Changes between Version 82 and Version 83 of Documentation/Forcings

Timestamp:

2018-02-28T11:59:16+01:00 (6 years ago)

Author:

jpolcher

Comment:

--

Documentation/Forcings

@@ -51 +51 @@
 
 (b) '''Two versions of the driver are available''' to do do both the reading and the inter/extrapolation of the forcing datasets. They differ in the way they interpret the time at which the diffrent variables are read in the forcing dataset (further called the timestamp, while the interval refers to the period between two succesive timestamps):
-* '''the old driver (src_driver/dim2driver)''', which assumes that scalar variables (Tair,Qair,Psurf,Wind) are instantaneous values at the timestamp, while flux variables (Rainf,Snowf, SWdown, LWdown) are mean values over the interval ending at the timestamp. This driver also requires that the first value of the annual files corresponds to 00UTC+dt_forcing (for instance 03UTC for a 3hourly forcing), and that the last value is for Dec 31st at 24UTC = 00UTC on Jan 1st of the following year. It also requires that all variables are averaged over the forcing time step and placed on the central time of the averaging period.    
+* '''the old driver (src_driver/dim2driver)''', which assumes that scalar variables (Tair,Qair,Psurf,Wind) are instantaneous values at the timestamp, while flux variables (Rainf,Snowf, SWdown, LWdown) are mean values over the interval ending at the timestamp. This driver also requires that the first value of the annual files corresponds to 00UTC+dt_forcing (for instance 03UTC for a 3hourly forcing), and that the last value is for Dec 31st at 24UTC = 00UTC on Jan 1st of the following year.
 * '''the new driver (src_driver/orchideedriver.f90) is more flexible''' and can normally deal with any temporal structure of the forcing datasets, provided the required attributes are inserted in the forcing files. Yet, orchideedriver is not yet compatible with libIGCM. This document [attachment:Description_Forcing_Files.pdf] provides details of how forcing files need to be prepared. It addresses in particular the issues one may face when the fluxes are averaged and scalar variables instantaneous. The driver needs to take this into account else the diurnal cycle can be misplaced. Some data providers have different convention for representing averaged fluxes and this type of details matter. 
 
 A longer discussion on the differences between the two drivers (dated in 2016) can be found on https://forge.ipsl.jussieu.fr/orchidee/wiki/Branches/Driver_Improvements.
 
-(c) '''Adding the cell_method attribute to files'''.
-It is a good practice to add the cell_method attribute to each of the forcing variable in order to document how they relate to the time axis. This attribute will say if the variable represents a point in time (instantaneous) or averages an interval (). Furthermore it can inform how the time interval over which the variable is valid relates to the discrete points in time the axis provides.
-The general form of the attribute follows the CF convention and is "cell_method = axname : method(position)" where the folowing values can be inserted :
- * '''axname : '''
- * '''method :'''
- * '''position :'''
+(c) '''Adding the cell_methods attribute to files'''.
+It is a good practice to add the cell_methods attribute to each of the forcing variable in order to document how they relate to the time axis and their representativity in time. This attribute will say if the variable represents a point in time (instantaneous) or averages  over an interval (mean). Furthermore it can inform how the time interval over which the variable is valid relates to the discrete points in time the axis provides.
+The general form of the attribute follows the CF convention and is "cell_methods = axname : method(position)" where the folowing values can be inserted :
+ * '''axname : ''' The name of the time axis within the file that is valid for the current variable
+ * '''method :''' instantaneous, mean
+ * '''position :''' start, centre, end
+Obviously for an instantaneous variable, the position information is not relevant as it is only the point in time provided by the axis which matters. On the other hand, for an average, it is essential to know if the point in time provided by the axis is at the beginning, middle or end of the interval used for computing the mean.
+More details and examples are provided in this document : https://forge.ipsl.jussieu.fr/orchidee/wiki/Branches/Driver_Improvements.
 
 (d) '''Usually, the forcing files are stored in the shared accounts''' in IGCM/SRF/METEO directory. The shared accounts are found: