REGIONAL FIRE MONITORING REQUIREMENTS IN SOUTH EAST ASIA

1
REGIONAL FIRE MONITORING REQUIREMENTS IN SOUTH
EAST ASIA

• Mastura Mahmud
• Universiti Kebangsaan Malaysia
• Malaysia

2
Forest Fires in SEA

• Fire is an indicator of land use and land cover
  changes. The rapid conversions of forests and
  deforestation has been occurring for the past few
  decades in South East Asia. This is the region
  where large areas of forests still exist. Further
  deforestation is expected in the future because
  of socio-economic pressures.

3
Continue

• Rapid rate of conversion of forests to
  agriculture, settlement and industrial areas put
  a high demand on land requirement.
• The change in human population and land use
  practices can affect the fire practices. Slash
  and burn agriculture is currently one of the main
  reasons of large scale fire activities in the SEA
  forests.
• The 1997/98 forest fires in Indonesia caught the
  worlds attention due to its magnitude and
  transboundary haze that clouded most of SEA.
  Climate variability such as the El Nino dry
  period also exacerbated the conditions.

4
Forest Coverage of the World Source World
Resource Institute

• Original forest cover
• Current forest cover
• Remaining frontier forest

5
Forest Fire Requirements in South East Asia

• A one stop shop center web site (SEARIN)
  utilizes the GOFC forest cover characteristics
  and changes information in the SEARIN region, and
  links to other sites for accurate, processed
  active fires and burnt scars that can be assessed
  freely of various spatial scale lower resolution
  (NOAA, DMSP or GMS) and higher resolutions
  (Landsat, SPOT, MODIS etc.). ?
• Operational products should be easily
  accessible a user friendly website for various
  users (researchers, public, government agencies).
  Improvement of data archiving is needed,
  especially the historical data of fire activities
  in relation to land use changes especially in the
  developing countries of SEA.
  ?

6
Continue 1

• Mapping burn scars areas should include a suite
  of satellites that can improve statistics of
  burned area and can validate burned area products
  
• Low resolution geostationary (GOES, GMS,
  MTSAT) and polar orbiters (NOAA AVHRR, DMSP)
• Moderate spatial resolution (SPOT-Vegetation,
  ASTR, MODIS, SeaWifs)
• Higher resolution Landsat and SPOT.

7
Continue 2

• Fire information e.g. detection, mapping of
  active fires, burn scars products are needed as
  well as estimates of aerosols and gas emissions
  provided as operational products.
• For developing countries, fire mappings and
  monitoring provide a means to
• Implement ghg inventories and for estimations of
  Carbon emissions (emission factors and emission
  rates)
• Improvement on the spatial resolution of current
  products

8
Continue 4

•   Operational automated burned area mapping
  system and data pre-processing, operational
  quality assessment is required for SEA. Other
  fire products that could be included should be
  fuel load, moisture content, fire intensity, fuel
  consumption.
•           Many GOES, DMSP, NOAA unprocessed
  historical archives are needed to study the
  history, trends and fire returns frequencies for
  climate change, carbon emission and land use
  change conversions and activities.
•      Near real time delivery of active fire (and
  burn scars), fire intensity, rate and direction
  of spread, smoke and post fire damage assessments
  information is important to researchers, the
  general public and management agencies. Fire
  detection needs to be robust, fast, accurate and
  automatic.

9
Continue 5

• Rapid response information for fire managers and
  control (high accuracy of locations and
  timeliness of data delivery).
• In situ data is needed to provide emission
  factors and validation of fuel loads (for trace
  gases and particulate emissions).
• Automated burned area mapping in algorithm for
  operation.
• Temporal scale requirements vary (fires burn for
  several days or are short lived). In tropical
  systems, fires have diurnal cycle and burn scars
  are short lived over a few weeks.
• Geostationary satellites can capture the diurnal
  cycle of fires, but polar orbiters are limited to
  at least only 2 images per day. There is no more
  fire detection on NOAA 16.
• Monthly, seasonal or annual emissions estimates
  are needed. Time series data collected
  periodically during burn seasons are needed.
  Emissions are important for FCCC interannual
  variability or trends in fire activity.

10
Fire Monitoring Facilities in SEA
11
Continue
12
Continue
13
Under the ASEAN Haze Action Plan, forest fire
monitoring is carried out by the Singapore Met.
Services (hotspot locations and smoke plumes NOAA
AVHRR detection) over the web. Detection areas
only cover Indonesia, Malaysia and Singapore,
excluding Thailand, Philippines and Indochina.
Detection of hotspots, smoke plumes and burnt
area scars are done individually by the relevant
governmental agencies in each country in SEA.
14
Opportunities for compiling a common database web
site

• Current available websites
• (I) SEARIN http//www.eoc.ukm.my/searin, with
  links to other sites which include hotspots
  images from MODIS, NOAA, SPOT, JRC and UMD (Open
  GIS historical fire locations).
• (ii) Asian Institute of Technology (Asian Centre
  for Remote Sensing) in Thailand at  
  http//www.acrors.ait.ac.th , available to
  Thailand, Philippines, Indochina for AVHRR images
  (with limitations).
• (ii) Improvement on present information
  needed Websites such as World Fire Web, Global
  Fire Monitoring Center (GFMC) comprehensive,
  but for the SEA region it focus on Indonesia and
  Malaysia (due to coverage of the NOAA image,
  obtained from the Singapore Met Services).
• 2 sites that locates the hotspot sites
  in Indonesia
• (a) Sumatra http//www.mdp.co.id/ffpc
  p.htm
• (b) Kalimantan (Integrated Forest
  Fire Management Project (IFFM/GTZ))
• (iii) GOIN (Bangkok) for a proposed
  DMSP-OLS image for SEA available in April 2000

15
Opportunities for development of a harmonized
dataset/map from existing programs/products in
each country

• Humid tropical equatorial forests have different
  times of fire regimes than the drier subtropical
  to tropical forests. So monitoring regimes of
  fires from Indonesia Malaysia (August-October)
  is different from the burning that occurs from
  April onwards in Vietnam or Thailand.
• Existing centres such as AIT and Singapore can
  harmonise these data to coincide with the fire
  season. For South Vietnam Indochina fire season
  is from April to June Indonesia Malaysia from
  July to October.

16
Prototype information network

• Spatial extent Area of fire monitoring should
  include SEA nations.
• User communities to be served Forestry
  agencies, government agencies, policy makers,
  research institutions, NGOs.
• Define user needs Burnt areas, active forest
  fires that transmit information on hotspots,
  plumes, burn scars at near-real time. Active
  fires useful for fighting teams forest managers
  in different regions in SEA. Higher revisit cycle
  needed to monitor variability in burning periods.
• Social impacts In terms of health, safety,
  deprivation of livelihood, relocation of villages
  driven from homes, transportation, biological
  impact, loss of biodiversity.
• Atmospheric impact air quality, aerosols,
  smoke plumes, ghg emissions, climate change etc.

17
Obstacles to data access

• Data available from the net, but limited in space
  resolution and temporal frequencies and burn scar
  areas. GOFC can facilitate relationships between
  space providers, processors and users.
• Lack of internet and network facilities (Vietnam
  / Cambodia) and expertise and openness of
  government policies.
• High operation costs information considered
  confidential by government agencies.

18
How GOFC can assist GOFC-SEA in capacity building

• Data collection and compilation
• Facilitates instruction manual and "off-line"
  help, provision of tool kit on data collection
  and compilation of historical data, logistics for
  inter-comparison of data accuracies from
  different sensor/satellites of the same events
  detected.
• Scientific interpretation, reference information.
• Reporting, validation protocols, quality
  documentation on final products.
• Training on calibration, validation, algorithms
  and accuracies of available products.

19
Regional Level Strategy

• Harmonize and develop regional base map
• Regional map level 1 (base map) exists.
  However, forest characteristics for base map is
  coarse, comprising of simple classification of 6
  categories. Level 2 data needed and standardised
  and validated (upland, swamps, mangroves,
  disturbed forests) relevant for forest fire
  monitoring. Useful for overlay of forest fire
  locations.
• Develop One-stop-shopping Regional Catalog
  through the SEARIN website links
• Conduct training on analysis methods
• Characterization, fire monitoring, bio
  processes mapping.
• Develop Policy Brief Products
• Inform policy makers regarding active
  fires by sending report that include
  recommendations esp. high risk areas (vulnerable,
  conservation) through ASEAN head of state
  meetings. Lobbies by NGOs and mass media to
  influence the policy makers.

20
Pilot Project Recommendations

• Data accuracy, ground truthing and validation of
  remote sensing products should be promoted
  through pilot projects in the region.
• Risk assessment database map for vegetation cover
  in watershed area, national park, peat swamp,
  protected areas etc. of potential fires and
  active recurring fires and assessed burnt scars
  at national level from superimposition of active
  fires on the LUC areas.
• Social impacts of forest fire health, air
  quality (at national level).

21
Proposed Projects for SEARIN
22
Suggested long term fire validation sites
(SEARIN) Other suggestions welcome
23
Other suggested long term calibration /
validation sites in SEA (Lisbon 2001)

• East Kalimantan (dipterocarp, degraded
  rainforest)
• Thailand Hue Khai Kheng
• Philippines Baguio (pine forest)
• Cambodia / Laos border (forest, cropland)

24
National and Regional Information Needs

• Rate of deforestation (fire related) needed for
  land use and carbon estimation purposes.
• Estimation of emissions and gases for climate
  change, human health, air quality impact studies.
• Near real time images of hotspots for policy
  makers and fire managers.
• Burn area maps

25
(No Transcript)
26
SEA Users Needs to GOFC

• Define resolutions required
• High, medium, low, hyperspectral?
• Define frequencies of images required for
  monitoring per day
• Every hour ? (MTSAT, low resolution)
• Specify number of images per day required.

27

•          Limitations of 1 km resolution
  NOAA-AVHRR data
• Coarse locations of the hotspots can be detected,
  but does not provide reliable estimates of the
  geographical extent of the fire activity.
•  The limited 2 - 4 passes per day is engulfed
  with cloud contamination.
•   Low frequency of satellite overpass, poor
  spatial resolution, sensor saturation etc.
•   Number of hotspots from the AVHRR data differs
  from the ESRIN data (due to their frequencies).

28

• Limitations (cost, access, accuracy of data)

• Gaps of data continuity fire activity
  incomplete (daily variabilities)
• AVHRR fire activity monitoring coarse
  resolution, daily overpass, clouds
• SPOT VGT mapping and monitoring high cost, low
  resolution, clouds
• Landsat TM cost, cloud, automated algorithm
  needed.
• RS/GIS training interpretation of area burned.
• Overcome limitations automated algorithm to
  interpret area burned, enhanced coordination
  amongst data suppliers (Govt agencies limited
  co-operation, NGOs, donor agencies).
• AVHRR hot spot mapping needs improved algorithm
  to reduce false alarms and cloud contamination.