“Statistics can be made to prove anything — even the truth.” (Noël Moynihan)

Two weeks ago in the 19th UN Forum on Forests in New York, delegates from the Ministry of Environment and Forestry of Indonesia presented their achievement of reducing deforestation in the country from 113,500 hectares during one annual period to 104,000 hectares during the next annual period. A net deforestation of 8.4 percent. Are congratulations in order? Not so fast. These numbers may not be the only statistics out there.

Global Forest Watch, which was launched by the World Resources Institute, an institution with a humble beginning dating back to the year 1997 when a network of NGOs launched an effort to compile up-to-date reports about the state of the forests in four pilot countries (Canada, Cameroon, Gabon, and Indonesia), also issues their annual studies which can be publicly accessed on their website. Two other Indonesian-based institutions, namely Auriga Nusantara, which created the Biomass Map of Indonesia and TheTreeMap, whose open-access platform Nusantara Atlas has been used by many to track forest fires in Indonesia, Malaysia, and Brunei Darussalam, have also contributed their shares on deforestation statistics.

Below are their numbers for the year 2022.

The gulf separating the numbers issued by the MoEF (KLHK) and the other three institutions (GFW, TheTreeMap, and Auriga) cannot be more obvious. What gives?

The reasons for this pretty much boil down to four things: minimum mapping unit, period when data are gathered, differences in the definition of a primary forest, and differences in the definition of deforestation itself.

1. Minimum mapping unit.

While the satellite images gathered by GFW, TheTreeMap, and Auriga use a minimum mapping unit size of 30 meter by 30 meter (0.09 hectare), following the world’s standards set by the Food and Agricultural Organization, the images gathered by the MoEF are 250 meter by 250 meter (6.25 hectare), resulting in fewer amounts of polygons for the same rendered areas. When blown up to detect deforestation which takes place in areas less than 6.25 hectare, the MoEF’s images, which have lower resolution, are more likely to miss deforestation events. Combined with the fact that the interpretation of the images is done manually in the case of the MoEF, which means a lot will depend on the skills of the technician, who is doing the task, instead of semi-automatically by machine learning as in the case of Auriga, you risk underestimating the amount of deforestation.

2. Period when data are gathered.

The MoEF gathers their data from June of a particular year to July of the following year, whereas the other three institutions follow calendar years (January to December). As deforestation is greatly influenced by regulations which often take some time to take effect after their issuance, for instance a government moratorium on opening new areas for palm oil plantations in what have been designated as natural forests, differences in choosing which months to gather field data, may lead to different results of deforestation. This in itself should not be a problem if all parties were willing to publicly release their data, not only the results, for the others to check and do some corresponding adjustment.

3. Definition of a primary forest.

How to define what can be categorized as a primary forest has also been a source of dispute between the MoEF and other institutions. In the MoEF’s terminology, natural forests which have been disturbed by logging or other human activities are categorized as secondary forests, when they could very well be non-intact primary forests in the GFW’s terminology. This seemingly trivial (mis)attribution has far reaching consequences. It brings down the total number of primary forests which the country has and, in turn, the total amount of deforestation taking place within them.

4. Definition of deforestation.

The definition of deforestation itself is quite different between that adopted by the MoEF and other institutions. For instance, the University of Maryland, the initiator behind the mapping of tree cover change around the world called the Global Land Analysis and Discovery, defines deforestation as any loss of at least thirty percent of tree canopy. The MoEF, on the other hand, defines deforestation simply as a change of class from forest to non-forest in their land cover type classification. And while the former only takes into account forests which have been felled as deforestation, the latter considers non-forest areas which have gone through reforestation programs in their equation as a gain in the size of forest areas, resulting in what is called net deforestation. These statistics fail to point out that it would take decades before newly-reforested forests can match old-growth forests in terms of their biodiversity richness and their functions to support complex ecosystems.

It is this net deforestation number of 8.4 percent mentioned earlier which was shown by the MoEF delegates to the UNFF19 forum attendants. Along with the introduction of an internal forest monitoring system and a land cover classification system with 23 different types in it, the forum attendants can perhaps be forgiven to — excuse the pun — miss the forest for the trees.

With all these differences and disputes, financial aid schemes which rely upon deforestation results certainly must take these into account. Unfortunately, bilateral agreements or partnerships may overlook some stringency measures to which private enterprises have been subjected in the voluntary carbon market. Norway’s International Climate and Forest Initiative (NICFI), for instance, cited the deforestation number released by the MoEF during 2021/2022 on their webpage without a disclaimer at how this number was arrived at or whether there have been other institutions which release their own studies with numbers which are double these. (Set up by the Government of Norway as an initiative to help save the world’s tropical forests with up to USD 276 millions every year at their discretion to disburse to forest-rich countries such as Indonesia, the NICFI has so far contributed USD 156 millions to Indonesia’s Environment Fund through three stages of result-based payments.)

All things aside, assuming that the GFW’s deforestation number is accurate, with the loss of almost a quarter of a million hectare of primary forest in just a year, Indonesia’s forests are disappearing faster than those in Peru and Colombia, though the country fared better than Brazil, Democratic Republic of Congo, and Bolivia did for that year.

However, a bigger look on the past two decades reveals that the top three countries which lost their humid tropical primary forests have been persistently the same: Brazil, Indonesia, and Democratic Republic of Congo.

The fact that the countries which own the largest rainforests in the world are also the ones which lost them the most should not be surprising. For generations, governments in these countries have relied on nature’s riches for income. Unfortunately, so far nature has only been valued for the commodities existing within; a purely extractive value. The value of nature as infrastructure is entirely omitted from the national accounting when it should have actually been valued the most.

The current carbon market may provide a short term solution for this problem although the mechanism is still not without flaws. For instance, within the soon-to-be-launched method by Verified Carbon Standard or Verra, called VM00048, jurisdictional baseline templates will be used nation-wide to measure how impactful a certain forest protection program is compared to a business-as-usual scenario. It is thus extremely important to ensure that, when applied in countries with many islands such as Indonesia, this method will not undervalue islands which now may still be HFLD (High Forest, Low Deforestation) but are undergoing a massive transition to be HFHD (High Forest, High Deforestation).

As for the long term solution, we at Forestbase work towards bringing nature to market for the full value of its ecosystem services stack: from carbon to biodiversity, from water to erosion control, from soil integrity to pollination. That can only be done by building precedent cases that involve the nature land itself in the financial instrumentation.

‍