| | 146 | |
| | 147 | = MINUTES |
| | 148 | |
| | 149 | === 13.10.2016 === |
| | 150 | present: Bertrand, Alex, Anne Sofie, Palmira, Nicolas, Nicolas, Josefine, Philippe P. & Sebastiaan (minutes) |
| | 151 | |
| | 152 | - Both Nicolas and Alex observed problems with the growth of deciduous trees when calculating growth respiration as a function of the labile biomass pool. From the equations, the cause of this problem is not obvious but the calculation uses lab_fac which was earlier found to function the way it was implemented (-> there is a min and max value and a smooth transition in between. With the current parameters, the smooth transition, however, does not cover the full range between min and max. As an alternative growth respiration can be calculated as a function of gpp. This seems to work but it introduces small inconsistencies in the code which is a scientific issue rather than a technical bug. Given that the code works with the relationship between growth respiration and gpp it was decided to document this issue on the wiki reporting the merge between CN and CAN and continue with the merge. Given this problem was not apparent in CAN, there is a small chance that it disappears after introducing the CAN canopy-layering. If the problem persist it will be dealt with later. |
| | 153 | |
| | 154 | - It was agreed that Alex, Anne Sofie and Seb (starting 10/11) will continue debugging until the CN-CAN version seems to work from a technical point of view (no new merging). At that point the code will be transferred to the current CN version (ORCHIDEE-CN-CAN will be merged into the branch ORCHIDEE-CN). Further developments and the remaining merges will be done on that version. The goals is to have a set of branches that is really close to each other. |
| | 155 | |
| | 156 | - Sechiba-CAN is running behind the trunk in terms of the way the variables are initialized. It was agreed that we will first try to get the code running. Josefine will then clean the initialization. In case of extreme debugging frustration, the order of these tasks might be swapped. |
| | 157 | |
| | 158 | - As it looks now, it is possible to have a working CN-CAN version by the end of this year. Therefore, it is time to re-animate the DESIRE project. Bertrand will contact Ben. Forcing files will then be prepared so that as soon as the code is ready we can start working towards the objectives of the project. |
| | 159 | |
| | 160 | - Preliminary agenda for 28, 29 and 30 November: (1) Discuss the transition from CN-CAN to TRUNK+CN+CAN. Identify bottlenecks, conflicts and issues that need to be fixed first. Work on this transition. (2) PP, JG, YM, JP and SL discuss the reduction of the dimensionality of variables that now have a npts (kjpindex) and nvm dimension. This seems an essential development to be able to run state-of-art vegetation structure resulting from LULCC. (3) Discuss the multi-column version. |
| | 161 | |
| | 162 | [BR] |
| | 163 | [BR] |
