Estimation of carbon sequestration potential in soybean farms using remote sensing plant indices (Case study of Gorgan county, Golestan province)

Nowadays, concerns about the amount of carbon emission to atmosphere and its effects on the climate are increasing. In the new approach to agricultural ecosystem management, carbon sequestration service is being replaced by carbon emissions. Therefore, providing accurate information on the spatial distribution of biomass and carbon sequestration potential of ecosystems is an essential issue. The purpose of this study was to using vegetation index to estimate the potential of carbon sequestration in soybean (Glycine max L.) biomass in the fields of Gorgan county.

This study was conducted in soybean fields of Gorgan county, located in Golestan province during 2016-2017. In order to carry out this study, we were used satellite images of Sentinel 2 on 17.9.2017. Also, the accuracy of the images was checked after geometric and radiometric corrections using 250 ground control points. In the field section, plant samples from of 150 points were randomly prepared using 0.5 × 0.5 quadrats in the stage of maximum vegetative growth of soybean and transferred to the agricultural research laboratory of Gorgan University of Agricultural Sciences and Natural Resources. Then, the dry weight of plant samples was calculated and the amount of carbon in the plant biomass (stems and leaves) was determined using the combustion method. To prepare the land use map of soybean cultivation area, the supervised classification was used according to the maximum and minimum reliability algorithm. The accuracy of this classification was assessed by value of overall accuracy and Kappa coefficient. The studied vegetation indices were NDVI (Normalized Difference Vegetation Index), DVI (Deference Vegetation Index), RVI (Ratio Vegetation Index) and SAVI (Soil Adjusted Vegetation index). Then, a regression relationship was established between plant indices and biomass and carbon sequestration potential in SPSS 16 software. After selecting the best vegetation index, plant biomass and carbon sequestration potential maps were prepared using ArcGIS 10.6 software. Then, using classification methods, the final layer was divided into four classes of biomass and carbon sequestration potential.

In this study, the area of soybean cultivation in Gorgan county was estimated as 12333.71 hectares. The results showed that the value of overall accuracy coefficient in the classification method was the minimum distance from the mean, 87% and in the maximum probability method was 92% and the value of Kappa coefficient was 0.79 and 0.93, respectively. Based on the results of regression analysis, DVI index was selected for biomass mapping and carbon sequestration potential due to the highest coefficient of determination (0.86) and the lowest amount of RMSE. Based on the results, the highest amount of carbon sequestration potential was obtained from 1924.78 to 2526.18 kg.ha-1 in the central and eastern parts of this region, and the lowest value estimated as 919.43 to 1313.83 kg.ha-1 that related to western regions of Gorgan.

In this experiment, DVI index indicated a better estimation of carbon sequestration potential in soybean agroecosystems. Due to the high accuracy, easiness and low cost of remote sensing technology, especially calculation of vegetation indices, this approach can used to estimate the biomass and carbon sequestration potential of crops in agroecosystems.