These calculations
require activity data at a scale of resolution finer than countrywide
(for example, statistics on crop areas for the various soil climatic
regions), and new emission factors that match these
scales of resolution. These inputs can be time consuming and
expensive to derive, and that iswhywewanted to identify the inputs
that had the most effect on the uncertainty in the total emissions.
We under took a sensitivity analysis to do this. Once we had identified
the inputs that influenced uncertainty the most, we explored
the effect of reducing their uncertainty by reducing the standard
deviation of the PDFs that we used to describe them by 50% in turn.