Fig. 2.

Sensitivity contributions to enhanced gross primary production response to ENSO. a Differences in the regression coefficient of gross primary production (GPP) anomalies $\left({\mathrm{\Delta \gamma }}_{\mathrm{GPP}}^{\mathrm{ENSO}}\right)$ from September to February for the December–February Niño3.4 index between the preindustrial experiment and ECP4.5. b Sum of changes in sensitivities based on each term in Eq. (4). c–f Each term in Eq. (4) plotted as a contribution to future changes in the response to ENSO and the GPP sensitivity changes to c temperature $\left({\mathrm{\Delta \gamma }}_{\mathrm{GPP}}^{\mathrm{Temp}}\right)$ and d precipitation $\left({\mathrm{\Delta \gamma }}_{\mathrm{GPP}}^{\mathrm{Prec}}\right)$; the GPP response to ENSO due to e temperature $\left({\mathrm{\Delta \gamma }}_{\mathrm{Temp}}^{\mathrm{ENSO}}\right)$ and f precipitation $\left({\mathrm{\Delta \gamma }}_{\mathrm{Prec}}^{\mathrm{ENSO}}\right)$. Units are kgC m⁻² yr⁻¹ °C⁻¹