DOGAMI Interpretive Map Series

[Go back] | Publications Center | DOGAMI Home

Publication Preview

IMS-57, Landslide hazard and risk study of central and western Multnomah County, Oregon, by William J. Burns, Nancy C. Calhoun, Jon J. Franczyk, Kassandra O. Lindsey, and Lina Ma

OVERVIEW: The study identifies areas of Multnomah County where landslides are more likely to happen – and how many people and buildings are at risk when they do. The study, which included the cities of Portland, Gresham, Troutdale, Fairview, and Wood Village, shows areas where taking action to reduce landslide risk could make a big difference. Read the news release.

Publication Downloads:

Related publications:



At least 1,700 landslides have occurred within the City of Portland during the last 90 years (1928–2016). Of these landslides, approximately 830 occurred during the severe storms in 1996. From these historical data, we estimate an average of 20 landslides per year in the City of Portland. We estimate annual loss from landslides in the City of Portland ranges from $1.5M (million) to $3M. In years with extreme winter storms, this estimate can increase to approximately $64M to $81M. These historical data are a clear indication of a significant landslide risk and thus the need for continued landslide risk reduction.

In 2014, the Oregon Department of Geology and Mineral Industries (DOGAMI) submitted a grant application to Federal Emergency Management Agency (FEMA) and was granted funding to perform this study. The majority of the work on this project took place during 2015-2016. The purpose of the project was to assist the communities in the study area to better understand the landslide hazard and risk and to continue landslide risk reduction. Deliverables of the study include:

The study area includes the Cities of Portland, Gresham, Fairview, Wood Village, Troutdale, and portions of Multnomah County and covers approximately 300 square miles. The city of Portland is divided into risk reporting areas roughly defined by the nine neighborhood coalitions. Nearly one quarter of the people living in Oregon (~4 million people), live in the study area (~724,000 people). These people live and work in approximately 230,000 buildings worth approximately $75M with an additional $45M in land value.

First, we compiled existing detailed, lidar-based landslide inventories. These data were created and published during 2010–2012 by following the protocol of Burns and Madin (2009). Then, we updated the historical landslide inventory inside the City of Portland with data provided by the City. We created new, generalized bedrock and surficial engineering geology datasets as part of this study as the foundation of new susceptibility maps. The new shallow and deep landslide susceptibility maps are appropriate for use in landslide risk reduction activates such as updates to building codes and evaluation of storm water systems.

We performed two types of risk analysis: 1) hazard and asset exposure, and 2) Hazus® earthquake-triggered landslide risk analysis. We found that approximately $1.65B (billion) in land and buildings and almost 6,700 people are located on existing landslides. Also, 29,000 people live in the high-susceptibility hazard zone for shallow landslides, and nearly 8,000 people live in the high-susceptibility hazard zone for deep landslides in the study area. The second type of risk analysis, with Hazus, a risk modeling software package developed by FEMA, can be used to model a variety of earthquake, flood, and wind probabilistic hazards and/or hazard event scenarios. Because there is no Hazus landslide module, we used the earthquake module with and without earthquake-induced landslide hazards. Then we subtracted the earthquake-without-landslides model from the earthquake-with-landslides model so that the earthquake-induced landslide damage and losses could be examined separately. We found in some communities up to 25% of the modeled damage is from landslides triggered by earthquakes.

Although we cannot predict when the next landslide events will occur or how big they will be, we were able to provide a detailed understanding of landslide events in the past, the estimated scale of a potential disaster, the areas more or less susceptible to future landslides, and an estimate of what the damage and losses might be. All of these data confirm that landslide risk exists in the study area and thus that there is a strong need for continued landslide risk reduction. Landslide risk reduction can be performed in various ways. We provide recommendations and conclusions based on our findings. These recommendations are not comprehensive, but they should provide an adequate foundation for many of the risk management phases. The primary actions are related to awareness, regulations, and planning.

Geodatabase is Esri® version 10.1 format.
Metadata is embedded in the geodatabase and is also provided as separate .xml formatted files.

The GIS data included with this publication are in an ArcGIS version 10.1 file geodatabase.


Points. A compilation of center points of published or known historic (roughly 1849-2016) landslides and any available information about these landslides including damage and losses and several other attributes within central and western Multnomah County, Oregon.

Deposits.xml Polygons. Delinate landslide deposits (including debris flow fans and talus extent). This feature class is also used for identification of landslide type, assigning a confidence value of landslide identification and existence, and calculating the volume of the landslide. .xml
Scarps_Flanks.xml Lines. Scarp lines are used to estimate of future retrogressive failure distance behind the head scarp, aid in type classification, and assigning a confidence value for the landslide. This feature class represents lines of the uppermost extent of landslide head scarp and any internal scarps in the body of the landslide. .xml
Scarps.xml Polygons. Scarps and flanks polygons are used to identify the failure plane, aid in type classification, assigning a confidence value for the landslide, and calculating the volume of the landslide. These head scarps or uppermost scarps in many cases expose the primary failure plane (surface of rupture) and flanks or shear zones. .xml
Shallow_LS_Susceptibility.xml Raster. Shallow-landslide susceptibility zones were established from locations of shallow-landslide deposits and their associated head scarps, factor of safety calculations, and buffers. The symbology on this map includes high, moderate, and low susceptibility zones. The susceptibility zones were established from locations of shallow landslide deposits and their associated head scarps, factor of safety calculations, and buffers following protocol developed by Burns and other (2012). .xml
Polygons. Deep-landslide susceptibility zones are high and moderate susceptibility. The deep susceptibility zones were established based upon location and proximity to deep landslide deposits and head scarps, buffers along the landslide head scarps, susceptible geologic units, slope angles, and mapper judgement following protocol developed by Burns (2008). .xml

PLATES (PDFs, 60 by 40 inches each; scale 1:200,000)

Plate 1. Landslide inventory map of central and western Multnomah County, Oregon
high-resolution (50 MB) | low-resolution (19 MB)

Plate 2. Shallow landslide susceptibility map of central and western Multnomah County, Oregon
high-resolution (58 MB)| low-resolution (23 MB)

Plate 3. Deep landslide susceptibility map of central and western Multnomah County, Oregon
high-resolution (50 MB) | low-resolution (20 MB)