[[PageOutline]] Last edited [[Timestamp]] [[BR]] '''Author''' : Gurvan Madec, Yevgeny Aksenov, Matthieu Leclair '''ticket''' : #665 '''Branch''' : [https://forge.ipsl.jussieu.fr/nemo/browser/branches/DEV_r1837_mass_heat_salt_fluxes DEV_r1837_mass_heat_salt_fluxes ] ---- useful commands {{{ * check out of NEMO directory: svn co svn+ssh://gm@forge.ipsl.jussieu.fr/ipsl/forge/projets/nemo/svn/branches/DEV_r1837_mass_heat_salt_fluxes/NEMO * commit : svn ci -m "ticket:#665 blah_blah" list_of_file }}} ---- === Description === Revisit the formulation of the fluxes between ocean, ice and atmosphere : an exchange of water (mass exchange) is now explicitly associated with an exchange of heat and salt content. This modification simplifies the implementation of a fully embedded sea-ice. This update is the NOCS.4 task (Revisit of mass and salt fluxes) due by 2nd July 2010 The main modifications concern OPA, LIM2 and LIM3. They are: ==== (0) stylistic modification ==== style changes only have been put in '''revision 1855'''. Nevertheless we have return back to the original revision 1850 from which we start from the trunk. The style changes we be re-added at the end of the development, after the testing phase. The following is thus done directly from the trunk. ==== (1) heat content of Freezing/Melting ice ==== In LIM2 only, the computation of the heat content of F/M ice is added (this is already done in LIM3 which is based on internal enrgy conservation, like CICE). The changes exactly mirror the changes introduced at NOCS in NEMO v3.0 when implementing first the sea-ice embedment. The basic idea is to reference at rt0 all the mass exchanges between ice and ocean. NB: since emps becomes the salt flux, it is generally zero except with the salt exchange at the ocean-ice interface and when a SSS restoring in salinity is used (not when a equivalent freshwater flux formulation is used for the SSS restoring). Therefore emps done not appear anymore in all the sbc routine,. It is only set to zero in sbc_init, set to the ice-ocean salt flux in limsbc(_2), and modified in sbcssr if SSS retoring is used. '''LIM2''' : the modules involved in the change are : (see also the '''revision 1858''') • '''ice_2.F90''' : add arrays to describe the heat content associated with both snow and ice mass changes and also change the name of the mass changes. Thus the following lines : {{{ REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdqsnif !: Variation of snow mass REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdqicif !: Variation of ice mass }}} becomes {{{ REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdm_snw !: Variation of snow mass over 1 time step [Kg/m2] REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdq_snw !: heat content associated to rdm_snw [J/m2] REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdq_ice !: Variation of ice mass over 1 time step [Kg/m2] REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: rdq_ice !: heat content associated to rdm_snw [J/m2] }}} • '''thd_ice_2.F90''' : associated 1D array names (rdsnif_1d, rdmicif_1d) have been changed into (rdm_snw_1d, rdm_ice_1d) and (rdq_snw_1d, rdq_ice_1d) have been added. • '''limthd_2.F90''' : introduce the new variable names, and pass the new variables (rdq_snw, rdq_ice) to the lim_thd_zdf and lim_thd_lac routines Note 1: we suppress the change of sst_m from Celcius to Kelvin and vice versa by adding a " + rt0 " at the 2 places it is used. Note 2: zfontn has been suppressed as the heat associated with the melting of solid precipitation is now added to qns in sbc modules ('''==>> TO BE DONE in SBC''') • '''limthd_lac_2.F90''' : introduce the new variable name (rdm_ice), and add the update of rdq_ice : the last DO loop of lim_thd_lac routine is transformed as follows : {{{ DO ji = kideb , kiut dvlbq_1d (ji) = ( 1. - frld_1d(ji) ) * h_ice_1d(ji) - ( 1. - zfrl_old(ji) ) * zhice_old(ji) rdmicif_1d(ji) = rdmicif_1d(ji) + rhoic * dvlbq_1d(ji) END DO }}} becomes {{{ DO ji = kideb , kiut dvlbq_1d (ji) = ( 1. - frld_1d(ji) ) * h_ice_1d(ji) - ( 1. - zfrl_old(ji) ) * zhice_old(ji) rdm_ice_1d(ji) = rdm_ice_1d(ji) + rhoic * dvlbq_1d(ji) rdq_ice_1d(ji) = rdq_ice_1d(ji) + rcpic * dvlbq_1d(ji) * ( tfu_1d(ji) - rt0 ) ! heat content relative to rt0 END DO }}} • '''limthd_zdf_2.F90''' : introduce the new variable names (rdm_snw, rdm_ice), and and add the computation/update of rdq_snw, rdq_ice : § " 9.1. Surface ablation and update of snow thickness and qstbif_1d ", the lines {{{ !--volume of snow lost after surface melting and the associated mass dvsbq_1d(ji) = ( 1.0 - frld_1d(ji) ) * ( h_snow_1d(ji) - zhsnw_old(ji) - zsprecip(ji) ) dvsbq_1d(ji) = MIN( zzero , dvsbq_1d(ji) ) rdmsnif_1d(ji) = rhosn * dvsbq_1d(ji) }}} becomes {{{ !--volume of snow lost after surface melting and the associated mass dvsbq_1d(ji) = ( 1.0 - frld_1d(ji) ) * ( h_snow_1d(ji) - zhsnw_old(ji) - zsprecip(ji) ) dvsbq_1d(ji) = MIN( zzero , dvsbq_1d(ji) ) ztmp = rhosn * dvsbq_1d(ji) rdm_snw_1d(ji) = ztmp !--heat content of the water provided to the ocean (referenced to rt0) rdq_snw_1d(ji) = cpic * ztmp * ( rt0_snow - rt0 ) }}} § " 9.4. Case of total ablation (ice is gone but snow may be left) ", the lines : {{{ !---updating new ice thickness and computing the newly formed ice mass zhicnew = zihgnew * zhicnew rdmicif_1d(ji) = rdmicif_1d(ji) + ( 1.0 - frld_1d(ji) ) * ( zhicnew - h_ice_1d(ji) ) * rhoic !---updating new snow thickness and computing the newly formed snow mass zhsnfi = zhsn + zdhsnm h_snow_1d(ji) = MAX( zzero , zhsnfi ) rdmsnif_1d(ji) = rdmsnif_1d(ji) + ( 1.0 - frld_1d(ji) ) * ( h_snow_1d(ji) - zhsn ) * rhosn }}} becomes {{{ !---updating new ice thickness and computing the newly formed ice mass zhicnew = zihgnew * zhicnew ztmp = ( 1.0 - frld_1d(ji) ) * ( zhicnew - h_ice_1d(ji) ) * rhoic rdm_ice_1d(ji) = rdm_ice_1d(ji) + ztmp !---heat content of the water provided to the ocean (referenced to rt0) ! use of rt0_ice is OK for melting ice, in case of freezing tfu_1d should be used. This is done in 9.5 section (see below) rdq_ice_1d(ji) = cpic * ztmp * ( rt0_ice - rt0 ) !---updating new snow thickness and computing the newly formed snow mass zhsnfi = zhsn + zdhsnm h_snow_1d(ji) = MAX( zzero , zhsnfi ) ztmp = ( 1.0 - frld_1d(ji) ) * ( h_snow_1d(ji) - zhsn ) * rhosn rdm_snw_1d(ji) = rdm_snw_1d(ji) + ztmp !---updating the heat content of the water provided to the ocean (referenced to rt0) rdq_snw_1d(ji) = rdq_snw_1d(ji) + cpic * ztmp * ( rt0_snow - rt0 ) }}} § " 9.5. Update internal temperature and ice thickness. ", the following lines have been added at the end : {{{ ! update the ice heat content given to the ocean in freezing case (part from rt0_ice to tfu_1d) ztmp = ( 1. - zidhb ) * rhoic * dvbbq_1d(ji) rdqicif_1d(ji) = rdqicif_1d(ji) + cpic * ztmp * ( tfu_1d(ji) - rt0_ice ) }}} § " 10. Surface accretion. ", the lines : {{{ !--- volume change of ice and snow (used for ocean-ice freshwater flux computation) rdmicif_1d(ji) = rdmicif_1d(ji) + ( 1.0 - frld_1d(ji) ) * ( zhicnew - h_ice_1d (ji) ) * rhoic rdmsnif_1d(ji) = rdmsnif_1d(ji) + ( 1.0 - frld_1d(ji) ) * ( zhsnnew - h_snow_1d(ji) ) * rhosn }}} becomes {{{ !--- volume change of ice and snow (used for ocean-ice freshwater flux computation) ztmp = ( 1.0 - frld_1d(ji) ) * ( zhicnew - h_ice_1d (ji) ) * rhoic rdm_ice_1d(ji) = rdm_ice_1d(ji) + ztmp rdq_ice_1d(ji) = rdq_ice_1d(ji) + cpic * ztmp * ( tfu_1d(ji) - rt0 ) !!gm BUG ?? snow ==> only needed for nn_ice_embd == 0 (standard levitating sea-ice) ztmp = ( 1.0 - frld_1d(ji) ) * ( zhsnnew - h_snow_1d(ji) ) * rhosn rdm_snw_1d(ji) = rdm_snw_1d(ji) + ztmp rdq_snw_1d(ji) = rdq_snw_1d(ji) + cpic * ztmp * ( rt0_snow - rt0 ) }}} and, at last, § " 11. Lateral ablation (Changes in sea/ice fraction) ", the lines : {{{ !--variation of ice volume and ice mass dvlbq_1d(ji) = zihic * ( zfrl_old(ji) - frld_1d(ji) ) * h_ice_1d(ji) rdmicif_1d(ji) = rdmicif_1d(ji) + dvlbq_1d(ji) * rhoic !--variation of snow volume and snow mass zdvsnvol = zihsn * ( zfrl_old(ji) - frld_1d(ji) ) * h_snow_1d(ji) rdmsnif_1d(ji) = rdmsnif_1d(ji) + zdvsnvol * rhosn }}} becomes {{{ !--variation of ice volume and ice mass dvlbq_1d(ji) = zihic * ( zfrl_old(ji) - frld_1d(ji) ) * h_ice_1d(ji) ztmp = dvlbq_1d(ji) * rhoic rdm_ice_1d(ji) = rdm_ice_1d(ji) + ztmp !!gm !!gm This should be split in two parts: !!gm 1- heat required to bring sea-ice at tfu : this part should be added to the heat flux taken from the ocean !!gm cpic * ztmp * 0.5 * ( tbif_1d(ji,2) + tbif_1d(ji,3) - 2.* rt0_ice ) !!gm 2- heat content of lateral ablation referenced to rt0 : this part only put in rdq_ice_1d !!gm cpic * ztmp * ( rt0_ice - rt0 ) !!gm Currently we put all the heat in rdq_ice_1d rdq_ice_1d(ji) = rdq_ice_1d(ji) + cpic * ztmp * 0.5 * ( tbif_1d(ji,2) + tbif_1d(ji,3) - 2.* rt0 ) ! !--variation of snow volume and snow mass zdvsnvol = zihsn * ( zfrl_old(ji) - frld_1d(ji) ) * h_snow_1d(ji) ztmp = zdvsnvol * rhosn rdm_snw_1d(ji) = rdm_snw_1d(ji) + ztmp !!gm !!gm This should be split in two parts: !!gm 1- heat required to bring snow at tfu : this part should be added to the heat flux taken from the ocean !!gm cpic * ztmp * ( tbif_1d(ji,1) - rt0_snow ) !!gm 2- heat content of lateral ablation referenced to rt0 : this part only put in rdqicif_1d !!gm cpic * ztmp * ( rt0_snow - rt0 ) !!gm Currently we put all the heat in rdqicif_1d rdq_snw_1d(ji) = rdq_snw_1d(ji) + cpic * ztmp * ( tbif_1d(ji,1) - rt0 ) }}} • '''phycst.F90''' : style changes + : {{{ - update of name: rau0r ==>> r1_rau0 ; ro0cpr = r1_rau0_rcp - new variable: xlsn = lfus * rhosn (for lim3, already defined in lim2) lfus = xlsn / rhosn (for lim2, already defined in lim3) }}} z1_rau0 implies changes in '''dynzdf_exp.F90''' ; '''dynzdf_imp.F90''' and '''eosbn2.F90''' r1_rau0_rcp implies changes in '''trabbc.F90''' ; '''traqsr.F90''' and '''eosbn2.F90''' CAUTION : there is a need to homogenize the definition of constant for sea-ice at least between LIM-2 and LIM-3. This must be coordonated with CICE variables ==== (2) LIM-3 specific changes see '''revision XXXX''' ==== In LIM3 case, the problem is somewhat different. Indeed, LIM3 is written in an energy conserving way (Bitz & Lipscomb 1999). Therefore the heat exchanged in both atmosphere and ocean is the internal energy (i.e. the heat flux through the interface + the heat content of the mass flux referenced to rt0). There is no need to introduce rdq_ice and rdq_snw fields. Five key elements have been changed : '''1- Addition of internal energy of snowfall over the ice'''. Reason: Tatm is used inside lim3 to compute the internal energy of snowfall. In coupled mode this will not be possible. Therefore, and in better agreement with SBC philosophy, we introduce qen_snw provided by sbcblk_c.. or sbccpl. • '''sbc_ice.F90''' : add the internal energie of snow in both LIM2 and LIM3 cases (even if in lim3 it will not be used) : {{{ REAL(wp), PUBLIC, DIMENSION(jpi,jpj) :: qen_snw !: internal energy of snowfall (heat content - Latent) [W/m2] }}} idem in thd_ice ( version _1d) • '''sbcice_lim.F90, sbcblk_core.F90 and sbcblk_clio.F90''' : add qen_snw to the call argument of blk_ice_c..., and add its computation (p_qen_s) in the 2 bulk formulae module: {{{ p_qen_s(:,:) = sf(jp_snow)%fnow(:,:) * ( cpic * ( sf(jp_tair)%fnow(:,:) - rtt - lfus ) }}} • '''sbccpl.F90''' : add the computation of qen_snw CAUTION : BUG correction snow CP is cpic not rcp.... correction made. • '''limthd.F90, limthd_ent.F90 and limthd_dh''' : pass qen_snw from 2D to 1D in limthd, replace the use of tatm by the use of qen_snw in limthd_dh : {{{ !!gm p_qen_s(:,:) = sf(jp_snow)%fnow(:,:) * ( cpic * ( sf(jp_tair)%fnow(:,:) - rtt ) - lfus ) zqprec = - p_qen_s(:,:) * rhosn / MAX( sprecip_1d(ji), 1.e-20 ) !!gm instead of : ! zqprec = rhosn * ( cpic * ( rtt - tatm_ice_1d(ji) ) + lfus ) ! energy of fallen snow !!gm }}} and in limthd_ent : {{{ !!gm p_qen_s(:,:) = sf(jp_snow)%fnow(:,:) * ( cpic * ( sf(jp_tair)%fnow(:,:) - rtt ) - lfus ) !!gm so rhosn * ( cpic * ( tatm - rtt ) - lfus ) = - p_qen_s(:,:) * rhosn / MAX( sprecip_1d(ji), 1.e-20 ) !!gm therefore: qm0(ji,1) = - ( ( 1. - snswi(ji) ) * p_qen_s(:,:) / MAX( sprecip_1d(ji), 1.e-20 ) & - snswi(ji) * ( cpic * ( t_s_b(ji,1) - rtt ) - lfus ) ) * rhosn * zthick0(ji,1) !!gm instead of : ! qm0(ji,1) = rhosn * ( cpic * ( rtt - ( 1. - snswi(ji) ) * tatm_ice_1d(ji) & ! & - snswi(ji) * t_s_b(ji,1) ) + lfus ) * zthick0(ji,1) !!gm }}} as a consequence, the declaration of tatm in key_lim3 case is suppressed from sbc_ice, and its initialisation from both sbcblk_core and clio. NB: in clio bulk, we replace the 2D ztatm array by a local scalar ztair '''OPEN QUESTION ==>''' Sublimation in coupled mode !!!! ===>>> Seb and Martin help required '''OPEN QUESTION ==>''' Coupled mode check the value of cpic used in the atmosphere as well as rcp !!! :-) '''2- Change in the snowfall partition between leads and ice'''. Reason: in LIM3 part of the snow that fall over ice is transported by the wind and kept in lead area. The parameterisation in coded deeply in LIM3 routine, whereas in coupled mode we need to have the information in sbccpl. Furthermore, with the partitioning at sbc level, LIM2 can also benefite from it. Introduce betas in the SBC, no more in LIM-3 alone. '''reason (1) :''' make it available in coupled mode (sbccpl.F90 will have to use it) ==>> Following change in LIM3 & sbccpl • '''sbc_oce.F90,sbcmod.F90, ice.F90 ''' : in namelist namsbc, add the following variable: {{{ REAL(wp), PUBLIC :: rn_psnw = 1.e0 !: coef. of lead/ice partition of snowfall (>0 and <=1) }}} add the lecture of the new variable and its print in sbcmod.F90 and remove the old variable (betas) from ice.F90 '''namelist ==> DON'T FORGET TO add the new variable in the ocean ref namsbc namelists !!! ''' '''namelist ==> DON'T FORGET TO remove betas from the lim3 ref namicethd namelists !!! ''' • '''limthd.F90 ''' : remove the lecture from namicethd namelist. Change the budget over lead from: {{{ zfontn = sprecip(ji,jj) * lfus ! energy of melting zfnsol = qns(ji,jj) ! total non solar flux qldif(ji,jj) = tms(ji,jj) * ( qsr(ji,jj) & & + zfnsol + fdtcn(ji,jj) - zfontn & & + ( 1.0 - zindb ) * fsbbq(ji,jj) ) & & * ( 1.0 - at_i(ji,jj) ) * rdt_ice }}} to the following lines: (with za_i = at_i ; and za_o = 1 - at_i = pfrld {{{ ! snowfall is partitionned between leads and ice. If snow fall was uniform, ! a fraction (1-at_i) would fall into leads but because of the winds, ! more snow falls on leads than on sea ice. Therefore a greater fraction ! (1-ai)**p (p=rn_psnw) of the total mass of snow (0< psnw < 1) falls in leads. ! ! leads ! ice ! NEW partition ! (1-ai)**p = zzc ! 1-(1-ai)**p = 1-zzc ! OLD partition ! (1-ai) = za_o ! ai = za_i ! difference ! zzc-za_o ! 1-zcc-za_i ! NEW sprecp ! sprecip*zzc/za_o ! sprecip*(1-zzc)/za_i ! new internal energy of snowfall over ice : qen_snw * (1-zzc)/za_i (idem sprecip) ! Since the snow internal energy over lead associated with the OLD snow partitioning ! has already been put in qns (see limblk_c..), we will update qns only with the ! difference between new and old snowfall over lead : qen_snw*(zzc-za_o) ! NB: In reality, psnw depends on wind speed, and should decrease with ! increasing wind speed but here, it is considered as a constant. ! An average value is 0.66 is suggested (Martin Vancoppenolle, personal comm. 2006) ! partitioning of snownfall : the snowfall over icy area is modified zzc = za_o**rn_psnw zzc_ice = sprecip(ji,jj) * ( 1.e0 - zzc ) / MAX( za_i, 1.e-20 ) * zinda zzc_oce = sprecip(ji,jj) * zzc / MAX( za_o, 1.e-20 ) * ( 1.e0 - zinda ) ! sprecip(ji,jj) = zzc_ice * sprecip(ji,jj) ! update snowfall and its internal energy over icy area qen_snw(ji,jj) = zzc_ice * qen_snw(ji,jj) ! (used only after over icy area) ! lead energy budget (qldif) except lateral accretion (done in limthd_lac) qldif(ji,jj) = tms(ji,jj) * rdt_ice & ! over 1 time step & * ( ( qsr(ji,jj) + qns(ji,jj) ) * za_o & ! total heat over lead & (-?)+ qen_snw(ji,jj) * ( za_o - zcc ) & ! heat due to snowfall change & + fdtcn(ji,jj) * za_o & ! turbulent ice-ocean heat !!gm ? & + fdtcn(ji,jj) * za_i & ! turbulent ice-ocean heat & + fsbbq(ji,jj) *( 1.- zindb ) * za_o ) ! residual heat from previous step ! sprecip(ji,jj) = zzc_ice * sprecip(ji,jj) ! update snowfall and its internal energy over icy area qen_snw(ji,jj) = zzc_ice * qen_snw(ji,jj) ! (used only after over icy area) !!gm end }}} Notes : (1) we now update here the snow partition between ice and ocean. (2) both qen_snw and sprecip are modified: suppression of partitioning in other limthd_... modules. (3) the line starting with the comment '!!gm ?' is there to emphasise a potential '''BUG'''. Potential because I'm not completely sure. see just below. (4) this include a bug correction, see just below: '''BUG''' correction : In LIM-3 a betas exponent has been introduced when computing the solid precipitation over the sea ice. Part of the snow that falls over sea-ice is put in the lead areas by the wind. Nevertheless, the latent heat flux due to snow melting is computed with sprecip *(1-at_i) whereas the actual solid precipitation over lead area is sprecip*(1-at_i**betas). If betas /=1, some latent heat is missing (see the ticket open on that purpose #670). '''POTENTIAL BUG''' It seems to me that the among of heat exchange through turbulent fluxes between the ice and the ocean should be fdtcn(ji,jj) * za_i, and not fdtcn(ji,jj) * za_o. Indeed the ice-ocean surface of contact is proportional to at_i (=za_i), not za_o ! '''===>> TO BE CHECKED with Martin''' • '''limthd_dh.F90 ''' : remove the snow partitioning: lines {{{ zcoeff = ( 1.0 - ( 1.0 - at_i_b(ji) )**betas ) / at_i_b(ji) zdh_s_pre(ji) = zcoeff * sprecip_1d(ji) * rdt_ice / rhosn }}} becomes : {{{ zdh_s_pre(ji) = sprecip_1d(ji) * rdt_ice / rhosn }}} • '''limsbc.F90 ''' : remove the lecture from namicethd namelist. Change the budget over lead from: {{{ & + sprecip(ji,jj) * ( 1. - pfrld(ji,jj)**betas ) & ! remove snow is intercepted by sea-ice (previous frld) }}} becomes : {{{ & + sprecip(ji,jj) * ( 1. - pfrld(ji,jj) ) ! remove snow intercepted by sea-ice (before thermo) }}} • '''sbccpl.F90 ''' : add the partitioning of snow : '''reason (2) :''' make it also avalable in LIM-2. ==>> the Following change is LIM-2 • '''limthd_2.F90 ''' : remove the lecture from namicethd namelist. Change the budget over lead from: '''3- Modify the salt flux computation from an equivalent freshwater flux to a salt flux''' '''4- update the ice-ocean interface to salt flux''' '''5- Add the coupled case in the interface ice-atmos and atmos-ocean''' '''BUG''' correction : limthd : use fse3t_m instead of fse3t '''==>> ONGOING work''' ==== (3) Non-solar flux including the heat content of mass exchanges (see '''revision 1859''') ==== the definition of qns is modified as follows: the non solar part of the surface heat flux takes now also into account the heat content changes due to the change in volume associated with evaporation, precipitation, ice freezing and ice melting. This heat content is evaluated using the temperature (expressed in Celcius) of the mass of water exchanged between the different media (ocean, atmosphere and ice). Note that the latent heat flux associated to the meltin of solid precipitation is now taken into account directly in qns in sbc... routines, instead of been added in limthd(_2) (cf. step (1)) • ocean-atmosphere mass exchanges : # liquid precipitation are assumed to be at the air temperature expressed in Celcius) # solid precipitation are assumed to be at least below the the melting temperature of snow (rtt_snow) : min( Tair , rtt_snow ) # evaporation are at the SST. Note that sst_m is used. This is thus different from the former implicit value assumed for both precip and evap (use of tn in trasbc.F90) We therefore need to add the air temperature to the sea-ice, change the definition of emps (now a salt flux), and add the heat content of the mass flux (emp) to the non-solar heat flux. The associated modified modules are : sbcmod.F90 ; sb • '''sbcmod.F90''' : initialize emps to zero. The following lines have been added {{{ emps(:,:) = 0.e0 ! the salt flux will be computed (i.e. will be non-zero) only if ! ! sea-ice is present, or lk_vvl=F, or surface salt restoring is used. }}} • '''sbcblk_core.F90''' : add in qns the latent flux of solid precipitation + heat content of precip. and evap. In blk_oce_core routine, the following lines {{{ !CDIR COLLAPSE qns(:,:) = zqlw(:,:) - zqsb(:,:) - zqla(:,:) ! Downward Non Solar flux !CDIR COLLAPSE emp (:,:) = zevap(:,:) - sf(jp_prec)%fnow(:,:) * rn_pfac * tmask(:,:,1) }}} becomes {{{ !CDIR COLLAPSE qns(:,:) = zqlw(:,:) - zqsb(:,:) - zqla(:,:) & ! Downward Non Solar flux & - sf(jp_snow)%fnow(:,:) * lfus ! including melting of solid precip !CDIR COLLAPSE emp (:,:) = ( zevap(:,:) & ! mass flux (evap. - precip.) & - sf(jp_prec)%fnow(:,:) * rn_pfac ) * tmask(:,:,1) !CDIR COLLAPSE qns(:,:) = qns(:,:) & ! heat content due to mass flux (T in Celcius) & - zevap(:,:) * pst(ji,jj) * rcp & ! evaporation at SST & + ( sf(jp_prec)%fnow(:,:) - sf(jp_snow)%fnow(:,:) ) & ! liquid precip at Tair & * ( sf(jp_tair)%fnow(:,:) - rt0 ) * rcp & & + sf(jp_snow)%fnow(:,:) & ! solid precip at min(Tair,Tsnow) & * ( MIN( sf(jp_tair)%fnow(:,:), rt0_snow ) - rt0 ) * cpic }}} • '''sbcblk_clio.F90''' : add in qns the latent flux of solid precipitation (only if sea-ice present, otherwise the solid precipitation are not computed) + heat content of precip. and evap. In blk_oce_clio routine, the following lines {{{ !CDIR COLLAPSE !CDIR NOVERRCHK DO jj = 1, jpj !CDIR NOVERRCHK DO ji = 1, jpi qns (ji,jj) = zqlw(ji,jj) - zqsb(ji,jj) - zqla(ji,jj) ! Downward Non Solar flux emp (ji,jj) = zqla(ji,jj) / cevap - sf(jp_prec)%fnow(ji,jj) / rday * tmask(ji,jj,1) END DO END DO emps(:,:) = emp(:,:) }}} becomes {{{ zcevap = rcp / cevap ! convert zqla ==> evap (Kg/m2/s) ==> m/s ==> W/m2 zcprec = rcp / rday ! convert prec ( mm/day ==> m/s) ==> W/m2 !CDIR COLLAPSE qns(:,:) = zqlw(:,:) - zqsb(:,:) - zqla(:,:) ! Downward Non Solar flux !CDIR COLLAPSE emp(:,:) = zqla(:,:) / cevap & ! freshwater flux & - sf(jp_prec)%fnow(:,:) / rday * tmask(:,:,1) ! ! add heat content of mass exhanges !CDIR COLLAPSE qns(:,:) = qns(:,:) & & - zqla(:,:) * pst(:,:) * zcevap & ! evaporation at SST in Celcius & + sf(jp_prec)%fnow(:,:) * sf(jp_tair)%fnow(:,:) * zcprec ! precipitation at Tair in Celcius ! NB: if sea-ice model, the snow precip are computed and the associated heat is added to qns (see blk_ice_clio) }}} and in blk_ice_clio, add the following lines: {{{ ! ----------------------------------------------------------------------------- ! ! Correct the OCEAN non solar flux with the existence of solid precipitation ! ! ---------------=====--------------------------------------------------------- ! !CDIR COLLAPSE qns(:,:) = qns(:,:) & ! update the non-solar heat flux with: & - p_spr(:,:) * lfus & ! melting solid precip & + p_spr(:,:) * MIN( sf(jp_tair)%fnow(:,:), rt0_snow - rt0 ) * cpic & ! solid P at least below melting & - p_spr(:,:) * sf(jp_tair)%fnow(:,:) * rcp ! remove solid precip }}} • '''sbccpl.F90''' : the latent flux of solid precipitation is already taken into account. The heat content of the mass flux [precip. - evap (-runoff) (-calving)] should be added but only over open ocean (the modification does not concern the surface ice fluxes). It is not a simple story. To be perfectly conservative between atmos. and ocean, the atmosphere should provide the heat content associated with the mass flux. This is not done with the current atmospheric component used. Therefore an approximation should be done: # evaporation : Since the atmosphere has seen the mean SST over the last coupling time interval. This previous SST is the one that should be used to compute the heat content of evaporation. It is not available but can be added to sbcssm module and in its restart. For the moment, we simply use the sst_m field. # precipitation : as in former forced runs, the temperature of precipitation is assumed to be at the SST: we use sst_m field. # runoff & calving : use sst_m as we have no better information In sbc_cpl_rcv routine (ocean-atmos. only), move the emp part before the qns part and add the following lines {{{ zcoef = xlsn / rhosn ! qns update over free ocean with: qns(:,:) = qns(:,:) - frcv(:,:,jpr_snow) * zcoef ! remove the latent heat flux of melting solid precipitation & - emp(:,:) * sst_m(:,:) * rcp ! remove the heat content of mass flux (assumed to be at SST) }}} In sbc_cpl_ice_flx routine (ocean-atmos-ice), the following lines {{{ IF( lk_diaar5 ) zcptn(:,:) = rcp * tn(:,:,1) ... zcoef = xlsn / rhosn ztmp(:,:) = p_frld(:,:,1) * zsnow(:,:) * zcoef pqns_tot(:,:) = pqns_tot(:,:) - ztmp(:,:) }}} becomes {{{ zcptn(:,:) = rcp * sst_m(:,:) ... zcoef = xlsn / rhosn ! qns_tot update over free ocean with: ztmp(:,:) = p_frld(:,:,1) * zsnow(:,:) * zcoef pqns_tot(:,:) = pqns_tot(:,:) & & - ztmp(:,:) & ! remove the latent heat flux of solid precipitation melting & - ( pemp_tot(:,:) & ! remove the heat content of mass flux (assumed to be at SST) & - pemp_ice(:,:) * p_frld(:,:,1) ) * zcptn(:,:) }}} • '''sbcana.F90 ; sbcflx.F90 ; sbcfwb.F90 ; sbcrnf.F90''' : add in qns the heat content of the mass flux taken at SST. immediate changes • '''sbcssr''' : add a change in qns if the SSS restoring is converted into an emp ; update emps as a salt flux is SSS restoring is just a salinity restoring term ==== (4) Update ocean forcing and ice forcing ==== see '''revision 1859''' except for LIM-3 (limsbc.F90) updated in '''revision XXXX''' Modification of trasbc and limsbc(_2) modules in order to take into account the above changes (heat content of water exchanges (LIM2) and explicit salt flux (LIM2 & 3) and coupled interface for LIM3) • ocean-atmosphere mass exchanges : In the un-approximated case (lk_vvl=T) the ocean receives at its surface (1) a salt flux associate with Freezing and Melting of sea-ice, and (2) a heat flux which includes the budget of the heat content of all mass exchanged with the atmosphere and sea-ice. In the linear free surface case (lk_vvl=F), the model volume of the ocean is assumed to be constant, therefore an extra term appears in both T and S forcing term, the concentration/dilution term : the volume (associated to mass flux) removed or add to the ocean is replaced by a same volume of ocean water at Tn and Sn, so that the model ocean volume remains constant. '''trasbc.F90''' T and S forcing terms becomes symmetric. The following lines : {{{ DO jj = 2, jpj DO ji = fs_2, fs_jpim1 ! vector opt. #if ! defined key_zco zse3t = 1. / fse3t(ji,jj,1) #endif IF( lk_vvl) THEN zta = ro0cpr * qns(ji,jj) * zse3t & ! temperature : heat flux & - emp(ji,jj) * zsrau * tn(ji,jj,1) * zse3t ! & cooling/heating effet of EMP flux zsa = 0.e0 ! No salinity concent./dilut. effect ELSE zta = ro0cpr * qns(ji,jj) * zse3t ! temperature : heat flux zsa = emps(ji,jj) * zsrau * sn(ji,jj,1) * zse3t ! salinity : concent./dilut. effect ENDIF ta(ji,jj,1) = ta(ji,jj,1) + zta ! add the trend to the general tracer trend sa(ji,jj,1) = sa(ji,jj,1) + zsa END DO END DO }}} becomes {{{ IF( lk_vvl ) THEN ! Variable Volume Layers case ===>> heat content of mass flux in qns DO jj = 2, jpj DO ji = fs_2, fs_jpim1 ! vector opt. z1_e3t_rau0 = 1./ ( fse3t(ji,jj,1) * rau0 ) ta(ji,jj,1) = ta(ji,jj,1) + z1_e3t_rau0 * qns (ji,jj) * r1_rcp ! non solar heat flux sa(ji,jj,1) = sa(ji,jj,1) + z1_e3t_rau0 * emps(ji,jj) ! salt flux (freezing/melting) END DO END DO ! ELSE ! Constant Volume layers case ===>> Concentration dillution effect DO jj = 2, jpj DO ji = fs_2, fs_jpim1 ! vector opt. z1_e3t_rau0 = 1./ ( fse3t(ji,jj,1) * rau0 ) ta(ji,jj,1) = ta(ji,jj,1) + z1_e3t_rau0 * ( qns (ji,jj) * r1_rcp & ! non solar heat flux & + emp (ji,jj) * tn(ji,jj,1) ) ! concent./dilut. effect sa(ji,jj,1) = sa(ji,jj,1) + z1_e3t_rau0 * ( emps(ji,jj) & ! salt flux (freezing/melting) & + emp (ji,jj) * sn(ji,jj,1) ) ! concent./dilut. effect END DO END DO ENDIF }}} • ocean-ice mass exchanges : Ice-IF : '''sbcice_if.F90''' : no changes except style as there is no mass flux exchanges in ice-if case. LIM-2 : '''limsbc_2.F90''' : {{{ .... lot of things.... }}} becomes {{{ ! ! - computation the solar flux at ocean surface #if defined key_coupled zqsr = qsr_tot(ji,jj) + ( fstric(ji,jj) - qsr_ice(ji,jj,1) ) * ( 1.0 - pfrld(ji,jj) ) #else zqsr = pfrld(ji,jj) * qsr(ji,jj) + ( 1. - pfrld(ji,jj) ) * fstric(ji,jj) #endif ! ! - computation the non solar heat flux at ocean surface zqns = - ( 1. - thcm(ji,jj) ) * zqsr & ! part of the solar energy used in leads & + iflt * ( fscmbq(ji,jj) + ffltbif(ji,jj) ) & & + ifral * ( ial * qcmif(ji,jj) + (1 - ial) * qldif(ji,jj) ) * r1_rdt_ice & & + ifrdv * ( qfvbq(ji,jj) + qdtcn(ji,jj) ) * r1_rdt_ice ! - store residual heat flux (put in the ocean at the next time-step) fsbbq(ji,jj) = ( 1.0 - ( ifvt + iflt ) ) * fscmbq(ji,jj) ! ??? ! ! - heat content of mass exchanged between ocean and sea-ice zqhc = ( rdq_snw(ji,jj) + rdq_ice(ji,jj) ) * r1_rdt_ice ! heat flux due to sown & ice heat content exchanges ! qsr(ji,jj) = zqsr ! solar heat flux qns(ji,jj) = zqns - fdtcn(ji,jj) + zqhc ! non solar heat flux ! !------------------------------------------! ! ! mass flux at the ocean surface ! ! !------------------------------------------! ! ! mass flux at the ocean-atmosphere interface (open ocean fraction = leads area) #if defined key_coupled ! ! coupled mode: zemp = + emp_tot(ji,jj) & ! net mass flux over the grid cell (ice+ocean area) & - emp_ice(ji,jj) * ( 1. - pfrld(ji,jj) ) ! minus the mass flux intercepted by sea-ice #else ! ! forced mode: zemp = + emp(ji,jj) * frld(ji,jj) & ! mass flux over open ocean fraction & - tprecip(ji,jj) * ( 1. - frld(ji,jj) ) & ! liquid precip. over ice reaches directly the ocean & + sprecip(ji,jj) * ( 1. - pfrld(ji,jj) ) & ! snow is intercepted by sea-ice (previous frld) #endif ! ! mass flux at the ocean/ice interface (sea ice fraction) zemp_snw = rdm_snw(ji,jj) * r1_rdt_ice ! snow melting = pure water that enters the ocean zfmm = rdm_ice(ji,jj) * r1_rdt_ice ! Freezing minus Melting (F-M) ! salt flux at the ice/ocean interface (sea ice fraction) [PSU*kg/m2/s] zfsalt = - sice_r(ji,jj) * zfmm ! F-M salt exchange zcd = soce_r(ji,jj) * zfmm ! concentration/dilution term due to F-M ! ! salt flux only : add concentration dilution term in salt flux and no F-M term in volume flux ! salt and mass fluxes : non concentartion dilution term in salt flux and add F-M term in volume flux emps(ji,jj) = zfsalt + zswitch * zcd ! salt flux (+ C/D if no ice/ocean mass exchange) emp (ji,jj) = zemp + zemp_snw + ( 1.- zswitch) * zfmm ! mass flux (- F/M mass flux if no ice/ocean mass exchange) ! }}} LIM-3 : '''limsbc.F90''' : see '''revision XXX'''. Add soce_r and sice_r as in LIM-2 case In routine '''lim_sbc_flx''' the computation of the heat, salt and mass fluxes has been completely revisited. In particular, the coupled case should be now OK, except for the albedo computation which has to be checked. The heat flux is now computed as follows: ''' ===>>>> CAUTION heat content still not added ! ''' {{{ ! !== solar heat flux at ocean surface ==! (N.B. use of frld before it changes due to the thermodynanics) #if defined key_coupled zqsr = qsr_tot(ji,jj) & ! total over a cell & - ( 1. - pfrld(ji,jj) ) * ( qsr_ice(ji,jj,1) - fstric(ji,jj) ) ! minus the fraction intercepted by ice #else zqsr = pfrld(ji,jj) * qsr(ji,jj) & ! fraction over open ocean & + ( 1. - pfrld(ji,jj) ) * fstric(ji,jj) ! plus the faction passing through the ice #endif fscmbq(ji,jj) = ( 1.0 - pfrld(ji,jj) ) * fstric(ji,jj) !!gm ??? to be checked ! !== non solar heat flux at ocean surface ==! zqns = - zqsr & ! remove the solar part & + iflt * ( fscmbq(ji,jj) ) & ! total abl -> fscmbq is given to the ocean & + ifral * ( ial * qcmif(ji,jj) + (1 - ial) * qldif(ji,jj) ) * r1_rdt_ice & & + ifrdv * ( qfvbq(ji,jj) + qdtcn(ji,jj) ) * r1_rdt_ice & & + fhmec(ji,jj) & ! snow melt in ridging & + fheat_rpo(ji,jj) & ! porous ridged ice formation & + fheat_res(ji,jj) ! residual heat flux (correction of ice thickness) ! fscmbq Part of the solar radiation transmitted through the ice and going to the ocean computed in limthd_zdf.F90 ! ffltbif Total heat content of the ice (brine pockets+ice) / delta_t ! qcmif Energy needed to bring the ocean surface layer until its freezing (ok) ! qldif heat balance of the lead (or of the open ocean) ! qfvbq i think this is wrong! ---> Array used to store energy in case of total lateral ablation ! qfvbq latent heat uptake/release after accretion/ablation ! qdtcn Energy from the turbulent oceanic heat flux heat flux coming in the lead IF( num_sal == 2 ) zqns = zqns + fhbri(ji,jj) ! brine drainage ! bottom radiative component is sent to the computation of the oceanic heat flux !== store residual heat flux == ! (put in the ocean at the next time-step) fsbbq(ji,jj) = ( 1.0 - ( ifvt + iflt ) ) * fscmbq(ji,jj) !== heat fluxes at the ocean interface ==! qsr(ji,jj) = zqsr ! solar heat flux qns(ji,jj) = zqns - fdtcn(ji,jj) ! non solar heat flux ! ! fdtcn : turbulent oceanic heat flux !lim2 ! ! - heat content of mass exchanged between ocean and sea-ice ! zqhc = ( rdq_snw(ji,jj) + rdq_ice(ji,jj) ) * r1_rdt_ice ! heat flux due to snow & ice heat content exchanges ! ! ! qsr(ji,jj) = zqsr ! solar heat flux ! qns(ji,jj) = zqns - fdtcn(ji,jj) + zqhc ! non solar heat flux ! !lim2 }}} The mass and salt fluxes are now computed as follows: {{{ !== mass flux at the ocean-atmosphere interface ==! (open ocean fraction = leads area) #if defined key_coupled ! ! coupled mode: zemp = emp_tot(ji,jj) & ! net mass flux over the grid cell (ice+ocean area) & - emp_ice(ji,jj) * ( 1. - pfrld(ji,jj) ) ! minus the mass flux intercepted by sea-ice #else ! ! forced mode: zemp = emp(ji,jj) * ( 1.0 - at_i(ji,jj) ) & ! mass flux over open ocean fraction & - tprecip(ji,jj) * at_i(ji,jj) & ! add liquid precip. over ice reaches directly the ocean & + sprecip(ji,jj) * ( 1. - pfrld(ji,jj)**betas ) & ! remove snow is intercepted by sea-ice (previous frld) #endif ! !== mass flux at the ocean/ice-snow interface ==! (sea ice fraction) zemp_snw = rdm_snw(ji,jj) * r1_rdt_ice & ! snow melting = pure water that enters the ocean & - fmmec (ji,jj) ! snow falling when ridging ! zemp_ice = + rdm_ice(ji,jj) * r1_rdt_ice ! Freezing minus Melting (F-M) ! !== salt flux at the ice/ocean interface ==! (sea ice fraction) [PSU*kg/m2/s] zfsalt = fsalt_res(ji,jj) & ! Residual salt flux due to correction of ice thickness & + fseqv (ji,jj) & ! Equivalent salt flux due to ice growth/melt & + fsbri (ji,jj) * zinda & ! brine drainage & + fsalt_rpo(ji,jj) ! porous ridged ice formation ! zcd = soce_r(ji,jj) * zemp_ice ! concentration/dilution term due to F-M ! !== total salt and mass fluxes at the ocean interface ==! ! zswitch=1, salt flux only : add concentration dilution term in salt flux and no F-M term in volume flux ! =0, salt and mass fluxes : no concentration dilution term in salt flux and add F-M term in volume flux emps(ji,jj) = zfsalt + zswitch * zcd ! salt flux (+ C/D if no ice/ocean mass exchange) emp (ji,jj) = zemp + zemp_snw + ( 1.- zswitch) * zemp_ice ! mass flux (- F/M mass flux if no ice/ocean mass exchange) ! fsbri(ji,jj) = zinda * fsbri(ji,jj) ! correcting brine and salt fluxes (diag.) ! }}} '''(4) change emp and emps names''' The emps field is suppress by introducing '''fsalt''', the salt flux at the ocean surface (only non zero below sea-ice) A symmetric name is chosen for the emp : '''fmass'''. Indeed emp is a mass flux expressed in Kg/m^2^/s. its name emp is miss leading: emp does not only represent the Evaporation minus Precipitation budget, as it also include the freezing or melting of sea-ice, and also sometime the runoff. Therefore using fmass appears more meaningful. in case of linear free surface (no key_vvl defined), the salt flux take into account the concentration/dilution term '''(5) update other parts of the system : C1D_SRC ; NST_SRC ; OFF_SRC ; TOP_SRC ...'' '''CAUTION''' : limsbc_2 : Baltic sea ice salinity changed from 2 PSU to 0.5 PSU !!!! change in limsbc.F90 to be done. Furthermore, only the orca2 case has been coded. All other orca configurations are missing. This should be introduce by a lat/lon definition of the baltic area. '''N.B.''' • LIM-3 needs the air temperature ... This is related to the heat content change when Melting the fallen snow. Think how to take this in a different way, so that in coupled mode it works ! • introduce (as was done for sea-ice embedment) the air temperature as an argument of both blk_ice_core and blk_ice_clio routines • Philosophical change: in near future it appears that the use of key_vvl should be replaced by key_linssh. In other words, the defaut option of NEMO will be a non-linear free surface (and thus variable volume for the model levels). This has to be discussed prior to its implementation. • In the same idea : the embedded sea-ice should become the defaut case, with in option the 2 levitating cases. A test should then force a levitating sea-ice if key_linssh defined • Idea for LIM-3 optimisation : instead of 2D => 1D transformation, is it possible to introduce a 3D (including ice categories) => 1D transformation. Gain: longer vector so clear in vector computer, less obvious on scalar one, except that the limthd... will become 1 gigantic done loop => better also for scalar. '''CAUTION''' • in coupled mode, the atmosphere and the ocean must share the same latent heat flux of solid precipitation • in coupled mode, the atmosphere should provide not only the mass flux [e-p(-runoff)(-calvin)] but also the heat content associated with this mass flux computed with a temperature in Celcius. ---- === Testing === Testing could consider (where appropriate) other configurations in addition to NVTK]. ||NVTK Tested||!'''YES/NO!'''|| ||Other model configurations||!'''YES/NO!'''|| ||Processor configurations tested||[ Enter processor configs tested here ]|| ||If adding new functionality please confirm that the [[BR]]New code doesn't change results when it is switched off [[BR]]and !''works!'' when switched on||!'''YES/NO/NA!'''|| (Answering UNSURE is likely to generate further questions from reviewers.) 'Please add further summary details here' * Processor configurations tested * etc---- === Bit Comparability === ||Does this change preserve answers in your tested standard configurations (to the last bit) ?||!'''YES/NO !'''|| ||Does this change bit compare across various processor configurations. (1xM, Nx1 and MxN are recommended)||!'''YES/NO!'''|| ||Is this change expected to preserve answers in all possible model configurations?||!'''YES/NO!'''|| ||Is this change expected to preserve all diagnostics? [[BR]]!,,!''Preserving answers in model runs does not necessarily imply preserved diagnostics. !''||!'''YES/NO!'''|| If you answered !'''NO!''' to any of the above, please provide further details: * Which routine(s) are causing the difference? * Why the changes are not protected by a logical switch or new section-version * What is needed to achieve regression with the previous model release (e.g. a regression branch, hand-edits etc). If this is not possible, explain why not. * What do you expect to see occur in the test harness jobs? * Which diagnostics have you altered and why have they changed?Please add details here........ ---- === System Changes === ||Does your change alter namelists?||!'''YES/NO !'''|| ||Does your change require a change in compiler options?||!'''YES/NO !'''|| If any of these apply, please document the changes required here....... ---- === Resources === !''Please !''summarize!'' any changes in runtime or memory use caused by this change......!'' ---- === IPR issues === ||Has the code been wholly (100%) produced by NEMO developers staff working exclusively on NEMO?||!'''YES/ NO !'''|| If No: * Identify the collaboration agreement details * Ensure the code routine header is in accordance with the agreement, (Copyright/Redistribution etc).Add further details here if required..........