This repository contains analysis and plotting scripts to reproduce the emergent constraint results presented in:
Keenan et al. 2021: A constraint on historic growth in global photosynthesis due to increasing CO2.
Nature - https://www.nature.com/articles/s41586-021-04096-9
This paper was retracted in early 2022 due to an issue that affect the results presented in Figure 1. See the retraction notice here: https://www.nature.com/articles/s41586-022-04869-w
We have removed the underlying code from the repository to avoid proliferation of the error in other analyses, but leave the description below for anyone wishing to dig deeper.
Full information on the methods used in this study are attached to this paper and are available online at the link provided above; this includes information about datasets used, as well as the motivation and reasoning behind the analysis.
The scripts, written in Matlab, are:
This script derives the relationship between Sland and Beta^{GPP}, and performs variance normalization to extract the partial response.
This code, called by 'varianceNormalization.m' uses the emergent constraint between the partial response of Sland to Beta^{GPP} across models to derive the constrained Beta^{GPP}
Running A_varianceNormalization.m will produce the following figures reported in Keenan et al. 2021:
*Figure 1a-d
*ED Figure 1
*ED Figure 2
*ED Figure 3
*ED Figure 6
A_varianceNormalization.m calls B_calc_EC_andPlot.m
./figures/emergent contains the figures produced if the save_figures flag is set to 1.
./TRENDYv6_derived contains derived output from the TRENDY model simulations. TRENDY model simulations are not publically available but can be obtained through request to Prof. Sitch ([email protected])
./dataIntermediates contains output from the scripts included here, extracted from the data contained in ./TRENDYv6_derived
./functions contains plotting code and the prediction error code.