The atmosphere exhibits variability across different time scales. Currently, in the field of atmospheric science, statistical filtering is one of the most widely used methods for extracting signals on certain time scales. However, signal extraction based on traditional statistical filters may be sensitive to missing data points, which are particularly common in meteorological data. To address this issue, this study applies a new type of temporal filters based on a one-dimensional convolution neural network (1D-CNN) and examines its performance on reducing such uncertainties. As an example, we investigate the advantages of a 1D-CNN bandpass filter in extracting quasi-biweekly-to-intraseasonal signals (10–60 days) from temperature data provided by the Hong Kong Observatory. The results show that the 1D-CNN achieves accuracies similar to a 121-point Lanczos filter. In addition, the 1D-CNN filter allows a maximum of 10 missing data points within the 60-point window length, while keeping its accuracy higher than 80% (R² > 0.8). This indicates that the 1D-CNN model works well even when missing data points exist in the time series. This study highlights another potential for applying machine learning algorithms in atmospheric and climate research, which will be useful for future research involving incomplete time series and real-time filtering.

Precipitation events have been occurring more frequently in the hyper-arid region of the Taklamakan Desert (TD) under recent climate change. However, in this water-limited environment, the microphysical characteristics of precipitation, as well as their link to rainfall intensity, remain unclear. To address this, this study utilizes dual-frequency precipitation radar (DPR) data of the Global Precipitation Measurement (GPM) satellite from 2014 to 2023 to analyze the microphysical characteristics of different precipitation types (stratiform and convective) in the TD during the summer. The results show that liquid water path (LWP) is a key factor influencing precipitation type: when LWP is insufficient, stratiform precipitation is more likely to occur (84.1%), while convective precipitation is difficult to occur (15.9%). Microphysical process analysis indicates that in convective precipitation, abundant low-level moisture leads to the growth of liquid particles primarily through the collision–coalescence process (59.7%), resulting in larger raindrop diameters (1.7 mm) and lower concentrations (31.9 mm⁻¹ m⁻³). In contrast, stratiform precipitation, with limited LWP, primarily involves the melting and breaking-up of high-level ice-phase particles, leading to smaller raindrop diameters (1.2 mm) and higher concentrations (34.3 mm⁻¹ m⁻³). The warm rain process plays a significant role in raindrop formation in both types of precipitation. The greater (lesser) the amount of LWP, the larger (smaller) the contribution of collision–coalescence (break-up) processes, and the larger (smaller) the raindrop diameter and precipitation intensity.

Global surface warming is recognized for its spatial asymmetry, which drives the complexity of climate change impacts. However, solely relying on average surface temperature change (\({\Delta T}_{{mean}}\)) as an indicator of climate change may obscure this asymmetry. Here, going beyond \({\Delta T}_{{mean}}\), we quantify the long-term change in global warming asymmetry and its implications for climate mitigation. Results reveal a significant increase in warming asymmetry (0.18–0.29 °C³/decade) since 1900. Among 58 IPCC reference regions, the Arctic Ocean, Northwest North America, and Russian Regions contribute only 48% of this trend, with other Northern Hemisphere mid-high latitudes accounting for an additional 35%. Climate models project an emission-dependent nonlinear acceleration of warming asymmetry, with trends under SSP 5-8.5 being approximately 3–8 times greater than those under SSP2-4.5 and SSP1-2.6. Consequently, the global surface temperature change deviates from its original Gaussian distribution toward a highly skewed bimodal distribution. This increasing asymmetry implies a widening “divergence” within the climate system that is dependent on emission levels, including the gap between the \({\Delta T}_{{mean}}\) and those observed across the majority of regions. Without rigorous emission reductions, this gap could foster cognitive biases in public awareness toward climate change, undermining public acceptance of climate mitigation policies.