Examination of Uganda’s above ground biomass distribution, affected by fires
Dissertation 2016 - Maria Foka.pdf (6.387Mb)
MetadataShow full item record
Aboveground biomass (AGB) observation and estimation is important for the assessment of ecosystems’ carbon stocks. Uganda’s tropical savannas and woodlands are highly affected by fires. This research attempted to examine the two way relation of fires and vegetation in Uganda. The developed approach, was based on the combination of AGB reference datasets, radar L-band images (ALOS PALSAR) and active fire data (FIRMS). The regression analysis showed that the there is a strong correlation between the AGB and the HV backscatter signal, with R2=0.55 and P>0.000, within a 95% confidence interval. The estimation of the AGB, which was conducted by applying a published equation and using the reference datasets and the HV backscatter signal image, resulted an RMSE =27.61 mg /ha. The HV backscatter signal starts saturating for AGB values >100mg/ha. According to the fire frequency histogram the number of fires is gradually decreased as the AGB value is increased. A small peak of the fire frequency in high values of AGB indicates the occurrence of mixed pixels. The final products was an AGB map of Uganda for 2007 and a fire frequency map for the time period 2002-2007. Further and more accurate investigation regarding the variables and patterns, which affect the AGB could result to the improvement of the management policies in developing countries.