Re: [dinosaur] Would non-avian dinosaur survive through the whole Cenozoic?

["David Marjanovic" <david.marjanovic@gmx.at>]

Gesendet:ÂDonnerstag, 05. Dezember 2019 um 20:46 Uhr
Von:Â"Mailing" <mailinglistinformation@gmail.com>

> Outcrop area is not necessarily a good proxy for predicting (terrestrial) 
> diversity, see e.g. Dunhill (2012), Dunhill et al. (2014), Walker et al. 
> (2017), and getting outcrop area/exposure area on a global scale is a 
> non-trivial task (even less so the density of fossils per formation).

Well, the worst-case scenario is not that we have to use a worse proxy. The 
worst-case scenario is that we simply cannot predict diversity unless we 
perform a non-trivial task.

> I am not so sure what you mean with "actually known fossils". The results are 
> obviously based on what is known at the moment - but that is trivial and also 
> holds true for the opposite interpretation of the fossil record. So, I think 
> I misunderstand you here: What do you mean by this?

As I said: if there was a radiation going on in Africa, we're not likely to 
find out anytime soon.

I'm not trying to argue that I can show there was no decline in origination 
rates; I'm trying to argue nobody has shown a reason to think there was a 
decline.

> The shape of the curves is determined by the mathematical properties of the 
> involved models. For a model with constant speciation/cladogenetic rates (I 
> will stick to speciation since that is the commonly used term) and extinction 
> rates (with spec. rates > ext. rates) you expect a linear increase in the 
> number of speciation events in log space through time. If speciation rate 
> decreases through time and is ultimately surpassed by extinction rate, you 
> expect a quadratic relationship (also holds true for the opposite case, where 
> speciation rate increases through time + extinction rate remains constant).
> You then fit your models to the data and assess which model performs best. 
> That will tell you whether non-avian dinosaurs as a whole were in decline or 
> not (in terms of speciational capability). You can add various covariates to 
> account for other effects involved (e.g., sampling bias, extrinsic controls 
> on speciation dynamics, etc.), but it is not necessary for the model per se.

Yes, yes. My problem with this is the long-term view all authors so far appear 
to have taken. Shouldn't we expect extinction rates, and probably also 
origination rates, to vary much faster and much more erratically than the 
smooth curves in the paper? Environments aren't stable, or show stable trends, 
over tens of millions of years, so why should origination or extinction rates 
do that?