Docstoc

Remote Sensing in the Forest Service - PowerPoint

Document Sample
Remote Sensing in the Forest Service - PowerPoint Powered By Docstoc
					USDA Forest Service Remote Sensing Applications Center
Operational Satellite-Based Remote Sensing Fire Support
                        Programs




                  WRAP/FETS Project Meeting
                         Boise, Idaho
                 August 31-September 1, 2009
Remote Sensing Applications Center (RSAC)
• National Technical Center
    •   Detached Washington Office – Engineering
    •   Geospatial Management Organization (GMO)

• Mission: Provide national assistance to agency units and deputy
  areas in applying the most advanced geospatial technology
  toward improved monitoring and mapping of natural resources
    USDA Forest Service Organization
                         National Headquarters in Washington, DC
                         9 Regional Offices
                         National Forest System
                           155 National Forests and 20 National Grasslands
                              600 Ranger Districts
                           193 million acres in CONUS, AK and PR
   Research
      7 stations
      50 field offices
   State and Private
    Forestry
      Fire & Aviation
      Forest Health
       Protection
   International
    Programs
Remote Sensing Applications Center (RSAC)
Rapid Disturbance Assessment and Services Program
    Tactical and strategic remote sensing support Forest Service
     units and external partners in response to disturbance events
    Includes operational active wildland fire management and
     post-fire assessment support programs/projects
RSAC Areas of Interest for Operational Air Quality
Management Support
    Provide high temporal observations/value-added products from
     various satellites (MODIS, AURA, VIIRS, etc.) to support USFS
     and other agency air quality managers
       Identify source/receptor relationships
           Exceptional events in relation to states
           Natural and anthropogenic sources of smoke relative to Class 1 areas
           Rx fire effects on urban areas

    Provide satellite-based ozone monitoring products, particularly in
     the south Sierras and the southeast U.S.
       Integrate/analyze results with ongoing USFS ground-based ozone
        monitoring efforts

    Provide fire detection, burn scar and burn severity geospatial
     data products (moderate and high resolution) for existing and
     future emissions tracking/forecasting applications
       WRAP/FETS
       Blue Sky Smoke Modeling Framework
    RSAC Satellite-Based Fire Mapping Programs
Active Fire Mapping
    MODIS Direct Readout/Active Fire Mapping Program
            Near real-time fire detection, characterization and monitoring
            Support interagency strategic fire management and planning

Post-fire Mapping
    Monitoring Trends in Burn Severity (MTBS)
            Assess frequency, extent and severity of all current/historical large
             wildland fires
            Monitor national fire policy effectiveness
    Burned Area Emergency Response (BAER) Imagery Support
            Emergency assessment of burn severity
            Support emergency stabilization activities by USFS BAER teams
    Rapid Assessment of Vegetation Condition after Wildfire (RAVG)
            Rapid assessment of forest damage
            Revegetation decision support data/information for USFS forest
             silviculturists
Forest Service MODIS Active Fire Mapping Program
   Satellite-based operational detection
    and monitoring of wildland fire
    activity in CONUS, Alaska, Hawaii &
    Canada

• Generate and provide "value added”
    fire mapping and visualization
    products, and geospatial data

• Facilitates decision support for
    strategic wildfire planning and
    response for U.S. and Canadian fire
    agencies
         Prioritize allocation of fire suppression
          assets
         Focus tactical airborne reconnaissance
          assets
         Supports several fire-related applications
          and decision support systems
  Forest Service MODIS Active Fire Mapping Program
  Currently Leveraged Sensor Assets
  Sensor         Platform         Spatial            Temporal       Fire Algorithm       Data Source
                   Type          Resolution         Resolution
                                 (Reflectance/TIR        (per
                                      Bands)         instrument)

  MODIS        Polar orbiting    250m, 500m,        2 times daily   MOD14/MYD14         Direct Readout;
                                   1km/1km                                               NASA Rapid
                                                                                           Response
                                                                                            System
  AVHRR        Polar orbiting      1km/1km          2 times daily       FIMMA           NOAA NESDIS


   GOES        Geostationary       1km/4km            4 times         WF-ABBA           NOAA NESDIS
                                                      hourly
  VIIRS*       Polar orbiting     375m/750m         2 times daily        TBD            Direct Readout;
                                                                                             Rapid
                                                                                           Response
 GOES-R#       Geostationary        500m,             4 times            TBD            Direct Readout;
                                   1km/2km            hourly                            NOAA NESDIS


* VIIRS launch on NPOESS Preparatory Project (NPP) mission in June 2011 and subsequent NPOESS missions
# GOES-R launch in 2015 and subsequent missions
MODIS Fire & Thermal Anomalies (MOD14/MYD14)

                                         •   Detects & characterizes fire at
                                             1km spatial resolution at time
                                             of satellite overpass

                                         •   Leverages response of 4m and
                                             11m bands to fire

                                         •   Absolute thresholds and
                                             contextual analysis

                                         •   Fire detection is affected by
                                             several variables
                                                  Atmospheric conditions, view angle,
                                                   land cover, overpass time relative to
                                                   fire activity, etc.

                                         •   Fire activity smaller than 1km
                                             can be detected
                    Central Idaho                 100m2 flaming fire (50%
                    Aug 12, 2007
                    Central Idaho                  probability)
                      Color Composite
               False Central 2007
                    Aug 12, Idaho                 50m2 flaming fire in ideal viewing
                   (1-KM 12, 2007
               False Aug detection
                      Color Composite              conditions (~100% probability)
                 True Color inyellow)
                 centroids inComposite
               (Active fires orange)
                AFM Operational Fire Product Deliverables
            Value-Added Mapping/Visualization Products and Geospatial Data
            • Printable regional fire detection
               maps
            • Interactive web fire detection maps
            • Google Earth fire KML/KMZ products
            • Fire detection GIS datasets
            • True & False color image subsets
            • Fire detection analysis/summary
               products




Northern California – June 27, 2008
A Look At Recent Fire Activity With MODIS

   Southern California – August 29, 2009 1515 MDT
   Aqua MODIS True Color Composite
A Look At Recent Fire Activity With MODIS

   Southern California – August 29, 2009 1515 MDT
   Aqua MODIS False Color Composite
A Look At Recent Fire Activity With MODIS
Station Fire – August 31, 2009 1300 MDT
    RSAC Satellite-Based Fire Mapping Programs
Active Fire Mapping
    MODIS Direct Readout/Active Fire Mapping Program
            Near real-time fire detection, characterization and monitoring
            Support interagency strategic fire management and planning

Post-fire Mapping
    Monitoring Trends in Burn Severity (MTBS)
            Assess frequency, extent and severity of all current/historical large
             wildland fires
            Monitor national fire policy effectiveness
    Burned Area Emergency Response (BAER) Imagery Support
            Emergency assessment of burn severity
            Support emergency stabilization activities by USFS BAER teams
    Rapid Assessment of Vegetation Condition after Wildfire (RAVG)
            Rapid assessment of forest damage
            Revegetation decision support data/information for USFS forest
             silviculturists
MTBS Project Overview
   Consistently map the location, extent and burn severity of
    large fires on all lands in the CONUS, AK, HI and PR from
    1984 and 2010
     > 1,000 acres in the western United States
     > 500 acres in the eastern United States

   Initiated in FY 2006
   Sponsored by the interagency Wildland Fire Leadership
    Council (WFLC)
     One element of a strategy monitoring the effectiveness of NFP and
      HFRA

   Jointly implemented by USFS RSAC and USGS EROS
     Equal funding from USDA-FS and DOI
Monitoring Trends in Burn Severity
Leveraged Sensor Assets
    Sensor         Platform Type           Spatial             Temporal          Data Source
                                          Resolution          Resolution
                                        (Reflectance Bands)   (per instrument)

 Landsat 4 TM        Polar orbiting            30m               16 days         USGS EROS
 Landsat 5 TM        Polar orbiting            30m               16 days         USGS EROS
   Landsat 7         Polar orbiting            30m               16 days         USGS EROS
     ETM+
           1
    LDCM             Polar orbiting            30m               16 days         USGS EROS

1 – LDCM launch currently scheduled for December 2012

      Why Landsat???
       Operational precedent
             NPS fire atlases
             USFS/DOI BAER Programs
         Landsat TM/ETM data record (1984 to present)
             Consistent data record spanning ecologically and possibly climatically
              significant time frame
         Resolution synergy
             Spatial and spectral resolutions comparable to other national scale
              data
Burn Severity Mapping Primer
Exploiting Spectral Response Curves…




    Visible     NIR                    SWIR
            Burn Severity Mapping Primer
               Burn scar delineation and burn severity characterization
                are based on Landsat TM/ETM image pairs and the
                differenced Normalized Burn Ratio (dNBR)
               See “Landscape Assessment” section in Fire Effects
                Monitoring and Inventory Protocol (FIREMON)
                Landsat           NBR
Pre-fire




                                                         dNBR              Burn Severity


                       6/8/2005
                                        Difference
Post-fire




                                                        Normalized Burn Ratio (NBR)
                                                     NBR = (NIR – SWIR) / (NIR + SWIR)
                      6/14/2007
                                                 Differenced Normalized Burn Ratio (dNBR)
                                                        dNBR = Pre NBR – Post NBR
MTBS Fire Mapping Methods Overview
    Compile a single MTBS fire occurrence database (FOD)
     from existing data sources
      28,000+ fire occurrences between 1984-2008
      Numerous duplicates, spatial anomalies, etc
    Based on FOD and prescribed assessment strategy,
     select pre and post-fire Landsat scenes
      Determined by primarily biophysical setting of fire
    Landsat TM/ETM+ data ordered and acquired from
     EROS
      All Landsat data are free (since late 2008)
    Perform necessary pre-processing, create and subset
     dNBR images, burn scar delineation, threshold dNBR
     images into burn severity classes
    Metadata, map products, burn severity GIS data
     analysis, and reporting summary
MTBS Fire Occurrence Locations




                                     1984-2008
                                 28,000+ target fires
                                 Numerous duplicates
                                 and28,000+ fires
                                    spatial anomalies
                                   for 1984-2008
MTBS Methods – Assessment Strategy
    Based on fire type
      Extended Assessment (EA)
          Severity based on post-fire assessment at peak of green of next
           growing season
          Forests/shrublands
      Initial Assessment (IA)
          Severity based on immediate post-fire assessment
          Grasslands/shrublands
      “Single Scene” Assessment
          Lack of suitable pre-fire imagery or other factors; use post-fire
           NBR
          Conducted on a limited basis (EAs and IAs)
MTBS Geospatial Products
 Fire Level Datasets                        Prefire Image           Postfire Image


    Available at
     http://www.mtbs.gov
    Pre/Post-fire Landsat imagery
       Bands 1-5, 7
    Burn scar boundary
       Vector delineation of burned
        area extent based on image
        analysis
                                         Burn Severity Indices   Thematic Burn Severity

    Continuous burn severity indices                            Unburned
                                                                 to Low

        dNBR/RdNBR                                              Low
                                                                 Moderate
    Thematic burn severity data                                 High

       6 classes (unburned to low,
        low, moderate, high, increased
        vegetation response, non-
        mappable areas)
MTBS Production Status




                            1984-2007
                           ~9,600 fires
                            completed
                         81.6 million acres
    MTBS Geospatial Products
Composite MTBS Fire Datasets
 National
      Burn Scar Boundaries
          ESRI Shapefile w/metadata
          Fire attributes
      Fire Occurrence Database
          ESRI Shapefile w/metadata
          Geographic centroid of burn
           scar boundaries for each fire
          Fire and MTBS processing
           attributes


   Regional
      Thematic Burn Severity
       Image Mosaics
          6 class thematic GeoTiffs
           w/metadata
          Available by MTBS mapping
           zone/year
                                           2005 burn severity mosaics by MTBS zones
    Burned Area Emergency Response (BAER)
   Fast track emergency assessment
     • Fires ranging from 100s of acres to 100s of
       thousands of acres in size

• Assess fire effects on the soil and watershed
  hydrologic function (erosion and flood
  potential)
• Prescribe and implement emergency
  stabilization measures to mitigate potential
  hazards to:
                                                     Straw waddles on Pike-San Isabel NF
     • Life
     • Property
     • Long-term soil productivity
     • Water quality
     • Natural resources

• BAER response plan is required within 7 days
  of fire containment                                    Contour logging on Shasta-Trinity NF


• Remote sensing support is critical in
  generating BAER team response plans
    Burned Area Emergency Response Imagery Support
• Provide rapid delivery of imagery and derived                 Prefire
    products to Forest Service BAER teams
         Burned Area Reflectance Classification (BARC),
          preliminary vegetation burn severity, based on
          dNBR/dNDVI assessment
         BAER teams use BARC with field observations and
          other spatial data to generate a soil burn severity
          map
                                                                Postfire
   RSAC remote sensing support provided at or
    immediately after fire containment
         Emergency assessment

   RSAC provides imagery and data products
    within 24 hours of image acquisition
                                                                BARC
   532 fires completed since 2001 (13.6 million
    acres)
   USFS and DOI BAER support programs
    combined -736 fires completed since 2001
    (25.7 million acres)
Burned Area Emergency Response Imagery Support
Leveraged Sensor Assets – Use Best Available
    Sensor         Platform Type           Spatial              Temporal              Data Source
                                          Resolution          Resolution (per
                                       (Reflectance Bands)       instrument)

 Landsat 5 TM       Polar orbiting            30m                 16 days             USGS EROS
   Landsat 7        Polar orbiting            30m                 16 days             USGS EROS
     ETM+
    AWiFS           Polar orbiting            56m                  5 days            USDA-FAS-SIA


    SPOT 4          Polar orbiting            20m                 2-3 days            SPOT Image
                                                                 (pointable)
    SPOT 5          Polar orbiting         10m/20m                2-3 days            SPOT Image
                                                                 (pointable)
            1
    ASTER           Polar orbiting         15m/30m                4-16 days           NASA/USGS
                                                                 (pointable)            EROS
  NASA AMS         Airborne (UAV)            ~ 21m                    --                  NASA

           2
    LDCM            Polar orbiting            30m                 16 days             USGS EROS

1 – Anomalous SWIR band issues since April 2008; 2 - LDCM launch currently scheduled for December 2012
 Creation of the BARC
Prefire
Postfire
dNBR
BARC
                                Black Pine 2 Fire
                                  Sawtooth NF
                                  73,000 Acres
                        Normalized Burn Ratio (NBR)
                        Differenced Normalized Burn Ratio (dNBR)
                        NBR = (NIR – SWIR) / (NIR + SWIR)
                        dNBR = Pre NBR – Post NBR




                        Normalized Difference Vegetation Index (NDVI)
                        Differenced NDVI (dNDVI)
                        NDVI = (NIR – Red) / (NIR + Red)
                        dNDVI = Pre NDVI – Post NDVI
Unburned to Low
Low
Moderate
                          dNDVI is utilized when appropriate
High
                          SWIR band is not available
Burned Area Emergency Response Imagery Support
BAER Support Rapid Delivery Products
Prefire Image         Postfire Image




                                         3D Visualizations



dNBR or dNDVI Image   BARC Image




                                       Map Products
    Post-Fire Vegetation Management on National Forest
    System Lands
   Agency resources to support reforestation in burned areas
    are limited, consequently, better prioritization is needed
   USFS Region 5 facilitated a protocol for rapid identification
    of deforestation and reforestation need on National Forest
    System lands
      Leverage   relationship of satellite-based burn severity observations
       to Composite Burn Index (CBI) field plot data that quantify fire
       effects to vegetation (deforestation)
      Spatially   represent forested vs. deforested areas following wildfire
            Based on Landfire EVT
      Silviculturistsuse results as a tool to identify and prioritize land
       suitable for reforestation
            Natural Recovery
            Assisted Recovery (planting, seeding or site preparation for natural
             recovery)
    Rapid Assessment of Vegetation Condition after Wildfire
    (RAVG)
• Provide Landsat imagery and derived products                 Postfire Image
    characterizing fire effects on forest resources
    to silviculturists (based on USFS protocol)
         Assessment based on RdNBR – a relative version of
          dNBR
         Regression models based on CBI field plot data and
          RdNBR used to derive post-fire forest condition
          layers:                                              RdNBR
               Burn severity
               % basal area loss
               % canopy cover loss
   Conducted on fires where > 1,000 acres of
    NFS forested land is affected
   RAVG data and summary products are                         7 Class Basal Area Mortality
    provided within 30 days of fire containment
   142 fires completed since 2007 (~4.5 million
    acres)
Rapid Assessment of Vegetation Condition after Wildfire
Primary RAVG Product Deliverables
• Prefire Landsat image    • % Change in Basal Area
• Postfire Landsat image   • % Change in Canopy Cover
• Burn Scar Boundary       • Burn Severity
• dNBR and RdNBR           • Forested/Deforested map and tabular summaries
                             by land status and slope
Additional Information
          RSAC Fire Mapping Support Programs

           MODIS Active Fire Mapping Program
              http://activefiremaps.fs.fed.us

        Monitoring Trends in Burn Severity Project
                   http://www.mtbs.gov

             BAER Imagery Support Program
             http://www.fs.fed.us/eng/rsac/baer

    Rapid Assessment of Vegetation after Fire Program
      http://www.fs.fed.us/postfirevegcondition/index.php