DISCUSSION

The Regatta Point beds were deposited during or close to the warmest known interval of the Tertiary—the Early Eocene Climatic Optimum (Zachos et al. 2001). The distinct Southern Ocean water masses and fronts of the present day had not evolved, and 'cool subtropical' water flowed to high latitudes (Nelson and Cooke 2001). There is heightened interest in this period as a potential analogue of the enhanced greenhouse warming that the world is currently experiencing (e.g., Wing et al. 2003). In Australia, there have been several publications covering the Early Eocene climate (Greenwood and Christophel 2004), although primary published data are limited.

Pioneering oxygen isotope studies of the Southern Ocean (Shackleton and Kennett 1975) suggested sea surface temperatures for the southern Australian Early Eocene of around 15-17ºC. There have been a much broader range of estimates since, from climate modelling and palaeobotanical evidence. For instance, Nix (1982) argued that southern Australia may have been borderline mesothermal (c. 14ºC) around the Early Eocene.

Early Eocene plant macrofossil sites on the southeastern Australian mainland lie about 4-5º of latitude further north than Regatta Point today–as they would have in the Eocene. Based on low Early Eocene thermal gradients (e.g., Greenwood and Wing 1995) sea surface temperatures in this region would have been no more than 2ºC warmer than Tasmania. Early Eocene mainland sites include Hotham Heights, Brandy Creek, and Deans Marsh. Greenwood et al. (2003) published MAT estimates for Hotham Heights of 17.9ºC, based on the proportion of entire-margined leaves (the Australian relationship later published as Greenwood et al. 2004) in an unpublished taxonomy, and 17.8ºC, based on leaf length and the current relationship between leaf length and MAT in Australian forests. The stated average length of leaves for this deposit of 78 mm is approximately the boundary between the microphyll and notophyll classes of leaf size (Webb 1959). Greenwood et al. (2003) extrapolated their results from Hotham Heights and other localities to infer that MAT for the lowland of southeastern Australia around the Early Eocene was in the 20-25ºC range. They noted that this conclusion was "consistent with the observation by Macphail et al. (1994) that the early Eocene was the acme of development of lowland megathermal species-rich rainforest in southeastern Australia," and further that this was "much higher" than recent computer climate model suggestions of MAT of <10ºC (Sewall et al. 2000; Shellito et al. 2003).

Greenwood and Christophel (2004, figure 18.1) acknowledged that coastal communities around Tasmania included the "mesothermal-megathermal to megathermal mangrove palm Nypa" but mapped the dry-land vegetation as "Microphyll Fern Forest" (Webb et al. 1984 give the present-day average MAT of this forest type as about 12-13ºC) and "Broad–leaved deciduous forest and temperate mixed conifer" (BDF). The basis of this decision is unclear. Greenwood and Christophel (2004) mapped the south-eastern Australian mainland as having notophyll-mesophyll vine forest (which they equated with "mesothermal-megathermal rainforest").

Carpenter et al. (2004) published an overview of Hotham Heights, which they regarded as preserving vegetation growing in an upland region, approximately 800 m above sea level. In terms of overall diversity, the prominence of Lauraceae and Proteaceae, and the absence of Myrtaceae macrofossils, the site is similar to Regatta Point. A similar diversity of conifers was reported for Hotham Heights (although without Cupressaceae) but they were apparently generally uncommon elements. The abundance of the cycad Bowenia in Tasmania contrasts with the absence of cycads at Hotham Heights. They concluded that "all lines of evidence are consistent for the prevalence of a wet, mesotherm environment at Mt Hotham in the Early Eocene" and that the lowlands had an "abundance of taxa that indicate the presence of vegetation with a megatherm character."

The evidence cited above indicates a surprising level of uncertainty about Early Eocene temperatures in southern Australia: ranging from a cool 10ºC (or less) to about 17ºC from modelling and oxygen isotope results, and up to as much as 25ºC based on palaeobotany. Greenwood et al.'s (2004) Australian-based leaf margin calibration is a welcome addition to our knowledge, and they stated that previous estimates of MAT for Australian floras based on the east Asian correlation should be revised, for the Early-Middle Eocene essentially downwards by around six degrees (Interestingly, at around the same time, Kowalski and Dilcher, 2003, suggested that "current methods of inferring paleotemperatures from fossil floras yield underestimates of 2.5–10ºC). However, Greenwood et al. (2004) also stated that the "most conservative approach" would be to use the Australian relationship as a minimum, and non-Australian as a maximum. Thus, leaf margin analysis using their MAT figures and their stated standard error of about 2ºC, suggests Hotham Heights MAT ranged between around 15.9 and 25.8ºC, and lowland temperatures 4–5ºC warmer still.

The actual evidence for megathermal temperatures ever being experienced in the southeastern Australian lowlands is slim. Despite the "megathermal" conclusions attributed to Macphail et al. (1994) above, in reality he wrote of "megathermal in character," to which a potentially broader range of MATs might be attributed. Various workers (e.g., Basinger et al. 1994; Plaziat et al. 2001) have emphasised that in an essentially ice-free world, the winter temperature minima which probably control the polewards limits of many apparently "tropical " taxa today, would have been relaxed. Thus this aspect of climate, rather than MAT, is more relevant. It is hard to reconcile Greenwood and Christophel's (2004) conclusion of (presumably sea-level) microphyll forest existing on Tasmania, while notophyll-mesophyll forest was on the south-eastern mainland. This would imply a higher thermal gradient than today.

The precision of leaf-length based estimates of MAT in Greenwood et al. (2003) also needs to be considered carefully. Leaf length is also a function of precipitation (e.g., Wilf et al. 1998) and the relationship which holds today in Australia may not have held in the past, or indeed, elsewhere today. For example, Schneider et al. (2003) documented rainforest vegetation from 2950 m in Venezuela which is, based on species or individuals, equally dominated by microphyll and notophyll sized leaves (similar to Hotham Heights). They cited the MAT in their area to be 14.9ºC at 2300 m. Based on a lapse rate of 0.55–0.60ºC per 100 m (Meyer 1992), the MAT at 2950 m would be about 11ºC.