Shannon Mason edited UM_bias.tex  almost 10 years ago

Commit id: 55a1298c88d47e6e8409b76b910bdc60dd772516

deletions | additions      

       

\subsection{Bias decomposition} Total bias is substantially lower than in ACCESS1.3 ($> 40.$ W/m2), but still very significant.  Summary:  \begin{itemize}  \item Over-production of U3 & U4 compensate for infrequent, too-then U5 (low clouds)  \item U10 is strongly over-produced: seems to be a recently-introduced compensating bias.  \item U9 is too optically thin and too infrequent  \item Mid-topped cloud biases are substantially improved: U6 occurs about the right amount of time, but isn't optically thick enough; U7 isn't frequent or thick enough.  \item U1 now contributes to the SW bias (in GA1.0/ACCESS1.3 H1 was a key compensating cloud regime).  \end{itemize}