Invited review articleWhen did a Mediterranean-type climate originate in southwestern Australia?

- Projections on the time of origin of a Mediterranean-type climate (MTC) in southwestern Australia (SWA) have varied from 30 to 3 million years ago (Ma)
- There is support for the proposition that a MTC originated in northwestern Australia and migrated south 30 Ma as Australia drifted north to its current position
- The evolution of a number of morphological-physiological traits of indigenous plant lineages can be linked to the advent of a hot, dry summer, mild wet winter and intensified fire regime, though the fixity of the relationships is often unclear
- These point to the advent of a MTC, at least on a limited scale, in the 30-10-Ma period, particularly 20-10 Ma but possibly even earlier than 30 Ma
- Stratigraphic and chemical approaches from non-biological disciplines hold promise of further refining these findings while all such approaches can be applied to other mediterranean regions

Knowing the environments under which biota have evolved is essential for understanding the functional traits that they possess. Here, we ask when a Mediterranean-type climate (MTC) originated in Western Australia that might help to explain some of its special plant adaptations to summer drought and heat, mild wet winters and intense summer fires. Periodic drought and fire can be traced back to the Cretaceous in southwestern Australia (SWA) but its seasonality is unknown. Previous estimates of the origin of the MTC in SWA have varied from 30 to 3 million years ago (Ma). An early proposal was that it originated in northwestern Australia 30 Ma and migrated south to its present location for which we find independent support. We collated 47 cases of what is currently known about the time of origin of adaptive responses to summer drought and heat, wet winters and periodic intense fire among flowering plants in SWA. About 15% arose in the 40-30-million-year (My) period, 32% in the 30-20-My period and 47% in the 20-10-My period, suggesting that a MTC may have been widespread in SWA by the mid-Miocene, but there may have been local appearances (proto-MTC) up to 20 My earlier. Uncertainties remain about the location of lineages at various times, the fitness benefits of apparent adaptive traits and their confinement to a MTC, and the significance of the origin versus the onset of diversification of traits. The current dearth of C4 grasses in SWA is consistent with other indications of the absence of a prior history of a summer-dominant rainfall, suggesting that the current MTC arose from a uniform-rainfall climate. Other lines of evidence that might refine these findings include the chemical composition of, and application of geochronometers to, laterites, carbonates and corals that have been used with success elsewhere. We conclude by noting that there may have been a mixture of climates for much of the time, indicated by the persistence of a few rainforest species up to 3 Ma, while there have been strong climatic oscillations about a MTC 'mean' over the last 5 My.