We conducted a sensitivity test of Joint UK Land Environment Simulator (JULES), in which the influence of to biophysical parameters on to find influential input parameters in the simulation of gross primary productivity (GPP) and ecosystem respiration (Re) was investigated for in two typical ecosystems inof Korea. For this test, we employed the whole-year observation of Using the eddy-covariance fluxes measured in 2006observations of eddy fluxes, then we evaluated the performance of the JULES in two major plant functional types in Korea: at two KoFlux sites: (1) a deciduous forest in complex terrain in Gwangneung and (2) a farmland with heterogeneous mosaic patches in Haenam. Our analysis showed that the simulated GPP and Re were most sensitive to leaf nitrogen concentration and wood biomass parameter at for the deciduous forest canopyin Gwangneung. At For the mixed farmland in Haenamcanopy, the modeled GPP was most sensitive to soil moisture content at saturation, whereas the modeled Re was most sensitive to leaf nitrogen concentration. At both sites, the model significantly overestimated both GPP and Re when the default values of input parameters were adopted. If we cConsidering the fact that the observed leaf nitrogen concentration wais only about 60% of its default value, the significant portion of the model’s overestimation can be attributed to such a discrepancy in the contrast of input parameters. Our finding suggests demonstrates that the above-mentioned key biophysical properties of the two ecosystems should be carefully evaluated carefully xamined prior to any realistically simulation and interpretation of ecosystem carbon exchange in Korea. in validating ecosystem models.