Read here. Climate models used by the IPCC and the national climate agencies are notoriously bad at making predictions of precipitation, including snow during the Northern Hemisphere winter months. The primary resaon for this prediction failure is the lack of input from natural ocean and atmospheric cycles. The models' perfect ignorance of the Atlantic Multidecadal Oscillation (AMO) is a case in point.
Shen et al. conducted research on snow coverage across the Tibetan Plateau over the last 200 years. They confirmed that snow variation was closely associated with the AMO, which climate models totally ignore.
"A paper published in the journal Geophysical Research Letters finds a strong influence of shifts in the Atlantic Multidecadal Oscillation (AMO) on changes in snow cover of the Tibetan Plateau over the past 200 years. Major shifts occurred in the 1840s, 1880s, 1920s, and 1960s...Ocean oscillations such as the AMO are not incorporated in climate models, but nonetheless have large effects upon climate change as demonstrated by this paper and others...also finds not surprisingly that cold phases are associated with more snow and warm phases with less snow...A plausible mechanism linking the North Atlantic climate to Asian monsoon is presented." [Caiming Shen, Wei-Chyung Wang, Gang Zeng 2011: Geophysical Research Letters]
Additional oscillation-cycle, failed-prediction and peer-reviewed postings.