Charcoal production for urban energy consumption is a main driver of forest degradation in sub Saharan Africa. Urban growth projections for the continent suggest that the relevance of this process will increase in the coming decades. Forest degradation associated to charcoal production is difficult to monitor and commonly overlooked and underrepresented in forest cover change and carbon emission estimates. We use a multitemporal dataset of very high-resolution remote sensing images to map kiln locations in a representative study area of tropical woodlands in central Mozambique. The resulting maps provided a characterization of the spatial extent and temporal dynamics of charcoal production. Using an indirect approach we combine kiln maps and field information on charcoal making to describe the magnitude and intensity of forest degradation linked to charcoal production, including aboveground biomass and carbon emissions. Our findings reveal that forest degradation associated to charcoal production in the study area is largely independent from deforestation driven by agricultural expansion and that its impact on forest cover change is in the same order of magnitude as deforestation. Our work illustrates the feasibility of using estimates of urban charcoal consumption to establish a link between urban energy demands and forest degradation. This kind of approach has potential to reduce uncertainties in forest cover change and carbon emission assessments in sub-Saharan Africa.

No abstract

No abstract

Recent work has shown that ridge regression (RR) is Pareto to partial least squares (PLS) and principal component regression (PCR) when the variance indicator Euclidian norm of the regression coefficients, //p//, is plotted against the bias indicator root mean square error of calibration (RMSEC). Simplex optimization demonstrates that RR is Pareto for several other spectral data sets when //p// is used with RMSEC and the root mean square error of evaluation (RMSEE) as optimization criteria. From this investigation, it was observed that while RR is Pareto optimal, PLS and PCR harmonious models are near equivalent to harmonious RR models. Additionally, it was found that RR is Pareto robust, i.e., models formed at one temperature were then used to predict samples at another temperature. Wavelength selection is commonly performed to improve analysis results such that bias indicators RMSEC, RMSEE, root mean square error of validation, or root mean square error of cross-validation decrease using a subset of wavelengths. Just as critical to an analysis of selected wavelengths is an assessment of variance. Using wavelengths deemed optimal in a previous study, this paper reports on the variance/bias tradeoff. An approach that forms the Pareto model with a Pareto wavelength subset is suggested.