Definitions for ORCHIDEE

Fractional land covers :


The specification of the fractions of land-cover which interact with the atmosphere have been split into two parts in SECHIBA :
This distinction is needed as the equations for each of them are very different and the second type would probably need a separate surface model for each type. If all these surface types would have been described within the same structure the amount of useless data which would have been carried around the different components of the surface scheme would have been huge. IF statement would have been needed to switch from one calculation to the other or loops with empty operations would have been needed. The amount of data stored with no information at all would have been large (transpiration and interception over glaciers for instance !).

veget(nbindex,nvm) :

This variable gives for all points the fraction of each vegetation type and bare soil. The first type is bare soil and then all the vegetation types. The values of nvm will change as the model evolves and so will the vegetation type at each value of the index. As a consequence SUM(veget(nbindex,:)) is always smaller or equal to 1.

fraction_ice(nbindex) :

The fraction of the mesh covered by glaciers. As for the moment this is the only non hydrologically or biologically active surface we have the following condition :
fraction_ice(nbindex) + SUM(veget(nbindex,:)) = 1
These are the variables describing the tiles of the surface which interact with the atmosphere in SECHIBA. This is insufficient as one may have vegetation which is dormant or is covered by another types and thus plays a role in the vegetation dynamic or the carbon cycle. For instance when a tree looses its leaves it is the bare soil beneath which starts to interact with the atmosphere but still STOMATE and LPJ need to know about this vegetation. Thus we need another variable : veget_max.

veget_max(nbindex, nvm) :

This variables gives the fraction of the mesh for each vegetation type weather it is dormant or not. This variables fulfills the same conditions as veget except that its sum may be larger than one when some vegetation types overlap each other (Trees above grass for instance). Nevertheless be careful because in the case where the vegetation types do not overlap we still need to fulfill the following condition :
fraction_ice(nbindex) + SUM(veget_max(nbindex,:)) = 1
veget can thus be derived from veget_max considering the status of the various vegetation types. This will be done once a day in the module slowproc. It is desired that veget and veget_max cover the same vegetation types but there is no strict need for it. One may consider that on the STOMATE side the plant functional types be chosen as a function of their phenology and dynamic while in SECHIBA it is done as function of their characteristics for the interaction with the atmosphere. This level of generality may lead to some complexity but it is too early to discard this option.
Jan Polcher
Last modified: Fri Sep 15 12:52:43 CEST 2000