Skip to contents

Example data from Sulman et al. (2018)

sulman2018.Rd

Monthly soil model outputs from a series of simple, idealized experiments looking at the effects of litter addition and removal treatments.

Usage

sulman2018

Format

A data frame of 7 columns:

model

Model name

clay

Soil clay level (all "medium" in these data)

litter

Litter quality level (all "highquality" in these data)

experiment

Name of experiment: "control", "litter_removal", "total_addition_100" (doubling), or "total_addition_30" (30% addition)

month

Month number of simulation, integer

name

Output variable: "total_protectedC", "total_unprotectedC", "total_C", "total_microbeC", "total_litterC", or "CO2flux" (all kgC/m2)

value

Model output value

Source

Sulman et al.: Multiple models and experiments underscore large uncertainty in soil carbon dynamics, Biogeochemistry 141:109–123, 2018. https://doi.org/10.1007/s10533-018-0509-z.

Downloaded 31 May 2025 from https://doi.org/10.6084/m9.figshare.6981842.

Note

This dataset is a small extract from the Sulman (2018) data, as it contains results from only three models (CORPSE, MEND, and MIMICS) and four treatments (control, no litter, 30 addition, and 100 high-quality litter scenario.

Examples

# Access the data
sulman2018
#> # A tibble: 48,042 × 7
#>    model  clay   litter      experiment month name                 value
#>    <chr>  <chr>  <chr>       <chr>      <dbl> <chr>                <dbl>
#>  1 CORPSE medium highquality control        0 total_protectedC   5.10   
#>  2 CORPSE medium highquality control        0 total_unprotectedC 0.378  
#>  3 CORPSE medium highquality control        0 total_C            5.50   
#>  4 CORPSE medium highquality control        0 total_microbeC     0.0234 
#>  5 CORPSE medium highquality control        0 total_litterC      0      
#>  6 CORPSE medium highquality control        0 CO2flux            0.00136
#>  7 CORPSE medium highquality control        1 total_protectedC   5.10   
#>  8 CORPSE medium highquality control        1 total_unprotectedC 0.378  
#>  9 CORPSE medium highquality control        1 total_C            5.50   
#> 10 CORPSE medium highquality control        1 total_microbeC     0.0233 
#> # ℹ 48,032 more rows

# Reconstruct figure 3 from Sulman et al. (2018)
# For this, we need to combine control and treatment data...
x <- subset(sulman2018, experiment == "total_addition_100")
x_control <- subset(sulman2018, experiment == "control")
names(x_control)[names(x_control) == "value"] <- "control_value"
x_control$experiment <- NULL
dat <- merge(x, x_control)
# ... and compute the log response ratio
dat$log_response <- log(dat$value) - log(dat$control_value)
library(ggplot2)
ggplot(dat, aes(month, log_response, color = model)) +
 geom_line() + facet_wrap(~name, scales = "free") +
 scale_color_manual(values = c("CORPSE" = "blue", "MEND" = "red", "MIMICS" = "purple"))