Item

Abstract

Crop phenology changes are important indicators of climate change. Climate change impacts on crop phenology are generally investigated through statistical analysis of the relationship between growth period length and growth period mean temperature. However, growth periods may be either earlier or later in a given year; hence, changes in mean temperature indicate both the effects of climate change and those attributable to seasonal temperature differences. Failure to consider temperature change resulting from seasonal shifts can lead to biased estimation of warming trends and their corresponding impact on phenology. We evaluated this potential bias in rice phenology change in 892 phenology series from China by applying time series regression control for phenological dates. The results indicate that the true magnitudes of climate change for early rice, late rice, and single rice are 0.20–0.56, 0.23–0.86, and 0.28–0.38 K/decade, after correction for the effects of seasonal shifts. The effects of seasonal shifts of growth periods led to underestimates of the magnitude of climate change by 0.16–0.22 and 0.05–0.08 K/decade for early rice and single rice, respectively, and an overestimate of the effect for late rice of 0.02–0.06 K/decade. Correspondingly, the net warming impacts on growth period length after correcting for the effects of seasonal shifts were − 2.7 d/K for early rice, − 4.8 d/K for late rice, and − 3.1 d/K for single rice, which were weaker for early and single rice, but stronger for late rice, relative to previous reports. Changes in growth period length were most closely associated with variation in phenological dates, while their relationship with climate change was less pronounced. Our results indicate that earlier phenological dates and prolonged-duration cultivars have been adopted to offset the impact of climate change, providing further evidence of active adaptation of rice cultivation practice to climate change in China.