Abstract
The numerous samples of fossil wood in Jurassic to Miocene sediments in New Zealand remain poorly understood and could potentially provide a better understanding of forest paleoecology and paleoclimatology in the region. More than 100 samples of fossil wood were sourced from 24 sites in the South Island, Chatham Islands, and Auckland Islands, ranging in age from the Jurassic to Miocene. This material was sectioned using hard rock thin sectioning methods for each of the three standard orientations for wood identification. Three sites from the Jurassic (Catlins, Low Hills and Productus Creek) contained fossil wood. Material from the Catlins was identified as having both abietinean and araucarian radial pitting, though as cross-fields were not well-preserved, further identification was not possible. The sample from Productus Creek was identified as Araucarioxylon Kraus (=Agathoxylon Hartig.). Material from Low Hills was identified as Planoxylon Stopes, the first time this genus has been identified from the Jurassic of New Zealand, extending its history in New Zealand back approximately 60 million years.
Eight sites from the Cretaceous were investigated (Pitt Island/Rangiauria, Conway River, Mount Somers, Oaro, Port Robinson, Quail Flat, Roa Mine and Whakapohai River). New wood material from the mid Cretaceous Tupuangi Formation at Waihere Bay, Pitt Island/Rangiauria, indicated a more diverse conifer flora than previously identified. Samples from Waihere Bay yielded a range of gymnosperm genera, including Agathoxylon, Protophyllocladoxylon Kräusel, Cupressinoxylon Göppert and abietinean-type woods. Samples from Conway River were identified as Planoxylon and an abietinean-type wood. One sample from Mount Somers had features that indicated it was Agathoxylon. Samples from Oaro had features attributable to Planoxylon and Agathoxylon. One sample from each of the Port Robinson, Quail Flat, Roa Mine and Whakapohai River sites were determined to be ‘Group C’, abietinean-type, Taxodioxylon Hartig and Agathoxylon respectively. This is the first record of Protophyllocladoxylon Kräusel and Cupressinoxylon from the Cretaceous of New Zealand, as well as the first record of Taxodioxylon from Cretaceous deposits in mainland New Zealand.
No new samples from Paleocene sites were investigated.
Two sites from the Eocene, Hampden Beach and Boulder Hill, each yielded a single sample. The sample from Hampden Beach was identified as Palmoxylon, but did not have the features necessary for further identification. The material from Boulder Hill was identified as legume, possibly subfamily Caesalpinioideae, the first Eocene record of legume fossil wood in Australasia. The Eocene is an important time in the diversification of the Leguminosae, and this sample has features quite unlike those found in modern New Zealand representatives of this family.
Two Oligocene sites were studied. The Cosy Dell Lime Pit yielded unidentifiable angiospermous material, while material from Pomahaka was identified as Agathoxylon, the first record of this genus from Oligocene aged strata in New Zealand.
Miocene sites from the South Island included: Burnside Quarry, Cosy Dell Coal Pit, Karitane, Macraes, New Vale, as well as several sites with outcrops of Kowai Gravels. Material from Burnside Quarry was identified as Casuarinaceae (either Casuarina L. or Allocasuarina) and samples from the extensive Kowai Gravels were identified as Myrtaceae (Eucalyptoxylon) and Scrophulariaceae (cf. Myoporum). Material from Karitane was tentatively attributed to Violaceae (Melicytus). Gymnosperms were identified from the Cosy Dell Coal Pit, Macraes and Newvale. Material from the sub-Antarctic Adams Island, Auckland Islands was also studied and identified as Araliaceae (possibly Pseudopanax or Raukaua). This is the first fossil record of wood from Araliaceae, Scrophulariaceae, and Violaceae from New Zealand and the first definitive record of Myrtaceae fossil wood from the South Island.
In addition to the identification of new fossil wood material from sites in southern Zealandia, the viability of using data from fossil wood material to determine paleoclimate was tested, using two methods. The first, the coexistence method, is based on the environmental range of the nearest living generic relatives for fossil species. This method utilised the known fossil species from the Miocene Landslip Hill, Southland and was conducted using two different sets of data. The first dataset was from the Paleoflora Database, which contained the temperature range of the modern species; the second dataset was the geographical range of the modern species, obtained from the Plants of the World Online database. Both resulted in estimates for mean annual temperature (MAT) that were within or slightly warmer (18–20 °C and 20–25 °C respectively) than previous estimations for the MAT of the Miocene of New Zealand (17–21 °C). It is likely that with further study to identify more of the fossil genera present at the site in the Miocene and refinement of the nearest living relatives to species level, rather than generic or familial level, that this method would prove useful for other fossil assemblages in New Zealand.
The second method, angiosperm physiognomy, uses the known relationship between climate and wood anatomical structures in modern taxa to estimate climate from fossil anatomical features. Two sets of equations derived from Wiemann et al. (1998, 1999, 2001) and Martínez-Cabrera HI & Cevallos-Ferriz (2008) respectively, were first tested on modern data and then on fossil wood data from the Miocene. It was found that the equations that best estimated the climatic parameters at the modern test sites did not return reasonable estimations (none were within ~5 °C) from the fossil material.
A principal component analysis (PCA) was also conducted on the anatomical variables of wood from the modern indigenous angiosperm flora of both New Zealand (108 species), and other regions globally, to determine how similar or divergent the wood characteristics of New Zealand are. The PCA showed that there is considerable difference in angiospermous wood anatomical features when comparing the New Zealand flora to that of other regions of the world, particularly in the presence of rays greater than 10 seriate; an incidence more than twice that of any other region. Given this disparity between the wood anatomy of the New Zealand flora and that of other regions, it is suggested that a unique set of models need to be created to determine paleoclimate variables from the wood anatomy of New Zealand angiosperms.
The identification of 54 samples of Mesozoic and Cenozoic fossil wood has provided new data on the diversity of the fossil forest flora of New Zealand and has given insights into the timing of evolution, diversification and extinction of various taxa. Methods for determining paleoclimate from fossil wood need to be revisited, given that analyses of the modern wood flora of New Zealand showed that the wood anatomy of native angiosperms is distinctive.