Revision d5606ef2
Added by Benoit Parmentier over 10 years ago
climate/research/oregon/interpolation/contribution_of_covariates_paper_interpolation.R | ||
---|---|---|
4 | 4 |
#different covariates using two baselines. Accuracy methods are added in the the function script to evaluate results. |
5 | 5 |
#Figures, tables and data for the contribution of covariate paper are also produced in the script. |
6 | 6 |
#AUTHOR: Benoit Parmentier |
7 |
#MODIFIED ON: 09/03/2014
|
|
7 |
#MODIFIED ON: 09/11/2014
|
|
8 | 8 |
#Version: 5 |
9 | 9 |
#PROJECT: Environmental Layers project |
10 | 10 |
################################################################################################# |
... | ... | |
69 | 69 |
met_stations_outfiles_obj_file<-"/data/project/layers/commons/data_workflow/output_data_365d_gam_fus_lst_test_run_07172013/met_stations_outfiles_obj_gam_fusion__365d_gam_fus_lst_test_run_07172013.RData" |
70 | 70 |
CRS_locs_WGS84<-CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +towgs84=0,0,0") #Station coords WGS84 |
71 | 71 |
y_var_name <- "dailyTmax" |
72 |
out_prefix<-"_09032014"
|
|
72 |
out_prefix<-"_09112014"
|
|
73 | 73 |
out_dir<-"/home/parmentier/Data/IPLANT_project/paper_contribution_covar_analyses_tables_fig" |
74 | 74 |
#setwd(out_dir) |
75 | 75 |
|
... | ... | |
270 | 270 |
interp_area_WGS84 <-spTransform(interp_area,CRS_locs_WGS84) # Project from WGS84 to new coord. system |
271 | 271 |
|
272 | 272 |
usa_map <- getData('GADM', country='USA', level=1) #Get US map |
273 |
#usa_map <- getData('GADM', country='USA') #Get US map |
|
274 |
#usa_map <- getData(name=GADM) #Get US map |
|
275 |
#usa_map <- load_obj(list.files(pattern="^USA.*.RData")) |
|
273 | 276 |
usa_map_2 <- usa_map[usa_map$NAME_1!="Alaska",] #remove Alaska |
274 | 277 |
usa_map_2 <- usa_map_2[usa_map_2$NAME_1!="Hawaii",] #remove Hawai |
275 | 278 |
usa_map_OR <- usa_map_2[usa_map_2$NAME_1=="Oregon",] #get OR |
... | ... | |
279 | 282 |
|
280 | 283 |
#set up the output file to plot |
281 | 284 |
res_pix<-960 |
285 |
res_pix<-1200 |
|
286 |
|
|
282 | 287 |
col_mfrow<-1 |
283 | 288 |
row_mfrow<-1 |
284 | 289 |
png(filename=paste("Figure1_contribution_covariates_study_area_",out_prefix,".png",sep=""), |
... | ... | |
299 | 304 |
#opar <- par(fig=c(0.85,0.95,0.8, 0.9), new=TRUE) |
300 | 305 |
#opar <- par(fig=c(0.8,0.95,0.75, 0.9), new=TRUE) |
301 | 306 |
#opar <- par(fig=c(0.75,0.95,0.7, 0.9), new=TRUE) |
302 |
opar <- par(fig=c(0.65,0.9,0.8, 0.92), new=TRUE) |
|
307 |
#opar <- par(fig=c(0.65,0.9,0.8, 0.92), new=TRUE) |
|
308 |
opar <- par(fig=c(0.66,0.91,0.82, 0.93), new=TRUE) |
|
303 | 309 |
|
304 | 310 |
plot(usa_map_2,border="black") #border and lwd are options of graphics package polygon object |
305 | 311 |
plot(usa_map_OR,col="dark grey",add=T) |
... | ... | |
330 | 336 |
#temp.colors <- matlab.like(no_brks) |
331 | 337 |
temp.colors <- colorRampPalette(c('blue', 'khaki', 'red')) |
332 | 338 |
|
333 |
png(filename=paste("Figure_3_paper_Comparison_daily_monthly_mean_lst",out_prefix,".png",sep=""),width=980,height=480) |
|
339 |
png(filename=paste("Figure_3_paper_Comparison_daily_monthly_mean_lst",out_prefix,".png",sep=""), |
|
340 |
width=1200,height=600) |
|
334 | 341 |
par(mfrow=c(1,2)) |
335 | 342 |
plot(lst_md,col=temp.colors(25),axes=F) #use axes=F to remove lat and lon or coordinates |
336 | 343 |
plot(interp_area,add=TRUE) |
... | ... | |
340 | 347 |
title("(b) Mean for month of January") |
341 | 348 |
dev.off() |
342 | 349 |
|
343 |
png(filename=paste("Figure_3_paper_Comparison_daily_monthly_mean_lst",out_prefix,".png",sep=""),width=980,height=480)
|
|
350 |
png(filename=paste("Figure_3_paper_Comparison_daily_monthly_mean_lst",out_prefix,".png",sep=""),width=1000,height=500)
|
|
344 | 351 |
par(mfrow=c(1,2)) |
345 | 352 |
|
346 | 353 |
p <-levelplot(lst_md,col.regions=temp.colors(25),axes=F,margin=F,scales = list(draw = FALSE), #use list(draw=F) so as not to dispaly coordinates!!! |
... | ... | |
417 | 424 |
y_lab_text <- "MAE (°C)" |
418 | 425 |
add_CI <- c(TRUE,TRUE,TRUE) |
419 | 426 |
#Now set up plotting device |
420 |
res_pix<-480 |
|
427 |
#res_pix<- 960 |
|
428 |
#res_pix<- 480 |
|
429 |
res_pix <- 600 #it will be time 2 |
|
421 | 430 |
col_mfrow<-2 |
422 | 431 |
row_mfrow<-1 |
423 | 432 |
png_file_name<- paste("Figure_4_accuracy_and_distance_to_closest_fitting_station_",out_prefix,".png", sep="") |
... | ... | |
492 | 501 |
##### plot figure 5 for paper |
493 | 502 |
#### |
494 | 503 |
|
495 |
res_pix<-480 |
|
504 |
#res_pix<-480 |
|
505 |
#res_pix<-960 |
|
506 |
res_pix<- 600 #width will be time 2 |
|
496 | 507 |
col_mfrow<-2 |
497 | 508 |
row_mfrow<-1 |
498 | 509 |
png_file_name<- paste("Figure_5_proportion_of_holdout_and_accuracy_",out_prefix,".png", sep="") |
... | ... | |
616 | 627 |
############### |
617 | 628 |
#### plot figure 6 |
618 | 629 |
#set up the output file to plot |
619 |
res_pix<-480
|
|
630 |
res_pix <- 600
|
|
620 | 631 |
col_mfrow<-2 |
621 | 632 |
row_mfrow<-1 |
622 | 633 |
png_file_name<- paste("Figure_6_overtraining_tendency_and_holdout_proportion_",out_prefix,".png", sep="") |
... | ... | |
727 | 738 |
y_range<-range(unlist(month_data_list)) |
728 | 739 |
|
729 | 740 |
#Now plot figure 14 |
730 |
res_pix<-480
|
|
741 |
res_pix<-500
|
|
731 | 742 |
col_mfrow<-2 |
732 | 743 |
#row_mfrow<-2 |
733 | 744 |
row_mfrow<-1 |
... | ... | |
863 | 874 |
layout_m<-c(4,3) #one row two columns |
864 | 875 |
|
865 | 876 |
png(paste("Figure_7a_spatial_pattern_tmax_prediction_models_baseline1_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
866 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
877 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
867 | 878 |
|
868 | 879 |
p<- levelplot(pred_temp_s,main="Interpolated Surfaces Model Comparison baseline 1", ylab=NULL,xlab=NULL, |
869 | 880 |
par.settings = list(axis.text = list(font = 2, cex = 1.3),layout=layout_m, |
... | ... | |
903 | 914 |
layout_m<-c(4,3) #one row two columns |
904 | 915 |
|
905 | 916 |
png(paste("Figure_7b_spatial_correlogram_tmax_prediction_models_baseline1_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
906 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
917 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
907 | 918 |
|
908 | 919 |
p<-xyplot(data ~ lag | which, dd,type="b", |
909 | 920 |
par.settings = list(axis.text = list(font = 2, cex = 1.3),layout=layout_m, |
... | ... | |
955 | 966 |
layout_m<-c(1,2) #one row two columns |
956 | 967 |
|
957 | 968 |
png(paste("Figure_8a_spatial_pattern_tmax_prediction_models_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
958 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
969 |
height=510*layout_m[1],width=510*layout_m[2])
|
|
959 | 970 |
#X11(height=7*layout_m[1],width=7*layout_m[2]) |
960 | 971 |
|
961 | 972 |
names_layers <-c("mod1 = s(lat,long)+s(elev)","mod4 = s(lat,long)+s(LST)") |
... | ... | |
973 | 984 |
#col.regions=temp.colors(25)) |
974 | 985 |
dev.off() |
975 | 986 |
|
976 |
|
|
977 | 987 |
diff<-raster(lf1$mod1)-raster(lf1$mod4) |
978 | 988 |
names_layers <- c("mod1-mod4") |
979 | 989 |
diff<-stack(diff) |
... | ... | |
981 | 991 |
|
982 | 992 |
|
983 | 993 |
png(paste("Figure_8b_spatial_pattern_tmax_prediction_models_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
984 |
height=530*1,width=534*1)
|
|
994 |
height=510*1,width=510*1)
|
|
985 | 995 |
|
986 | 996 |
#plot(diff,col=temp.colors(100),main=names_layers) |
987 |
levelplot(diff,main="(b) Difference between models", ylab=NULL,xlab=NULL, |
|
997 |
p_diff<- levelplot(diff,main="(b) Difference between models", ylab=NULL,xlab=NULL,
|
|
988 | 998 |
margin=F, |
989 | 999 |
par.settings = list(axis.text = list(font = 2, cex = 1.3),layout=c(1,1), |
990 | 1000 |
par.main.text=list(font=2,cex=2),strip.background=list(col="white")),par.strip.text=list(font=2,cex=1.5), |
991 | 1001 |
names.attr=names_layers,col.regions=temp.colors) |
1002 |
print(p_diff) |
|
1003 |
dev.off() |
|
992 | 1004 |
|
1005 |
layout_m <- c(1,3) |
|
1006 |
png(paste("Figure_8_spatial_pattern_tmax_prediction_models_gam_levelplot_and_difference_",date_selected,out_prefix,".png", sep=""), |
|
1007 |
height=500*layout_m[1],width=500*layout_m[2]) |
|
1008 |
grid.arrange(p,p_diff,ncol=3) |
|
993 | 1009 |
dev.off() |
994 | 1010 |
|
995 | 1011 |
############################### |
... | ... | |
1163 | 1179 |
names(tb_mod4_LST_rec3) <- c("month","p_val_rec3") |
1164 | 1180 |
tb_mod4_LST_rec3$month <- 1:12 |
1165 | 1181 |
|
1166 |
res_pix<-480
|
|
1182 |
res_pix <- 500
|
|
1167 | 1183 |
layout_m <- c(1,3) # works if set to this?? ok set the resolution... |
1168 | 1184 |
#date_selected <- c("20100101","20100901") |
1169 | 1185 |
png(paste("Figure_15_paper_","LST_elev_",out_prefix,".png", sep=""), |
... | ... | |
1300 | 1316 |
layout_m <- c(2,2) # works if set to this?? ok set the resolution... |
1301 | 1317 |
#date_selected <- c("20100101","20100901") |
1302 | 1318 |
png(paste("Figure10_paper_","_variogram_",out_prefix,".png", sep=""), |
1303 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1319 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
1304 | 1320 |
#height=480*6,width=480*4) |
1305 | 1321 |
|
1306 | 1322 |
p_hist <-histogram(tb_variogram$model, |
... | ... | |
1378 | 1394 |
|
1379 | 1395 |
layout_m<-c(4,3) # works if set to this?? ok set the resolution... |
1380 | 1396 |
png(paste("Figure_11_paper_","_residuals_",date_selected,"_",out_prefix,".png", sep=""), |
1381 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1397 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
1382 | 1398 |
|
1383 | 1399 |
grid.arrange(list_p[[1]],list_p[[2]],list_p[[3]], |
1384 | 1400 |
list_p[[4]],list_p[[5]],list_p[[6]], |
... | ... | |
1389 | 1405 |
|
1390 | 1406 |
layout_m<-c(1,3) # works if set to this?? ok set the resolution... |
1391 | 1407 |
png(paste("Figure_11a_paper_","_residuals_",date_selected,"_",out_prefix,".png", sep=""), |
1392 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1408 |
height=520*layout_m[1],width=520*layout_m[2])
|
|
1393 | 1409 |
|
1394 | 1410 |
grid.arrange(list_p[[1]],list_p[[2]],list_p[[3]], |
1395 | 1411 |
ncol=3) |
1396 | 1412 |
dev.off() |
1397 | 1413 |
layout_m<-c(1,3) # works if set to this?? ok set the resolution... |
1398 | 1414 |
png(paste("Figure_11b_paper_","_residuals_",date_selected,"_",out_prefix,".png", sep=""), |
1399 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1415 |
height=520*layout_m[1],width=520*layout_m[2])
|
|
1400 | 1416 |
|
1401 | 1417 |
grid.arrange(list_p[[4]],list_p[[5]],list_p[[6]], |
1402 | 1418 |
ncol=3) |
1403 | 1419 |
dev.off() |
1404 | 1420 |
layout_m<-c(1,3) # works if set to this?? ok set the resolution... |
1405 | 1421 |
png(paste("Figure_11c_paper_","_residuals_",date_selected,"_",out_prefix,".png", sep=""), |
1406 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1422 |
height=520*layout_m[1],width=520*layout_m[2])
|
|
1407 | 1423 |
|
1408 | 1424 |
grid.arrange(list_p[[7]],list_p[[8]],list_p[[9]], |
1409 | 1425 |
ncol=3) |
1410 | 1426 |
dev.off() |
1411 | 1427 |
layout_m<-c(1,3) # works if set to this?? ok set the resolution... |
1412 | 1428 |
png(paste("Figure_11d_paper_","_residuals_",date_selected,"_",out_prefix,".png", sep=""), |
1413 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1429 |
height=520*layout_m[1],width=520*layout_m[2])
|
|
1414 | 1430 |
|
1415 | 1431 |
grid.arrange(list_p[[10]], |
1416 | 1432 |
ncol=3) |
... | ... | |
1500 | 1516 |
|
1501 | 1517 |
layout_m<-c(1,3) # works if set to this?? ok set the resolution... |
1502 | 1518 |
png(paste("Figure12_paper_","average_MAE_","_",out_prefix,".png", sep=""), |
1503 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1504 |
X11(height=7*layout_m[1],width=7*layout_m[2]) |
|
1519 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
1520 |
#X11(height=7*layout_m[1],width=7*layout_m[2])
|
|
1505 | 1521 |
|
1506 | 1522 |
grid.arrange(list_p_mae[[1]],list_p_mae[[2]],list_p_mae[[3]],ncol=3) |
1507 |
savePlot(paste("Figure12_paper_","average_MAE_","_",out_prefix,".png", sep=""), |
|
1508 |
type="png") |
|
1523 |
#savePlot(paste("Figure12_paper_","average_MAE_","_",out_prefix,".png", sep=""),
|
|
1524 |
# type="png")
|
|
1509 | 1525 |
|
1510 | 1526 |
dev.off() |
1511 | 1527 |
|
... | ... | |
1531 | 1547 |
|
1532 | 1548 |
layout_m <- c(1,3) # works if set to this?? ok set the resolution... |
1533 | 1549 |
png(paste("Figure_13_paper_","residuals_MAE_",out_prefix,".png", sep=""), |
1534 |
height=480*layout_m[1],width=480*layout_m[2])
|
|
1550 |
height=500*layout_m[1],width=500*layout_m[2])
|
|
1535 | 1551 |
#height=480*6,width=480*4) |
1536 | 1552 |
|
1537 | 1553 |
p_bw1<-bwplot(data_v_ag$res_mod1~elev_rcstat,do.out=F,ylim=c(-15,15), |
... | ... | |
1601 | 1617 |
#max_val<- -10 |
1602 | 1618 |
#min_val <- 0 |
1603 | 1619 |
layout_m<-c(1,3) #one row, three columns |
1604 |
res_pix <- 480
|
|
1620 |
res_pix <- 500
|
|
1605 | 1621 |
png(paste("Figure_0a_graphic_abstact_spatial_pattern_tmax_prediction_models_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
1606 | 1622 |
height=res_pix*layout_m[1],width=res_pix*layout_m[2]) |
1607 | 1623 |
|
... | ... | |
1646 | 1662 |
dd$lag <- mydata$lag |
1647 | 1663 |
|
1648 | 1664 |
layout_m<-c(1,3) #one row three columns |
1649 |
res_pix <- 480
|
|
1665 |
res_pix <- 500
|
|
1650 | 1666 |
png(paste("Figure_0b_graphic_abstract_spatial_correlogram_tmax_prediction_models_gam_levelplot_",date_selected,out_prefix,".png", sep=""), |
1651 | 1667 |
height=res_pix*layout_m[1],width=res_pix*layout_m[2]) |
1652 | 1668 |
|
Also available in: Unified diff
proof RS paper contribution of covar, increasing resolution of figures