative depletion of mobile elements to immobile elements. Mobile-to-immobile ratios are mostly used in many indices. The major oxides commonly known as alkali and alkaline oxides SiO2, Na2O, K2O, CaO, and MgO are considered to be mobile (decreasing) while the remaining major oxides Al2O3, Fe2O3, and TiO2 (known as sesquioxides) are considered to be immobile (constant). Therefore, these weathering indices have been accordingly used and described below along with their corresponding values shown in Fig. 6.

Fig. 6
figure 6

Values of studied weathering indices for all tephras

Full size image

The Ruxton Ratio, RR (Ruxton 1968), keeps aluminium immobile during the weathering process. The RR for a fresh sample is > 10, and this value decreases with the increase of weathering. The RR values evaluated out to be in the range 2.9–6.1 indicating higher weathering for basaltic TS (2.9) and PB (3.9), andesite basaltic TBKAc (3.2) and TB (3.3–3.6) samples; followed by intermediate weathering for dacitic RM (4.6) and basalt andesitic MA (4.7) deposits and relatively lower for rhyolitic and dacitic GKA (5.1), TBKA—m and f (~ 5.6), BPA (~ 5.9) and WKA (6.1) samples.

The Weathering Index of Parker, WIP (Parker 1970), used for silicate rocks is based on the proportions of alkali and alkaline earth metals (Na, K, Ca, and Mg) remaining after weathering. For a given rock or tephra, the optimum fresh and weathered values of WIP are > 100 and 0 respectively. The decreasing order of weathering pattern by WIP was as follows: RM ≈ TB2 ≈ TS (4040.9–4068.7) > TB1 ≈ TBKAc (4117.2–4148.7) > MA (4230.3) > GKA (4443.0) > TBKAm ≈ TBKAf (4561.8–4597.6) > BPA1 (4601.5) > PB (4796.5) > BPA2 (4893.8) > WKA (5116.7). Our WIP data were in the zonal range of those published by Fiantis et al. (2010) (~ 3800) and much higher to those published by Souri et al., (2006) (2.8–11.0). Previously, Price and Velbel (2003) reported that the WIP of felsic metamorphic rocks was 2625–15,079. The seemingly low WIP values for RM, MA, and GKA samples could be explained by their phreatomagmatic and hydrothermal alterations (Flores et al. 2014; Lowe and Balks 2019; Heap and Violay 2021), making these samples to be altered by weathering processes. Accordingly among all the samples, the WKA was the least altered and TS the most altered.

The Index of Laterization, IOL (Schellmann 1981), based on sesquioxides (Al2O3 + Fe2O3) to SiO2 mass ratio showed the following decreasing order of weathering intensity: TS ≈ PB (35.8–36.1) > TBKAc ≈ TB2 (31.7–32.6) > TB1 (29.2) > MA (26.5) > RM (21.6) > GKA (18.1) > TBKAm ≈ TBKAf (17.3–17.5) > WKA ≈ BPA1 ≈ BPA2 (15.3–16.6). The trend suggested granular Taupo and Auckland basalts to be most weathered and the rhyolitic and dacitic TBKAm, TBKAf, WKA, and BPA samples to be the least weathered.

The Chemical Index of Alteration, CIA (Nesbit and Young 1982), measures the extent of conversion of feldspars to clayey minerals such as kaolinite. The CIA values for fresh tephras are ≤ 50, and these values are ≥ 100 for heavily weathered tephras (Fontoura et al. 2023). The CIA values ranged from 50.0 to 66.0 for the tephras, indicating that the samples were slightly weathered. In the A–CN–K diagram and SiO2–Al2O3–Fe2O3 plot (Fig. 7), the samples were located in the slightly weathered region and near the SiO2 portion respectively, suggesting that the samples were fresh or very less weathered and nearing kaolinization indicating the beginning stages of weathering.

Fig. 7
figure 7

a A–CN–K and (A—Al2O3; C—CaO; N—NaO; K—K2O; Plag.—Plagioclase; K-Feld—Potassium Feldspar; Biot.—Biotite; Mus.—Muscovite; Ill—Illite; Smec. Smectite; Kao.—Kaolin; Chl.—Chlorite and Gibb.—Gibbsite), b S–A–F (S—SiO2; A—Al2O3 and F—Fe2O3) ternary plots

Full size image

The CIA using K2O was modified by excluding K2O resulting in an index known as the Chemical Index of Weathering, CIW (Harnois 1988). The optimum value of CIW varies as ≤ 50 and increases up to 100 as the weathering intensity increases. The CIW values of the studied tephras ranged from 51.9 to 75.3, indicating slight weathering and lesser clay minerals. Likewise, the Plagioclase Alteration Index PIA (Fedo et al. 1995) developed as an alteration to the CIW attains values of 50 in non-weathered tephras and 100 for clayey minerals such as kaolinite, illite, and gibbsite containing tephras (similar to ones obtained by CIA formula). The values obtained ranged from 50.0 to 71.3 indicating slight weathering.

Lastly, the Chemical Weathering Index, CWI (Sueoka 1988), is based on the increasing tendency of sesquioxides with weathering. Based on the CWI values obtained ranging between 15.8 and 22.8%, the rhyolitic samples such as WKA seemed to be least weathered or fresh followed by minor or slight alterations in BPA, TBKAm, and moderate ones in GKA and dacitic TBKAf and RM samples. The basalt andesitic TBKAc, TB, MA, and basaltic TS and PB positioned on the higher end of the weathering state based on their CWI values due to higher inclusion of sesquioxides (Al2O3, Fe2O3, and TiO2) as a result of their weathering processes.

Out of the studied weathering indices, the IOL and CWI tended to show an increasing trend with the increase in weathering intensity whereas the RR, WIP, CIA, CIW, and PIA showcased a decrement with the increase in weathering. Further on, the trends for weathering indices namely WIP, RR, IOL, and CWI agreed with their expected trends of weathering whereas those for CIA, CIW, and PIA showed the opposite behaviour.

Using the above information, the relationship between chemical indices and weathering state was established, as shown in Fig. 8. Based on the international classification for weathering (Hatheway 2009), the tephras could be classified as fresh or slightly weathered (Table 5). Three types of correlations: low (R2 = 0.34 for WIP), moderate (R2 = 0.49–0.70 for CIA, PIA, CIW), and high (R2 = 0.83–0.89 for IOL and CWI) seemed to exist. Though the trend for RR seemed to be an expectedly negatively correlated state (moderate correlation of R2 = 0.70) with the weathering state, RR has to be considered appropriately further due to the involvement of only two components (SiO2 and Al2O3).

Fig. 8
figure 8

Relationships between weathering state and used weathering indices (a–d) and suited weathering indices (WIP vs. CWI) for the investigated tephras (e)

Full size image
Table 5 Classification of rock weathering
Full size table

The WIP, CIA, CIW, and PIA are based on the groups of mobile oxides. The decreasing trends and corresponding values of CIA, CIW, and PIA from silty sandy Rotorua rhyolites to gravelly sandy Auckland basalts suggested that feldspars conversion (primary minerals) to clays (secondary minerals) was lesser. Because CIW and PIA do not consider aluminum associated with K-feldspar, they yielded higher values for K-feldspar-rich tephras, irrespective of their chemical weathering or not (Fedo et al. 1995). This is evident from the correlation values of CIW (R2 = 0.68), PIA (R2 = 0.56), and CIA (R2 = 0.49).

The WIP showed the lowest correlation (R2 = 0.34) with the weathering state in comparison to the other weathering indices (Fig. 8b). The WIP is the only index to consider aluminum as mobile in comparison to the remaining indices and is purely based on the alkali and alkaline earth elements. This characteristic was used to study and get the best fit through chemically heterogeneous tephras in this study (Price and Velbel 2003). Based on the principle of WIP, as the values get smaller, the samples are more weathered and older. Also, with more weathering, the SiO2 is expected to be lost along with mobile Na2O, K2O, CaO, and MgO, thus decreasing WIP (and increasing CWI). Also, the trends for CWI and IOL increased with weathering due to the depletion of Si and increment of Fe and Ti (Fig. 8c).

Therefore, in order to properly characterize different tephra deposits, the following conditions were considered: parent rocks (homogeneous or heterogeneous), aluminum mobility, and many mobile oxides. The selection of weathering indices such as CIA, CIW, and PIA were avoided due to opposite weathering trends observed. Also, weathering indices RR, VRI, and IOL were not selected despite possessing higher correlations against the weathering state due to lesser mobile elemental oxides in their formulation. Considering the stated, the weathering indices WIP and CWI were selected for further assessment of engineering properties. The correlation between them is shown in Fig. 8e.

3.2.2 Mineralogical Composition by X-ray Diffraction (XRD)

In supplement to the major oxides, the array of minerals was established through the X-Ray Diffraction (XRD)—Rietveld analysis. Using the same, the relative amounts of crystalline phases and amorphous contents in the tephras are presented in Table 6.

Table 6 Modal composition results of XRD-Rietveld analyses for studied tephras
Full size table

An illustrative example showing the amorphous phase and other crystalline phases is presented for rhyolitic WKA in Fig. 9. The presence of amorphous content is evident as a background hump in the XRD pattern of WKA tephra. This was also supported by high glass content ranging between 57.9 and 86.2% in rhyolitic and dacitic samples (WKA, BPA1, BPA2, GKA TBKAm, and TBKAf) as obtained from Rietveld refinement (Table 6). The basalt–andesitic TB (TB1 and TB2) and basaltic PB and MA samples constituted intermediate amounts of glass content ranging between 21.2 and 38.4%, whereas the dacitic RM contained comparatively lesser glass content of 13.9%. No amorphous content was detected in the TS sample. As also evident, the glass content of rhyolites and dacites was comparatively higher than the andesites and basalts.

Fig. 9
figure 9

Illustration of amorphous and detected crystalline phases in the XRD pattern of rhyolitic White Kaharoa Ash (WKA)

Full size image

In addition, the differences among the various tephra types could also be accounted for by the relative abundances of minerals. For rhyolitic and dacitic Rotorua tephras WKA, GKA, BPA, GKA, TBKAm, and TBKAf; silica minerals such as quartz and cristobalite, alkali and plagioclase feldspars were the primary minerals whereas biotite (mica) and augite (pyroxene) were present as accessories (Leonard et al. 2002; Ewart and Cole 1967). In comparison, for basalt–andesitic Rotorua samples TBKAc, TB1, and TB2 and basaltic Taupo and Auckland (TS and PB) samples, plagioclase feldspars, pyroxenes and olivines (forsterite for PB soil) were the essential minerals and secondary minerals such as hematite in TS (thus explaining its red colour) were also seen (Hiess et al. 2007; Searle 1961; Brown et al. 1994). Lastly, the dacitic RM and basalt–andesitic MA constituted a higher proportion of silica minerals indicating influences of other weathering mechanisms as discussed further. The overall susceptibility of the minerals is identified as olivine > pyroxene > mica (biotite) > plagioclase feldspar (Ca to Na) > alkali (K) feldspar > quartz (Bowen 1928). This series indicates olivine to be the most and quartz to be the least susceptible to weathering. In other words, tephra containing a lower amount of olivine or a higher amount of quartz shows relatively lower weathering potential (Goldich 1938). This is reflected in the mineralogical compositions (Table 6), wherein rhyolitic and dacitic tephras such as WKA, BPA1, BPA2, GKA, TBKAm, and TBKAf were relatively less weathered than basalt andesitic tephras such as TBKAc, TB1 and TB2 and basaltic TS and PB.

As noted above, the alkali feldspar is the last feldspar to weather before quartz. Orthoclase was present in lesser amounts in basalt andesites and basalts in comparison to plagioclases, indicating that the plagioclases suffered chemical weathering primarily in these tephras. Sanidine was the alkali feldspar of rhyolites and dacites studied. The Rotorua rhyolites and dacites were placed closer to albite (Ab) and in the intermediate region of albite (Ab)–orthoclase (Or) space, whereas the Rotorua basaltic andesites, Taupo, and Auckland basalts were positioned in the proximity of anorthite (An) region in the albite (Ab)–anorthite (An) plane (Fig. 10a). To get a better viewpoint of their compositional analyses, all the tephras were plotted in the Q–A–P (Quartz–Alkalis–Plagioclase) plane (Fig. 10b). All of them plotted parallel to the Q–P plane, indicating a lack of alkali feldspars in their composition.

Fig. 10
figure 10

Chemical analyses of the studied tephras a classification in Ab–An–Or space (from XRF data) and b weathering state in QAP space (from XRD data)

Full size image

Moreover, the analysis of the weathering state based on quartz and feldspar contents for granitic rocks was used to explain the differences in weathering conditions of studied tephras (Nesbitt et al. 1997; Torabi-Kaveh et al. 2023). The Rotorua basalt andesites TBKA (m and f), TB (1 and 2), basaltic Taupo TS, and Auckland PB plotted very close to the Plagioclase (P) section indicating slight chemical weathering in terms of mechanical breakdown, and the samples composed of bedrocks (feldspars) rather than clay minerals or quartz (Menendez et al. 2022) Slightly above them along the same Q–P plane lay the Rotorua rhyolites and dacites series which was relatively less coarser, and more quartz and alkali feldspar enriched.

On the other hand, the dacitic Rotorua RM sample belonging to a similar eruption source (i.e. OVC) as Rotorua rhyolites contained comparatively higher quartz. The RM consisted of abundant crushed sands and silt fractions (quartz and plagioclase abundant) and secondary aluminous clay minerals such as muscovite. Likewise, the basalt andesitic MA from Auckland plotted closest to the Q region in the Q–A–P plane, showing significant enrichment of quartz in proportion to feldspars. The phreatic and phreatomagmatic nature of RM (Lowe and Balks 2019) explained the mechanical disintegration of rhyolitic ash and country lithics (quartz and feldspar enrichment) and the mixing of clay and silt lake sediments (clay enrichment). Being a surge deposit like RM, the MA tephra constituted an amalgamation of basaltic lapilli with country rock lithics ultimately leading to quartz and plagioclase abundance (Flores et al. 2014). The higher locations of RM and MA in A–CN–K and Q–A–P plots, thus, indicated that external influences such as mechanical breakdown and intermixing were dominant factors and the mineralogy was not entirely controlled through chemical breakdown yet.

The studied tephras, therefore, did not appear to be achieving steady-state weathering conditions yet and were relatively fresh or slightly weathered in nature making them ideal geomaterials for further geotechnical characterization. The weathering indices and relative percentages of silica (such as quartz and cristobalite) and feldspar minerals were further used to study the effect of chemical composition on the degree of particle breakage resulting from compaction tests, as discussed in Sect. 4.

4 Compaction Characteristics and Particle Breakage Evaluation

To facilitate the use of tephras as backfilling materials, it is important to ensure sufficient densification of the constituent particles. Further investigation involving density check was carried out on six tephra samples using the standard Proctor compaction method (ASTM 2012). The samples included Rotorua rhyolitic and dacitic (WKA, GKA, TBKA, RM) and Auckland basalt andesitic and basaltic (MA and PB) tephras. Figure 11 shows the dry density (ρd) and void ratio (e) versus water content (w) relationships; essentially, the maximum dry density (ρdmax) and optimum water content (wopt.) varied from 1.32 to 1.71 g/cm3 and 10.6 to 24.9% (lower range water content, Liu et al. 2022) for the investigated tephras. These values are also presented in Table 7.

Fig. 11
figure 11

Compaction curves obtained for different tephras: a Variations in dry density with water content; and b variations in void ratio with water content

Full size image
Table 7 Compaction parameters of studied tephras
Full size table

Three types of compaction curves were obtained (Fig. 11): one-and-a-half peak (for WKA, GKA, and TBKA), parabolic (for RM and MA), and gentle peak (PB). The Rotorua rhyolites WKA, GKA, and TBKA being well-graded and non-plastic silty sands (14–22% fines), as also manifested by their uniformity coefficient (Cu) and curvature coefficient (Cc) values in Table 1, were more compactable owing to a wide range of particle sizes enabling better interlocking. From about w = 2.5–3.5%, their ρd decreased with the increase in w and reached a minimum at about w = 9.0–12.0%. With the further increase in w, ρd increased for values of w up to approximately 16–17% (WKA and TBKA) and 24.5–27.5% (GKA). Beyond this w, the ρd decreased for WKA, TBKA, and GKA. The ρd (at starting w) was nearly identical to ρdmax (at wopt.), hence the curves are defined as one-and-a-half peaks that are typical of sand-dominant tephras, containing quartz and feldspars (Lee and Suedkamp 1972).

For Auckland MA and Rotorua RM samples containing 50–70% slightly plastic fines (PI = 3.7 and 13.0%), the compaction curves were more regular or parabolic in shape. The ρd increased, reached a peak, and, thereafter, decreased with increasing w. This behavior was attributable to the water lubrication theory for fine-grained soils.

The saturation values ranged from 72 to 85%. for silty-sands (WKA, TBKA, and GKA) and silty tephras (RM and MA) of Rotorua and Auckland. In contrast, the inclusion of water was ineffective in controlling the density of poorly graded sand PB. This was apparent by the very small increase in ρd with the increase in w (Fig. 11a) and a moderate saturation value of 50%. Expectedly, the rhyolitic (WKA, GKA, TBKA) and dacitic (RM) silty sands from Rotorua were placed below Auckland basalt–andesitic silt (MA) and basaltic sand (PB) due to their lower Gs values. The differences between tephras from different geological origins and physio–chemical states are thus evident.

As the tephras are void-filled slightly-weathered materials, they are susceptible to breakage after impact loading such as compaction. After compaction, the tested tephras were oven-dried and, then, sieved to quantify the particle breakage. The breakage indices defined by Hardin (1985) and Miura and Yagi (1997) were used to cover the entire range of particle sizes as well as the generation of fines after compaction respectively.

Figure 12 presents the PSD curves following the compaction tests. For the tested tephras, the differences in positions of the grain size distribution curves (before and after compaction) were very small. This meant that the amounts of particle crushing following the compaction process were very similar. The average of the particle breakages at different water contents was, therefore, used. As also reported in Table 7, the highest post-compaction breakage by Hardin’s index was shown by fine-grained MA and RM, then friable GKA; followed by well-grained TBKA and WKA, and lastly poorly-graded PB. A slight change in breakage pattern was shown through average breakage values by Miura and Yagi’s index with the friable GKA more crushable than fine-grained MA and RM; followed by well-graded WKA and TBKA and then poorly-graded PB.

Fig. 12
figure 12

Shifts in the particle size distributions after compaction at tested water contents for the tested tephras

Full size image

Higher breakage in fine-grained soils such as RM and MA could be explained by sliding and rotation between a larger number of particles and the contact forces being weaker leading to breakage at the contact points upon compaction. In addition, the nature of deposition of these tephras (base surge deposits) was also accountable for the higher breakage. In comparison, the well-graded WKA, TBKA, and GKA showcased lesser breakage owing to better contact forces among a wider range of particle sizes. However, although being well-graded, higher breakage in GKA was due to associated weathering effects leading to pulverization and softer particles (as elaborated earlier in Sect. 3.2). For uniformly graded PB, the interlocking forces between the uniformly sized particles were negligible and additional water played very little role in bringing them together leading to insufficient compaction.

Figure 13 shows the relationship between ρd and breakages of different tephras. From this figure, it could be understood that the breakage in general shows an increase with the increase in ρd with correspondingly higher values around wopt. condition than at the drier or wetter conditions. This fact implied that the transmission of compaction effort was most utilized and effective in optimum conditions. The wopt. and ρdmax obtained can, thus, be further utilized as density parameters for their usage as backfilling materials.

Fig. 13
figure 13

Relations of overall particle size distribution (Hardin’s breakage index) and increment in fines content (Miura and Yagi breakage index) and dry density after compaction

Full size image

5 Inter-Relating Physical, Chemical, Compaction, and Other Geotechnical Characteristics

This section describes the inter-relationships developed between the physical, chemical, and compaction aspects of the airfall tephra deposits. As seen, an evaluation of the physical and chemical properties enabled the assessment of the weathering state of the tephras, which further impacted compaction characteristics and can also possibly influence shear strength behaviour. The natural existence and weathering processes dictated a broader range of particle types from gravels to clays to be present in the tephra samples. The tephras also, therefore, had different maximum densities and optimum moisture contents according to the Gs of the particles, PSD, and grain shape of particles. The majority of them had sub-angular to angular grain shapes, as indicated through morphology (Fig. 3), thereby aiding compaction.

Figure 14 shows the relationship between median particle size (D50) with the ρdmax and wopt. for the studied tephras. The ρdmax increased with increasing D50. Correspondingly, the wopt. showed a decrease with increasing D50. The differences in achievable ρdmax can be attributed to their respective Gs values. The ρdmax of Auckland samples (PB and MA) was higher than Rotorua samples (WKA, GKA, TBKA, and RM) due to higher Gs associated with the former. Concerning wopt., the water content required to achieve optimum conditions was expectedly higher for fine-grained (> 50% fines) RM and MA tephras followed by silty sandy WKA, GKA, and TBKA and then sandy PB with almost zero fines.

Fig. 14
figure 14

Variation of median particle size D50 (mm) and specific gravity Gs with maximum dry density ρdmax and optimum water content wopt.

Full size image

The D50 (= 1.63 mm) of Auckland sandy PB was substantially higher than the remaining tephras possessing D50 in the range of 0–0.50 mm. The coarser uniform structure of PB, therefore, subsequently required lesser wopt.. Taking Gs = 2.682 of New Brighton Sand (NBS, hard-grained beach sand, Sood et al. 2022a) as the dividing line, the rhyolitic samples WKA, GKA, TBKA, and dacitic RM fell on the lighter side, and basalt andesitic MA and basaltic PB leaned towards the heavier side. Being from the same eruption source, the difference between Rotorua samples was due to the inclusion of clayey minerals such as muscovite in RM in comparison to non-plastic WKA, GKA, and TBKA. Though MA also included slightly plastic fines in higher proportion, the lower plasticity, and higher Gs did not permit lower ρdmax. Instead, higher amounts of heavier accessory minerals (augite) shifted Auckland samples PB and MA to the heavier side.

As compaction results in particle breakage, suited weathering indices WIP and CWI and mineralogical composition in terms of silica and feldspar minerals after respective weathering processes were compared against breakage values. The average breakage values were plotted in terms of all particle sizes (Hardin 1985) and fine sizes (Miura and Yagi 1997). The weathering state and average breakage values for TBKA were computed by summing the respective values at its three fraction particle sizes (TBKAc, TBKAm, and TBKAf).

As established through XRD-Rietveld analyses, the amount of crystalline silica minerals such as quartz and cristobalite were highest in RM and MA; followed by GKA, TBKA, WKA, and least in PB. Considering feldspars, the Auckland samples contained lower feldspars than the Rotorua samples. As silica mineral is the last mineral to crystallize in any tephra, its resistance to weathering is highest in comparison to feldspars. However, due to external factors such as deposition and later intermixing, the balancing of silica and feldspar minerals might get affected in the host tephra. Therefore, the silica and feldspars present in the current weathered state of tephra were cumulative effects of (i) deposition manner, (ii) chemical breakdown of inherent minerals, and (iii) external intermixing if any.

As indicated by the cumulative plots of WIP and CWI (Fig. 15) against quartz–feldspar balancing (lower WIP and higher CWI values indicated higher weathering), the RM and MA samples showed higher quartz (PB being an exception showing very low silica content) and correspondingly lower feldspars than WKA, TBKA and GKA. The GKA was silica enriched in comparison to WKA and TBKA (being similar source products) due to the hydrothermal action (Lowe and Balks 2019; Heap and Violay 2021) of RM ash falling on it. The resulting pulverization of GKA, therefore, gave it a friable and more weathered structure (thus, color change to golden) than WKA and TBKA.

Fig. 15
figure 15

Weathering state in terms of WIP and CWI against Silica and Feldspar minerals (arrows indicate weathering progression)

Full size image

A higher WIP meant a greater amount of alkalis (Na2O + K2O, which are removed later for rhyolites and dacites) remaining after weathering and therefore lesser weathering. Also, the concentration of sesquioxides (TiO2 + Al2O3 + Fe2O3) for all chemical components (CWI) was highest for PB and lowest for WKA. In other words, this also meant lesser loss of SiO2 from the WKA sample and, therefore, least weathering in comparison to PB and remaining tephra samples. By using WIP and CWI as such, the effect of both alkalis and sesquioxides was taken into consideration.

All of these factors lead to higher breakage values for more weathered GKA, RM, and MA samples in comparison to TBKA, WKA, and PB samples which were less weathered and identical individually in bulk mineralogical aspect. These trends presented in terms of weathering state (evaluated by WIP and CWI) against breakage (computed by the Hardin, and Miura and Yagi indices) values (Fig. 16) showed identical orders. The comparatively lower breakage for uniformly-graded PB in comparison to the remaining tephras could be attributed to the presence of heavier minerals such as augite and forsterite, thus making particles extremely hard and leading to an increase in their Gs.

Fig. 16
figure 16

Weathering state and associated particles breakages in terms of Hardin (1985) and Miura and Yagi (1997) indices (arrows indicate weathering progression)

Full size image

As such, considering the different geological origins of the studied tephras, the physio–chemical and compaction parameters were interlinked and seemed to be interdependent. Different weathering tendencies and mineralogical abundancies were found to control the geotechnical properties such as compaction parameters (ρdmax and wopt.) and post-compaction breakage.

Furthermore, to highlight the effect of physical and chemical parameters on the geotechnical characteristics, the variation in shear strength of the compacted tephras (Dc ≈ 90%) was studied (Sood et al. 2022b). It was found that the shear strength of the tephras differed with the resultant particle sizing, weathering state, and mineralogy of the deposits. As shown in Fig. 17, for the monotonic triaxial shear tests carried under drained ‘D’ and undrained ‘U’ conditions, the effective stress paths q versus p’ became asymptotic to the failure line. The failure line or gradient of these stress paths, M, was correlated to the interparticle friction angle φ using the following equation:

Fig. 17
figure 17

Stress ratios q/p’ versus εq and stress paths q versus p’ curves for the tephras WKA, GKA and MA under monotonic triaxial conditions (Sood et al. 2022b)

Full size image
$$M={\text{q}}/\mathrm{p{\prime}}=(6{\text{sin}}\varphi )/(3-{\text{sin}}\varphi )$$
(2)

For the range of confining pressures 50–200 kPa studied, the φ of WKA tephra (41.5°) was higher than that of GKA (35.7°) and MA (35.5°) tephras (Heap and Violay 2021). This was due to the granular and least weathered nature of WKA, also leading to a higher resistance among the studied tephras.

In accordance, the highest post shear breakage Br. or lowest friction angle φ for basalt andesitic MA in comparison to the other tephras, was as a result of its highest weathering state as detected from its lowest WIP value (Fig. 18). This was relatively eventuated through the highest amounts of feldspars in MA, with higher silica accountable due to lithics intermixing. Thus, the chemical composition seemed to affect the shear strength of the studied tephras.

Fig. 18
figure 18

Geotechnical parameters such as post shear breakage Br. and friction angle related to weathering index WIP (arrows indicate weathering progression, Sood et al. 2022b)

Full size image

6 Conclusions

As such, in this study, eleven tephra samples (basalt to rhyolitic) collected from the North Island of New Zealand were investigated to characterize their physical (grain size, specific gravity, and morphology), chemical (elemental and mineralogy using X-ray fluorescence and X-ray diffraction) and compaction characteristics (maximum dry density, optimum water content, and particle breakage). The following conclusions were accordingly drawn.

  1. i.

    Depending on the geology of the sites and weathering induced, the tephras presented considerable differences in their physical and chemical properties. The samples included a wide range of particle sizes from gravel, sand, and silt to slight clay and chemical variability from basalts to rhyolites. For certain Rotorua deposits such as those collected from close proximities between Tarawera basaltic and Kaharoa rhyolitic depositions (TBKA, TB, and BPA), admixing was evident in deviations from their specific gravity (Gs) values. Thus, in addition to checking the density of a natural volcanic deposit, Gs was also used for checking intermixing.

  2. ii.

    The elemental oxides and mineralogical compositions were used as the primary indicators to assess the weathering state of the tephra deposits. The major oxides expressed in terms of weathering indices such as WIP, CWI, and A–CN–K plots, and minerals assemblages (quartz–feldspar balancing) in Q–A–P plots along with depositional settings were used to properly evaluate the weathering state.

  3. iii.

    All of them being slightly weathered enabled their utilization as potential backfilling geomaterials for engineering applications. The applicability of these tephras for backfilling was further checked by compaction tests. The well-graded tephras with non-plastic fines (WKA, TBKA, and GKA) and fine-grained tephras (MA and RM) with plastic fines showcased better compaction properties than uniformly graded PB tephra with no fines content.

  4. iv.

    The post-compaction breakage values fell below 10%, therefore, encouraging the use of these deposits as structural fills. The higher post-compaction breakage of fine-grained tephras than well-grained and uniformly-grained tephras was accountable to factors such as deposition, intermixing, and current weathering state. The test results also confirmed that the higher degree of weathering led to higher particle breakage. Yet, the degree of weathering and, thus the associated particle breakage, was inevitably linked to the quartz and feldspar contents. It was seen that more weathered volcanic deposits such as RM and MA, having enriched silica (quartz) content and/or low feldspars content due to their mode of deposition (base surge) and intermixing, showed higher particle breakage in comparison to the other deposits. The GKA deposit, also showcased higher breakage, in response to pulverization and intermixing by falling ash and countryside lithics.

  5. v.

    The quantification of the weathering state using chemical weathering indices and mineralogical compositions (silica–feldspar balancing), therefore, had a measurable impact on ground investigation practices. This approach enabled the interpretation of different tephra profiles as well as their effects on engineering properties such as compactability, breakage, and shear strength.

Nevertheless, before the use of compacted tephras as structural fills can be further recommended, it is of paramount importance to understand their response when subjected to extreme rain and flood events due to climate change, and associated potential accelerated weathering processes in combination with earthquake conditions. But firstly it was important to distinguish different tephra types on the fundamental basis of their physical and chemical properties, as achieved in this paper. To proceed further, complementary and detailed laboratory investigations of their geotechnical engineering properties (strength, compressibility, collapsibility, liquefaction potential, etc.) have been conducted by the authors, the results of which will be available in the forthcoming papers.