Not a geologist are you John?
So many errors, such a short article:
Firstly Ephesus lost its access to the sea because of the silting up of the Menderes River. Nothing to do with sea level change. It's actually one of the few places in the Eastern Med and Aegean where sea levels *aren't* a major factor - over much of Greece, recent massive changes in observed sea level are down to tectonic movements. Yet remove the tectonic movements and sea levels still change, so there is a climatic effect
You wrote: 'Yet, with the notable exception of the extinction of the dinosaurs, it seems life has happily trundled along through it all.'
You mean it's trundled happily along - APART from the repeated mass extinctions observed in the geological record; the biggest of which are: end Ordovician, late Devonian, end Permian, end Triassic, and end Cretaceous, some of which were much more serious than the one which killed off T-Rex and friends. There are also the PreCambrian Snowball Earth in the Proterozoic and the Oxygen Catastrophe in the PalaeoProterozoic.
Increased greenhouse gases the best explanation for the Palaeocene - Eocene Maximum; a relatively minor, recent mass extinction which did huge damage to oceanic diversity. It has also been proposed as the driver for the end Triassic extinction which saw something like 20 - 25% of species go to the wall.
Finally, you'll be glad to know, climate models DO include factors such as water vapour and they include climatic information from deep ocean oozes, O16/18 isotope ratios in shells, ice cores, lake sediments, tree rings... to help establish climate records going back hundreds of thousands of years.
It's also worth noting that climate scientists have been desperately trying to tighten up their models. Last year the Open University released a PC client to let everyday users help resolve some issues with the parameters in existing climate models. Users could download a model then their computer would change some of the parameters, run the model and transmit information back to the central server where it could be compared against historical records to see if the model became more accurate or less accurate. More information and download links here:
With so many mistakes, it kind of makes me wonder why you wrote the article?