Land use change refers to the modification of the Earth's surface by humans. Land use/land cover change (in short, land change), especially the clearing of tree cover, is a major source of increased carbon dioxide (CO2) emissions contributing to anthropogenic climate change. In this study, carbon densities and changes in aboveground tree carbon (agc) across different land cover types were mapped in the Taita Hills, Kenya, using field measurements, airborne laser scanning (ALS) data and classified satellite imagery. An existing biomass map was used for retrieving carbon densities for a part of the study area. For the lowland area, another biomass map was created with a regression model based on field measurements of 415 trees on 61 plots and metrics calculated from discrete return ALS data. Explanatory variables in the linear regression model were the standard deviation and 60 % height percentiles of return elevations. Carbon fraction was calculated as 47 % of aboveground biomass. 11 land cover classes were classified from a satellite image with an object-based approach. Overall classification accuracy was 71.1 % with most confusion between the cropland and shrubland classes and shrubland and thicket. Based on the biomass maps, carbon densities were calculated for different land cover classes. Mean carbon densities were 89.0 Mg C ha-1 for indigenous broadleaved forests, 29.0 Mg C ha-1 for plantation forests, 15.6 Mg C ha-1 for woodland, 5.5 Mg C ha-1 for thicket, 3.2 Mg C ha-1 for shrubland, 8.1 Mg C ha-1 for croplands above 1220 meters above sea level (m a.s.l.) and 2.3 Mg C ha-1 for croplands below 1220 m a.s.l.. Land cover maps from 1987, 1992, 2003 and 2011 were used for studying the impact of land change on aboveground carbon stocks. A reduction in carbon storage was observed between years 1987, 1992 and 2003. An increase in total carbon stocks from 2003 to 2011 was observed as a results of increased proportion of woodland, plantation forest and broadleaved forest. These changes should be further verified in a spatially explicit way. More detailed data should be used in order to understand the full complexity of the dynamics between land change and carbon stocks in the heterogeneous landscape of the Taita Hills.
