156 | | [[Image(error_LITTER_MET_AB_inferior_to_0.2%.gif)]] [[Image(error_LITTER_MET_BE_inferior_to_0.5%.gif)]] |
157 | | [[Image(error_CARBON_ACTIVE_inferior_to_0.5%.gif)]] [[Image(error_CARBON_SLOW_inferior_to_0.5%.gif)]] |
158 | | [[Image(error_CARBON_PASSIVE_inferior_to_2%.gif)]] |
159 | | |
| 157 | [[Image(error_LITTER_MET_AB_inferior_to_0.2%.gif)]] [[Image(error_LITTER_MET_BE_inferior_to_0.5%.gif)]] |
| 158 | [[Image(error_CARBON_ACTIVE_inferior_to_0.5%.gif)]] [[[Image(error_CARBON_SLOW_inferior_to_0.5%.gif)]] |
| 159 | [[Image(error_CARBON_PASSIVE_inferior_to_2%.gif)]] [[Image(error_stock.gif)]] |
| 160 | |
| 161 | The vertical axis correspond to the error in percentage. Notice that the carbon stock could be considered |
| 162 | at equilibrium before the passive pool is at equilibrium (we got more than 1000 years difference!). [[BR]] |
| 163 | |
| 164 | For this reason, we have chosen in a first time step to implement a stopping condition based on the maximum |
| 165 | of the relative errors for each pool. It is still not satisfactory. The next step will be to implement a stopping |
| 166 | condition based on the carbon flux. |
| 216 | |
| 217 | Forcesoil is very performing. Let's take the example of the passive pool.[[BR]] |
| 218 | |
| 219 | ORCHIDEE+forcesoil runs only 500Y to find a error of 0.05%.[[BR]] |
| 220 | |
| 221 | The algebraic method have to run 1500Y to have a similar error (and 500Y more to find the exact result) |
| 222 | and ORCHIDEE alone got the same value as forcesoil after 6200Y. [[BR]] |
| 223 | |
| 224 | Our analytic method is 3 to 4 more faster than ORCHIDEE alone but less performing than forcesoil. [[BR]] |
| 225 | |
| 226 | The analytic method found the same results after 500Y than ORCHIDEE+forcesoil in 51Y.[[BR]] |
| 227 | (see /dmnfs12/cont003/p529sol/IGCM_OUT/OL2/50Y_ORC_FORCESOIL_2/) |
| 228 | |
| 229 | |
| 230 | === Another test === |
| 231 | |
| 232 | A another test was made using the same point and the pft 6. The results obtained are globally the same : they |
| 233 | are not detailed but could be found here : /dmnfs12/cont003/p529sol/IGCM_OUT/OL2/spinup_7000Y_pft6.tar |
| 234 | |
| 235 | |
| 236 | |
| 237 | == Summary == |
| 238 | |
| 239 | === Work done === |
| 240 | |
| 241 | - Implementation of the method of PASIM in ORCHIDEE : adaptation to 7 pools + flux to the fire[[BR]] |
| 242 | |
| 243 | - Implementation of a stopping condition based on the maximum of all the relative errors calculated for each pool[[BR]] |
| 244 | |
| 245 | - Modification of the script and job of OOL_SEC_STO : if the stopping condition chosen by the user is reached before the end of the run, the job will stop[[BR]] |
| 246 | |
| 247 | - Tests made on the temperate regions for pfts 10 and 6 |
| 248 | |
| 249 | === To be done === |
| 250 | |
| 251 | - New tests on a boreal region, tropical region and global[[BR]] |
| 252 | |
| 253 | - Implementation of a new stopping condition based on the evaluation of the flux and not on the stock |