| 412 | | Describes r713X. There are two different ways of looking at a root profile in the code. It could reflect "structure" or "function". Root structure is probably how most of us think about roots (i.e. digging a whole and observing where the roots are). When thinking at root structure the profile should be relatively constant over time. (A logic time integrator to determine this constancy is longevity_root as the profile cannot grow faster then the roots grow and die -> in ORCHIDEE we simply fixed it over time). |
| 413 | | |
| 414 | | In ORCHIDEE root structure is used in the calculation of k_lab which is water infiltration along roots (accounted for in hydrol.f90) and the input of soil carbon and nitrogen at depth due to the turnover of roots which is accounted for stomate_soil_carbon_discretization.f90. Furthermore, it is used to calculate the root temperature in stomate_resp.f90. |
| 415 | | |
| 416 | | When thinking about root function it is not so important where the roots are located but it is more important at which depth the roots will be active. The function approach is used in the dynamic root profile where the plants can take most of the soil water from the layers where the soil water happens to be (feature added by D.Zhu in r4363. I'm not sure which versions this is in but we have it in the 4.x). This way of looking at the roots is similar to how we look at the canopy where we have a lot of leaves at places in the canopy where little light can penetrate and where a large part of the photosynthesis is taken care of by the leaves in the top layers of the canopy. |
| 417 | | |
| 418 | | The exponentially decreasing root mass with depth which is the default in ORCHIDEE, is based on root mass observations and, therefore, reflects a structural approach. ORCHIDEE uses the same root profile to calculate plant water uptake, infiltration and som inputs. Hence, root function and root structure are considered to be the same. |
| 419 | | |
| 420 | | The dynamic root profile in ORCHIDEE follows a functional approach based on soil water availability. This seems a fair approach. Plants probably have way more roots than they need exactly to be able to quickly adjust to changes in the soil moisture environment (likewise they have way too many leaves to adjust to the fact that the Earth is rotating and leaves are continuously moving from the wind). If we pass this functional root profile (which is hyper dynamic as it may change every 30 minutes) to the rest of the code, we would introduce scientific inconsistencies in the calculation of infiltration, temperature and som_input. Note: yet another functional approach could be based on nutrient availability as that is another function of the roots. |
| 421 | | |
| 422 | | Ideally the code should thus use a different root profile depending on what we want to use it for. |
| 423 | | |
| 424 | | Root_profiles can and is calculated in the absence of vegetation. Having a root profile even in the absence of the PFT is convenient for when land cover changes happen. The downside is that the history files become unintuitive. To have the best of both worlds we calculate it for every pixel but will do some cleaning before writing the history files which is done in hydrol_main. |
| 425 | | |
| 426 | | The approach that has been used in previous versions of the trunk describes the root profile as a truncated exponentially decreasing function of depth. The function is truncated at a prescribed depth ('''maxdepth'''). The shape of the root profile is prescribed by the pft-dependent parameter '''humcste'''. The nodes and interfaces of the soil layers follow the scheme proposed by De Rosnay (PhD thesis). In r4363 a second option was added and a dynamic root profile can be calculated by setting the flag '''ok_dynroot''' to y. The dynamic root profile optimizes soil moisture usage and was modified from Dan Zhu's initial implementation to account for soil layer depth. The root profile that is now calculated in hydrol.f90 is passed to the other subroutines that make use of it. |
| | 412 | Describes r7149. There are two different ways of looking at a root profile in the code. It could reflect "structure" or "function". Root structure is probably how most of us think about roots (i.e. digging a whole and observing where the roots are). When thinking at root structure the profile should be relatively constant over time. A logic time integrator to determine this constancy would be longevity_root as the profile cannot grow faster then the roots grow and die. In ORCHIDEE the structure profile is fixed it over time because the parameter that describes it ('''humcste''') and the rooting depth are constant over time. Note that when '''ok_soil_carbon_discretization''' is used, the depth of the roots may depend on the thickness of the active layer. Likewise it would not be unlogic to make rooting depth a function of plant biomass as it could be expected that small plants need some time (and biomass) to develop a deep root system. When doing so the literature on this topic should be checked as in semi-arid regions plants probably propagate through vegetative reproduction because that would give them the opportunity to develop deep roots before becoming independent from the parent plant. |
| | 413 | |
| | 414 | In r7149 a new subroutine calles hydrol_root_profile was added to hydrol.f90. This subroutine caldculates the rooting depth based on the PFT-dependent variable '''max_root_depth''' and the thickness of the active layer as calculated in stomate_soil_carbon_discretization. The nodes and interfaces of the soil layers follow the scheme proposed by De Rosnay (PhD thesis). This is calculated in vertical_soil.f90. The subroutine continues with calculating root structure as an exponentially decreasing function with depth which is the default in ORCHIDEE. This approach is based on root mass observations and, therefore, reflects a structural approach. The structural approach is used in the calculation of kfact_root which is water infiltration along roots (accounted for in hydrol.f90) and the input of soil carbon and nitrogen at depth due to the turnover of roots which is accounted for stomate_soil_carbon_discretization.f90. Furthermore, it is used to calculate the root temperature in stomate_resp.f90. |
| | 415 | |
| | 416 | When thinking about root function it is not so important where the roots are located but it is more important at which depth the roots will be active. This seems a fair approach. Plants probably have way more roots than they need exactly to be able to quickly adjust to changes in the soil moisture environment. The function approach is used in the functional root profile where the plants can take most of the soil water from the layers where the soil water happens to be (feature added by D.Zhu in r4363. I'm not sure which versions this is in but we had it in trunk 4.0). The nodes and interfaces of the soil layers follow the scheme proposed by De Rosnay (PhD thesis). This is calculated in vertical_soil.f90. This way of looking at the roots is similar to how we look at the canopy where we have a lot of leaves at places in the canopy where little light can penetrate and where a large part of the photosynthesis is taken care of by the leaves in the top layers of the canopy. This functional approach is used in hydrol.f90 to calculate the water uptake from the soil for transpiration. It is also used in hydraulic architecture to calculate the water supply to the stomata and also the uptake of soil water from each soil layer. |
| | 417 | |
| | 418 | In ok_soil_carbon_discretization many more soil depths are introduced. When testing this functionality with trunk 4.0 some of the options could be removed and the use of different soil depths could be checked for consistency, i.e., are the same nodes and interfaces used as in the rest of the model? Is it really a new depth or just another name of the same depth? |
| 429 | | || ok_dynroot || ok_hydraul_arch || ok_soil_discretization || Comments || |
| 430 | | || - || - || - || Default (r7135). Nodes and interfaces of the discretization have been checked. Extensively tested at the global scale. || |
| 431 | | || - || - || + || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
| 432 | | || - || + || - || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
| 433 | | || - || + || + || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
| 434 | | || + || - || - || To be checked. || |
| 435 | | || + || - || + || Model made to crash. The dynamic root profile may conflict with the way som_input is distributed. || |
| 436 | | || + || + || - || To be checked. || |
| 437 | | || + || + || + || Model made to crash. The dynamic root profile may conflict with the way som_input is distributed.|| |
| | 421 | || ok_hydraul_arch || ok_soil_discretization || Comments || |
| | 422 | || - || - || Default (r7135). Nodes and interfaces of the discretization have been checked. Extensively tested at the global scale. || |
| | 423 | || - || + || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
| | 424 | || + || - || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
| | 425 | || + || + || Nodes and interfaces of the discretization have been checked. Tested for a limited test case. || |
