It has now been in operation for over a decade, helping to keep a vast area of forest intact by generating credits to be sold in an often volatile carbon credit market.
Deforestation avoidance projects, especially the Kariba REDD+ project, have received their fair share of media attention recently. I want to go on the record to tell the in-depth story of Kariba, looking at the history of this project, the lessons we have learned along the way, and be fully transparent about the good, the bad and the ugly of trying to build a commercial market for supporting climate action.
It's a day in early 2010, a time not long after the failure of the COP15 climate summit in Copenhagen and before the first vuvuzelas were blown for the World Cup in South Africa.
Carbon markets are young, with most players focused on the Clean Development Mechanism (CDM) established through the Kyoto Protocol, an international treaty on climate change signed in 1997. The Kyoto Protocol came into force in early 2005 around the same time the EU launched its Emissions Trading Scheme (EU ETS).
At this time, most climate action projects focused on building out renewable energy and addressing sources of methane to reduce greenhouse gases. Land-based projects, such as those involving agriculture and forestry, had been excluded from the EU ETS. As a result, almost no forestry projects are being developed. The challenge of reducing deforestation is largely excluded from Kyoto Protocol discussions.
I had been working for four years at South Pole, as one of five co-founders who shared a background in environmental sciences and a desire to build a business to help address the greatest challenge of the 21st century, climate change.
We received an email from Zimbabwe. Someone wants to work with us to develop a REDD+ project, a relatively new UN climate initiative that was designed to provide a mechanism to reduce emissions from deforestation and forest degradation in developing countries, incentivised by the sale of carbon credits.
Out of curiosity, we reply. Until then, we had mostly developed projects in places like Thailand and Kenya, China and Colombia, where two of the co-founders were living. REDD+ was new territory for us and the whole market.
We fly out to a Game conservation area in northeast Zimbabwe to learn more. We talk to people, to better understand the threats to forests in that park as well as in the eastern highlands as a result of agriculture policies. Over several years, civil unrest and the macroeconomic situation had forced workers to leave commercial farms in central Zimbabwe causing massive unemployment. Desperate people were returning to their communal lands, causing deforestation across rural Zimbabwe to spike. Someone in our team buys a 1 trillion Zimbabwe dollar bill as a souvenir. The IMF is looking to yet again bail out the country's failing economy.
We start reviewing the methodologies for REDD+ projects developed by Verra, the only carbon standard for deforestation avoidance projects available back then. We find a recently written methodology, VM9, which has been used by one other project in Asia. Using high level approximations with Landsat satellite imagery, we conclude that it does not make sense to develop a REDD+ project in eastern Zimbabwe - the area is too small and the opportunity cost of land use is too high.
We were disappointed but we still had the area known as Songo, near Lake Kariba, identified as a potential project. We repeat the carbon eligibility work and find that this area has potential.
Preparing the ground
A feasibility study typically involves several steps. First and foremost, we need to be confident that there is an environmental problem that can be solved through implementing a REDD+ project. This is what is known as a theory of change, where the underlying drivers and agents of deforestation are identified and a series of interventions are designed to reduce deforestation.
This will only work if the project is delivered by a capable partner, with support from engaged local communities. Once we are persuaded of this, we can review whether it qualifies as a REDD+ project based on detailed methodologies and explore technical feasibility.
It is also important that the land is eligible, and the potential volume of carbon credits generated (if successful) is sufficient to cover the costs of developing the project. We develop an initial feasibility study on Kariba for our client Carbon Green Investments (CGI), who starts discussions with stakeholders in surrounding communities - the Hurungwe, Binga, Nyaminyami and Mbire Rural District Councils. These communities are interested in a forest conservation project that would also address two other serious problems they face: food security and water scarcity. Few NGOs were active in their areas at that time, and those that were focused mainly on providing emergency food aid and basic health interventions, part of short-term programs supported by donations.
As a result of this study, we agree that the prerequisites to developing a REDD+ project are in place:
CGI comes to agreement with the Rural District Councils on how the net revenues from carbon credit sales will be split, after taking into account administration and sales costs (CGA graph below). It was decided that CGI would retain 30% of the revenue in return for setting up the project and providing the seed funding, with the remainder split between local structures. These included Carbon Green Africa - the team that delivered the deforestation avoidance and social development project activities on the ground - the four Rural District Councils (RDCs), community projects, a community 'rainy day' fund, and two concession leaseholders.
On July 1, 2011 the Kariba project officially starts its operations. It was first operated with the help of two Zimbabwean consulting companies, Black Crystal Consulting and Environment Africa. After a while, Carbon Green Africa (CGA) was founded to be the permanent project operator, taking over the day-to-day work of preventing forest fires, tackling poaching and illegal deforestation, and supporting social and economic development activities with local communities.
Meanwhile, South Pole and CGI agreed on a contract for consulting services, to develop the technical Project Design Document (PDD). In simple terms, a PDD is like the initial blueprint to describe the detailed plan for a project, and how it meets the requirements of a particular carbon project standard.
We started to prepare the project for certification, initially using a set of standards known as the Climate, Community and Biodiversity (CCB) Standards, which were developed by a group of NGOs including Conservation International, CARE, The Nature Conservancy, Rainforest Alliance, and the Wildlife Conservation Society.
In parallel and in addition to this, we also looked into developing the project in line with the Verified Carbon Standard (VCS) or Verra, as the organisation is called today. For our proposed project to be validated under Verra, more time was needed to determine the baseline analysis and set up monitoring protocols. Although both the CCB Standards and the VCS are today administered by Verra, back in 2011 these were two separate organisations, and had independent certification processes and templates.
We managed to get CCB Standards validation for the project done first. Under the CCB Standards, a project developer must show that there is a net positive climate impact across 15 criteria. This was done using traditional satellite imagery to establish a baseline deforestation rate (which did not have to be quantified, nor did a reference area need to be established). Using a simplified calculation aimed at showing the project's net climate positive impact, we estimated a net climate benefit of above 1 million tons of emission reductions per year. We were confident that the project would generate enough carbon credits for such a large area to be viable as a project.
What we didn't know at that time was how hard it was going to be to sell the credits!
We then proceeded with the VCS validation. The VCS PDD required more work. We used the first version of something known as Methodology 9 (VM9), which employed something called a Cumulative Deforestation Model (CDM). This model takes into account the fact that deforestation does not occur in a linear manner:
Predicting the future is challenging, so to develop the model for this S-shaped curve of deforestation under the Verra methodology, our team had to classify data from satellite imagery by visually confirming, on the ground, if a specific point in the image is showing a forest area, non-forest area, or cloud cover/cloud shade. This is something we call 'ground-truthing'.
Because we had already been through the process of developing a project baseline using satellite imagery with work done for the CCB Standards, we asked Verra for a methodology deviation. This can occur when the auditing body for the project checks and approves a conservative deviation from what the methodology prescribes. The auditor we worked with initially (Environmental Services Inc) approved the deviation1. However, when Verra reviewed that proposed deviation it was rejected, because a deviation is not allowed under Methodology 9 if it relates to the way the baseline is determined.
We were in a situation where the preparation of the REDD+ project had already incurred substantial operating costs for CGI. Because this was still the early days of the market, Verra told us that a different approach could potentially be implemented – provided that we propose the development of a new methodology. However, we knew from previous projects that developing a new or modified methodology would cost about US$200,000 and take up to two years, as it requires multiple expert reviews and public consultations.
South Pole and CGI agreed that it would be best to proceed using the existing VCS Methodology 9 as it was written, without deviation. This meant that we now had to take thousands of grid sample points from satellite imagery and compare them to on-the-ground observations, in order to put these into the Cumulative Deforestation Model. It would still take time and significant effort, but substantially less than two years.
We completed the modelling work and entered it in the final VCS PDD, which was then audited by an independent third party and approved by Verra. It resulted in a forecast of 6.6 million carbon credits per year.
However, as we will see later on in the story, such large volumes were neither generated nor issued: the final volume of issued carbon credits was approximately 3.6 million tons per year.
Kariba: boreholes for clean water, with solar-powered pumps
Meanwhile, the project implementation was going ahead at full speed. Conscious of the fact that the project had not yet sold any carbon credits, it prioritised the activities that they felt would make the largest impact on deforestation according to the theory of change (detailed in the PDD 'blueprint'). These were also those that were among the most sought-after activities by the local communities.
At this stage, the project focused mostly on three interventions:
“Pfumvudza farming" at Kariba
Apart from these activities, the project introduced early burning at the start of the dry season, to manage risks from annual wildfires (which emits carbon dioxide). Measures were also put in place to tackle poaching, as poachers would often start fires to scare animals out of the bush.
In particular, the support for conservation farming was a big success. The project initially started by training a small number of “champion farmers", who would then go on to train neighbours in their local wards. These practices are now called “pfumvudza farming" and have been so successful that they have spread across Zimbabwe, doubling yields and reducing the risk of slash and burn in many other areas outside the project.
Unfortunately, because there was still no money coming to the project from carbon credit sales, there were many other activities that had to be postponed across the four rural district council areas.
On October 15, 2012, the project was officially registered in the Markit environmental registry (today, the Verra registry) and everyone was elated and relieved.
Carbon finance uses a results-based model, so carbon credits are only issued to a project when reduced emissions can be evidenced – and audited. This means that a REDD+ project only generates carbon credits when there is a measurable impact – one that takes into account the predicted rates of deforestation, and shows what has happened in the project area as well as the comparative reference area. All this also needs to be successfully verified by a third party auditor.
This requires detailed project monitoring, which includes:
Monitoring project tree biomass and area deforestation at Kariba
Finally, just before Christmas in 2012, the project's first monitoring, reporting and verification (MRV) statement, audited by Environmental Services, Inc, was accepted by Verra. One year later, in December 2013, the project generated its very first carbon credits for sale.
Following three years of hard work, we were elated.
It is worth noting here that while the PDD predicted that the project might generate 6.6 million carbon credits a year on average, this did not happen. In that same vein, it is important to understand that climate action projects never claim credits based on initial evaluations (such as what was provided in the PDD) but only after the monitoring, third party audit/verification, and final validation of a standard has taken place.
In our initial work, we used default value assumptions for forest carbon and soil carbon that were higher than what we measured on the ground. The volume of credits issued were tied to these more accurate numbers.
As a result, about 3.6 million carbon credits were generated each year on average from December 2013 onwards, although the exact volume varies a little from year to year – this is due to changes in forest carbon stock in the permanent sampling plots, as well as changes in the soil carbon samples collected.
Training and awareness raising about the Kariba project
Setting up one of the world's first deforestation avoidance projects to generate carbon credits was a big challenge. But we didn't realise at the time that, for many years, the most difficult part would be selling them to the market.
At the beginning of 2013, and shortly after the dramatic end of the Kyoto Protocol, South Pole was asked by CGI to do the marketing and sales of the carbon credits for Kariba.
The terms of our agreements have varied over the years, adjusting to market conditions and the project's needs. The very first contract gave South Pole the right to buy 400,000 VERs per year from the project at US$3.50/ton. We would share any sale revenue above this price 50/50. It quickly became clear that the market price at that time was significantly lower and the carbon certificates were unsellable at that price.
Why is that?
As mentioned, just a few months earlier, the Kyoto Protocol's first commitment period had expired, with no agreement among governments for an extension. This meant that many CDM projects around the world all of a sudden found themselves without a buyer of their carbon credits, leading to an over-supply that – as we will see later – lasted until around 2019 and pushed down prices substantially.
Between 2013-2019, the transaction volumes of Kariba VERs remained low. The project requires at least US$60,000 per month just to keep minimal operations, staff and offices running (which is not much if you consider that the project is nearly the size of Puerto Rico…). Operating the project at full scale is much more expensive.
In this depressed market, CGI was not able to generate sufficient revenue to run basic activities, which it had been contracted to do in return for its 30% benefit share. It was proposed that South Pole - in contrast to a standard revenue-sharing model - would purchase significant volumes of carbon credits on stock.
For South Pole, a company with 150 employees at the time, and without a strong balance sheet, such a step was unprecedented and represented a massive risk. Nonetheless, we believed that keeping this project alive was essential. We decided to front-load the capital and purchase credits on stock without having a buyer - essentially providing a high-risk, asset-backed loan to the project that we never thought would get 'repaid'. These transactions occurred several times (in 2015, 2017, 2018, and - as part of an envisaged transaction to scale up the project that is still ongoing in 2021). We also did an additional early commitment to credit purchases in 2019 (see below).
To be clear, if South Pole had not purchased these carbon credits in 2015-2019, the project would not have survived. Let's not forget that these were years of terrible civil unrest, massive inflation and no access to hard currency in Zimbabwe. There was no way for the project to get finance any other way.
This was not a speculative investment but the only way to keep the project going. We could have applied such a strategy to the wider market (which at the time was over supplied and had very little demand) had we expected prices to rise, but we did not.
South Pole held on to these credits for up to six years as a non-productive asset. Several efforts to sell larger amounts of Kariba REDD+ VERs failed. For instance, South Pole negotiated with a major Italian company for nearly a year – to no avail. Also, a REDD+ project really does not have a guarantee or asset it can offer in return for a loan other than carbon credits, an asset that – as shown – was completely illiquid during those years.
Soon, another problem emerged for South Pole: the early carbon credits that we had purchased gradually lost value because – in the eyes of buyers – later “vintages" of credits are more attractive than older ones. Double whammy.
On site, frustration grew. Not much revenue was coming in, and subsequently, not much money was being distributed to the local communities. Considering the large project area, with 400,000 people, it was clear that some were feeling that the project was producing woefully few benefits.
It was only towards the end of 2019 that we started to see a welcome change in the voluntary carbon market. Carbon prices - at least for mid-sized transactions - slowly rose. Companies around the world were finally seeing the need for climate action and the “Greta Thunberg effect" was taking root.
Yet, Kariba was still in dire straits and South Pole again committed to an at-risk purchase of carbon credits at the end of 2019. This time, the purchase volume was paid over roughly two years. The price for the credits was calculated by multiplying the most likely sales price based on advanced discussions with some buyers, with 75% going to the project. This gave the project the certainty that for about two years, a secure revenue flow was guaranteed per month while South Pole took the sales risk – in other words, the possibility that some buyers would not follow through on their purchase intention. Unfortunately, this also happened.
In 2021, we started discussing an investment in new projects. At that point, the project asked us if we would agree to another carbon credit offtake transaction to support the project's financials, given the experience with years of low prices and high volatility. They felt that such a transaction was also needed to show communities that a substantial amount of funding had been secured for the project, no matter where carbon markets were going.
South Pole agreed to buy VERs from the project. This allowed the project to substantially increase its activities in 2022 and is hoping to also do so in 2023. The funds go to communities, the RDCs and the leaseholders, and increases the amount paid into the longevity fund, as detailed above.
In late 2021, driven by renewed private sector interest in climate action and demand for carbon credits, prices exploded. The price of REDD+ VERs, which in 2019 hovered around US$ 1 or less, jumped to US$ 8 or sometimes more. Secondary markets developed, such as the CBL market tracker. The Kariba carbon asset, some of which South Pole had held onto for a long time, was worth a lot more all of a sudden. It came as a surprise, but South Pole's firm trust in the value of protecting forests had finally paid off.
For all other carbon credits, South Pole and CGI had agreed on a contract where South Pole received a 25% commission for work that included the technical carbon work, all third-party verification and auditing fees, prepayment of the registry fees, and sales and marketing of the credits.
The result is that South Pole essentially has two types of Kariba carbon credit transactions:
It is important to note that South Pole now strictly uses type the commission model for all applications that have a direct interaction with a retail customer, such as websites that allow a checkout with the option to compensate for emissions by funding climate action.
In 2021, carbon markets looked more attractive than ever before, with both volumes and prices rising, and buyers interested in nature-based solutions. This is when CGI and South Pole started looking at other project areas in Zimbabwe to expand forest protection.
Kariba had taught us a lot, and we wanted to make sure that new projects had solid financing – and reliable revenue streams – from the get-go. We considered jointly investing in some new REDD+ projects as partners. We signed a Term Sheet for an equity transaction and started the more thorough investment analysis that comes with it. Of course, being equity partners requires a whole new level of due diligence.
For a variety of reasons,including the Zimbabwe government's foreign investment regulations, Verra's new consolidated REDD + methodology and our own diligence processes, we decided to decline the opportunity at the time. Subsequently in 2023, South Pole terminated all commercial agreements with CGI.
In light of market requirements and best practice, we agreed at the time to further upgrade the documenting and reporting processes of the current Kariba project.
“Deforestation is a wicked problem with multiple underlying drivers and no perfect solution" (Steve Zwick).
As Steve Zwick, long-time managing editor of Ecosystem Marketplace, hints at, there may always be a level of uncertainty in methodologies and solutions that are trying to do something that we all recognize to be inherently uncertain: predicting the trajectory of human behaviour into the future.
Along with growing interest in climate action in 2022 came a growing interest in (and scrutiny of) some of carbon methodologies – and some of the market leaders. This, of course, is welcome: the vast majority of the REDD+ community - South Pole included - have been continuously testing, challenging, and improving efforts to develop credible baselines as well as ways to monitor, report, and verify solutions for avoiding deforestation – methods that have been peer reviewed and that have received public scrutiny for over a decade.
We welcome and seek out constructive criticism wherever we can. This is why we are also continuously improving our own processes in line with how projects, standards, and climate science are evolving – as is the rest of the market. The fact that performance is being evaluated and lessons learned is not cause for alarm, but rather a sign of a healthy, “self-learning" system. This is how science works.
Recently, however, we have become quite concerned that comments made based on limited desk-research, and citing sources that have been proven to not accurately show deforestation rates in African Dry Forests, for example, can not only deter developers and other actors from taking risks to protect our ecosystems at scale, but also create such distrust of climate action that it may stop many of us from acting altogether.
Throughout its 10 years of existence, the Kariba project has undergone periodic assessments, and has been successfully verified five times. This biannual monitoring, reporting, verification (MRV) and final validation is a must in order to be able to issue carbon credits in the first place.
The Cumulative Deforestation Model is generated to predict a 'baseline deforestation rate' by taking deforestation data points from the project's reference area during the decade leading up to the start of the project. In other words, we observe a decade of deforestation in the reference area and use that to try to predict the next 10 years.
Contrary to the critique around the rigour of REDD+ methodologies and their ability to ensure real impact, Verra's methodology requires that – on top of periodic MRVs – the Cumulative Deforestation Model is also updated at the end of the project's first 10-year crediting period to compare the predicted deforestation rates with the observed deforestation rates. Revalidation is currently done every 10 years across the full 30 year duration of the Kariba project, but Verra will be shortening this to every 6 years.
We call this revalidation a 'true up' moment, a built-in safety mechanism that ensures that the project is on the right track. If a higher rate of deforestation is detected in the reference area than in the project area, then the Cumulative Deforestation model is corrected upwards. Conversely, if the deforestation rate has slowed down, the model has to be adjusted downwards.
So how does this technical work relate to carbon credits? In short, it balances out the issuances of carbon credits over the entire lifespan of the project – 30-years in the case of Kariba – ensuring that future credit issuances will be adjusted to reflect the actual pace of deforestation on the ground. If observed deforestation was higher, more credits will be issued in the next crediting period of the project, and vice versa.
Reassessing a project of Kariba's size takes time. This is why South Pole started preparing for the baseline revalidation in early 2022, using satellite imagery from late 2021. We received the first results of this preliminary (unaudited) review in mid and late 2022, see graph:
Our initial findings show that the cumulative deforestation model will have to be adjusted downwards, i.e. the deforestation rate in the reference area has slowed down. This is a wonderful, albeit rather surprising, development.
And this is when the true-up mechanism of Methodology 9 comes into play: In order for all of the project's emission reductions to date to match the observed deforestation on the ground, we do not expect many credits to be issued by the Kariba project for a number of years. It might be longer if deforestation rates continue dropping in the right direction – which in itself would be a great outcome for Zimbabwe.
At the end of that period (of limited credit issuances), the number of verified carbon credits issued at any time by the project, during its 30-year lifespan, will again be in line with the emission reductions modelled according to the updated Cumulative Deforestation Model.
As science, technologies and markets evolve, we evolve with them. We have most of the data that allows us now to formally start the baseline revalidation.
However, Verra has not been resting on its laurels either, and has already made significant advances in developing a new, consolidated REDD+ methodology incorporating scientific updates into how forest protection projects can be further improved. This methodology will likely be mandatory for all REDD+ projects, new and old, from 2025 onwards.
While this is welcome news, it also means that our team is faced with yet another dilemma: do we revalidate Kariba in 2023 for the period 2021-2025, just to move the project to the new consolidated methodology in 2025? Or do we wait, and develop the project with the consolidated methodology in 2025?
At the same time, the Zimbabwean government is taking an interest in protecting the country's forests, so we may also consider moving to a so-called jurisdictional REDD+ scheme. This requires a significant amount of work, however - in other countries it has taken several years and is still not fully concluded anywhere, so we don't have a good model to follow. This is truly uncharted territory, something we are familiar with.
Global warming, unfortunately, does not follow project development timelines or government agendas, which is why we want to move as fast as possible, to make every year count. We are consulting with Verra on a recommended approach.
Meanwhile, for the communities on the ground, this poses a severe problem: a number of activities that keep deforestation rates down may not get funded. These include moringa (conservation) cultivation, nutrition gardens, borehole maintenance, early burning and anti-poaching. Other activities need regular and predictable revenue streams from carbon credit sales to deliver social benefits, such as the expansion and maintenance of schools and health clinics.
As we look ahead, let's not miss the forest for the trees: we must halt and reverse deforestation within this decade. The REDD+ mechanism is currently the only forest conservation finance instrument of any significant scale. And – as our team has experienced first-hand during our long-standing commitment to Kariba – it is proving to be effective in protecting our ecosystems and using revenues to improve the lives and livelihoods of people who depend on them, so much so that the project should be able to sustain itself without carbon finance after its full 30 year lifespan.
During its first 10 years, the Kariba Project was financed via the carbon market, loans by committed Zimbabweans (CGI), and through South Pole's commitments. On top of making a genuine difference in forest conservation, the project has managed to create additional revenue streams for the communities via activities that got off the ground thanks to carbon credits: these include the sales of Moringa products, and vegetable and honey production (each generating about US$60,000 per year in revenue). Tourism in the area is also doing better than before the project, largely thanks to Kariba attracting more wildlife.
Many other activities financed by carbon credits have also taken off: The conservation farming approach has been a clear success, with farmers trained by the project heading off to spread the successful practices. In fact, the training financed by Kariba proved so effective that it has now spread across the entire country - including by trained Agritex officers.
A successful REDD+ project is set up in a way that - in the long run - it should be able to sustain itself without carbon finance.
When it comes to Kariba REDD+, we are not quite there yet: high population density and growth, coupled with low levels of economic development mean that there is still an increased risk of forest being cut down if carbon finance stops.
The communities will clearly be concerned that this source of revenue might be paused. They wonder how their project will survive and we have to discuss the various options with them.
The project has already shown an impressive impact on biodiversity and tourism – as successfully certified and verified multiple times by Verra, both under the Verified Carbon Standard and the Climate, Community & Biodiversity Standard – but a lot more can still be done.
We hope that this long tale of the Kariba REDD+ project will inspire others to support forest protection and to learn from our experience – bumps in the road and all.
1 The related documentation can be found on the Verra webpage, showing the estimated total emission reductions from avoided deforestation at 1.7 million tons of COe per year.
