Sand and dust is a typical weather disaster which outbreaks in arid and semi-arid areas globally.This natural phenomenon， which is the result of stormy winds， raises a lot of dust from desert surfaces and decreases visibility to less than 1 km.The dust aerosol generated from dust storm dominates the aerosol loading in the troposphere and has comprehensive impacts on the global environment， weather， climate and ecology.Monitoring sand and dust from space using satellite remote sensing has become one of the most important issues in this field.However， sand and dust is difficult to accurately characterize by using single-band and linear models.Feng Yun-4A （FY-4A） imagery provides a good data source for timely and accurate monitoring of sand and dust.The machine learning models are important tools in sand and dust monitoring and forecasting.In this paper， the Normalized Difference Dust Index （NDDI）， Random Forests （RF） and Convolutional Neural Networks （CNN） were employed to monitor sand and dust based on the Advanced Geostationary Radiation Imager （AGRI） of geostationary FY-4A meteorological satellite in the Tarim Basin.The results showed that， sand and dust can be identified by NDDIAGRI thresholds calculated using AGRI data.The determination of the NDDIAGRI thresholds were obtained through statistical analysis of pixels， but it is necessary to take different thresholds for different times AGRI data.There are some identification errors in the cross region of cloud and land， and some vegetation coverage and desert by the NDDIAGRI thresholds.The values of Precision， Recall， and F1-score of testing samples were all 100%； and the accuracy of cross validation of training samples was 99.5% for the sand and dust model of RF.The Loss and Accuracy in the estimation obtained using the CNN algorithm were about 0.1% and 99.9%， respectively， versus the training samples and testing samples.Both RF and CNN models have the ability and robustness to be used in sand and dust monitoring.The efficiency of two models had been checked using new dust events.Results show that the CNN algorithm preforms better than RF algorithm in identifying the junction of dust and non-dust.The RF and CNN algorithm have identification errors in some parts of sand and dust monitoring process， such as the mixed area of dust and clouds， and the Gobi area.The research results of this paper provide an important basis application of machine learning combined with FY-4A meteorological satellite data to monitor sand and dust operational.
Hong JIANG
,
Qing HE
,
Xiaoqing ZENG
,
Ye TANG
,
Keming ZHAO
,
Xinying DOU
. Sand and Dust Monitoring Using FY-4A Satellite Data based on the Random Forests and Convolutional Neural Networks[J]. Plateau Meteorology, 2021
, 40(3)
: 680
-689
.
DOI: 10.7522/j.issn.1000-0534.2020.00060
[1]Breiman L， 2001.Random forests［J］.Machine Learning， 45（1）： 5-32.DOI： 10.1023/A： 1010933404324.
[2]Fukushima K， 1980.Neocognitron： A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position［J］.Biological Cybernetics， 36（4）： 193-202.DOI： 10.1007/bf00344251.
[3]Kaufman Y J， Tanré， Didier， al et， 2001.Absorption of sunlight by dust as inferred from satellite and ground-based remote sensing［J］.Geophysical Research Letters， 28（8）： 1479-1482.DOI： 10.1029/2000GL012647.
[4]Lecun Y， Bottou L， Bengio Y， al et， 1998.Gradient-based learning applied to document recognition［J］.Proceedings of the IEEE， 86（11）： 2278-2324.DOI： 10.1109/5.726791.
[5]Legrand M， Plana-Fattori A， N’Doumé C， 2001.Satellite detection of dust using the IR imagery of Meteosat.1.Infrared difference dust index［J］.Journal of Geophysical Research， 106（16）： 18251-18274.DOI： 10.1029/2000jd900749.
[6]Ma Y， Gong W， Wang P， al et， 2011.New dust aerosol identification method for spaceborne lidar measurements［J］.Journal of Quantitative Spectroscopy and Radiative Transfer， 112（2）： 338-345.DOI： 10.1016/j.jqsrt.2010.08.004.
[7]Murguia M I C， Yearim Q H， Rivas-Perea P， al et， 2011.Dust storm detection using a Neural Network with uncertainty and ambiguity output analysis［J］.Pattern Recognition， 6718： 305-313.DOI： 10.1007/978-3-642-21587-2_33.
[8]Norton C C， Mosher F R， Hinton B， al et， 1980.A model for calculating desert aerosol turbidity over the oceans from geostationary satellite data［J］.Journal of Applied Meteorology， 19（6）： 633-644.DOI： 10.1175/1520-0450（1980）019<0633： AMFCDA>2.0.CO； 2.
[9]Qu J J， Hao X， Kafatos M， al et， 2006.Asian dust storm monitoring combining terra and aqua MODIS SRB measurements［J］.IEEE Geoscience and Remote Sensing Letters， 3（4）： 484-486.DOI： 10.1109/lgrs.2006.877752.
[10]Samadi M， Boloorani A D， Alavipanah S K， al et， 2014.Global dust detection index （GDDI）； a new remotely sensed methodology for dust storms detection［J］.Journal of Environmental Health Science Engineering， 12（1）： 20-34.DOI： 10.1186/2052-336X-12-20.
[11]Shenk W E， Curran R J， 1974.The Detection of dust storms over land and water with satellite visible and infrared measurements［J］.Monthly Weather Review， 102（12）： 830-837.DOI： 10.1175/1520-0493（1974）1022.0.CO； 2.
[12]Souri A H， Vajedian S， 2015.Dust storm detection using random forests and physical-based approaches over the Middle East［J］.Journal of Earth System Science， 124（5）： 1127-1141.DOI： 10.1007/s12040-015-0585-6.
[13]Steven A A， 1989.Using the radiative temperature difference at 3.7 μm and 11 μm to tract dust outbreaks［J］.Remote Sensing of Environment， 27（2）： 129-133.DOI： 10.1016/0034-4257（89）90012-6.
[14]Zhang P， Lu N M， Hu X Q， al et， 2006.Identification and physical retrieval of dust storm using three MODIS thermal IR channels［J］.Global and Planetary Change， 52（1/4）： 197-206.DOI： 10.1016/j.gloplacha.2006.02.014.
[15]Francois， 2018.Python深度学习［M］.北京： 人民邮电出版社， 11-15.
[16]曹广真， 张鹏， 胡秀清， 等， 2016.静止与极轨气象卫星监测沙尘的融合算法研究［J］.气象科技进展， 6（1）： 116-119.DOI： 10. 3969/j.issn.2095-1973.2016.01.018.
[17]崔林丽， 郭巍， 葛伟强， 等， 2020.FY-4A卫星云顶参数精度检验及台风应用研究［J］.高原气象， 39（1）： 196-203.DOI： 10.7522/j.issn.1000-0534.2019.00065.
[18]韩涛， 李耀辉， 郭铌， 2005.基于EOS/MODIS资料的沙尘遥感监测模型研究［J］.高原气象， 24（5）： 757-764.
[19]胡秀清， 卢乃锰， 张鹏， 2007.利用静止气象卫星红外通道遥感监测中国沙尘暴［J］.应用气象学报， 18（3）： 266-275.DOI： 10.11898/1001-7313.20070302.
[20]黄守友， 徐国强， 2020.FY4A的LMIE闪电数据对云信息初始化的影响及数值试验［J］.高原气象， 39（2）： 378-392.DOI： 10. 7522/j.issn.1000-0534.2019.00110.
[21]李富刚， 李仑格， 林春英， 等， 2008.西北地区气溶胶的源和汇与沙尘暴研究综述［J］.中国沙漠， 28（3）： 586-591.DOI： 1000-694X（2008）03-0586-06.
[22]李海萍， 熊利亚， 庄大方， 2003.中国沙尘灾害遥感监测研究现状及发展趋势［J］.地理科学进展， 22（1）： 45-52.DOI： 10.3969/j.issn.1007-6301.2003.01.006.
[23]李令军， 高庆生， 2001.2000年北京沙尘暴源地解析［J］.环境科学研究， 14（2）： 1-3.DOI： 10.13198/j.res.2001.02.3.lilj.001.
[24]毛东雷， 蔡富艳， 赵枫， 等， 2018.塔克拉玛干沙漠南缘近4年沙尘天气下的气象要素相关性分析［J］.高原气象， 37（4）： 1120-1128.DOI： 10.7522/j.issn.1000-0534.2018.00010.
[25]石广玉， 赵思雄， 2003.沙尘暴研究中的若干科学问题［J］.大气科学， 27（4）： 591-606.DOI： 10.3878/j.issn.1006-9895.2003. 04.11.
[26]史莹莹， 张镭， 田鹏飞， 等， 2018.黄土高原半干旱区沙尘气溶胶光学和微物理特征［J］.高原气象， 37（1）： 286-295.DOI： 10. 7522/j.issn.1000-0534.2017.00024.
[27]杨志华， 张旭， 毛炜峄， 2014.基于FY-3B/VIRR数据的新疆沙尘天气遥感监测应用研究［J］.沙漠与绿洲气象， 8（5）： 48-52.DOI： 10.3969/j.issn.1002-0799.2014.05.009.
[28]张宝林， 2018.沙尘天气及沙尘气溶胶影响的研究进展［J］.气象科技进展， 8（1）： 22-27.DOI： 10.3969/j.ossn.2095-1973. 2018.01.003.
[29]张芝娟， 陈斌， 贾瑞， 等， 2019.全球不同类型气溶胶光学厚度的时空分布特征［J］.高原气象， 38（3）： 660-672.DOI： 10.7522/j.issn.1000-0534.2019.00002.
[30]郑新江， 陆文杰， 罗敬宁， 2001.气象卫星多通道信息监测沙尘暴的研究［J］.遥感学报， 5（4）： 300-306.DOI： 0.11834/jrs. 20010410.
