We present a comparison of techniques for estimating atmospheric emissions from fires using Australia's 2009 "Black Saturday" wildfires as a case study. Most of the fires started on Saturday the 7th of February 2009 (a date now known as "Black Saturday") and then spread rapidly, fanned by gale force winds, creating several firestorms and killing 173 people. The fires continued into early March, when rain and cooler conditions allowed the fires to be extinguished. In this study, we compare two new techniques (and one more established method) to estimate the total emissions of a number of atmospheric trace gases from these fires. One of the new techniques is a "bottom-up" technique that combines existing inventories of fuel loads, combustion efficiencies and emission factors with an estimate of burned area derived from MODIS rapid response daily fire counts. The other new method is a "top-down" approach using MODIS aerosol optical depth as a proxy for total amounts of trace gases emitted by the fires. There are significant differences between the estimates of emissions from these fires using the different methods, highlighting the uncertainties associated with fire emission estimates. These differences are discussed along with their likely causes and used as a vehicle to explore the merits of the different methods, and further constrain fire emissions in the future. (c) 2012 Elsevier Ltd. All rights reserved.