MODELING THE IMPACT OF LAND USE/COVER, CLIMATE CHANGE, AND SOIL MANAGEMENT PRACTICE’S ON SOIL ORGANIC CARBON STOCK IN NORTHWEST AND CENTRAL RIFT VALLEY OF ETHIOPIA

Abstract

The stock of soil organic carbon (SOC) has been impacted by changes in land use and land cover (LULC), climate change (CC), and soil management (SM) practices. Soil organic carbon loss increases when forest converted to agricultural land, and likewise SM practice like crop rotation increase SOC while tillage decrease, but the impact of CC on SOC stock is not direct. To monitor and predict the dynamics of SOC stock at various scales, soil carbon models (SCM) are essential. In Ethiopia, SCM were not validated well due to lack of skill and adequate data to use the model. In other word, SCM were not explicitly utilized in the context of CC, LULC changes, and SM practices. Thus, the objective of this study is Evaluate the likely impacts of LULCC, changing Climate, and Soil Management Practices on SOC stock of Anjeni Watershed using a modeling approach. RothC and CQESTR models were used for this study. to assess the impact of CC, LULC change, slope gradient, and the RothC model performance on SOC stock of Anjeni watershed (AW) i.e., northwest Ethiopia as well as evaluate the CQESTR model for estimating SOC response to tillage and residue removal in Melkasa agricultural research center (MARC) i.e., in central rift valley. Current and historical soil carbon, bulk density, clay content, land use, crop yield, crop residue and climate data were used as inputs to evaluate the performance of RothC and CQESTR models. The RothC model was calibrated using long-term SOC, land management, and climatic data from the AW, while CQESTR used tillage and crop residue management data from MARC. The RothC correlation coefficient between simulated and observed SOC in 1997 and 2021 were 0.77 and 0.86, respectively, whereas CQESTR model was able to simulate SOC change with correlation coefficient value of 0.89 and mean square deviation of 0.008, indicating that both models effectively characterized the SOC. The RothC model simulated higher current and projected SOC for grass/fallow land as compared to the cultivated land. The RothC model predicted/simulated slightly different SOC stock under different slope gradients, with values ranging from 4.15 ton/ha, in the lower slope gradient to 2.1 ton/ha in the upper slope gradient. xx xx The projected temperature for the year (2022-2052) indicated rising that result in increasing of SOC from the baseline for e.g., in upper slope gradient of GFDL by 0.64 and 2.1 ton/ha using IPSL (SSP5-8.5) and MPI (SSP5-2.6), respectively. The CQESTR model result in MARC indicated a declining trend of SOC under CP both in topsoil layer (0-15) and lower layer (15-30) cm soil depths, which is due to practices of tillage and crop residue removal from cultivated land in maize bean intercropping. Variations in SOC concentration was noted in the lower layer soil depths under CA, while the topsoil layer SOC for CA increased over time. The predicted SOC under CP decrease to less than 1% by 2030 from about 1.25% in the 2000s. Overall, improved SM practices including CA, soil water conservation, and the expansion of fallow lands are responsible for the anticipated SOC stock increase in both locations, which is beneficial for increasing agricultural productivity. The models tested simulated the SOC under different land use, climate scenario and soil management practices reasonably well, and hence can be used as decision support tools under Ethiopia context. However, calibration and validation of other models would be crucial and more has to be done to resolve critical shortage of long-term data across locations in the country. Without such data, the use of such decision-support tools can’t be possible.