Understanding the creeping behavior of loess is of great importance as large-scale loess landslides are closely related with creep behavior. At present, it is still challenging to predict and estimate the long-term stability of such landslides. This is in a large degree due to the poor understanding of moisture control on creep behavior of loess. The purpose of this study is to decipher the loess creep behavior under various moisture contents (MCs) using loess specimens obtained from Baqiao landslide, Xi’an of China, using multi-loading triaxial creep tests under different MCs of 9%, 12%, 15%, 18% and 21%. Based on the laboratory test results, a series of relationships between the creep rate at the steady-state creep stage and the initial strain and initial shear modulus are revealed. Meanwhile, a method for obtaining the long-term strength of loess specimens, namely, the Steady-state Creep Rate Slope Method (SCRSM), is proposed. SCRSM resolves the issue in several conventional methods such as the Isochronous Stress-Strain Curve Method, the Tangent Method of Steady Creep Rate when MCs are of concern. Such an improvement is primarily due to a better way of finding the inflection point of the steady-state rate. It is found that SCRSM is robust and accurate to determine the long-term strength of loess specimens. Furthermore, we propose a modified Burgers model with a newly introduced nonlinear parameter n to overcome deficiencies of conventional creep models. This modified Burgers model is flexible to fit the creep test curves of loess, and can describe the curves at the accelerated creep stage more accurately. Lastly, the main factors triggering the Baqiao landslide is analyzed considering stratum lithology, rainfall and excavation. In general, this study provides a basis for understanding the evolutional process of loess landslides as well as guidelines for prevention, controlling and prediction of loess landslides.

了解黄土的蠕变行为非常重要，因为大规模的黄土滑坡与蠕变行为密切相关。目前，预测和估计此类滑坡的长期稳定性仍具有挑战性。这在很大程度上是由于对水分控制对黄土蠕变行为的了解不足。这项研究的目的是使用从西安八桥滑坡获得的黄土标本，通过在9％，12％， 15％，18％和21％。根据实验室测试结果，揭示了稳态蠕变阶段的蠕变速率与初始应变和初始剪切模量之间的一系列关系。同时，提出了一种获得黄土试样长期强度的方法，即稳态蠕变速率斜率法（SCRSM）。SCRSM通过多种常规方法解决了这个问题，例如，等时应力-应变曲线法，稳态蠕变速率切线法（当MC受到关注时）。这种改进主要是由于找到稳态速率拐点的更好方法。发现SCRSM是可靠且准确的，可以确定黄土标本的长期强度。此外，我们提出了带有新引入的非线性参数的改进的Burgers模型 当关注MC时，稳态蠕变速率的切线方法。这种改进主要是由于找到稳态速率拐点的更好方法。发现SCRSM是可靠且准确的，可以确定黄土标本的长期强度。此外，我们提出了带有新引入的非线性参数的改进的Burgers模型 当关注MC时，稳态蠕变速率的切线方法。这种改进主要是由于找到稳态速率拐点的更好方法。发现SCRSM是可靠且准确的，可确定黄土样品的长期强度。此外，我们提出了带有新引入的非线性参数的改进的Burgers模型n克服了常规蠕变模型的不足。这种修改后的Burgers模型可以灵活地拟合黄土的蠕变测试曲线，并且可以更准确地描述加速蠕变阶段的曲线。最后，从地层岩性，降雨和开挖的角度分析了引起the桥滑坡的主要因素。总的来说，这项研究为了解黄土滑坡的演化过程以及黄土滑坡的预防，控制和预测提供了基础。