Differences of MMF and USLE Models for Soil Loss Prediction along BTC and SCP Pipelines

The main goal of this study is to assess the Morgan-Morgan-Finney (MMF) and the universal soil loss equation (USLE) erosion models in the prediction of soil degradation along the corridor of oil and gas pipelines. In the comparative analysis, the MMF model revealed a larger coefficient of variation (COV) in predicted soil loss rates. Based on the pair-sample $t$-test, the predictions of the two models were significantly different in the spatial distribution of soil loss along the rights-of-way (RoW) of the pipelines. Sensitivity of the MMF and USLE models to terrain morphometric elements was also assessed. Slope gradient was one of the controlling factors of erosion processes, but not of the soil loss rates. The MMF and USLE models did not reveal any sensitivity to slope aspects. In terms of elevation, the MMF model revealed higher soil loss rates in the lower elevations than with the USLE model, leading to the conclusion that the USLE model is more sensitive to elevation change than the MMF model. The USLE model revealed higher sensitivity to the terrain curvature than the MMF model because it had larger variations within concave and flat terrain curvature types. Both models were sensitive to increasing vegetation cover (VC) percentage. Both models revealed different sensitivities; therefore, better understanding of these sensitivities may contribute to the selection of the most suitable model, depending on the terrain, to yield the highest soil loss prediction accuracy. Qualitative validation of the spatial distribution of USLE- and MMF-predicted erosion-prone areas was performed using 6 years of ongoing surveillance and measurement of erosion occurrences. Quantitative validation of the predicted soil loss was performed using 3 years of monitoring of field erosion plots. The USLE model performed better than the MMF model in terms of the frequency ratio of erosion occurrences within the critical erosion classes (soil loss $>10\mathrm{t}/\mathrm{ha}$). The USLE-predicted soil loss rates were more reliable than the MMF rates not only in terms of spatial distributions of critical erosion classes, but also in quantitative terms of soil loss rates because of the high correlation with the soil loss measurements of field erosion plots. The number of erosion-prone pipeline segments realistically predicted by the USLE model, e.g., soil loss more than $10\mathrm{t}/\mathrm{ha}$, was 88, whereas the MMF model predicted only 76 erosion-prone pipeline segments. The regression analysis between the total of 354 USLE and MMF erosion-prone segments revealed an $R^2$ equal to 0.33, which means that the predictions by the USLE and MMF erosion models are significantly different on the level of pipeline segments.