By Marc Baker, and Jo Anderson, Founders, Carbon Tanzania
We are sitting on the Chingole Ridge in southern Tanzania, discussing recent articles on carbon credits. About 100 metres away – and 80 metres lower – we can see a dense, green forest. Yet here it is gravely; a stunted, Miombo woodland. If you were to look from space these forests would, depending on the time of year, present very different images. From such satellite pictures, you could even conclude that we are not sitting in a forest at all. But working on the ground here in Tanzania, we know that different types of forest exist in close proximity and form what is known as a ‘mosaic’ landscape. Established forest science knows this too and it is these context-specific datasets which underpin our projects’ operational methodologies.
Established forest science knows this too and it is these context-specific datasets which underpin our projects’ operational methodologies.
On 18th January the Guardian published articles based on an investigation conducted with Die Zeit and SourceMaterial critiquing the validity of carbon credits generated from forest conservation projects verified by the regulatory body, Verra. The investigation is based on the findings of published scientific papers by West et al. 2020, Guizar-Coutiño et al. 2022 and an unpublished paper by West et al. 2023. While there are many other climate mitigation activities from which carbon credits can be produced, forest conservation using REDD+ approaches, has been implemented on the largest scale.
The Guardian provides some of the world’s best and most comprehensive climate reporting; but they missed the mark this time. As founders of Carbon Tanzania, we have seen first-hand the power of the voluntary carbon market to deliver finance and combat deforestation which is often in direct conflict with the publication’s reporting. We can add much nuance to this discussion as two of our projects, Makame Savannah and Ntakata Mountains, are cited by the unpublished West et al. 2023 study and so are indirectly accused by the Guardian of being “worthless”.
A methodology is only as good as the data
We recognise the ‘causal inference’ methodology used by West et al. 2023 to be sound and a potentially useful approach, especially once it has been through a proper peer-review process. However, until such a review has been conducted, the way West et al. applied the methodology is still experimental and will benefit from, and likely be adjusted through, peer-review scrutiny.
Moreover, a computer-model is only as good as the data it is given and in this case, there are some glaring errors associated with the datasets chosen for the analyses used by West et al.
It is no revelation that different input data yields different output results. Yet it is important to raise awareness of the shortcomings of the specific input data used in these studies, especially as, in this case, it has formed the foundation of a scientific paper and subsequently news articles with somewhat sensational claims.
An example source for the data used by West et al. is Protected Planet – a public source for spatial data on the world’s protected areas. However, Protected Planet data is not an accurate depiction of Tanzania’s protected areas. For instance, it classifies the Maasai Steppe – where our Makame project is located – as protected. This misclassification can lead to a false assumption that the Maasai Steppe enjoys the same level of protection as a National Park, and that zero deforestation would have occurred in Makame in the absence of the project. This is incorrect. In reality, the entire area is village land and much of the landscape is already fragmented by shifting agriculture. Today, the area remains extremely vulnerable to further deforestation and degradation.
Similarly, the study also used the Global Forest Watch dataset which is known to perform poorly in many contexts, particularly in Tanzania. Indeed, section 4.2.4 of Verra’s technical review of the two West et al. papers includes three images of open forest in Tanzania from 2013, 2018 and 2019. These images very clearly reveal the area of forest lost to be significantly greater than the area detected in the Global Forest Watch dataset. In fact, Global Forest Watch itself, along with many related scientific papers, makes clear that it should never be relied upon to offer precise information for audit purposes, although it does provide a good first approximation of global and national trends.
Our current, conservative methodology
Both the Makame and Ntakata projects used baselines in line with Verra’s VM 0007 methodology for avoiding unplanned deforestation.
These baselines were painstakingly created over many months of analysis by some of the world’s leading forest carbon scientists (Rebecca Dickson and David Shoch of Terra Carbon). As such, the process to set this baseline involves two steps:
Establishing the likely deforestation rate through analysis of real, observed historical deforestation. This involves a rigorous classification of land in an area deemed similar to the project area which is certainly not an easy, or quick, task in dryland, mosaic forests. This area is known as the ‘reference region’. For this, we used the IDRISI TERRSET software (Eastman 2011) to classify a time series of Landsat imagery (satellite image program NASA/USGS) and then measured past deforestation rates. The resulting classifications meet strict accuracy requirements and are a far superior, more appropriate dataset for dryland ecosystems in this area than Global Forest Watch and Protected Planet. Covering millions of hectares, our reference regions for both the Makame and Ntakata projects are significantly larger than is technically required. However, we wanted to ensure the baselines could be as close as possible to the ‘jurisdictional’ areas on which future baselines will be calculated (see step 2 below).
Establishing the likely deforestation location(s) inside the project area. This involved producing a detailed risk map to predict the likelihood of future deforestation in specific geographical areas of our project sites, as well as the leakage belts (an area equal to and adjacent to the project where deforestation might be shifted) in the absence of the project. For this work, we used the peer-reviewed Land Change Modeler (LCM) within TERRSET to create a detailed deforestation risk map which we overlaid onto the project areas.
In addition, our detailed, geospatial work was ‘ground truthed’ on site visits by the geospatial team (and project auditors). As further insurance, deductions (between 20-30% of total credit issuances depending on the year) were applied to account for leakage and risk.
All of the assumptions and methodologies are described in detail in the publically available Makame and Ntakata Project Design Documents. We stand behind the resulting baselines for each project that this two-step process produced, until proven otherwise. Given the issues and discrepancies of the datasets highlighted previously, the scientific studies cited by the Guardian do not yet currently offer such proof.
We are committed to integrity and continual improvement
At Carbon Tanzania, we are guided by an unwavering commitment to both integrity and improvement.
First and foremost, our baselines have been designed and developed by the leading forestry scientists in the world and in compliance with the requirements of the standards and methodologies set forth by Verra.
The volume of credits we issue is always based upon conservative calculations as demonstrated above. This is a foundational principle of our work at Carbon Tanzania and of the approaches of carbon accounting designed by Terra Carbon.
As mentioned previously, we do not issue credits in correspondence to the full volume of emissions avoided. In numbers, this means that all our projects have a risk buffer, 20% for Makame and 13% for Ntakata, to account for variation in outcome, should it occur.
We do not rest on our laurels. All of our projects are verified annually and only issue ex-post (after the fact) verified emission reductions (carbon credits).
Forward thinking, our methodologies are already designed to fit into the new jurisdictional baselining methodologies being finalised by Verra. We are working closely with the government of Tanzania, Space intelligence and other landscape partners on developing a national deforestation baseline and associated risk map. This work is almost complete. We will adopt this risk map for our projects and ‘nest’ our existing projects as soon as practically possible under the Verra certification (currently anticipated to be towards the end of 2024).
Why not wait for improvement before buying credits?
Verra is committed to continually improving its methodologies and we support them in this. At present, it is working on a new methodology to set baseline deforestation rates at a jurisdictional level rather than at project level, as it is at the moment. This is going to be called the Jurisdictional and Nested REDD+ (JNR) Framework.
This does not mean that all previous credits are invalid or worthless. JNR simply means the baseline average will now be spread over a wider region, either at a national or sub-national level. The previous annual emissions reductions (carbon credits) are still in place, what JNR will allow for is greater scale and consistent carbon accounting frameworks across the landscape.
In line with our commitment to continual improvement, we support this move to jurisdictional deforestation baselines and indeed we have been preparing for it for years. However we continue to back existing credits for they are based on the best science and validated baseline models available now. The newer approaches will simply ensure greater efficiency of data collection and analysis.
We continue to back the voluntary carbon market
Our critiques of datasets are applicable to anyone looking to assess project baselines on a wider scale. The data used must be specific, taking into account governance structures, soil types, drivers of deforestation, political boundaries and complex ecological systems. Without a consideration of the on-the-ground environment, it is not possible to draw accurate conclusions.
Carbon markets are a tool to channel finance to where it is most needed. Without corporate demand for credits and our profit sharing model, many people we work with in Tanzania would not have had access to the same opportunities.
Faraja Oswald Alberto is a Finance Officer for the Ntakata Mountains project in western Tanzania. She has seen first hand how carbon finance has changed the area as she works to develop short and long term financial plans with her local community. Speaking from Ntakata, Faraja says.
“Before the start of the Ntakata Mountains forest protection project, there was an invasion and massive clearing of forest areas. Our lands were badly damaged. After that, the community decided to make a plan for the best use of land and implemented a forest carbon project. Gradually, the environment began to improve as the community received carbon finance to support sustainable projects and forest conservation.”
Indeed, Inaccurate critiques can create real consequences for those on the ground because without carbon finance, many local projects – from conservation to healthcare – would not be viable. Faraja continues:
“Over 25,000 people within the eight villages of the project areas benefit from developments such as [health] clinics, schools and health insurance. Also the presence of modern classrooms and food for students in schools vastly improves education in the community.
Village Game Scouts are now fully employed by their respective villages to protect the forests and are paid a monthly salary from the carbon credit revenue. Groups of entrepreneurs benefit from small loans made possible by carbon finance from Cocoba (Community Conservation Banks) to run their various wealth-producing activities. This is improving the local, community economy.”
Supuk Olekao is Manager of the Makame WMA, representing five Maasai villages, Irkiushoibor, Makame, Katikati, Ndedo and Ngabolo, and their communities.
“We set up a community conservation area, or Wildlife Management Area (WMA) in 2009 to stop the invasion of our land and its deforestation by nearby subsistence agriculturalists. However, the WMA management authority was unable to put it into practice because of lack of finance. We had the organisation, the ideas and the people, but not the resources to really protect our land and our forests.
“We partnered with Carbon Tanzania to set up a forest carbon project in 2016. The financial revenues from the carbon credits we now earn from protecting our forests in the way we always have, mean that we now have the resources to ensure our land is not invaded and our forests stay standing. Importantly, the carbon finance also enables health and education for our communities, and we can protect our livelihoods and our culture as Maasai.”
As we enter a new chapter for valuing nature, the coming years will see a radical change in how our global economic systems account for, and include, nature. As a mechanism to drive this shift, we continue to back the voluntary carbon markets in our collective transition to a nature-based economy.
Privacy & Cookies Policy
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.