Changes in the sizes of precipitation events in the context of global climate change may have profound impacts on ecosystem productivity in arid and semiarid grasslands. However, we still have little knowledge about to what extent grassland productivity will respond to an individual precipitation event. In this study, we quantified the duration, the maximum, and the time-integrated amount of the response of daily gross primary productivity(GPP) to an individual precipitation event and their variations with different sizes of precipitation events in a typical temperate steppe in Inner Mongolia, China. Results showed that the duration of GPP-response(τR) and the maximum absolute GPP-response(GPPmax) increased linearly with the sizes of precipitation events(Pes), driving a corresponding increase in time-integrated amount of the GPP-response(GPPtotal) because variations of GPPtotal were largely explained by τR and GPPmax. The relative contributions of these two parameters to GPPtotal were strongly Pes-dependent. The GPPmax contributed more to the variations of GPPtotal when Pes was relatively small(<20 mm), whereas τR was the main driver to the variations of GPPtotal when Pes was relatively large. In addition, a threshold size of at least 5 mm of precipitation was required to induce a GPP-response for the temperate steppe in this study. Our work has important implications for the modeling community to obtain an advanced understanding of productivity-response of grassland ecosystems to altered precipitation regimes. Changes in the sizes of precipitation events in the context of global climate change may have profound affect on ecosystem productivity in arid and semiarid grasslands. However, we still have little knowledge about to what extent grassland productivity will respond to an individual precipitation event. study, we quantified the duration, the maximum, and the time-integrated amount of the response of daily gross primary productivity (GPP) to an individual precipitation event and their variations with different sizes of precipitation events in a typical temperate steppe in Inner Mongolia, China. Results showed that the duration of GPP-response (τR) and the maximum absolute GPP-response (GPPmax) increased linearly with the sizes of precipitation events (Pes), driving a corresponding increase in time-integrated amount of the GPP-response (GPPtotal) because variations of GPPtotal were substantially explained by τR and GPPmax. The relative contributions of these two parameters to GPPtotal wer The GPPmax contributed more to the variations of GPPtotal when Pes was relatively small (<20 mm), while τR was the main driver to the variations of GPPtotal when Pes was relatively large. In addition, a threshold size of at least 5 mm of precipitation was required to induce a GPP-response for the temperate steppe in this study. Our work has important implications for the modeling community to obtain an advanced understanding of productivity-response of grassland ecosystems to altered precipitation regimes.