Remote Sensing Geomorphology of Akaroa Volcano and Detailed Mapping of Okains Bay, Banks Peninsula, NZ

Transcription

1 Remote Sensing Geomorphology of Akaroa Volcano and Detailed Mapping of Okains Bay, Banks Peninsula, NZ Daniel J. Hobbs 1,2 1. Department of Geological Sciences, University of Canterbury, Christchurch, New Zealand 2. Department of Geology, Middlebury College, Middlebury, VT USA Abstract Remote sensing geomorphology of the highly eroded cone of a Miocene composite volcano in conjunction with detailed geologic mapping of a subsequent radial valleys provides insight into the controls on bay geomorphology at Okains Bay and similar valleys on Akaroa Volcano, Banks Peninsula, South Island, New Zealand. Akaroa Volcano is one of two large Miocene volcanoes forming Banks Peninsula. The other volcanic structure, Lyttelton Volcano, has been well studied and is known to be a complex assemblage of basaltic and hawaiite lavas likely originating from over a dozen major eruptive centres. The previous interpretations of Akaroa Volcano consider the structure to be a rapidly formed, singular cone primarily composed of a single, highly variable formation. Digital terrain models (DTM) and primary volcanic features of Akaroa were used to identify 10 potential volcanic summits. These summits are associated with distinct sectors of the cone flank and suggest multiple eruptive sequences sourced from separate eruptive centres. The morphology of cone sectors seems to control the location and orientation of large radial valley and ridges, thereby affecting the geomorphology of bays such as Okains Bay. Detailed geologic mapping of Okains Bay does not preclude the findings of remote sensing geomorphic analysis, which suggest the Okains Bay area marks the transition or intersection between two distinct cone sectors. Introduction Akaroa, an extinct and heavily eroded Miocene composite shield volcano, forms the largest portion of Banks Peninsula as the largest and most prominent volcanic feature in the Canterbury Province of New Zealand (Figure 1a). Okains Bay is one of the many radial bays and inlets which ring Akaroa (Figure 1b). Since the first thorough geologic publication on Banks Peninsula by Julius von Haast in 1879, numerous studies have described Akaroa s structure (Shelley, 1988), stratigraphy (Sewell, 1988), and geochemical history (Stipp & McDougall, 1968; Price & Taylor, 1980; Sewell, 1988; Timm et al., 2009; Hartung, 2011). However, much of this research has focused on Akaroa Harbour, leaving most of Akaroa, including Okains Bay, relatively unstudied. Significant work to determine the exact location and extent of lavas, vents, and volcanic features has yet to be completed. Likewise, the factors which control the peninsula s present morphology, particularly the formation and location of modern bay, are poorly understood. The need for coherent, consolidated

2 2 D.J. Hobbs research on Akaroa is clearly evident by the incredibly low level of existing geologic maps (Figure 1a). This paper provides a detailed look at the stratigraphy of Okains Bay and the current geomorphology of Akaroa to better understand the eruptive and erosive history of the volcano. Remote Sensing Geomorphology Akaroa Volcano is an excellent subject for remote sensing analysis as the area is well preserved except for erosion strongly controlled by the original structure. As with many older volcanoes, chemical alteration and erosion of the central magma conduits have caused the flanks of Akaroa Volcano to be significantly less eroded than the central region (Figure 1). For this reason, this study largely avoids structures found within the central zones and instead relies on the better preserved flank features to provide more accurate information on original volcanic structures. Remote analysis of DTM models of both active and ancient volcanic edifices has proven useful for extrapolating past structures from imagery of current structures at a number of volcanoes (Székely & Karátson, 2004), including the adjacent Lyttelton volcano (Hampton & Cole, 2009). For long extinct cones in particular, these techniques are highly useful for identifying radial patterns of topographic (ridge and valley) and volcanic (dyke, lava flow, cone sectors) structures that are hypabyssal, constructional, and erosional in origin. Akaroa is known to host similar topographic and volcanic features which are used in this paper to evaluate patterns identify volcanic summits, eruptive centres, and secondary cone sectors. This paper uses the primary volcanic landform identification scheme outlined by Hampton and Cole (2009) and summarized below. Primary Volcanic Feature Identification Primary volcanic features form during and after volcanic activity and are closely related to the original morphology of a volcano just prior to its extinction (Székely & Karátson, 2004). These structures can be categorized by their origin into constructional, erosional, and hypabyssal features. Constructional volcanic features form externally as a result of volcanic activity during the growth of a volcano. Lava flows from the central vent are the primary constructional feature of large volcanoes; however, flows from parasitic vents, pyroclastics, and scoria cones also fall within this category. Hypabyssal volcanic features such as dykes and sills are those which form internally during volcanic activity, usually as a result of magma injections into areas of weaknesses related to gravitational forces. Dyke swarms near the surface can form radially from a cone structure as the increase in localized gravitational forces focus on eruptive centres, creating a local weakness into which magma can be preferentially injected. Most important to this study are erosional volcanic features, which can form early in a volcano s life and become accentuated with time. During the active phase of a volcano, drainage pathways are strongly influenced by cone morphology, extending radially from

3 3 D.J. Hobbs eruptive centres. These pathways eventually form well defined valleys and, consequently, ridges. As degradation of the volcanic edifice continues after extinction, radial valleys and ridges often become the most prominent erosive features (Hampton and Cole, 2009). One type of erosional feature common to volcanic cones is the planeze (Figure 2). Planezes are triangular dips slope features found on the flanks of eroding volcanoes and form from the incision of drainage pathways at their margins (Cotton, 1944). While drainage pathways on the surface and alongside planezes often extend radially from a volcanic summit, the escarpments at the edges of planezes usually do not point to an eruptive centre. Instead, escarpments converge at a point or merge into a ridge which is oriented toward a volcanic centre. The need to separate radial from non-radial erosional features associated with planezes can complicate the process of summit identification. A firm understanding of planeze formation and identification is therefore necessary for accurate interpretation of remote sensing geomorphology of eroding volcanoes Primary Volcanic Feature Analysis Remote sensing geomorphology analysis for this study was conducted in the ~30km by ~30km area roughly east of Lake Forsyth and Pigeon Bay (Figure 3). Primary erosional axes were identified from relatively straight valley and ridge segments on the eroded flank of Akaroa at least 1km in length. Hillshade (Figure 1a) and aspect models (Figure 3) of slopes on Akaroa were constructed in ArcGIS from a 10m resolution DTM and used to identify the location and orientations of ridge and valley features. The selection of radial flank structures for this analysis was determined at the author s discretion. This slightly subjective approach allows for misleading features such as planezes outlines, secondary erosional features, tributary drainage paths, and other noise to be ignored in favour of primary erosional landforms. With the primary erosional axes identified, a best fit line was applied to represent the general orientation of each feature. Best fit lines were then projected no more than 16km (approximate radius of Akaroa cone) inward towards the central region of Akaroa. Zones of convergence, defined by where five or more primary erosional feature orientation projections converged within a 500m by 500m region, were then identified (Figure 3). Group-based Geologic Mapping Detailed mapping of Okains Bay (Figure 4) is used in this study as an in-depth example of valley system formation at Akaroa Volcano. Most of the field data collected for this map was conducted by small groups of students. Over the course of several days, each of these teams closely studied volcanic deposits along ~1km stretches of cliff and hillside in Okains Bay (Figure 4). The mapped areas include the lower cliff face adjacent to the thin, several kilometre long lagoon; Chorlton Road on the hillside above the lagoon; the cliffs and lower hillside from the end of Okains

4 4 D.J. Hobbs Bay Road to the small island in the middle of Okains Bay; and the cliffs bounding the two smaller bays in the southern corner of Okains Bay (referred to as Little Okains Bay to the west and Hermit Bay to the east) (Figure 4). At these locations, students recorded the extent and appearance of major volcanic deposits and while collected samples (Figure 5) from these deposits for geochemical analysis. Due to the limited mapping extent and the large number of students participating, students were able to confidently map at a very fine resolution. Rock samples for geochemical analysis were also collected from domes near Okains Bay, including Goat Rock, Panama Rock, and Le Bons Bay Peak as well as from outcrops along the high central ridge that separates Little Okains Bay from Okains Bay proper (Figure 5). A number of samples were taken from lavas and dykes along the lengths of Big Hill Rd and Cameron Track as well as in the Panama Rock and Goat Rock areas (Figure 5). The process of digitizing each field maps also involved significant collaboration with and within each small group. Teams were responsible for scanning their field maps into ArcGIS Desktop and georeferencing them. Students then replicated and edited their digitized maps. Digitized sections and field notes were collected from each group to produce a composite map, and samples were catalogued in preparation for geochemical analysis. XRF major and trace element analysis was processed in conjunction with Joshua Johnson (2012, unpublished manuscript) and subsequently used to correlate distinct lava flows and deposits by their geochemistry (Figure 6). In some locations maps from different groups overlapped and conflicted. These problem areas were all located along the margins of groups mapping areas and were thus easily resolved by consulting with students from each groups. The geologic focus of the field mapping exercises conducted by students was on volcanic deposits in the hills and headwalls of Okains Bay. For additional context of the area, four ages of Quaternary beach and dune ridges in the main Okains valley (Stephenson & Shulmeister, 1999) were included in the final map (Figure 4). Results Remote Sensing Geomorphology Primary erosional features highlight ten zones of convergence (Figure 3), many of which are located inside the central valley of Akaroa. A total of 92 valley and ridge projections are used in the analysis, 72 of which are part of a zone of convergence. An average of ~7 projections associated with each zone of convergence. In some instances, projections of certain axes orientations are part of two different zones of convergence. Lava overprinting, as explained later, is thought to be the primary cause of this overlap.

5 5 D.J. Hobbs Okains Bay Mapping More than 30 lava flows were identified based on structural and textual evidence (Figure 4). The flows exhibit a gentle northeastwardly dip which is oriented a few degrees more east than the major valley structures defining the area. The discrepancy between orientations of stratigraphy and erosional features contributes to an overall southeastward younging direction. Based on unit descriptions, some flows were correlated between mapping groups. Deposits are defined by a wide range of structures, textures, and compositions (Appendix 1). Lava flows up to several meters thick range from pahoehoe to blocky, massive to heavily brecciated, apheric to porphyritic, and picrite basalt to mugearite (Figure 4). Scoria deposits are exposed in numerous, several meter high sea caves which line base the cliffs of headwalls near Little Okains, and along the Okains Lagoon, and extend about 500 meters inland on the southeastern wall of the Okains valley. For full descriptions of all these units, see Appendix 1. Discussion Eruptive Centre Locations As suggested by previous studies of other volcanoes (Hampton and Cole, 2009), eruptive centres are likely to have been located where projections of several nearby primary erosive axes converge. Having identified ten zones of convergence around central Akaroa region (Figure 3), the possibility of past eruptive centres at these locations can be assessed. Many zones of convergence coincide with large trachytic structures and deposits compatible with explosive volcanism. Located within zone 3 (Figure 3) is Pulpit Rock, a prominent trachytic body which may be a preserved plug and source of cone 3 lavas. A cursory observation of the small peninsula and southern coastline of Anchorage Bay (Figure 1b) near Location 5 (Figure 3) reveals extensive scoriaceous deposits indicative of early stage explosive activity. If this was indeed an early eruptive centre, it is quite likely that activity continued in this area and contributed to the lavas comprising cone 5. Based on the scoriaceous deposits within French Bay (Figure 1b), the eruptive centre of cone 6 appears to be another site of early activity. The southern coastline associated with this cone appears to follow an arc different from the more distal cone 7 (Figure 3) (Hartung, 2011; Trent, in prep). A distinctive, ~20m thick lava flow runs along the crest of southeastern ridge of Akaroa (Figure 7). These lavas appear to top the ridges associated with cone 7 (Figure 3), indicating the stratigraphically young ridge lavas are a product of eruptive centre 7. Centre 8 (Figure 3) is located ~700m north of Purple Peak, another prominent trachytic feature. Purple Peak could in fact be related to eruptive centre 8 as a secondary vent within a larger vent complex. Alternatively, the lavas of cone 8 sector could have been sourced from Purple Peak but consistently diverted away from

6 6 D.J. Hobbs their initial radial path by a prominent topographic feature during their emplacement. In the North Island of New Zealand, the historical lava flows of Ngauruhoe take a sharp ~90 turn when they reach the lower northern flank, creating L shaped flows (Hobden et al., 2002) which could serve as a more extreme analogue for the lavas of cone 8. Centres 1 and 2 (Figure 3) sit in a large valley system, confined by roughly circular erosional rim. This valley system is completely separate from the other centres and does not match the typical flank morphology of eastern Akaroa. These centres could be part of another large cone structure that borders the main Akaroa cone along the north-south trending ridge between them. South of this dividing ridge, is a north-northeast trending valley which divides opposing flanks associated with the Little River valley system to the west and the Akaroa Valley system to the east. The geomorphological signals associated with valley and ridge systems in the northern Akaroa region between Pigeon Bay and Bay (Figure 3) are noticeably complex and do not clearly indicate the position of one or more volcanic summits. This area is at the margins of the Akaroa cone and as with the marginal eruptive centres 1 through 4, the erosional axes projections here are not as well oriented towards the centre of Akaroa as those to the southeast (Figure 3). The type of erosive axis identification analysis used by this study is intended for conical volcanic structures and in areas like northern and northwesten Akaroa, the applicability of this analysis is limited. What can be inferred, however, is that this complexity reflects the combined influence of cone-building lavas from Akaroa and more complex structures generated by earlier and contemporaneous volcanism to the west, in the Mt. Herbert region (Sewell, 1988; Timm et al., 2009). Drainage pathways through Mt. Herbert Group deposits conceived prior to the emplacement of cone-building Akaroa lavas probably remained active and relatively undiverted throughout the growth of Akaroa. Although erosional axes projections produce an imprecise, noisy signal in this region, it is appears that eruptive centres may have existed in the large valley inland of Pigeon Bay and the slopes north of Barrys Bay (Figure 1b, 3). Cone Sector Geomorphology The probable lateral extent of cone sectors for each zone of convergence were identified (Figure 8) based primarily on the locations and orientations of valley and ridge features associated with the zone (Figure 3, 9). The basal footprint of cones sectors are defined by short sections of arcuate coastline which extended different distances from the centre of Akaroa. The inward extent of cone sectors is considered to represent the erosional crater rim of the sector and is partially defined by the larger erosional rim surrounding central Akaroa (Figure 8). In many instances, erosional axes and their cone sectors appear to overlap (Figure 8). This is because some valley and ridge systems record the cone morphology of more than one eruptive centre. Lava sequences associated with a specific volcanic summit should produce radial valleys and

7 7 D.J. Hobbs ridges as they erode. Younger lavas will either follow these older drainage pathways or produce new pathways in a different orientation. Depending on the extent of overlap and erosion, a number of complex patterns can occur. Where younger, overprinting lavas do not extend, radial erosive systems from older cones will be preserved. Thus, the orientation of a valley axis may be controlled by younger eruptive centres at the top of a valley system and older systems lower down. That is to say, the age of valley axis orientations can transition from oldest at the base of a valley to youngest at the top. (Behncke and Neri, 2003; Branca, 2003) This phenomenon is especially apparent in the southeast region of Akaroa where the overlap of cone sectors produces valley systems with orientations which alternate between north-south and northeast-southwest (Figure 3). Where cone sectors overlap it is also expected that, based on simple geometry, relatively large valleys form at their intersection. These valleys inherently serve as drainage pathways and are enhanced with time. Signs of these phenomenon are apparent in the areas between cone sectors 1/2 and 3/4 (Peraki Bay), cones 7 and 8 (Otanerito Bay), and cones 8 and 9 (Hickory Bay) (Figure 1b, 9). The likely existence of intercone valleys at Akaroa provides further evidence for a distinct multicone or subcone growth pattern. Formation of Okains Bay Significant differences in the orientations of features east and west of Okains Bay (Figure 3) suggest that the erosional axes on either side are controlled by different volcanic summits. The fact that a large valley exists at Okains further suggests that this is an area at the margin or intersection of two distinct cone sectors (Figure 8). This interpretation does not require the existing lavas to be from different eruptive centres as lava overprinting by the top-most deposits can produce erosional axes orientations that are different than of older lavas below. The structural complexity of an area at the margins or intersection of cone sectors means that interpretations of the relative origin of units along the southern margin of Okains Bay from remote sensing geomorphology alone are inconclusive. Extensive geochemical analysis by Johnson (2012, unpublished manuscript) has found little difference in composition between lavas on the east and west side of the Okains valley. Stratigraphic trends of increasing total alkali-silica ratios on either side of the valley imply rapid evolution of several distinct magma batches from picritic basalt to mugearite with time. These findings, however are also inconclusive as to the relative origin of lavas which were mapped in detail. Implications for Akaroa s Formation and Erosion A regional northeast-southwest trend of volcanism is observed from the locations of the ten identified eruptive centres and the two regions which likely hosted similar vents. This trend fits well

8 8 D.J. Hobbs with the observations which indicate volcanic activity migrated progressively from Lyttelton in the northwest to Akaroa in the southeast (Stipp & McDougall, 1968; Price & Taylor, 1980; Sewell, 1988)over the period between ~12.3 Ma and ~6.8Ma (Timm et al., 2009). The idea of multiple eruptive centres also works well with many other key models of volcanic evolution applied to Banks Peninsula. Lyttelton volcano, similar in origin, structure, and composition is thought to have multiple eruptive centres (Hampton & Cole, 2009). It is a reasonable assumption that Akaroa could have a similar dispersal of vent structures, especially given the geomorphologic evidence. Johnson (2012, unpublished manuscript) proposes that deposits on Akaroa record the evolution of small magma batches which evolve separately but similarly from basaltic to trachytic compositions. The surface expression of discrete magma batches is very likely the distinct volcanic centres identified by this study. Akaroa Volcano has a distinct structure that is defined by a deeply eroded central region which extends southward beyond the erosional crater rim as Akaroa Harbour (Figure 1b). It is assumed that there existed more, larger, and/or older eruptive centres near the centre of Akaroa which could explain the increased erosion in this area and the general conical shape of the volcano. Erosional axis analysis does not clearly indicate a high concentration of eruptive centres (Figure 3); however, the fact that the entire northwestern region of Akaroa is unsuitable for this kind of analysis must account for some of this discrepancy. Perhaps a more interesting is why Akaroa Harbour formed where it did. The first clue is the relatively small volume of remaining deposits in the area associated with cones 5 and 6 (Figure 8). The area is at the end of the southeastward volcanic trend associated with Banks Peninsula (Figure 8) and probably did not experience as many overlapping series of lava deposition as areas to the northwest. The considerable distance between eruptive centres 5 and 6 (Figure 3) also suggest that relatively little lava reached the southern end of the harbour. With a small amount of lava input, the area between cones 5 and 6 was likely always a low point in the greater Akaroa cone structure. As a natural low point between two cones, this area would have been a major drainage pathway, creating a positive feedback loop for continued erosion in this area. Based on the distant eruptive centre locations (Figure 3) and inferred low output of cones 5 and 6 (Figure 8), it seems inevitable that Akaroa Harbour formed to the south. Future Research In addition to providing a new, comprehensive geomorphologic interpretation of the patterns of growth and erosion at Akaroa, this study highlights numerous opportunities for future research. The logical first priority is a more definitive control on the locations of eruptive centres associated with Akaroa. Locating specific vent structures enables the matching of vent and flow

9 9 D.J. Hobbs lithologies as well as provides a more detailed understanding of Akaroa geomorphology. Areas with complex erosional axes systems, namely northern Akaroa (Figure 3), are of particular interest for this kind of close scrutiny. Additionally, significantly more work could be devoted to understanding how the age, composition, and eruptive centre of deposits in Akaroa relate to the erosional axes orientations of valley and ridge structures over time. A comprehensive understanding of these relationships would allow for the creation of a detailed growth model showing when and where new eruptive centres and their deposits were emplaced. Conclusions Remote geomorphologic analysis of digital terrain models is a powerful technique for interpreting the existing structure of a heavily eroded volcanic edifice. From this analysis, the approximate locations of ten likely volcanic summits (Figure 3) are identified and interpreted to represent the existence of numerous eruptive centres within the proximal region of Akaroa. These centres are thought to have produced structurally distinct cone sectors with their own set of radial lavas and drainage pathways (Figure 9). The extent of cone sectors (Figure 8) is extrapolated from the locations of associated valley and ridge systems, coastlines with maximum extents that follow a similar arcuate pattern, and the morphology of Akaroa s erosional crater rim. The northwestsoutheast trend of distinct eruptive centres (Figure 3) is consistent with previous models of volcanic structures, migration of surface activity, and magma evolution on Banks Peninsula. From eruptive centre and cone sector locations (Figure 9), the major factors controlling the current morphology of Akaroa are inferred. Where cone sectors overlap, deep and often complex valley systems emerge as a result of focused drainage and erosion (Figure 8). These complex end products are observed in several locations on the flanks of Akaroa, including Okains Bay. The relative input of lavas into a specific region is thought to affect the size and morphology of valley systems. Akaroa Harbour (Figure 1) is thought to be one such example of a deep valley forming where lava supply is significantly limited in an area between two cones. This research can serve as the basis and impetus of future studies of Akaroa focused on refining the location of eruptive centres and understanding their role in the structural evolution of the volcano. References Behncke, B., Neri, M. (2003). The July August 2001 eruption of Mt Etna (Sicily). Bull. Volcanol. 65,

10 10 D.J. Hobbs Branca, S., (2003). Geological and geomorphological evolution of the Etna volcano NE flank and relationships between lava flow invasions and erosional processes in the Alcantara Valley (Italy). Geomorphology, 53, Cotton, C.A. (1944). Volcanoes as landscape forms. Whitcombe & Tombs limited, Christchurch, New Zealand Haast, J. von. (1879). Geology of the Provinces of Canterbury and Westland, New Zealand: a report comprising the results of official explorations. NZETC. Christchurch: Times. Retrieved from Hampton, S. J., & Cole, J. W. (2009). Lyttelton Volcano, Banks Peninsula, New Zealand: Primary volcanic landforms and eruptive centre identification. Geomorphology, 104(3-4), Elsevier B.V. doi: /j.geomorph Hartung, E. (2011). Early magmatism and the formation of a Daly Gap at Akaroa Shield Volcano, New Zealand. Hobden, B., Houghton, B., & Nairn, I. (2002). Growth of a young, frequently active composite cone: Ngauruhoe volcano, New Zealand. Bulletin of Volcanology, 64(6), doi: /s Price, R., & Taylor, S. (1980). Petrology and Geochemistry of the Banks Peninsula Volcanoes, South Island, New Zealand. Contributions to Mineralogy and Petrology, 72, Sewell, R. J. (1988). Late Miocene volcanic stratigraphy of central Banks Peninsula, Canterbury, New Zealand. New Zealand Journal of Geology and Geophysics, (February 2012), Shelley, D. (1988). Radial dikes of Lyttelton Volcano their structure, form, and petrography Radial dikes of Lyttelton Volcano - and petrography. New Zealand Journal of Geology and Geophysics, 31, Stipp, J., & McDougall, I. (1968). Geochronology of the Banks Peninsula volcanoes, New Zealand. New Zealand journal of geology and, Retrieved from Székely, B., & Karátson, D. (2004). DEM-based morphometry as a tool for reconstructing primary volcanic landforms: examples from the Börzsöny Mountains, Hungary. Geomorphology, 63(1-2), doi: /j.geomorph Timm, C., Hoernle, K., Van Den Bogaard, P., Bindeman, I., & Weaver, S. (2009). Geochemical Evolution of Intraplate Volcanism at Banks Peninsula, New Zealand: Interaction Between Asthenospheric and Lithospheric Melts. Journal of Petrology, 50(6), doi: /petrology/egp

11 11 D.J. Hobbs 329 Figure 1. A) Simplified geologic map of Banks Peninsula (modified from Hampton and Cole, 2009) within the context of New Zealand (inset). B) Hillshade of Banks Peninsula created from a 10m by 10m resolution DTM.

12 12 D.J. Hobbs 330 Figure 2. Cartoon of planeze formation (modified from Cotton, 1944).

13 13 D.J. Hobbs 331 Figure 3. Slope aspect map of Akaroa highlighting the convergence of projected valley and ridge orientations. Bold lines represent primary erosional valley (red) and ridge (green) orientations and thin, black lines are projections of these orientations. Numbered boxes are where five or more projections converge within a 500m by 500m square, representing an inferred topographic high and possible eruptive centre. Contours have 50m spacings from sea level. 1km grid spacing.

14 14 D.J. Hobbs 332 Figure 4. Detailed geologic map of Okains Bay in which individual lava flows are outlined and also grouped by composition in colour. Field mapping was conducted by students in the 2012 Frontiers Abroad programme and digitized for this paper. Known compositions were determined by XRF analysis of samples within those units and assigned rock types according to. Inferred rock types are extrapolated from the magma batch model proposed by Johnson (2012, unpublished manuscript) which describes small, rapidly evolving magma batches which produce short basalt through benmorite lava sequences that are stratigraphically repeated with each new batch. Radiocarbon and four ages of Quaternary beach and dune ridges within the Okains Valley are from Stephenson & Shulmeister, (1999). Thick grey lines represent lava flows identified exclusively from aerial photographs but not observed in person.

15 15 D.J. Hobbs Figure 5. Sample locations of geochemical analysis. The location of samples from Le Bons Bay Peak and Goat Rock are generalized and are in fact more spread out than shown.

16 16 D.J. Hobbs Figure 6. Rock type geochemistry of samples from the Okains Bay area (Figure 5) (Johnson, unpublished manuscript). Most samples plot within the picrite basalt to mugearite range. Samples with benmorite signature were taken from dyke-cut scoria while samples with trachyte signatures are from Panama Rock. Figure 7. Google Earth image of southeast Akaroa. Highlighted is the large lava flow forming the valley ridge above Akaroa Township and radial ridges extending towards the Flea Bay area. These lavas are thought to be associated with eruptive centre 7 (Figure 3).

17 17 D.J. Hobbs Figure 8. Cone sectors of each eruptive centre. Cones 1/2 and 3/4 represent cones for two eruptive centres.

18 18 D.J. Hobbs

19 19 D.J. Hobbs Figure m by 500m zones of convergence of erosional valley and ridge projections within Akaroa and their associated cone sectors.