Abstract: 【Objective】 Hunan is one of the main regions of double cropping rice in China, and the total rice yield was the highest in the country in 2014. This study characterized the spatial and temporal variations of the potential and reported statistical yields as well as their gaps of the early and late rice in the double cropping rice areas of Hunan province. 【Method】The historical weather (1981-2010), soil, crop growth and statistical yield data of Hunan province were collected and used to parameterize the rice model ORYZA v3 through the standardized calibration and validation processes. The performance of the model with parameters of rice cultivars was evaluated with satisfied decision coefficient (R²), D-index, root mean square error (RMSE), normalized root mean square error (NRMSE), mean absolute error (MAE) between simulated and observed values. The verified model was used to compute the potential yields in each area in the period from 1981 to 2010, and then the temporal and spatial variations of the potential yields and their gaps (in both absolute and relative value) to actual yields were characterized by using ArcGIS. 【Result】The ORYZA v3 estimations highly agree with observations in the rice development phenology and yields for both of the early and late rice in Hunan province. The verified model could be used in the study of rice potential yields in Hunan province. The spatial distribution of potential yields of early and late rice were both higher in north area than in the south area. High potential yields appeared in Wugang and Shaoyang while low-yielding area was located in hilly south area. Within the study area, the inter-year variation of the potential yields of double cropping rice was slightly higher in the eastern region than that in the western region while the stability of the early rice was higher than that of the late rice. The potential yield of double cropping rice in Hunan province presented a decreasing tendency from north to south and also from 1981 to 2010, especially more significant for the late rice. The biggest yield gap of the double cropping rice appeared in the north areas and the actual yield could be increased furthermore. The yield gaps were smaller in the south area than the north, and they were relatively larger in the late rice than in the early rice, indicating that late rice yield could be improved more in this area. During the period from 1981 to 2010, the yield gaps of double cropping rice in Hunan province decreased gradually. The reduction was more obvious in the late rice than that in the early rice.【Conclusion】The potential yields of double-cropping rice decreased as the average temperature increased under the background of climate change. Meanwhile, the actual production increased continuously due to varieties improvement, technology optimization and increase of inputs. Thus yield gaps between early rice and late rice both decreased during the study period excepted a few stations like Daoxian and Yuanjiang. The yield gaps of the early rice were generally higher than those of late rice with the lower reduction rates over the period from 1981 to 2010. The positive effect of increasing sunshine hours in the growing season of early rice could partially compensate the negative effect of temperature rise. Additionally, the yield gaps of the early rice presented an increasing trend in more areas than those in the late rice. Generally, the actual yield of early rice has a larger gap to close in the future in the double cropping rice area of Hunan province.

Key words: Hunan province, double cropping rice, yield potential, yield gap, temporal and spatial distribution characteristic