期刊
ATMOSPHERIC MEASUREMENT TECHNIQUES
卷 14, 期 12, 页码 7749-7773出版社
COPERNICUS GESELLSCHAFT MBH
DOI: 10.5194/amt-14-7749-2021
关键词
-
资金
- National Aeronautics and Space Administration [80NM0018D0004]
An improved cloud detection algorithm based on feedforward artificial neural network was presented for Aura Microwave Limb Sounder (MLS). The algorithm showed significant improvement in cloud classification performance compared to the current MLS cloudiness flag. It successfully detected over 93% of cloud-affected profiles with low false positives, providing reliable global cloud cover maps and predictions for cloud top pressure.
An improved cloud detection algorithm for the Aura Microwave Limb Sounder (MLS) is presented. This new algorithm is based on a feedforward artificial neural network and uses as input, for each MLS limb scan, a vector consisting of 1710 brightness temperatures provided by MLS observations from 15 different tangent altitudes and up to 13 spectral channels in each of 10 different MLS bands. The model has been trained on global cloud properties reported by Aqua's Moderate Resolution Imaging Spectroradiometer (MODIS). In total, the colocated MLS-MODIS data set consists of 162 117 combined scenes sampled on 208 d over 2005-2020. A comparison to the current MLS cloudiness flag used in Level 2 processing reveals a huge improvement in classification performance. For previously unseen data, the algorithm successfully detects > 93 % of profiles affected by clouds, up from similar to 16 % for the Level 2 flagging. At the same time, false positives reported for actually clear profiles are comparable to the Level 2 results. The classification performance is not dependent on geolocation but slightly decreases over low-cloud-cover regions. The new cloudiness flag is applied to determine average global cloud cover maps over 2015-2019, successfully reproducing the spatial patterns of mid-level to high clouds seen in MODIS data. It is also applied to four example cloud fields to illustrate its reliable performance for different cloud structures with varying degrees of complexity. Training a similar model on MODIS-retrieved cloud top pressure (p(CT)) yields reliable predictions with correlation coefficients > 0.82. It is shown that the model can correctly identify > 85 % of profiles with p(CT) < 400 hPa Similar to the cloud classification model, global maps and example cloud fields are provided, which reveal good agreement with MODIS results. The combination of the cloudiness flag and predicted cloud top pressure provides the means to identify MLS profiles in the presence of high-reaching convection. (C) 2020 California Institute of Technology. Government sponsorship acknowledged.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据