main: Run revisitmici analysis.

In tamsinedwards/revisitmici: Analysis code for Edwards et al. (2019) Nature, "Revisiting Antarctic ice loss due to marine ice cliff instability"

Description

This is the main steering function to run the analysis for Edwards et al. (2019) "Revisiting Antarctic ice loss due to marine ice cliff instability", Nature, producing figures and numerical output for Antarctic projections at 2100. Uses lhs package to generate maximin Latin Hypercube design points. All results for figures and tables can be generated except timeseries projections (Figures 2, 3) and emulator validation (Figure 6). Results are given for Figure 4 and ED Figure 5, but the manuscript figures are not plotted.

Usage

main(analysis_type, test = FALSE, calib_era = "threeEras",
  discrep_PLIO = NULL, discrep_LIG = NULL, discrep_present = NULL,
  VCLIF_max = 5)

Arguments

analysis_type	Analysis type. Options: "MICI", "NoMICI" or "SimHigh".
test	If TRUE, sets emulator ensemble size to N = 500 instead of 10,000.
calib_era	Which eras to calibrate with. Default = "threeEras" (Pliocene, Last interglacial and 1992-2017). Other options = "palaeo" (Pliocene and LIG only), "present" (1992-2017 only). If "palaeo" or "present", then only plot Figure 1 (as ED Figure 1c assumes all three eras used and other ED figures do not change).
discrep_PLIO	Pliocene model discrepancy in metres. Default = 5. Set to 0 or 10 for sensitivity analyses in ED Figure 5. If provided, then only plot Figure 1.
discrep_LIG	Last Interglacial model discrepancy in metres. Default = 2. Set to 0 or 4 for sensitivity analyses in ED Figure 5. If provided, then only plot Figure 1.
discrep_present	1992-2017 model discrepancy in cm. Default = 0.5. Set to 0 or 1 for sensitivity analyses in ED Figure 5. If provided, then only plot Figure 1.
VCLIF_max	Maximum ice wastage parameter value in km per year. Default = 5 and cannot be larger. Sensitivity analysis in ED Figure 5 uses 4 km/a.

Details

All argument combinations to reproduce results are given in "Examples" below. N.B. Currently does not check if any calibration arguments are provided but unused in analysis type: will just ignore them.

Value

Creates a ./results/ directory with the output of the analysis. $DATESTAMP is the current date and time in YY-mm-dd–HH-MM-SS format.

Main_Fig*.pdf: Main paper figures in pdf format.

ED_Fig*.tiff: Extended Data figures in TIFF format.

output–$DATESTAMP.txt: Human-readable analysis settings and results.

data–$DATESTAMP.txt: CSV of main results (if emulating).

stats–$DATESTAMP.txt: output of emulation from do_emulation.R (if emulating).

Examples

## Not run: 
# Main projections:
main("MICI") # Figure 1a; ED Figures 1, 2a,b, 3, 4a, 7
main("NoMICI") # Figure 1b; ED Figure 4c
main("SimHigh") # ED Figure 2b,c

# Sensitivity tests for ED Figure 5:
main("MICI", calib_era = "palaeo")
main("MICI", calib_era = "present")
main("MICI", discrep_PLIO = 10, discrep_LIG = 4, discrep_present = 1)
main("MICI", discrep_PLIO = 0, discrep_LIG = 0, discrep_present = 0)
main("MICI", calib_era = "palaeo", discrep_PLIO = 0, discrep_LIG = 0)
main("MICI", VCLIF_max = 4)
main("NoMICI", calib_era = "palaeo")
main("NoMICI", calib_era = "present")
main("NoMICI", discrep_PLIO = 10, discrep_LIG = 4, discrep_present = 1)
main("NoMICI", discrep_PLIO = 0, discrep_LIG = 0, discrep_present = 0)
main("NoMICI", calib_era = "palaeo", discrep_PLIO = 0, discrep_LIG = 0)

## End(Not run)

tamsinedwards/revisitmici documentation built on May 13, 2019, 9:53 p.m.