Spatio-temporal dynamic evaluation of agricultural drought based on a three-dimensional identification method in Northwest China

Drought evolution is a complex dynamic process, showing continuous characteristics in space-time. Currently, the researches on drought dynamics from multiple dimensions are limited and should be improved further. Therefore, a three-dimensional identification method was adopted to investigate the spatiotemporal patterns of classified drought events (namely two-month drought events, within-year drought events, and cross-year drought events) in northwest China during1961-2018. Furthermore, more spatial variables of drought event, such as start and end location, drought orientation and rotation, were extracted to describe the dynamic migration trajectories of drought events. Significantly, drought rotation helped to reveal the forcing mechanisms (such as the influence of climate and land surface) behind their spatial development. The results showed that a total of 169 drought events were identified over 1961-2018, and drought variables (duration, area, severity, and migration distance) represented fluctuating decreasing trend with the rate of - 0.135/10a, - 0.006/10a, - 0.16/10a, and - 5.67/ 10a. The dynamic migration process of individual drought event was objectively and effectively depicted from a three-dimensional perspective, which was consistent with the historical drought records in study area. Regarding the classified events, drought magnitude was more severe in 1961-1980 than that in 1981-2018, and three primary migration patterns (east-to-east, west-to-west, and south-to-south) were identified in northwest China. Within-year drought events mostly started in the first three quarters and finished in the second half year, whilst the situation was reversed for the cross-year drought events. Spatially, Alxa League of Inner Mongolia and northwest Qinghai Province were two main drought centers. Additionally, drought events characterized by clockwise rotation migration were mostly distributed in the plateau and southeast climate zones, and that was anticlockwise rotation in the westerly climate zone. Noticeably, the east-to-east pathways were mainly located in the east part of the study area, while the west-to-west and south-to-south trajectories were dominated in the west part.