MUHAMMAD BARIK & JENNIFER ADAMWASHINGTON STATE UNIVERSITY

LANDSLIDE SUSCEPTIBILITY MAPPING TO INFORM LANDUSE MANAGEMENT DECISIONS IN AN

ALTERED CLIMATE

Steve Burges Retirement SymposiumMarch 26th, 2010

Availability: the open doorInspiration, optimismBalancing direction and self directionThe art of questioning and listeningBeing widely read and widely acceptingThe initial projectLife after scienceCelebration

Mentoring Tips

Land Use Change:Logging has increased landslide frequency by

2-23 times in the Pacific Northwest (Swanson and Dyress 1975, Jakob 2000, Guthrie 2002, Montgomery et al. 2000).

Climate Change: PNW winters are expected to become wetter;

precipitation events are expected to become more extreme (Mote and Salathe 2010).

Impacts on Riparian Health:Resulting sediment negatively affects riparian

ecosystems, i.e., reduced success of spawning and rearing of salmon (Cederholm et al. 1981; Hartman et al. 1996).

Motivation

Forest Management ObjectiveIncreasing economic viability while preserving

the natural environment.“Zoned” management approach

Previous Best Management Practice StudiesImpacts on landslides are site specificNo incorporation of climate change effects into

long term plans

Improved Forest Management Practices

To provide high resolution maps of the susceptibility of landslide activity to timber extraction under historical and future climate conditions.

Objective

How is landslide activity affected by timber extraction and how does this impact vary over a range of topographic, soil, and vegetation conditions?

How will landslide susceptibility to timber extraction respond to projected climate change?

Research Questions

The Olympic Experimental State Forest (OESF)

Source: DNR

“Unzoned” Management Approach

The Queets River Basin

The Distributed Hydrology Soil Vegetation Model (DHSVM) (Wigmosta et al. 1994), with a sediment module (Doten et al. 2006) was used for this study.

DHSVM mass wasting is stochastic in nature.

Model

HILLSLOPE EROSION

Soil Moisture Content

CHANNEL ROUTI NGPrecipitationLeaf Drip

Infiltration and Saturation Excess Runoff

DHSVM

Q

Qsed

Q

Qsed

Sediment

MASS WASTI NG

Erosion

Deposition

ROADEROSION

Sediment

Channel Flow

Infinite Slope Model

Uses Factor of Safety Approach

Doten et al 2006

Hydrologic calibration and evaluation (NS = 0.52, Volume Error = 22%; other studies looking into reasons behind poor model performance)

Evaluation of mass wasting module over sub-basins

Model Calibration and Evaluation

Slide Year Historic Landslides TotalSurface Area(m2)

Total Surface Area(m2) of All Cells Factor Safety <1

(From Modeled Run)

Sub-basin 1 10614 11400Sub-basin 2 15257 13678

1 23

Factors considered: slope, soil, vegetation

* The primary factors triggering harvesting-related shallow landslides (Watson et al. 1999).

Factors affecting landslide susceptibility

Watson et al. 1999

Elevation class (m)

Slope Class (Degree)

Soil Classes Vegetation Classes

0-500 0-10 Sand Deciduous Broadleaf

<500 11-20 Silty Loam Mixed forest

21-30 Loam Coastal conifer

31-40 Silty clay Loam Mesic conifer

40-50 Talus

>50

Logging Scenarios for Model Simulation

Properties changed to simulate logging:

1.Root cohesion 2.Vegetation Surcharge 3.Fractional coverage

Selection of Logging Scenario

Clear-cutting done in20-30 degree sloperange.

Results: Slope Classes

Results: Soil Classes

Results: Vegetation Classes

Results: Elevation Classes

Landslide Susceptibility Map for Historical ClimateWeighted

indices calculated for each category of each class

Used to determine the susceptibility class

Comparison to actual landslides in logged areas

All the polygons are harvested areas processed from 1990 Landsat-TM image. Weights were calculated for each cell on the harvested area and three susceptibility classes are created.

Red marks are all historical landslides between 1990 to 1997, collected from DNR HZP inventories.

Climate change effectsCGCM(B1) 2045

Climate change effectsCGCM(A1B) 2045

Results indicate that 30 to 50 degree slopes range and certain types of soils (e.g. talus, sandy) are most vulnerable for logging-induced landslides.

For 2045 projected climate areas with high landslide risk increased on average 7.1% and 10.7% for B1 and A1B carbon emission scenarios, respectively.

Ongoing Work:Model inputs and calibrationMore extensive model evaluation Isolate effects of soil and terrain factors Isolate effects of precipitation versus temperature

changesMore realistic post-logging effects Impacts on riparian habitat

Summary

Factor of safety calculation

CS = Soil cohesionCr = Root cohesionФ= Angle of internal frictiond= Depth of soilm= Saturated depth of soilS = Surface slopeq0 = Vegetable surcharge

Weight Calculation

Wi= The weight given to the ith class of a particular thematic layer

Npix(Si)=The number of slides pixels in a certain thematic class

Npix(Ni)=The total number of pixels in a certain thematic class.

n= The number of classes in the Thematic map

Yin and Yan (1988), Saha et al. (2005)

Weight for a particular cell W = ƩWi

Susceptibility Class Segmentation

No of landslides cell in the

susceptibility classNo. of total cells in the

susceptibility classPercentage of landslides in a susceptibility class

Low(<.05) 621 28049 2.2

Medium(.05-.79) 617 25099 2.5

High(>0.79) 627 19021 3.3

Index segmentation and evaluation

Frequency of slides in different susceptibility classes.

LSI value had the range from -3.24 to 2.21. This range was divided into three susceptibility classes based on cumulative frequency values of LSI on slide areas ( Saha et al. 2005). The breaks were done at 33 and 67%.

classesCGCM_3.1t47 (A1B)

CGCM_3.1t47 (B1)

CNRM-cm3 (A1B)

CNRM-cm3 (B1)

(a)Elevation(m)0-500 2.3 1.9 2.3 2.6>500 4.6 5.4 3.0 4.4(b)slope(Degree)<10 US* US* US* US*

10-20 8.5 7.1 8.0 8.720-30 6.2 9.1 10.7 6.030-40 2.3 5.0 2.9 2.740-50 1.0 1.0 0.6 2.5>50 0.2 0.1 0.2 0.2(c)Soil Sand 12.3 16.9 8.5 19.2Silty Loam 1.2 1.4 1.5 2.1Loam 4.4 2.8 1.9 3.2silty clay Loam 7.7 9.0 6.0 7.7Clay 3.6 10.9 14.5 10.9Talus 9.2 12.0 11.2 8.5(d) VegetationDeciduous Broadleaf 10.8 12.5 10.9 11.2Mixed forest 1.4 5.1 9.8 4.8Coastal conifer forest 0.2 0.6 0.3 0.5Mesic conifer forest 5.2 6.4 5.1 6.7

Climate change scenarios

Increment of slides in harvested areas for different climate change scenarios

Susceptibility Class

Historical

CGCM_A1B

Percentage change

CGCM_B1

Percentage change

CNRM_A1B

Percentage change

CNRM_B1

Percentage change

Low 4120646 4130782 0.25 4131625 0.27 4130782 0.25 4130782 0.25

Medium 3224217 2783816 -13.66 3078979 -4.50 2750346 -14.70 2772799 -14.00

High 4187537 4617802 10.27 4321796 3.21 4651272 11.07 4628819 10.54

Climate Change Effect

Change in percentage of areas in different susceptibility classes for different climate change scenarios with respect to the historical scenario. For all the future climate change scenarios areas increased under the high susceptibility class.