Estimates of NOx Emission Factors from GOME-2 Measurements for the Major Types of Open Biomass Burning - Towards an Interdisciplinary Approach in Earth System Science - page 65

Fig. 2 Best fitting least-squares regression lines for evergreen broadleaf forest (light green),
woody savannas (orange), savannas (yellow), and croplands (red) within Africa north of Equator
(ANE)

NO x are estimated at 1.83, 1.48, 2.96, and 0.72 g kg − 1 for tropical forest, savanna
and grassland, crop residue, and boreal forest, respectively.
From the data in Fig. 3 it is apparent that the arithmetic means of satellite-based
EFs for tropical forest, crop residue, and boreal forest fall within one standard
deviation of means reported by Andreae and Merlet ( 2001 ) and/or Akagi et al.
( 2011 ). However, the EFs for savanna and grassland obtained in this study are
lower by a factor of 2.5 when compared to the other two values. A possible
explanation for this might be the seasonal cycle in NO x emissions rates (Mebust and
Cohen 2013 ). It is possible that the fuel for laboratory-controlled combustion
experiments has been mainly taken during that part of the

re season, where the
highest emission rates are expected. In contrast, the satellite-derived EFs are based
on the entire fire season, and thus, represent rather average values.

Fig. 3 Emission factors (EFs) of NO x (g kg − 1 ) for tropical forest (light green), savanna and
grassland (ochre), crop residue (red), and boreal forest (dark green). Satellite-based EFs are
represented by squares, reported values from Andreae and Merlet ( 2001 ) are indicated by a cross,
and the asterisk denotes EFs from Akagi et al. ( 2011 )