Uncertainty characterisation and validation are critical phases to generate any Essential Climate Variable (ECV), and therefore both have been included as key deliverables of the ESA CCI programme [1]. All products generated by the CCI are required to have an associated per pixel uncertainty characterisation. This paper describes both the uncertainty characterisation framework and the related uncertainty validation exercise of the Fire-CCI project.

Data from wide field-of-view sensors have provided a wealth of information of the Earth's surface for many years. Of the more recent efforts, the National Aeronautics and Space Administration Earth Observing System Moderate Resolution Imaging Spectroradiometer (MODIS) instrument has been operational since 2000 for Terra (mid-2002 for Aqua). The wide field of view and high temporal frequency of MODIS provide near daily global coverage. However, this coverage comes at the cost of some level of spatial accuracy due to known effects of pixel size growth at increasing view zenith angles (VZAs). Further complicating the matter of variable viewing geometry is the application of a fixed grid for the geolocation of MODIS observations. An accurate understanding of this process is necessary to characterize the effective spatial resolution of daily MODIS gridded products. In this paper, we estimate the point spread function (PSF) of nominal 250-m MODIS gridded surface reflectance products from sequences of daily images over man-made large-size targets. Our results suggest that, in near optimal locations, the resolution of those MODIS gridded products varies between 344 and 835 m along rows and between 292 and 523 m along columns when the VZA ranges from 0° (nadir view) to 55 °, respectively. We also discuss some implications of the reliance on a global fixed grid, such as the MODIS sinusoidal grid, on the relation between location and spatial resolution.