Predicting the amount of sequestrated carbon in rangeland soil under the effect of soil sampling depth and height using the response surface methodology (RSM)

Background and objectives: Organic carbon is one of the important parameters to determine soil fertility, and production ability and is a mind index for showing soil quality, especially in dry and semi-dry lands. The rangelands are composed of about 50 % of the world's lands and have more than 33 % of carbon reservoirs on earth. Considering the ranges as the most important terrestrial ecosystems, which are natural reserves of carbon, are very considered plus how and amount of carbon sequestration in them is very important. Methodology: In this research, after selecting rangelands in the southern domain of Damavand mountain, considering the geographic condition and regional plant cover, 4 height groups relative to sea level with equal type and composition of plant cover plus medium slope based on physical presence and region natural resources office data composed of height group 1 (2504-2664 m), group 2 (2730-2896 m), group 3 (2923-3050 m) and group 4 (3119-3545 m) for sampling soil at different depths were selected. At all height groups, 13 random points were determined and 3 samples from each point were dug at 0-15 cm and 15-30 cm depth. After sequestrated carbon measuring at all samples, the obtained data were analyzed. The amount of carbon sequestration was determined with the multiplication of soil organic carbon and bulk density and depth at hectare unit. Then regression equations and response surface method (rsm) in the design software platform were created. They were compared and their accuracy in predicting carbon sequestration was identified as a more accurate method. The coefficient of determination (R2) between the measured and predicted data was used to evaluate the accuracy of regression models and the response surface methodology. Results: The results of analysis of variance (ANOVA) showed that only sampling depth had a significant effect on soil carbon sequestration, but the effect of sampling height and the interaction of depth and height were meaningless. In general, the rate of carbon sequestration at a depth of 15-30 cm was more than 0-15 cm, and at both depths, height group 4 (3119-3545 m) had the highest soil carbon sequestration. In addition, after height group 4, the highest carbon sequestration was related to height group 1 (2504-2664 m). The highest amount of carbon sequestration as 604654 g/m2 was related to the height group 4 and the depth of 15-30 cm and the lowest about 225000 g/m2 was related to height group 2 and the depth of 0-15 cm. In modeling, the quadratic model of the response surface method predicted soil carbon sequestration rate with higher accuracy (R2 = 0.7363) than the linear model of the same method (R2 = 0.6014) and the less accurate regression model (R2 = 0.428). In addition, the response surface method succeeded in optimizing the inputs as a height of 3500 meters and a second depth in order to find the amount of carbon sequestration in the soil as 808811.699 (g/m2) with a satisfaction of 0.976. Conclusion: The response surface methodology (rsm) as a novel tool has satisfied operation at prediction and optimization of rangeland soil carbon sequestration under the different sampling depths and heights. Furthermore, it will be used as an intelligent tool for predicting the different parameters at studied ranges and rangeland science like above and underground biomass volume, covering the range of plant species, and so on.  od in the design software platform. They were compared and their accuracy in predicting carbon sequestration was identified as a more accurate method. The coefficient of determination (R2) between the measured and predicted data was used to evaluate the accuracy of regression models and the response surface methodology. The results of the analysis of variance (ANOVA) showed that only sampling depth had a significant effect on soil carbon sequestration, but the effect of sampling height and the interaction of depth and height were meaningless. In general, the rate of carbon sequestration at a depth of 15-30 cm was more than 0-15 cm, and at both depths, elevation category 4 (3119-3545 m) had the highest soil carbon sequestration. In addition, after elevation category 4, the highest carbon sequestration was related to elevation category 1 (2504-2664 m). The highest amount of carbon sequestration (604654 gr/m2) was related to the height category 4 and the depth of 15-30 cm. In modeling, the quadratic model of the response surface method predicted soil carbon sequestration rate with higher accuracy (R2 = 0.7363) than the linear model of the same method (R2 = 0.6014) and the less accurate regression model (R2 = 0.428). In addition, the response surface method succeeded in optimizing the inputs at a height of 3500 meters and a second depth in order to find the amount of carbon sequestration in the soil as 808811.699 gr/m2 with a satisfaction of 0.976.