In the climate-carbon cycle system, the terrestrial ecosystem feedback is significant. In studies on feedback analysis, the ecosystem feedback is divided into the sensitivity of carbon storage to atmospheric CO₂ concentration (β_L), and temperature change (γ_L). Although ecosystems include many nonlinear processes, the scenario-and time-dependency of β_L and β_L have not been explicitly discussed. To check the validity of this simplification and its robustness, we carried out two, 1% per year (p.a.) increase and RCP4.5 scenario, experiments, using a process-based terrestrial ecosystem model forced by an existing GCM output, combined with an energy moisture balance model for the latter experiment, with 300 ensemble members perturbing twelve important and a priori unconstrained parameters. In the 1% p.a. experiment, β_L peaked around 500 ppm and then gradually decreased with increasing CO₂ level, while β_L decreased with some inter-annual variability as temperature increases. The time-dependency of β_L was small (at least for CO₂ level > 550 ppm), but that of β_L was significant, and the effect of this was larger than that of the nonlinear term (i.e., combined effect of CO₂ and temperature change). The scenario-dependency is also significant, but the effect in the estimated carbon uptake was smaller than that by the time-dependency. By investigating the background of this effect, we found that in the 1% p.a. CO₂ increase scenario, the maximum photosynthesis rate and specific leaf area (SLA, leaf area per unit dry mass) had the most significant contribution to both of β_L and β_L, and the contributions are dependent on the climate state (i.e., temperature and atmospheric CO₂ level). For carbon uptake in both experiments, SLA and coefficient of plant respiration were most significant. We also attempted to constrain the ensemble members, and found that for net primary production, soil carbon and soil respiration, the default parameter set was already well-tuned, while observations of leaf area index (LAI) strongly constrains β_L, β_L, airborne fraction and ecosystem carbon balance as the default model overestimated the LAI.