-
Notifications
You must be signed in to change notification settings - Fork 0
/
timesSeries_DBEST.R
1183 lines (1014 loc) · 49.4 KB
/
timesSeries_DBEST.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
setwd("C:/R_workspace")
library(DBEST)
library(zoo)
library(lubridate)
library(tidyverse)
library(RColorBrewer)
dataDate1 = "2022_06_30" # using for the latest version of the time series csv
dataDate2 = "2022_08_12" # using for the latest version of the stand attribute dataset
dataDate3 = "2022_08_04" # using for the latest version of the climate variables csv
# Bringing in vegetation index time series
inputDF = read.csv(paste("NBR_timeSeriesDF2000_",dataDate1,".csv",sep = ""),header = TRUE)
TS_DF = inputDF[order(inputDF$X),]
EVI2_DF = read.csv(paste("EVI2_timeSeriesDF2000_",dataDate1,".csv",sep = ""),header = TRUE)
EVI2_DF = EVI2_DF[order(EVI2_DF$X),]
B5_DF = read.csv(paste("B5_timeSeriesDF2000_",dataDate1,".csv",sep = ""),header = TRUE)
B5_DF = B5_DF[order(B5_DF$X),]
TCC_DF = read.csv("AllCleanedData_ONLYtcc.csv",header = TRUE)
TCC_DF = TCC_DF[order(TCC_DF$UniqueID),]
# sznLength_DF = read.csv("seasonLength_timeSeriesDF_2022_07_14.csv",header = TRUE)
# sznLength_DF = sznLength_DF[order(sznLength_DF$X),]
#
# peakDay_DF = read.csv("PeakDay_timeSeriesDF_2022_07_14.csv",header = TRUE)
# peakDay_DF = peakDay_DF[order(peakDay_DF$X),]
#
# midSznLength_DF = read.csv("midSeasonLength_timeSeriesDF_2022_07_14.csv",header = TRUE)
# midSznLength_DF = midSznLength_DF[order(midSznLength_DF$X),]
#
# EVI2area_DF = read.csv("EVI2_Area_timeSeriesDF_2022_07_14.csv",header = TRUE)
# EVI2area_DF = EVI2area_DF[order(EVI2area_DF$X),]
numSamples = length(inputDF$X)
reformatClimateData = function(unformattedDataframe,Nsamples){
newDF <- data.frame(matrix(nrow = Nsamples, ncol = 39,byrow = TRUE))
colnames(newDF) = c('UniqueID',as.character(seq(1984,2021)))
rownames(newDF) = unique(unformattedDataframe$'UniqueID')
lastIncluded = 0
for (multiplier in seq(1,38)){
newDF[,(multiplier+1)] = unformattedDataframe$'mean'[(lastIncluded+1):(Nsamples*multiplier)]
lastIncluded = Nsamples*multiplier
}
newDF$UniqueID = unique(unformattedDataframe$'UniqueID')
return(newDF)
}
freezeDays_DF = read.csv(paste("gridmetMetrics/freeze_TS_",dataDate3,".csv", sep = ''),header = TRUE)
freezeDays_DF = freezeDays_DF[,4:5]
freezeDays_DF = reformatClimateData(freezeDays_DF,numSamples)
GDD_DF = read.csv(paste("gridmetMetrics/GDD_TS_",dataDate3,".csv", sep = ''),header = TRUE)
GDD_DF = GDD_DF[,4:5]
GDD_DF = reformatClimateData(GDD_DF,numSamples)
meanVPD_DF = read.csv(paste("gridmetMetrics/meanVPD_TS_",dataDate3,".csv", sep = ''),header = TRUE)
meanVPD_DF = meanVPD_DF[,4:5]
meanVPD_DF = reformatClimateData(meanVPD_DF,numSamples)
totalPrecip_DF = read.csv(paste("gridmetMetrics/totalPrecip_TS_",dataDate3,".csv", sep = ''),header = TRUE)
totalPrecip_DF = totalPrecip_DF[,4:5]
totalPrecip_DF = reformatClimateData(totalPrecip_DF,numSamples)
meanTemp_DF = read.csv(paste("gridmetMetrics/meanTemp_TS_",dataDate3,".csv", sep = ''),header = TRUE)
meanTemp_DF = reformatClimateData(meanTemp_DF,numSamples)
# Bringing in topographic, soil, locational, and ecotonal data
ecoRegionInfo = read.csv(paste("standAttribute_values_",dataDate2,".csv",sep = ""),header = TRUE)
## creating the dates for the time series index
compositeMonth = 2 # first month used for annual composite
startYear = colnames(TS_DF)[2]
startYear = substr(startYear, 2,5)
startDate = paste(startYear,'-',compositeMonth,'-01',sep = '')
startDate = as.Date.character(startDate,format = '%Y-%m-%d')
endYear = colnames(TS_DF)[length(TS_DF)]
endYear = substr(endYear, 2,5)
endDate = paste(endYear,'-',compositeMonth,'-01',sep = '')
endDate = as.Date.character(endDate,format = '%Y-%m-%d')
dates = seq(startDate,endDate, by = 'years')
# Dataframe to store recovery metrics for each stand
rMetrics <- data.frame(matrix(nrow = length(rownames(TS_DF)), ncol = 14,byrow = TRUE))
colnames(rMetrics) = c('YearOfCut','Recovery_min','Recovery_max','Recovery_mag',
'Disturbance_mag','Y2R_seg','Recovery_slope','NBRy3','NBRy5',
'NBRy7','NBRy10','pre_CutMean','post_CutMean','Y2R_80'
)
rownames(rMetrics) = TS_DF$X
# Dataframe to store auxiliary data for plotting (start dates for segments, etc.)
auxSegData <- data.frame(matrix(nrow = length(rownames(TS_DF)), ncol = 6,byrow = TRUE))
colnames(auxSegData) = c('DistStartIndex','DistEndIndex','RecoveryStartIndex','RecoveryEndIndex','gainLossGap','fitRMSE')
rownames(auxSegData) = TS_DF$X
## A function to calculate recovery metrics, will be looped over the output
# objects from DBEST, outputs to be passed to plotting function
# also outputs index values for the segments I'm selecting to an auxiliary dataframe
calcRecoveryMetrics = function(DBEST_object, index) {
changeSeg_VIdif = DBEST_object$Change
# Need to find the greatest magnitude decline in the time series first
distIndex = which.min(changeSeg_VIdif) # index value for breakpoint arrays for disturbance
distSegEndDate = dates[DBEST_object$End[distIndex]] # date object of end of disturbance segment
distSegStartDate = dates[DBEST_object$Start[distIndex]] # date object of start of disturbance segment
distStartIndex = DBEST_object$Start[distIndex]
distEndIndex = DBEST_object$End[distIndex]
# index value for breakpoint arrays for recovery, should be an increasing segment directly after disturbance
gainSegIndices = which(changeSeg_VIdif > 0)
gainSegStartDates = dates[DBEST_object$Start[gainSegIndices]]
## need to find the gain segment directly after the disturbance
# making a list of the possible dates for possible gain segments
potentialGainSegStartDates = gainSegStartDates[gainSegStartDates >= distSegEndDate]
# adding the disturbance end date to the list
potentialGainSegStartDates = sort(potentialGainSegStartDates,decreasing = FALSE)
gainIndex = gainSegIndices[which(gainSegStartDates == potentialGainSegStartDates[1])]
gainSegStartDate = dates[DBEST_object$Start[gainIndex]]
gainSegEndDate = dates[DBEST_object$End[gainIndex]]
gainSegEndIndex = DBEST_object$End[gainIndex]
gainSegStartIndex = DBEST_object$Start[gainIndex]
## Adding a statement to remove data from bad time series
# i.e time series with no gain, no loss, or only one segment
if ((length(changeSeg_VIdif) <= 1) |
(length(changeSeg_VIdif[changeSeg_VIdif < 0]) == 0) |
(length(changeSeg_VIdif[changeSeg_VIdif > 0]) == 0) |
(length(gainIndex) == 0)
) {
rMetricsArray = rep(NA,14)
auxSegArray = rep(NA,6)
} else {
# Calculating recovery metrics and segmentation outputs
ts_len = length(DBEST_object$Data)
DistStartY = as.double(substr(as.character.Date(distSegStartDate), start = 0, stop = 4))
recovMin = as.double(DBEST_object$Data[distEndIndex]) # Recovery min
recovMax = as.double(DBEST_object$Data[gainSegEndIndex]) # Recovery max
recoveryMag = as.double(recovMax - recovMin) # Recovery magnitude
# Using Max() and Min() in these calculations to enable the 5 year calculation period without going out of range of the dataset
preDistMean = as.double(mean(DBEST_object$Data[max(1,(distStartIndex-5)):distStartIndex])) # Pre-disturbance mean VI value
PostRecovMean = as.double(quantile(DBEST_object$Data[gainSegEndIndex:ts_len],c(0.85))) # Post-recovery mean VI value
distMag = as.double(preDistMean - DBEST_object$Data[distEndIndex]) # Disturbance magnitude
Y2R = as.double(gainSegEndIndex - distEndIndex) # Years to recovery
Y2R80 = min(which(output$Data[distEndIndex:ts_len] >= (0.8*preDistMean))) - 1
Y2R80 = min(Y2R80,(ts_len-distEndIndex))
recovSlope = as.double(recoveryMag / Y2R) # Recovery slope
NBR_year3 = as.double(DBEST_object$Data[(min((distEndIndex+3),ts_len))]) # absolute NBR value at 3,5,7,10 years post disturbance
NBR_year5 = as.double(DBEST_object$Data[(min((distEndIndex+5),ts_len))])
NBR_year7 = as.double(DBEST_object$Data[(min((distEndIndex+7),ts_len))])
NBR_year10 = as.double(DBEST_object$Data[(min((distEndIndex+10),ts_len))])
gapInGainLoss = as.double(DBEST_object$Start[gainIndex] - DBEST_object$End[distIndex]) # Gap between the end of the disturbance segment and the start of the gain segment
RMSE = sqrt(mean((DBEST_object$Fit - DBEST_object$Data)^2))
## Adding recovery metrics to an array to be returned from the function
rMetricsArray = c(DistStartY,recovMin,recovMax,recoveryMag,distMag,Y2R,recovSlope,NBR_year3,NBR_year5,NBR_year7,NBR_year10)
## Adding auxiliary data to dataframe
auxSegArray = c(DBEST_object$Start[distIndex],DBEST_object$End[distIndex],DBEST_object$Start[gainIndex],gainSegEndIndex,gapInGainLoss,RMSE)
rMetricsArray = append(rMetricsArray,c(preDistMean,PostRecovMean,Y2R80)) # These have to be calculated seperately, since their indexing data
} # if a time series doesn't have a gain or loss it'll crash without this
return(list(rMetricsArray,auxSegArray))
}
# Storing the max and min values of the input vegetation index / band, to be used to scale DBEST
VImax = max(TS_DF[2:length(TS_DF[1,])],na.rm = TRUE)
VImin = min(TS_DF[2:length(TS_DF[1,])],na.rm = TRUE)
span = VImax - VImin
## Function to plot the DBEST output correctly
plot_DBEST = function(DBEST_object,index,metricDF,auxDF,minVIval) {
# getting output device set up
svg(filename = paste('C:/R_workspace/images/ClearCutPlotsOnly/plot_',index,'_TimeSeriesGraph.svg',sep = ''),
width = 4,
height = 2,
pointsize = 4,
onefile = FALSE
)
stableColor = "#3B999B"
# plotting the generalization lines furthest back
plot.default(dates,as.vector(DBEST_object$Fit),
type = 'l',
xlab = 'YEAR',
ylab = 'NBR',
main = paste("DBEST Result : Stand #",index,sep = ""),
lwd = 3,
col = stableColor,
ylim = c(VImin,VImax),
axes = FALSE,
)
# getting axes customized
box()
axis.Date(1,at = seq(from = (startDate+365), to = endDate, by = '5 years'),labels = TRUE)
axis(2)
distColor = '#EE5873'
# adding lines for the disturbance segments
for (i in c(1,3)) {
segments(x0 = dates[auxDF[index+1,i]],
y0 = DBEST_object$Fit[auxDF[index+1,i]],
x1 = dates[auxDF[index+1,i+1]],
y1 = DBEST_object$Fit[auxDF[index+1,i+1]],
col = distColor,
lty = 1,
lwd = 3
)
}
# adding the original data last
# raw time-series from original dataframe
rawValues = as.numeric(as.vector(TS_DF[(index+1),2:length(TS_DF)]))
# adding to plot
lines(dates,as.vector(rawValues),
col = 'black',
lwd = 1,
lty = 2
)
# adding bar to indicate disturbance duration
segments(x0 = dates[auxDF[index+1,2]],
y0 = minVIval,
x1 = dates[auxDF[index+1,4]],
y1 = minVIval,
col = 'red',
lty = 1,
lwd = 4.5
)
# adding text to the graph for the disturbance duration
text(x = dates[round((mean(c(auxDF[index+1,2],auxDF[index+1,4]))),digits = 0)],
y = (minVIval + abs(minVIval*0.1)),
labels = paste(as.character(metricDF$Y2R_seg[index+1]),'Year Recovery'),
adj = 0.5
)
#
# # adding lines to indicate pre-disturbance and post-recovery means
# abline(h = metricDF$pre_CutMean[(index+1)],
# col = 'azure3',
# lty = 3,
# lwd = 1
# )
# abline(h = metricDF$post_CutMean[(index+1)],
# col = 'azure4',
# lty = 3,
# lwd = 1
# )
#
# # adding line to show NBRy5 value
# segments(x0 = dates[(auxDF[[index+1,2]]+5)],
# y0 = minVIval,
# x1 = dates[(auxDF[[index+1,2]]+5)],
# y1 = DBEST_object$Data[(auxDF[[index+1,2]]+5)]
# )
#
# recovery min / max indicators on axis
points(x = c(as.Date(as.yearmon(dates[1]) -0.95, frac = 1),
as.Date(as.yearmon(dates[1]) -0.99, frac = 1)
),
y = c(metricDF[index+1,2],
metricDF[index+1,3]
),
col = c('orange','green'),
pch = '-',
cex = 4
)
# adding a legend
legend(x = dates[(length(dates)-8)],
y = -0.1,
legend = c('Original Time-series','DBEST Generalization','Disturbance Event','Recovery Min','Recovery Max'),
col = c('black',stableColor,distColor,'orange','green'),
lty = c(2,1,1,NA,NA),
pch = c(NA,NA,NA,'-','-'),
pt.cex = c(NA,NA,NA,4,4),
lwd = c(1,2,2,NA,NA),
bg = 'light grey'
)
# closing the graphical device
dev.off()
}
# a function that will make it easier to look up stands in google earth
GE_coords = function(standID) {
attributeEntry = ecoRegionInfo[which(ecoRegionInfo$UniqueID == standID),]
cat("Stand #",standID," : ",attributeEntry$lat,", ",attributeEntry$long, sep = "")
}
# implimenting DBEST, plotting results, and storing recovery metrics
for (i in TS_DF$X) {
## creating the time series object
NBRvals = as.numeric(as.vector(TS_DF[(i+1),2:length(TS_DF)]))
ts1 = zoo(x = NBRvals,order.by = dates)
ts1 = na.locf(ts1, na.rm = FALSE, fromLast = TRUE) # filling in end values
ts1 = na.locf(ts1, na.rm = FALSE, fromLast = FALSE)
generalized = DBEST(data = ts1,
data.type = 'non-cyclical',
algorithm = "generalization",
breakpoints.no = 4,
first.level.shift = (span*0.15),
second.level.shift = (span*0.20),
duration = 5,
distance.threshold = 'default',
alpha = 0.05,
plot = 'off'
)
output = DBEST(data = generalized$Fit,
data.type = 'non-cyclical',
algorithm = "change detection",
breakpoints.no = 2,
first.level.shift = (span*0.30),
second.level.shift = (span*0.35),
duration = 5,
distance.threshold = (span*0.10),
alpha = 0.05,
plot = 'off'
)
# ### REMOVE WHEN DONE TESTING ###
# plot.default(ts1,type = 'l', col = 'black', ylim = c(-0.2,0.8))
# lines(generalized$Fit,type = 'l', col = 'blue')
# lines(output$Fit,type = 'l', col = 'red')
# Calculate recovery metrics and auxilary data
listOfArrays = calcRecoveryMetrics(output,i)
rMetrics[(i+1),] = listOfArrays[[1]]
auxSegData[(i+1),] = listOfArrays[[2]]
# Plotting
plot_DBEST(output,i,rMetrics,auxSegData,VImin) ##### rerun when needed #####
cat("Stand #",i, "complete","\n")
}
#### Post Segmentation evaluation of the time series data / metrics ####
# need to add the identifier as a column for tracking things during filtering
rMetrics$'UniqueID' = row.names(rMetrics)
rMetrics$'K_shift' = rMetrics$post_CutMean - rMetrics$pre_CutMean
auxSegData$'UniqueID' = row.names(rMetrics)
auxSegData$'DistDuration' = auxSegData$DistEndIndex - auxSegData$DistStartIndex
# Experimenting with other recovery
rMetrics$'recovPcent3y' = (rMetrics$NBRy3 - rMetrics$Recovery_min) / (abs(rMetrics$Disturbance_mag)) * 100
rMetrics$'recovPcent5y' = (rMetrics$NBRy5 - rMetrics$Recovery_min) / (abs(rMetrics$Disturbance_mag)) * 100
rMetrics$'recovPcent7y' = (rMetrics$NBRy7 - rMetrics$Recovery_min) / (abs(rMetrics$Disturbance_mag)) * 100
rMetrics$'recovPcent10y' = (rMetrics$NBRy10 - rMetrics$Recovery_min) / (abs(rMetrics$Disturbance_mag)) * 100
# A function that finds index values for 3, 5, 7, 10 years post disturbance for a single stand
returnSyncVIs = function(standID,VItext,ts_length,sumMetric = FALSE) {
VI_values = get(paste(VItext,'_DF',sep = ''))[which(get(paste(VItext,'_DF',sep = ''))[,1] == standID),2:length(get(paste(VItext,'_DF',sep = ''))[1,])]
segmentationInfo = auxSegData[which(auxSegData$UniqueID == standID),]
if (sumMetric == FALSE){
if (length(VI_values) == 38) {
timeSeries = zoo(x = as.numeric(VI_values),order.by = dates)
timeSeries = na.locf(timeSeries, na.rm = FALSE, fromLast = TRUE) # filling in end values
timeSeries = na.locf(timeSeries, na.rm = FALSE, fromLast = FALSE)
VI_year3 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+3),ts_length))])
VI_year5 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+5),ts_length))])
VI_year7 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+7),ts_length))])
VI_year10 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+10),ts_length))])
} else if (length(VI_values) < 38){
startY = substr(colnames(VI_values)[1],start = 2, stop = 5)
endY = substr(colnames(VI_values)[length(colnames(VI_values))],start = 2, stop = 5)
holder = as.vector(rep(NA,ts_length))
timeSeries = zoo(x = holder,order.by = dates)
startIndex = match(startY,year(timeSeries))
endIndex = match(endY,year(timeSeries))
timeSeries[startIndex:endIndex] = as.numeric(VI_values)
VI_year3 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+3),ts_length))])
VI_year5 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+5),ts_length))])
VI_year7 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+7),ts_length))])
VI_year10 = as.numeric(timeSeries[(min((segmentationInfo$DistEndIndex+10),ts_length))])
} else {
cat('Please use time series less than or equal to 38 years')
}
} else if (is.na(segmentationInfo$fitRMSE) == FALSE) {
startY = substr(colnames(VI_values)[1],start = 1, stop = 5)
endY = substr(colnames(VI_values)[length(colnames(VI_values))],start = 1, stop = 5)
holder = as.vector(rep(NA,ts_length))
timeSeries = zoo(x = holder,order.by = dates)
startIndex = match(startY,year(timeSeries))
endIndex = match(endY,year(timeSeries))
timeSeries[startIndex:endIndex] = as.numeric(VI_values)
VI_year3 = as.numeric(sum(timeSeries[segmentationInfo$DistEndIndex:(min((segmentationInfo$DistEndIndex+3),ts_length))],na.rm = TRUE))
VI_year5 = as.numeric(sum(timeSeries[segmentationInfo$DistEndIndex:(min((segmentationInfo$DistEndIndex+5),ts_length))],na.rm = TRUE))
VI_year7 = as.numeric(sum(timeSeries[segmentationInfo$DistEndIndex:(min((segmentationInfo$DistEndIndex+7),ts_length))],na.rm = TRUE))
VI_year10 = as.numeric(sum(timeSeries[segmentationInfo$DistEndIndex:(min((segmentationInfo$DistEndIndex+10),ts_length))],na.rm = TRUE))
} else {
VI_year3 = NA
VI_year5 = NA
VI_year7 = NA
VI_year10 = NA
}
return(c(VI_year3,VI_year5,VI_year7,VI_year10))
}
rMetrics$'EVI2y3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'EVI2y5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'EVI2y7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'EVI2y10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'B5y3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'B5y5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'B5y7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'B5y10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'TCCy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'TCCy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'TCCy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'TCCy10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'freezeDaysy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'freezeDaysy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'freezeDaysy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'freezeDaysy10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'GDDy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'GDDy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'GDDy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'GDDy10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanVPDy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanVPDy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanVPDy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanVPDy10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'totalPrecipy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'totalPrecipy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'totalPrecipy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'totalPrecipy10' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanTempy3' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanTempy5' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanTempy7' = rep(NA,length(rMetrics$UniqueID))
rMetrics$'meanTempy10' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'sznLengthy3' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'sznLengthy5' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'sznLengthy7' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'sznLengthy10' = rep(NA,length(rMetrics$UniqueID))
#
# rMetrics$'midSznLengthy3' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'midSznLengthy5' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'midSznLengthy7' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'midSznLengthy10' = rep(NA,length(rMetrics$UniqueID))
#
# rMetrics$'peakDayy3' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'peakDayy5' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'peakDayy7' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'peakDayy10' = rep(NA,length(rMetrics$UniqueID))
#
# rMetrics$'EVI2areay3' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'EVI2areay5' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'EVI2areay7' = rep(NA,length(rMetrics$UniqueID))
# rMetrics$'EVI2areay10' = rep(NA,length(rMetrics$UniqueID))
for (standNum in rMetrics$UniqueID) {
for (VI in c('EVI2','B5','TCC')) {
output = returnSyncVIs(standNum, VI, 38)
startCol = match(paste(VI,'y3',sep = ''),colnames(rMetrics))
endCol = startCol + 3
standIndex = which(rMetrics$UniqueID == standNum)
rMetrics[standIndex,startCol:endCol] = output
}
}
for (standNum in rMetrics$UniqueID) {
for (VI in c('freezeDays','GDD','meanVPD','totalPrecip','meanTemp')) {
output = returnSyncVIs(standNum, VI, 38, TRUE)
startCol = match(paste(VI,'y3',sep = ''),colnames(rMetrics))
endCol = startCol + 3
standIndex = which(rMetrics$UniqueID == standNum)
rMetrics[standIndex,startCol:endCol] = output
}
}
### Adding in the temporal change in these metrics ###
rMetrics$'freezeDaysChange' = apply(freezeDays_DF[34:39], MARGIN = 1,FUN = mean) - apply(freezeDays_DF[2:7], MARGIN = 1,FUN = mean)
rMetrics$'GDDChange' = apply(GDD_DF[34:39], MARGIN = 1,FUN = mean) - apply(GDD_DF[2:7], MARGIN = 1,FUN = mean)
rMetrics$'meanVPDChange' = apply(meanVPD_DF[34:39], MARGIN = 1,FUN = mean) - apply(meanVPD_DF[2:7], MARGIN = 1,FUN = mean)
rMetrics$'totalPrecipChange' = apply(totalPrecip_DF[34:39], MARGIN = 1,FUN = mean) - apply(totalPrecip_DF[2:7], MARGIN = 1,FUN = mean)
rMetrics$'meanTempChange' = apply(meanTemp_DF[34:39], MARGIN = 1,FUN = mean) - apply(meanTemp_DF[2:7], MARGIN = 1,FUN = mean)
#plottingDF$'meanTempChange'
rMetrics$'NonPSAy3' = atan2(rMetrics$EVI2y3,rMetrics$B5y3)
rMetrics$'NonPSAy5' = atan2(rMetrics$EVI2y5,rMetrics$B5y5)
rMetrics$'NonPSAy7' = atan2(rMetrics$EVI2y7,rMetrics$B5y7)
rMetrics$'NonPSAy10' = atan2(rMetrics$EVI2y10,rMetrics$B5y10)
# export table I can add back into arcgis pro
exportDF = rMetrics[,c(14,32:length(colnames(rMetrics)))]
write.csv(exportDF,file = "additionalVIvals2.csv")
write.csv(rMetrics,file = "RecoveryMetrics_output.csv")
#### Cleaning data based on practical relationships ####
# Number of TS which did not have qualifying segments for metric calculation
areNAindicies = which(is.na(rMetrics$Y2R_seg) == TRUE)
areNAstands = rMetrics$UniqueID[areNAindicies]
# Number of TS which had large gaps between disturbance and recovery segments
bigGapIndices = which(auxSegData$gainLossGap > 5)
bigGapLengths = auxSegData$gainLossGap[bigGapIndices]
bigGapStands = auxSegData$UniqueID[bigGapIndices]
# Number of TS which had disturbances lasting longer than three years
longDistStands = auxSegData$UniqueID[which(auxSegData$DistDuration > 4)]
# Experimental, there are many stands which because of segmentation or noise have longer disturbances
# seem to be able to keep the good ones by including a disturbance magnitude condition
expFilter = auxSegData$UniqueID[which((auxSegData$DistDuration > 4) & (rMetrics$Disturbance_mag < 0.3))]
# Removing low magnitude disturbance
smalldist = rMetrics$UniqueID[which(rMetrics$Disturbance_mag < 0.20)]
# Going to remove stands that were probably not pine prior to disturbace
hardwoodCNV = rMetrics$UniqueID[which(rMetrics$pre_CutMean < 0.20)]
# Number of TS which had a recovery within 4 years or less
shortRecovStands = rMetrics$UniqueID[which(rMetrics$Y2R_seg < 3)]
# An array which will hold all UniqueID's of stands removed from statistics
badStands = sort(as.numeric(unique(c(areNAstands,bigGapStands,longDistStands,shortRecovStands))),decreasing = FALSE)
badStands2 = sort(as.numeric(unique(c(areNAstands,longDistStands,shortRecovStands,hardwoodCNV,smalldist))),decreasing = FALSE)
# # trying out an outlier test
# findExtremeStands = function(metricValues,IDs,c){
# stats = summary(metricValues)
# anIQR = IQR(metricValues, na.rm = TRUE)
# lower = stats[2] - (c*anIQR)
# upper = stats[5] + (c*anIQR)
# indices = which(metricValues > upper | metricValues < lower)
# return(IDs[indices])
# } # can use either 1 or 3 for c (3 for extreme values)
## Creating a new dataframe with only good values
goodDataDF = rMetrics[which(is.na(rMetrics$YearOfCut) == FALSE),]
goodDataDF$fit_RMSE = auxSegData$fitRMSE[which(is.na(auxSegData$DistStartIndex) == FALSE)]
#
# goodDataDF = goodDataDF[-which(goodDataDF$UniqueID == badStands2),]
for (index in goodDataDF$UniqueID){
ID = as.numeric(index)
IDindex = which(goodDataDF$UniqueID == ID)
bool = ID %in% as.numeric(badStands2)
if (bool == TRUE){
goodDataDF = goodDataDF[-IDindex,]
}
}
#### the above code is a much simplier and more transparent way of selecting the good stands
# for (ID in goodDataDF$UniqueID) {
# gapLength = auxSegData$gainLossGap[which(auxSegData$UniqueID == ID)]
# distDuration = as.numeric(auxSegData$DistDuration[which(auxSegData$UniqueID == ID)])
# recovDuration = goodDataDF$Y2R[goodDataDF$UniqueID == ID]
# if ((distDuration > 3) | (recovDuration < 5)) {
# indexValue = which(goodDataDF$UniqueID == ID)
# goodDataDF = goodDataDF[-indexValue,]
# }
# } # removing stands with large gaps between disturbance end and recovery start
# # Copying only the good files into a seperate directory within all the plot images folder
# allPlotsDirPath = "C:/R_workspace/images/ClearCutPlotsOnly"
# goodDirPath = "C:/R_workspace/images/ClearCutPlotsOnly/CleanedData"
# for (standI in goodDataDF$UniqueID){
# filePath = paste('plot_',standI,'_TimeSeriesGraph.svg',sep = '')
# file.copy(from = paste(allPlotsDirPath,'/',filePath, sep = ''),
# to = goodDirPath,
# overwrite = TRUE,
# recursive = FALSE,
# copy.mode = TRUE,
# copy.date = TRUE
# )
# }
#
# # Copying only the bad stands graphs into a seperate directory within all the plot images folder
# badDirPath = "C:/R_workspace/images/ClearCutPlotsOnly/BadStands"
# for (standI in badStands2){
# filePath = paste('plot_',as.character(standI),'_TimeSeriesGraph.svg',sep = '')
# file.copy(from = paste(allPlotsDirPath,'/',filePath, sep = ''),
# to = badDirPath,
# overwrite = TRUE,
# recursive = FALSE,
# copy.mode = TRUE,
# copy.date = TRUE
# )
# }
# merging the ecoregion info and good data for plotting #
ecoRegionInfo$intLat = floor(ecoRegionInfo$lat)
ecoRegionInfo$ownership_mode = replace_na(ecoRegionInfo$ownership_mode,999)
ecoRegionInfo$ownership_mode = as.factor(ecoRegionInfo$ownership_mode)
plottingDF = merge(x = goodDataDF,
y = ecoRegionInfo,
by.x = 'UniqueID',
by.y = 'UniqueID',
all.x = TRUE
)
SSURGOdata = read.csv(file = 'allStandsSSURGOdata.csv')
plottingDF = merge(x = plottingDF,
y = SSURGOdata,
by.x = 'UniqueID',
by.y = 'UniqueID',
all.x = TRUE
)
SCMCdata = read.csv(file = 'cleanedData_SCMC.csv')
colnames(SCMCdata) = c('UniqueID','groupID','memProb')
plottingDF = merge(x = plottingDF,
y = SCMCdata,
by.x = 'UniqueID',
by.y = 'UniqueID',
all.x = TRUE
)
# Setting some of the categorical data columns to be factors #
plottingDF$aspect_mode = as.factor(plottingDF$aspect_mode)
plottingDF$soil0_mode = as.factor(plottingDF$soil0_mode)
plottingDF$soil30_mode = as.factor(plottingDF$soil30_mode)
plottingDF$soil100_mode = as.factor(plottingDF$soil100_mode)
plottingDF$ecoR_code = as.character(plottingDF$Classified)
plottingDF$plantation_class = as.factor(plottingDF$plantation_class)
plottingDF$soilGG_mode = as.factor(plottingDF$soilGG_mode)
plottingDF$groupID = as.factor(plottingDF$groupID)
plottingDF$'rel_Kshift' = ((plottingDF$post_CutMean - plottingDF$pre_CutMean) / plottingDF$pre_CutMean) * 100 #(plottingDF$post_CutMean / plottingDF$pre_CutMean) * 100 #
write.csv(plottingDF,file = "RecoveryMetrics_output.csv")
#### Experimenting with CART analysis ####
library(rpart)
library(rpart.plot)
library(ranger)
library(lctools)
library(dplyr)
library(corrplot)
# Looking at correlation matrix for all variables
onlyNumeric = unlist(lapply(plottingDF, is.numeric))
correlationMatrix = cor(plottingDF[,onlyNumeric],
use = 'everything'
)
corrplot(correlationMatrix)
# Determining spatial autocorrelation
spatialWeights = w.matrix(Coords = data.frame(plottingDF$long,plottingDF$lat),
Bandwidth = 50,
WType = 'Binary',
family = 'adaptive'
)
morans = moransI.w(plottingDF$K_shift,spatialWeights)
localMorans = l.moransI(Coords = data.frame(plottingDF$long,plottingDF$lat),
Bandwidth = 500,
x = plottingDF$K_shift,
WType = 'Binary',
scatter.plot = TRUE,
family = 'adpative'
)
plot.default(plottingDF$pre_CutMean, plottingDF$rel_Kshift, col = plottingDF$YearOfCut)
testModel = lm(formula = 'rel_Kshift ~ pre_CutMean', data = log(plottingDF[,c('rel_Kshift', 'pre_CutMean')]))
plot.default(plottingDF$rel_Kshift, exp(testModel$fitted.values))
abline(0,1)
# Model formula
formulaString = "recovPcent5y ~ aspect_mode + elevation_mean + slope_mean +
soil0_mode + soil30_mode + soil100_mode + pH30_mean +
ecoR_code + ownership_mode + freezeDaysy3 + freezeDaysy5 +
freezeDaysy7 + freezeDaysy10 + GDDy3 + GDDy5 + GDDy7 +
GDDy10 + meanVPDy3 + meanVPDy5 + meanVPDy7 + meanVPDy10 +
totalPrecipy3 + totalPrecipy5 + totalPrecipy7 +
totalPrecipy10 + meanTempy3 + meanTempy5 + meanTempy7 +
meanTempy10 + freezeDaysChange + GDDChange + meanVPDChange +
totalPrecipChange + meanTempChange + plantation_class +
soilGG_mode + SI_25 + SI_50 + Recovery_min + YearOfCut +
pre_CutMean + lat + long"
# # Seperating train from test data 70% train, 30% test
# trainIndices = sample(x = (1:length(plottingDF$UniqueID)), size = (0.7*length(plottingDF$UniqueID)))
# trainData = plottingDF[trainIndices,]
# testIndices = (1:length(plottingDF$UniqueID))[-trainIndices]
# testData = plottingDF[testIndices,]
# Fitting the regression CART model for K_shift
CART_fit = rpart(formula = formulaString,
data = plottingDF,
method = 'anova',
cp = 0
)
rsq.rpart(CART_fit)
# Finding the lowest cp value within 1SE of the lowest cross validated error
lowERRcpI = as.numeric(which.min(CART_fit$cptable[,4]))
lowERRcp = as.numeric(CART_fit$cptable[lowERRcpI,1])
optimumCP = which(CART_fit$cptable[,4] <= (CART_fit$cptable[lowERRcpI,4] + CART_fit$cptable[lowERRcpI,5]))
optimumCP = as.double(CART_fit$cptable[min(optimumCP),1])
# CART_test = rpart.predict(CART_fit, type = 'matrix')
#
# # Going to look at the highest residuals from this model
# CARTResids = sort(abs(residuals(CART_fit)),decreasing = TRUE)
# residQs = quantile(CARTResids,c(0.80,0.85,0.90,0.95))
# Qresids = CARTResids[which(CARTResids >= residQs[4])]
# residsI = as.numeric(names(Qresids))
# residsStands = plottingDF$UniqueID[residsI]
# # Testing the regression tree
# CART_test = rpart.predict(object = CART_fit,
# type = 'matrix',
# newdata = testData
# )
# # RMSE of the test fit
# CART_rmse = sqrt(mean((CART_test - testData$K_shift)^2))
#
# # Finding the number of splits with the lowest complexity
# lowestCP = CART_fit$cptable[which.min(CART_fit$cptable[,'xerror']),'CP']
# Pruning the tree to the lowest complexity parameter
CART_fit2 = prune.rpart(CART_fit, cp = optimumCP)
paste('Pruned CART R^2',round(1-min(CART_fit2$cptable[,4]),digits = 4))
rpart.plot(CART_fit2)
# CART_test2 = rpart.predict(object = CART_fit2,
# type = 'matrix',
# newdata = testData
# )
# CART2_rmse = sqrt(mean((CART_test2 - testData$K_shift)^2))
noSIstands = which(is.na(plottingDF$SI_25))
### Trying out random forest ###
RF_1 = ranger(formula = formulaString,
data = plottingDF[c(-178,-3448,-2596,-noSIstands),],
num.trees = 2000,
importance = 'impurity',
replace = TRUE,
respect.unordered.factors = TRUE,
oob.error = TRUE,
num.threads = 8,
write.forest = TRUE
)
# Checking out accuracy of Random Forest
paste("Random Forest R2",round(RF_1$r.squared, digits = 3))
plot.default(plottingDF$recovPcent5y [c(-178,-3448,-2596,-noSIstands)],RF_1$predictions) #xlim = c(-50,200),ylim = c(-50,200)
abline(0,1)
## Plotting variable importance c(RF_1$variable.importance,CART_fit$variable.importance)
varIdata = sqrt(sort(CART_fit2$variable.importance, decreasing = TRUE))[0:floor(0.6*length(RF_1$variable.importance))]
varGraph = ggplot(mapping = aes(x = reorder(names(varIdata),as.numeric(varIdata)), y = as.numeric(varIdata))) +
geom_bar(width = 1,
stat="identity",
colour = "black",
fill="lightblue",
position = position_dodge()
)+
theme(aspect.ratio = 2/3)+
coord_polar(theta = "x", start=0)+
labs(x = NULL,y = 'Variable Importance Metric')
print(varGraph)
## Plotting variable importance c(RF_1$variable.importance,CART_fit$variable.importance)
varIdata = sqrt(sort(RF_1$variable.importance, decreasing = TRUE))[0:floor(0.6*length(RF_1$variable.importance))]
varGraph = ggplot(mapping = aes(x = reorder(names(varIdata),as.numeric(varIdata)), y = as.numeric(varIdata),fill = as.numeric(varIdata))) +
geom_bar(width = 1,
stat="identity",
colour = "black",
position = position_dodge()
)+
coord_flip()+
scale_fill_continuous(low="blue", high="red")+
theme(aspect.ratio = 2/3,
text = element_text(size = 16)
)+
guides(fill=guide_legend(title=""))+
labs(x = NULL,y = 'Gini Variable Importance')
print(varGraph)
# Going to try the CART model with binned K_shift rather than a continuous regression
test = ecdf(plottingDF$K_shift)
calcConfusion = function(confusionMatrix){
inputDim = dim(confusionMatrix)
accuracyTable = matrix(ncol = 3,nrow = inputDim[1])
colnames(accuracyTable) = c('Users','producers','overall')
rownames(accuracyTable) = seq(1,inputDim[1])
for (classN in seq(1,inputDim[1])){
accuracyTable[classN,1] = confusionMatrix[classN,classN] / sum(confusionMatrix[classN,])
accuracyTable[classN,2] = confusionMatrix[classN,classN] / sum(confusionMatrix[,classN])
}
total = sum(confusionMatrix[,c(seq(1,inputDim[1]))])
accuracyTable[1,3] = sum(diag(confusionMatrix)) / total
return(accuracyTable)
}
### trying to bin values with a given number of bins (equal number of samples per bin)
rowOrder = data.frame(sort(plottingDF$K_shift,decreasing = FALSE,index.return = TRUE))[,2]
plottingDF$'binKshift'[rowOrder] = ntile(sort(plottingDF$K_shift, decreasing = FALSE), n = 3)
#plottingDF$'binKshift'[c(91,2374,3393)] = 2 # the above code misclassifies 3 values as negative
plottingDF$'binKshift' = as.factor(plottingDF$'binKshift')
### this code can be used to bin values based on quantiles
# KshiftQs = quantile(plottingDF$K_shift,c(0.0,0.1241812,0.345,0.540,0.75,1.0), digits = 3) # this is just for < 0 OR > 0
# binnedKshift = data.frame(findInterval(plottingDF$K_shift,KshiftQs,all.inside = TRUE))
# plottingDF$'binKshift' = as.factor(binnedKshift$findInterval.plottingDF.K_shift..KshiftQs.)
# I'm trying out removing the other soil modes (soil0_mode + soil100_mode)
formulaString2 = "binKshift ~ aspect_mode + elevation_mean + slope_mean +
soil0_mode + soil30_mode + soil100_mode + pH30_mean +
ecoR_code + ownership_mode + freezeDaysy3 + freezeDaysy5 +
freezeDaysy7 + freezeDaysy10 + GDDy3 + GDDy5 + GDDy7 + GDDy10 +
meanVPDy3 + meanVPDy5 + meanVPDy7 + meanVPDy10 +
totalPrecipy3 + totalPrecipy5 + totalPrecipy7 + totalPrecipy10 +
meanTempy3 + meanTempy5 + meanTempy7 + meanTempy10 +
freezeDaysChange + GDDChange + meanVPDChange +
totalPrecipChange + meanTempChange + plantation_class +
soilGG_mode"
CART_fit3 = rpart(formula = formulaString2,
data = plottingDF,
method = 'class',
cp = 0
)
paste('Pruned binned K_shift CART R^2',round(1-min(CART_fit3$cptable[,4]),digits = 4))
CART_fit3 = prune.rpart(CART_fit3,cp = 0.00470925)
CART_confM = table(predict(CART_fit3, type = 'class'),plottingDF$binKshift)
print(CART_confM)
print(calcConfusion(CART_confM))
### Trying out random forest for binned K_shift ###
RF_2 = ranger(formula = formulaString2,
data = plottingDF[c(-178,-3448,-2596,-noSIstands),],
num.trees = 2000,
importance = 'impurity_corrected',
replace = TRUE,
respect.unordered.factors = TRUE,
oob.error = TRUE,
num.threads = 8,
write.forest = TRUE,
classification = TRUE,
verbose = TRUE
)
print(RF_2$confusion.matrix)
print(calcConfusion(RF_2$confusion.matrix))
#### Temporal trend analysis ####
library(Kendall)
yearsMedian = aggregate.data.frame(plottingDF$K_shift,
by = list(plottingDF$YearOfCut),
FUN = median,
simplify = TRUE
)
trendTest = Kendall(yearsMedian$Group.1,yearsMedian$x)
summary(trendTest)
plot.default(yearsMedian$Group.1,yearsMedian$x)
temporalTrend = lm('x ~ Group.1',data = yearsMedian)
summary(temporalTrend)
abline(temporalTrend$coefficients)
#### Looking at regional groups ####
group1 = c(8,6,4,5,3)
group2 = c(7,2,1)
temporalVariable = 'freezeDaysy3'
## Looking at temporal trend for coastal regions
rows1 = plottingDF$groupID %in% group1
G1data = plottingDF[rows1,]
yearsMedianG1 = aggregate.data.frame(G1data[temporalVariable],
by = list(G1data$YearOfCut),
FUN = median,
simplify = TRUE
)
trendTestG1 = Kendall(yearsMedianG1[,1], yearsMedianG1[,2])
summary(trendTestG1)
g1zoo = zoo(x = yearsMedianG1[,2],
order.by = yearsMedianG1$Group.1
)
g1ACF = acf(x = g1zoo,
type = 'correlation',
plot = TRUE
)
print(g1ACF)
plot.default(yearsMedianG1[,1], yearsMedianG1[,2], type = 'l', main = paste('Coastal Stands',temporalVariable,'Temporal Trend'))
temporalTrendG1 = lm(paste(temporalVariable,' ~ Group.1', sep = ''),data = yearsMedianG1)
abline(temporalTrendG1$coefficients, col = 'blue')
## Looking at temporal trend for inland regions
rows2 = plottingDF$groupID %in% group2
G2data = plottingDF[rows2,]
yearsMedianG2 = aggregate.data.frame(G2data[temporalVariable],
by = list(G2data$YearOfCut),
FUN = median,
simplify = TRUE
)
trendTestG2 = Kendall(yearsMedianG2[,1], yearsMedianG2[,2])
summary(trendTestG2)
g2zoo = zoo(x = yearsMedianG2[,2],
order.by = yearsMedianG1$Group.1
)
g2ACF = acf(x = g2zoo,
type = 'correlation',
plot = TRUE
)
print(g2ACF)
plot.default(yearsMedianG2[,1], yearsMedianG2[,2], type = 'l', main = paste('In-land Piedmont Stands',temporalVariable,'Temporal Trend'))
temporalTrendG2 = lm(paste(temporalVariable,' ~ Group.1', sep = ''),data = yearsMedianG2)
abline(temporalTrendG2$coefficients, col = 'blue')
## Plot to look at cluster temporal trends together
combinedTrends = ggplot(mapping = aes_string(x = 'YearOfCut', y = temporalVariable, color = 'groupID'),
data = plottingDF
)+
geom_point(shape = 4)+
stat_smooth(method = 'lm')
print(combinedTrends)
## Looking at TCC values
onlyTCCrows = which(is.na(plottingDF$TCCy10) == FALSE)
onlyTCC_DF = plottingDF[onlyTCCrows,]
onlyNumeric = unlist(lapply(onlyTCC_DF, is.numeric))
correlationMatrix = cor(onlyTCC_DF[,onlyNumeric],
use = 'everything'
)
corrplot(correlationMatrix)
## Boxplots to describe spatial cluster
arcmapSymb = c('#78AAFF','#FF6455','#7DDC55','#FFB400','#C864E1','#BEA064','#FABEC8','#AFAFAF')
clusterAttri = ggplot(data = plottingDF[-104,],
mapping = aes_string(y = 'pre_CutMean', fill = 'groupID')
)+
geom_boxplot(notch = TRUE,
varwidth = TRUE,
coef = 1.0
)+
scale_fill_manual(values = arcmapSymb
)+
theme(axis.text.x=element_blank(), #remove x axis labels
axis.ticks.x=element_blank()
)+
labs(xlab = '', fill = 'Cluster ID')
print(clusterAttri)
## Plot for relationship between pre-disturbance mean and relative K_shift
Krelation = ggplot(data = plottingDF,
mapping = aes_string(y = 'pre_CutMean', x = 'rel_Kshift') # color = 'groupID'
)+
#geom_point(shape = 4, stroke = 0.967)+
geom_hex()+
theme(legend.position = 'right',
aspect.ratio = 2/3,