README.md

In HannesDeDeurwaerder/SWIFT: SWIFT: Tracking stable water isotopic fluctuation along the length of a plant

SWIFT (0.1) Release notes.

The SWIFT model provides multiple functions for tracking stable water (2H and 18O) isotopic fluctuations and variance along the lenght of a plant. Functions SWIFT _ SB and SWIFT _ H can respectively be used to calculate the isotopic signature over time at the stem base of the plant, or at a by the user defined height and time. Additional functions are SoilRootCond, PSI0calc and Bprep which in this order help the user to define (i) the soil to root conductivity for every defined soil layer, (ii) the water potential at stem base for every timestep, and (iii) the root length distribution for every defined soil layer.

Quicklinks

• Quick start and tutorials
• Installation
• Dependencies
• Examples
• Thanks

Quick start and tutorials

An in depth description of the formula used within the SWIFT model can be found in De Deurwaerder et al (2020). The example provided, in combination with the provided comments in the R-script should enable the user to succesfully run the model.

Installation

Use the devtools package to install the current development version from R.

Install from Github repository

# PACKAGE DEVTOOLS REQUIRED
#--------------------------
require(devtools)

# INSTALL PACKAGE FROM GITHUB REPOSITORY
#---------------------------------------
install_github("HannesDeDeurwaerder/SWIFT")

Dependencies

SWIFT has no other dependencies other than the base R installation.

Examples

The code below defines a simple example of the use of the SWIFT model, based on the example provided in the paper of De Deurwaerder et al (2020)

Run the SWIFT model

# PARAMETERS INITIALISATION
#--------------------------
n   <- 5  # Multiple number of days studied, needed for spin up of the model
tF  <- 60  # Time frequence of measurements per hour [in measurments per h] 
t   <- seq(0,24*n,length.out = 24*tF*n)     # Discrete time vector [in h]
dZ  <- 0.001    # Thickness of sampled layer [in m] 
L   <- 1        # maximum soil depth [in m]
Z   <- seq(dZ,L,dZ)  # Discrete depth vector centered [in m]
kr  <- 10*10^(-10)   # root membrane permeability [in s-1] 
                     # (source: Leuschner et al, 2004)

DBH <- 0.213  # Diameter at breast height [in m]
LA  <- 0.136  # Lumen Fraction, i.e. lumen area/sapwood area [%]
              # (derived from Zanne et al, 2010) [table 2, F-value]            
Ax  <- LA * ((1.582*((DBH*100)^1.764)) /10^4)    
            # Total lumen area  of the studied tree [m²],i.e. the lumen area 
    # fraction multiplied by sapwood area estimated from the DBH 
            # (Meinzer et al, 2001) 
ARtot <- exp(0.88*log(pi*2*(100*DBH/2)^2)-2)  
    # ARtot in [m^2] via DBH[in cm] (Cermak et al, 2006)

R0  <- 438688      # Derived from (Huang et al, 2017)
betaCom <-0.976    # Derived from (Huang et al, 2017)
beta <- 0.976      # Derived from (Huang et al, 2017)


# SOURCE THE SWIFT MODEL
#-----------------------
require(SWIFT)

# PREPARE ADITIONAL DATASETS
#---------------------------

    # a. Diurnal sapflow pattern

        data(SFday)     
            # SF for one day, time frequency: every min, expressed in
                # [kg h-1]. This data is derived from Huang et al, 2017
    uch <- 60*60*1000       # unit conversion: h to sec; from kg to m3
        SF  <- c( rep(SFday,n) )/(uch)
            # repetition of SF day over n prefered days

    # b. Soil water potential curve with depth, from Meissner et al, 2012
        CTpsi <- 101.97     # Conversion factor between MPa and m H2O
        Apsi  <- 198.4455;  Bpsi <- 448.9092;  Cpsi <- 255.5937;
                            # Apsi, Bpsi and Cpsi are values derived from the Meissner
                            # et al 2012 data from optimizing the function defined in
                            # the paper of De Deurwaerder et al. (In Review)
        PSIs  <- ((Apsi + Bpsi *log(Z) - Cpsi*Z^2) /10000) * CTpsi

    # c.Isotopic water isotopic signature with depth from Meissner et al, 2012
        lsoil   <- -73.98008;  msoil <- 0.148735;  soildev <- 0.001     
                # lsoil and msoil are values derived from the Meissner et al 2012 
                # data from optimizing the function defined in the paper of 
                # De Deurwaerder et al. (In Review) 
        D2Hsoil <- lsoil * (Z+soildev)^msoil


# REMARK
#-------
# NOTE that the data of Meissner et al 2012 is accessible and 
# can be sourced as provided below:
data(MeissnerData)


# USING THE SWIFT MODEL FUNCTIONS
#--------------------------------
    # a. Root density distribution 
        B <- Bprep(betaCom, R0, Z)      # output is in [m m^-3]

    # b. The soil-root conductance
        k <- SoilRootCond(B, kr, PSIs, Z, 'Silt Loam')

# c. The ARi
      devio=beta^(100*Z)*(1-beta^(100*dZ))
      ARi=ARtot*devio/sum(devio)

    # d. Isotopic signature fluctuation at stem base
        StemBase <- SWIFT_SB(ARi, D2Hsoil,  k,  PSIs,  SF, t, Z)

    # e. Isotopic signature at specific height and time
        tstud <- seq(2*(24*tF) , 3*(24*tF) ,1)  
                  # Provides the data of day 3 of the modeled tree 
                  # (We selected day 3 to assure proper spin up of the model). 
        hom<- 1.3   # measured at standard coring height, 1.3 m
        D2Htree <- SWIFT_H( Ax, StemBase,  hom, SF, tstud, tF)

    # e. The water potential at stem base
        PSI0vec <- PSI0calc(ARi,  k,  PSIs,  SF, t, Z)/CTpsi    
          # CTpsi to convert from m to MPa



# MAKING A SIMPLE PLOT OF THE SWIFT OUTPUT
#-----------------------------------------
      plot(t,StemBase, xlim= c( min(tstud)/tF, max(tstud)/tF), type='l', 
      lty=2,col='grey', ylab='d2H [permil, vsmow]', xlab='timesteps', 
      mgp = c(2, 0.8, 0))   
            # Plot the signature fluctuations at the stem base

      lines( tstud/tF, D2Htree, col='blue') 
            # Plot the signature fluctuations at the height defined by the user

        legend( 'bottomleft', c('at stembase',paste0('at ',hom,' m')), 
        lty=c(2,1), col=c('grey', 'blue'), bty='n', cex=0.7)    
          # Add a legend

Thanks

Special thanks to Marco D. Visser for all his appreciated help in creating this SWIFT package. I also thank all other co-authors for their help in develloping the SWIFT model.

HannesDeDeurwaerder/SWIFT documentation built on March 22, 2021, 4:49 p.m.