Re: [arts-users] FASTEM in ARTS
Dear Patrick and Oliver, Thank you for your advice! Indeed, I had used the 2.2.64 stable version assuming it was a safe choice for a beginner. Now I tried Patrick's piece of code for FASTEM call with the current ARTS version from Oliver's GitHub (2.3.1234, I assume), and everything worked flawlessly. I appreciate your help a lot. Best regards, Natalia Natalia Bliankinshtein Research Council Officer/Agent du Conseil de recherches NRC Aerospace/CNRC Aérospatiale National Research Council Canada/Conceil National de recherches Canada 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 Bldg/Edifice U-61 | Rm/Ch. 233A 1920 Research Private Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 natalia.bliankinsht...@nrc-cnrc.gc.ca From: Patrick Eriksson Sent: March 7, 2019 2:21 PM To: Bliankinshtein, Natalia; arts_users.mi@lists.uni-hamburg.de Subject: Re: [arts-users] FASTEM in ARTS Dear Natalia, Try this: VectorCreate( trv ) NumericCreate( wind_speed ) NumericCreate( wind_dir ) AgendaSet( iy_surface_agenda ){ SurfaceDummy specular_losCalc ppathCalc( rte_pos=rtp_pos, rte_los=specular_los ) ArrayOfStringSet( iy_aux_vars, ["Optical depth"] ) iyEmissionStandard transmittanceFromIy_aux( transmittance=trv ) surfaceFastem( wind_speed=wind_speed, wind_direction=wind_dir, transmittance=trv ) iySurfaceRtpropCalc } This should work. I have not actually made a test run (due to lack of time) but hopefully I have not missed any detail. (If you use this, you can ignore surface_rtprop_agenda). FASTEM does not fit into the pattern of ARTS, as it wants the transmission as input. This makes the agenda definition a bit messy. FASTEM 6 is default. My suggestion uses the default for salinity, adopt if you want to use another value. You need a quite recent version of ARTS 2.3 for this. I added the method transmittanceFromIy_aux not a long time ago. Bye, Patrick On 2019-03-07 19:22, Bliankinshtein, Natalia wrote: > Dear Patrick, > > Thanks for your prompt reply! I would like to use FASTEM in a simple 1D > forward model configuration, so no retrievals of surface properties will be > needed. I am particularly interested in FASTEM6 - not sure how its versions > align with those of ARTS. Also, I think a simple prescribing of surface wind, > temperature and salinity as Numerics would be sufficient for my purposes. > > I have started building my controlfile upon the classroom exercise #2 > (rtcalc) included in the release package, so if you could show me how to > include FASTEM into that example, that would be just great. > Please let me know if you need more info about the planned usage to give me a > hint on this. > > Thank you for your help, > Natalia > > > > Natalia Bliankinshtein > > Research Council Officer/Agent du Conseil de recherches > NRC Aerospace/CNRC Aérospatiale > National Research Council Canada/Conceil National de recherches Canada > 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 > Bldg/Edifice U-61 | Rm/Ch. 233A > 1920 Research Private > Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 > natalia.bliankinsht...@nrc-cnrc.gc.ca > > > From: Patrick Eriksson > Sent: March 7, 2019 2:43 AM > To: Bliankinshtein, Natalia; arts_users.mi@lists.uni-hamburg.de > Subject: Re: [arts-users] FASTEM in ARTS > > Dear Natalia, > > Yes, FASTEM is at hand, but it is a bit tricky to give a general answer. > It depends on e.g. > > * Do you want to retrieve any surface properties? > > * Do you want to set specific wind speeds etc, or do you want to obtain > wind speed etc. by interpolation from some data? > > Can you give us some more information? > > Bye, > > Patrick > > > > On 2019-03-06 20:51, Bliankinshtein, Natalia wrote: >> Dear ARTS users and developers, >> >> >> I have recently started learning ARTS and appreciate its extreme >> flexibility. >> >> According to the documentation, ARTS does not include an option of using >> FASTEM6 water surface model, however I see some evidence people have >> used it, for example here in Mr. Oliver Lemke's repository >> https://github.com/olemke/arts/tree/master/3rdparty/fastem and in Mr. >> Stuart Fox's email to the list in November 2018. >> >> >> I am struggling though to come up with a controlfile that would call >> FASTEM6. So I am kindly asking if anyone could please share a minimum >> working example of how to do that? >> >> >> Also, a workaround that could work for my purposes would be reading >> surface reflectivities from a file, which I tried. It seems to me, >> however, that surface emissivities are always computed as >> (1-reflectivity), which is not the case for FAST
Re: [arts-users] FASTEM in ARTS
Hi Natalia, Since you mentioned using classroom exercise #2 from the release package and Patrick suggesting to switch to the current development version instead, I'd like to point out that the exercises are not part of the ARTS package anymore. But you can find a separate arts-lectures package with up to date controlfiles in our SVN repository: svn co https://arts.mi.uni-hamburg.de/svn/rt/arts-lectures/trunk arts-lectures Cheers, Oliver > On 7 Mar 2019, at 20:21, Patrick Eriksson > wrote: > > Dear Natalia, > > Try this: > > VectorCreate( trv ) > NumericCreate( wind_speed ) > NumericCreate( wind_dir ) > > AgendaSet( iy_surface_agenda ){ > SurfaceDummy > specular_losCalc > ppathCalc( rte_pos=rtp_pos, rte_los=specular_los ) > ArrayOfStringSet( iy_aux_vars, ["Optical depth"] ) > iyEmissionStandard > transmittanceFromIy_aux( transmittance=trv ) > surfaceFastem( wind_speed=wind_speed, wind_direction=wind_dir, > transmittance=trv ) > iySurfaceRtpropCalc > } > > This should work. I have not actually made a test run (due to lack of time) > but hopefully I have not missed any detail. (If you use this, you can ignore > surface_rtprop_agenda). > > FASTEM does not fit into the pattern of ARTS, as it wants the transmission as > input. This makes the agenda definition a bit messy. > > FASTEM 6 is default. > > My suggestion uses the default for salinity, adopt if you want to use another > value. > > You need a quite recent version of ARTS 2.3 for this. I added the method > transmittanceFromIy_aux not a long time ago. > > Bye, > > Patrick > > > > On 2019-03-07 19:22, Bliankinshtein, Natalia wrote: >> Dear Patrick, >> Thanks for your prompt reply! I would like to use FASTEM in a simple 1D >> forward model configuration, so no retrievals of surface properties will be >> needed. I am particularly interested in FASTEM6 - not sure how its versions >> align with those of ARTS. Also, I think a simple prescribing of surface >> wind, temperature and salinity as Numerics would be sufficient for my >> purposes. >> I have started building my controlfile upon the classroom exercise #2 >> (rtcalc) included in the release package, so if you could show me how to >> include FASTEM into that example, that would be just great. >> Please let me know if you need more info about the planned usage to give me >> a hint on this. >> Thank you for your help, >> Natalia >> Natalia Bliankinshtein >> Research Council Officer/Agent du Conseil de recherches >> NRC Aerospace/CNRC Aérospatiale >> National Research Council Canada/Conceil National de recherches Canada >> 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 >> Bldg/Edifice U-61 | Rm/Ch. 233A >> 1920 Research Private >> Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 >> natalia.bliankinsht...@nrc-cnrc.gc.ca >> >> From: Patrick Eriksson >> Sent: March 7, 2019 2:43 AM >> To: Bliankinshtein, Natalia; arts_users.mi@lists.uni-hamburg.de >> Subject: Re: [arts-users] FASTEM in ARTS >> Dear Natalia, >> Yes, FASTEM is at hand, but it is a bit tricky to give a general answer. >> It depends on e.g. >> * Do you want to retrieve any surface properties? >> * Do you want to set specific wind speeds etc, or do you want to obtain >> wind speed etc. by interpolation from some data? >> Can you give us some more information? >> Bye, >> Patrick >> On 2019-03-06 20:51, Bliankinshtein, Natalia wrote: >>> Dear ARTS users and developers, >>> >>> >>> I have recently started learning ARTS and appreciate its extreme >>> flexibility. >>> >>> According to the documentation, ARTS does not include an option of using >>> FASTEM6 water surface model, however I see some evidence people have >>> used it, for example here in Mr. Oliver Lemke's repository >>> https://github.com/olemke/arts/tree/master/3rdparty/fastem and in Mr. >>> Stuart Fox's email to the list in November 2018. >>> >>> >>> I am struggling though to come up with a controlfile that would call >>> FASTEM6. So I am kindly asking if anyone could please share a minimum >>> working example of how to do that? >>> >>> >>> Also, a workaround that could work for my purposes would be reading >>> surface reflectivities from a file, which I tried. It seems to me, >>> however, that surface emissivities are always computed as >>> (1-reflectivity), which is not the case for FASTEM. Thus my question: is >>> it possible to
Re: [arts-users] FASTEM in ARTS
Dear Natalia, Try this: VectorCreate( trv ) NumericCreate( wind_speed ) NumericCreate( wind_dir ) AgendaSet( iy_surface_agenda ){ SurfaceDummy specular_losCalc ppathCalc( rte_pos=rtp_pos, rte_los=specular_los ) ArrayOfStringSet( iy_aux_vars, ["Optical depth"] ) iyEmissionStandard transmittanceFromIy_aux( transmittance=trv ) surfaceFastem( wind_speed=wind_speed, wind_direction=wind_dir, transmittance=trv ) iySurfaceRtpropCalc } This should work. I have not actually made a test run (due to lack of time) but hopefully I have not missed any detail. (If you use this, you can ignore surface_rtprop_agenda). FASTEM does not fit into the pattern of ARTS, as it wants the transmission as input. This makes the agenda definition a bit messy. FASTEM 6 is default. My suggestion uses the default for salinity, adopt if you want to use another value. You need a quite recent version of ARTS 2.3 for this. I added the method transmittanceFromIy_aux not a long time ago. Bye, Patrick On 2019-03-07 19:22, Bliankinshtein, Natalia wrote: Dear Patrick, Thanks for your prompt reply! I would like to use FASTEM in a simple 1D forward model configuration, so no retrievals of surface properties will be needed. I am particularly interested in FASTEM6 - not sure how its versions align with those of ARTS. Also, I think a simple prescribing of surface wind, temperature and salinity as Numerics would be sufficient for my purposes. I have started building my controlfile upon the classroom exercise #2 (rtcalc) included in the release package, so if you could show me how to include FASTEM into that example, that would be just great. Please let me know if you need more info about the planned usage to give me a hint on this. Thank you for your help, Natalia Natalia Bliankinshtein Research Council Officer/Agent du Conseil de recherches NRC Aerospace/CNRC Aérospatiale National Research Council Canada/Conceil National de recherches Canada 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 Bldg/Edifice U-61 | Rm/Ch. 233A 1920 Research Private Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 natalia.bliankinsht...@nrc-cnrc.gc.ca From: Patrick Eriksson Sent: March 7, 2019 2:43 AM To: Bliankinshtein, Natalia; arts_users.mi@lists.uni-hamburg.de Subject: Re: [arts-users] FASTEM in ARTS Dear Natalia, Yes, FASTEM is at hand, but it is a bit tricky to give a general answer. It depends on e.g. * Do you want to retrieve any surface properties? * Do you want to set specific wind speeds etc, or do you want to obtain wind speed etc. by interpolation from some data? Can you give us some more information? Bye, Patrick On 2019-03-06 20:51, Bliankinshtein, Natalia wrote: Dear ARTS users and developers, I have recently started learning ARTS and appreciate its extreme flexibility. According to the documentation, ARTS does not include an option of using FASTEM6 water surface model, however I see some evidence people have used it, for example here in Mr. Oliver Lemke's repository https://github.com/olemke/arts/tree/master/3rdparty/fastem and in Mr. Stuart Fox's email to the list in November 2018. I am struggling though to come up with a controlfile that would call FASTEM6. So I am kindly asking if anyone could please share a minimum working example of how to do that? Also, a workaround that could work for my purposes would be reading surface reflectivities from a file, which I tried. It seems to me, however, that surface emissivities are always computed as (1-reflectivity), which is not the case for FASTEM. Thus my question: is it possible to override surface emissivity and to read it from a file instead? Looking forward to hear your valuable advice. Best regards, Natalia *Natalia Bliankinshtein* Research Council Officer/Agent du Conseil de recherches NRC Aerospace/CNRC Aérospatiale National Research Council Canada/Conceil National de recherches Canada 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 Bldg/Edifice U-61 | Rm/Ch. 233A 1920 Research Private Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 natalia.bliankinsht...@nrc-cnrc.gc.ca ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi
Re: [arts-users] FASTEM in ARTS
Dear Patrick, Thanks for your prompt reply! I would like to use FASTEM in a simple 1D forward model configuration, so no retrievals of surface properties will be needed. I am particularly interested in FASTEM6 - not sure how its versions align with those of ARTS. Also, I think a simple prescribing of surface wind, temperature and salinity as Numerics would be sufficient for my purposes. I have started building my controlfile upon the classroom exercise #2 (rtcalc) included in the release package, so if you could show me how to include FASTEM into that example, that would be just great. Please let me know if you need more info about the planned usage to give me a hint on this. Thank you for your help, Natalia Natalia Bliankinshtein Research Council Officer/Agent du Conseil de recherches NRC Aerospace/CNRC Aérospatiale National Research Council Canada/Conceil National de recherches Canada 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 Bldg/Edifice U-61 | Rm/Ch. 233A 1920 Research Private Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 natalia.bliankinsht...@nrc-cnrc.gc.ca From: Patrick Eriksson Sent: March 7, 2019 2:43 AM To: Bliankinshtein, Natalia; arts_users.mi@lists.uni-hamburg.de Subject: Re: [arts-users] FASTEM in ARTS Dear Natalia, Yes, FASTEM is at hand, but it is a bit tricky to give a general answer. It depends on e.g. * Do you want to retrieve any surface properties? * Do you want to set specific wind speeds etc, or do you want to obtain wind speed etc. by interpolation from some data? Can you give us some more information? Bye, Patrick On 2019-03-06 20:51, Bliankinshtein, Natalia wrote: > Dear ARTS users and developers, > > > I have recently started learning ARTS and appreciate its extreme > flexibility. > > According to the documentation, ARTS does not include an option of using > FASTEM6 water surface model, however I see some evidence people have > used it, for example here in Mr. Oliver Lemke's repository > https://github.com/olemke/arts/tree/master/3rdparty/fastem and in Mr. > Stuart Fox's email to the list in November 2018. > > > I am struggling though to come up with a controlfile that would call > FASTEM6. So I am kindly asking if anyone could please share a minimum > working example of how to do that? > > > Also, a workaround that could work for my purposes would be reading > surface reflectivities from a file, which I tried. It seems to me, > however, that surface emissivities are always computed as > (1-reflectivity), which is not the case for FASTEM. Thus my question: is > it possible to override surface emissivity and to read it from a file > instead? > > > Looking forward to hear your valuable advice. > > > Best regards, > > Natalia > > > > *Natalia Bliankinshtein* > > Research Council Officer/Agent du Conseil de recherches > NRC Aerospace/CNRC Aérospatiale > National Research Council Canada/Conceil National de recherches Canada > 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 > Bldg/Edifice U-61 | Rm/Ch. 233A > 1920 Research Private > Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 > natalia.bliankinsht...@nrc-cnrc.gc.ca > > > > ___ > arts_users.mi mailing list > arts_users.mi@lists.uni-hamburg.de > https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi > ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi
Re: [arts-users] FASTEM in ARTS
Dear Natalia, Yes, FASTEM is at hand, but it is a bit tricky to give a general answer. It depends on e.g. * Do you want to retrieve any surface properties? * Do you want to set specific wind speeds etc, or do you want to obtain wind speed etc. by interpolation from some data? Can you give us some more information? Bye, Patrick On 2019-03-06 20:51, Bliankinshtein, Natalia wrote: Dear ARTS users and developers, I have recently started learning ARTS and appreciate its extreme flexibility. According to the documentation, ARTS does not include an option of using FASTEM6 water surface model, however I see some evidence people have used it, for example here in Mr. Oliver Lemke's repository https://github.com/olemke/arts/tree/master/3rdparty/fastem and in Mr. Stuart Fox's email to the list in November 2018. I am struggling though to come up with a controlfile that would call FASTEM6. So I am kindly asking if anyone could please share a minimum working example of how to do that? Also, a workaround that could work for my purposes would be reading surface reflectivities from a file, which I tried. It seems to me, however, that surface emissivities are always computed as (1-reflectivity), which is not the case for FASTEM. Thus my question: is it possible to override surface emissivity and to read it from a file instead? Looking forward to hear your valuable advice. Best regards, Natalia *Natalia Bliankinshtein* Research Council Officer/Agent du Conseil de recherches NRC Aerospace/CNRC Aérospatiale National Research Council Canada/Conceil National de recherches Canada 1200 Montreal Rd, Ottawa (Ontario) K1A 0R6 Bldg/Edifice U-61 | Rm/Ch. 233A 1920 Research Private Tel.: (613) 998-5349 | Cellphone: (343) 549-4262 natalia.bliankinsht...@nrc-cnrc.gc.ca ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi
Re: [arts-users] FASTEM in ARTS
Thanks – that works fine. Stuart From: Richard Larsson Sent: 08 November 2018 17:01 To: Fox, Stuart Cc: arts users mi Subject: Re: [arts-users] FASTEM in ARTS Hi Stuart, A development model problem. I am not sure this will solve your issues completely, but can you try adding SurfaceDummy at the start of your iy_surface_agenda? The new variables you are seeing are part of a method Patrick is working on for calculating the Jacobian of surface features. SurfaceDummy exists as a method to ignore said feature and have ARTS work as before. See line 180 in controlfiles/general/agendas.arts for an example. Probably, ARTS-2.3.638 did not have said line because the feature is new since ARTS-2-3-1010. With hope, //Richard Den tors 8 nov. 2018 kl 17:40 skrev Fox, Stuart mailto:stuart@metoffice.gov.uk>>: I’m currently trying to use the ARTS version of FASTEM via iySurfaceFastem in iy_surface_agenda. This used to work, but on the latest version of the ARTS trunk (2.3.1146) it fails with: Run-time error in method: AgendaSet The agenda iy_surface_agenda must generate the output WSV dsurface_rmatrix_dx, but it does not. It only generates: diy_dx iy auto_iySurfaceFastem_gin3_fastem_version The section of my controlfile that sets the surface is: NumericCreate(wind_speed) NumericCreate(salinity) NumericCreate(wind_dir) ReadXML(wind_speed,"surface_wind_speed.xml") ReadXML(wind_dir, "surface_wind_dir.xml") ReadXML(salinity, "surface_salinity.xml") AgendaSet(iy_surface_agenda){ iySurfaceFastem(salinity=salinity, wind_speed=wind_speed, wind_direction=wind_dir) } This used to work (certainly it worked at version 2.3.638), so I’m guessing something has changed (it looks like it’s related to calculating surface jacobians). Is this a bug, or do I need to add something extra to iy_surface_agenda to calculate the required quantities? Regards, Stuart ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de<mailto:arts_users.mi@lists.uni-hamburg.de> https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi
Re: [arts-users] FASTEM in ARTS
Hi Stuart, A development model problem. I am not sure this will solve your issues completely, but can you try adding SurfaceDummy at the start of your iy_surface_agenda? The new variables you are seeing are part of a method Patrick is working on for calculating the Jacobian of surface features. SurfaceDummy exists as a method to ignore said feature and have ARTS work as before. See line 180 in controlfiles/general/agendas.arts for an example. Probably, ARTS-2.3.638 did not have said line because the feature is new since ARTS-2-3-1010. With hope, //Richard Den tors 8 nov. 2018 kl 17:40 skrev Fox, Stuart : > I’m currently trying to use the ARTS version of FASTEM via iySurfaceFastem > in iy_surface_agenda. This used to work, but on the latest version of the > ARTS trunk (2.3.1146) it fails with: > > Run-time error in method: AgendaSet > > The agenda iy_surface_agenda must generate the output WSV > dsurface_rmatrix_dx, > > but it does not. It only generates: > > diy_dx > > iy > > auto_iySurfaceFastem_gin3_fastem_version > > > > The section of my controlfile that sets the surface is: > > NumericCreate(wind_speed) > > NumericCreate(salinity) > > NumericCreate(wind_dir) > > ReadXML(wind_speed,"surface_wind_speed.xml") > > ReadXML(wind_dir, "surface_wind_dir.xml") > > ReadXML(salinity, "surface_salinity.xml") > > AgendaSet(iy_surface_agenda){ > > iySurfaceFastem(salinity=salinity, wind_speed=wind_speed, > wind_direction=wind_dir) > > } > > > > This used to work (certainly it worked at version 2.3.638), so I’m > guessing something has changed (it looks like it’s related to calculating > surface jacobians). Is this a bug, or do I need to add something extra to > iy_surface_agenda to calculate the required quantities? > > > > Regards, > > > > Stuart > > > ___ > arts_users.mi mailing list > arts_users.mi@lists.uni-hamburg.de > https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi > ___ arts_users.mi mailing list arts_users.mi@lists.uni-hamburg.de https://mailman.rrz.uni-hamburg.de/mailman/listinfo/arts_users.mi