Projection of Low Cloud Variation Through Robust Meteorological Linkage and Its Comparison With CMIP6 Models at the SACOL Site

Low clouds significantly influence Earth's energy budget by reflecting solar radiation. Consequently, inadequate representation of these clouds in models introduces the largest uncertainty in predicting future climate change. This study investigates low cloud cover (LCC) variation using 6 years (2014–2019) of high-precision ground-based Ka-band Zenith Radar (KAZR) observations at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL). We analyze the relationship between observed low cloud properties and four large-scale meteorological factors: 700 hPa relative humidity, estimated inversion strength, low-level wind shear, and 700 hPa vertical velocity. These factors are identified as key parameters influencing low cloud evolution over this semi-arid region. We utilize principal component analysis to integrate these parameters into a single meteorological predictor (PC1) and establish a robust linkage between meteorological conditions and low cloud properties. By comparing LCC fluctuations derived from the meteorological factors with those directly simulated by models over the same period, we assess the projected LCC trends under various carbon emission scenarios. Contrary to the declining LCC projected by CMIP6 models outcomes, the LCC form PC1 shows a rising tendency by 2100 under global warming. This discrepancy implies that CMIP6 models may exaggerate the extent of future warming at the SACOL site. Our approach can be applied to a broader global distribution of low clouds to examine the differences between low cloud variations constrained by meteorological fields and those from direct model simulations. This will enhance our understanding of low cloud feedback on future climate change.