cc: d.viner@uea.ac.uk, aidan.burton@ncl.ac.uk, buishand@knmi.nl
date: Mon, 18 Sep 2000 11:47:50 +0100
from: "Chris Kilsby" <c.g.kilsby@ncl.ac.uk>
subject: relative humidity bias
to: jmmurphy@meto.gov.uk, phillip.reid@uea.ac.uk, p.jones@uea.ac.uk

Dear James and Wrincle colleagues,

we are rather concerned about the RH bias in the HadCM2/3 
output. My concern is that it may be caused by a feature of the 
parameterizations, and we are not performing a fair comparison 
wiuth the NCEP data. Adri's comments below lend weight to this. 
Any comments on how we should proceed ?

Chris
------- Forwarded Message Follows -------
From:           	"Adri Buishand <buishand@knmi.nl>" <Adri.Buishand@knmi.nl>
To:             	"Chris Kilsby" <c.g.kilsby@ncl.ac.uk>
Date sent:      	Mon, 18 Sep 2000 12:21:25 +0000
Subject:        	relative humidity bias
Priority:       	normal

Dear Chris,

I had a short discussion with Erik van Meijgaard about the relative
humidity bias in the HADCM simulations. The HADCM2 simulation 
shows a cold
bias at 700 hPa.  A temperature bias of 2 degrees has already an 
effect of
15% on the saturated vapour pressure and may thus substantially 
contribute
to the bias in relative humidity. 

There is, however, little temperature bias in the winter season. Erik 
told
me that GCMs work with the relative humidity over ice if the 
temperature
is below zero. Have these relative humidities been converted to 
relative
humidities over water? The difference between the saturated vapour
pressure over water and ice is about 10% if the temperature is - 10
degrees (average winter temperature at 700 hPa for the grid-point
considered).

Wrong parametrizations were also suggested as causes for the 
relative
humidity bias. In the old low-resolution UKMO GCM, the zonally 
averaged
700 hPa relative humidity at 50 N is about 70% (Fig. 1b in Mitchell 
and
Ingram, J. Climate, 5(1992), 5-21), which is comparable with that in 
the
HADCM2 simulation.

We do not have the relative humidity in the ECHAM/OPYC3 
simulations.
Over the Netherlands and Northern Germany there is a bias in the
specific humidity of about 10%, whereas the bias in 700hPa 
temperature is
less than 1 degree. We have the impression that it looks 
somewhat better
than the HADCM simulations.

Best regards,

Adri



Chris Kilsby   Lecturer in Hydrology
Water Resource Systems Research Laboratory
Dept of Civil Engineering, University of Newcastle upon Tyne
NE1 7RU    United Kingdom 
Tel + 191-222-5614  Fax + 191-222-6669
http://www.staff.ncl.ac.uk/c.g.kilsby
