Himalayan Disasters Script 

Amanda Rumsey:
Nepal is highly susceptible to landslides due to its mountainous topography, 
monsoon rain, seismic activity, and underdeveloped infrastructure. In April 2015, the magnitude
7.8 Gorkha earthquake struck Nepal, causing over 9,000 casualties and created over $1 billion in
damages.  Because of this earthquake, rainfall-triggered landslides are likely to emerge as a 
significant induced hazard during the monsoon season.

Previous research efforts have utilized landslide event data for hazard assessment and landslide
modeling, but landslide event databases are limited in scope and size due to non-reporting biases
associated with the lack of knowledge of events in under-populated areas and with the grouping 
of events with their primary triggering hazards.

With current underestimation of landslide impacts, it is critical to develop a landslide event
inventory that has the ability to accurately quantify landslide event information. Conventional
methods for landslide event detection rely on visual interpretation of high-resolution imagery
with manual digitization by analysts. These methods are resource intensive and are plagued by
spatial and temporal constraints. 

Jessica Fayne:
The DEVELOP team at Goddard Space Flight Center has developed an automated landslide detection
tool called the Sudden Landslide Identification Product (SLIP). SLIP helps researchers identify
landslides at a 60 meter scale or larger and helps them recognize changes in bare earth such as
increased soil moisture and snow melt. SLIP also creates a catalog detailing where and 
approximately when the land changes occur.

The SLIP program is simple, can run on any operating system, and is completely free, which
eliminates the need for complex and expensive software. The program first searches the United
States Geological Survey database to download a collection of 11 Landsat 8 images. The red bands
from the 10 oldest images are composited to produce a cloud-free red band baseline image. 
The red band from the newest image is then compared to the red band  baseline image using the 
percent change formula. Pixels with a red increase greater than 40% represent a significant change 
in vegetation. 

Next, the near infrared bands and the short wave infrared bands from the 10 oldest images are
composited separately to produce cloud-free baseline images. The near infrared and short wave
bands from the newest image are then subtracted from the infrared baseline images to show
increases or decreases in moisture. Areas with significant change in these band combinations
represent changes in soil moisture. 

Finally, a slope layer derived from the Shuttle Radar Topography Mission (SRTM) and
the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) elevation models
is used to identify  areas that are more likely to experience mass wasting due to high slopes.
The slope layer is also used to filter out areas of flat terrain such as river bottoms and 
farms, which regularly experience reflectance changes and moisture increases.

Pixels that show a percent red increase greater than 40%, a significant soil moisture increase,
and a slope of greater than 15% are flagged as landslide events and changes in bare earth to 
produce the final SLIP product.

Jordan Scheffler:
Currently, the Goddard DEVELOP team is validating the SLIP algorithm. To do this, first the raster
that is output from the SLIP product is converted into a polygon feature class using ArcMap. 
Next, the polygons are given a 60 meter buffer that clusters the individual polygons into larger
landslide features. After the buffers are created, 50 or more of the largest features are
extracted and validated by visual inspection with pan sharpened Landsat 8 true color imagery.

Future work on SLIP will include the creation of an online web platform that will feature 
the calibrated outputs of the SLIP product. Additionally, the algorithm is currently being
tested with runoff and soil moisture data from the Global Land Data Assimilation System (GLDAS)
to calibrate the SLIP product.