TERN Ecosystem Surveillance Plots Kakadu National Park
Alex Held_Achievements of AusCover - TERN's remote sensing data facility
1. AusCover
The National Satellite Image-based
Biophysical Data and Ground Validation
Facility
2. AusCover Facility
Production and delivery of nationally consistent long‐time series of
satellite‐based biophysical map products and next generation remote
sensing research data that is validated for Australian conditions.
3. Capabilities to be Delivered
Remote Sensing Data Delivery Backbone
• Solving data formats, interoperability, data-policy,
• Physical storage, efficient delivery to end-users etc.
Data Production Network (via Central Console and 7 Regional “Nodes”)
• Nationally-consistent, standard biophysical data products,
• Metadata and technical support documents
• Specialised space-borne, airborne & in-situ research-grade data
Ground Validation Program and Instrumentation
• Set national standards & field validation protocols,
• TERN SuperSites and Environmental transects etc.
• International Benchmarking (eg CEOS WGCV LPV)
4. The AusCover Team & Network
Securing Data and Collaborations
via Links to International Agencies and
Networks
Darwin Node
Stefan Maier (CDU)
Brisbane Node
Kasper Johansen UQ)
Peter Scarth (DSITIA
Perth Node
Merv Lynch (Curtin)
Sydney Node
Alfredo Huete (UTS)
Adelaide Node
Megan Lewis (Adel. U.) Canberra Node
Alex Held (CSIRO)
7 Regional Nodes Melbourne Node Medhavy Thankappan (GA)
Ian Grant (BoM)
11 Institutions Simon Jones (RMIT)
28+ Researchers and
Students
..over 500 TBytes of
ground, airborne and
satellite data
5. Progress
• Extensive field campaigns and airborne data acquisitions in
2012. Six out of eight AusCover Test-sites covered
• Drafting first field validation manual “Green Book”.
• Consolidation of continental-scale datasets, associated
metadata, and portal across AusCover data system.
• Held several agency briefings across “TERN Themes” (Carbon,
Fire, Biodiversity..)
• Collaboration with Google Earth Outreach (“Earth Engine and
Smartphone apps) and IIASA (GEOwiki)
• Initiated establishment of demonstration “forest sensing and
phenocam network”
• Field validation program continuing for key data products (e.g.
aboveground biomass and GEOSS Ecosystems)
& many others..
8. AusCover Good Practice Guidelines, Contents
• Review of Validation Campaigns
• Representativeness and Sampling Design
• Sensor Calibration
– Vicarious calibration
– Atmospheric correction
– Radiometric
– Optical sensors
– Best practice guidelines for calibration and validation of SAR data and derived
biophysical products
• Validation of AusCover Biophysical RS products
– LAI, Fpar
– Fractional Cover
• Validation of key variables derived from biophysical RS products
– Phenology
– Leaf spectroscopy
– Foliar chemistry
– Persistent green vegetation index
– Biomass
• AusCover specifications and quality assurance steps for hyperspectral and LIDAR data
• Good Practice Field Data Management and Delivery
• National examples of spectral dataset exchange
• National examples of CalVal
• Appendices with protocols for field spectroscopy, FPC, LAI, etc.
9. Satellite and Airborne-Derived Data Products for
use in Ecosystem Science and NRM
National/Regional Scale Time-series
• Green Land-Cover (NDVI-EVI)
• Fractional Cover (green vegetation, non-photosynthetic
vegetation, bare ground)
• Fraction Absorbed PAR (Photosynthetically Active Radiation)
• Burnt Area
• Forest cover
• Land Surface Temperature
• Broad Ecosystem Maps
State-based (to be extended nationally)
• Fraction Projected Cover (FPC)
In-Situ Validation and TERN SuperSite Information
• Leaf-Area Index
• Vegetation 3D Structure & biomass
• Spectral Reflectance (nadir-corrected)
Science Data
• Airborne + Spaceborne hyperspectral
• Lidar
• SAR
13. Linking Climate and Land-Cover
Dynamics
Gridded Climate Data: http://data.auscover.org.au/xwiki/bin/view/Product+pages/WebHome
14. Dynamics of Fractional Cover Validation
(a CSIRO, ABARES C4C, and TERN AusCover Project)
• “Fractional Cover Product” – Used for land condition and trends, soil erosion
– Vegetation indices based unmixing of monitoring
photosynthetic vegetation, non- photosynthetic – Available from http://www-data.iwis.csiro.au/
vegetation and bare soil – Guerschman et al, 2009 (RSE)
– Based on MODIS reflectance (MCD43A4)
– Australia, 8-day, 500 meter, 2000-2012
– Semi-operational, updated weekly
15. Adoption by South
Australia of remotely
sensed soil exposure
data to improve soil
erosion risk monitoring
Green vegetation / non-photosynthetic vegetation
/ soil from a dry to a wet time, as measured by
the Relative Spectral Mixture Analysis index
(originally by Greg Okin, developed by AusCover
facilitated collaboration between Greg Okin and
AusCover SA)
17. Comparing Land-Cover dynamics with known land-
management practices across Australian Wildlife
Conservancy Sanctuary sites
NVIS classes
(protected and
non-protected)
MODIS satellite
data (protected
and non-protected)
Hummock Grassland
protected and non-
protected
MODIS EVI
Decadal MODIS time series
2000 2006 2011
18. Using International Data to Monitor Land-cover and productivity: e.g. Southern
Hemisphere Crop NDVI Anomaly, September 15th, 2012
Australia: Wheat
Brazil/Argentina:
Wheat
Crop NDVI Anomaly
South Africa: Wheat
Non Cropland
-0.4 0 0.4
Worse than normal Better than
normal normal
Current season crop development (2012)
Median season development (2000-2011)
22. Use of Terrestrial Scanning Lidar
to Characterise 3D Structure of Canopies
• Evaluating Ground-based lidar technologies designed specifically for forest and
vegetation assessment
• The CSIRO-patented Echidna® has key differences from scanning rangefinders
– Digitizes the full return waveform
– Has variable beam divergence
– Uses full hemispherical scanning and beyond
– Linear response and calibration
AGU 2012 22
23. Visualisation of Echidna Forest
Measurements in California
Gap probability to range (dark = 1, light = 0, green indicates partial return
Plate Carré (equal-angle, “Andrieu”) projection
AGU 2012 23
26. Demonstration of In-Situ Sensor Network at
Whroo Flux Tower Site - Victoria
• Ten-node wireless sensor network measuring
Photosynthetically Active Radiation (PAR),
temperature and humidity
• Five minute sample frequency
27. Following through the Fire
TERN Theme….
Dr Geoff Garrett (photo at left) speaks at the opening of the TERN–Queensland Government briefing on
fire risk, which was attended (photo at right) by Fergus Adrian (left) and Bernard Trembath (right) from the
Queensland Department of Fire and Emergency Services
28. Sample: TERN/AusCover Fire Products
Product Name Availability Validation
Burned Area – MODIS (CDU 250m), Australia Partial 2
Fire Frequency – MODIS (CDU 250m), Australia 2013/Q4
Fire Frequency – AVHRR (1km), Australia Available 1
Fire Severity – MODIS (CDU 250m), Australia 2013/Q4
Sub-pixel Fire Patchiness – MODIS (CDU 250m), Australia 2013/Q4
TOA Reflectance Red/NIR – MODIS (for manual burned area Partial 3
mapping), North and Central Australia
Thermal Anomalies (Fire Hotspots) – MODIS (LPDAAC Available 2
MOD14A2 1km), Australia – not near real-time
Burned Area – MODIS (LPDAAC MCD45A1 500m), Australia Available 1
Burned Area – MODIS (University of Maryland MCD64A1 Available 1
500m), Australia
1 Product accuracy is assessed from a small (typically < 30) set of locations and time periods by comparison with in-situ or other suitable reference data.
2 Product accuracy is estimated over a significant set of locations and time periods by comparison with reference in situ or other suitable reference data.
3 Uncertainties are characterized in a statistically robust way over multiple locations and time periods representing the complete range of environmental conditions.
29. Grassland curing
(Bureau of Meteorology)
• Bushfire CRC curing project maps (A, B, C, D)
• National maps and state cut-outs
32. Sample: Fire Research Enabled through TERN
Oliveira et al. (in preparation). Characterization of fire in tropical savannas of northern Australia using
satellite measurements of fire radiative power.
Edwards et al. (accepted). Remote sensing of fire severity in north Australian tropical savannas.
Price et al 2013. Role of weather and fuel in stopping fire spread in tropical savannas.
Maier et al. 2013. Sensitivity of the MODIS fire detection algorithm (MOD14) in the savanna region of the
Northern Territory, Australia.
Maier & Russell-Smith 2012. Measuring and Monitoring of Contemporary Fire Regimes in Australia using
Satellite Remote Sensing.
Maier et al. 2011. Characterising Bush Fires in Australia’s Top End using MODIS Active Fire Observations.
33. AusCover Facility Coordination Office
Contact Details
Alex Held – Facility Director
Rowena Smith – Facility Coordinator
Matt Paget – Data and Systems Coordinator
CSIRO Marine and Atmospheric Research
GPO Box 3023
Canberra, ACT 2601
Australia
Tel.: 61-2-6246 5630
Fax: 61-2-6246 5988
Email: alex.held@csiro.au / rowena.smith@csiro.au / matt.paget@csiro.au