Fig. 9: Reconstruction of RCP4.5 principle component using FEM-BV-VARX representations of RCP2.6 and RCP8.5 climate model output. | Communications Earth & Environment

Fig. 9: Reconstruction of RCP4.5 principle component using FEM-BV-VARX representations of RCP2.6 and RCP8.5 climate model output.

From: A machine learning approach to rapidly project climate responses under a multitude of net-zero emission pathways

Fig. 9

a Projection of RCP2.6 and RCP8.5 FEM-BV-VARX models using their associated exogenous concentrations, compared to PCs of all three scenarios (black dashed). b Projection of RCP2.6 and RCP8.5 FEM-BV-VARX models using RCP4.5 exogenous concentrations. c Centred 3-year moving average of the squared distance between the FEM-BV-VARX projections using RCP4.5 exogenous concentrations and the PCs of the associated scenarios (RCP2.6, RCP8.5) from which the FEM-BV-VARX models were learnt. d Weights for combining time varying FEM-BV-VARX parameters. e FEM-BV-VARX projection of RCP4.5 (magenta) using the ws(t) weighted average of the RCP2.6 an RCP8.5 model parameters, compared to the first PC of each scenario (black dashed).

Back to article page