Renoster allege that the Southern Cardamom REDD+ Project (SCRP) has been using an inflated baseline – this claim is not credible. Their analysis presents a limited, inaccurate view of the reasonableness of the Southern Cardamom baseline.

Based on multi-modal analysis that incorporates high quality geospatial data, and which contains critical information that is unavailable through satellite data, the baseline under which the SCRP operates well reflects the actual threat level faced by the project. 

Renoster’s findings are flawed – here’s why: 

  1. The project uses the national FREL (Forest Reference Level) as its baseline, as allowed in the methodology. It is true that deforestation drivers at the national level may not always align entirely with local deforestation drivers, but this does not necessarily discredit the use of FRELs for project-level baselines – a more rigorous analysis and understanding of local circumstances is required to make such an assumption. Renoster’s analysis is too limited in its scope and quality of underlying geospatial data to make this assumption, and there are circumstances that exist locally that cannot be analyzed from desk-based geospatial data alone.

  2. Renoster uses Global Forest Watch (Hansen et al.) data which is known to be highly uncertain – particularly in Cambodia – and should not be used for project-level analyses. The Global Forest Watch team has been very clear that their datasets should only be used for national/regional analyses, not to assess local deforestation (see blog here) and that you should avoid making strong claims with this data. These findings are echoed in an independent accuracy analysis of Hansen et al. data in Cambodia published as part of this new research (in pre-print); the analysis finds that Hansen et al. data has an uncertainty of ±59% within this context. Such uncertainty is mainly due to problems mapping seasonal forests and misclassification of plantations as forests – together resulting in the prediction of far less deforestation than in reality (thus giving the appearance of overstated baselines).

  3. Renoster analyzes baselines using only one methodology. However, there are many different methods that can be used to assess baselines – as outlined in the paper (see pre-print). Each method has limitations – particularly when applied using low quality geospatial dataset as Renoster has. Instead, a series of methods and higher quality datasets must be used to assess a project’s baseline.

  4. The Southern Cardamom REDD+ Project is at higher risk than it appears. There is a history of extensive Economic Land Concession (ELC) proposals (covering nearly 1 million ha; see pre-print Supplementary Figure 5 and Table 6) within the project area and buffer regions. Ultimately all ELCs to date have been revoked as a result of campaigning involving Wildlife Alliance. In the absence of the project, these ELCs would have been awarded, resulting in roughly half of the project area being designated as concessions today. Such concessions have on average much higher rates of deforestation than the baseline used for the project. Renoster’s analysis fails to consider this.