Tratamiento en sistema de lagunas.

The locations and details of the six records obtained from unexposed locations, used in this analysis, are given in Figure 5.1 and Table 5.1. The first four locations are near enough to the

Table 5.1: Details of unexposed locations (j= 1. . .6). The variablerj denotes the distance

from the vent (km), θj denotes the angular direction from the vent (radians anticlockwise

from east),ndenotes the number of tephras recognized in the location.

IDj Location Latitude Longitude rj θj n Age range (cal. yr BP) Source

1 Lake Umutekai 39◦05.27’S 174◦08.22’E 24.15 1.34 104 1400 - 11500 (Turner et al., 2008a)

2 Lake Rotokare 39◦_{27.10’S 174}◦_{24.64’E 34.49 -0.49 42 490 - 7050 (Turner et al., 2009)}

3 Near Source 39◦_{16.59’S 174}◦_{06.09’E 3.59 0.44 23 94 - 2200 (Turner et al., 2009)}

4 Eltham Swamp 39◦25.52’S 174◦20.09’E 27.83 -0.54 177 3000 - 36900 (Tinkler, 2013)

5 Lake Rangatauanui 39◦_{26.04’S 175}◦_{22.45’E 114.39 -0.11 37 2100 - 24500 (Moebis, 2010)}

6 Auckland 276.29 1.37 40 9800 - 75000 (Green et al., 2014)

- Lake Pupuke 36◦47.25’S 175◦46.25’E 286.00 1.38 15 (Molloy et al., 2009)

- Onepoto Basin 36◦_{48.29’S 174}◦_{45.02’E 282.72 1.38 21 (Shane and Hoverd, 2002)}

- Orakei Basin 36◦52.00’S 174◦48.40’E 277.62 1.36 18 (Molloy et al., 2009)

- Hopua Crater 36◦_{55.46’S 174}◦_{47.05’E 270.24 1.36 11 (Molloy et al., 2009)}

- Pukaki Crater 36◦_{58.58’S 174}◦_{48.37’E 265.01 1.35 28 (Sandiford et al., 2001;} Shane, 2005)

volcano (<35 km) to potentially record both minor unnamed eruptions and those producing

larger tephra falls. The remaining two unexposed sites are located much further from the

volcano (> 100 km). Although the distal sites include a wide age range of tephras, due to

their proximity, they are less likely to contain a given tephra. The thinnest tephra observable in a core is 0.5 mm.

The Auckland record (the sixth record in the analysis) arises from the combination of multiple

tephra records collected from within maar depressions in Auckland, New Zealand. A maar

is a crater formed by a volcanic eruption that is (typically) subsequently filled with water and sediment to become a lake or a swamp. This analysis uses the detailed tephra records established in Molloy et al. (2009) from Lake Pupuke, Hopua Basin, and Orakei Basin. In Chapter 4 these records were combined with tephra records from Onepoto Basin (Shane and Hoverd, 2002) and Pukaki crater (Sandiford et al., 2001; Shane, 2005) to identify 40 distinct Mt Taranaki events. The Auckland sites are located more than 260 km from Mt Taranaki

and only contain very fine tephra (∼2 mm thick). To avoid letting the distal data dominate

the analysis, the records from the five Auckland sites are aggregated. The average tephra thickness observed, the average distance from the vent, and the average azimuth of the five Auckland sites in Chapter 4, provide the sixth record for this analysis.

The approximate age ranges in Table 5.1 indicate an age overlap between many of the cores. A particular recorded eruption may have left a deposit at all of the sites, or any subset. The local search algorithm presented in Chapter 4 is utilized to find the most likely arrangement of tephras observed across the six unexposed locations. In addition to thickness measurements, the records contain geochemical data (where available), and estimated ages and associated

age errors for each tephra. The age and associated age errors form the basis of the matching algorithm, and arrangements are constrained by considering the geochemical compositions and stratigraphic relationship among tephras.

In Chapter 4 the matching algorithm was illustrated through application to the five Auckland records discussed above. The only geochemical constraint applied there was that tephras must be sourced from the same volcano. In the case of the records used here however, all of the tephras are sourced from Mt Taranaki so that constraint is no longer applicable. There is, however, more detailed geochemistry data available in the form of titanomagnetite chemistry. Employing the approach of Turner et al. (2009), matches among tephra are permitted or

ruled-out through principal component analysis on the major elements (TiO2, Al2O3, MgO)

of the titanomagnetite compositions. The following principal components (PCs) are computed for each sample from each tephra:

PC1 = 0.18 TiO2−0.71 Al2O3−0.68 MgO, (5.1)

PC2 = −0.96 TiO2+ 0.03 Al2O3−0.28 MgO. (5.2)

These PCs differ from those in Turner et al. (2009) because of the inclusion of the Eltham Swamp and Lake Rangatauanui tephras. Not all of the tephras have geochemistry data

available. In particular none of the tephras sourced from the Auckland region have titano-

magnetite geochemistry available. While the Auckland tephras do have geochemistry data available, it is in the form of glass chemistry so is not comparable to the other records. An eruption consists of many geochemistries; different geochemistries occur at different phases of eruption, so can be found in different combinations at different sites. Ellipses, centered at the mean principal component values, with axis length equal to four times the standard deviation, are constructed for each tephra. Only candidate age matches among tephras with overlapping ellipses are permitted.

Figure 5.3 shows an example for a section of the record aged 5500 - 6000 cal yr BP. Candidate matches among three or more tephras are allowed, provided there is at least one tephra with geochemistry consistent with all other tephras. For example, a match between Rotokare-35 and Eltham-19 would be permitted only with the inclusion of Umutekai-52. This is keeping with the nature of a magma as an aggregate of individual magmas. As deposits in different cores tend to reflect different wind conditions, they are often from different phases of an eruption. The requirement of at least one overlapping tephra ensures that there is a temporal

Figure 5.3: Geochemistry of tephras aged 5500 - 6000 cal yr BP. Labels indicate the source (Lake Umutekai (Um), Lake Rotokare (Ro), and Eltham swamp (El)) and the number for

each of the tephras.

−3 −2 −1 0 1 2 3 4 5 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 Um−49 Um−52 Ro−32 Ro−33 Ro−35 El−18 El−19 PC1 = 0.18TiO₂ − 0.71 Al₂O₃ − 0.68 MgO PC2 = −0.96TiO 2 − 0.03 Al 2 O 3 − 0.28 MgO

magma evolution path through the eruption. It also gives the matching algorithm greater flexibility and ensures that the search for the optimal arrangement is not over-constrained. In any case, the estimated ages for the tephras provide the true control over the arrange- ment. The geochemistry constraints serve to rule out infeasible matches to which we take a conservative approach.

Checking the geochemical correlations among tephra is done automatically through adapta- tion of the automated matching procedure in Chapter 4. There are a total of 270 distinct events identified by the matching algorithm. The final arrangement is given in Appendix D.