File attachment:
| Title | Improved Methods for Assessment and Prediction of Grassland Curing Satellite Based Curing Methods and Mapping - final report |
| Publication Type | Report |
| Year of Publication | 2010 |
| Authors | Newnham, GJ, Grant, IF, Martin, D, Anderson, SAJ |
| Abstract | The degree of grassland curing (the proportion of cured and/or dead material in a grassland fuel complex) has a significant effect on the ease of ignition, rate of spread and difficulty of suppression of grassfires. There is a critical need for accurate curing information as an input to Australian and New Zealand fire behaviour models and fire danger rating systems. Curing is currently assessed by networks of agency personnel using visual estimates which are spatially sparse, subjective and of limited accuracy. Satellite based curing maps have been produced for south-eastern Australia and Western Australia for several years, but these are based on research restricted to Victorian grasslands, in which curing was estimated visually or indirectly. The component of Project A1.4 reported here has developed satellite-based approaches to the mapping of curing that can provide timely, spatially explicit, objective curing maps covering all of Australia and New Zealand. The research was based on an extensive dataset of objective curing measurements collected over up to four seasons at 39 sites across Australia and 16 sites in New Zealand, with considerable support from fire agency personnel. The research focused on estimating curing with data from the recent MODIS satellite instrument, as a prototype of future satellite systems that will replace the AVHRR instrument on which current satellite curing maps are based. The project investigated multispectral and temporal approaches to minimising the influence of variations in the relative proportions of grass, evergreen vegetation and soil. Both approaches showed a modest performance benefit in curing estimation over the field dataset. However, the advantages of these methods are expected to be more apparent in regions with more mixed vegetation cover proportions than the project field sites. The performance of four candidate methods to map curing over the landscape was evaluated through agency feedback during a pilot trail conducted in southern Australia and New Zealand over the 2009/2010 curing season. On the basis of results at the project field sites and agency feedback, methods to map curing using MODIS satellite data are recommended for Australia and New Zealand. Compared to visual estimates, these satellite methods showed greater accuracy at the project field sites, and provide a basis for efficient, objective and spatially continuous curing maps. The project has surveyed end users on their requirements of an operational system in terms of map format, spatial resolution, frequency, timeliness and delivery mechanism. Given the urgent need for objective and spatially continuous curing information, it is recommended that a system to produce and distribute MODIS curing maps be set up within an agency experienced in the distribution of national scale satellite products. This will enable a more extensive evaluation by end users of their utility in an operational setting in upcoming fire seasons and allow any shortfalls not exposed by the field dataset to be addressed. The potential of the other methods identified by the project to provide alternative curing estimates should be borne in mind during this wider evaluation, particularly in semi-arid or wooded regions where only limited qualitative validation has been available. |
