The Tibetan Plateau (TP) has experienced an overall rapid warming and moistening; however, the knowledge of TP climate regionalization and of its spatial-temporal variations is far behind the rapid climate change and subsequent environmental responses. To address the threats of frequent compound processes, like heatwaves or flash droughts, we analyze the interaction of atmospheric processes and climate subsystems, propose a novel data-driven climate diagnostics approach, and generate a time series of multi-scale local climate zonings, which characterize the spatial-temporal variations of TP climate and provides an in-depth understanding of TP climate change from 1979 to 2018. The interpretation of this data driven approach is supported by Holdridge’s life-zones, Budyko’s physical framework of geobotanical biomes driven by the surface fluxes of water supply (precipitation) and water demand (net radiation), and the Köppen phenomenological climate classification. Three different local climate patterns are identified: (i) The main driver of local climate change in the Qiangtang area (central-western TP) is shifting from water supply to demand dominance. (ii) In the Qaidam area (north-eastern TP), the humid region expands accompanied with a contracting arid region; this trend of warming and moistening expands from the east westwards. (iii) Hengduan Mountains area (south-eastern TP) becomes warmer and wetter but with frequent local climate variations.

青藏高原整体快速增暖湿润；然而，对青藏高原气候区划及其时空变化的认识远远落后于快速的气候变化和随后的环境响应。为了解决热浪或突发干旱等频繁复合过程的威胁，我们分析了大气过程和气候子系统的相互作用，提出了一种新颖的数据驱动的气候诊断方法，并生成了一系列多尺度局部气候分区的时间序列，描绘了青藏高原气候的时空变化，并提供了对1979年至2018年青藏高原气候变化的深入了解。这种数据驱动方法的解释得到了Holdridge的生命带、Budyko的地表驱动的地球植物生物群落物理框架的支持供水通量（降水）和需水量（净辐射），以及柯本现象学气候分类。确定了三种不同的局地气候模式：（i）羌塘地区（高原中西部）局地气候变化的主要驱动因素正在从供水主导转向需求主导。(ii) 柴达木地区（高原东北部）湿润区扩大，干旱区缩小；这种变暖和湿润的趋势是从东向西扩展的。(iii) 横断山脉地区（高原东南部）变暖、湿润，但局部气候变化频繁。