Basic HTML Version
3-74
ANNEX II – Limitations of the preliminary biomass carbon map for the DRC
The preliminary biomass carbon map generated has a number of limitations, for example
regarding the age and resolution of the underlying data, the way the data has been processed
and interpreted and the lack of field plot data to verify the final dataset.
Baccini
et al.
(2008) used MODIS Nadir bidirectional reflectance distribution function adjusted
reflectances (NBAR) data from the years 2000 to 2003 to generate a biomass dataset. However,
in the ten years since 2000, an estimated 3.4 million hectares of recent forest cover were lost
(Forêts d’Afrique Centrale Évaluées par Télédétection (FACET), 2010), a development that is not
reflected in Baccini
et al.
(2008). Forest degradation may have occurred as well since the year
2000, but can be more difficult to track from remotely sensed data. It is therefore uncertain how
well forest degradation from 2000 to 2003, is reflected in Baccini
et al.
(2008).
The resolution of the dataset generated by Baccini
et al.
(2008) is currently limited to 1km by
1 km. For use of the resulting map at national and sub-national scale this resolution may not be
sufficient.
In addition, and despite the fact that techniques for the collection and interpretation of remotely
sensed data are improving rapidly, errors can still occur in the generation and interpretation of
remotely sensed data. Baccini
et al.
(2008)
predicted above-ground biomass using a regression-
tree model and latest lidar metrics from the Geoscience Laser Altimetry System (GLAS), which
are sensitive to vegetation structure, to cross-validate the model results. However, data
interpretation needs to take into account a number of influencing factors, which can vary across
space, such as altitude of the terrain, and whose omission or inaccurate consideration can result
in misinterpretation of the data.
Since completion of the main analyses, two new biomass carbon datasets have become
available: Saatchi
et al.
(2011) published their benchmark map of forest carbon stocks in tropical
regions across three continents and Baccini
et al.
(2012) published an improved pantropical
carbon-density map.
Both the Baccini
et al.
(2008) and the Saatchi
et al.
(2011) datasets are based on remotely
sensed data from the early 2000’s, while Baccini
et al.
(2012) is based on data from the period
2007-2008. No field plot data from the DRC was used in Baccini
et al.
(2008) or Saatchi
et al.
(2011), while Baccini
et al.
(2012) used some field plot data from the DRC to characterize above-
ground biomass within GLAS footprints. Baccini
et al.
(2012) estimate the above-ground biomass
carbon in the DRC at about 22 Gt. This figure is higher than the figure provided by Saatchi
et al.
(2011) of 18.9 Gt above-ground biomass carbon in the DRC and the FAO FRA 2010 estimate of
19.6 Gt (FAO, 2010). According to the preliminary map presented in this report, which is based on