Thermal sharpening of Sentinel 3 images for water status mapping in large grapevines
1H. Huryna, 2Y. Cohen, 1N. Panov, 1A. Karnieli, 4W. Kustas, 5A. Torres-Rue , 1N. Agam
1. Institute of Agricultural Engineering, Agricultural Research Organization (ARO), Israel; Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede-Boker campus, Israel
2. Institute of Agricultural Engineering, Agricultural Research Organization (ARO), Israel
3. Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede-Boqer campus, Israel
4. Hydrology and Remote Sensing Laboratory, Agricultural Research Services, USDA, Maryland, USA
5. Water Research Laboratory, Utah State University, Utah, USA
High-resolution thermal images (TI) coupled with measured atmospheric conditions have been utilized to map the within-field water status variability. While the spatial resolution of spaceborne TIs is not suitable for monitoring crop conditions of individual fields, the spatial resolution of spaceborne VIS-NIR images is often fine enough for field and intra-field scale applications. Consequently, thermal sharpening techniques have been developed to sharpen TI to the spatial resolution of the VIS-NIR images. Here we propose to utilize the new satellites Sentinel 2 and 3 (S2, S3) which have revisit time of 2-3 days and sharpen the S3 TI images using VIS-NIR images from S2 to produce high resolution TIs at a frequent revisit time suitable for water status monitoring of agricultural fields. Two thermal sharpening methods, namely the DMA (data mining approach , Gao et al., 2002) and a modified TsHARP (Agam et al., 2007) were applied on two types of imagery data: 1) simulated Landsat images based on airborne thermal and VIS-NIR images taken above a large vineyard in the Central Valley in California (two different dates in 2016) and 2) S2 and S3 images of agricultural regions (one around the vineyard in Gallo valley; July, 2017). The S2 and S3 images were selected such that their time acquisition would match the Landsat passage time in order to validate the sharpening results.
With starting spatial resolutions of 120 and 60 m of the simulated data to targeted resolutions of 60 and 30 m, respectively, RMSE of 1.3 – 1.5 Celsius degrees was obtained by the two methods, where the TsHARP consistently obtained slightly better RMSEs. S3 images were sharpened by S2 images from 1 km to 240, 120 and 60 m. Comparison with Landsat images, yielded RMSE of 1.0-1.7 Celsius degrees for images taken from different dates.
To our knowledge, there is no previous publication on the use of satellite thermal sharpening for vineyards. These initial results demonstrate that thermal sharpening methods could be used to map temperature variability in large agricultural areas and could be adequate for water status variability mapping in large vineyards.
