########################################################################################################################## # ------------------------------------------------- Data Processing ---------------------------------------------------- # # R code related to # # Title: Plant diversity effects on forage quality, yield and revenues of semi-natural grasslands # # Authors: Schaub. S., Finger, R., Leiber, F., Probst, S., Kreuzer, M., Weigelt, A., Buchmann, N., Scherer-Lorenzen, M. # ########################################################################################################################## ############################# #### I. table of contents ############################# # 0. pre-settings # 1. data manipulation ## 1.a interpolation (replication) ## 1.b compute quality-adjusted yield ## 1.c compute annual variables: ### 1.c.i annual biomass yield ### 1.c.ii annual average forage quality, ### 1.c.iii annual quality-adjusted yield and ### 1.c.iv annual revenues ## 1.d create CSV files for Stata # me = metabolizable energy, mpp = milk production potential, cp = crude protein, ucp = utilizable crude protein, # om = organic matter, ndf = neutral detergent fiber, bm = biomass ############################# #### 0. pre-settings ############################# # clear workspace: rm(list = ls()) # set working directory (note this directory needs to be specified): setwd("H:/Jena_Management_Experiment/Data_R_Stata_for_Submission") # install and load packages: # install.packages("tidyverse") # load packages: require(tidyverse) ############################# #### 1. data manipulation ############################# # load data (DOI xxx) dat1 <-read.csv("schaub_etal_2019_data_jenamanagement.csv", header=T, sep=",") #---------------------------------------------------------------------------------------------------------- # a) interpolation (replication) #---------------------------------------------------------------------------------------------------------- # create scalas with days between 1. cut and X. cut of management intensities C4F100 and C4F200: daysC2 <- 43 # number of days between first and second cut daysC3 <- 92 # number of days between first and third cut daysC4 <- 132 # number of days between first and last cut dat1 <- dat1 %>% mutate( # i. computing slopes between first and last cut: slope_om = ifelse(cuts>2,(om_content_last-om_content_first)/daysC4,NA), slope_ndf = ifelse(cuts>2,(ndf_content_last-ndf_content_first)/daysC4,NA), slope_cp = ifelse(cuts>2,(cp_content_last-cp_content_first)/daysC4,NA), slope_me = ifelse(cuts>2,(me_content_last-me_content_first)/daysC4,NA), slope_mpp = ifelse(cuts>2,(mpp_last-mpp_first)/daysC4,NA), # ii. linearly interpolating forage quality for 2. cut: om_content_int1_test = ifelse(cuts>2,om_content_first+slope_om*daysC2,NA), ndf_content_int1_test = ifelse(cuts>2,ndf_content_first+slope_ndf*daysC2,NA), cp_content_int1_test = ifelse(cuts>2,cp_content_first+slope_cp*daysC2,NA), me_content_int1_test = ifelse(cuts>2,me_content_first+slope_me*daysC2,NA), mpp_int1_test = ifelse(cuts>2,mpp_first+slope_mpp*daysC2,NA), # ii. linearly interpolating forage quality for 3. cut: om_content_int2_test = ifelse(cuts>2,om_content_first+slope_om*daysC3,NA), ndf_content_int2_test = ifelse(cuts>2,ndf_content_first+slope_ndf*daysC3,NA), cp_content_int2_test = ifelse(cuts>2,cp_content_first+slope_cp*daysC3,NA), me_content_int2_test = ifelse(cuts>2,me_content_first+slope_me*daysC3,NA), mpp_int2_test = ifelse(cuts>2,mpp_first+slope_mpp*daysC3,NA)) # deleting interpolated test variables dat1 <- dat1 %>% select(-c(slope_om:mpp_int2_test)) #----------------------------------------------------------------------------- # b) compute quality-adjusted yield (= biomass yield * forage quality) #----------------------------------------------------------------------------- # identify samples with very small biomass yield or with missing biomass yield information delete <- dat1 %>% mutate( bm_delete_c1 = ifelse((cuts==1 & is.na(bm_yield_last ))|(cuts==1 & bm_yield_last<=1 ),NA,1), bm_delete_c2 = ifelse((cuts==2 & is.na(bm_yield_first))|(cuts==2 & bm_yield_first<=1)| (cuts==2 & is.na(bm_yield_last ))|(cuts==2 & bm_yield_last<=1 ),NA,1), bm_delete_c4 = ifelse((cuts==4 & is.na(bm_yield_first))|(cuts==4 & bm_yield_first<=1)| (cuts==4 & is.na(bm_yield_int1 ))|(cuts==4 & bm_yield_int1 <=1)| (cuts==4 & is.na(bm_yield_int2 ))|(cuts==4 & bm_yield_int2 <=1)| (cuts==4 & is.na(bm_yield_last ))|(cuts==4 & bm_yield_last <=1),NA,1), bm_delete = bm_delete_c1*bm_delete_c2*bm_delete_c4) %>% filter(is.na(bm_delete)) %>% select(-c(bm_delete_c1:bm_delete)) # identify the plant diversity of samples that will be deleted length((delete %>% filter(sowndiv==1))$sowndiv) length((delete %>% filter(sowndiv==2))$sowndiv) length((delete %>% filter(sowndiv==4))$sowndiv) length((delete %>% filter(sowndiv==8))$sowndiv) length((delete %>% filter(sowndiv==16))$sowndiv) length((delete %>% filter(sowndiv==60))$sowndiv) # extract sample code code_delete <- delete$samplecode # delete observations from the data with very small biomass yield dat1 <- dat1[ !(dat1$samplecode %in% code_delete), ] # compute quality-adjusted yield: dat1 <- dat1 %>% mutate( # om yield om_yield_first = bm_yield_first * om_content_first/1000, om_yield_int1 = bm_yield_int1 * om_content_int1/1000, om_yield_int2 = bm_yield_int2 * om_content_int2/1000, om_yield_last = bm_yield_last * om_content_last/1000, # ndf yield ndf_yield_first = bm_yield_first * ndf_content_first/1000, ndf_yield_int1 = bm_yield_int1 * ndf_content_int1/1000, ndf_yield_int2 = bm_yield_int2 * ndf_content_int2/1000, ndf_yield_last = bm_yield_last * ndf_content_last/1000, # cp yield cp_yield_first = bm_yield_first * cp_content_first/1000, cp_yield_int1 = bm_yield_int1 * cp_content_int1/1000, cp_yield_int2 = bm_yield_int2 * cp_content_int2/1000, cp_yield_last = bm_yield_last * cp_content_last/1000, # ucp yield ucp_yield_first = bm_yield_first * ucp_content_first/1000, # me yield me_yield_first = bm_yield_first * me_content_first/1000, me_yield_int1 = bm_yield_int1 * me_content_int1/1000, me_yield_int2 = bm_yield_int2 * me_content_int2/1000, me_yield_last = bm_yield_last * me_content_last/1000, # mpp yield per m^2 mpp_yield_first = bm_yield_first * mpp_first/1000, mpp_yield_int1 = bm_yield_int1 * mpp_int1/1000, mpp_yield_int2 = bm_yield_int2 * mpp_int2/1000, mpp_yield_last = bm_yield_last * mpp_last/1000, # mpp yield per ha mpp_yield_first_ha = mpp_yield_first*10000, mpp_yield_int1_ha = mpp_yield_int1*10000, mpp_yield_int2_ha = mpp_yield_int2*10000, mpp_yield_last_ha = mpp_yield_last*10000) #----------------------------------------------------------------------------- # c) compute: # i. annual biomass yield # ii. annual average forage quality, # iii. annual quality-adjusted yield and # iv. annual revenues #----------------------------------------------------------------------------- dat1 <- dat1 %>% mutate( # i. annual biomass yield: bm_yield_year = ifelse(cuts==1 & !is.na(bm_yield_last), bm_yield_last, ifelse(cuts==2 & !is.na(bm_yield_first) & !is.na(bm_yield_last), bm_yield_first+bm_yield_last, ifelse(cuts==4 & !is.na(bm_yield_first) & !is.na(bm_yield_int1)& !is.na(bm_yield_int2) & !is.na(bm_yield_last), bm_yield_first+bm_yield_int1+bm_yield_int2+bm_yield_last, NA))), # ii. annual average forage quality: # om content om_content_year = ifelse(cuts==1 & !is.na(om_content_last) & !is.na(bm_yield_year), om_content_last, ifelse(cuts==2 & !is.na(om_content_first) & !is.na(om_content_last) & !is.na(bm_yield_year), om_content_first * (bm_yield_first/bm_yield_year) + om_content_last * (bm_yield_last/bm_yield_year), ifelse(cuts==4 & !is.na(om_content_first) & !is.na(om_content_int1) & !is.na(om_content_int2) & !is.na(om_content_last) & !is.na(bm_yield_year), om_content_first*(bm_yield_first/bm_yield_year)+om_content_int1*(bm_yield_int1/bm_yield_year)+ om_content_int2 *(bm_yield_int2 /bm_yield_year)+om_content_last*(bm_yield_last/bm_yield_year),NA))), # ndf content ndf_content_year = ifelse(cuts==1 & !is.na(ndf_content_last) & !is.na(bm_yield_year), ndf_content_last, ifelse(cuts==2 & !is.na(ndf_content_first) & !is.na(ndf_content_last) & !is.na(bm_yield_year), ndf_content_first * (bm_yield_first/bm_yield_year) + ndf_content_last * (bm_yield_last/bm_yield_year), ifelse(cuts==4 & !is.na(ndf_content_first) & !is.na(ndf_content_int1) & !is.na(ndf_content_int2) & !is.na(ndf_content_last) & !is.na(bm_yield_year), ndf_content_first*(bm_yield_first/bm_yield_year)+ndf_content_int1*(bm_yield_int1/bm_yield_year)+ ndf_content_int2 *(bm_yield_int2 /bm_yield_year)+ndf_content_last*(bm_yield_last/bm_yield_year),NA))), # cp content cp_content_year = ifelse(cuts==1 & !is.na(cp_content_last) & !is.na(bm_yield_year), cp_content_last, ifelse(cuts==2 & !is.na(cp_content_first) & !is.na(cp_content_last) & !is.na(bm_yield_year), cp_content_first * (bm_yield_first/bm_yield_year) + cp_content_last * (bm_yield_last/bm_yield_year), ifelse(cuts==4 & !is.na(cp_content_first) & !is.na(cp_content_int1) & !is.na(cp_content_int2) & !is.na(cp_content_last) & !is.na(bm_yield_year), cp_content_first*(bm_yield_first/bm_yield_year)+cp_content_int1*(bm_yield_int1/bm_yield_year)+ cp_content_int2 *(bm_yield_int2 /bm_yield_year)+cp_content_last*(bm_yield_last/bm_yield_year),NA))), # me content me_content_year = ifelse(cuts==1 & !is.na(me_content_last) & !is.na(bm_yield_year), me_content_last, ifelse(cuts==2 & !is.na(me_content_first) & !is.na(me_content_last) & !is.na(bm_yield_year), me_content_first * (bm_yield_first/bm_yield_year) + me_content_last * (bm_yield_last/bm_yield_year), ifelse(cuts==4 & !is.na(me_content_first) & !is.na(me_content_int1) & !is.na(me_content_int2) & !is.na(me_content_last) & !is.na(bm_yield_year), me_content_first*(bm_yield_first/bm_yield_year)+me_content_int1*(bm_yield_int1/bm_yield_year)+ me_content_int2 *(bm_yield_int2 /bm_yield_year)+me_content_last*(bm_yield_last/bm_yield_year),NA))), # mpp mpp_year = ifelse(cuts==1 & !is.na(mpp_last) & !is.na(bm_yield_year), mpp_last, ifelse(cuts==2 & !is.na(mpp_first) & !is.na(mpp_last) & !is.na(bm_yield_year), mpp_first * (bm_yield_first/bm_yield_year) + mpp_last * (bm_yield_last/bm_yield_year), ifelse(cuts==4 & !is.na(mpp_first) & !is.na(mpp_int1) & !is.na(mpp_int2) & !is.na(mpp_last) & !is.na(bm_yield_year), mpp_first*(bm_yield_first/bm_yield_year)+mpp_int1*(bm_yield_int1/bm_yield_year)+ mpp_int2 *(bm_yield_int2 /bm_yield_year)+mpp_last*(bm_yield_last/bm_yield_year),NA))), # iii. annual quality-adjusted yield: # om yield om_yield_year = ifelse(cuts==1 & !is.na(om_yield_last), om_yield_last, ifelse(cuts==2 & !is.na(om_yield_first) & !is.na(om_yield_last), om_yield_first+om_yield_last, ifelse(cuts==4 & !is.na(om_yield_first) & !is.na(om_yield_int1)& !is.na(om_yield_int2) & !is.na(om_yield_last), om_yield_first+om_yield_int1+om_yield_int2+om_yield_last, NA))), # ndf yield ndf_yield_year = ifelse(cuts==1 & !is.na(ndf_yield_last), ndf_yield_last, ifelse(cuts==2 & !is.na(ndf_yield_first) & !is.na(ndf_yield_last), ndf_yield_first+ndf_yield_last, ifelse(cuts==4 & !is.na(ndf_yield_first) & !is.na(ndf_yield_int1)& !is.na(ndf_yield_int2) & !is.na(ndf_yield_last), ndf_yield_first+ndf_yield_int1+ndf_yield_int2+ndf_yield_last, NA))), # cp yield cp_yield_year = ifelse(cuts==1 & !is.na(cp_yield_last), cp_yield_last, ifelse(cuts==2 & !is.na(cp_yield_first) & !is.na(cp_yield_last), cp_yield_first+cp_yield_last, ifelse(cuts==4 & !is.na(cp_yield_first) & !is.na(cp_yield_int1)& !is.na(cp_yield_int2) & !is.na(cp_yield_last), cp_yield_first+cp_yield_int1+cp_yield_int2+cp_yield_last, NA))), # me yield me_yield_year = ifelse(cuts==1 & !is.na(me_yield_last), me_yield_last, ifelse(cuts==2 & !is.na(me_yield_first) & !is.na(me_yield_last), me_yield_first+me_yield_last, ifelse(cuts==4 & !is.na(me_yield_first) & !is.na(me_yield_int1)& !is.na(me_yield_int2) & !is.na(me_yield_last), me_yield_first+me_yield_int1+me_yield_int2+me_yield_last, NA))), # mpp yield mpp_yield_year = ifelse(cuts==1 & !is.na(mpp_yield_last), mpp_yield_last, ifelse(cuts==2 & !is.na(mpp_yield_first) & !is.na(mpp_yield_last), mpp_yield_first+mpp_yield_last, ifelse(cuts==4 & !is.na(mpp_yield_first) & !is.na(mpp_yield_int1)& !is.na(mpp_yield_int2) & !is.na(mpp_yield_last), mpp_yield_first+mpp_yield_int1+mpp_yield_int2+mpp_yield_last, NA))), # mpp yield mpp_yield_year_ha = ifelse(cuts==1 & !is.na(mpp_yield_last_ha), mpp_yield_last_ha, ifelse(cuts==2 & !is.na(mpp_yield_first_ha) & !is.na(mpp_yield_last_ha), mpp_yield_first_ha+mpp_yield_last_ha, ifelse(cuts==4 & !is.na(mpp_yield_first_ha) & !is.na(mpp_yield_int1_ha)& !is.na(mpp_yield_int2_ha) & !is.na(mpp_yield_last_ha), mpp_yield_first_ha+mpp_yield_int1_ha+mpp_yield_int2_ha+mpp_yield_last_ha, NA))), # iv. annual revenues: revenue_year = mpp_yield_year_ha * 31/100 # price: 31 Euro 100 kg^-1 ) #----------------------------------------------------------------------------- # d) create CSV files for Stata #----------------------------------------------------------------------------- # convert treatment into numeric factors: dat1 <- dat1 %>% mutate(treatment = ifelse(treatment=="C1F0",1, ifelse(treatment=="C2F0",2, ifelse(treatment=="C2F100",3, ifelse(treatment=="C4F100",4,5))))) # set working directory (note this directory needs to be specified): setwd("H:/Jena_Management_Experiment/Data_R_Stata_for_Submission/Stata_data_input") # biomass yield dat1 %>% select(c(samplecode:cuts),bm_yield_year) %>% filter(!is.na(bm_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_bm_yield.csv', row.names = F) # forage quality dat1 %>% select(c(samplecode:cuts),bm_yield_year,om_content_year) %>% filter(!is.na(bm_yield_year),!is.na(om_content_year)) %>% write.csv('schaub_etal_2019_data_input_stata_om_content.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,ndf_content_year) %>% filter(!is.na(bm_yield_year),!is.na(ndf_content_year)) %>% write.csv('schaub_etal_2019_data_input_stata_ndf_content.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,cp_content_year) %>% filter(!is.na(bm_yield_year),!is.na(cp_content_year)) %>% write.csv('schaub_etal_2019_data_input_stata_cp_content.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,ucp_content_first) %>% filter(!is.na(bm_yield_year),!is.na(ucp_content_first)) %>% write.csv('schaub_etal_2019_data_input_stata_ucp_content.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,me_content_year) %>% filter(!is.na(bm_yield_year),!is.na(me_content_year)) %>% write.csv('schaub_etal_2019_data_input_stata_me_content.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,mpp_year) %>% filter(!is.na(bm_yield_year),!is.na(mpp_year)) %>% write.csv('schaub_etal_2019_data_input_stata_mpp.csv', row.names = F) # quality-adjusted yield dat1 %>% select(c(samplecode:cuts),bm_yield_year,om_yield_year) %>% filter(!is.na(bm_yield_year),!is.na(om_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_om_yield.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,ndf_yield_year) %>% filter(!is.na(bm_yield_year),!is.na(ndf_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_ndf_yield.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,cp_yield_year) %>% filter(!is.na(bm_yield_year),!is.na(cp_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_cp_yield.csv', row.names = F) dat1 %>% filter(!is.na(bm_yield_year),!is.na(ucp_yield_first)) %>% select(c(samplecode:cuts),bm_yield_first,ucp_yield_first) %>% write.csv('schaub_etal_2019_data_input_stata_ucp_yield.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,me_yield_year) %>% filter(!is.na(bm_yield_year),!is.na(me_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_me_yield.csv', row.names = F) dat1 %>% select(c(samplecode:cuts),bm_yield_year,mpp_yield_year) %>% filter(!is.na(bm_yield_year),!is.na(mpp_yield_year)) %>% write.csv('schaub_etal_2019_data_input_stata_mpp_yield.csv', row.names = F) # revenues dat1 %>% select(c(samplecode:cuts),bm_yield_year,revenue_year) %>% filter(!is.na(bm_yield_year),!is.na(revenue_year)) %>% write.csv('schaub_etal_2019_data_input_stata_revenue.csv', row.names = F)