Vegetation Monitoring Products Derived from
NOAA Satellite Data

Monthly NDVI Browse Service

Normalised Difference Vegetation Index (NDVI) products are produced by the Bureau of Meteorology for the Australian region using measurements from the Advanced Very High Resolution Radiometer (AVHRR) on board the USA's NOAA polar orbiting meteorological satellites. The reflectance measured from Channel 1 (visible: 0.58 - 0.68 microns) and Channel 2 (near infrared: 0.725 - 1.0 microns) are used to calculate the index:

NDVI = (Ch2-Ch1)/(Ch2+Ch1)

Figure 1 below shows the spectral reflectance response characteristic for green vegetation, soil and water compared with the bandwidth of the AVHRR Channels 1 and 2 used to create the index.

Figure 1: Spectral reflectance characteristics of common earth surface materials
[Source: Remote Sensing Digital Image Analysis, J.Richards,1986 ]

The differential reflectance in these bands provide a means of monitoring density and vigour of green vegetation growth using the spectral reflectivity of solar radiation. Green leaves commonly have larger reflectances in the near infrared than in the visible range. As the leaves come under water stress, become diseased or die back, they become more yellow and reflect significantly less in the near infrared range. Clouds, water, and snow have larger reflectances in the visible than in the near infrared while the difference is almost zero for rock and bare soil. Vegetation NDVI typically ranges from 0.1 up to 0.6, with higher values associated with greater density and greenness of the plant canopy. Surrounding soil and rock values are close to zero while the differential for water bodies such as rivers and dams have the opposite trend to vegetation and the index is negative. A range of errors such as scattering by dust and aerosols, Rayleigh scattering, subpixel-sized clouds, plus large solar zenith angles and large scan angles all act to increase Ch1 with respect to Ch2 and reduce the computed NDVI.

NOAA-14's radiometer is a 5-channel instrument which scans continuously at a maximum ground resolution of 1.1km and swath width of approximately 2,400 km. Local Area Coverage (LAC) data received in Melbourne is used to produce NDVI values from each orbit. Typically 2 sequential daytime orbits covering most of Australia are available for processing in near realtime each day. Each pixel is an average of the differential reflectance over the range of vegetation, soil types, water bodies and other surfaces within the pixel footprint.

Maximum Value Composite NDVI

Monthly Maximum Value Composite (MVC) NDVI maps in mercator projection are produced by taking the of highest pixel value for the month from all the daily composites created from the individual orbits. This minimises data gaps in any particular composite due to cloud interference or missing data and overcomes some of the systematic errors that reduce the index value. The MVC NDVI maps have been archived since January 1997. Maximum composites for running 5-day and running satellite synchronous (SATSYNC) 9-day NDVI have also been trialed.

The Bureau interest in NDVI products is in their potential to support:

• derivation of related curing index maps for fire weather services support
• drought monitoring and climate services support for primary industry organisations
• hydrological applications such as flood monitoring for inland rivers.

A range of satellite derived vegetation indices have been widely used to classify land cover, estimate crop acreage, and detect plant stress. The USA National Oceananic and Atmospheric Administration (NOAA) produces global estimates of NDVI, in non realtime, based on the reduced resolution Global Area Coverage (GAC) data from the NOAA orbiting satellites. This data is used to study features such as continental land cover, global vegetation, primary production (eg crop monitoring) vegetation dynamics, desert encroachment, deforestation, bushfire scaring, forest fire potential, insect breading grounds and flooding. CSIRO also produce NDVI from higher resolution LANDSAT data, however the time between orbits (approximately every 16 days) precludes daily monitoring required to support some of the Bureau's potential applications.

Fortnightly and yearly NDVI images (Dept Environment and Heritage)

Maximum Value Composite Difference NDVI

The Bureau is also trialing Maximum Value Composite Differential (MVCD) NDVI products. These involve taking the difference between two consecutive monthly MVC maps (or the current and the previous 'year-month') to create monthly or annual MVCD maps. The examples below show time-sequential and regional trends across the continent.

May/June 1999 MVCD of NDVI
June 1998/1999 MVCD of NDVI

Daily, 5 Day or satellite synchronous 9 Day MVC NDVI maps can also been used to construct running MVCD maps for any required daily time period. MVCDs at these time steps are particularly relevant where vegetation growth response to a rainfall event occurs rapidly or the response is short lived. MVCD maps based on the 9 day satellite synchronous NDVI for a range of time periods are shown below:

1-Day Running SATSYNC-MVCD NDVI
2-Day Running SATSYNC-MVCD NDVI
3-Day Running SATSYNC-MVCD NDVI
4-Day Running SATSYNC-MVCD NDVI
5-Day Running SATSYNC-MVCD NDVI
6-Day Running SATSYNC-MVCD NDVI
7-Day Running SATSYNC-MVCD NDVI
8-Day Running SATSYNC-MVCD NDVI
9-Day Running SATSYNC-MVCD NDVI

Status of current systematic error correction methodologies

A cloud clearing algorithm is currently used to discard any pixel contaminated by cloud from the NDVI calculation.

Maximum Value Composite (MVC) techniques are used over a period of a month to overcome the following systematic errors that work to reduce NDVI value.

• Large solar zenith angles
• Large scan angles
• Atmospheric effects
• Cloud shadows

The MVC technique also overcomes any data gaps caused by missed data due to cloud coverage. Work is currently being done to evaluate the impact of other issues such as:

• Calibration drift in the radiometer sensor
• Navigation errors

so that further improvements can be implemented.

Other potential products

Grassland Curing Index (GCI)
The GCI has potential in supporting Bureau fire weather services. It uses the same basic algorithms as the NDVI and requires additional work on corrections for vegetation type using overlays to mask all except appropriate areas of grassland vegetation. However to produce maps for shorter periods than a month the geometry errors, that are currently avoided by using the MVC technique over a whole month, need to be reduced by introducing filters that restrict use of data based on large solar zenith or scan angles.

Examples of CGI products: SE Australia GCI Map | SW WA GCI Map (WASTAC)

A proposal: A Strategy to Improve Grassland Curing Information, June 1998, has been developed by a working group from the Victorian CFA, SA Country Fire Service, Victorian Dept of Natural Resources and Environment,WA Dept of Land Administration, Monash University Dept Geography, CSIRO Division of Atmospheric Research and the Bureau. The aim is to improve the quality, scope and availablity of GCI information to users on a national basis, including remotely sensed information.

Maximum Value Composite Difference (MVCD) NDVI products
MVCD NDVI products for specialised users could be developed. However, corrections for satellite radiometer calibration drift may be needed especially where the data used spans large time periods. Difference products are based on less than monthly (eg 5 day or 9 Day) NDVI also require additional geometry correction algorithms as mentioned above for GCI

Time series data
Time series monthly , 9 day or 5 day MVC NDVI data can be produced for specific locations. The quality of this type of product would be improved once navigation problems are corrected. Also time series based on the 9 day or 5 day composites NDVI require the additional geometry correction algorithms as well.

NDVI Grids (AXF format)
Another potential product for specialist users are the grids for Monthly, Satellite Synchronous 9 day, 5 day or daily NDVI data in format suitable for use input to GIS systems.

References

D. O'Brien, Monitoring grassland curing from space: A proposal for development of operational products and electronic distribution to Australian fire agencies, CSIRO Division of Atmospheric Research, 1998

J. Barber and G.W. Paltridge, Fuel Moisture Content of Vegetation from AVHRR - an Operational Fire Potential Monitoring System, Proceedings 1st Australian AVHRR Workshop October 22-24, 1986

K. B. Kidwell, NOAA Global Vegetation Index User's Guide, July 1997, U.S. Department of Commerce NOAA National Environmental Satellite data and Information Service.

R. Johnston, Applications of AVHRR in Agriculture- A Review, Proceedings 7th Australian Remote Sensing Conference, 1994

R. Hosking, Grassland Curing Index- A district Model that allows Forecasting, National Parks and Wildlife Service, Braidwood NSW.

W. G. Tuddenham and D. Griersmith NOAA satellite data for flood monitoring

W. G Tuddenham, J.F. LeMarshall, B.J. Rouse and E.E. Ebert, The Real Time Generation and Processing of NDVI from NOAA-11: A perspective view from the Bureau of Meteorology, Proceedings 7th Australian Remote Sensing Conference, 1994