Hello Randy,
For projection_x/y_coordinates the convention says "The x (abscissa) and y (ordinate) rectangular coordinates are identified by the standard_name attribute values projection_x_coordinate and projection_y_coordinate respectively." -- I suppose, for clarity, this should be amended to read "The x (abscissa) and y (ordinate) rectangular coordinates with units conforming to metres are .....", to signal the limitations in applicability of these terms more clearly. I'm not sure what you want to suggest with your question about how specific the names should be. At the moment we have projection_x/y_coordinate as generic up to a point, and a range of grid mappings. The diagram I sent is relevant to the one under discussion, namely "geostationary", which currently covers two options. It could be replaced by a more generic mapping for spherical coordinates centred on an orbiting platform, but there is understandable reluctance to introduce such generalisations without a specific need. The analogue of "abscissa" and "ordinate" distances would, I think, be "polar/elevation" and "azimuthal" angles, but I can see that there would be an awkward clash with existing usage if we had standard names tied to these concepts. An alternative would be to have projection_x/y_angle and use the attributes of the grid_mapping to indicate which is which. I would prefer to avoid relying on terminology which is specific to the instrument architecture (e.g. sweep angle and fixed angle) and instead refer to the elevation and azimuthal angles. However, if this is seen as making life too difficult for the people working with geostationary data, it would be enough to just say how the "fixed" and "sweep" angles relate to the broader standard. Incidentally, the use of "perspective_point_height" doesn't quite fit the usage of "height" in the convention. In general height refers to distance from the actual surface, while "altitude" is used to refer to height above the geoid. We could avoid introducing the shape of the surface here by referring to the "perspective_point_radius". regards, Martin ________________________________ From: Randy Horne <[email protected]> Sent: 30 April 2018 15:04 To: Juckes, Martin (STFC,RAL,RALSP) Cc: Daniel Lee; [email protected] Subject: Re: [CF-metadata] Fix Geostationary projection, including proposal for two new standard names Hi Martin: Your diagram and accompanying narrative look good to me (other than spelling). It would be helpful for the definition of the “sweep angle” grid mapping parameter for the geostationary projection to be enhanced to include a graphic like this. Is it your thinking that the new standard names be defined so they are specific to the geostationary projection ? The new angular coordinate standard names could be used by multiple projections as is the case with the existing standard names projection_x_coordinate and projection_y_coordinate. Incidentally, the sweep angle for the GEOS satellites should be set to y ( see CF trac item 72), as Meteosat imaging scan is north/south. The sweep angle for GOES-R is east/west because the GOES-R imaging scan is east/west. v/r randy > On Apr 30, 2018, at 5:59 AM, Martin Juckes - UKRI STFC > <[email protected]> wrote: > > Hello All, > > > I've made a diagram of the geometry to try to get my understanding of these > rotations clearer: > > http://bit.ly/geosr_scan_geometry > > > The blue point in the sketch results from rotating the viewing vector through > an angle y about the axis S_x, which is the tangent to the satellite orbit, > and then through an angle x about S*_y, which is an axis formed by taking the > axis S_y and rotating it through angle x. > > > The above sentence describes the angles in a way which is related to the > functioning of the instrument, and I agree with Daniel that this is not an > ideal approach for the standard. > > > To respond to Jim's question about parametric coordinates: I think it would > be a shame to describe two components of a spherical polar coordinate system > as "parametric" ... they are spherical angles which are very widely used. The > challenge is to come up with a definition which is accurate and concise. > > > For the GEOS-R geometry we can define the angles in terms of a coordinate > axis (the tangent to the orbit, S_x) and a reference plane (the plane defined > by the Earth's polar axis and the vector from the centre of the Earth to the > satellite). Once the axis and reference plane are defined, y is the azimuthal > angle relative to the line from the satellite to the Earth's centre, and x is > pi/2 minus the polar angle. i.e. x is pi/2 minus the angle between the line > of sight and S_x. y is the angle between the line from the satellite to the > Earth's centre and the perpendicular projection of the line of sight onto the > reference plane. > > > For the alternative gimbal geometry a similar definition applies except that > the coordinate axis (S_y) is parallel to the Earth's axis and the reference > plane is the Earth's equatorial plane. Now y is pi/2 minus the angle between > the line of sight and S_y. x is the angle between the line from the satellite > to the Earth's centre and the perpendicular projection of the line of sight > onto the reference plane. > > > regards, > > Martin > > > > > ________________________________ > From: CF-metadata <[email protected]> on behalf of Martin > Juckes - UKRI STFC <[email protected]> > Sent: 20 April 2018 17:07 > To: Randy Horne; Daniel Lee > Cc: [email protected] > Subject: Re: [CF-metadata] Fix Geostationary projection, including proposal > for two new standard names > > Hello Randy, > > > the sweep_angle_axis parameter appears to be used to distinguish between the > two axis configurations, with a value of "x" to indicate that the GEOS series > have the sweep axis aligned with "x", which here means East/West. > > > You've answered my questions now, thanks for your patience. > > > Given that the angles are components of the well known spherical coordinate > system, which is, I believe defined in ISO 80000-2:2009, I feel it would be a > mistake to omit reference to that in the definition. I appreciate that the > complexity of the formulae (which I've attempted to work out from first > principles, with results below) means that, for practical applications, you > want to work with existing tools and procedures rather than starting from a > fundamental definition. On the other hand, we are trying to set up a standard > using universal language. > > > regards, > > Martin > > > > Working from first principles, the relationship between latitude and relative > longitude on the Earth and the projection coordinates are (for the case of > coordinate axis orientated E/W): > > cos(lat) * sin(lon) = r1 * sin( y ); > > sin(lat) = r1 * cos(y) * sin(x); > > where r1 is a solution of the quadratic: > > r1*r1 - 2*Rs*r1*cos(x)*cos(y) + Rs*Rs -1 = 0. > > > > > ________________________________ > From: Randy Horne <[email protected]> > Sent: 20 April 2018 15:15 > To: Daniel Lee > Cc: Juckes, Martin (STFC,RAL,RALSP); [email protected] > Subject: Re: [CF-metadata] Fix Geostationary projection, including proposal > for two new standard names > > Dear Martin: > > Note that the CF “geostationary projection” has been design to accommodate > both GOES-R and Meteosat (and Himawari, etc.), The ways the different > imagers do their scans is different causing the same (x.y) angular > coordinates to resolve to a different earth location. A projection > parameter “sweep angle” is included to allow the same projection to work for > both imager designs. > > A GOES-R writeup on this projection is located at: > https://www.goes-r.gov/users/docs/PUG-L1b-vol3.pdf, paragraph 5.1.2 ABI Fixed > Grid. > > v/r > > randy > > > >> On Apr 20, 2018, at 10:06 AM, Daniel Lee <[email protected]> wrote: >> >> Hi Martin, >> >> Yes, you need both coordinates in order to find the position viewed on the >> Earth's surface - it's not a simple distance measurement. >> >> I'm a bit confused about your use of azimuth vs. polar - do you mean >> meridional and zonal? >> >> I'd be cautious about bringing in too much information about how >> geostationary satellites function into the grid mapping. Not all satellites >> use the same scanning principals - technology changes and there have already >> been a number of different geostationary satellites deployed with different >> scanning characteristics, so that the geostationary constellation currently >> contains satellites with different scanning methods. This will surely >> continue to change in the future. >> >> At the end of the day, describing geostationary satellite data with the >> method NOAA is currently using is well-established and has worked well for >> decades in several data formats, so I don't think we'll benefit much from >> changing anything except for maybe the name of the variables. The >> methodology should clearly remain the same. >> >> Cheers, >> Daniel >> >>> -----Original Message----- >>> From: CF-metadata [mailto:[email protected]] On Behalf >>> Of Martin Juckes - UKRI STFC >>> Sent: 20 April 2018 15:41 >>> To: Randy Horne <[email protected]> >>> Cc: [email protected] >>> Subject: Re: [CF-metadata] Fix Geostationary projection, including proposal >>> for two new standard names >>> >>> Hello Randy, >>> >>> >>> thanks, that clears up a lot of my confusion. >>> >>> >>> Since the coordinates are N/S and E/W aligned, at least at the origin, it >>> may >>> be better to include this in the names. "x" and "y" are generally used for >>> coordinates which have an arbitrary orientation relative to the Earth's axis >>> which then needs to be specified in additional attribute values. >>> >>> >>> Your answer does not completely define the angles for me. If we consider a >>> point (A) which is, for example, at 45N at the same longitude, then it >>> angular >>> distance in the N/S direction is uniquely defined, but if we take another >>> point >>> (B) 45degrees to the east, then we have two angles and their values will >>> depend on the definition of the coordinate system. >>> >>> >>> I've found some documentation on geostationary satellites which suggests >>> that the viewing angles are related to the gimbal system, with an outer >>> "sweep" axis and an inner "fixed-angle" axis (this is from >>> proj4.org/projections/geos.html). Relating this back to the mathematical >>> terminology of spherical coordinates that I'm familiar with, I believe the >>> angle >>> of rotation around the sweep axis is the azimuthal angle and the rotation >>> around the fixed axis is the polar angle. I.e. we have a spherical >>> coordinate >>> system relative to the sweep axis. >>> >>> >>> The proj4.org document also states that the GEOS series have the sweep axis >>> aligned E/W, which would imply that projection_y_angular is an azimuthal >>> angle and projection_x_angular is a polar angle. With this information (and >>> the height of the satellite) I would be able to calculate the two angles for >>> point B. I'll get a different answer if projection_y_angular is the polar >>> angle >>> and projection_x_angular is azimuthal, so it is important to know which is >>> which. Perhaps projection_polar_angle, projection_azimuthal_angle would >>> be better? >>> >>> >>> The calculation would indeed be complicated, as has already been >>> emphasized below, but I think it is worth going back to the fundamentals >>> here, and stating the underlying assumptions behind the coordinate system. >>> E.g. the fact that it is defined relative to a fixed point above the Earth's >>> surface which corresponds to an ideal satellite position. >>> >>> >>> regards, >>> >>> Martin >>> >>> >>> >>> ________________________________ >>> From: Randy Horne <[email protected]> >>> Sent: 20 April 2018 13:56 >>> To: Juckes, Martin (STFC,RAL,RALSP) >>> Cc: [email protected] >>> Subject: Re: [CF-metadata] Fix Geostationary projection, including proposal >>> for two new standard names >>> >>> Hi Martin: >>> >>> RE: I agree with Jim that a little more basic information is needed about >>> what >>> the angles are. I may be misinterpreting the discussion, but I had imagined >>> that the angles as components of a spherical coordinate system centred on >>> the satellite, with the nadir at (0,0) ... is that correct? >>> >>> The projection_x_angular_coordinate and projection_y_angular coordinates >>> are the angular distances from the satellite’s nadir in the E/W an N/S >>> direction, respectively, from the ideal location of the imaging instrument >>> in >>> geostationary orbit. >>> >>> >>> v/r >>> >>> randy >>> >>> >>> >>>> On Apr 20, 2018, at 4:06 AM, Martin Juckes - UKRI STFC >>> <[email protected]> wrote: >>>> >>>> I agree with Jim that a little more basic information is needed about what >>> the angles are. I may be misinterpreting the discussion, but I had imagined >>> that the angles as components of a spherical coordinate system centred on >>> the satellite, with the nadir at (0,0) ... is that correct? >>> >>> _____________________________________ >>> >>> Randy C Horne ([email protected]) Principal Engineer, Excalibur >>> Laboratories Inc. >>> voice & fax: (321) 952.5100 >>> cell: (321) 693.1074 >>> url: http://www.excaliburlabs.com >>> >>> >>> >>> >>> _______________________________________________ >>> CF-metadata mailing list >>> [email protected] >>> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata >> >> Any email message from EUMETSAT is sent in good faith but shall neither be >> binding nor construed as constituting a commitment by EUMETSAT, except where >> provided for in a written agreement or contract or if explicitly stated in >> the email. Please note that any views or opinions presented in this email >> are solely those of the sender and do not necessarily represent those of >> EUMETSAT. This message and any attachments are intended for the sole use of >> the addressee(s) and may contain confidential and privileged information. >> Any unauthorised use, disclosure, dissemination or distribution (in whole or >> in part) of its contents is not permitted. If you received this message in >> error, please notify the sender and delete it from your system. > > _____________________________________ > > Randy C Horne ([email protected]) > Principal Engineer, Excalibur Laboratories Inc. > voice & fax: (321) 952.5100 > cell: (321) 693.1074 > url: http://www.excaliburlabs.com > > > > > _______________________________________________ > CF-metadata mailing list > [email protected] > http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata _____________________________________ Randy C Horne ([email protected]) Principal Engineer, Excalibur Laboratories Inc. voice & fax: (321) 952.5100 cell: (321) 693.1074 url: http://www.excaliburlabs.com _______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
