-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy path5.3_Uncertainty_RandomWalk.R
487 lines (406 loc) · 23.4 KB
/
5.3_Uncertainty_RandomWalk.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
### This script loads in MCMC fit outputs and assesses uncertainties
library(ecoforecastR)
library(rjags)
library(coda)
#setwd("../Aedes2019Forecast/")
source("0_County_subset.R")
source("1.0_Climate_data_import.R")
source("2.1_Run_Fit.R")
source("2.2_Basic_Diagnostics_Function.R")
source("4.1_GLM_Fit.R")
source("5.1_Create_Met_Ensemble.R")
path = "./RandomWalk_Out/"
aedes.data <- subset_aedes_data()
data.fit <- aedes.data$data.training
counties <- aedes.data$subset.counties
clim.dat.monthly <- get_daymet()
######################################
# load in aegypti model fits for each county & counts, store in list
aegypti <- list()
aegypti.files <- list.files(path = path, pattern="^aegypti") # load in files with "aegypti" at the beginning
for(i in aegypti.files){
aegypti[[i]] <- mget(load(file.path("RandomWalk_Out", i))) # get objects in each RData file, store
}
# load in albopictus model fits for each county & counts, store in list
albopictus <- list()
albopictus.files <- list.files(path = path, pattern="^albopictus") # load in files with "aegypti" at the beginning
for(i in albopictus.files){
albopictus[[i]] <- mget(load(file.path("RandomWalk_Out", i))) # get objects in each RData file, store
}
######################################
# plot by site with aggregated counts and confidence intervals
## site (s = 1-11)
## time length of fit (typically 1-end of fit)
## x limit (default x = 1-end of fit)
## NOTE the following MUST be included before calling plot.run() (see lines 59-65 and 152-161)
sp = aegypti #or albopictus
state <- as.matrix(sp[[s]]$predict)
ci.state <- apply(state,2,quantile,c(0.025,0.5,0.975))
time <- 1:ncol(ci.state)
summary(sp)
plot.run <- function(s,time,x=c(1,length(time)), ...){
plot(time, ci.state[2,],main=gsub('_', ' ',counties[s]), xlab = "Year - Month", xaxt = 'n', pch="",xlim=x,...)
ciEnvelope(time,ci.state[1,],ci.state[3,],col="lightblue")
points(time,sp[[s]]$y,pch=16,type="p",col='black')
# axis(1, at = at, labels = year.month[at])
}
par(mfrow=c(2,2))
# aegypti plots
sp <- aegypti
for(s in 1:length(sp)){
state <- as.matrix(sp[[s]]$predict)
ci.state <- apply(state,2,quantile,c(0.025,0.5,0.975))
time <- 1:ncol(ci.state)
plot.run(s,time)
}
# albopictus plots
sp <- albopictus
for(s in 1:length(sp)){
state <- as.matrix(sp[[s]]$predict)
ci.state <- apply(state,2,quantile,c(0.025,0.5,0.975))
time <- 1:ncol(ci.state)
plot.run(s,time)
}
######################################
##` @param IC Initial Conditions
##` @param mu counts
##` @param beta beta coefficients on met data
##` @param ens ensemble of drivers
##` @param Q Process error (default = 0 for deterministic runs)
##` @param n Size of Monte Carlo ensemble
##` @param NT Length of forecast (months)
forecast <- function(IC,tau=0,n,NT){
N <- matrix(NA,n,NT) # store results
Nprev <- IC # initialize with initial conditions
for(t in 1:NT){
mu <- Nprev
negative.check <- rnorm(n,mu,tau) # get new prediction
N[,t] <- (abs(negative.check) + negative.check) / 2 # check for negative predictions
Nprev <- N[,t]
}
return(N)
}
###################################### GET MET DATA & DEFINE INPUTS FUNCTION
met <- create_met_ensemble(n.ens = 100) # get driver ensemble members
# calculate mean of all inputs
# get the input data for the forecast function based on
## species (sp = aegypti or albopictus)
## site (s = 1-11)
# months into the future (NT = 1-12)
calc.inputs <- function(sp,s,NT){
met.ens <- met[[s]]
# only work with 1:NT months
for(i in 1:length(met.ens)){
met.ens[[i]] <- met.ens[[i]][seq(1:NT),]
}
# # subset met ensemble to variables used in historical fit
met.ensemble <- list(prcp = met.ens$prcp,
tmax = met.ens$tmax,
RH = met.ens$RH)
# rows = months, cols = met variables
met.means <- matrix(NA,NT,length(met.ensemble))
for(i in 1:length(met.ensemble)){
met.means[,i] <- apply(met.ensemble[[i]],1,mean) # drivers
}
colnames(met.means) <- c("prcp", "tmax", "RH")
met.prcp <- matrix(met.ens$prcp, ncol=1, byrow=FALSE)
met.tmax <- matrix(met.ens$tmax, ncol=1, byrow=FALSE)
met.RH <- matrix(met.ens$RH, ncol=1, byrow=FALSE)
parameters<- as.matrix(sp[[s]]$params)#, ncol=1, byrow=FALSE) # 5 cols, 20001 rows
met.values<- cbind(rep(1, 2001), rep(met.prcp, 2001),
rep(met.tmax, 2001), rep(met.RH, 2001)) # 4 cols, 60 x 2001 rows
met.values<- met.values[1:100004,] #length of vars$parameters aka the length of beta and nrow in ens to be the same
met.means <- cbind(rep(1, NT), met.means) # 4 cols, 6 rows
params <- as.matrix(sp[[s]]$params)
param.mean <- apply(params,2,mean) # betas & tau
IC <- as.matrix(sp[[s]]$out$predict) # initial conditions
# inputs <- list(met.means,param.mean,IC, parameters, met.values)
inputs <- list(met.means,param.mean,IC, params, met.ensemble)
names(inputs) <- c("met.means","param.mean","IC", "parameters", "met.values")
return(inputs)
}
##### DETERMINISTIC FORECAST FOR AEGYPTI, FLORIDA Osceola, 6 MONTHS #####
s <- 8
NT <- 12
vars <- calc.inputs(aegypti,s,NT)
N.det <- forecast(IC=mean(vars$IC[,ncol(vars$IC)]), # mean of initial conditions @ last time step of fit
# beta=t(as.matrix(vars$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(vars$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(vars$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(vars$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=1, # 1 run to generate deterministic prediction
NT=NT) # forecast NT months into the future
par(mfrow=c(1,1))
# define inputs necessary
sp <- aegypti # species
state <- as.matrix(sp[[s]]$predict) # get predictions
ci.state <- apply(state,2,quantile,c(0.025,0.5,0.975)) # confidence intervals for predictions
time <- 1:ncol(ci.state) # time length of fit
time2 <- seq(ncol(ci.state)+1,ncol(ci.state)+NT) # length of forecast (beginning @ next time step after length of fit ends)
time3 = seq(1:time2[length(time2)])
# plot
x.start <- ncol(ci.state)-15
plot.run(s,time, x=c(x.start,ncol(ci.state)+NT)) # plot for Florida Lee for length of fit, set xlim to show deterministic forecast
lines(time2,N.det,col="red",lwd=2,type="l") # plot deterministic forecast
######################################
### Initial Condition Uncertainty
Nmc <- 500 # number of samples to run
draw <- sample.int(nrow(vars$parameters), Nmc, replace = TRUE)
N.I <- forecast(IC=vars$IC[draw,ncol(vars$IC)], # sample initial conditions @ last time step of fit
# beta=t(as.matrix(vars$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(vars$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(vars$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(vars$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=Nmc, # Nmc samples for propogation
NT=NT) # forecast 6 months into the future
N.I.ci = apply(N.I,2,quantile, c(0.025, 0.5, 0.975))
plot.run(s,time,x=c(x.start,ncol(ci.state)+NT)) # plot for Florida Lee for length of fit
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="red")
lines(time2, N.I.ci[2,], lwd=2, type="b")
######### Parameter Uncertainty-- changing beta from mean parametervalues to samples
N.IP <- forecast(IC=vars$IC[draw,ncol(vars$IC)], # sample initial conditions @ last time step of fit
# beta=vars$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(vars$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(vars$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(vars$met.means[,"RH"]), # matrix of mean rh from 5.1
# tau=0, # process error off (note tau must be a SD, not a precision)
n=Nmc, # Nmc runs
NT=NT) # forecast NT months into the future
N.IP.ci = apply(N.IP,2,quantile, c(0.025, 0.5, 0.975))
plot.run(s,time,x=c(x.start,ncol(ci.state)+NT)) # plot for Florida Lee for length of fit
ecoforecastR::ciEnvelope(time2,N.IP.ci[1,], N.IP.ci[3,], col="lightblue")
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="red")
lines(time2, N.IP.ci[2,], lwd=2, type="b")
######### Driver Uncertainty-- changing ens from mean met values to met samples
# random sample of met ensemble
draw.met <- sample.int(ncol(vars$met.values$prcp), Nmc, replace = TRUE)
N.IPD <- forecast(IC=vars$IC[draw,ncol(vars$IC)], # sample initial conditions @ last time step of fit
# beta=vars$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(vars$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(vars$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(vars$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=Nmc,
NT=NT) # forecast 6 months into the future
N.IPD.ci = apply(N.IPD,2,quantile, c(0.025, 0.5, 0.975))
plot.run(s,time,x=c(x.start,ncol(ci.state)+NT))# plot for Florida Lee for length of fit
ecoforecastR::ciEnvelope(time2,N.IPD.ci[1,], N.IPD.ci[3,], col="lightsalmon")
ecoforecastR::ciEnvelope(time2,N.IP.ci[1,], N.IP.ci[3,], col="indianred2")
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="violetred4")
lines(time2, N.IP.ci[2,], lwd=2, type="l")
legend('topleft',
legend = c('Forecast','Inital Conditions','Parameter','Driver','Process'),
col = c('black', 'violetred4', 'indianred2','lightsalmon','peachpuff1'),
lwd = c(2,4,4,4,4),
cex = 0.75,
inset = 0.02)
######### Process uncertainty tau = 1/sqrt(vars$param.mean[5])
N.IPDE <- forecast(IC=vars$IC[draw,ncol(vars$IC)], # sample initial conditions @ last time step of fit
# beta=vars$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(vars$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(vars$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(vars$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=1/sqrt(vars$parameters[draw,5]), # process error off (note tau must be a SD, not a precision)
n=Nmc,
NT=NT) # forecast 6 months into the future
N.IPDE.ci = apply(N.IPDE, 2, quantile, c(0.025, 0.5, 0.975))
plot.run(s,time,x=c(x.start,ncol(ci.state)+NT),ylim=c(0,500))# plot for Florida Lee for length of fit
ecoforecastR::ciEnvelope(time2,N.IPDE.ci[1,], N.IPDE.ci[3,], col="peachpuff1")
ecoforecastR::ciEnvelope(time2,N.IPD.ci[1,], N.IPD.ci[3,], col="lightsalmon")
ecoforecastR::ciEnvelope(time2,N.IP.ci[1,], N.IP.ci[3,], col="indianred2")
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="violetred4")
lines(time2, N.IP.ci[2,], lwd=2, type="l")
legend('topleft',
legend = c('Forecast','Inital Conditions','Parameter','Driver','Process'),
col = c('black', 'violetred4', 'indianred2','lightsalmon','peachpuff1'),
lwd = c(2,4,4,4,4),
cex = 0.75,
inset = 0.02)
############################## FOR ALL SITES
par(mfrow=c(2,2))
sp <- aegypti
nmc = 500
sites = c(4,10) ##CHANGED THIS TO ONLY RUN CA ORANGE AND FL HILLSBORO, CHANGE BACK TO 1:11 FOR ALL
Nt = 12
aegypti.forecast.mean <- aegypti.forecast.var <- list()
for(S in sites){
var <- calc.inputs(sp, S, Nt)
st <- as.matrix(sp[[S]]$predict) # get predictions
ci.state <- apply(st,2,quantile,c(0.025,0.5,0.975)) # confidence intervals for predictions
time <- 1:ncol(ci.state) # time length of fit
time2 <- seq(ncol(ci.state)+1,ncol(ci.state)+Nt) # length of forecast (beginning @ next time step after length of fit ends)
time3 = seq(1:time2[length(time2)])
x.start = ncol(ci.state)-15
draw <- sample.int(nrow(var$parameters), nmc, replace = TRUE)
draw.met <- sample.int(ncol(var$met.values$prcp), nmc, replace = TRUE)
N.det <- forecast(IC=mean(var$IC[,ncol(var$IC)]), # mean of initial conditions @ last time step of fit
# beta=t(as.matrix(var$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=1, # 1 run to generate deterministic prediction
NT=Nt) # forecast NT months into the future
N.I <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=t(as.matrix(var$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # Nmc samples for propogation
NT=Nt) # forecast 6 months into the future
N.IP <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # Nmc runs
NT=Nt) # forecast NT months into the future
N.IPD <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(var$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # 1 run to generate deterministic prediction
NT=Nt) # forecast 6 months into the future
N.IPDE <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(var$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=1/sqrt(var$parameters[draw,5]), # process error off (note tau must be a SD, not a precision)
n=nmc,
NT=Nt)
N.I.ci = apply(N.I,2,quantile, c(0.025, 0.5, 0.975))
N.IP.ci = apply(N.IP,2,quantile, c(0.025, 0.5, 0.975))
N.IPD.ci = apply(N.IPD,2,quantile, c(0.025, 0.5, 0.975))
N.IPDE.ci = apply(N.IPDE, 2, quantile, c(0.025, 0.5, 0.975))
aegypti.forecast.mean[[S]] <- apply(N.IPDE, 2, mean)
aegypti.forecast.var[[S]] <- apply(N.IPDE, 2, var)
# historical observation dates
county <- counties[S]
historical.fit <- aedes.data$data.training %>%
filter(state_county == county) %>%
unite("year_month", year, month, sep = "-")
dates.training <- unique(historical.fit$year_month)
# 2018 observations
fit.2018 <- aedes.data$data.validation %>%
filter(state_county == county) %>%
unite("year_month", year, month, sep = "-")
dates.2018 <- unique(fit.2018$year_month)
# indexing vector, only puts 7 dates on the axis
all.dates <- c(dates.training, dates.2018)
plot.run(S,time,x=c(x.start,ncol(ci.state)+Nt),ylim=c(0,max(N.IPDE.ci)), xaxt = "n", ylab = "Individual Aedes aegypti")
ecoforecastR::ciEnvelope(time2,N.IPDE.ci[1,], N.IPDE.ci[3,], col="peachpuff1")
ecoforecastR::ciEnvelope(time2,N.IPD.ci[1,], N.IPD.ci[3,], col="lightsalmon")
ecoforecastR::ciEnvelope(time2,N.IP.ci[1,], N.IP.ci[3,], col="indianred2")
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="violetred4")
lines(time2, N.IP.ci[2,], lwd=2, type="l")
legend('topleft',
legend = c('Inital Conditions','Process'),
col = c( 'violetred4', 'peachpuff1'),
lwd = c(4,4),
cex = 0.75)
axis(1, at = at, labels = all.dates[at])
}
sp <- albopictus
nmc = 500
sites = c(4,10) ##CHANGED THIS TO ONLY RUN CA ORANGE AND FL HILLSBORO, CHANGE BACK TO 1:11 FOR ALL
Nt = 12
albopictus.forecast.mean <- albopictus.forecast.var <- list()
for(S in sites){
var <- calc.inputs(sp, S, Nt)
st <- as.matrix(sp[[S]]$predict) # get predictions
ci.state <- apply(st,2,quantile,c(0.025,0.5,0.975)) # confidence intervals for predictions
time <- 1:ncol(ci.state) # time length of fit
time2 <- seq(ncol(ci.state)+1,ncol(ci.state)+Nt) # length of forecast (beginning @ next time step after length of fit ends)
time3 = seq(1:time2[length(time2)])
x.start = ncol(ci.state)-15
draw <- sample.int(nrow(var$parameters), nmc, replace = TRUE)
draw.met <- sample.int(ncol(var$met.values$prcp), nmc, replace = TRUE)
N.det <- forecast(IC=mean(var$IC[,ncol(var$IC)]), # mean of initial conditions @ last time step of fit
# beta=t(as.matrix(var$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=1, # 1 run to generate deterministic prediction
NT=Nt) # forecast NT months into the future
N.I <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=t(as.matrix(var$param.mean[-5])), # mean of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # Nmc samples for propogation
NT=Nt) # forecast 6 months into the future
N.IP <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.means[,"prcp"]), # matrix of mean precip from 5.1
# tmax=as.matrix(var$met.means[,"tmax"]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.means[,"RH"]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # Nmc runs
NT=Nt) # forecast NT months into the future
N.IPD <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(var$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=0, # process error off (note tau must be a SD, not a precision)
n=nmc, # 1 run to generate deterministic prediction
NT=Nt) # forecast 6 months into the future
N.IPDE <- forecast(IC=var$IC[draw,ncol(var$IC)], # sample initial conditions @ last time step of fit
# beta=var$parameters[draw, -5], # sample of fitted beta values (remove tau from matrix)
# prcp=as.matrix(var$met.values$prcp[,draw.met]), # matrix of ensemble precip from 5.1
# tmax=as.matrix(var$met.values$tmax[,draw.met]), # matrix of mean tmax from 5.1
# rh=as.matrix(var$met.values$RH[,draw.met]), # matrix of mean rh from 5.1
tau=1/sqrt(var$parameters[draw,5]), # process error off (note tau must be a SD, not a precision)
n=nmc,
NT=Nt)
N.I.ci = apply(N.I,2,quantile, c(0.025, 0.5, 0.975))
N.IP.ci = apply(N.IP,2,quantile, c(0.025, 0.5, 0.975))
N.IPD.ci = apply(N.IPD,2,quantile, c(0.025, 0.5, 0.975))
N.IPDE.ci = apply(N.IPDE, 2, quantile, c(0.025, 0.5, 0.975))
albopictus.forecast.mean[[S]] <- apply(N.IPDE, 2, mean)
albopictus.forecast.var[[S]] <- apply(N.IPDE, 2, var)
# historical observation dates
county <- counties[S]
historical.fit <- aedes.data$data.training %>%
filter(state_county == county) %>%
unite("year_month", year, month, sep = "-")
dates.training <- unique(historical.fit$year_month)
# 2018 observations
fit.2018 <- aedes.data$data.validation %>%
filter(state_county == county) %>%
unite("year_month", year, month, sep = "-")
dates.2018 <- unique(fit.2018$year_month)
# indexing vector, only puts 7 dates on the axis
all.dates <- c(dates.training, dates.2018)
at <- seq(1, length(all.dates), length.out = 14)
plot.run(S,time,x=c(x.start,ncol(ci.state)+Nt),ylim=c(0,max(N.IPDE.ci)), ylab = "Individual Aedes albopictus")
ecoforecastR::ciEnvelope(time2,N.IPDE.ci[1,], N.IPDE.ci[3,], col="honeydew2")
ecoforecastR::ciEnvelope(time2,N.IPD.ci[1,], N.IPD.ci[3,], col="honeydew3")
ecoforecastR::ciEnvelope(time2,N.IP.ci[1,], N.IP.ci[3,], col="honeydew4")
ecoforecastR::ciEnvelope(time2,N.I.ci[1,], N.I.ci[3,], col="gray30")
lines(time2, N.IP.ci[2,], lwd=2, type="l")
legend('topleft',
legend = c('Inital Conditions','Process'),
col = c( 'gray30', 'honeydew2'),
lwd = c(4,4),
cex = 0.75)
axis(1, at = at, labels = all.dates[at])
}
names(aegypti.forecast.mean) <- counties
names(aegypti.forecast.var) <- counties
names(albopictus.forecast.mean) <- counties
names(albopictus.forecast.var) <- counties
save(aegypti.forecast.mean,
aegypti.forecast.var,
albopictus.forecast.mean,
albopictus.forecast.var,
file = "forecast.mean.var.RData")