partitionNEEGL.RdPartition NEE fluxes into GP and Reco using the daytime method.
partitionNEEGL(ds, NEEVar = if (!missing(NEEVar.s)) NEEVar.s else paste0("NEE",
suffixDash, "_f"), TempVar = if (!missing(TempVar.s)) TempVar.s else "Tair_f",
VPDVar = if (!missing(VPDVar.s)) VPDVar.s else "VPD_f",
RadVar = if (!missing(RadVar.s)) RadVar.s else "Rg_f",
suffix = if (!missing(Suffix.s)) Suffix.s else "",
NEEVar.s, TempVar.s, VPDVar.s, RadVar.s,
Suffix.s, ..., controlGLPart = partGLControl(),
isVerbose = TRUE, nRecInDay = 48L, lrcFitter = RectangularLRCFitter())dataset with all the specified input columns and full days in equidistant times
Variable of NEE
Filled air or soil temperature variable (degC)
Filled Vapor Pressure Deficit - VPD - (hPa)
Filled radiation variable
string inserted into column names before
identifier for NEE column defaults
(see sEddyProc_sMDSGapFillAfterUstar).
deprecated
deprecated
deprecated
deprecated
deprecated
identifier for NEE column defaults
(see sEddyProc_sMDSGapFillAfterUstar).
further arguments to
partGLExtractStandardData, such as PotRadVar
further default parameters,
see partGLControl
set to FALSE to suppress output messages
number of records within one day (for half-hourly data its 48)
R5 class instance
responsible for fitting the light response curve.
Current possibilities are RectangularLRCFitter(),
NonrectangularLRCFitter(),
and LogisticSigmoidLRCFitter().
Daytime-based partitioning of measured net ecosystem fluxes into gross primary production (GPP) and ecosystem respiration (Reco)
The fit to the light-response-curve is done by default using the Rectangular
hyperbolic function, as in Lasslop et al. (2010)
Alternative fittings can be used by providing the corresponding subclass of
LightResponseCurveFitter-class to lrcFitter argument.
(see LightResponseCurveFitter_predictGPP)
While the extrapolation uses filled data, the parameter optimization
may use only measured data, i.e. with specified quality flag.
Even with using filled VPD, there may be large gaps that have not been
filled.
With the common case where VPD is missing for fitting the LRC, by default
(with controlGLPart$isRefitMissingVPDWithNeglectVPDEffect = TRUE)
is to redo the estimation of LRC parameters with neglecting the VPD-effect.
Next, in the predictions (rows) with missing VPD are then replaced
with predictions
based on LRC-fits that neglected the VPD effect.
predicted ecosystem respiration: mumol CO2/m2/s
predicted gross primary production mumol CO2/m2/s
Further light response curve (LRC) parameters and
their standard deviation depend on the used LRC
(e.g. for the non-rectangular LRC
see NonrectangularLRCFitter_getParameterNames).
They are estimated for windows and are reported with the first record
of the window
NEE filtered for nighttime records (others NA)
NEE filtered for daytime records (others NA)
temperature after filtering for quality flag degree Celsius
vapour pressure deficit after filtering for quality flag, hPa
basal respiration estimated from nighttime (W / m2)
quality flag: 0: good parameter fit, 1: some parameters out of range, required refit, 2: next parameter estimate is more than two weeks away
records until or after closest record that has a parameter estimate associated
information why LRC-fit was not
successful or was rejected, see result of
LightResponseCurveFitter_fitLRC
predicted GPP at VPD = 0 and PAR = 2000: a surrogate for maximum photosynthetic capacity
list object of fitting results including iOpt and covParms
same as FP_OPT_VPD holding optimization results with fit neglecting the VPD effect
Lasslop G, Reichstein M, Papale D, et al. (2010) Separation of net ecosystem exchange into assimilation and respiration using a light response curve approach: critical issues and global evaluation. Global Change Biology, Volume 16, Issue 1, Pages 187-208
partGLFitNightTimeTRespSens
partGLFitLRCWindows
partGLInterpolateFluxes