README.md

In fronzek/IRSanalysis: Plot and analyse impact response surfaces

IRSanalysis

An R package to plot impact response surfaces, 2-dimensional contour plots of tabled data that define values (in z-dim) to pairs of coordinates (x- and y-dimensions). These can be results of model simulations from a sensitivity analysis of two variables that have been perturbed jointly, e.g. the sensitivity of an impact model to changes in temperature and precipitation. This package was published as a supplement to Fronzek et al. (2022). Earlier versions of the code were used by Fronzek et al. (2011), Pirttioja et al. (2015), Fronzek et al. (2018), Fronzek et al. (2019) and Pirttioja et al. (2019).

Install

# from github:
devtools::install_github("fronzek/IRSanalysis")

Simple example

## winter wheat yield emulator by Olesen et al. (2007, DOI:10.1007/s10584-006-9216-1)
## based on DAISY simulations; returns DM yield in t ha-1
# default values for temperature (Ta, T3) and precipitation (P2, P3) are just examples
# deltaT, Pchange - temperature change (oC), precipitation change (%)
# Ta, T3 - annual (Ta), May-July seasonal temperature (T3) (oC)
# P2, P3 - seasonal precipitation sum (mm), P2 Feb-April, P3 May-July
# CO2 - atmospheric CO2 concentration (ppm)
# soilwhc - soil water-holding capacity (mm)
wheatYield <- function(deltaT=0, Pchange=0, Ta=10, T3=12, P2=200, P3=300, CO2=360, soilwhc=150) {
    return(
        1.66 +0.270*sqrt(CO2) + 0.0113*soilwhc
        +0.523*(Ta+deltaT) -0.0224*(Ta+deltaT)*(Ta+deltaT)
        -0.316*(T3+deltaT)
        +0.0158*(100+Pchange)*P2/100
        -0.0042*(100+Pchange)*P3/100
    )
}
## prepare some example sensitivity data using a simple winter wheat yield model
# create a table with all T (-2 -- +7oC) and P (-30 -- +40%) change combinations
datYieldEx <- data.table(expand.grid(deltaT=-2:7,Pchange=(-3:4)*10))
# Ta=10oC, T3=12oC, P2=200 mm, P3=300 mm, CO2=360 ppm, soilwhc=150 mm
datYieldEx[,yield:=wheatYield(deltaT=deltaT,Pchange=Pchange)]
# calculate the %-change in yield relative to the unperturbed estimate
datYieldEx[,yieldChange:=100*(yield/datYieldEx[deltaT==0&Pchange==0,yield]-1)]
plotIRS(dat=datYieldEx, var="yield", levels=seq(6,13.5,.5),
    title="DAISY emulator\nWinter wheat yield (t DM ha-1)", colSc="normal")
addLegend(levels=seq(6,13.5,.5),add=TRUE, pos=c(7.2,-32,7.5,42))

fronzek/IRSanalysis documentation built on Jan. 28, 2023, 2:38 a.m.