Forest Landslide

[Christopher]

AbstractDespite the large number of recent advances and developments in landslide susceptibility mapping (LSM) there is still a lack of studies focusing on specific aspects of LSM model sensitivity. For example, the influence of factors such as the survey scale of the landslide conditioning variables (LCVs), the resolution of the mapping unit (MUR) and the optimal number and ranking of LCVs have never been investigated analytically, especially on large data sets.

In this paper we attempt this experimentation concentrating on the impact of model tuning choice on the final result, rather than on the comparison of methodologies. To this end, we adopt a simple implementation of the random forest (RF), a machine learning technique, to produce an ensemble of landslide susceptibility maps for a set of different model settings, input data types and scales. Random forest is a combination of Bayesian trees that relates a set of predictors to the actual landslide occurrence. Being it a nonparametric model, it is possible to incorporate a range of numerical or categorical data layers and there is no need to select unimodal training data as for example in linear discriminant analysis. Many widely acknowledged landslide predisposing factors are taken into account as mainly related to the lithology, the land use, the geomorphology, the structural and anthropogenic constraints. In addition, for each factor we also include in the predictors set a measure of the standard deviation (for numerical variables) or the variety (for categorical ones) over the map unit.

As in other systems, the use of RF enables one to estimate the relative importance of the single input parameters and to select the optimal configuration of the classification model. The model is initially applied using the complete set of input variables, then an iterative process is implemented and progressively smaller subsets of the parameter space are considered. The impact of scale and accuracy of input variables, as well as the effect of the random component of the RF model on the susceptibility results, are also examined. The model is tested in the Arno River basin (central Italy). We find that the dimension of parameter space, the mapping unit (scale) and the training process strongly influence the classification accuracy and the prediction process.

This, in turn, implies that a careful sensitivity analysis making use of traditional and new tools should always be performed before producing final susceptibility maps at all levels and scales.

Trees and forests can also increase landslide risk by imposing load on unstable slopes and via wind-related effects; they are unlikely to prevent or minimise deep landslides or slides on very steep slopes. However, they can make a positive contribution in various situations as this policy brief explains.

Findings of the research, led by associate forest engineering associate professor Catalina Segura and graduate student Arianna Goodman of the Oregon State University College of Forestry, were published in the journal Earth Surface Processes and Landforms.

Probing the factors behind landslide frequency and magnitude is crucial because slides occur in all 50 states, causing an average of more than 25 deaths per year, according to the United States Geological Survey. The USGS puts the total annual average economic damage resulting from landslides at greater than $1 billion.

Focusing on the Lookout Creek watershed in western Oregon, a research team that included Julia Jones of the OSU College of Earth, Ocean, and Atmospheric Sciences and Frederick Swanson of the U.S. Forest Service examined a decades-long history of old-growth clear-cutting and associated road construction and how those practices affected flooding, landslides, big pieces of wood jamming up waterways, and channel change.

The scientists note that three zones of distinct and contrasting geologic history comprise the Lookout Creek watershed: one zone with relatively smooth terrain and U-shaped valleys; another characterized by irregular topography, rough surfaces and moderate steepness; and a third featuring V-shaped valleys, steep slopes and narrow drainages.

Even small floods caused landslides and stream channel changes during the first 15 years of road construction and logging, and amid ongoing logging in the early part of the time period between large flood events, she said.

Landscape effects were negligible in 2011 for the third largest flood event on record, the researchers found; by that time clear-cut areas of the forest had been replanted and the new trees were 20 to 70 years old.

About the OSU College of Forestry: For a century, the College of Forestry has been a world class center of teaching, learning and research. It offers graduate and undergraduate degree programs in sustaining ecosystems, managing forests and manufacturing wood products; conducts basic and applied research on the nature and use of forests; and operates more than 15,000 acres of college forests.

In all, the slide dislodged enough material to cover 600 football fields 10 feet deep, not only plastering the land but plugging the North Fork of the Stillaguamish River, forming a temporary lake. The landscape was gray and bald with broken trees jack-strawed in every direction and sand, gravel, mud and clay.

This area is closed to the public, and there are no plans to rebuild the homes that once stood here. In recent visits, a team of Seattle Times journalists, escorted by scientists from the Stillaguamish Tribe, witnessed the changes in this landscape in just a decade.

The slide area is alive with coyotes, elk, deer and beavers. Salmon are returning to the North Fork in the same numbers and places as before, even right in the debris field. Cottonwoods, willows and alders quill the land and a pelage of golden moss covers the ground. Fir trees have grown high overhead.

It is hard to appreciate the scale of the destruction today: An eyewitness to the March 22, 2014, landslide described a wall of mud and water moving across the North Fork, destroying some 35 homes and killing 43 people, burying Highway 530, then finally slopping back on itself, its energy spent.

With a swing of a machete, Jason Griffith, environmental program manager for the Stillaguamish Tribe, hacked into the thick forest already growing on the slide area as Charlotte Scofield, fisheries biologist for the tribe, cut her way in with loppers.

Cottonwood, alder and willow are masters of the reset: When a river floods and scours and creates new land with its gravel bars, when a volcano goes off, when a landslide levels an open, sunny space, these are the species that move right in. And by their presence they in turn transform the land. Willows and cottonwoods stabilize and hold river banks. Alders fix nitrogen from the air into nodules on their roots, providing crucial food for plants.

Ten years have passed since the landslide, and a new forest is emerging. Mosses, lichens and fungi are covering, stabilizing and enriching the sandy hummocks. Young trees are sending their roots deep into the churned layers, further stabilizing the landscape. Beavers are thriving. Elk and deer are eating emerging flowering plants and lichens.

Deciduous saplings such as alder, cottonwood and willow thrive in disturbed areas with lots of light. Red alder forms special nodules on its roots that contain a symbiotic bacteria, Frankia, which fixes nitrogen.

Lichens are fungi paired with a photosynthesizing partner, either green algae or cyanobacteria. Some lichens fix nitrogen from the air, returning this vital nutrient to the soil for other plants.

Here is an all-they-can-eat cafe perfect for their needs. They in turn are prey for cougars and other predators. (Not for nothing do biologists call beavers the Milk Duds of the forest.) Ducks were also using the ponds and wetlands created by the beavers.

Tom Hinckley, professor emeritus at the University of Washington, said the multispecies recovery underway here reminds him, at a smaller scale, of the regrowth he and others studied at Mount St. Helens after the blast on May 18, 1980. The same processes are at work here.

A long-term Pacific Northwest study of landslides, clear-cutting timber and building roads shows that forest management history has a greater impact on how often landslides occur and how severe they are compared to how much water is coursing through a watershed.

Findings of the U.S. National Science Foundation-supported research, led by Catalina Segura and Arianna Goodman of Oregon State University, were published in the journal Earth Surface Processes and Landforms. "The study highlights the importance of land-use dynamics on natural processes such as landslides," said Justin Lawrence, a program director in NSF's Division of Earth Sciences. "This team could improve the way forests are managed in the future."

Probing the factors behind landside frequency and magnitude is crucial because slides occur in all 50 states, causing an average of more than 25 deaths per year, according to the U.S. Geological Survey. The USGS puts the total annual average economic damage resulting from landslides at greater than $1 billion.

Focusing on the Lookout Creek watershed in western Oregon, the researchers examined a decades-long history of old-growth clear-cutting and associated road construction and how those practices affected flooding, landslides, big pieces of wood jamming up waterways, and channel change.

The 64-square-kilometer Lookout Creek watershed is part of the H.J. Andrews Experimental Forest, an NSF Long-Term Ecological Research site. Logging and road building began in the Lookout Creek area in 1950 and largely ceased by the 1980s, enabling the scientists to track forest management practices' impact on slides and floods during and after the period of active management.

The Washington Geological Survey works to increase public and scientific understanding of landslide hazards in Washington State. Please visit our Geologic Information Portal and Geologic Hazard Maps page for the most up-to-date listing of all of our hazard maps.

3a8082e126