利用多源數(shù)據(jù)建立GA-BP算法模型估算PM<sub>2.5</sub>的研究

doi:10.12306/2021jms.0014

首頁(yè) > 過(guò)刊瀏覽>2021年第41卷第3期 >314-322. DOI:10.12306/2021jms.0014

利用多源數(shù)據(jù)建立GA-BP算法模型估算PM2.5的研究
DOI:
                        10.12306/2021jms.0014
                    
CSTR:
                        
                    
作者:
                        
                        
                    
作者單位:
作者簡(jiǎn)介:
通訊作者:
中圖分類號(hào):X513
基金項(xiàng)目:國(guó)家重點(diǎn)研發(fā)計(jì)劃項(xiàng)目（2018YFC1506502）；國(guó)家自然科學(xué)基金資助項(xiàng)目（42075073；40975019）

Research on estimating PM_2.5 by building GA-BP algorithm model based on multi-source data

Author:

Affiliation:

Fund Project:

摘要

圖/表

訪問(wèn)統(tǒng)計(jì)

參考文獻(xiàn)

相似文獻(xiàn)

引證文獻(xiàn)

資源附件

文章評(píng)論

摘要:

利用MODIS AOD（Aerosol Optical Depth）3 km分辨率的L2產(chǎn)品并輔以地面氣象測(cè)量站點(diǎn)和環(huán)保監(jiān)測(cè)站點(diǎn)的逐小時(shí)數(shù)據(jù)，對(duì)2017—2018年南京地面各站點(diǎn)進(jìn)行數(shù)據(jù)匹配，分析估算PM_2.5的各相關(guān)組合因子，然后利用GA-BP神經(jīng)網(wǎng)絡(luò)算法構(gòu)建衛(wèi)星數(shù)據(jù)輔以地面氣象數(shù)據(jù)來(lái)估算PM_2.5質(zhì)量濃度的機(jī)器學(xué)習(xí)模型。結(jié)果表明：（1）GA-BP神經(jīng)網(wǎng)絡(luò)算法對(duì)PM_2.5進(jìn)行估算是有效可行的，且比BP效果改善明顯。（2）在多源數(shù)據(jù)的各輸入變量中，選擇AOD變量加輔助變量的GA-BP算法模型共構(gòu)建了9組分季節(jié)試驗(yàn)，其中應(yīng)用在2017年數(shù)據(jù)的試驗(yàn)6最好，表現(xiàn)為秋季>冬季>夏季>春季，秋季R²最大為0.91（RMSE為11.79 μg·m^-3），春季R²最小為0.65（RMSE為8.67 μg·m^-3）。

Abstract:

By using MODIS AOD (Aerosol Optical Depth) Level 2 products with ground spatial resolution of 3 km and supplemented by hourly data from ground weather measurement sites and environmental monitoring sites, the data of Nanjing ground stations from 2017 to 2018 was matched and the relevant combination factors of estimating PM_2.5 were analyzed. Moreover, the GA-BP neural network algorithm was applied to construct machine-learning model of estimating PM_2.5 mass concentration based on satellite data supplemented by ground weather data. Results show that: (1) GA-BP network algorithm is effective and feasible to estimate PM_2.5, and the improving effect is significantly better than BP algorithm. (2) Among the input variables of the multi-source data, the GA-BP algorithm model who select AOD variable and ground ancillary variables constructs nine groups of tests in order to estimate PM_2.5 seasonally. Among them, test six for data of 2017 is the best overall, which performs as autumn>winter>summer>spring. The maximum R² in autumn is 0.91 (RMSE is 11.79 μg·m^-3) while the minimum R² in spring is 0.65 (RMSE is 8.67 μg·m^-3).

參考文獻(xiàn)

相似文獻(xiàn)

引證文獻(xiàn)

引用本文

胡娟,鄭軍,許文龍,邱玉珺,胡方超.利用多源數(shù)據(jù)建立GA-BP算法模型估算PM_2.5的研究.氣象科學(xué),2021,41(3):314-322 HU Juan, ZHENG Jun, XU Wenlong, QIU Yujun, HU Fangchao. Research on estimating PM_2.5 by building GA-BP algorithm model based on multi-source data. Journal of the Meteorological Sciences,2021,41(3):314-322

復(fù)制

文章指標(biāo)

點(diǎn)擊次數(shù):
下載次數(shù):
HTML閱讀次數(shù):
引用次數(shù):

歷史

收稿日期:2020-09-25
最后修改日期:
錄用日期:
在線發(fā)布日期: 2021-07-05
出版日期:

文章二維碼

您是第位訪問(wèn)者
氣象科學(xué) ® 2026 版權(quán)所有
技術(shù)支持：北京勤云科技發(fā)展有限公司

抽查舔水白紧大视频,冲喜娘子超甜哒激情五月,亚洲黄色图片偷拍网站,亚洲三级黄片免费播放,91在线观看喷潮红桃,日韩精品在线观看一区二区,久久久久久久久久久久黄色片子,日产精品99久久久久久久久,欧洲亚洲另类图片小说

引用本文

分享

相關(guān)視頻

文章指標(biāo)

歷史

文章二維碼

抽查舔水白紧大视频,冲喜娘子超甜哒 激情五月,亚洲黄色图片偷拍网站,亚洲三级黄片免费播放,91在线观看喷潮红桃,日韩精品在线观看一区二区,久久久久久久久久久久黄色片子,日产精品99久久久久久久久,欧洲亚洲另类图片小说

引用本文

分享

相關(guān)視頻

文章指標(biāo)

歷史

文章二維碼

抽查舔水白紧大视频,冲喜娘子超甜哒激情五月,亚洲黄色图片偷拍网站,亚洲三级黄片免费播放,91在线观看喷潮红桃,日韩精品在线观看一区二区,久久久久久久久久久久黄色片子,日产精品99久久久久久久久,欧洲亚洲另类图片小说