June 29, 2018

Serious dataset errors undermine claim of support from Norway

( - Indonesian Peat Agency Chief Nazir Foead made a questionable claim at the Oslo Tropical Forest Forum (Jun 28) by stating that the improved quality of Indonesia’s peat maps was due to support from Norway, and for that he was grateful to this country.

This spurious claim needs correcting, bearing in mind that the Norwegian-funded LiDAR mapping in parts of the agency’s targeted peat restoration areas contained serious errors on both a technical and legal level.  

As previously reported by in November last year, Indonesian Environment and Forestry Minister Siti Nurbaya, who is also a spatial expert, confirmed that she had decided to not yet use the datasets from the Norwegian-funded LiDAR mapping, considering the level of bias in them. 

When asked for an update on this matter (Jun 29), the chairmain of the ministry’s improved peat governance team Bambang Hendroyono, who is also the ministry’s secretary general, reaffirmed that among the approved peat recovery legal documents for almost 300 forestry and palm oil concessions, there were none at all that use datasets derived from the Norwegian-funded LiDAR mapping.

“As the peat authority, there is no new policy formulated by our minister based on datasets from the results of the Norwegian-funded LiDAR mapping,” the secretary general asserted. 

It should be noted the only consideration in the formation of the peat agency by President Joko Widodo in early January 2016 was the acceleration of the restoration of 2015’s burned peatlands totalling approximately two million hectares.

The Indonesian Environment and Forestry Ministry produced 2015’s burned forest and land dataset, among other relevant peat maps funded by the Indonesian state budget, confirming that almost all of the peat agency's targeted peat restoration map published in mid-September 2016 was based on the ministry’s datasets.

Meanwhile, Karliansyah, the ministry’s director general in charge of peat protection and recovery, declared in a written message (Jun 29) that all peat hydrological units and peat ecosystem maps were made using the Indonesian budget. 

Dataset errors 

One of the datasets produced by the Norwegian-funded LiDAR mapping team under the coordination of the World Resources Institute (WRI) consisted of land cover data based on aerial photos. The objective of this dataset, as stated in the LiDAR mapping factsheet, was “to see the current state of peatlands.”

As it turns out, the WRI-coordinated LiDAR mapping team was found to have committed a grave error by leaving out almost all of the evidence of 2015’s burned peatlands in the land cover dataset it produced. The motivation behind the omission of this evidence is unclear. 

The photos below show the legal and ground evidence of 2015’s burned peatlands, as documented by the Indonesian Ministry of the Environment and Forestry. These were later classified by the Norwegian-funded LiDAR mapping as peat forests, thereby overlooking the reality of 2015’s burned peatlands.

Bizarrely, the Norwegian-supported LiDAR mapping team's misclassification of the burned peatlands went even deeper, categorizing them as secondary peat forests of high, medium and low density. This further reinforces the notion that the WRI-coordinated LiDAR mapping ignored the facts on the ground regarding 2015’s burned peatlands. 

Given these significant flaws in the Norwegian-supported mapping, the peat agency chief’s assertion that this mapping helped improve the quality of Indonesia’s peat maps should be rejected, from both a legal and technical perspective.

The following photos, taken by the team, also depict 2015’s burned peatlands, which went on to be wrongly classified by the Norwegian-funded LiDAR mapping team as peat forests. 

Considering that the main purpose of the Norwegian-LiDAR mapping was to identify sites for peat restoration interventions, this makes the neglect of ground-based evidence of 2015’s burned peatlands an even more momentous legal mistake, especially because decisions on such interventions in peat areas rely heavily on a land cover dataset.

Seeing that the land cover dataset produced by the WRI-coordinated LiDAR mapping team was completely inaccurate, and in contradiction to the legal and ground-based facts, it follows that the Norwegian-supported mapping did not result in accurate peat maps in any way. 

All of this makes the claim that Norway has bolstered the peat agency’s peat restoration acceleration mandate nonsensical. On the contrary, the misclassification of 2015’s burned peatlands as non-burned peatlands by the Norwegian-LiDAR mapping team has arguably impeded peat restoration efforts.

In fact, the data used by the peat agency - derived from the WRI-coordinated mapping team’s dataset - can be considered manipulative due to its wilful neglect of almost all the evidence of 2015’s burned peatlands. 

Any doubts about whether the land cover dataset produced by the WRI-coordinated mapping team is truly manipulative can be dispelled by looking at time-series data from Google Earth as well as other relevant satellite data. This evidence is accessible to everyone.