TEMPORAL-SPATIAL CHANGES IN AEROSOLS DURING WINTER HAZE OF WUHAN： A TYPICAL HIGH-HUMID INLAND CITY IN CENTRAL CHINA

Accurate wind profile estimation, especially for the lowest few hundred meters of the atmosphere, is of great significance for the weather, climate, and renewable energy sector. We propose a novel method that combines the power-law method with the random forest algorithm to extend wind profiles beyond the surface layer. Compared with the traditional algorithm, this method has better stability and spatial applicability and can be used to obtain the wind profiles on different land cover types.

To better understand the Asian aerosol environment, we studied distributions and trends of aerosol with different sizes and types. Over the past 2 decades, dust, sulfate, and sea salt aerosol decreased by 5.51 %, 3.07 %, and 9.80 %, whereas organic carbon and black carbon aerosol increased by 17.09 % and 6.23 %, respectively. The increase in carbonaceous aerosols was a feature of Asia. An exception is found in East Asia, where the carbonaceous aerosols reduced, owing largely to China's efforts.

Wind energy is one of the most essential clean and renewable forms of energy in today’s world. However, the traditional power law method generally estimates the hub-height wind speed by assuming a constant exponent between surface and hub-height wind speeds. This inevitably leads to significant uncertainties in estimating the wind speed profile. To minimize the uncertainties, we here use a machine learning algorithm known as random forest to estimate the wind speed at hub height.

Aerosol parameter retrievals have always been a research focus. In this study, we used an advanced aerosol algorithms (GRASP, developed by Oleg Dubovik) to test the ability of DPC/Gaofen-5 (the first polarized multi-angle payload developed in China) images to obtain aerosol parameters. The results show that DPC/GRASP achieves good results (R > 0.9). This research will contribute to the development of hardware and algorithms for aerosols

The vertical distribution of the aerosol extinction coefficient (EC) measured by lidar systems has been used to retrieve the profile of particle matter with a diameter of less than 2.5 μm (PM_2.5). However, the traditional linear model cannot consider the influence of multiple meteorological variables sufficiently, which then causes low inversion accuracy. In this study, the machine learning algorithms which can input multiple features are used to solve this constraint.