cc: dave lister <d.lister@uea.ac.uk>
date: Tue, 23 Oct 2007 18:43:38 +0100
from: Ian Harris <i.harris@uea.ac.uk>
subject: CRU TS 3.00 Precip
to: Phil Jones <p.jones@uea.ac.uk>

<x-flowed>
Hi

As a result of trying to produce the 'wet' secondary parameter, I  
discovered that I was using an inappropriate algorithm to convert  
percentage anomalies to real values. No idea where I got it from  
though I can say with certainty that I didn't just make it up!

In the following, N is the Normal, A is the anomaly and V is the  
'real' value.

I was using:

V = N(A+N)/100

However, the anomdtb.f90 program (which creates the original  
anomalies at the start of the process) uses:

A = 1000((V/N)-1)

which translates to:

V = N(A+1000)/1000


This only affects Pre and Wet, I have re-run Pre. An example of the  
difference is that, for January 2001, the maximum was 14408 and is  
now 7174! Obviously trends will not be significantly impacted so I  
don't think the Nat Geog stuff needs amending.

Apologies for any inconvenience (did Dimitrious use these files too?).

Cheers

Harry
Ian "Harry" Harris
Climatic Research Unit
School of Environmental Sciences
University of East Anglia
Norwich NR4 7TJ
United Kingdom


</x-flowed>
