During JPSS-1 VIIRS testing at Raytheon El Segundo, a larger than expected radiometric response nonlinearity was discovered in Day-Nigh Band (DNB). In addition, the DNB nonlinearity is aggregation mode dependent, where the most severe non-linear behavior are the aggregation modes used at high scan angles (<~50 degree). The DNB aggregation strategy was subsequently modified to remove modes with the most significant non-linearity. We characterized the DNB radiometric response using pre-launch tests with the modified aggregation strategy. The test data show the DNB non-linearity varies at each gain stages, detectors and aggregation modes. The non-linearity is most significant in the Low Gain Stage (LGS) and could vary from sample-to-sample. The non-linearity is also more significant in EV than in calibration view samples. The HGS nonlinearity is difficult to quantify due to the higher uncertainty in determining source radiance. Since the radiometric response non-linearity is most significant at low dn ranges, it presents challenge in DNB cross-stage calibration, an critical path to calibration DNB’s High Gain Stage (HGS) for nighttime imagery. Based on the radiometric characterization, we estimated the DNB on-orbit calibration accuracy and compared the expected DNB calibration accuracy using operational calibration approaches. The analysis showed the non-linearity will result in cross-stage gain ratio bias, and have the most significant impact on HGS. The HGS calibration accuracy can be improved when either SD data or only the more linearly behaved EV pixels are used in cross-stage calibration. Due to constrain in test data, we were not able to achieve a satisfactory accuracy and uniformity for the JPSS-1 DNB nighttime imagery quality. The JPSS-1 DNB nonlinearity is a challenging calibration issue which will likely require special attention after JPSS-1 launch.