Empirical Criteria for Rainfall-Induced Landslides

1
Empirical Criteria for Rainfall-Induced Landslides

• Professor Steve Kramer

2
Landslide Records

• Can take an empirical approach to the problem
• Record observed landslides along with
  precipitation date Caine (1980)
• Attempted to find the threshold level of rainfall
  needed to trigger instability
• Total amount of rainfall not a good predictor by
  itself
• Highest instantaneous intensity not a good
  predictor by itself
• A combination of the two is effective

3
Landslide Records

• Caine (1980)
• 73 sets of intensity-duration data that produced
  shallow landslides and debris flows worldwide

65 mm/hr
25 mm/hr
14 mm/hr
4 mm/hr
1.6 mm/hr
1 min
10 min
1 hr
1 day
10 days
4
Landslide Records

• Many others have collected data and proposed
  alternative thresholds
• Intensity-Duration thresholds proposed at local,
  regional, and global scales
• Global models independent of local
  morphological, lithological, and land-use
  conditions of local or regional rainfall
  patterns, and history
• Regional models scale of few to several
  thousand square kilometers exhibit similar
  meteorological, climatic, physiographic, and soil
  characteristics
• Local models scale of a few to several hundred
  square kilometers consider local climatic
  regime, geomorphological setting

5
Landslide Records

• Global Model
• Guzzetti et al. (2007)
• Collected database of rainfall events that
  produced shallow slides
• 2,626 slide events from 1917-2005 97.5 between
  1950-2005

6
Landslide Records

• Global Model
• Rainfall Information
• Intensity and duration of rainfall event that
  resulted in shallow slides and debris flows
• Cumulative amount of precipitation for the event
• Measures of antecedent precipitation (where
  available)

7
Landslide Records

• Global Model
• Landslide Information
• Type and depth of failure descriptions often
  ambiguous
• Number of slope failures usually qualitative
  descriptions of time, date of landslide
• Lithology generic description of main rock type

8
Landslide Records

• Global Model
• Climate Information
• Broad Categories
• Tropical rainforest (Af), monsoon (Am), wet-dry
  savanna (Aw)
• Arid semi-arid steppe (Bsk), middle latitude
  desert (Bwk)
• Humid subtropical east coast (Cfa), mild marine
  west coast (Cfb), warm (Cwa), Mediterranean warm
  (Csa), Mediterranean cool (Csb)
• Humid Continental mild summers (Dfb), hot
  summers (Dwa), subarctic (Dfc)
• Tundra (ET)
• Highland (H)

9
Landslide Records

• Global Model

NOTE Squares are debris flows, and circles are
shallow slides
10
Landslide Records

• Global Model
• Durations range from 7 minutes to 5 weeks
• Intensities range from 0.12 mm/hr to 713 mm/hr
• General trend is apparent consistent with Caine
• No significant difference between shallow slides
  and debris flows
• Smooth threshold curve implied

11
Landslide Records

• Global Model
• Statistical analysis of data shows different
  pattern

48 hrs
12
Landslide Records

• Global Model
• Data suggests
• Higher intensity threshold for short events (D lt
  48 hours)
• Lower intensity threshold for long events (D gt 48
  hours)
• 48 hours is the duration for which antecedent
  precipitation and soil moisture conditions become
  important
• Beyond 48 hours, intensity stays relatively
  constant (I 0.2 mm/hr) regardless of duration
  suggests that other things (groundwater recharge
  etc.) play an important role
• Can develop single model or composite model

13
Landslide Records

• Global Model
• Guzzetti et al. (2008)
• Single Model
• Composite Model

14
Landslide Records

• Global Model
• Guzzetti et al. (2008)
• Can also break
• out different sets
• of climatic conditions

15
Landslide Records

• Local Model
• Landslide data in Seattle dates back to 1897
• Most slides occur in November-April period
  January is the most common month

16
Landslide Records

• Local Model
• Used short-term and longer term rainfall metrics
• 3-day cumulative precipitation
• 15-day cumulative precipitation (prior to 3-day
  period)
• Developed composite boundary

NOTE Points represent events with 3 or more
slides
17
Landslide Records

• Local Model
• With additional data, inclined boundaries
  extrapolated

18
Landslide Records

• Local Model
• 2000-2001 season

19
Landslide Records

• Local Model
• 2001-2002 season

20
Landslide Records

• Local Model
• 2002-2003 season

21
Landslide Records

• Local Model
• For the three periods (2000-01, 2001-02, 2002-03)
• 89 of Seattle slides were part of events above
  threshold ( 3 slides)
• 65 of all events (including surrounding area and
  lt 3 slide events) were above threshold
• Majority of slides below threshold had human
  influence (fill, plugged culvert, broken water
  pipe, excavation, improper drainage, etc.)
• Indicates lower threshold may be required for
  developed sites

22
Landslide Records

• Local Model - Forecasting
• Rainfall statistics can be monitored
• Near real-time plots of 3-day and 15-day values
  can indication potential for sliding

http//landslides.usgs.gov/monitoring/seattle/rtd/
plot.php