Disney’s commitment to Mickey Mouse REDD: Conservation International’s trick baseline for the Alto Mayo project in Peru

Tweet about this on Twitter16Flattr the authorShare on Reddit0Share on Google+3Share on Facebook10Share on LinkedIn0Share on StumbleUpon50Digg thisEmail this to someone

The Alto Mayo Protected Forest in the Peruvian Amazon covers about 182,000 hectares. Although it became a protected area in 1987, it remained under serious threat. Today it is the site of a REDD project run by Conservation International with funding from Walt Disney.

In March 2013, Walt Disney bought 437,000 verified carbon units from the project at a total cost of US$3.5 million (US$8 per credit). Here’s a project set up by a major US NGO, with backing from a major US corporation, that has already sold hundreds of thousands of REDD carbon credits. The project is verified and validated under the Verified Carbon Standard (VCS).

The company that carried out the verification and validation process, Scientific Certification Systems, wrote in its validation report that,

The Project correctly applies the selected methodology element and is in conformance with all applicable requirements of the Verified Carbon Standard (VCS). The Project is designed to lead to reductions of GHG emissions that are real, measurable and give long-term benefits to the mitigation of climate change.

From 2009 to 2012, the project generated almost three million tons of emissions reductions. In order to generate all these REDD carbon credits, Conservation International has to be able to argue that in the absence of the REDD project, more deforestation would have taken place. Part of the process involves establishing the baseline – or what would have happened in the absence of the project. As REDD-Monitor has pointed out in the past, this is by definition tricky, because it’s measuring something that didn’t actually take place.

Of course, project developers would like to be able to argue that without their project deforestation would be rapid so that they can argue that deforestation is much slower with the project and they can generate more carbon credits (what would have happened minus what actually happened). That’s why VCS has developed a series of extremely complicated methodologies to determine baselines in different situations.

In the case of the Alto Mayo project, the methodology was VM0015, “Methodology for Avoided Unplanned Deforestation” (at only 208 pages long it makes for fascinating reading). The process of establishing the baseline for the Alto Mayo is explained in a Project Description (dated 15 June 2012, and available on the VCS website).

The process of establishing the baseline is described on page 106 of the Project Description:

Land-cover change data for the reference region were mapped by Conservation International, via analysis of Landsat-5 and Landsat-7 images, for the reference period of 1996 to 2006.

From an analysis of these maps, Conservation International calculated the area of forest and the area of forest lost for the period 1996 to 2006:

2013-04-26-151039_817x288_scrot

Between 1996 and 2001, the rate of forest loss was 0.12% per year. Between 2001 and 2006, the rate increased to 0.36% per year.

But if the rate of deforestation were to remain so low without a REDD project, Conservation International would be struggling to generate many carbon credits. Conservation International therefore introduced a new methodology in order to increase the baseline. On page 122 of the Project Description, Conservation International explain that they will use a “Cumulative Deforestation Model” as described in the VCS Methodology for Avoided Mosaic Deforestation of Tropical Forests (VM0009). Using this methodology, Conservation International managed to dramatically increase the baseline deforestation rate.

The rate of deforestation, illustrated by the green line has now miraculously increased to 1% per year – three times the rate the Conservation International found from the satellite data between 2001 and 2006.

Conservation International produced a series of maps of baseline deforestation using this new model. Here’s what Conservation International want us to believe would happen without the 20-year REDD project, which started in 2008:

2013-04-26-153640_883x688_scrot

The red area is forest loss. It looks horrific. That’s the purpose of these maps. But compare the map above of “what would have happened in 2008 in the absence of the project” with the actual deforestation between 2001 and 2006. Deforestation is also coloured red:

2013-04-26-154128_730x744_scrot

The deforestation between 2001 to 2006 is difficult to see. How is the huge area of deforestation coloured bright red on the 2008 map supposed to have happened in the two years between 2006 and 2008? And why did SCS verify and validate what seems to be a trick baseline? Or perhaps, since Walt Disney is buying the credits, we should call it a Mickey Mouse baseline.


PHOTO Credit: Conservation International.

Tweet about this on Twitter16Flattr the authorShare on Reddit0Share on Google+3Share on Facebook10Share on LinkedIn0Share on StumbleUpon50Digg thisEmail this to someone
Bookmark the permalink.

17 Comments

  1. Excellent posting Chris. We see the inner workings of a typical REDD fraud – and surely no-one should be surprised that CI is the main perpetrator or that SCS is prepared to verify this garbage – in the same way that they have been certifying non-sustainable logging operations for the Forest Stewardship Council for the last 20 years.

  2. Thank you for this message.
    It seems that what we have here is a good illustration of what this VCS VM0009 methodology is made for: to create hot air.
    The Mai Ndombe project that has been validated in DRC is even a worse example. Based on a reference region located in Bas-Congo (about 700km from the project area and close to Matadi international harbour)and on what happened there between 1987 and 2007, they use this cumulative model to project about 15 times more emissions than what has been observed on the project area between 1990 and 2000.
    Who says that VCS was credible?
    This project has been validated by DNV with the same type of argument: “It is DNV’s opinion that the project activity “The Mai Ndombe REDD+ Poject” as described in the VCS PD, meets all relevant VCSA requirements for the VCS project and correctly applies the VCS/CDM methodology for Avoided Mosaic Deforestation of Tropical Forests (VM0009)”, version 2.0.”
    And if the problem was coming from the methodology itself?
    The funniest part is that DRC government is now working with them to extend the project to all Bandundu province in DRC. REDD+ the low hanging fruit…

  3. Britaldo Soares Filho

    VCS REDD baseline, no matter linear trend or cumulative, is totally delusional. How can someone predict the future? Deforestation models were not meant for that. See my presentation at http://www.csr.ufmg.br/dinamica/redd/REDD_CIFOR.pdf

  4. Although nobody can tell how the future will be, I would argue that are some ways to predict what they would likely be, given some constrains and uncertainties (of course). REDD baseline are supposed to take into account causes and underlying causes of deforestation. If such attributes are somewhat similar to the ones we observed current observed and we observed in the past, it is likely that the same patterns could be reproduced observed in the future. I do realize that this issue is context dependent and for some areas a deforestation projection based on retrospective trends would make no sense, but for others it could. Unfortunately, it is not hard to find inexperienced carbon project developers or even poorly trained carbon auditors that ignore these issues, what ultimately would result in low-technical quality REDD projects.
    Under Soares Filho presentation, the deforestation projections from Juma project do look like terrible, but it is important to realize that that was the third REDD project validated under CCBS in the world, and if you compare it to the first one from Aceh, Indonesia (where they used a simple average of historical deforestation), it was a great improvement, at some level. Anyways, Juma project don’t have a VCS certification and they will never will with that model, because the methodology now cannot be considered appropriate enough (it is not VCS approved and likely results in a lot of “hot air”…) VCS REDD methodologies require baseline to be validated against past deforestation observations (mostly using Figure of Merits), what is not perfect, but it is something that should not be ignored… The Surui deforestation model is one of the best that I could find, it is based on the economic interaction among the community associated with land-uses and it is more likely to represent future trends (although because it is good the its reading is not that easy for readers without some fundamental modeling and statistical background).
    Anyways, baselines need to be revised every 10 years. That is a way to correct future changes in deforestation trends. I guess the bottom line is that it is hard to predict land-use changes, but some techniques and knowledge can help to us to develop more likely scenarios (and that is why it is called prediction for that matter). If we don’t consider these predictions, how could we select and reward conservation projects that are implemented in “high deforestation risk” areas and their performance?

  5. A well-known company (with no mandatory emissions cap) donating funds to voluntarily reduce deforestation by working with poor farmers with a respected NGO following international best-practice standards…and all audited by a third party? sounds horrible!

  6. @me – Could you please re-read the post? I think you’ve misunderstood it. If Disney wants to hand over its cash to Conservation International so that they can conserve forests, then that may well be a good thing. But the point is that Conservation International is selling carbon credits to Disney that are based on a trick baseline. Is that really “following international best-practice standards”? The wonderful third party audit completely failed to notice the trick baseline. And if the auditors did notice, they decided it didn’t matter. Doesn’t this worry you?

  7. no, sorry. I wish there were more projects helping protect the forests.

  8. @me (#7) – Are you not concerned about climate change? If so, then wouldn’t it be better to protect the forests without generating carbon credits because that would mean a genuine reduction in emissions (rather than trading off a reduction in one place for continued emissions somewhere else)?

    If the credits are based on a Mickey Mouse baseline as in the case of Alto Mayo, that just makes matters worse, because it would actually lead to an increase in emissions.

  9. @me – “Respected NGO”? Which one is that, then?

    Chris – can we have a posting explaining the role of the REDD certifiers, and how this all works? I’ve heard a lot about some of these companies, such as SCS, certifying things that shouldn’t get certified (well, they get paid for it, so of course they are going to), but how does this work with REDD projects?

    Setting aside the important question about whether the VCS methodology is any good in the first place, who is checking the certifiers to ensure that they are not cutting corners or just selling their ‘certificates’ for profit? Who would someone complain to if, for example, they believed that SCS had been grossly incompetent in not spotting the obvious errors with the baselines and projections for this Disney project? Is there a regulator to whom they must account, and can they be struck off the register if they misbehave?

  10. @treefellas Complaints would have to be made to the accreditation board which accredited the company (ie ANSI, CDM EB). The accreditor could then look into the accusations and potentially suspend the verifier.

    @chrislang Voluntary carbon credits don’t pose a risk for ‘hot air’, so the concern is much lessened in this ‘mickey mouse’ scenario. It would be great if REDD (or whatever other means of forest conservation you can come up with) could be fully funded by ODA, but its pretty clear that there is a funding gap that is only going to be filled if the private sector becomes involved. Pragmatically, I don’t think you are going to see companies committing to ambitious climate legislation (or politicians pushing ambitious climate change legislation) without at least some form of off-set involved where companies both have to reduce their own emissions but can also off-set a certain percentage.

    In the absence of anything in the present, I don’t see how a voluntary initiative like this hurts, or how you could argue there are no actual reductions since Disney is not using this for any mandatory emission reductions.

  11. @al perkins (#10) – You agree that this looks like a Mickey Mouse baseline, but then argue that it doesn’t matter much because the carbon credits are voluntary. I thought part of the point of the voluntary market was to pave the way for a compliance market. Currently there is no REDD compliance market, but the problem of guesstimating baselines will remain.

    If Disney had handed over the money anyway – regardless of the carbon credits – wouldn’t that have been even better? The project would certainly have been cheaper without all the modelling and baseline calculations.

  12. Britaldo Soares Filho

    Simulation models are no crystal ball. They are planning tools, useful to perform ex ante evaluation of potential outcomes from a variety of scenarios. The use of forward-looking baselines, being under VCS methodology or not, to fix credits for REDD is an irresponsible temerity because, in this case, every cent counts and thus there will be always a tendency to inflate the projection, which, anyhow, is just a speculation about the future.

    Maybe I am the culprit to have developed the methods for simulating deforestation that are being applied, not need to mention in a wrong way, to REDD projects, so it is my role to deconstruct this “fantasyland”. Is not there where Mickey lives?

    The text states that the project followed precisely the complex methodology of VCS. But, what about if its fundamentals are unsound? As a result, all the methodology becomes flawed. By the way, people who have never studied deforestation modeling, nor published a scientific paper on it, designed most of this methodology, and now have become the REDD senior experts.

    Saving protected areas is very noble intent, indeed, but the means do not justify the end. Those REDD projects, together with their unwise methods, are undermining the credibility of REDD and tarnishing the reputation of institutions with a long history in conservation. Only the fool or the money blind does not see or does not want to see this mess.

  13. @al perkins

    Thanks for your reply. Presumably the CDM could not ‘suspend’ a certifier from carrying out these voluntary carbon market certifications, only from carrying out CDM assessments? And does the ANSI specifically have sets of standards for REDD projects that accredited organisations have to comply with? Does ANSI have any expertise in REDD or carbon credit certification issues?

    Any credible certification system should have an oversight body that ensures proper procedures, that there is compliance with the standards (in this case, the VCS standards) and mechanisms for redress and accountability – but it seems to me that in the case of certification of voluntary REDD projects, these are completely absent.

  14. I do understand Cris and Soares Filho’s points and I partially agree with you. I also don’t belive that VCS methodologies are sound enough (assuming that baseline deforestation prejections can ever be called sound), and I do share my concerns about non-experts validating such methods… But maybe we don’t need a crystal ball to implement such payment for environmental services, we just need estimations that somewhat make sense, based on similar causes and underlying causes of deforestation, and if they don’t, then VCS auditors should be trained well enough to indentify them and require corrections.
    Anyway, this was a good discussion, thank you for sharing your point of view.

  15. Chris, This is a fantastic post. One point worth noting is that the Alto Mayo has a sister project in the Congo that, I believe, did not pass the VSC verification process last summer. Also, part of the Disney investment went to fund that project as well, making the cost per credit less than $8.
    Thanks again for such an insightful post.

  16. There’s an interesting recent article on Ecosystem Marketplace about Alto Mayo:

    Disney Helps Dreams Come
    True In Peru’s Alto Mayo Forest
    .

    The article features lots of discussion about Disney’s funding and how Conservation International is working to protect the forest, but nothing about baselines. Mickey Mouse’s role is not mentioned…

  17. I think that the validity of the Cumulative Deforestation Model and therefore VM0009 in general is an important topic. It does seem to generate results that intuitively seem unlikely and risky given that deforestation can decline dramatically on its own due to changes in market forces. However, in fairness to CI, from my reading, they only included the CDM model in their project description as a means of showing the conservativeness of the linear model (green line) they did choose to apply. Also not all the data from the project description is presented in this post. The CDM model that they didn’t use was built from observations from 6 different time periods, not just the two time periods shown in the chart, so it is entirely possible that the two rates shown in the chart are only a rough summary of what is a logistic curve only apparent when examining more points in time.

    If you take just the data from the chart presented in this post, CI saw a tripling of deforestation between the first and second five year period of thier baseline. If that trend continued they would have a deforestation rate above 1% just for the first 5 years of the project and 3% over the 10 year period. They also found that the underlying data could fit a logistic CDM, which they decided not to use. Instead, they simply made a linear model, which if projected over the next ten years results in an average annual deforestation rate of a little under 1% per year. Finally, this baseline is only valid for 10 years, at which time the rates will be reassessed based on new data.

    The challenge for REDD projects in frontier areas is that they know that deforestation rates will increase over time (as evidenced by the CI baseline). There are three options. One is to simply take a historical average, but this will severally underestimate the effect of the project in most frontier areas and perhaps leave the project without enough finances to succeed in the face of increasing deforestation pressure. The other option is to try to model the future rates, which is admittedly a tricky business and thus why CI attempted to justify the rate they did apply by showing the CDM alternative. The third option, which I think is more justifiable is to simply expand the reference region until it captures areas that have undergone the complete transition expected in the project area, take the average deforestation rate from that reference region and apply it to a model of the future location of deforestation for the remaining forest in the reference region. However, the reference region size required to accomplish this might be beyond the capabilities of an individual REDD project and the boundaries have to be constructed in an unbiased manner. As national REDD programs come online, this will be easier to do if a country has wall to wall mapping of thier baseline and decides at the national level how to divide the country into appropriate large scale reference regions.

    Another possible solution is to shorten the validity of the projected baseline to 5 years.

    National level accounting also offers the promise of facilitating something along the lines of stock flow accounting at the subnational level, which will enable REDD actors in frontier areas to benefit from carbon stock based payments to help offset the fact that the historical baseline may underestimate thier future deforestation pressure.

    I seldom see potential solutions to problems with REDD discussed on this website, but I hope that other REDD observers see the verified portions of the voluntary carbon market as a useful grounds for testing approaches and ideas and not just as an object of scorn. Disney is not under any obligation to offset emissions and by supporting this project, they have enabled CI to test a wide variety of activities to protect forests. If they had decided to simply fund a protected area without attaching it to REDD, they would have escaped the critisms of this website, but not had anywhere near the ability to measure the impact of thier investment. Even if you only have non-carbon interests in forests, the MRV approach used in REDD is one of the best means to measure the effect of forest conservation activities precisely because it relys on counterfactuals. The quality of these counterfactuals is debatable in some cases, but in this case, CI did choose a much more conservative baseline option than what was available to them.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>