In Depth Coverage

The contribution of landsliding and deforestation to land-cover change

 Landsliding is a dominant process that may influence the short- and long-term dynamics of carbon in many steep landscapes worldwide. First, landsliding contributes to the redistribution of plant biomass, soil organic matter, and saprolite.  Second, landsliding influences the rates at which ecosystems reorganize, soils develop, and bedrock weathers on the recently created substrates.  Thus landsliding may play an important but unappreciated role in the large-scale dynamics of carbon through its effect on ecosystems and soils, the coupling between landslide deposits and streams, and ultimately the residence time of carbon in mountainscapes. The long-term goal of this proposal is to characterize carbon pools and fluxes in the Sierra de Las Minas (SLM), a small mountain range in eastern Guatemala designated as a Biosphere Reserve to protect its elevated diversity and rich water resources. The specific objectives are a) to characterize the spatial distribution of landslides and land-cover types, and b) to establish the relationship between landslide deposits and the stream network.  Because the elevated diversity of land-cover types and the large variability in landslide scars and deposit sizes found in the SLM, we propose to investigate the usefulness of WorldView-2 images for mapping purposes.  With their high spatial and spectral and likely temporal resolution, WorldView-2 images may help overcome the challenges outlined above, and in combination with GIS data address the two aims of this proposal.  Expected results include land-cover and landslide maps and a variety of spatial analyses. The proposed work represents a novel and invaluable contribution to the societal understanding of climate change, hazard assessment, and conservation studies.

• Link With Climate Change

  Open

   The links of this project to climate change are two-fold.  First, the proposed work will generate a better understanding of landsliding and its role in the carbon cycle and will help incorporate previously unaccounted carbon sources and sinks into national and global carbon budgets.  Second, in combination with previous mapping efforts it will help understand the long-term effects of climate change on landsliding and deforestation in a unique setting. 

   
  

• Objectives

  Open

   The long-term goal of this proposal is to characterize carbon pools and carbon fluxes associated with deforestation and landsliding in the Sierra de Las Minas (SLM), The specific objectives driving this proposal are a) to characterize the spatial distribution of landslides and land-cover types, and b) to establish the relationship between landslide deposits and the stream network.

• Work plan

  Open

   Addressing the two specific aims of this proposal involves three steps: a) compiling a database of known cover types and control points, b) classifying the images, and c) conducting a variety of spatial analyses in a GIS environment.  In turn, each of these steps will involve the development of various tasks described under Technical and Scientific Approach and Methods Proposed.

• Schedule

  Open

  Database compilation - January 2010 - March 2010

  Image classification - March 2010 - August 2010

  Spatial analyses - September 2010 - December 2010

• Technical and Scientific Approach & Methods Proposed

  Open

  a)   Cover types and control point databases.  Synthesis of work conducted in the SLM will help identify cover types, as well as locations where they have been reported.  The vast majority of this information is found in unpublished theses and reports.  The locations will be entered in a single geodatabase and will receive a unique ID.  Some of these locations may already bee in a shape file, but other may need to be digitized from maps or coordinates provided in tables.  In a separate file information on cover type, species composition, and author will be entered and identified with the corresponding ID. These databases can inform about the number of classes to be used in unsupervised classifications or be used in a supervised classification, as well as in the accuracy assessment of the classifications.

  b) Image classification.  This will involve pre-processing of the images, and the classification per se. Working in mountainous terrain is always a challenge given the rough characteristics of the terrain and the presence of shaded areas.  Ideally we would like to obtain orthorectified images, and therefore proceed with the atmospheric and terrain corrections, the latter to eliminate the effects of shadows. We will use a combination of bands, band indices, and ancillary data to run both unsupervised and supervised classifications.  Ancillary data includes geological substrate, elevation, and geomorphic setting.  We will evaluate the accuracy of the classifications using control points from the database, as well as ground truthing.  We will use ERDAS-Imagine to complete this step.

  c) Spatial analyses. Mapping landslide tracks includes the initiation and transport zone, and in some instances a scoured stream channel and deposits.  These will be mapped separately when possible.  The initiation zone is usually an area of complete scour whereas the transport zone is an area of scour, and in some instances of deposition along the landslide path.  This database will be used in combination of an ArcGIS Hydro data model, new GIS routines, and field work to examine the relationship between land use, transport and deposition of the material mobilized by landsliding, and the fluvial system.

• Results

  Open

  This project will generate a land-cover map, including one on landslides, and will become a crucial for estimating carbon budgets in the SLM.

• Deliverables

  Open

  Expected deliverables include land-cover and landslide maps and a variety of spatial analyses.

• Use of Satellite Imagery and GIS Solutions

  Open

  Since landsliding and deforestation affect vast areas that are usually remote and/or steep and thus difficult to access, the proposed work relies heavily on the analysis of remotely sensed data to map forests, land-use types, and landslides.  This will be accompanied by extensive use of spatial analyses in a GIS environment.

• Local Actions

  Open

  The proposed work will not only contribute to the synthesis of valuable data accumulated over the years by FDN and UV, but most importantly will result in the first map for the SLM based on the automated classification of remotely sensed data.  This map is a key input for the estimation of carbon pools and fluxes associated with landsliding, and will ultimately help FDN, the UV, and Guatemala to incorporate previously unaccounted carbon sources and sinks into regional and national carbon budgets.

   

  The local communities will also benefit from this project in that landslide mapping can help identify areas of increased hazard thereby contributing to planning efforts.

   

  The rich water resources of the SLM are beginning to be tapped for hydroelectric power.  Our work can be important in their planning efforts by helping identify areas of increased hazard.

• Miscellaneous information / interesting details on project

  Open

   The PI has recently received a National Science Foundation-NSF award to conduct her landslide-carbon research project in the Sierra de las Minas.  Two papers provide background information about the prevalence of landsliding in Central America and the importance of this process in mountainous regions worldwide:

  Restrepo, C.,  L. Walker, A. Shiels, R. Bussman, L. Claessens, S. Fisch, P. Lozano, G. Negi, L. Paolini, G. Poveda, C. Ramos-Scharron, M. Richter, E. Velazquez.  2009. Landsliding and its multi-scale influence on mountainscapes. BioScience 59:685-698. 

  Restrepo, C. and N. Alvarez.  2006. Landslides and their impact on land-cover in the mountains of Mexico and Central America.  Biotropica 38:446-457.

• Region Name

  Open
  Sierra de Las Minas

• Partners involved in project

  Open
  Fundacion Defensores de la Naturaleza (FDN) and Universidad del Valle (UV)