Knowledge Resource Center for Ecological Environment in Arid Area
| 基于Himawari-8/AHI数据的气溶胶光学厚度反演和沙尘监测研究 | |
| 其他题名 | Aerosol Retrieval and Dust Detection over Land from Himawari-8/AHI data |
| 佘璐 | |
| 出版年 | 2018 |
| 学位类型 | 博士 |
| 导师 | 薛勇 |
| 学位授予单位 | 中国科学院大学 |
| 中文摘要 | 陆地上空的大气气溶胶遥感反演一直是大气环境研究的热点问题，也是一项具有挑战性的科学问题。多年来学者们对气溶胶反演算法进行了大量的探索，然而目前国际上成熟的算法和产品大都是基于极轨卫星观测数据，其观测频次有限，单景影像覆盖范围较低，难以满足对时空变化巨大的大气气溶胶的研究需求。新一代静止卫星Himawari-8搭载的新型传感器Advanced Himawari Imagery （AHI）具有极高的观测频次和多波段设置，对于气溶胶的时空变化研究尤其是沙尘气溶胶的动态监测具有极大的优势。本文针对AHI数据特征，构建了基于多时相观测和最优估计技术的气溶胶光学厚度（Aerosol optical depth，AOD）反演算法（Aerosol Retrieval algorithm based on multiple observations and Optimal Estimation technique，AR-OE），并利用AHI热红外观测数据以及反演的气溶胶光学厚度构建了针对中国大陆地区的沙尘识别方法和沙尘指数。本文从大气辐射传输模型出发，通过耦合地表二向性反射率分布函数（Bidirectional Reflectance Distribution Function，BRDF）构建了非朗伯前向模型。基于地表BRDF特性稳定即短时间内不变的假设，通过联合使用多时相的卫星观测，并引入地表BRDF形状先验知识以约束地表变量，构建了基于最优估计技术的气溶胶反演算法。在气溶胶模型的选取上，本文利用地基AErosol RObotic NETwork（AERONET）站点长时间的气溶胶特性参数观测数据，利用聚类分析方法构建了代表研究区典型气溶胶特征的六种气溶胶类型，用于支撑气溶胶反演。为了验证算法的可行性，本文利用2016年3月-2017年2月期间共12个月的AHI数据对AR-OE反演算法进行测试，并利用研究区域内48个地基AERONET站点的观测数据对AOD反演结果进行了系统验证。验证结果表明，AR-OE算法对于不同类型地表下垫面均有较好的适用性，能应用于不同的大气状况的AOD反演。与MODerate-resolution Imaging Spectroradiometer（MODIS） Collection 6版气溶胶产品的比较显示，两者反演的AOD空间分布具有良好的一致性；与日本气象局（Japan Meteorological Agency，JMA）官方发布的基于AHI数据反演的Level 2气溶胶产品（JMA-AOD）相比，AR-OE反演结果（OE-AOD）具有更大的覆盖度，尤其是对于亮地表地区和太阳角度较大的情况。与AERONET的AOD观测的定量比较表明，OE-AOD具有较高精度， OE-AOD与AERONET-AOD两者相关系数达到0.88，回归方程为AODAHI=0.97*AODAERONET+0.02，约69.6%的反演结果落入±0.02*AODAERONET±0.05的期望误差（expect error，EE）以内。同时段的JMA-AOD与AERONET-AOD的相关系数仅为0.82，回归方程为AODAHI=0.60*AODAERONET+0.08，仅59.1%的反演结果落入EE内。此外，对反演结果的分区域验证发现在不同区域算法的反演精度存在明显差异：华北地区和朝鲜半岛地区反演结果具有较高精度，华北地区的反演结果与AERONET观测结果的相关系数达0.94，但是表现出轻微的系统高估现象；而朝鲜半岛区域相关系数为0.87，有77%的反演结果在EE以内。东南亚地区由于地表先验知识存在较大不确定性，该地区AOD反演精度相比于其他地区偏低，并且有轻微的低估。区域性的低估（高估）主要是由气溶胶类型的不准确估计导致的。总体而言，AOD反演结果的验证表明AR-OE算法具有较广泛的适用性和较为理想的精度。另一方面，基于我国沙漠面积大，沙尘天气频繁的环境现实，本文利用AHI热红外波段观测数据，通过沙尘气溶胶的光谱辐射特性分析，构建了基于多个亮温差组合的阈值沙尘识别算法，并基于地表亮温存在区域差异的考虑，提出了随归一化植被指数（NDVI）与地表高程变化的动态阈值。在沙尘识别的基础之上，本文进一步利用可见光与近红外波段的反射率，热红外波段的亮温，以及反演的气溶胶光学厚度构建了沙尘指数以评估沙尘强度。沙尘识别结果与地基AERONET站点观测结果比较显示，沙尘识别的精度达到84%，正确率为77%。沙尘指数与地基能见度数据的比较显示两者拟合的回归方程为要y=1.67*exp(-0.25*x)+0.47，相关系数达到0.81。表明本文提出的沙尘识别方法能够有效的识别沙尘并对沙尘强度进行合理的评估。 |
| 英文摘要 | Atmospheric aerosol has significant impacts on Earth's radiation balance, hydrological cycle and biogeochemical cycles. As one of the most important type of aerosol in mass and aerosol optical depth (AOD), dust also exert severe impacts on humans and environment. The new generation geostationary satellite Himawari-8 carring an Advanced Himawari Imager (AHI), which has high frequency of observation and multiple channels, makes it unique advantages to monitor atmospheric aerosol distribution and dust transport. In this dissertation, we presented an aerosol retrieval algorithm based on multiple AHI observations and optimal estimation technique (AR-OE). In addition, a dust detection and intensity estimation methods were presented based on the brightness temperature (BT) of thermal channels of AHI observation. The aerosol retrieval algorithm was based on a non-Lambertian forward model coupled with a surface bidirectional reflectance factor (BRF) model. Our retrieval is based on the assumption that the surface bidirectional reflective properties are invariable during a short time period, while aerosol properties change. The BRDF shape were obtained from MODIS BRDF/Albedo products and employed as prior knowledge. Using multiple AHI observations, the AOD and BRDF coefficients were jointly retrieved using an optimal estimation method.The AOD retrieval results were validated using ground-based measurements (AErosol RObotic NETwork (AERONET) sites) and were cross-compared with satellite product (MODIS Collection 6.0 AODs). A total of one year AHI data were used for this validation and comparison. The validation of the AOD with AERONET measurements shows a high correlation coefficient: R=0.88, RMSE=0.17, the liner regression function is AODAHI=0.97*AODAERONET+0.02, and approximately 69.6% AOD retrieval results within the expected error (EE) of (±0.02*AODAERONET±0.05). In addition, Level 2 AOD product provided by Japan Meteorological Agency (JMA-AOD) were also validated against with AERONET measurements. The liner regression function between JMA-AOD and AERONET-AOD is AODAHI=0.60*AODAERONET+0.08, and the correlation coefficient is 0.82, and 59.1% AOD retrieval results within the EE. JMA-AOD shows lower accuracy than OE-AOD according to the validation against AERONET-AOD. In addition, OE-AOD were also compared with MODIS Collection 6.0 AODs, and it shows a high consistency. As there are huge areas of desert in China, which emit huge amount of mineral dust every year, causing abundant dust storms, especially during spring seasons. It is significant important to monitor dust aerosol with high frequency. In this paper, simple dust detection and intensity estimation methods using AHI data are developed. Based on the differences of thermal radiation characteristics between dust and other typical objects, brightness temperature difference (BTD) among four channels are used together for dust detection. Considering thermal radiation variation of dust particles over different land cover types, a dynamic threshold scheme for dust detection is adopted. An enhanced dust intensity index (EDII) is developed based on the reflectance of visible/near-infrared bands, BT of thermal-infrared bands, and retrieved AOD and is applied to the detected dust area. The dust detection results were compared quantitatively with dust identification results from the AERONET AOD and ?ngstr?m exponent, achieving a total dust detection accuracy of 84%. A good agreement is obtained between EDII and the visibility data from National Climatic Data Center ground measurements, with a correlation coefficient of 0.81, indicating the effectiveness of EDII in dust monitoring. |
| 中文关键词 | 气溶胶光学厚度 ; 地表二向反射率分布函数 ; 最优估计 ; 沙尘识别 ; 沙尘指数 |
| 英文关键词 | Aerosol optical depth (AOD) bidirectional reflectance distribution function (BRDF) optimal estimation (OE) dust detection dust index |
| 语种 | 中文 |
| 国家 | 中国 |
| 来源学科分类 | 地图学与地理信息系统 |
| 来源机构 | 中国科学院遥感与数字地球研究所 |
| 资源类型 | 学位论文 |
| 条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/288207 |
| 推荐引用方式 GB/T 7714 | 佘璐. 基于Himawari-8/AHI数据的气溶胶光学厚度反演和沙尘监测研究[D]. 中国科学院大学,2018. |
| 条目包含的文件 | 条目无相关文件。 | |||||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [佘璐]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [佘璐]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [佘璐]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
