基于MODIS数据验证分析雾参数反演算法和影响因素

Verification of the retrieval algorithm and analysis of influencing factors of fog physical parameters based on MODIS data

基于遥感影像对雾进行定量分析是近年来的研究热点之一。地面能见度、雾顶高度、垂直总水汽含量和有效粒子半径是对雾进行定量分析的重要参数。选取MODIS传感器数据为数据源,以发生于江苏省西南部的一次冬季平流辐射雾为研究对象,结合来自南京信息工程大学的地面观测数据,对基于大气辐射传输方程模型和SBDART辐射传输的雾物理参数反演算法所反演的地面能见度和雾顶高度进行地面实测数据验证;分析影响这2个物理参数精度变化的原因;在序列影像雾物理参数反演结果基础上,进一步结合地面高程和植被指数对雾物理参数影响因素进行分析。结果表明,该次反演的雾能见度和雾顶高度与地面观测结果相关系数分别为0.908和0.980,由于SBDART模型设置的参数以及BP神经网络训练次数等原因,夜间反演的能见度值偏大;此次大雾物理参数与地面高程和植被指数之间具有一定的关系,地面高程与能见度呈正相关,植被指数和水汽含量呈正相关,与雾有效粒子半径和光学厚度呈负相关;雾区物理参数易受水体和土地利用等影响。

Retrieving fog physical parameters becomes one of the major hot spots of study in recent years based on remote sensing data. The visibility, top height of fog, effective particle radius, and liquid water path （LWP） of fog are the fundamental physical parameters for fog monitoring. In this study, the authors retrieved fog physical parameters from southwest Jiangsu Province according to the path model of fog radioactive phenomena and SBDART based on the MODIS images. The authors verified the visibility and top height of fog according to t-he data from the Nanjing Information Engineering University and analyzed the influencing factors for the changes of physical parameters. The results showed that the correlation coefficient of visibility and top height of fog was 0.908 3 and 0. 980 7, indicating that the retrieval of remote sensing data was feasible. The study also found positive correlations between the fog physical parameters, the surface elevation and vegetation index. The vegetation index was negatively correlated with the radius and optical depth and positively correlated with the liquid water. There was a positively correlation between the visibility and the surface elevation.