Interpretation of WEL files

[TobiasRosskopf]

Since Wave version 1.6.2 (commit 4d4712e) nearly all time series in WEL-files are interpreted as block right (except title includes "_VOL"):

BlueM.Wave/source/FileFormats/WEL.vb

Lines 207 to 212 in 037010d

	'Series in a WEL file are always interpreted as BlockRight, with the exception of volume
	If ts.Title.EndsWith("_VOL") Then
	ts.Interpretation = BlueM.Wave.TimeSeries.InterpretationEnum.Instantaneous
	Else
	ts.Interpretation = BlueM.Wave.TimeSeries.InterpretationEnum.BlockRight
	End If

Why is there a differation between volume curves (from BlueM.Sim) and everything else?

Could it be an enhencement to catch up multiple curve types from BlueM.Sim (and maybe other simulation models) and map it to the diffrent interpretation types?

Such as:

curve type	unit	appendices (BlueM.Sim)	interpretation
(presipitation) heights	mm	_NIE, _NEF, _ETP, _IZP, _TRA, _EVA, _ETA, _SCH	blockright
temperature	°C	_TEM	instantaneous
soil moisture	mm/m	_BOF	instantaneous
discharge	m3/s	_QOB, _QBA, _QIN, _1AB, _1ZU	instantaneous
water level	m	_WSP	instantaneous
velocity	m/s	_ v	instantaneous
volume	m3	_VOL	instantaneous
everything else	-		undefined

(Maybe uncomplete)

[jamaa]

The current implementation of how WEL series are interpreted actually corresponds to how Talsim calculates/outputs its results. All results are the average for the timestep, except for volume, which is instantaneous for the timestamp (don't ask me why). Are you sure BlueM.Sim is different?

[jamaa]

Actually it does seem to be different for BlueM.Sim:

Die Interpretation der ausgegeben Ganglinien variiert mit dem ausgegebenen Wert. So gilt z.B. für Temperatur und Abfluss die Interpretation "Block rechts", für Niederschlag und Verdunstung aber "Summenlinie pro Zeitschritt".

https://wiki.bluemodel.org/index.php/WEL-Format_(BlueM)

Unfortunately the information in the wiki is incomplete. Do you know more? We would have to figure out a way to distinguish BlueM WEL files from Talsim WEL files in order to implement this though.

maybe other simulation models

Unfortunately interpretation is often neglected when storing time series (and also in model documentation!), most file formats do not even support specifying it. But whatever we can find out should also be added to Wave. Let me know if you have any info.

For example, I have seen $ib Funktion-Interpretation: Summenlinie in the header of UVF files, but I have never been able to find any specification for this. I have the impression that only some software adds this kind of information (I don't even know which software), basically in the form of a comment, and we would be guessing on how exactly to read it. So until now, it has been ignored. There is a TODO in the code to at least store these comments as metadata, but I'm not sure how useful that would actually be to anyone

BlueM.Wave/source/FileFormats/UVF.vb

Lines 143 to 146 in 037010d

	If Zeile.StartsWith("$") Then
	'Kommentarzeile
	'TODO: store comments as metadata
	Continue Do

Because of this and other issues with file formats, I have sometimes even thought about defining a new time series file format, but you know how it is with standards :-) https://xkcd.com/927/

[jamaa]

To get back to the original question:

• The current implementation sets the interpretation of series read from WEL files according to what is defined by Talsim
• It is not entirely clear what the interpretation of WEL files output by BlueM.Sim is (where did you even get the info in your table at the top, @TobiasRosskopf ?)
• If we can clarify what interpretation(s) BlueM.Sim WEL files use, I would be happy to to implement it