GICC-Rhone Project
Background Information
Model simulations from four climate models were used to provide six future climate change
scenarios resulting from a presumed doubling of the atmospheric concentration of
CO2. The
four models which were utilized are from the LMD (Laboratoire Météorologique Dynamique),
CNRM (Centre National Recherches Météorologiques), HC (Hadley Centre) and UR (University of
Reading). In addition, relatively high spatial resolution (HR) CNRM and LMD model outputs were
used to make a total of six model datasets. Each model provided a baseline
(1xCO2) present
climate (PC) and a future (2xCO2) modified
climate (MC) which were based on a 10 year average
for the years 2050-2060 after an initial spin-up period. All model scenarios were assumed to have
an equal probability for the current study. The monthly average atmospheric forcing
comprising a set of six variables
was extracted from the model output datasets:
incoming solar radiation,
incoming atmospheric longwave radiation,
air temperature, wind speed,
specific humidity and total precipitation.
The present climate from the Rhone database was then used together with
the GCM anomalies to calculate 6 future climates. These climates
were then used to drive the coupled ISBA-MODCOU model in order
to explore the impact of possible future climate change on
the hydrological cycle of the Rhone basin.
For more information, please see the Rhone-GICC report.
Forcing Database Information
Basic information on the Rhone Atmospheric forcing database is presented
in this section relative to the GICC Project: the generation of Modified
Climate Scenarios for the period August, 1981, though July, 1998.
This page also contains information
about the GICC-Rhone CDROMs which are currently available.
Atmospheric Forcing Variables
|
- Using a time step of 3 hours over 17 years at 1471 grid points
|
PC |
Rg |
Rat |
Pr |
Pn |
Ta |
qa |
Va |
MC Type I |
Rg |
Rat |
Pr* |
Pn* |
Ta* |
qa |
Va |
MC Type II |
Rg* |
Rat* |
Pr* |
Pn* |
Ta* |
qa* |
Va* |
|
* = Modified Climate variable = Present Climate variable + effect of anomalies
VV_YEAR.MODEL.RES.forc
|
Atmospheric Forcing Variable
|
|
year start (August 1) to year end (July 31)
|
|
Global Climate Model (GCM)
|
|
spatial resolution of GCM simulation (High or Low Resolution)
|
MC Files, Types I and II:
- Using a time step of 3 hours over 17 years at 1471 grid points
|
variable |
units |
VV |
YEAR |
MODEL |
RES |
Rg |
W m-2 |
RV |
8182 |
CNRM |
HR |
Rat |
W m-2 |
RI |
8283 |
CNRM |
LR |
Pr |
kg m-2
s-1 |
RR |
. |
LMD |
HR |
Pn |
kg m-2
s-1 |
SS |
. |
LMD |
LR |
Ta |
K |
TA |
. |
HC |
LR |
Va |
m s-1 |
UA |
9697 |
UR |
LR |
qq |
kg kg-1 |
QA |
9798 |
|
|
|
CIG Files, Types I and II:
- Using a time step of 1 day over 17 years at 1471 grid points
|
variable |
units |
VV |
YEAR |
MODEL |
RES |
Pr |
kg m-2
d-1 |
RRJ |
8182 |
CNRM |
HR |
Pn |
kg m-2
d-1 |
SSJ |
8283 |
CNRM |
LR |
ETP |
kg m-2
d-1 |
ETPJ |
. |
LMD |
HR |
Ta(max) |
K |
TMAXJ |
. |
LMD |
LR |
Ta(min) |
K |
TMINJ |
. |
HC |
LR |
|
|
|
9798 |
UR |
LR |
|
The PC, MC-I and MC-II files are available in single-precision
binary format at a temporal resolution of 3 hours for 17 years
at 1471 grid points.
Due to the large data file size, however, the files have also been compressed
using a simple linear algorithm. This format change causes a small
loss in precision, however, the impact on modeled discharge was
found to be negligible. This format change results in a signifigant space
savings. Note, this format change is applied to the PC, MC-I and MC-II databases,
not the CIG files (which are in UNIX compressed ASCII format).
The transform is simply expressed as
where V* is an INTEGER*2 (or short) variable and
the original variable V is REAL*4 (usual default). The
symbols are defined as
- INT = integer
- i = space index (i = 1, 1471)
- j = time steps each year [j = 1, (365 or 366)x8]
- k = year index (k=1, 17)
- Io = 32766 (approximate range of INTEGER*2 variables)
- Vi j k* INTEGER*2 transformed variable
- Vi j k REAL*4 original variable (true value)
- Vk max REAL*4 file maximum
(for all i and j)
- Vk min REAL*4 file minimum
(for all i and j)
Note that the single yearly-domain maximum and minimum
for each variable are also provided in a seperate file (maximum and
minimum for each variable for each file).
The inflation algorithm named inflate.f90 (f77) is provided in both FORTRAN90
and FORTRAN77. The newly formatted files are compressed using gzip.
The file naming convention is the same as above, except
that the extension _RED is added to denote REDuced format:
VV_YEAR.MODEL.RES_RED.forc
The files using the same naming convention but with
the extension _MAX-MIN each contain 2 values on one line,
the maximum and minimum annual-domain value for each variable
for each year:
VV_YEAR.MODEL.RES_MAX-MIN.forc
These values are read in by the provided inflation program along
with the reduced format data to produce (by default) REAL*4 variables
in binary format. The user can change this output format simply by editing the
file inflate.f90 (f77).
Please note that there is also a simple bit-swapping program
which is provided for converting the _RED INTEGER*2 files
to the binary format needed by some machines (if you compile
inflate.f90 and it can not read the _RED files on your
machine) or you don't get the correct transformation (see README.inflate
file in the /progs directory on the CDROMs).
This directory contains selected figures from the above report as color images
(they are listed below). Monthly anomalies for the 6 GCM model simulations
for the Rhone basin and various observed-climate fields are shown.