Comparison of MODIS, SEVIRI and INSAT-3D Land Surface Temperature (LST)

Journal of Radar and Optical Remote Sensing and GIS

مقاله 1، دوره 3، شماره 4، اسفند 2020، صفحه 7-22   

نویسندگان

Mehdi Gholamnia^{* 1}؛ Salman Ahmadi²؛ Reza Khandan³؛ Seyed kazem Alavipanah⁴؛ Ali Darvishi Boloorani⁵؛ Saeid Hamzehe⁵

¹Assistance Professor, Department of surveying Engineering, Sanandaj Branch,Islamic Azad University

²Assistant Professor of Eengineering Faculty, University of Kurdistan

³Faculty of Geography, Department of Remote Sensing and GIS, University of Tehran, Iran

⁴Professor, Faculty of Geography, Department of Remote Sensing and GIS, University of Tehran, Iran

⁵Assistant Professor, Faculty of Geography, Department of Remote Sensing and GIS, University of Tehran, Iran

چکیده

The accuracy of retrieved LST from satellites is of great importance. Among different LST validation methods, a cross-calibration procedure is highly cost-effective and applicable. The Indian National Satellite-3D series (INSAT-3D) and Meteosat Second Generation (MSG) are two geostationary satellites that which provide LST products with high temporal resolution. Considering MODIS as the reference (polar orbit that is onboard Aqua and Terra satellites), the comparison of the LST products of these geostationary satellites was evaluated from 4th March to 1st September 2015. For this purpose mean LST ratios were calculated for both MODIS-Imager (from INSAT-D) and MODIS-SEVIRI. Then the behavior of their mean LST ratio was analyzed for the exciting four major land covers and five elevation classes in the study area. The results showed that Imager data underestimated and overestimated the LST in comparison to MODIS data during the day and night time respectively. The SEVIRI LSTs underestimated the LST in both day and night time in comparison with MODIS products. In order to model the discrepancies between MODIS-Imager and MODIS-SEVIRI, for each land cover a multilinear regression model was fitted based on slope, aspect, azimuth, and View Zenith Angle (VZA). The results showed that barren, Shrub, grass, and cereal crops had low RMSEs in model fitting, respectively.

کلیدواژه‌ها

LST؛ Geostationary Satellite؛ Land cover؛ Elevation؛ Remote Sensing