Carbon stock assessment of date palm using remote sensing coupled with field-based measurements in Abu Dhabi (United Arab Emirates)

This study aims at developing a remote sensing based model to estimate carbon stock (CS) in date palm plantations in the United Arab Emirates (UAE). Data from Landsat 8 OLI were used to assess the correlation between spectral reflectance and different vegetation indices on one side, and aboveground biomass (AGB) derived from ground measurements on the other. AGB and CS (ton ha⁻¹) were estimated using traditional destructive methods of standard sampling techniques and allometric equations developed in a previous study for date palms in the area. The relationships between the estimated AGB and parameters derived from remote sensing (RS) data were tested using single and multiple linear regression analysis. The results indicated a significant correlation with certain RS parameters. For mature palms class alone (>10 years), the correlation with single bands was only significant with SWIR1 and SWIR2 while the correlation was significant with all tested VI’s except for TCB vegetation index. A combination of bands R, SWIR1, and SWIR2 improved the determination of this class to a coefficient of determination (R²) value of 0.961. However, for the medium and young age classes (10–5 and less than 5 years), the correlation was not significant (with the exception of SWIR1 and TCB index for medium class), where the use of higher spatial resolution can be a good alternative. On the other hand, for mixed ages (young, medium and mature palms), the strongest correlations were found using SWIR2 single band and the SR vegetation index; having R² values of 0.753 and 0.871, respectively. The R² was improved to 0.952 by using stepwise multiple regression combining DVI, NDI, and RVI vegetation indices. Finally, results obtained showed that CS represented 53.87% of the total AGB in date palms. Subsequently, the average amount of carbon stock (CS) for both mature and mixed classes was calculated at 15.81 and 10.3 ton ha⁻¹ respectively.