将如下段落用规范英语润色Figure 3 shows a comparison of observations from sediment traps and simulated annual POC fluxes corresponding to sediment trap locations and depths for the Control and for the ten sensitiv

Figure 3 illustrates a comparison between observations from sediment traps and simulated annual POC fluxes corresponding to sediment trap locations and depths for both the Control and the ten sensitive experiments with RLM added. A type II linear regression of log10-transformed POC flux estimates from all experiments against log10-transformed observations yields a regression slope close to 0.4, indicating that the model underestimated POC fluxes in the mesopelagic ocean and slightly overestimated POC fluxes in the deep ocean (Lima 2014). The correlation coefficient between model outputs and observations obtained is slightly less than 0.6. Despite the errors and uncertainties of model estimates and sediment traps, the simulation results can still be considered reasonable when compared with other circulation models combined with biogeochemical models. Figure 4's Taylor diagram summarizes the fit between simulation results of all experiments and observation results, with a relative standard deviation near 1.5, indicating that the predictions have more spatial variability in POC fluxes than the sediment trap observations