Extended usage of the gap filling algorithm
#+++ load libraries used in this vignette
library(REddyProc)
library(dplyr)
#+++ Add time stamp in POSIX time format
EddyDataWithPosix.F <- fConvertTimeToPosix(Example_DETha98, 'YDH',Year.s = 'Year'
,Day.s = 'DoY',Hour.s = 'Hour')
#+++ Add some (non-sense) example vectors:
#+++ Quality flag vector (e.g. from applying ustar filter)
QF.V.n <- rep(c(1, 0, 1, 0, 1, 0, 0, 0, 0, 0), nrow(EddyDataWithPosix.F) / 10)
#+++ Dummy step function vector to simulate e.g. high / low water table
Step.V.n <- ifelse(EddyDataWithPosix.F$DoY < 200 | EddyDataWithPosix.F$DoY > 250, 0, 1)
#+++ Initialize new sEddyProc processing class with more columns
EddyTest.C <- sEddyProc$new('DE-Tha', cbind(EddyDataWithPosix.F, Step = Step.V.n
, QF = QF.V.n), c('NEE', 'LE', 'H', 'Rg', 'Tair', 'Tsoil', 'rH', 'VPD', 'QF', 'Step'))
#+++ Gap fill variable with (non-default) variables and limits
# including preselection of data with quality flag QF == 0
EddyTest.C$sMDSGapFill('LE', QFVar.s = 'QF', QFValue.n = 0, V1.s = 'Rg'
, T1.n = 30, V2.s = 'Tsoil', T2.n = 2, 'Step', 0.1)
#+++ Use individual gap filling subroutines with different window sizes
# and up to five variables and limits
EddyTest.C$sFillInit('NEE') #Initialize 'NEE' as variable to fill
Result_Step1.F <- EddyTest.C$sFillLUT(3, 'Rg', 50, 'rH', 30, 'Tair', 2.5
, 'Tsoil', 2, 'Step', 0.5)
Result_Step2.F <- EddyTest.C$sFillLUT(6, 'Tair', 2.5, 'VPD', 3, 'Step', 0.5)
Result_Step3.F <- EddyTest.C$sFillMDC(3)
#Individual fill result columns are called 'VAR_...'
EddyTest.C$sPlotHHFluxesY('VAR_fall', Year.i = 1998)

Explicit demonstration of MDS algorithm for NEE gap filling
EddyTestMDS.C <- sEddyProc$new('DE-Tha', EddyDataWithPosix.F
, c('NEE', 'Rg', 'Tair', 'VPD'))
#Initialize 'NEE' as variable to fill
EddyTestMDS.C$sFillInit('NEE')
# Set variables and tolerance intervals
V1.s = 'Rg'; T1.n = 50 # Global radiation 'Rg' within +/-50 W m-2
V2.s = 'VPD'; T2.n = 5 # Vapour pressure deficit 'VPD' within 5 hPa
V3.s = 'Tair'; T3.n = 2.5 # Air temperature 'Tair' within +/-2.5 degC
# Step 1: Look-up table with window size +/-7 days
Result_Step1.F <- EddyTestMDS.C$sFillLUT(7, V1.s, T1.n, V2.s, T2.n, V3.s, T3.n)
# Step 2: Look-up table with window size +/-14 days
Result_Step2.F <- EddyTestMDS.C$sFillLUT(14, V1.s, T1.n, V2.s, T2.n, V3.s, T3.n)
# Step 3: Look-up table with window size +/-7 days, Rg only
Result_Step3.F <- EddyTestMDS.C$sFillLUT(7, V1.s, T1.n)
# Step 4: Mean diurnal course with window size 0 (same day)
Result_Step4.F <- EddyTestMDS.C$sFillMDC(0)
# Step 5: Mean diurnal course with window size +/-1, +/-2 days
Result_Step5a.F <- EddyTestMDS.C$sFillMDC(1)
Result_Step5b.F <- EddyTestMDS.C$sFillMDC(2)
# Step 6: Look-up table with window size +/-21, +/-28, ..., +/-70
for (WinDays.i in seq(21, 70, 7) ) Result_Step6.F <-
EddyTestMDS.C$sFillLUT(WinDays.i, V1.s, T1.n, V2.s, T2.n, V3.s, T3.n)
# Step 7: Look-up table with window size +/-14, +/-21, ..., +/-70, Rg only
for (WinDays.i in seq(14, 70, 7) ) Result_Step7.F <-
EddyTestMDS.C$sFillLUT(WinDays.i, V1.s, T1.n)
# Step 8: Mean diurnal course with window size +/-7, +/-14, ..., +/-210 days
for (WinDays.i in seq(7, 210, 7) ) Result_Step8.F <- EddyTestMDS.C$sFillMDC(WinDays.i)
# Export results, columns are named 'VAR_'
FilledEddyData.F <- EddyTestMDS.C$sExportResults()
head(FilledEddyData.F)
## VAR_orig VAR_f VAR_fqc VAR_fall VAR_fall_qc VAR_fnum VAR_fsd
## 1 -1.21 -1.210000 0 1.373704 1 54 2.531604
## 2 1.72 1.720000 0 1.396364 1 55 2.513678
## 3 NA 1.006569 1 1.006569 1 58 3.311918
## 4 NA 1.061343 1 1.061343 1 67 3.164355
## 5 2.55 2.550000 0 1.071176 1 68 3.141698
## 6 NA 1.205441 1 1.205441 1 68 2.884610
## VAR_fmeth VAR_fwin
## 1 1 14
## 2 1 14
## 3 1 14
## 4 1 14
## 5 1 14
## 6 1 14