Skip to main content
. 2013 May 28;371(1991):20120287. doi: 10.1098/rsta.2012.0287

Figure 1.

Figure 1.

Temperature Tq is observed with likelihood q. If the temperature cdf is not changing with time (g is constant) then temperature T* is from the same distribution as temperature Tq, and is observed with likelihood q+dq. This is the path dg=0. If the temperature cdf is changing with time, due to change in g(t), then temperature T* is observed at time t2 with the same likelihood q as temperature Tq was observed at t1. This is the path dq=0. (Online version in colour.)