SWEETPO Software v1.2

A sweet potato production simulation model

SWEETPO is a crop growth model that simulates dry mass assimilation in different sweet potato (Ipomea batatas) varieties. The model simulates daily dry matter increments in storage roots and total biomass under potential crop growth conditions.

Abstract

The goal of this survey is to analyse ……….

The Pre-processing section show the procedure used to prepare the data for analysis. The steps of the procedure include:

download weather file and load data
setting data type of the fields
selecting a subset of fields.
creating a data frame derived of the original data

The outputs of the survey are made up of three plots:

Fig. 1 : Dry matter production
Fig. 2 : Canopy cover
Fig. 3 : ……..

The weather data used in this survey can be downloaded from here: http://inrm.cip.cgiar.org/home/sweetpo/climate.prn

Pre-processing

Download weather file

fileurl<-"http://inrm.cip.cgiar.org/home/sweetpo/climate.prn";
targetfile<-paste(getwd(),"/climate.prn",sep="")
download.file(fileurl,targetfile) # use this function for windows users
#download.file(fileurl,targetfile, method="curl") # use this function for Mac Os X users

Load weather dataset

climate <- read.table("climate.prn", header = TRUE)

show structure of weather dataset

str(climate)

## 'data.frame':    365 obs. of  5 variables:
##  $ Fecha: int  1 2 3 4 5 6 7 8 9 10 ...
##  $ Tmax : num  23.9 25.4 23 22.7 24 ...
##  $ Tmin : num  22.9 24.4 22.1 21.9 23.1 ...
##  $ Prec : num  0 0.17 0.52 0.02 0 0 0 0 0.03 0 ...
##  $ Rad  : num  25.2 25.3 16.4 16.9 25.6 ...

show the first six records of weather dataset

head(climate)

##   Fecha  Tmax  Tmin Prec   Rad
## 1     1 23.88 22.94 0.00 25.23
## 2     2 25.44 24.38 0.17 25.28
## 3     3 22.98 22.07 0.52 16.43
## 4     4 22.70 21.92 0.02 16.95
## 5     5 23.98 23.06 0.00 25.61
## 6     6 24.37 23.34 0.00 25.45

Set crop parameters

N : Plant density (1/m2)

fcl : Maximum ground cover

F0 : Initial light interception capacity per plant (MJ/m2)

R0 : Initial relative growth rate (1/(°C*d))

M : Maximum tuberization index (0-1)% (%)

A : Thermal time at max root growth rate (inflection pt) (oCd)

b : Slope harvest index

DMCont : Dry matter content (%)

LUE : Average light use efficiency (g/MJ)

EDay : Emergence day (day)

v : Dry matter variability

t0 : Base temperature for sweetpotato growth (°C)

N      <- 3.33;
fcl    <- 0.6275;
F0     <- 0.0082;
R0     <- 0.00898;
M      <- 0.48;
A      <- 624;
b      <- -9.45;
DMCont <- 0.4405;
LUE    <- 2.531;
EDay   <- 5;
v      <- 0.075;
t0     <- 12;

Select a subset of fields from weather dataset

MinTemp : Minimum temperature (oC)

MaxTemp : Maximum temperature (oC)

Precipit : Precipitation (mm)

Radiation : Radiation (MJ/m2)

MinTemp<-as.numeric(climate$Tmin)
MaxTemp<-as.numeric(climate$Tmax)
Precipit<-as.numeric(climate$Prec)
Radiation<-as.numeric(climate$Rad)

Set time conditions

totaldays : Total days of simulation

mm : Month simulation starts

dd : Day simulation starts

numrep : Number of runs

totaldays<-130;
mm<-5;
dd<-7;
numrep<-20;

Processing

SDate<-switch(mm, dd+0, dd+31,dd+59,dd+90,dd+120)

dWtot_mat<-matrix(0.0,totaldays,numrep)
dWtb_mat<-matrix(0.0,totaldays,numrep)
dWtbf_mat<-matrix(0.0,totaldays,numrep)
dFlinti_mat<-matrix(0.0,totaldays,numrep)

for(irep in 1:numrep)
{
  t_ac<-0.001
  dWtot<-0.0
  day<-switch(mm, dd+0, dd+31,dd+59,dd+90,dd+120,dd+151,dd+181,dd+212,dd+243,dd+273,dd+304,dd+334)
  rnd=runif(1)
  rdm = (log((1+rnd)/(1-rnd)))/1.82 ## log: natural logarithm
  for(i in 1:totaldays)
  {
    SDay<-i
    Flinti <- (fcl*N*F0*exp(R0*t_ac))/(N*F0*exp(R0*t_ac)+1-N*F0)
    HI <- M/(1.0+((t_ac/A)^b))
    dWtb <- dWtot*HI;
    dWtbf <- dWtb/DMCont;
    if(SDay>=EDay){DAE<-SDay-EDay}else{DAE<-0}
    Flint1 <- rdm*v*Flinti+Flinti
    Tmax <- MaxTemp[day]
    Tmin <- MinTemp[day]
    PAR  <- Radiation[day]*0.5
    Tx<-(Tmax+Tmin)/2.0
    Tm <- 40.0  # Maximum Temperature (oC)
    To <- 23.0  # Optimum temperature (oC)
    dW<-(LUE*Flint1*PAR)/100.0
    W <- if(Tx>=t0 & Tx<To) 1-(((Tx-To)/(To-t0))^2) else (if(Tx>=To & Tx<=Tm) -1*((Tx-Tm)/(Tm-To)) else 0.0 )
    Te <- if(DAE<=0.0) 0.0 else (if(DAE>0.0 & Tx<t0) 0.0 else (if(Tx>=t0 & Tx<=Tm) (Tx-t0)*W else 0.0))
    t_ac=t_ac+Te
    dWtot=dWtot+dW

    dWtot_mat[i,irep]=dWtot
    dWtb_mat[i,irep]=dWtb
    dWtbf_mat[i,irep]=dWtbf
    dFlinti_mat[i,irep]=Flint1
  }
}

Post-processing

average

XdWtot<-rowMeans(dWtot_mat)
XdWtb<-rowMeans(dWtb_mat)
XdWtbf<-rowMeans(dWtbf_mat)
XdFlinti<-rowMeans(dFlinti_mat)

sample standard deviation

SDdWtot<-apply(dWtot_mat,1,sd)
SDdWtb<-apply(dWtb_mat,1,sd)
SDdWtbf<-apply(dWtbf_mat,1,sd)
SDdFlinti<-apply(dFlinti_mat,1,sd)

confidence interval (95% confidence level)

LCI_dWtot<-XdWtot - qnorm(0.975)*(SDdWtot/numrep)
LCS_dWtot<-XdWtot + qnorm(0.975)*(SDdWtot/numrep)

LCI_dWtb<-XdWtb - qnorm(0.975)*(SDdWtb/numrep)
LCS_dWtb<-XdWtb + qnorm(0.975)*(SDdWtb/numrep)

LCI_dWtbf<-XdWtbf - qnorm(0.975)*(SDdWtbf/numrep)
LCS_dWtbf<-XdWtbf + qnorm(0.975)*(SDdWtbf/numrep)

LCI_dFlinti<-XdFlinti - qnorm(0.975)*(SDdFlinti/numrep)
LCS_dFlinti<-XdFlinti + qnorm(0.975)*(SDdFlinti/numrep)

Output

The first figure depict the total dry matter production (tn/ha). we use a 95% confidence level for finding the confidence interval.

plot(XdWtot,type="l",main="Dry matter production",col="black",xlab="Days",ylab="tn/ha")
lines(LCI_dWtot,col="blue")
lines(LCS_dWtot,col="red")

plot of chunk unnamed-chunk-12

plot(XdWtb,type="l",main="Root dry matter",col="black",xlab="Days",ylab="tn/ha")
lines(LCI_dWtb,col="blue")
lines(LCS_dWtb,col="red")

plot of chunk unnamed-chunk-13

plot(XdWtbf,type="l",main="Root fresh matter",col="black",xlab="Days",ylab="tn/ha")
lines(LCI_dWtbf,col="blue")
lines(LCS_dWtbf,col="red")

plot of chunk unnamed-chunk-14

plot(XdFlinti,type="l",main="Canopy cover",col="black",xlab="Days",ylab="tn/ha")
lines(LCI_dFlinti,col="blue")
lines(LCS_dFlinti,col="red")

plot of chunk unnamed-chunk-15

Conclusion

…………. …………. ………….