REALU vs. REDD+: Carbon and biodiversity in the Afromontane landscapes of SW Ethiopia

•The scope of REDD+ in SW Ethiopia should be expanded to all land use practices.•80% of the landscape’s carbon stock in SW Ethiopia is stored in non-forest land uses.•Carbon stocks and tree species richness were positively correlated.•Tree communities and Hill’s indices proved useful biodiversity co-benefit indicators.•Tree community information improved stratification, reducing plot numbers by 23%.

A major constraint to current REDD+ initiatives is the arbitrariness of the operational definition of ‘forest’. The UN-REDD+ framework ignores the high potential of other (tree containing) land uses for carbon sequestration and the generation of co-benefits, in contrast to a whole landscape approach known as ‘Reducing Emissions from All Land Uses’ (REALU). To indicate this potential, carbon stocks and biodiversity were assessed at the landscape scale around Jimma, SW Ethiopia. Notwithstanding the indigenous forests’ high carbon stocks per hectare (336.96 ± 120.6 Mg C ha−1), 80% of the total carbon stock of the 106 ha assessed landscape window was found to be stored in non-forest land uses. Especially local agroforestry systems, such as managed semi-forest coffee production fragments (179.92 ± 38.5 Mg C ha−1), homegardens (106.26 ± 16.8 Mg C ha−1) and pole wood plantations (112.36 ± 23.9 Mg ha−1) store large amounts of carbon. Together with local maize fields (69.45 ± 12.3 Mg C ha−1), teff fields (69.51 ± 6.5 Mg C ha−1) and grazing lands (76.90 ± 10.3 Mg C ha−1) these non-forest land uses play an important role in a prospective ‘high-carbon-stock rural-development’ under a REALU framework. A REDD+ approach, applying a stringent forest definition that disqualifies for agriculture and a 30% canopy cover threshold, would only account for 24 Gg C in the 106 ha landscape window. Applying the canopy cover threshold alone, would triple the amount to 72 Gg C, while a REALU project would account for the whole 116 Gg landscape carbon. The biodiversity assessment, based on tree community composition and Hill’s diversity indices, indicated important biodiversity co-benefits under a REALU approach. Biodiversity indices such as the Hill’s N1 diversity of abundant tree species showed a strong positive correlation with local carbon stocks (R2 = 0.56). Since the cover of indigenous forest is reduced to only 7% of the landscape window’s area, most of the remaining trees and biodiversity can be found within the surrounding landscape mosaic. Semi-forest coffee production fragments for example still harbour almost half of all encountered native tree species. An interesting result of the simultaneous biodiversity assessment was that a regrouping of plots based on tree biodiversity provides a more precise and cost efficient way to assess carbon stocks, reducing the number of plots needed with 23%, to achieve the same precision level.