CHEMICAL FORECASTS WITH THE ECHAM
MODEL
Martin Schultz (martin.schultz@dkrz.de, ph: +49 40 41173-308)
Hans-Stefan Bauer (stefan.bauer@dkrz.de, ph: -404)
We will provide daily results from 4 day forecast simulations
with the ECHAM GCM nudged to ECMWF forecast fields in 6-hourly
time resolution. These
will be ready for download in the field (Hongkong and Yakota)
before 8 am local time and then be valid as 3 day forecasts.
The model simulations will be carried out in T106 resolution
(~1.1x1.1 deg) and include Radon-222, several CO tracers tagged
by emission region and type (anthropogenic or biomass burning)
and several climate state variables (especially cloud cover,
precipitation).
The data shall be transferred into the field electronically
via ftp in two formats:
(1) a series of compressed postscript
files with standardized plots zoomed in over the North Pacific
and on at least 4 standard levels (850, 700, 500, 300 hPa)
(2) a set of 3-dimensional fields
in vis5d format which can be readily animated on a laptop computer
which we will bring into the field but could also be converted
for use with the CAVE system.
We plan to transmit plots for at least 6 time steps (0h,
+12h, +24h, +36h, +48h, +60h relative to local noon in the field)
which means about 150MB per day. The 3-dimensional data sets
will also be on the order of 150MB per day if we transfer 6 "codes" (variables)
for the greater Pacific region with 6h time intervals. It will
be easy to adjust the number of fields and amount of data depending
on their usefulness and the available connection. We will be
able to adjust the output almost on a daily notice but we would
like to standardize our products early in the field phase.
We will bring laptop computers to retrieve and process
the data, but we would like access to a color printer and a beamer
for producing plots and animations. Postprocessing of data in
the field should not take longer than 1 hour.
Martin Schultz will be in Hongkong and Hans-Stefan Bauer
will be in Yakota to handle our data and prepare presentations
for the flight planning sessions.
After the TRACE-P experiment, we will rerun the model
using the ECMWF analysis fields and compare the results with
the forecasted fields.
|