Examining Spring Phenological Responses To Temperature Variations During Different Periods In Subtropical And Tropical China

Using ground observed Melia azedarach budburst, first leaf unfolding, and first flowering dates and daily mean temperature data (1981 to 2014) at 37 stations in China's subtropical and tropical zones, we identified the positive and negative temperature effects on spring phenology from winter and spring, and their spatial and seasonal variations. Then, we grouped the spatial variations into temperature effect phases. Lastly, we examined relationships between mean temperature during the positive effect phase and effective accumulated temperature during the negative effect phase. Results show that Melia azedarach spring phenology is positively influenced (delayed) by preseason winter temperatures and negatively influenced (advanced) by preseason spring temperatures. The negative spring temperature effect on the three phenophases was detected at all stations, while the positive winter temperature effect was apparent only at 51%-78% of stations. The number of stations with the positive winter temperature effect is larger in colder areas and for earlier occurring phenophases, and smaller in warmer areas and for later occurring phenophases. In addition, either negative responses of spring phenology to spring temperature or positive responses of spring phenology to winter temperature are more sensitive at warmer locations than colder locations. Further analyses indicate that the lower the mean temperature during the positive winter temperature effect phase at a specific station, the less the effective accumulated temperature needed for budburst, first leaf unfolding, and first flowering during the negative spring temperature effect phase at the same station. This implies that effective accumulated temperature during the negative effect phase depends on temperature status during the positive effect phase. Our study highlights that the interannual variation of Melia azedarach spring phenology is triggered by the net interaction impacts of winter and spring temperatures.