Determination and modelling of above ground and below ground carbon contents in forest, grassland and agriculture areas with association by using satallite images in some sub-basins of Artvin, Erzurum and Bayburt

In the present study, it was aimed to determine above and below ground carbon contents in some land uses of 11 sub-basins in the Coruh River Basin using satellite images. For this, 49 samples were taken for above ground biomass (AGB), 1196 root samples from 299 different points for below ground biomass (BGB) and 666 samples for soil carbon in the grassland areas. Biomass and carbon contents in other land uses were calculated using related coefficients in the literature. The reflectance values of satellite images were used to develop regression models and thus to estimate the annual net biomass amounts in productive and degraded forest. The coefficient of determination (R2) of regression model for the AGB was found to be 0,72 based on ecological factors in grassland areas. Similarly, R2 values of the developed models for the AGB based on satellite images were between 0,67 to 0,80. On the other hand, above ground carbon (AGC) and below ground carbon (BGC) contents ranged from 0,48 ton/ha to 0,94 ton/ha and 1,70 ton/ha to 2,76 ton/ha, respectively. Moreover, annual net biomass productions in the productive and degraded forest were modeled using satellite images with R2 values between 0,62 and 0,80. The total amount of carbon in agricultural areas was calculated as 467.997 tons. The total biomass, total amount of carbon (including soil carbon) and the annual net carbon production were 6.803.967, 14.589.276 and 139.852 tons, respectively. The carbon contents for the unit area changed from 31,3 tons/ha to 121,6 tons/ha. Total annual net carbon production was calculated as 139.852 tons and the unit values were between 0.26 ton/ha and 0.67 ton/ha. Finally, the carbon efficiency values were calculated for the all land uses. It was seen that the performance ranking for the land use classes in terms of carbon efficiency were productive forest, degraded forest, agriculture and grassland. In conclusion, this study suggests that AGB contents can be effectively estimated using remote sensing data.