Fig. 1: Schematic overview of the framework developed to explore the influence of slow driver trajectories and/or noise on the timing of ATDCs.
From: Earlier collapse of Anthropocene ecosystems driven by multiple faster and noisier drivers
a, The four systems models simulated in this study (see section on Overview of systems models). b, Schematic representation of a system dynamics model (Lake Phosphorus model) with its external slow (blue and green) and noisy (red/orange) drivers depicted in colour (see Generation of future scenarios). c, Depiction of the four experiment types (section on Generation of future scenarios), ranging from changes in the primary baseline driver only (experiment 1), changes in all slow drivers and noise inputs simultaneously (experiment 4, where ‘a’ and ‘b’ represent noise profiles that are uncoupled or coupled to the primary driver trajectory, respectively): darker colours schematically represent steeper trajectories and/or higher noise levels. d, The two linear techniques used to check whether outcomes shift into a functionally different state (section on Time-series breakpoint detection)—the top panel is applied to Lake Chilika, Easter Island and TRIFFID, where the systems collapse from high quantitative outcome states to low quantitative outcome states and the bottom panel is applied to Lake Phosphorus (where lake phosphorus concentrations shift from low to high). e, Depiction of the time-series breakpoint date recognition (section on Time-series breakpoint detection). The Easter Island icon in a is made by Roundicons and the remaining three icons are made by Freekpik, as sourced from www.flaticon.com.
