New Zealand limestone purity

Abstract

Geographic coordinates and lithostratigraphic names have been assigned to 2255 limestone samples that were previously analysed for soluble carbonate content. This permits, for the first time, a comparative assessment of the purity of onland New Zealand limestones by location and age. The three purest limestone populations have median CaCO₃ contents of 93–97 wt%, and are from: (1) Cambrian to Early Cretaceous basement terranes, (2) Eocene–Pliocene limestones of the Chatham Islands and (3) the Eocene Ototara Limestone. These have CaCO₃ contents comparable with carbonates found in isolated submarine bank settings that lack appreciable siliciclastic detritus and biogenic silica. By contrast, New Zealand Oligocene–Early Miocene limestones (Whangarei, Te Kuiti, Mahurangi, Nile Group, Otekaike and Forest Hill stratigraphic units) have median CaCO₃ contents of 71–91 wt%, comparable with New Zealand Paleocene–Eocene and Middle Miocene–Pleistocene limestones. Petrographic data from a subset of 52 Otekaike and Forest Hill limestones confirm that the non-carbonate content is mostly siliciclastic sand, not glauconite. The limestone purity data support, but do not on their own prove, a hypothesis of a terrigenous source (emergent land) during the Late Oligocene-Early–Miocene maximum marine inundation of Zealandia.

Keywords:

• New Zealand
• limestone
• calcium carbonate
• purity
• paleogeography
• Oligocene
• Miocene

Introduction

Limestones are widespread in New Zealand (Fig. 1). Their geological, economic and scenic importance is indicated by an extensive stratigraphic, paleontological, paleo-environmental and geopreservation literature, as well as widespread quarrying (Morgan 1919; Nelson 1978; Kamp & Nelson 1988; Christie et al. 2001; Hood et al. 2003a,b; Kenny & Hayward 2009 and references therein). One aspect of New Zealand limestone geology that has received little attention is limestone purity.

Figure 1 Location of analysed New Zealand limestones referred to in the text.

In this short synthesis paper, we summarise the purity of onland New Zealand limestones as revealed by their quantitatively determined soluble carbonate content. Earlier compilations of total rock limestone chemistry made by Morgan (1919) and Kitt (1962) reported wt% CaCO₃ as determined by weight loss in cold dilute acid. This method is simple, is reliably standardised over large numbers of samples and different vintages of analysis and is also comparable with modern colorimetric methods of carbonate determination.

Our digital compilation is the first to assign accurate geographic coordinates and stratigraphic names to Morgan's and Kitt's data, and this is, therefore, the first paper to make and present a New Zealand-wide synthesis and comparison of limestone purity between different formations. We also present volumetric petrographic data on insoluble residues in limestones of Otago and Southland, and discuss the relevance of chemical and petrographic limestone purity to the Oligocene–Miocene paleogeography of Zealandia.

Data

Limestone purity

By definition, limestones contain more than 50% CaCO₃ (e.g. Andrews 1982) and include marlstones. The small density differences between calcite, quartz and oligoclase feldspar (relative densities 2.71, 2.65 and 2.67, respectively) result in differences of <2% when total carbonate is calculated on a volume % compared with a weight % basis. Any rocks with a measured CaCO₃ content <50% have been excluded from this paper.

Morgan (1919) catalogued New Zealand limestone occurrences and presented c. 1002 whole-rock carbonate analyses. However, no geographic coordinates were given (there was no national grid in Morgan's day), and ages and stratigraphic names were not explicitly listed. A chemical data compilation of 1191 limestone analyses by Kitt (1962) was even more sparsely documented, with no ages or formations given and with only some grid references.

Based on written location descriptions and quarry names, we were able to confidently assign geographic coordinates and lithostratigraphic names to a total of 661 samples analysed by Morgan (1919), and 952 by Kitt (1962). In addition, 682 limestones analysed by other workers that have stratigraphic and age information have been included in this study. These datasets are from Cooper (1966), Waterhouse, (1966), Warren (1969), Moore (1975), Moore & Belliss (1979), Moore & Hatton (1985), Browne & Wezenberg (1988), Smith et al. (1989) and unpublished data from the New Zealand Geological Survey Sedimentology Laboratory files. In rare cases where MgCO₃ analyses were additionally reported, we have, for this paper, added wt% MgCO₃ to wt% CaCO₃ to give a total soluble carbonate fraction. For this study, we made just one new CaCO₃ analysis ourselves: P80187, an Otekaike Limestone from the Hakataramea Valley, Otago (NZMG E2323589, N5613729), which has 89.1 wt% CaCO₃, as determined by cold HCl dissolution.

After removing duplicates, we were left with a digitised dataset of 2255 New Zealand limestone samples analysed for CaCO₃ labelled with 1 : 250 000 scale lithostratigraphic names as used by the QMAP project, e.g. Rattenbury et al. (2006), Lee et al. (2011), Rattenbury & Isaac (2012). Some limestone formations are diachronous (e.g. Amuri) and accuracy and precision of age are limited to what can be inferred from lithostratigraphy. Furthermore, the bioclast content and textures (grainstone, micrite, etc.) of individual samples in our dataset are unknown and, again, generalisations must be made on the overall character of the rock unit. Sample numbers, grid references, locations, rock types, stratigraphic names and carbonate analyses have all been entered into the Petlab database (http://pet.gns.cri.nz). The data are also listed in the Supplementary Table File of this paper.

The vast majority of samples used in this paper were collected by people other than ourselves, and the analyses were done over the course of a century, with unknown sample sizes and in a number of different laboratories. So as not to over-interpret the data, we have categorised the samples into 5 wt% bins (Fig. 2). The relatively large amount of chemical data gives us confidence that it is representative of the formations. If there is any bias in the sampling it might be that purer limestones have been included at the expense of impure ones as the purpose of the investigations of Morgan (1919) and others was to locate, identify and analyse good limestone resources.

Figure 2 Histograms of wt% carbonate content (CaCO₃ plus, where reported, MgCO₃) for selected limestone populations of Table 1. Class interval 5 wt%, lower cut-off for limestones 50 wt% CaCO₃. A–B, isolated carbonate bank references; C–D, New Zealand basement terranes and Chatham Islands; E–J, Oligocene–earliest Miocene (mainly Late Oligocene) limestones from time of maximum marine inundation of Zealandia, in north–south order; K–P, pre- and post-New Zealand Oligocene limestones, in north–south order. Vertical axis of histograms has been scaled so area of bars is the same for all diagrams. Dots and lines along the x-axis are median and interquartile ranges from Table 1. EP, Eastern Province; M, Mangaheia; R, Rangitikei; TK, Te Kuiti; WP, Western Province; Wh, Whangarei.

Limestone petrography

The non-carbonate content of New Zealand Oligocene–Miocene limestones comprises varying proportions of glauconite (used here in a loose, undifferentiated sense for perigenic, authigenic and/or allogenic green phyllosilicate minerals), siliciclastic sand grains and mud matrix (Fig. 3; Ward & Lewis 1975; Nelson & Hume 1987; Hood et al. 2003a,b). In order to provide more information on the nature of the insoluble limestone residues, volumetric data from South Island limestone sites were compiled.

Figure 3 Photomicrograph of P78764 Otekaike Limestone. Plane polarised light, scale bar 1 mm. Porosity is blue, glauconite very dark green to brown, angular quartz and feldspar grains white, rest of slide is carbonate (light to dark grey). Carbonate part includes numerous foraminifera.

Table 1 Summary statistics of limestone CaCO₃ datasets used in this paper. Major groupings listed in order of decreasing median CaCO₃.

Population	Age	n	CaCO3 max	75th percentile	CaCO3 median	25th percentile	CaCO3 min	% VHP samples
New Zealand limestones
Basement terranes	Cam–Cret	367	100.0	98.7	96.7	90.5	50.0	31
Chatham Islands	Eoc–Plio	52	100.0	97.5	95.0	89.6	50.5	6
Ototara	Eocene	77	99.5	96.4	93.4	89.0	58.7	4
Nile Group	Oligocene	341	99.1	95.5	90.9	79.8	50.4	1
Te Kuiti	Oligocene	169	99.5	95.0	90.4	78.5	50.0	7
Whangarei	Oligocene	35	98.5	93.2	89.3	78.1	65.9	0
Mangaheia	Plio–Pleist	368	99.6	92.7	86.0	78.4	51.3	2
Amuri	Eocene	27	96.5	88.4	85.0	75.9	58.1	0
Tolaga	Miocene	86	99.0	91.5	84.7	73.6	50.0	1
Forest Hill	Olig–Mio	122	99.4	90.7	82.4	72.2	50.0	2
Otakou–Motunau	Miocene	71	97.4	91.5	80.8	68.2	51.2	0
Amuri	Early Olig	52	96.1	84.5	79.8	66.6	50.5	0
Otekaike–Weka Pass	Oligocene	193	98.4	90.0	78.6	66.5	50.0	0
Amuri	Paleocene	41	93.9	85.0	75.4	69.5	50.5	0
Mangatu	Pal–Eocene	13	83.5	76.3	75.0	70.0	63.8	0
Weber	Oligocene	20	94.0	84.9	73.0	66.0	50.5	0
Mahurangi	Oligocene	209	98.0	77.0	71.0	66.7	53.5	0
Rangitikei	Plio–Pleist	12	99.0	85.9	68.4	62.1	53.0	8
Reference data
Cook Islands and Niue	Mio–Holo	40	100.0	99.1	98.3	97.1	80.5	45
Shatsky Rise	Cret–Holo	163	100.0	96.9	96.0	94.9	71.6	3
Ontong Java Plateau	Cret–Mio	102	100.0	96.9	95.0	92.9	70.1	10
Bahamas	Mio–Pleist	370	99.8	95.6	94.6	93.5	50.1	7
Maldives	Mio–Holo	541	99.2	95.6	93.8	91.9	82.9	2
Kerguelen Plateau	Cret–Olig	73	97.2	95.0	93.6	89.5	65.1	0

CaCO₃ concentrations in wt%. Lower cut-off for limestones = 50 wt%. See text for data sources. Cam, Cambrian; Cret, Cretaceous; Eoc, Eocene; Holo, Holocene; Mio, Miocene; Olig, Oligocene; Pal, Paleocene; Pleist, Pleistocene. % VHP = per cent of very high purity (> 98.5%) samples in the population.Reference datasets are Cook Islands and Niue (Kitt 1962), Shatsky Rise (Malone 2005), Ontong Java Plateau (Shipboard Scientific Party 2001), Bahamas (Shipboard Scientific Party 1990; Malone 2000), Maldives (Malone et al. 1990) and Kerguelen Plateau (Ehrmann 1991).

Smith et al. (1989) reported visually estimated modes of 46 thin sections of the Oligocene Otekaike Limestone of Canterbury and North Otago. Hyden (1979) gave point count data from 429 thin sections of Oligocene–Miocene Forest Hill Limestone of Southland. To augment these datasets, we point counted 12 samples of our own from Otago and Southland to give volumetric data from a total of 52 South Island limestone locations (Table 2). Despite the different count methods and parameters used by Hyden (1979), Smith et al. (1989) and ourselves, we believe the following seven parameters can be compared across datasets: clastic grains (> 60–100 µm in size) of: (1) bioclastic material, (2) glauconite (sensu lato), (3) siliciclastic material mainly quartz and feldspar; interstitial (matrix) grains of (4) mud (<60–100 µm in size), (5) micritic carbonate, (6) sparry calcite cement and (7) porosity. Some clastic grain and total rock parameters have been calculated and are also shown in Table 2 and Fig. 4. We acknowledge that authigenic glauconite may have a minor degraded terrigenous mud progenitor, and that it can be difficult to discriminate terrigenous mud from micrite in normal thin sections.

Figure 4 A, Histogram of the volume percent sand and mud content of 52 limestone sites in the South Island as a proportion of counted non-carbonate grains (= sand + mud + glauconite). Data from rightmost column in Table 2. Note that, because minor insoluble residues of siliceous and phosphatic microfossils are not shown in Table 2, the sand fractions in the histogram may be slight overestimates. B, Geographic distribution of the vol% sand content of Oligocene–Miocene Otekaike, Otakou and Forest Hill limestones as a proportion of all clastic grains (= bioclasts + glauconite + sand). Data from fifth columns from the right in Table 2. Blue colour on map = geological basement (Cambrian to Early Cretaceous rocks).

Table 2 Petrographic data for samples of Otekaike, Otakou and Forest Hill limestones.

Sample		n	Source	Lat °S	Long °E	Location	Clastic grains (vol%)			Interstitial material (vol%)				Calculated percentages
							Bioclast	Glauconite	Sand	Mud	Micrite	Cement	Pores	Bio/Clastic	Glauc/Clastic	Sand/Clastic	Sand+Mud/All	Glauc/All	CaCO3/All	Sand+Mud/Sand+Mud+Glauc
Otekaike Limestone
SL	K36/s25–29	5	S89	−43.6444	171.4120	Staveley	74	3	1	–	3	14	5	95	4	1	1	3	96	21
SL	I40/s30	1	S89	−44.6423	170.6453	Hakataramea	60	5	0	–	5	0	30	92	8	0	0	7	93	0
SL	J38/s24	1	S89	−44.1079	170.8143	Fairlie	73	5	5	–	15	0	2	88	6	6	5	5	90	50
SL	J38/s03	1	S89	−44.1127	170.7674	Fairlie	70	5	5	–	0	15	5	88	6	6	5	5	90	50
SL	I42/s01–02	2	S89	−45.2823	170.5101	Green Valley	67	5	5	–	3	10	10	87	6	7	5	6	89	50
SL	K36/s21–22	2	S89	−43.6757	171.3330	Chapmans Ck	65	5	5	–	7	8	10	86	7	7	6	6	88	50
P	78763	1	New	−43.7488	171.1549	Mesopotamia	71.5	5.3	4.3	1.6	12.0	2.0	3.3	89	6	5	6	6	88	53
SL	J36/s174–243	3	S89	−43.6910	171.2110	Hinds River	59	2	10	–	10	7	12	83	3	14	11	2	87	83
SL	K36/s13–24	5	S89	−43.5919	171.4555	Taylors Stream	66	3	9	–	3	5	14	84	4	12	11	3	86	77
SL	K36/s30–38	5	S89	−43.6761	171.2970	Blands Bluff	74	1	12	–	2	5	6	86	2	12	13	1	86	90
P	80187	1	New	−44.6620	170.6495	Hakataramea	56.3	7.9	5.0	2.6	3.0	0.6	24.3	82	11	7	10	11	79	49
OU	29565	1	New	−44.3268	170.6426	Hakataramea	67.0	7.6	7.6	4.6	3.6	5.6	3.6	82	9	9	13	8	79	62
SL	J38/s31–32	2	S89	−44.3294	170.8671	Pareora	52	6	15	–	5	17	5	72	8	20	16	6	78	71
P	78764	1	New	−44.1921	171.0457	Pleasant Point	49.3	11.9	7.3	0.6	12.3	3.3	15.0	72	17	11	9	14	77	40
SL	K36/s18–20	2	S89	−43.5594	171.7422	Curiosity Shop	59	19	1	–	1	6	14	74	24	2	2	22	76	7
SL	J40/s14	2	S89	−44.8024	170.9740	Waihao	54	17	6	–	1	16	6	71	22	7	7	17	76	27
SL	I40/s24–25	2	S89	−44.8065	170.5313	Waitaki	59	12	12	–	9	0	8	71	14	15	13	13	74	51
SL	I38/s09–10	2	S89	−44.1916	170.7044	Tengawai River	57	6	19	–	6	4	8	70	7	23	20	7	73	76
SL	J36/s39–41	3	S89	−43.7495	171.1559	Coal Creek	47	11	16	–	7	12	7	63	14	23	18	11	71	61
SL	J38/s23	1	S89	−44.3258	170.7569	Rocky Gully	50	10	20	–	5	10	5	63	12	25	21	11	68	67
SL	I37/s09–10	2	S89	−43.9914	170.7369	Stoneleigh Rd	45	2	28	–	15	0	10	60	3	37	30	3	67	91
SL	I40/s10–11	2	S89	−44.7060	170.3777	Awahokomo Stm	45	7	28	–	5	5	10	56	10	34	31	8	61	79
SL	J42/s12	1	S89	−45.2059	170.8861	All Day Bay	30	35	5	–	25	0	5	43	50	7	5	37	58	12
Otakou Group
OU	38210	1	New	−46.0653	169.9956	Milburn	62.0	4.6	13.0	0.0	2.3	17.3	0.6	78	6	16	13	5	82	74
P	79928	1	New	−45.8732	170.6398	Sandymount	67.3	3.9	16.3	4.6	2.0	1.6	4.0	77	4	19	22	4	74	84
P	80065	1	New	−45.9224	170.4685	Tunnel Beach	50.6	3.6	18.6	10.0	1.0	0.3	15.6	70	5	25	35	4	61	89
Forest Hill Limestone
OU	44143–150	8	H79	−45.9621	167.9433	Twinlaw	88.7	0.4	0.1	0.0	4.3	6.8	0.1	99	1	0	0	1	99	22
OU	44134–142	4	H79	−46.0039	168.2099	Pukemutu	83.1	0.0	0.4	1.0	5.0	11.5	0.0	99	0	1	1	0	99	100
OU	44078	1	H79	−46.1820	168.1145	Isla Bank	95.0	0.5	1.0	0.0	2.0	1.0	1.0	98	1	1	1	0	99	67
OU	44079	1	H79	−46.1819	168.0263	Ringway	89.0	0.5	1.0	0.5	10.0	1.0	0.5	98	1	1	1	1	98	75
OU	43674–706	33	H79	−46.1480	168.4131	Browns	91.0	0.5	0.3	1.7	1.2	3.9	2.3	99	1	0	2	1	97	83
P	78911	1	New	−46.2044	168.4396	Forest Hill reserve	71.6	2.6	1.3	0.3	1.6	11.6	10.6	95	3	2	2	3	95	38
OU	44160–169	8	H79	−45.6058	167.8565	Elmwood quarry	82.8	0.0	3.0	3.4	5.5	7.1	0.2	96	0	4	6	0	94	100
OU	43714–731	13	H79	−46.1000	168.3599	Lady Barkly quarry	83.7	1.5	1.7	3.2	8.2	2.1	0.7	96	2	2	5	1	94	76
OU	43588–592	4	H79	−46.2163	168.4296	Poynter Road	81.9	1.0	5.3	0.4	3.3	3.9	5.5	93	1	6	6	1	93	85
OU	43753–806	45	H79	−46.0571	168.3453	Centre Bush quarry	78.1	3.0	1.7	2.4	11.0	4.2	1.2	94	4	2	4	3	93	57
OU	44080	1	H79	−46.2453	167.9909	Ermedale	54.0	0.5	7.0	0.0	33.0	2.5	3.0	88	1	11	7	1	92	93
OU	44075–077	3	H79	−46.0756	167.6573	Alton Valley	75.8	3.3	4.3	0.7	14.0	1.2	1.2	90	5	5	5	3	92	60
OU	43606–664	42	H79	−46.1902	168.4317	Forest Hill quarry	71.6	2.5	4.2	2.5	13.4	7.4	0.6	89	4	7	6	2	91	73
OU	43598	1	H79	−46.2297	168.4273	Forest Hill reserve	62.0	0.5	9.0	0.0	1.0	27.0	1.0	86	1	13	9	1	90.	95
OU	44123–125	2	H79	−46.1783	168.6415	Marshall Creek	58.8	0.0	10.5	0.0	13.3	17.8	0.3	81	0	19	10	0	89	100
OU	43842–911	114	H79	−45.8055	168.2961	Castle Rock	75.8	4.4	4.0	2.5	7.1	6.6	0.5	90	5	5	7	4	89	59
OU	44081–096	9	H79	−45.8403	168.5898	Balfour quarry	43.9	8.0	3.8	0.6	34.4	9.4	0.3	79	14	7	4	8	88	35
OU	43520–584	61	H79	−46.2213	168.4682	Sharks Tooth Hill	75.9	4.9	4.6	2.9	8.2	2.9	1.6	88	6	6	8	5	87	60
OU	44021–070	23	H79	−46.0322	167.7132	Clifden	78.2	1.9	9.0	1.6	5.0	3.7	1.5	87	2	11	11	2	87	85
OU	43807–833	22	H79	−46.0228	168.3442	Fernhill quarry	70.3	4.1	3.3	3.3	12.3	4.4	1.1	87	9	4	7	6	87	52
OU	44154–55	2	H79	−45.7613	168.0138	Waihōaka	81.8	0.0	13.0	0.5	0.5	6.8	0.0	86	0	14	13	0	87	100
P	79926	1	New	−46.0200	167.7428	Clifden caves	56.6	5.6	6.3	1.0	7.0	11.6	11.6	83	8	9	9	6	85	57
OU	44108–121	8	H79	−46.1299	168.7489	Waimumu quarry	51.7	0.9	15.8	0.9	15.7	15.4	1.1	74	1	25	17	1	82	95
P	78907	1	New	−46.2017	167.9228	Pourakino River	61.3	2.6	21.3	1.6	6.0	5.0	2.0	72	3	25	23	3	74	90
OU	44185–186	2	H79	−45.8363	167.6432	Blackmount	53.0	1.5	27.0	1.0	2.5	16.0	0.5	64	2	34	28	1	71	95
P	78667	1	New	−45.8833	168.6059	Balfour	45.0	32.6	10.6	2.3	4.6	1.6	3.0	51	37	12	14	34	52	28

Means of data are presented for samples <1 km distant from each other. S89 = Smith et al. (1989) visual estimate from thin section matrix grain size cut-off not specified, H79 = Hyden (1979) 400 point counts per thin section matrix grains <100 µm, New = this study 300 point counts per thin section matrix grains <63 µm (Gazzi-Dickinson method, Ingersoll et al. 1984). Lat and long are NZGD49 projection. ‘0’ means below detection limit, ‘–’ means not recorded. Glauconite (sensu lato) counts include limonite replacing glauconite. Sand is mostly quartz, feldspar and mica. Mud includes silt and clay sized particles.

The main terrigenous material seen by Hyden (1979), Smith et al. (1989) and in this study was angular grains of quartz and feldspar with lesser muscovite, chlorite, magnetite and epidote. Ward & Lewis (1975) have noted angular grains of quartz, feldspar and other minerals in the Otekaike Limestone of South Canterbury.

Discussion

Limestone purity

We divided the onland New Zealand limestones into 18 populations based on geographic distribution and age (Fig. 1, Table 1). Thirteen of the populations contain 40 or more samples and the larger datasets can be considered more reliably representative of their parent units. Histograms of carbonate content are presented in Fig. 2.

Reference samples

The British Geological Survey definition of limestone purity is: > 98.5% CaCO₃ very high purity, 97.0–98.5% high purity, 93.5–97.0% medium purity, 85.0–93.5% low purity, 50.0–85.0% impure (Harrison 1992). For industrial (as opposed to agricultural) use, limestones generally have to be very high or high purity.

Almost all of the onland New Zealand limestone groupings show a wide range in carbonate content from near 50 to near 100% (Table 1, Fig. 2). Of the 2255 New Zealand limestones analysed, only 6% are very high purity, 8% are high purity, 17% are medium purity, 24% are low purity and 45% are impure. Based on median carbonate values, three limestone populations, from the Cambrian to Early Cretaceous basement terranes, from the Chatham Islands and from the Eocene Ototara Limestone, are classified as medium purity. Five populations, from Nile, Te Kuiti, Whangarei, Mangaheia and Eocene Amuri units are low purity and the remaining ten populations are impure. Very high purity limestones only occur in nine populations and most of these occurrences are in the basement.

To compare the purity of onland New Zealand limestones with very pure limestone deposits around the world, we selected carbonate analyses from five studies of Cretaceous to Holocene calc-oozes and chalks from the Ocean Drilling Programme (ODP) literature and from Kitt's (1962) Cook Islands and Niue data. These carbonates were deposited in geographically isolated (Cooks–Niue, Ontong Java, Shatsky) and/or continental carbonate bank settings (Bahamas, Maldives, Kerguelen) that, for the most part, lack epiclastic or tuffaceous input. The mainly tropical and pelagic limestones are not directly comparable with the bryozoan, echinoderm, algal and foraminifera Oligocene temperate bioclastic limestones of New Zealand (Nelson 1978), but do represent what might be considered to be end-member ‘pure’ limestone populations. All the reference limestone populations are highly positively skewed and unimodal with median carbonate contents > 93 wt% (Fig. 2A, 2B, Table 1). However, even these datasets contain a measurable and distinct non-carbonate component, probably comprising terrestrial and extraterrestrial dust as well as siliceous bioclasts.

Basement and Chathams limestones

The Permian to Early Cretaceous limestones of New Zealand's Eastern Province accreted terrane basement (EP in Fig. 2C) are among the purest in the country. This can be explained in terms of their formation as reefs on subsiding intra-oceanic volcanoes, isolated from significant terrigenous input (e.g. Spörli & Gregory 1980; Silberling et al. 1988). The Ordovician limestones in the Western Province are sandwiched between siliciclastic strata and we attribute their high purity to an inferred origin in a carbonate rise setting, along with their relatively great thickness (>350 m, Cooper & Druce 1975). The Western Province limestones are also the only New Zealand population to contain substantial dolomite (Christie et al. 2001). Stylolite and vein formation, and recrystallisation during diagenesis and metamorphism may also have enhanced the purity of individual samples of some Cretaceous and older New Zealand limestones, at least on the scale of individual analysed hand samples.

Limestone deposition on the Chatham Islands was broadly coeval with known episodes of volcanism, although detailed examination shows that limestone accumulated during short intervals of volcanic quiescence (James et al. 2011). Limestones do grade into calcareous tuffs and the fact that our dataset of analysed limestones is relatively pure (Fig. 2D) can probably be explained by a general sampling bias to analyse purer limestones, James et al.'s (2011) antipathetic relationship between volcanism and limestone deposition and/or the geographic isolation of the Chathams from land, which limited any clastic input.

Late Oligocene limestones

Limestones of Late Oligocene to Early Miocene age are widespread throughout New Zealand and formed during the interval of maximum marine inundation of Zealandia (Landis et al. 2008). Populations from six areas are shown in the green histograms in Fig. 2E–2J). These have substantially lower median and larger interquartile ranges of carbonate content than either the basement or Chathams limestones, indicative of overall higher insoluble impurities. The western groups (Te Kuiti–Whangarei, Nile and Forest Hill) are demonstrably more pure than the eastern ones (Mahurangi, Weber, Otekaike), as shown by their higher median carbonate content and positively skewed distributions. Some, but by no means all, of the non-carbonate content in these limestones is due to the presence of glauconite (Ward & Lewis 1975; Table 2, Fig. 3). However, Fig. 4A shows that, at most sample sites of South Island limestones, most of the insoluble residue consists of extra-basinal sand and mud rather than intra-basinal glauconite. Figure 4B shows that the percentage of quartzofeldspathic sand (compared with other sand-sized bioclasts and glauconite) is both measurable and widespread. Previous petrographic studies of North Island limestones have also shown appreciable terrigenous sand and mud within Oligocene Te Kuiti Group limestones (Nelson 1977; Nelson & Hume 1987; Nelson et al. 1994; Hood et al. 2003a,b). Nelson & Hume (1987) reported a bimodal distribution of Te Kuiti Group limestone purity with a minimum at 70% soluble carbonate, a result weakly discernable in our separate dataset (Fig. 2E).

Pre- and post-Oligocene limestones

Visually prominent limestones of Plio-Pleistocene age occur in the East Coast region of the North Island (Mangaheia Group, formerly ‘Te Aute’ limestones). Limestone is also a minor component in Neogene and Paleogene stratal successions in other New Zealand sedimentary basins. These are shown as orange histograms in Fig. 2K–2P. The Eocene Ototara Limestone stands out as the most pure of these populations (Fig. 2P), but all the others have appreciable amounts of insoluble material. Speculatively, tuffaceous (ash) input may have been a major control on Tolaga and Mangaheia Groups limestone purity as they were deposited east and southeast of belts of subduction-related Neogene volcanism (Carter et al. 2004).

Oligocene–Miocene paleogeography

The Late Oligocene–Early Miocene (late Duntroonian stage to Waitakian stage, 26–23 Ma) is generally recognised as the time of maximum marine inundation (MMI) of the Zealandia continent (King 2000; Landis et al. 2008 and references therein). The median siliciclastic sand/total clastic contents of the Otekaike and Forest Hill limestones of this age are 9% and 6%, respectively (data of Table 2; see also Fig. 4). The consistently angular shape of the grains (Fig. 3) argues against recycling from older siliciclastic strata. Thus, clastic grains were seemingly being supplied to areas of Otekaike and Forest Hill limestone deposition, and being locked up as the limestones were deposited. The geographic distribution of siliciclastic sand content in Otekaike and Forest Hill limestones (Fig. 4B) does not show any clear pattern or gradient, e.g. towards a paleo-shoreline. However, there is some suggestion that Otekaike limestones become more pure to the east and that Forest Hill limestones become more pure to the south. The geographically intermediate and slightly younger Otakou Group limestones include the Milburn Limestone. Landis et al. (2008) noted that local schist basement could not have contributed K-feldspar and hornblende to the sand fraction of the Milburn Limestone but did admit the possibility of contemporaneous erosion of a distant and southwesterly source to supply clastic detritus during limestone deposition.

Figure 5 shows the secular change in Cenozoic limestone purity with time in the eastern North Island and the eastern South Island as revealed by our datasets. In both regions, the Late Oligocene limestones (Weber and Otekaike) are among the least pure although there is no dramatic shift in limestone purity from the Paleocene to the Pliocene, across the maximum flooding interval. A point to be taken from Fig. 5 is that, with the presence of demonstrable pre- and post-Oligocene land in New Zealand, the purity of the Oligocene limestones does not increase, as might have been the case if the sediment supply had decreased or ceased for any significant period.

Figure 5 Box and whisker plots showing changes in limestone purity with time in A, North and B, South Island datasets. Eastern populations shown in black and red; western populations shown in blue. Thick black line in box is median, ends of boxes are 1st and 3rd quartiles, tips of whiskers are 9th and 91st percentiles.

As mentioned above, the Oligocene to Early Miocene limestone populations of western New Zealand (Whangarei, Te Kuiti, Nile, Forest Hill) are slightly more pure than those in eastern New Zealand (Mahurangi, Weber, Otekaike–Weka Pass) (Table 1, Fig. 2, blue boxes in Fig. 5). Without petrographic data to supplement the soluble carbonate data, it is difficult to know how to interpret this observation. However, it does suggest an overall different paleoenvironment for the Oligocene of western New Zealand versus that in eastern New Zealand. If sediment supply is the reason then perhaps overall less (and/or more channelised and localised) sediment was reaching areas of limestone deposition to the west than to the east. Speculatively, the western marine shelves could have been more faulted than those to the east, creating offshore sediment traps separated by areas of purer limestone deposition. The eastern basins are, in general, deeper water and may have facilitated more continuous, more even dispersal of mud-grade terrigenous sediment and siliceous microfossil material to areas of limestone deposition. Lewis & Belliss (1984) have identified Late Oligocene karstic surfaces in the Otekaike Limestone that indicates at least local emergence during limestone deposition.

We tentatively interpret the appreciable siliciclastic angular sand content of widespread New Zealand Oligocene–Miocene limestones, as revealed by both chemical (Figs 2, 5) and petrographic (Figs 3–4) data, to indicate the presence of nearby contemporaneous landmasses that steadily supplied clastic sediment to the Zealandia continental shelves and slopes where limestones accumulated. How persistent, high or areally extensive the landmasses might have been is beyond the scope of this study. Systematic provenance work on a basin-by-basin basis, integrated with other geological datasets (e.g. limestone depositional setting, fossil assemblages, diagenetic observations), would provide useful further data on this, and also on the various controls of limestone purity.

Conclusions

Limestones are marine sedimentary rocks indicative of high biological productivity and limited clastic input. A digital compilation of 2255 New Zealand limestone chemical analyses has allowed us to compare the single, simple parameter of limestone purity (soluble carbonate content) of onland lithostratigraphic units across the entire country.

Our study shows that most onland New Zealand limestone populations are impure, low purity and medium purity, with individual very high purity samples present in only some formations. The purest New Zealand limestone populations have median CaCO₃ contents of 93–97 wt% and come from Cambrian to Cretaceous basement terranes, Eocene–Pliocene strata of the Chatham Islands and the Eocene Ototara Limestone. Most of these limestones were deposited in geographically isolated marine settings and have carbonate contents comparable to calc-oozes found in carbonate banks remote from clastic sources.

By contrast, the majority of New Zealand Paleocene–Pleistocene limestones, including those deposited in the Late Oligocene–Early Miocene maximum marine transgression interval, are of low purity and/or impure (median CaCO₃ contents of 71–91 wt%). In the case of the Late Oligocene–Early Miocene Otekaike and Forest Hill limestones, impurities are identified as variable terrigenous sand (and glauconite) content. Although other explanations are possible, we interpret the widespread occurrence of impure Oligocene–Miocene limestones to support the presence of land during maximum marine inundation of Zealandia.

The limestone purity dataset doubtless can be put to many further uses. Spatial clustering and variability in purity on a local scale may have limestone resource assessment implications. The relationship of limestone purity to karst landscape development, bioproductivity and fossil assemblages, bedforms and sedimentation rate, and to tectonic setting can all now potentially be explored in future studies.

Supplementary data

Supplementary file 1. Excel spreadsheet containing location, carbonate and stratigraphic name data used in this paper.

Acknowledgements

Discussions with Kathie Marsaglia and Daphne Lee helped formulate our ideas. The manuscript was improved by comments from Greg Browne, Phil Glassey, Kathie Marsaglia and Cam Nelson. We acknowledge the use of ‘R’ routines on Patrick Wessa's website (wessa.net) to plot histograms. Supported by the New Zealand Marsden Fund.