Date of Award


Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science in Civil Engineering (MSCE)


Civil Engineering


Aaron Gallant

Second Committee Member

Carlos Vega-Posada

Third Committee Member

Sean Smith


The September 28 Mw 7.5 Palu-Donggala earthquake in Indonesia was the deadliest natural disaster in 2018. Five flowslides on the eastern side of Palu Valley, which were attributed to the substantial majority of deaths and economic losses in the region, occurred due to liquefaction of naturally dry alluvial deposits in gently sloping ground (2 to 6%). The requisite conditions for soil liquefaction are the presence of loose granular sediments and full saturation. Instigation of these catastrophic ground failures were linked to artificial saturation by an unlined agricultural irrigation canal. From a civil engineering perspective, the Palu ground failures on the east side of the valley highlighted the geotechnical risks imposed by land use and the built environment. Thus, assessment of the hazard that irrigation networks introduce must be thoughtfully considered, but also balanced with the economic necessity of irrigation systems. As the canal and agricultural land use extended along the entire eastern side of the valley, it remains poorly understood why flowslides occurred at isolated locations and were not widespread. The objective of this study was to address why large flowslides occurred at discrete locations adjacent to the canal and not others; it was hypothesized that staggered land use and variation in the density of secondary irrigation networks locally resulted in non-uniform or partial saturation of potentially liquefiable deposits on the eastern slopes in some regions of Palu Valley. Remote mapping of attributes potentially contributing to flowslide initiation, including surficial geology, topography, land use, and features contributing to—or potential indicators of—the extent of ground saturation, were compared with mapped ground displacements. The degree of instability was categorized based on the magnitude of displacement to define different displacement “events.” Spatial analyses were performed using the Getis-Ord statistic to identify clustering (i.e. hot and cold spots) of displacement events and compared with the aforementioned attributes. Mapping revealed that although the density of agricultural land use did not differ appreciably in regions where flowslides did not occur, the density of secondary irrigation networks extending from the main canal did. Flowslides initiated where three conditions were true: i.) agricultural land use was continuous on the eastern slopes; ii.) the secondary irrigation network was relatively dense; and iii.) the slopes extending from the canal were steeper than 2%. Analyses provided evidence that suggests decreasing the density of secondary irrigation networks and/or intermittent agricultural land use (e.g. incorporation of regions with dense vegetation) may be a potential avenue to mitigate future flowslide triggering in Palu Valley without compromising agricultural activity.

Files over 10MB may be slow to open. For best results, right-click and select "save as..."