Assessing Forest Carbon Stocks


A critical early step in our REDD project was to assess the current forest carbon stocks where we work in Kilwa District. This assessment underpins all of our REDD activities and determines the baseline against which progress can be measured.

We did this in 2010-11 through a survey of 25 'super-plots' that each cover 9 hectares and contain various sub-plots in a cluster arrangement. Using large-scale plots was important because the vegetation in Miombo Woodland varies from relatively open areas to patches of denser woodland (i.e. it is a very heterogeneous environment). This means that surveying smaller areas may result in too many plots virtually empty of trees and/or data that strongly departs from the statistical Normal Distribution. This plot design also allowed us to focus our efforts on surveying the biggest trees (which account for the most carbon). We counted and measured trees with DBH > 30cm in the 9 hectare large plot; smaller trees were only counted in the smaller sub-plots.

From this work we estimate that Kilwa District has approximately 30 million tonnes of carbon in aboveground stem biomass (see report). The sample plots we used have been permanently marked to support long term monitoring.

Carbon Loss from Fire

We estimate that upwards of 60% of the landscape in our REDD project area burns every year, which is mostly due to fires used to clear forest area for farmland burning out of control during the mid-to-late dry season. These wildfires are the most significant driver of deforestation in our REDD project area. A model developed by our partner, the University of Edinburgh (UoE), estimates between 0.5 and 1 tonne of carbon per hectare of dry forest can be lost from fire every year.

MCDI are therefore focusing our efforts to improve carbon stocks by working with communities to effectively manage forest fires. Depending on revenue earned and MCDI’s ability to expand our REDD work in south-eastern Tanzania, total emissions reductions could amount to anywhere between 520,000 – 1,850,000 tCO2e over a ten year period.

Monitoring Carbon Stocks

We need a powerful and robust methodology for detecting annual carbon savings from effective community based fire management. We expect improved fire management will lead to a slow habitat shift towards thicker forest, with some wooded savannah transforming into woodland, and some woodland becoming forest. The biomass increases from these transitions will be what the project sells in terms of carbon offsets. Specifically, we look at the changes in carbon stocks from avoided large tree mortality and from the regeneration of seedlings and saplings.

We decided to use the ‘GapFire’ model, which was developed by UoE to predict the response of forests to different fire regimes, to measure these changes over time. The model has already been used to measure the impacts of fire in Mozambique and Zimbabwe, and we worked with UoE to adapt the model for our project. This adapted GapFire model allows us to predict future carbon stocks (and thus carbon savings) in local miombo woodlands under improved community fire management. However, the actual values need to be verified using other monitoring methods, such as surveys of large permanent sample plots, monitoring individual trees (especially large trees) and using remote sensing technology.