This page is dedicated to the development and validation activities of the model as it evolves over time.

This is the version of the WRF-ORCHIDEE which ran from October 2016 to XXX 2017 under the supervision of Karine.

• Validation of the clouds and comparisons with previous version (WRF3.6.1 RUC, WRF3.6.1 DIFF) Document.

Simulations performed during the month of February 2017 under the supervision of Marc.

• Validation of the new T2/Q2 parameterization with previous version (WRF3.7.1 - ORCHIDEE ) Document.

Analysis of the continental water cycle as simulated by the latest version of ORCHIDEE forced by WRF and other data sets.

• A short presentation with the main results Document.

This section presents the simulation carried out with the WRF/ORCHIDEE configuration at tag MEDCORDEX-A over the ERA-I period (1979-2016).

The Monitoring board as produced to monitor the simulation.

The CliMAF atlas has been produced for different variables which can be compared to the E-OBS database (Version 16). Due to technical problems at IDRIS the maps can only be accessed until 2009. As soon as this will be resolved the links will be updated here.

1. Maximum Air Temperature (2m): Monthly Maps, Seasonal Maps , Annual Maps
2. Minimum Air Temperature (2m): Monthly Maps, Seasonal Maps , Annual Maps
3. Air Temperature (2m): Monthly Maps, Seasonal Maps , Annual Maps
4. Rainfall: Monthly Maps, Seasonal Maps , Annual Maps

For each of the FluxNet station within the domain the fluxes simulated by WRF and ORCHIDEE can be compared to the in-situ observations :fluxnetplots.pdf. One of the main information which can be drawn from this analysis is how the in-coming radiation is redistributed and rainfall used to feed evaporation. Some preliminary conclusions can be drawn :

• 74 stations can be positioned on the MEDCORDEX domain but the observations are of very contrasting quality.
• Generally the model underestimates rainfall over these stations.
• On the other hand net-radiation is generally overestimated.
• As a direct consequence both turbulent fluxes (latent and sensible heat fluxes) are overestimated.
• The graph of mean rainfall against evaporation illustrates the evaporation/rainfall efficiency. When moving from the observations (blue circles) the model (red circles) one notes that the model tends to evaporate more of the water received from the atmosphere.
• The graphic which relates net radiation to the sum of the turbulent fluxes illustrates the Bowen ration. It also shows that ORCHIDEE sends back more energy to the atmosphere than estimated in the observations.
• An underlying problem visible in these diagnostics is that the FluxNet observations do not close the energy balance. The last figure of page 2 shows that this issue is as large as the shift in the Bowen ratio. This has to be kept in mind.
• The time evolutions for the 72 stations shows the diversity of behavior and quality of the data.

One of the conclusion of this first analysis is that a more thoroughly verified version of the FluxNet data should be used.

These graphics are still rather preliminary as they will include an off-line simulation of ORCHIDEE with observational based forcing. Still some interesting results can be highlighted.

• The Loire : The annual cycle and total water budget is rather well reproduced.
• The Po and Rhone : WRF/ORCHIDEE does not provide enough water in the streams. Either there is too little precipitation on the Alps or too much evaporation.
• Further to the East the Danube seems to have a quite reasonable annual mean water cycle. But, ORCHIDEE does not deal properly with the melt water peak. Either it is an issue with the soil freezing or the water regulation which exists ?
• In the other rivers flowing to the Black Sea (Don, Dnepr, Dniestr) the model has too much water (too much rainfall or too little evaporation ?)

• RegIPSL : towards an IPSL Regional modelling infrastructure
• Validating the water cycle of the MEDCORDEX reference simulation.

Discussion of the various points which were corrected in the ORCHIDEE/WRF coupling.

Diagnostics were performed with CliMAF over the 1979-1989 which is common to MEDCORDEX-A and MEDCORDEX-B.

• Taux : : Eastward surface wind stress
• Tauy : Northward surface wind stress
• E-P : Evaporation minus precipitation
• Qnet : Net heat flux into the ocean
• Qsol : Solar net flux into the ocean

Namelist file used for WRF : namelist_wrf.pdf

Namelist file used for WPS : namelist_wps.pdf

configuration file : config_card.pdf

Presentation of the first analysis of the two simulation done during the retreat (2-4 September 2019) : analysis_sep19.pdf

wrforchnemo_analyse.pdf

The figures in this file compfog_out.pdf compare for the 10 regions of the Mediterranean seas the fresh water inflow from the continents.

WRFORCH and WRFORCHNEMO are compared to the FOG estimate (Wang & Polcher 2019) as well as two off-line simulations of ORCHIDEE.

In none of the cases do we have as much water flowing into the sea as proposed by FOG. The file also includes the riverine input in m/y (i.e. distributed over the area of the corresponding basin) in order to compare to table 11 of Jorda et al. (2017). It is clear that FOG with 0.21 is slightly above that estimate and the model with 0.08 is way too low.

The figure above provides the bias of WRFORCHNEMO with FOG in % over each of the basins of the MED. The dots provide for each GRDC station the bias, in % as well, of the simulated discharge. The message from this figure is that the errors of the model compared to these two estimates of the water flowing off the continents is quite similar. It reinforces the interpretation that the bias of the model when compared to FOG is real.

The figures in file basinfluxes_out.pdf compare the various surface fluxes, as well as SST and near surface wind, between WRFORCH and WRFORCHNEMO for the different basins of the MED. Observation derived fluxes have also been added.

Generally the difference between forced SST and coupling to NEMO is smaller than the inter-annual variability of climate.

The energy and water mass exchanges through the interface with the atmosphere were plotted using the units of tables 10 and 11 of Jorda et al. 2017. The graphics balance_out.pdf show that for the water exchange the model is close to the consensus. While for the energy exchanges there is a large impact of the coupling.

The table from Jorda et al. which serve as a reference for this analysis :

Euro-SW simulation results : plot

Altitudinal variation of meteorological variables : updated plot

Altitudinal variation of meteorological variables : plot

Comparison of meteorological variables over river basins in the 20KM and EuroSW simulations. : plot

The spatial pattern of monthly mean precipitation obtained from EURO-SW simulation and interpolated corresponding MEDCORDEX-B simulation (on EURO-SW grid) during 1996 and 1997 is shown in the plot. The month number is indicated in the bottom right corner of each panel.

2 test case simulations for a year 1995.

The slides are provided here : plots

3 test case simulations for a year 1995.

The slides are provided here : plots

Analysis of the first 11 months of 1999 (Euro-SW domain simulation).

The slides are provided here : Updated plots

Analysis of the first month of January simulated over the Euro-SW domain at a resolution of 3km.

The slides are provided here : Slides by Namendra

The updated slides are provided here : Slides by Namendra

Compared the Euro-SW metgrid datasets with the MEDCODEX-B datasets as well as with the ERA-Interim metgrid datasets.

The slides are provided here : Slides by Namendra

Comparisons of MEDCORDEX-B output with E-OBS observation

The slides are provided here : Slides by Namendra

Prediction skill of extreme precipitation events in WRFORCH and WRFORCHNEMO simulations

The slides are provided here : Slides by Namendra

Prediction skill of monthly maximum of daily precipitation in WRFORCH and WRFORCHNEMO simulations

The slides are provided here : Slides by Namendra