Effect of the altitudinal climate change on growing season length for deciduous broadleaved forest in southwest China

The altitudinal changes of plant phenology in response to climate change remain poorly understood in subtropical mountain areas. Using the satellite phenology and climate dataset (temperature, precipitation and solar radiation) from 2001 to 2016 in southwest China, we analyzed the spatiotemporal changes of climate and phenological characteristics of the growing season length (LOS), start of the growing season (SOS) and end of the growing season (EOS) for the deciduous broadleaf forest (DBF). Results show that LOS was shortened by 25 and 15.2 days/km rise in elevation, respectively, using two regression methods based on “Hopkins' bioclimatic law” (expressing LOS as a function of altitude, latitude and longitude) and altitudinal mean annual LOS. The majority of the shortened LOS towards high elevations was attributed to the postponed SOS and the advanced EOS as the elevation is higher and lower than 2.2–2.3 km, respectively. The recent climate warming has significantly prolonged LOS in the entire DBF area. This increase in LOS differs with altitude due to altitudinal heterogeneity of climate change. In the cold high mountain environment, changes of phenological parameters are more sensitive to climate warming, characterized by a significantly advanced SOS, postponed EOS and prolonged LOS driven by spring and autumn warming. In the warm environment of the low elevation areas, changes of phenological parameters are relatively smaller even though the temperature rise is greater than that in the cold high mountains. Furthermore, winter wetting can significantly weaken the advanced SOS and prolonged LOS at lower elevations in the warm south, but winter drying and declining solar radiation in spring can enhance the advanced SOS and prolonged LOS at the extremely high elevations in the cold north. These results highlight the critical need to include altitudinal heterogeneity when assessing phenological changes from remote sensing platforms