Pan-Tropical Mapping of Forest Cover

Uganda

The Woods Hole Research Center

In Depth Coverage

Pan-Tropical Mapping of Forest Cover and Associated Above-Ground Carbon Stock

The overall purpose of this project is to strengthen the participation of tropical and subtropical developing countries in the emerging carbon markets for avoiding or reducing emissions from deforestation and forest degradation (REDD). This is accomplished through two main outcomes:

  1. Generating pan-tropical forest cover map derived from the radar imagery as a baseline for subsequent change monitoring, and c) medium-resolution, pan-tropical biomass/carbon map based on the fusion of optical (MODIS), radar, and lidar (GLAS) data.
     
  2. Building national and regional capacity in developing countries through technology transfer, as well as through engagement in calibration and validation, and through integration of Outcome 1 data sets in the exploration of policy options.

Spot Imagery provided by Planet Action is essential to:

  1. calibrate the derived forest maps,
  2. better characterize forest degradation and biomass,
  3. create visual to increase the dialog between local communities / civil society on climatic change, forest & carbon management and the emerging carbon markets.

 

  • Link With Climate Change

    Open

    More than 25% of CO2 emissions are associated with land use change in the tropics. Without knowledge of the distribution of carbon stocks, countries are handicapped when attempting to identify the location and magnitude of carbon emissions associated with land use change. Using a combination of satellite imagery and fine scale forest inventories we will help tropical countries to develop their carbon monitoring system which will help them to get ready for the emerging carbon markets.

  • Objectives

    Open

    The Woods Hole Research Center is working to produce a high-resolution (ca. 20 m) pan-tropical map of forest cover (i.e., Americas, Africa, and South-East Asia) for the year 2007 derived  from Radar imagery (ALOS/PALSAR),  and a carbon map derived from MODIS imagery at 500m resolution.
    We propose to work in collaboration with tropical countries to test these new products for developing a forest and carbon monitoring systems in preparation for the emerging carbon markets under the REDD mechanism. One of our main partners is the Carbon Measurement Collaborative.
    We will focus on the following actions:

    1. Use of high resolution SPOT imagery to develop maps of forest distribution based on ALOS/PALSAR radar data
    2. Use of SPOT imagery to develop biomass maps and quantify forest degradation
    3. Build local, national, and regional capacity to use SPOT derived products with REDD stakeholders through workshops.

     

  • Work plan

    Open

    tbd

  • Schedule

    Open

    tbd

  • Technical and Scientific Approach & Methods Proposed

    Open

    1- Use of SPOT high-resolution data for calibration and validation of pan-tropical forest cover mapping activities

    To produce a pan-tropical, forest-cover product from the ALOS/PALSAR image data, reference information must be acquired for the purposes of training (i.e., calibration) and testing (i.e., validation) the landcover classification algorithms used in map generation.  This reference information (sometimes referred to as "ground truth") can be acquired from two primary sources.  The first source is from actual observations of current landcover conditions made in the field by trained personnel.  While field observations of landcover are generally very accurate, the information is often challenging and time consuming to obtain and, hence, tends to be cost prohibitive across large geographic extents (e.g., the pan-tropics).  The second source is from very high-resolution (< 2.5 m) passive optical satellite imagery.  The benefits of this second source include (1) reasonably high accuracy, (2) low relative cost per unit area, and (3) speed/efficiency of data collection.

    In order to collect a sufficient amount of training and testing information, we intend to compile a pan-tropical reference database derived principally from very high-resolution (< 2.5 m) optical imagery (e.g., SPOT/Quickbird/Ikonos) and/or aerial photography where available.  Very high-resolution optical imagery will be acquired for the 2007 time period to provide a spatially and temporally representative sample of forest cover conditions within each pan-tropical ecoregion.

    Image data acquired from the visible and near-infrared portions of the electromagnetic spectrum are required. Trained image analysts will collect sample points from each high-resolution scene using on-screen image interpretation and digitizing techniques. Sample points, and their associated class labels, i.e., forest, open water, etc., will be spatially linked with pixel-level data in the ALOS/PALSAR imagery. The resulting database will provide the basis for subsequent statistical classification and pan-tropical map generation.

    2- Use of SPOT high resolution data for calibration and validation of national biomass mapping activities:

    The Woods Hole Research Center is working to produce a pan-tropical map biomass map of (i.e., Americas, Africa, and South-East Asia) for the year 2007.  The map, derived from MODIS imagery acquired by the Acqua and Terra Land Observing Satellites, and will provide an unprecedented baseline for pan-tropical carbon monitoring at medium resolution. 

    Validation and calibration are essential components of remote sensing studies. In both cases, ground measurements are collected and then related to the remote sensing observations or model results. National level studies are commonly based on moderate resolution remote sensing and a mismatch exists between the sensor's field of view and the scale at which in situ measurements are collected. It is essential to properly scale-up field observations to match the spatial resolution of remote sensing data.

    High resolution remotely sensed data sets such as SPOT, KOMPSAT, Digital Globe have the capability to improve the characterization of land surfaces in terms of biomass by providing estimation of the spatial distribution of tree density at finer scale and also a better characterization of land cover and lands use types in very fragmented landscape. These data sets consequently provide an essential step toward the scaling up of biomass from field measurements to national and regional level. Specific methods describing how the SPOT imagery have been already used for calibration/validation for biomass mapping are provided in (Baccini et al, 2007, 2008). For this project we will select specific SPOT scenes based on a random stratified sample.

    Baccini A, Friedl M A,Woodcock C E & Zhu Z 2007. Scaling Field Data to  Calibrate and Validate Moderate Spatial Resolution Remote Sensing  Models. Photogrammetric Engineering and Remote Sensing 73(8), 945-954.

    Baccini, A., Laporte, N., Goetz, S. J., Sun, M., Dong, H., 2008. A first map of Tropical Africa's above-ground biomass derived from satellite imagery, Environmental Research Letters

  • Results

    Open
    1. Generating pan-tropical forest cover map derived from the radar imagery as a baseline for subsequent change monitoring,  using SPOT imagery for calibration and validation
    2. medium-resolution, pan-tropical biomass/carbon map based on the fusion of optical (MODIS), radar, and lidar (GLAS) data; using SPOT imagery for calibration
    3. building national and regional capacity in developing countries through technology transfer, as well as through engagement in calibration and validation, and through integration of 1), 2), & 3) in the exploration of policy options.
    4. Create visual to increase the dialog between local communities / civil society on climatic change, forest & carbon management and the emerging carbon markets.

  • Deliverables

    Open

    tbd

  • Use of Satellite Imagery and GIS Solutions

    Open

    tbd

  • Local Actions

    Open

    We propose to work in collaboration with tropical countries to develop forest and carbon monitoring systems in preparation of the emerging carbon markets under the REDD mechanism. One of our main partner is the Carbon Measurement Collaborative.

    The following actions:

    1. Test the use of SPOT imagery to develop radar maps of forest distribution.
    2. Test the use of SPOT imagery to develop map biomass and quantify forest degradation.
    3. Building local, national and regional capacity, to use SPOT derived products with the REDD stakeholders through workshops.

  • Miscellaneous information / interesting details on project

    Open

    The results from the first workshop organized in Uganda can be found at:

    http://www.whrc.org/africa/PAWAR/Workshops/index.htm

  • Region Name

    Open
    The project will focus on the tropical forests of the world, with use of SPOT imagery for selected countries. This project is targeting the Albertine Rift region where a training workshop will take place in Uganda. A description of the workshop is described at:
    http://www.whrc.org/test/africa/PAWAR/workshops/index.htm

  • Partners involved in project

    Open
    The Woods Hole Research Center, the National Forest Authority (NFA), the Uganda Wildlife Authority (UWA), the Jane Godall Institute, the Makerere University in Kampala, the Institute for Tropical Forest Conservation, Uganda and the "National Institut de la Conservation et de la Nature" in the Republic Democratic of Congo.