library(readr) library(readxl) library(dplyr) library(tidyr) library(ggplot2) library(gridExtra) library(car) library(caret) library(gstat) library(sp) library(stringr) library(corrplot) library(nlme) library(randomForest) library(caTools) library(scales) library(MASS) library(ggpubr) library(viridis) library(tidymodels) library(grid) #set working directory dir = 'D:/Research/Global_Urban_River_Emission' GHG<-read.csv(paste0(dir,'/urban_river_GHG.csv'),skip=0) #join monthly precipitation and temperature prectemp<-read.csv(paste0(dir,'/tempPrep.csv'),skip=0) GHG<-left_join(GHG,prectemp,by='COMID') GHG<- GHG%>%mutate( Temp=case_when(Mon%in%c('Jan')~tavg_01, #degree celcius Mon%in%c('Feb')~tavg_02, Mon%in%c('Mar')~tavg_03, Mon%in%c('Apr')~tavg_04, Mon%in%c('May')~tavg_05, Mon%in%c('Jun')~tavg_06, Mon%in%c('Jul')~tavg_07, Mon%in%c('Aug')~tavg_08, Mon%in%c('Sep')~tavg_09, Mon%in%c('Oct')~tavg_10, Mon%in%c('Nov')~tavg_11, Mon%in%c('Dec')~tavg_12, Mon%in%c('Feb/Apr/Jun/Aug/Oct/Dec')~(tavg_02+tavg_04+tavg_06+tavg_08+tavg_10+tavg_12)/6, Mon%in%c('Jun/Jul/Aug/Sep')~(tavg_06+tavg_07+tavg_08+tavg_09)/4, Mon%in%c('Sep/Dec/Mar/Jun')~(tavg_06+tavg_12+tavg_03+tavg_09)/4, Mon%in%c('Mar/Apr')~(tavg_03+tavg_04)/2, Mon%in%c('Apr/May')~(tavg_05+tavg_04)/2, Mon%in%c('Mar/Apr/May')~(tavg_03+tavg_04+tavg_05)/3, Mon%in%c('Jun/Jul/Aug')~(tavg_06+tavg_07+tavg_08)/3, Mon%in%c('Sep/Oct/Nov')~(tavg_09+tavg_10+tavg_11)/3, Mon%in%c('Dec/Jan/Feb')~(tavg_12+tavg_01+tavg_02)/3, Mon%in%c('Oct/Feb/Aug')~(tavg_10+tavg_02+tavg_08)/3, Mon%in%c('Jun/Sep/Apr')~(tavg_06+tavg_09+tavg_04)/3, Mon%in%c('Jul/Aug')~(tavg_07+tavg_08)/2, Mon%in%c('Sep/Oct')~(tavg_09+tavg_10)/2, Mon%in%c('Feb/Mar')~(tavg_02+tavg_03)/2, Mon%in%c('May/Jun/Jul/Aug/Sep/Oct')~(tavg_05+tavg_06+tavg_07+tavg_08+tavg_09+tavg_10)/6, Mon%in%c('Dec/Mar/Jul/Oct')~(tavg_12+tavg_03+tavg_07+tavg_10)/4, Mon%in%c('Dec/Apr/Jul/Sep')~(tavg_12+tavg_04+tavg_07+tavg_09)/4, Mon%in%c('Dec/Jun')~(tavg_12+tavg_06)/2, Mon%in%c('Jul/Jan')~(tavg_07+tavg_01)/2, Mon%in%c('Jan/Feb/May/Jun/Aug/Sep')~(tavg_01+tavg_02+tavg_05+tavg_06+tavg_08+tavg_09)/6, Mon%in%c('Feb/Mar/Apr/May')~(tavg_02+tavg_03+tavg_04+tavg_05)/4, Mon%in%c('Oct/Nov/Dec/Jan')~(tavg_10+tavg_11+tavg_12+tavg_01)/4, Mon%in%c('Sep/Jun')~(tavg_09+tavg_06)/2, Mon%in%c('Jul/Oct/Jan/Apr')~(tavg_07+tavg_10+tavg_01+tavg_04)/4, Mon%in%c('Jan/Jul')~(tavg_07+tavg_01)/2, Mon%in%c('Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep')~(tavg_01+tavg_02+tavg_03+tavg_04+tavg_05+tavg_06+tavg_07+tavg_08+tavg_09)/9, Mon%in%c('Jul/Aug/Sep')~(tavg_07+tavg_08+tavg_09)/3, Mon%in%c('Jun/Jul/Aug/Oct')~(tavg_07+tavg_08+tavg_06+tavg_10)/4, Mon%in%c('May/Sep')~(tavg_05+tavg_09)/2, Mon%in%c('Jul/Sep')~(tavg_07+tavg_09)/2, Mon%in%c('Dec/Jan/Feb/Mar/Apr/May')~(tavg_12+tavg_01+tavg_02+tavg_03+tavg_04+tavg_05)/6, Mon%in%c('Ann')~tavg_ann), Prec=case_when(Mon%in%c('Jan')~prec_01/31*365, #mm/mon Mon%in%c('Feb')~prec_02/28*365, Mon%in%c('Mar')~prec_03/31*365, Mon%in%c('Apr')~prec_04/30*365, Mon%in%c('May')~prec_05/31*365, Mon%in%c('Jun')~prec_06/30*365, Mon%in%c('Jul')~prec_07/31*365, Mon%in%c('Aug')~prec_08/31*365, Mon%in%c('Sep')~prec_09/30*365, Mon%in%c('Oct')~prec_10/31*365, Mon%in%c('Nov')~prec_11/30*365, Mon%in%c('Dec')~prec_12/31*365, Mon%in%c('Feb/Apr/Jun/Aug/Oct/Dec')~(prec_02/28*365+prec_04/30*365+prec_06/30*365+prec_08/31*365+prec_10/31*365+prec_12/31*365)/6, Mon%in%c('Jun/Jul/Aug/Sep')~(prec_06/30*365+prec_07/31*365+prec_08/31*365+prec_09/30*365)/4, Mon%in%c('Sep/Dec/Mar/Jun')~(prec_06/30*365+prec_12/31*365+prec_03/31*365+prec_09/30*365)/4, Mon%in%c('Mar/Apr')~(prec_03/31*365+prec_04/30*365)/2, Mon%in%c('Apr/May')~(prec_04/30*365+prec_05/31*365)/2, Mon%in%c('Mar/Apr/May')~(prec_03/31*365+prec_04/30*365+prec_05/31*365)/3, Mon%in%c('Jun/Jul/Aug')~(prec_06/30*365+prec_07/31*365+prec_08/31*365)/3, Mon%in%c('Sep/Oct/Nov')~(prec_09/30*365+prec_10/31*365+prec_11/30*365)/3, Mon%in%c('Dec/Jan/Feb')~(prec_12/31*365+prec_01/31*365+prec_02/28*365)/3, Mon%in%c('Oct/Feb/Aug')~(prec_10/31*365+prec_02/28*365+prec_08/31*365)/3, Mon%in%c('Jun/Sep/Apr')~(prec_06/30*365+prec_09/30*365+prec_04/30*365)/3, Mon%in%c('Jul/Aug')~(prec_07/31*365+prec_08/31*365)/2, Mon%in%c('Sep/Oct')~(prec_09/30*365+prec_10/31*365)/2, Mon%in%c('Feb/Mar')~(prec_02/28*365+prec_03/31*365)/2, Mon%in%c('May/Jun/Jul/Aug/Sep/Oct')~(prec_05/31*365+prec_06/30*365+prec_07/31*365+prec_08/31*365+prec_09/30*365+prec_10/31*365)/6, Mon%in%c('Dec/Mar/Jul/Oct')~(prec_12/31*365+prec_03/31*365+prec_07/31*365+prec_10/31*365)/4, Mon%in%c('Dec/Apr/Jul/Sep')~(prec_12/31*365+prec_04/30*365+prec_07/31*365+prec_09/30*365)/4, Mon%in%c('Dec/Jun')~(prec_12/31*365+prec_06/30*365)/2, Mon%in%c('Jul/Jan')~(prec_07/31*365+prec_01/31*365)/2, Mon%in%c('Jan/Feb/May/Jun/Aug/Sep')~(prec_01/31*365+prec_02/28*365+prec_05/31*365+prec_06/30*365+prec_08/31*365+prec_09/30*365)/6, Mon%in%c('Feb/Mar/Apr/May')~(prec_02/28*365+prec_03/31*365+prec_04/30*365+prec_05/31*365)/4, Mon%in%c('Oct/Nov/Dec/Jan')~(prec_10/31*365+prec_11/30*365+prec_12/31*365+prec_01/31*365)/4, Mon%in%c('Sep/Jun')~(prec_09/30*365+prec_06/30*365)/2, Mon%in%c('Jul/Oct/Jan/Apr')~(prec_07/31*365+prec_10/31*365+prec_01/31*365+prec_04/30*365)/4, Mon%in%c('Jan/Jul')~(prec_07/31*365+prec_01/31*365)/2, Mon%in%c('Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep')~(prec_01/31*365+prec_02/28*365+prec_03/31*365+prec_04/30*365+prec_05/31*365+prec_06/30*365+prec_07/31*365+prec_08/31*365+prec_09/30*365)/9, Mon%in%c('Jul/Aug/Sep')~(prec_07/31*365+prec_08/31*365+prec_09/30*365)/3, Mon%in%c('Jun/Jul/Aug/Oct')~(prec_07/31*365+prec_08/31*365+prec_06/30*365+prec_10/31*365)/4, Mon%in%c('May/Sep')~(prec_05/31*365+prec_09/30*365)/2, Mon%in%c('Jul/Sep')~(prec_07/31*365+prec_09/30*365)/2, Mon%in%c('Dec/Jan/Feb/Mar/Apr/May')~(prec_12/31*365+prec_01/31*365+prec_02/28*365+prec_03/31*365+prec_04/30*365+prec_05/31*365)/6, Mon%in%c('Ann')~prec_ann)) #mm/year GHG<-GHG[,-which(names(GHG)%in%paste0('tavg_',str_pad(c(1:12),2,pad=0)))] GHG<-GHG[,-which(names(GHG)%in%paste0('prec_',str_pad(c(1:12),2,pad=0)))] #join monthly GPP and NPP mongppnpp<-read.csv(paste0(dir,'/monthly_gpp_npp.csv'),skip=0) GHG<-left_join(GHG,mongppnpp,by='COMID') GHG<- GHG%>%mutate( GPP=case_when(Mon%in%c('Jan')~gpp_01, Mon%in%c('Feb')~gpp_02, Mon%in%c('Mar')~gpp_03, Mon%in%c('Apr')~gpp_04, Mon%in%c('May')~gpp_05, Mon%in%c('Jun')~gpp_06, Mon%in%c('Jul')~gpp_07, Mon%in%c('Aug')~gpp_08, Mon%in%c('Sep')~gpp_09, Mon%in%c('Oct')~gpp_10, Mon%in%c('Nov')~gpp_11, Mon%in%c('Dec')~gpp_12, Mon%in%c('Feb/Apr/Jun/Aug/Oct/Dec')~(gpp_02+gpp_04+gpp_06+gpp_08+gpp_10+gpp_12)/6, Mon%in%c('Jun/Jul/Aug/Sep')~(gpp_06+gpp_07+gpp_08+gpp_09)/4, Mon%in%c('Sep/Dec/Mar/Jun')~(gpp_06+gpp_12+gpp_03+gpp_09)/4, Mon%in%c('Mar/Apr')~(gpp_03+gpp_04)/2, Mon%in%c('Apr/May')~(gpp_05+gpp_04)/2, Mon%in%c('Mar/Apr/May')~(gpp_03+gpp_04+gpp_05)/3, Mon%in%c('Jun/Jul/Aug')~(gpp_06+gpp_07+gpp_08)/3, Mon%in%c('Sep/Oct/Nov')~(gpp_09+gpp_10+gpp_11)/3, Mon%in%c('Dec/Jan/Feb')~(gpp_12+gpp_01+gpp_02)/3, Mon%in%c('Oct/Feb/Aug')~(gpp_10+gpp_02+gpp_08)/3, Mon%in%c('Jun/Sep/Apr')~(gpp_06+gpp_09+gpp_04)/3, Mon%in%c('Jul/Aug')~(gpp_07+gpp_08)/2, Mon%in%c('Sep/Oct')~(gpp_09+gpp_10)/2, Mon%in%c('Feb/Mar')~(gpp_02+gpp_03)/2, Mon%in%c('May/Jun/Jul/Aug/Sep/Oct')~(gpp_05+gpp_06+gpp_07+gpp_08+gpp_09+gpp_10)/6, Mon%in%c('Dec/Mar/Jul/Oct')~(gpp_12+gpp_03+gpp_07+gpp_10)/4, Mon%in%c('Dec/Apr/Jul/Sep')~(gpp_12+gpp_04+gpp_07+gpp_09)/4, Mon%in%c('Dec/Jun')~(gpp_12+gpp_06)/2, Mon%in%c('Jul/Jan')~(gpp_07+gpp_01)/2, Mon%in%c('Jan/Feb/May/Jun/Aug/Sep')~(gpp_01+gpp_02+gpp_05+gpp_06+gpp_08+gpp_09)/6, Mon%in%c('Feb/Mar/Apr/May')~(gpp_02+gpp_03+gpp_04+gpp_05)/4, Mon%in%c('Oct/Nov/Dec/Jan')~(gpp_10+gpp_11+gpp_12+gpp_01)/4, Mon%in%c('Sep/Jun')~(gpp_09+gpp_06)/2, Mon%in%c('Jul/Oct/Jan/Apr')~(gpp_07+gpp_10+gpp_01+gpp_04)/4, Mon%in%c('Jan/Jul')~(gpp_07+gpp_01)/2, Mon%in%c('Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep')~(gpp_01+gpp_02+gpp_03+gpp_04+gpp_05+gpp_06+gpp_07+gpp_08+gpp_09)/9, Mon%in%c('Jul/Aug/Sep')~(gpp_07+gpp_08+gpp_09)/3, Mon%in%c('Jun/Jul/Aug/Oct')~(gpp_07+gpp_08+gpp_06+gpp_10)/4, Mon%in%c('May/Sep')~(gpp_05+gpp_09)/2, Mon%in%c('Jul/Sep')~(gpp_07+gpp_09)/2, Mon%in%c('Dec/Jan/Feb/Mar/Apr/May')~(gpp_12+gpp_01+gpp_02+gpp_03+gpp_04+gpp_05)/6, Mon%in%c('Ann')~GPP), NPP=case_when(Mon%in%c('Jan')~npp_01, Mon%in%c('Feb')~npp_02, Mon%in%c('Mar')~npp_03, Mon%in%c('Apr')~npp_04, Mon%in%c('May')~npp_05, Mon%in%c('Jun')~npp_06, Mon%in%c('Jul')~npp_07, Mon%in%c('Aug')~npp_08, Mon%in%c('Sep')~npp_09, Mon%in%c('Oct')~npp_10, Mon%in%c('Nov')~npp_11, Mon%in%c('Dec')~npp_12, Mon%in%c('Feb/Apr/Jun/Aug/Oct/Dec')~(npp_02+npp_04+npp_06+npp_08+npp_10+npp_12)/6, Mon%in%c('Jun/Jul/Aug/Sep')~(npp_06+npp_07+npp_08+npp_09)/4, Mon%in%c('Sep/Dec/Mar/Jun')~(npp_06+npp_12+npp_03+npp_09)/4, Mon%in%c('Mar/Apr')~(npp_03+npp_04)/2, Mon%in%c('Apr/May')~(npp_05+npp_04)/2, Mon%in%c('Mar/Apr/May')~(npp_03+npp_04+npp_05)/3, Mon%in%c('Jun/Jul/Aug')~(npp_06+npp_07+npp_08)/3, Mon%in%c('Sep/Oct/Nov')~(npp_09+npp_10+npp_11)/3, Mon%in%c('Dec/Jan/Feb')~(npp_12+npp_01+npp_02)/3, Mon%in%c('Oct/Feb/Aug')~(npp_10+npp_02+npp_08)/3, Mon%in%c('Jun/Sep/Apr')~(npp_06+npp_09+npp_04)/3, Mon%in%c('Jul/Aug')~(npp_07+npp_08)/2, Mon%in%c('Sep/Oct')~(npp_09+npp_10)/2, Mon%in%c('Feb/Mar')~(npp_02+npp_03)/2, Mon%in%c('May/Jun/Jul/Aug/Sep/Oct')~(npp_05+npp_06+npp_07+npp_08+npp_09+npp_10)/6, Mon%in%c('Dec/Mar/Jul/Oct')~(npp_12+npp_03+npp_07+npp_10)/4, Mon%in%c('Dec/Apr/Jul/Sep')~(npp_12+npp_04+npp_07+npp_09)/4, Mon%in%c('Dec/Jun')~(npp_12+npp_06)/2, Mon%in%c('Jul/Jan')~(npp_07+npp_01)/2, Mon%in%c('Jan/Feb/May/Jun/Aug/Sep')~(npp_01+npp_02+npp_05+npp_06+npp_08+npp_09)/6, Mon%in%c('Feb/Mar/Apr/May')~(npp_02+npp_03+npp_04+npp_05)/4, Mon%in%c('Oct/Nov/Dec/Jan')~(npp_10+npp_11+npp_12+npp_01)/4, Mon%in%c('Sep/Jun')~(npp_09+npp_06)/2, Mon%in%c('Jul/Oct/Jan/Apr')~(npp_07+npp_10+npp_01+npp_04)/4, Mon%in%c('Jan/Jul')~(npp_07+npp_01)/2, Mon%in%c('Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep')~(npp_01+npp_02+npp_03+npp_04+npp_05+npp_06+npp_07+npp_08+npp_09)/9, Mon%in%c('Jul/Aug/Sep')~(npp_07+npp_08+npp_09)/3, Mon%in%c('Jun/Jul/Aug/Oct')~(npp_07+npp_08+npp_06+npp_10)/4, Mon%in%c('May/Sep')~(npp_05+npp_09)/2, Mon%in%c('Jul/Sep')~(npp_07+npp_09)/2, Mon%in%c('Dec/Jan/Feb/Mar/Apr/May')~(npp_12+npp_01+npp_02+npp_03+npp_04+npp_05)/6, Mon%in%c('Ann')~NPP)) GHG<-GHG[,-which(names(GHG) %in% paste0('gpp_',str_pad(c(1:12),2,pad=0)))] GHG<-GHG[,-which(names(GHG) %in% paste0('npp_',str_pad(c(1:12),2,pad=0)))] #join monthly soil respiration soilResp<-read.csv(paste0(dir,'/soilResp.csv'),skip=0) GHG<-left_join(GHG,soilResp,by='COMID') GHG<- GHG%>%mutate( Soil_Resp=case_when(Mon%in%c('Jan')~SR_01*365, #g C m-2 d-1 Mon%in%c('Feb')~SR_02*365, Mon%in%c('Mar')~SR_03*365, Mon%in%c('Apr')~SR_04*365, Mon%in%c('May')~SR_05*365, Mon%in%c('Jun')~SR_06*365, Mon%in%c('Jul')~SR_07*365, Mon%in%c('Aug')~SR_08*365, Mon%in%c('Sep')~SR_09*365, Mon%in%c('Oct')~SR_10*365, Mon%in%c('Nov')~SR_11*365, Mon%in%c('Dec')~SR_12*365, Mon%in%c('Feb/Apr/Jun/Aug/Oct/Dec')~(SR_02*365+SR_04*365+SR_06*365+SR_08*365+SR_10*365+SR_12*365)/6, Mon%in%c('Jun/Jul/Aug/Sep')~(SR_06*365+SR_07*365+SR_08*365+SR_09*365)/4, Mon%in%c('Sep/Dec/Mar/Jun')~(SR_06*365+SR_12*365+SR_03*365+SR_09*365)/4, Mon%in%c('Mar/Apr')~(SR_03*365+SR_04*365)/2, Mon%in%c('Apr/May')~(SR_05*365+SR_04*365)/2, Mon%in%c('Mar/Apr/May')~(SR_03*365+SR_04*365+SR_05*365)/3, Mon%in%c('Jun/Jul/Aug')~(SR_06*365+SR_07*365+SR_08*365)/3, Mon%in%c('Sep/Oct/Nov')~(SR_09*365+SR_10*365+SR_11*365)/3, Mon%in%c('Dec/Jan/Feb')~(SR_12*365+SR_01*365+SR_02*365)/3, Mon%in%c('Oct/Feb/Aug')~(SR_10*365+SR_02*365+SR_08*365)/3, Mon%in%c('Jun/Sep/Apr')~(SR_06*365+SR_09*365+SR_04*365)/3, Mon%in%c('Jul/Aug')~(SR_07*365+SR_08*365)/2, Mon%in%c('Sep/Oct')~(SR_09*365+SR_10*365)/2, Mon%in%c('Feb/Mar')~(SR_02*365+SR_03*365)/2, Mon%in%c('May/Jun/Jul/Aug/Sep/Oct')~(SR_05*365+SR_06*365+SR_07*365+SR_08*365+SR_09*365+SR_10*365)/6, Mon%in%c('Dec/Mar/Jul/Oct')~(SR_12*365+SR_03*365+SR_07*365+SR_10*365)/4, Mon%in%c('Dec/Apr/Jul/Sep')~(SR_12*365+SR_04*365+SR_07*365+SR_09*365)/4, Mon%in%c('Dec/Jun')~(SR_12*365+SR_06*365)/2, Mon%in%c('Jul/Jan')~(SR_07*365+SR_01*365)/2, Mon%in%c('Jan/Feb/May/Jun/Aug/Sep')~(SR_01*365+SR_02*365+SR_05*365+SR_06*365+SR_08*365+SR_09*365)/6, Mon%in%c('Feb/Mar/Apr/May')~(SR_02*365+SR_03*365+SR_04*365+SR_05*365)/4, Mon%in%c('Oct/Nov/Dec/Jan')~(SR_10*365+SR_11*365+SR_12*365+SR_01*365)/4, Mon%in%c('Sep/Jun')~(SR_09*365+SR_06*365)/2, Mon%in%c('Jul/Oct/Jan/Apr')~(SR_07*365+SR_10*365+SR_01*365+SR_04*365)/4, Mon%in%c('Jan/Jul')~(SR_07*365+SR_01*365)/2, Mon%in%c('Jan/Feb/Mar/Apr/May/Jun/Jul/Aug/Sep')~(SR_01*365+SR_02*365+SR_03*365+SR_04*365+SR_05*365+SR_06*365+SR_07*365+SR_08*365+SR_09*365)/9, Mon%in%c('Jul/Aug/Sep')~(SR_07*365+SR_08*365+SR_09*365)/3, Mon%in%c('Jun/Jul/Aug/Oct')~(SR_07*365+SR_08*365+SR_06*365+SR_10*365)/4, Mon%in%c('May/Sep')~(SR_05*365+SR_09*365)/2, Mon%in%c('Jul/Sep')~(SR_07*365+SR_09*365)/2, Mon%in%c('Dec/Jan/Feb/Mar/Apr/May')~(SR_12*365+SR_01*365+SR_02*365+SR_03*365+SR_04*365+SR_05*365)/6, Mon%in%c('Ann')~SR_ann)) # g C m-2 yr-1 GHG<-GHG[,-which(names(GHG)%in%paste0('SR_',str_pad(c(1:12),2,pad=0)))] #join yearly GDP gdp<-read.csv(paste0(dir,'/GDP.csv'),skip=0) GHG<-left_join(GHG,gdp,by='COMID') GHG<- GHG%>%mutate( GDP=case_when(Year%in%c('1994')~gdp_1994, Year%in%c('1999')~gdp_1999, Year%in%c('2000')~gdp_2000, Year%in%c('2003')~gdp_2003, Year%in%c('2004')~gdp_2004, Year%in%c('2005')~gdp_2005, Year%in%c('2006')~gdp_2006, Year%in%c('2007')~gdp_2007, Year%in%c('2008')~gdp_2008, Year%in%c('2009')~gdp_2009, Year%in%c('2010')~gdp_2010, Year%in%c('2011')~gdp_2011, Year%in%c('2012')~gdp_2012, Year%in%c('2013')~gdp_2013, Year%in%c('2014')~gdp_2014, Year%in%c('2015')~gdp_2015, Year%in%c('2016')~gdp_2015, Year%in%c('2017')~gdp_2015, Year%in%c('2018')~gdp_2015, Year%in%c('2019')~gdp_2015, Year%in%c('2020')~gdp_2015, Year%in%c('2021')~gdp_2015, Year%in%c('2022')~gdp_2015, Year%in%c('1994/1995')~(gdp_1994+gdp_1995)/2, Year%in%c('2006/2007')~(gdp_2006+gdp_2007)/2, Year%in%c('2006/2007/2008/2009')~(gdp_2006+gdp_2007+gdp_2008+gdp_2009)/4, Year%in%c('2007/2009')~(gdp_2007+gdp_2009)/2, Year%in%c('2010/2011')~(gdp_2010+gdp_2011)/2, Year%in%c('2009/2010/2011')~(gdp_2009+gdp_2010+gdp_2011)/3, Year%in%c('2010/2011/2012/2013/2014')~(gdp_2010+gdp_2011+gdp_2012+gdp_2013+gdp_2014)/5, Year%in%c('2011/2012')~(gdp_2011+gdp_2012)/2, Year%in%c('2011/2014')~(gdp_2011+gdp_2014)/2, Year%in%c('2013/2014')~(gdp_2013+gdp_2014)/2, Year%in%c('2014/2015')~(gdp_2014+gdp_2015)/2, Year%in%c('2015/2016')~(gdp_2015+gdp_2015)/2, Year%in%c('2016/2017')~(gdp_2015+gdp_2015)/2, Year%in%c('2016/2017/2018/2019')~(gdp_2015+gdp_2015+gdp_2015+gdp_2015)/4, Year%in%c('2017/2018')~(gdp_2015+gdp_2015)/2, Year%in%c('2017/2018/2019')~(gdp_2015+gdp_2015+gdp_2015)/3, Year%in%c('2018/2019')~(gdp_2015+gdp_2015)/2, Year%in%c('2020/2021')~(gdp_2015+gdp_2015)/2, Year%in%c('2021/2022')~(gdp_2015+gdp_2015)/2, Year%in%c('2018/2019/2020')~(gdp_2015+gdp_2015+gdp_2015)/3)) #join yearly GDP_per_capita gdp_per_capita<-read.csv(paste0(dir,'/GDP_per_capita.csv'),skip=0) GHG<-left_join(GHG,gdp_per_capita,by='COMID') GHG<- GHG%>%mutate( GDP_per_capita=case_when( Year%in%c('1994')~gdp_per_capita_1994, Year%in%c('1999')~gdp_per_capita_1999, Year%in%c('2000')~gdp_per_capita_2000, Year%in%c('2003')~gdp_per_capita_2003, Year%in%c('2004')~gdp_per_capita_2004, Year%in%c('2005')~gdp_per_capita_2005, Year%in%c('2006')~gdp_per_capita_2006, Year%in%c('2007')~gdp_per_capita_2007, Year%in%c('2008')~gdp_per_capita_2008, Year%in%c('2009')~gdp_per_capita_2009, Year%in%c('2010')~gdp_per_capita_2010, Year%in%c('2011')~gdp_per_capita_2011, Year%in%c('2012')~gdp_per_capita_2012, Year%in%c('2013')~gdp_per_capita_2013, Year%in%c('2014')~gdp_per_capita_2014, Year%in%c('2015')~gdp_per_capita_2015, Year%in%c('2016')~gdp_per_capita_2015, Year%in%c('2017')~gdp_per_capita_2015, Year%in%c('2018')~gdp_per_capita_2015, Year%in%c('2019')~gdp_per_capita_2015, Year%in%c('2020')~gdp_per_capita_2015, Year%in%c('2021')~gdp_per_capita_2015, Year%in%c('2022')~gdp_per_capita_2015, Year%in%c('1994/1995')~(gdp_per_capita_1994+gdp_per_capita_1995)/2, Year%in%c('2006/2007')~(gdp_per_capita_2006+gdp_per_capita_2007)/2, Year%in%c('2006/2007/2008/2009')~(gdp_per_capita_2006+gdp_per_capita_2007+gdp_per_capita_2008+gdp_per_capita_2009)/4, Year%in%c('2007/2009')~(gdp_per_capita_2007+gdp_per_capita_2009)/2, Year%in%c('2010/2011')~(gdp_per_capita_2010+gdp_per_capita_2011)/2, Year%in%c('2009/2010/2011')~(gdp_per_capita_2009+gdp_per_capita_2010+gdp_per_capita_2011)/3, Year%in%c('2010/2011/2012/2013/2014')~(gdp_per_capita_2010+gdp_per_capita_2011+gdp_per_capita_2012+gdp_per_capita_2013+gdp_per_capita_2014)/5, Year%in%c('2011/2012')~(gdp_per_capita_2011+gdp_per_capita_2012)/2, Year%in%c('2011/2014')~(gdp_per_capita_2011+gdp_per_capita_2014)/2, Year%in%c('2013/2014')~(gdp_per_capita_2013+gdp_per_capita_2014)/2, Year%in%c('2014/2015')~(gdp_per_capita_2014+gdp_per_capita_2015)/2, Year%in%c('2015/2016')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2016/2017')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2016/2017/2018/2019')~(gdp_per_capita_2015+gdp_per_capita_2015+gdp_per_capita_2015+gdp_per_capita_2015)/4, Year%in%c('2017/2018')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2017/2018/2019')~(gdp_per_capita_2015+gdp_per_capita_2015+gdp_per_capita_2015)/3, Year%in%c('2018/2019')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2020/2021')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2021/2022')~(gdp_per_capita_2015+gdp_per_capita_2015)/2, Year%in%c('2018/2019/2020')~(gdp_per_capita_2015+gdp_per_capita_2015+gdp_per_capita_2015)/3)) #join yearly pop density, persons per km2 popdens<-read.csv(paste0(dir,'/popdens.csv'),skip=0) GHG<-left_join(GHG,popdens,by='COMID') GHG<- GHG%>%mutate( PopDen=case_when(Year%in%c('1994')~popdens_2000, Year%in%c('1999')~popdens_2000, Year%in%c('2000')~popdens_2000, Year%in%c('2003')~popdens_2000, Year%in%c('2004')~popdens_2000, Year%in%c('2005')~popdens_2005, Year%in%c('2006')~popdens_2005, Year%in%c('2007')~popdens_2005, Year%in%c('2008')~popdens_2005, Year%in%c('2009')~popdens_2005, Year%in%c('2010')~popdens_2010, Year%in%c('2011')~popdens_2010, Year%in%c('2012')~popdens_2010, Year%in%c('2013')~popdens_2010, Year%in%c('2014')~popdens_2010, Year%in%c('2015')~popdens_2015, Year%in%c('2016')~popdens_2015, Year%in%c('2017')~popdens_2015, Year%in%c('2018')~popdens_2015, Year%in%c('2019')~popdens_2015, Year%in%c('2020')~popdens_2015, Year%in%c('2021')~popdens_2015, Year%in%c('2022')~popdens_2015, Year%in%c('1994/1995')~(popdens_2000+popdens_2000)/2, Year%in%c('2006/2007')~(popdens_2005+popdens_2005)/2, Year%in%c('2006/2007/2008/2009')~(popdens_2005+popdens_2005+popdens_2005+popdens_2005)/4, Year%in%c('2007/2009')~(popdens_2005+popdens_2005)/2, Year%in%c('2010/2011')~(popdens_2010+popdens_2010)/2, Year%in%c('2009/2010/2011')~(popdens_2005+popdens_2010+popdens_2010)/3, Year%in%c('2010/2011/2012/2013/2014')~(popdens_2010+popdens_2010+popdens_2010+popdens_2010+popdens_2010)/5, Year%in%c('2011/2012')~(popdens_2010+popdens_2010)/2, Year%in%c('2011/2014')~(popdens_2010+popdens_2010)/2, Year%in%c('2013/2014')~(popdens_2010+popdens_2010)/2, Year%in%c('2014/2015')~(popdens_2010+popdens_2015)/2, Year%in%c('2015/2016')~(popdens_2015+popdens_2015)/2, Year%in%c('2016/2017')~(popdens_2015+popdens_2015)/2, Year%in%c('2016/2017/2018/2019')~(popdens_2015+popdens_2015+popdens_2015+popdens_2015)/4, Year%in%c('2017/2018')~(popdens_2015+popdens_2015)/2, Year%in%c('2017/2018/2019')~(popdens_2015+popdens_2015+popdens_2015)/3, Year%in%c('2018/2019')~(popdens_2015+popdens_2015)/2, Year%in%c('2020/2021')~(popdens_2015+popdens_2015)/2, Year%in%c('2021/2022')~(popdens_2015+popdens_2015)/2, Year%in%c('2018/2019/2020')~(popdens_2015+popdens_2015+popdens_2015)/3)) #join flow slope and Q flslopeQ<-read.csv(paste0(dir,'/FLslopeQ.csv'),skip=0) flslopeQ$V<-exp(-1.06+0.12*log(flslopeQ$annQ)) flslopeQ$k600<-2841*flslopeQ$Slope*flslopeQ$V+2.02 GHG<-left_join(GHG,flslopeQ[c('COMID','k600')],by='COMID') #join elevation (m) and slope (degree) elevSlope<-read.csv(paste0(dir,'/elevSlope.csv'),skip=0) elevSlope$slope<-tan(elevSlope$slope*pi/180) #m/m GHG<-left_join(GHG,elevSlope,by='COMID') #join watershed area basinarea=read.csv(paste0(dir,'/basinArea.csv')) #km2 GHG<-left_join(GHG,basinarea,by='COMID') #Variable completion GHG <- GHG %>% recipe(~.) %>% update_role(COMID,new_role = 'outcome') %>% add_role(GPP,NPP,Soil_Resp,new_role = 'terrestrial carbon flux')%>% add_role(GDP,GDP_per_capita,Pop,PopDen,new_role = 'economic')%>% step_impute_bag(GDP,GDP_per_capita,Pop,PopDen,impute_with = im_vars(has_role(match='economic')))%>% step_impute_bag(GPP,NPP,Soil_Resp,impute_with = imp_vars(has_role(match='terrestrial carbon flux'))) GHG juice(prep(GHG)) %>%view() mds<-GHG #CO2con_RFmodel CO2c<-data.frame(mds$CO2_con,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(CO2c)<-c('CO2_con','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') set.seed(0) sa<-sample(nrow(CO2c), nrow(CO2c)*0.85) train_CO2c<-CO2c[sa,] test_CO2c<-CO2c[-sa,] rfmodCO2c<-randomForest(CO2_con~Temp+Prec+GDP+GDP_per_capita+Pop+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope, data=train_CO2c,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_CO2c<-predict(rfmodCO2c,newdata=test_CO2c[,c('Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read.csv(filename) csv['CO2_con']<-predict(rfmodCO2c,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] write.csv(csv, paste0(dir,'/', mon, '_CO2con_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_CO2con_result.csv'), skip=0)) } #annual CO2con Dec$annC=(31*Jan$CO2_con+ 28*Feb$CO2_con+ 31*Mar$CO2_con+ 30*Apr$CO2_con+ 31*May$CO2_con+ 30*Jun$CO2_con+ 31*Jul$CO2_con+ 31*Aug$CO2_con+ 30*Sep$CO2_con+ 31*Oct$CO2_con+ 30*Nov$CO2_con+ 31*Dec$CO2_con)/365 city_CO2con<-data.frame(Dec$annC,Dec$OBJECTID) names(city_CO2con)<-c('annC','OBJECTID') result_CO2con <- city_CO2con %>% group_by(OBJECTID) %>% summarise(avg = mean(annC)) write.csv(result_CO2con, paste0(dir,'/city_CO2con.csv')) #CO2flux_RF model CO2f<-data.frame(mds$CO2_flux,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(CO2f)<-c('CO2_flux','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') CO2f$CO2_flux<-CO2f$CO2_flux+250 set.seed(0) sa<-sample(nrow(CO2f), nrow(CO2f)*0.85) train_CO2f<-CO2f[sa,] test_CO2f<-CO2f[-sa,] rfmodCO2f<-randomForest(CO2_flux~Temp+Prec+GDP+Pop+GDP_per_capita+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope, data=train_CO2f,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_CO2f<-predict(rfmodCO2f,newdata=test_CO2f[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read.csv(filename) csv['CO2_flux']<-predict(rfmodCO2f,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] River_area<-read.csv(paste0(dir,'/urban_riv_monthly_area.csv'),skip=0) River<-left_join(River,River_area,by='COMID') csv$CO2_flux<-csv$CO2_flux-250 csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] csv[csv$Temp < -2,]$CO2_flux<- 0 write.csv(csv, paste0(dir,'/', mon, '_CO2flux_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_CO2flux_result.csv'), skip=0)) } all_CO2f<-data.frame(Jan$OBJECTID,Jan$Temp,Feb$Temp,Mar$Temp,Apr$Temp,May$Temp,Jun$Temp, Jul$Temp,Aug$Temp,Sep$Temp,Oct$Temp,Nov$Temp,Dec$Temp, Jan$CO2_flux,Feb$CO2_flux,Mar$CO2_flux,Apr$CO2_flux,May$CO2_flux,Jun$CO2_flux, Jul$CO2_flux,Aug$CO2_flux,Sep$CO2_flux,Oct$CO2_flux,Nov$CO2_flux,Dec$CO2_flux, Jan$Jan_area,Feb$Feb_area,Mar$Mar_area,Apr$Apr_area,May$May_area,Jun$Jun_area, Jul$Jul_area,Aug$Aug_area,Sep$Sep_area,Oct$Oct_area,Nov$Nov_area,Dec$Dec_area) names(all_CO2f)<-c('OBJECTID','temp','temp','temp','temp','temp','temp', 'temp','temp','temp','temp','temp','temp', 'Jan_CO2_flux','Feb_CO2_flux','Mar_CO2_flux','Apr_CO2_flux','May_CO2_flux','Jun_CO2_flux', 'Jul_CO2_flux','Aug_CO2_flux','Sep_CO2_flux','Oct_CO2_flux','Nov_CO2_flux','Dec_CO2_flux', 'Jan_area','Feb_area','Mar_area','Apr_area','May_area','Jun_area', 'Jul_area','Aug_area','Sep_area','Oct_area','Nov_area','Dec_area') #annual CO2flux all_CO2f$annF=(31*all_CO2f$Jan_CO2_flux+ 28*all_CO2f$Feb_CO2_flux+ 31*all_CO2f$Mar_CO2_flux+ 30*all_CO2f$Apr_CO2_flux+ 31*all_CO2f$May_CO2_flux+ 30*all_CO2f$Jun_CO2_flux+ 31*all_CO2f$Jul_CO2_flux+ 31*all_CO2f$Aug_CO2_flux+ 30*all_CO2f$Sep_CO2_flux+ 31*all_CO2f$Oct_CO2_flux+ 30*all_CO2f$Nov_CO2_flux+ 31*all_CO2f$Dec_CO2_flux)/365 city_CO2flux<-data.frame(all_CO2f$annF,all_CO2f$OBJECTID) names(city_CO2flux)<-c('annF','OBJECTID') result_CO2flux <- city_CO2flux %>% group_by(OBJECTID) %>% summarise(avg = mean(annF)) write.csv(result_CO2flux, paste0(dir,'/city_CO2flux.csv')) #ice + ice-melt correction emission all_CO2f$annEE <- 0 annEE<-(31*all_CO2f$Jan_CO2_flux*all_CO2f$Jan_area*44*10^-12+ 28*all_CO2f$Feb_CO2_flux*all_CO2f$Feb_area*44*10^-12+ 31*all_CO2f$Mar_CO2_flux*all_CO2f$Mar_area*44*10^-12+ 30*all_CO2f$Apr_CO2_flux*all_CO2f$Apr_area*44*10^-12+ 31*all_CO2f$May_CO2_flux*all_CO2f$May_area*44*10^-12+ 30*all_CO2f$Jun_CO2_flux*all_CO2f$Jun_area*44*10^-12+ 31*all_CO2f$Jul_CO2_flux*all_CO2f$Jul_area*44*10^-12+ 31*all_CO2f$Aug_CO2_flux*all_CO2f$Aug_area*44*10^-12+ 30*all_CO2f$Sep_CO2_flux*all_CO2f$Sep_area*44*10^-12+ 31*all_CO2f$Oct_CO2_flux*all_CO2f$Oct_area*44*10^-12+ 30*all_CO2f$Nov_CO2_flux*all_CO2f$Nov_area*44*10^-12+ 31*all_CO2f$Dec_CO2_flux*all_CO2f$Dec_area*44*10^-12) for (i in 1:nrow(all_CO2f)) { if (any(all_CO2f$temp[i] < -2)) { all_CO2f$annEE[i] <- 1.17 * all_CO2f$annEE[i] } else { all_CO2f$annEE[i] <- all_CO2f$annEE[i] } } city_CO2emission<-data.frame(all_CO2f$annEE,all_CO2f$OBJECTID) names(city_CO2emission)<-c('annEE','OBJECTID') result_CO2emission <- city_CO2emission %>% group_by(OBJECTID) %>% summarise(cityEE = sum(annEE)) write.csv(result_CO2emission, paste0(dir,'/city_CO2emission.csv')) #CH4con_RFmodel CH4c<-data.frame(mds$CH4_con,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(CH4c)<-c('CH4_con','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') set.seed(0) sa<-sample(nrow(CH4c), nrow(CH4c)*0.85) train_CH4c<-CO2c[sa,] test_CH4c<-CO2c[-sa,] rfmodCH4c<-randomForest(CH4_con~Temp+Prec+GDP+Pop+GDP_per_capita+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope, data=train_CH4c,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_CH4c<-predict(rfmodCH4c,newdata=test_CH4c[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read.csv(filename) csv['CH4_con']<-predict(rfmodCH4c,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] write.csv(csv, paste0(dir,'/', mon, '_CH4con_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_CH4con_result.csv'), skip=0)) } #annual CH4con Dec$annC=(31*Jan$CH4_con+ 28*Feb$CH4_con+ 31*Mar$CH4_con+ 30*Apr$CH4_con+ 31*May$CH4_con+ 30*Jun$CH4_con+ 31*Jul$CH4_con+ 31*Aug$CH4_con+ 30*Sep$CH4_con+ 31*Oct$CH4_con+ 30*Nov$CH4_con+ 31*Dec$CH4_con)/365 city_CH4con<-data.frame(Dec$annC,Dec$OBJECTID) names(city_CH4con)<-c('annC','OBJECTID') result_CH4con <- city_CH4con %>% group_by(OBJECTID) %>% summarise(avg = mean(annC)) write.csv(result_CH4con, paste0(dir,'/city_CH4con.csv')) ##########Diffusive CH4flux_RF model########## dCH4f<-data.frame(mds$dCH4_flux,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(dCH4f)<-c('dCH4_flux','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') set.seed(0) sa<-sample(nrow(dCH4f), nrow(dCH4f)*0.85) train_dCH4f<-dCH4f[sa,] test_dCH4f<-dCH4f[-sa,] rfmoddCH4f<-randomForest(dCH4_flux~Temp+Prec+GDP+Pop+GDP_per_capita+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope data=train_dCH4f,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_dCH4f<-predict(rfmoddCH4f,newdata=test_dCH4f[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read_csv(filename) csv['dCH4_flux']<-predict(rfmoddCH4f,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] River_area<-read.csv(paste0(dir,'/urban_riv_monthly_area.csv'),skip=0) River<-left_join(River,River_area,by='COMID') csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] csv[csv$Temp < -2,]$dCH4_flux<- 0 write.csv(csv, paste0(dir,'/', mon, '_dCH4flux_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_dCH4flux_result.csv'), skip=0)) } all_dCH4f<-data.frame(Jan$OBJECTID,Jan$Temp,Feb$Temp,Mar$Temp,Apr$Temp,May$Temp,Jun$Temp, Jul$Temp,Aug$Temp,Sep$Temp,Oct$Temp,Nov$Temp,Dec$Temp, Jan$dCH4_flux,Feb$dCH4_flux,Mar$dCH4_flux,Apr$dCH4_flux,May$dCH4_flux,Jun$dCH4_flux, Jul$dCH4_flux,Aug$dCH4_flux,Sep$dCH4_flux,Oct$dCH4_flux,Nov$dCH4_flux,Dec$dCH4_flux, Jan$Jan_area,Feb$Feb_area,Mar$Mar_area,Apr$Apr_area,May$May_area,Jun$Jun_area, Jul$Jul_area,Aug$Aug_area,Sep$Sep_area,Oct$Oct_area,Nov$Nov_area,Dec$Dec_area) names(all_dCH4f)<-c('OBJECTID','temp','temp','temp','temp','temp','temp', 'temp','temp','temp','temp','temp','temp', 'Jan_dCH4_flux','Feb_dCH4_flux','Mar_dCH4_flux','Apr_dCH4_flux','May_dCH4_flux','Jun_dCH4_flux', 'Jul_dCH4_flux','Aug_dCH4_flux','Sep_dCH4_flux','Oct_dCH4_flux','Nov_dCH4_flux','Dec_dCH4_flux', 'Jan_area','Feb_area','Mar_area','Apr_area','May_area','Jun_area', 'Jul_area','Aug_area','Sep_area','Oct_area','Nov_area','Dec_area') #annual tCH4flux all_dCH4f$anntF= (31*(all_dCH4f$Jan_dCH4_flux+10^(0.089+1.1*all_dCH4f$Jan_dCH4_flux))+ 28*(all_dCH4f$Feb_dCH4_flux+10^(0.089+2.1*all_dCH4f$Feb_dCH4_flux))+ 31*(all_dCH4f$Mar_dCH4_flux+10^(0.089+2.1*all_dCH4f$Mar_dCH4_flux))+ 30*(all_dCH4f$Apr_dCH4_flux+10^(0.089+2.1*all_dCH4f$Apr_dCH4_flux))+ 31*(all_dCH4f$May_dCH4_flux+10^(0.089+2.1*all_dCH4f$May_dCH4_flux))+ 30*(all_dCH4f$Jun_dCH4_flux+10^(0.089+2.1*all_dCH4f$Jun_dCH4_flux))+ 31*(all_dCH4f$Jul_dCH4_flux+10^(0.089+2.1*all_dCH4f$Jul_dCH4_flux))+ 31*(all_dCH4f$Aug_dCH4_flux+10^(0.089+2.1*all_dCH4f$Aug_dCH4_flux))+ 30*(all_dCH4f$Sep_dCH4_flux+10^(0.089+2.1*all_dCH4f$Sep_dCH4_flux))+ 31*(all_dCH4f$Oct_dCH4_flux+10^(0.089+2.1*all_dCH4f$Oct_dCH4_flux))+ 30*(all_dCH4f$Nov_dCH4_flux+10^(0.089+2.1*all_dCH4f$Nov_dCH4_flux))+ 31*(all_dCH4f$Dec_dCH4_flux+10^(0.089+2.1*all_dCH4f$Dec_dCH4_flux)))/365 city_CH4flux<-data.frame(all_dCH4f$anntF,all_dCH4f$OBJECTID) names(city_CH4flux)<-c('anntF','OBJECTID') result_CH4flux <- city_CH4flux %>% group_by(OBJECTID) %>% summarise(avg = mean(anntF)) write.csv(result_CH4flux, paste0(dir,'/city_CH4flux.csv')) #ice + ice-melt correction emission all_dCH4f$annEE <- 0 annEE<-(31*(0.089+2.1*all_dCH4f$Jan_dCH4_flux)*all_dCH4f$Jan_area*16*10^-12+ 28*(0.089+2.1*all_dCH4f$Feb_dCH4_flux)*all_dCH4f$Feb_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$Mar_dCH4_flux)*all_dCH4f$Mar_area*16*10^-12+ 30*(0.089+2.1*all_dCH4f$Apr_dCH4_flux)*all_dCH4f$Apr_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$May_dCH4_flux)*all_dCH4f$May_area*16*10^-12+ 30*(0.089+2.1*all_dCH4f$Jun_dCH4_flux)*all_dCH4f$Jun_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$Jul_dCH4_flux)*all_dCH4f$Jul_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$Aug_dCH4_flux)*all_dCH4f$Aug_area*16*10^-12+ 30*(0.089+2.1*all_dCH4f$Sep_dCH4_flux)*all_dCH4f$Sep_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$Oct_dCH4_flux)*all_dCH4f$Oct_area*16*10^-12+ 30*(0.089+2.1*all_dCH4f$Nov_dCH4_flux)*all_dCH4f$Nov_area*16*10^-12+ 31*(0.089+2.1*all_dCH4f$Dec_dCH4_flux)*all_dCH4f$Dec_area*16*10^-12) for (i in 1:nrow(all_dCH4f)) { if (any(all_dCH4f$temp[i] < -2)) { all_dCH4f$annEE[i] <- 1.27 * all_dCH4f$annEE[i] } else { all_dCH4f$annEE[i] <- all_dCH4f$annEE[i] } } city_CH4emission<-data.frame(all_dCH4f$annEE,all$OBJECTID) names(city_CH4emission)<-c('annEE','OBJECTID') result_CH4emission <- city_CH4emission %>% group_by(OBJECTID) %>% summarise(cityEE=sum(annEE)) write.csv(result_CH4emission, paste0(dir,'/city_CH4emission.csv')) ##########N2Ocon_RF model########## N2Oc<-data.frame(mds$N2O_con,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(N2Oc)<-c('N2O_con','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') set.seed(0) sa<-sample(nrow(N2Oc), nrow(N2Oc)*0.85) train_N2Oc<-N2Oc[sa,] test_N2Oc<-N2Oc[-sa,] rfmodN2Oc<-randomForest(N2O_con~Temp+Prec+GDP+Pop+GDP_per_capita+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope data=train_N2Oc,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_N2Oc<-predict(rfmodN2Oc,newdata=test_N2Oc[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read.csv(filename) csv['N2O_con']<-predict(rfmodN2Oc,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] write.csv(csv, paste0(dir,'/', mon, '_N2Ocon_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_N2Ocon_result.csv'), skip=0)) } #annual N2Ocon Dec$annC= (31*Jan$N2O_con+ 28*Feb$N2O_con+ 31*Mar$N2O_con+ 30*Apr$N2O_con+ 31*May$N2O_con+ 30*Jun$N2O_con+ 31*Jul$N2O_con+ 31*Aug$N2O_con+ 30*Sep$N2O_con+ 31*Oct$N2O_con+ 30*Nov$N2O_con+ 31*Dec$N2O_con)/365 city_N2Ocon<-data.frame(Dec$annC,Dec$OBJECTID) names(city_N2Ocon)<-c('annC','OBJECTID') resultN2Ocon <- city_N2Ocon %>% group_by(OBJECTID) %>% summarise(avg = mean(annC)) write.csv(resultN2Ocon, paste0(dir,'/city_N2Ocon.csv')) #N2Oflux_RF model N2Of<-data.frame(mds$N2O_flux,mds$Temp,mds$Prec,mds$GDP,mds$GDP_per_capita,mds$Pop, mds$PopDen,mds$GPP,mds$NPP,mds$watershed_area,mds$Soil_Resp, mds$elevation,mds$slope) names(N2Of)<-c('N2O_flux','Temp','Prec','GDP','GDP_per_capita','Pop','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope') N2Of$N2O_flux<-N2Of$N2O_flux+35 set.seed(0) sa<-sample(nrow(N2Of), nrow(N2Of)*0.85) train_N2Of<-N2Of[sa,] test_N2Of<-N2Of[-sa,] rfmodN2Of<-randomForest(N2O_flux~Temp+Prec+GDP+Pop+GDP_per_capita+PopDen+GPP+NPP+watershed_area+ Soil_Resp+elevation+slope data=train_N2Of,importance=TRUE,na.action=na.omit,mtry=5,ntree=500) pred_rf_N2Of<-predict(rfmodN2Of,newdata=test_N2Of[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) for(mon in month.abb){ print(mon) filename = paste0(dir, '/', mon, '.csv') csv=read_csv(filename) csv['N2O_flux']<-predict(rfmodN2Of,newdata=csv[,c('Temp','Prec','GDP','Pop','GDP_per_capita','PopDen','GPP','NPP','watershed_area', 'Soil_Resp','elevation','slope')]) River<-read.csv(paste0(dir,'/River.csv'),skip=0) duplicates_river <- duplicated(River$COMID, fromLast = TRUE) River <- River[!duplicates_river, ] River_area<-read.csv(paste0(dir,'/urban_riv_monthly_area.csv'),skip=0) River<-left_join(River,River_area,by='COMID') csv$N2O_flux<-csv$N2O_flux-35 csv<-left_join(csv,River,by='COMID') duplicates_csv <- duplicated(csv$COMID, fromLast = TRUE) csv <- csv[!duplicates_csv, ] csv[csv$Temp < -2,]$N2O_flux<- 0 write.csv(csv, paste0(dir,'/', mon,'_N2Oflux_result.csv')) } months <- c("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec") for (mon in months) { assign(mon, read.csv(paste0(dir, '/', mon, '_N2Oflux_result.csv'), skip=0)) } all_N2Of<-data.frame(Jan$OBJECTID,Jan$Temp,Feb$Temp,Mar$Temp,Apr$Temp,May$Temp,Jun$Temp, Jul$Temp,Aug$Temp,Sep$Temp,Oct$Temp,Nov$Temp,Dec$Temp, Jan$N2O_flux,Feb$N2O_flux,Mar$N2O_flux,Apr$N2O_flux,May$N2O_flux,Jun$N2O_flux, Jul$N2O_flux,Aug$N2O_flux,Sep$N2O_flux,Oct$N2O_flux,Nov$N2O_flux,Dec$N2O_flux, Jan$Jan_area,Feb$Feb_area,Mar$Mar_area,Apr$Apr_area,May$May_area,Jun$Jun_area, Jul$Jul_area,Aug$Aug_area,Sep$Sep_area,Oct$Oct_area,Nov$Nov_area,Dec$Dec_area) names(all_N2Of)<-c('OBJECTID','temp','temp','temp','temp','temp','temp', 'temp','temp','temp','temp','temp','temp', 'Jan_N2O_flux','Feb_N2O_flux','Mar_N2O_flux','Apr_N2O_flux','May_N2O_flux','Jun_N2O_flux', 'Jul_N2O_flux','Aug_N2O_flux','Sep_N2O_flux','Oct_N2O_flux','Nov_N2O_flux','Dec_N2O_flux', 'Jan_area','Feb_area','Mar_area','Apr_area','May_area','Jun_area', 'Jul_area','Aug_area','Sep_area','Oct_area','Nov_area','Dec_area') #annual N2Oflux all_N2Of$annF= (31*all_N2Of$Jan_N2O_flux+ 28*all_N2Of$Feb_N2O_flux+ 31*all_N2Of$Mar_N2O_flux+ 30*all_N2Of$Apr_N2O_flux+ 31*all_N2Of$May_N2O_flux+ 30*all_N2Of$Jun_N2O_flux+ 31*all_N2Of$Jul_N2O_flux+ 31*all_N2Of$Aug_N2O_flux+ 30*all_N2Of$Sep_N2O_flux+ 31*all_N2Of$Oct_N2O_flux+ 30*all_N2Of$Nov_N2O_flux+ 31*all_N2Of$Dec_N2O_flux)/365 city_N2Oflux<-data.frame(all_N2Of$annF,all_N2Of$OBJECTID) names(city_N2Oflux)<-c('annF','OBJECTID') result_N2Oflux <- city_N2Oflux %>% group_by(OBJECTID) %>% summarise(avg = mean(annF)) write.csv(result_N2Oflux, paste0(dir,'/city_N2Oflux.csv')) #ice + ice-melt correction all_N2Of$annEE <- 0 annEE<- (31*all_N2Of$Jan_N2O_flux*all_N2Of$Jan_area*44*10^-15+ 28*all_N2Of$Feb_N2O_flux*all_N2Of$Feb_area*44*10^-15+ 31*all_N2Of$Mar_N2O_flux*all_N2Of$Mar_area*44*10^-15+ 30*all_N2Of$Apr_N2O_flux*all_N2Of$Apr_area*44*10^-15+ 31*all_N2Of$May_N2O_flux*all_N2Of$May_area*44*10^-15+ 30*all_N2Of$Jun_N2O_flux*all_N2Of$Jun_area*44*10^-15+ 31*all_N2Of$Jul_N2O_flux*all_N2Of$Jul_area*44*10^-15+ 31*all_N2Of$Aug_N2O_flux*all_N2Of$Aug_area*44*10^-15+ 30*all_N2Of$Sep_N2O_flux*all_N2Of$Sep_area*44*10^-15+ 31*all_N2Of$Oct_N2O_flux*all_N2Of$Oct_area*44*10^-15+ 30*all_N2Of$Nov_N2O_flux*all_N2Of$Nov_area*44*10^-15+ 31*all_N2Of$Dec_N2O_flux*all_N2Of$Dec_area*44*10^-15) for (i in 1:nrow(all_N2Of)) { if (any(all_N2Of$temp[i] < -2)) { all_N2Of$annEE[i] <- 1.17 * all_N2Of$annEE[i] } else { all_N2Of$annEE[i] <- all_N2Of$annEE[i] } } city_N2Oemission<-data.frame(all_N2Of$annEE,all_N2Of$OBJECTID) names(city_N2Oemission)<-c('annEE','OBJECTID') result_N2Oemission <- city_N2Oemission %>% group_by(OBJECTID) %>% summarise(cityEE = sum(annEE)) write.csv(result_N2Oemission, paste0(dir,'/city_N2Oemission.csv')) #Uncertainty #Uncertainty of CO2 emissions## all_log_CO2f