Context Navigation

soil_physic

Timestamp:: 2015-02-27T14:07:44+01:00 (9 years ago)
Author:: peylin
Comment:: --

Legend:

: Unmodified
: Added
: Removed
: Modified

DevelopmentActivities/CMIP6/DevelopmentsCMIP6/soil_physic

-                      v28
+                      v29
 Scheme: [https://forge.ipsl.jussieu.fr/orchidee/attachment/wiki/Meetings/CMIP6/Physic/soil_discretization.pptx][[BR]]
 The main principle are:
  -- The T scheme is always as depth or deeper than the W scheme
  -- W and T share the same common discretization for the common part: the layers are the same; For the nodes:  W has the firt node at the surface while T has it at the middle of the first layer; Else they have the same node position
  -- The upper part of the scheme follow a geometric discretization of reason 2 (like the current W scheme)
  -- Below a certain depth, depth_lin, we have for the W and T a linear discretization, with the depth of each layers equal to that of the lowest layer of the "geometrical part" of the scheme. Note that depth_lin may be equal to depth max with thus no linear discretization section.
  -- Below another depth, depth_geom, we have again the geometrical discretization with a reason xx to be defined. This allows to model a deep soil for permafrost modelling. Note that Depth_geom may also be equal to depth_max with thus no bottom geometrical discretization.
+The main principles are:
+ - The T scheme is always as depth or deeper than the W scheme
+ - W and T share the same common discretization for the common part: the layers are the same; For the nodes:  W has the firt node at the surface while T has it at the middle of the first layer; Else they have the same node position
+ - The upper part of the scheme follow a geometric discretization of reason 2 (like the current W scheme)
+ - Below a certain depth, depth_lin, we have for the W and T a linear discretization, with the depth of each layers equal to that of the lowest layer of the "geometrical part" of the scheme. Note that depth_lin may be equal to depth max with thus no linear discretization section.
+ - Below another depth, depth_geom, we have again the geometrical discretization with a reason xx to be defined. This allows to model a deep soil for permafrost modelling. Note that Depth_geom may also be equal to depth_max with thus no bottom geometrical discretization.
 The list of parameters to define the scheme would be:
 …
  - depth_Wmax : max depth of the W discretization, between [depth_lin, depth_geom]
 Diagnostic variable for the discretization would be:
+Diagnostic variable of the discretization would be:
  - the number of layer for W and T scheme : Name to be defined ?
  - the depth of the nodes for W and T
 …
 Modulation of the Thermal properties as a function of the Water content:
  -- Heat capacity of each layer (C) should be a function of the layer water content: thus a function of the layer total water content ("W" variable of CWRR divided by the layer depth)
  -- Conductivity between layers (Lambda) should be a function of the "Theta" of layer i and i+1 (weighted arythmetic mean using the depth of layers)
  -- For the part of the thermal scheme that is below the hydrology scheme: we should take a constant in time value of "Theta" and "water content", to avoid changing heat capacity over time and thus not conserving the energy (without any corresponding imput/output of energy by water flow); the values to choose is open for discussion/tests ?
+ - Heat capacity of each layer (C) should be a function of the layer water content: thus a function of the layer total water content ("W" variable of CWRR divided by the layer depth)
+ - Conductivity between layers (Lambda) should be a function of the "Theta" of layer i and i+1 (weighted arythmetic mean using the depth of layers)
+ - For the part of the thermal scheme that is below the hydrology scheme: we should take a constant in time value of "Theta" and "water content", to avoid changing heat capacity over time and thus not conserving the energy (without any corresponding imput/output of energy by water flow); the values to choose is open for discussion/tests ?