Smith, R, Adams, M, Maier, S, Craig, R, Kristina, A, Maling, I (2007). Estimating the area of stubble burning from the number of active fires detected by satellite. REMOTE SENSING OF ENVIRONMENT, 109(1), 95-106.
To estimate the area affected by stubble burning in southern Australia, use was made of observations from the MODIS (MODerate resolution Imaging Spectroradiometer) on the Terra and Aqua satellites. The burnt area (BA) was calculated from the number of active fires, known as fire hot spots (FHS) using parameters estimated from a survey of firms in the agricultural area of south-westem Australia. The study also served as a ground validation of the capability and limitations of the MODIS sensor and associated algorithm for detection of small agricultural fires. During the period from I March to 23 May 2005, 3240 unique FHS associated with stubble burning were detected. The majority of these FHS occurred in the afternoon in the last 3 weeks of April. To estimate the total area of stubble burning associated with these FHS, a survey was sent to 2066 farmers. This survey determined for each farm, the number of fields in which stubble was burnt, average size of field burnt (A), crop types burnt, dates and time of day of burning. Responses were received from 273 firms, 38% of whom reported over 500 stubble bums. The 3240 FHS were intersected with the polygons of the farm boundaries to determine the proportion (P) of stubble bums detected using MODIS. Only 13% (+/- 3%) of the stubble bums recorded in the farm survey were detected. Average field size burnt was 75 ha ( 6 ha). Total BA was calculated as: BA=A * 3240/P, which gave an estimated area of 1.87 million ha. This area was 27% of the total cropped area in south-westem Australia. This level of stubble burning was similar to that determined by a 1990s survey in the State of Victoria. Neither cloud cover nor field size was significantly correlated with the low number of stubble fires that were detected. Therefore it was concluded that many stubble bums went undetected because of the lack of coincidence between the time of the MODIS overpass and when stubble bums were initiated. Also the use of additional sensors such as the Advanced Very High Resolution Radiometer (AVHRR) on NOAA satellites with afternoon overpasses would improve the fraction of stubble bums detected. Across the whole of Australia where winter cropping occurs, there was a high correlation (r(2) = 0.96) between FHS and total cropped area in each State. This provided the basis for extrapolating the field results from south-western Australia, to estimate the total area of stubble burning in southern Australia for 2005. (C) 2006 Elsevier Inc. All rights reserved.