[Reddyproc-users] Flux Partitioning during polar summer with no night-time records

[Thomas Wutzler]

Dear Christina,

You may find better or faster help with specific problems by posting to
reddyproc-users at lists.r-forge.r-project.org

The interesting case with your data is, that you have no night-time
records during polar summer.
We did not think of this special case during implementing the flux
partitioning.

Usually, we REddyProc developer can not take the time to go through user
scripts.
Nevertheless in procrastinating my own work I notice some things:
- The input file is missing units in the row after column headings.
- You fill NEE twice, that may cause problems in the internal setup of
the class, I suggest to only use the filling after uStar
- The seasons are only relevant for the uStar threshold estimation. And
the uStar threshold estimation has to be done before gap-filling.
However,  with sparse data I recommend to stick to user-defined uStar
threshold cases. You can work with different scenarios and output
columns by providing argument suffix to the gapfilling of NEE and to the
further partitioning.
- Without proper uStar threshold distribution, you have to call
EProc$sGLFluxPartition separately for your uStar scenarios instead of
EProc$sGLFluxPartitionUStarScens
- When you specify a fixed temperature sensitivity you have to provide a
numerical value of activation energy in Kelvin (?fLloydTaylor)


For me the following line assumes a temperature sensitivity of 150K
(arbitrary, you should come up with your estimate)
EProc$sGLFluxPartition(controlGLPart = partGLControl(fixedTempSens =
data.frame(E0 = 150)))
complains about non-finite RRefNight

REddyProc, currently, assumes also in daytime-based partitioning to
estimate a prior for the temperature-respiration relationship from
night-time data.
You already specify the temperature sensitivity to avoid this problem. I
need to think how to implement also specifying a prior for the basal
respiration.

I formulated an issue in the development tracking system, and you need
to wait till we find time to look at this special case and come up with
a solution.
https://github.com/bgctw/REddyProc/issues/36

Yours
Thomas



On 12.11.19 19:44, Christina Braybrook wrote:
> Hello Dr. Wutzler,
> 
>  
> 
> I am a graduate student at Queen’s University, Kingston, Ontario, Canada
> working with Dr. Paul Treitz and Dr. Neal Scott. I am working through
> your post-processing workflow with the attached text file from our flux
> tower at Cape Bounty Arctic Watershed Observatory. I am aware that you
> need at least three months of data, so I have filled in with -9999
> values to achieve this requirement. I have also attached the script I
> have adapted for my txt file and there are errors associated with too
> finite records within the season when I attempt to estimate Ustar
> threshold that I believe are a result from the lack of values throughout
> the period specified.
> 
>  
> 
> Please let me know if you have any suggestions for working through this
> problem.
> 
>  
> 
> Thanks!
> 
>  
> 
> Christina
> 
>  
> 
> *Christina Braybrook, B.Sc.*
> 
> *
> 
> *
> 
> M.Sc. Candidate 
> 
> Department of Geography and Planning
> 
> Queen's University 
> 
> Kingston, ON, K7L 2N9 
> 
> Mackintosh-Corry Hall, D117
> 
>  
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5152 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.r-forge.r-project.org/pipermail/reddyproc-users/attachments/20191119/57077cbf/attachment.bin>