6. Investigating Harvesting: Desk-Based

The principle of investigating legality at the point of harvest is quite simple. It involves comparing official reference data that reveals what harvesting is allowed, and under what conditions, with observations of what is actually happening in the forest.

The greatest challenge is in accessing the necessary information. Official reference data, that determines what is allowed, will commonly be held by governments, which are often reluctant to disclose it. Determining what is actually happening, on the other hand, can present technical, logistical and security challenges. This section explains where these kinds of datasets can be found, and how they can be compared with one another at each step of the investigation to identify illegality.

Defining a target

Investigations begin with indicative evidence, or a hypothesis. Indicative evidence may take the form of testimony from a community that illegal logging is taking place in its territory. Or it might take the form of a news article quoting a government official, stating that most plantation companies in a given district are clearing forest without the necessary timber harvesting permits. This evidence defines a target or targets: whether it is a named company, a group of companies, or a type of company. Where there is no clear information on the perpetrators, the target might be a geographic area, or even a particular species that is subjected to over-extraction.

A target may be arrived at by working back from the market. Trade data may identify a specific company, that is engaged in harvesting, as among the leading exporters to sensitive markets. In such cases the preliminary evidence that the company is engaged in illegality may not be strong, but its significance within the sector and supply chain may merit investigation. This would particularly be the case where the rates of illegality are known to be high within the source country. Where the investigation has begun by identifying retailers or importers of high-risk products, the target may be identified by working back systematically through their supply chain. For such cases, it may be appropriate to begin the investigation with processes described in Section 10: Tracking timber to end market.

Obtaining permit data

As mentioned, official data can be difficult to obtain. To obtain it, it is critical to cast the net wide, both in terms of the data that is sought and the places in which it is sought.

Aside from permits that are specific to the target company or area, it is equally important to gather as much contextual data as possible because the comparison between different datasets can provide important answers. Key examples would be aggregated data on timber harvesting in a given region, and spatial plans or forest zoning that designate areas for logging or conversion to agriculture. It is also important to bear in mind that information on a license area of interest may often be included in documents relating to neighbouring areas.

The internet is the most accessible source of relevant permit information. Data may be published by government agencies themselves, on their websites. It may also have been obtained and published by third-parties in the past, such as newspapers or NGOs. For example, permit information (including boundaries and licensee names) is now available for many forest countries via the World Resources Institute’s Global Forest Watch website. Reports from conservation organisations, regarding protected areas or general land use planning, also often have detailed maps of adjoining logging, mining or plantation companies. Companies may also publish information about permits they have obtained, including in annual reports and official announcements.

An especially rich source of information are ‘prospectuses’ published by companies in advance of stock exchange listings. Where they are members of a certification scheme, such as the Roundtable on Sustainable Palm Oil or Forest Stewardship Council, websites of the scheme or individual certifiers often contain useful information. Intelligent use of search terms and an appreciation of the limitations of search engines is essential when carrying out online research (See Online Sources of Information).

Some information may be in the public domain, but not on the internet. NGOs, particularly those local to the area of interest, often hold unpublished data they have obtained from the government in the course of their work. Communities can present a particularly rich source of permit data, which they may have been given during consultation processes, by the government or companies. Even in areas where the rights of communities are weak, there sometimes exists a responsibility to provide information to them. In many cases community members will obtain employment from companies that are operating in or adjacent to their territories, which provides further access to information.

Some governments publish relevant information only in hardcopy, either as announcements in newspapers or in official journals. Where information is not in the public domain, it must be sought directly from relevant government agencies. However, in most regions lack of transparency and collusion between officials and companies present challenges. In many countries data management is also poor, and records may not be complete even if they are accessible. Data may be deliberately disorganised and even falsified to avoid thorough scrutiny. Nonetheless, obtaining data through formal channels can support a robust evidence base.

It is important to note that as companies are subject to a range of different regulations, permission invariably comes from a range of sources, within different government departments and at different levels of government, from the local to the national. Where some sources may be reluctant to release information, others may be more forthcoming. Some countries, such as Peru and Indonesia, have introduced Freedom of Information laws, which give citizens the legal right to access certain types of information (See Freedom of Information Laws).

Analysing reference documents: What do the permits tell you?

The next step is to compare the permits with (a) the regulations governing them, and (b) each other. This will identify illegalities in the permitting process itself, and whether permits are missing, incomplete, or issued out of turn.

Research by NGOs, government and research institutes in almost every forested country provide summaries of how the permitting process should function in practice. The permit data that has been obtained should be ordered and cross-checked against this, highlighting any deviances from the process on paper. Though there might often be permits missing, such a finding should be treated with caution, as the permit may exist but not have been obtained. The significance of the finding varies according to the importance of the permit. For example, a missing Environmental Impact Assessment or forest management plan is a critical finding; other bureaucratic requirements may be less so.

Subsequent to this structural comparison, the content of the permits should be interrogated. Documents that are integral to the right to harvest – Environmental Impact Assessments, forest management plans, contracts and others – will contain narrative data that can be compared to the regulatory framework. This process will entail a more detailed understanding of the content of the regulations and regulatory framework, which can be complex. Reference to legal analysis and, where possible, expert legal advice at this stage can be useful in determining some subtle, but serious, forms of illegality. As with the structural permit analysis, the important findings may not be in what is included, but in what is not included. For example, where legally-required social obligations to communities are not included in contracts, or where there is evidence that communities were not consulted during Environmental Impact Assessments.

Reference to legal analysis can be useful in determining some subtle, but serious, forms of illegality

In some cases, the permit data may even provide concrete evidence that companies have violated the law by beginning operations before the permits were obtained. This is particularly the case with Environmental Impact Assessments which, where done properly, should provide some analysis of current conditions in the concession or targeted area. In Indonesia, landcover analyses within assessment documents have shown that deforestation for plantation development began before the assessment process. In Sarawak, Environmental Impact Assessments have shown that logging companies began re-entry logging before they were legally entitled to do so (see Case Study 2).

At this stage the research process should seek to identify data that may not immediately be useful, but will become so as the investigation progresses. Critically important datasets that will be found in the permit data include:

  • Projections for volumes of timber that will be harvested in a given area. This can later be compared with estimates of volumes harvested based on fieldwork, or volumes exported. This is significant in identifying the under-declaration of volumes to avoid taxes, or the over-declaration of volumes to facilitate log laundering into concessions.
  • Concession boundaries. These will later be compared to land cover change using satellite data, and GPS data from fieldwork. Where they are found in permits they are likely to require digitisation before such analysis can be carried out. It is notable that different permits may contain different boundaries for the same concession, so they should be treated with caution.
  • Cutting plans that define which blocks can be cut, and when. This can also be compared to the reality on the ground using both satellite analysis and fieldwork.
  • Areas that are off-limits for cutting, either in cutting plans, forest management plans, Environmental Impact Assessments or other documents. Again, these can be compared to satellite imagery and fieldwork evidence.

The methods employed by Greenpeace to identify illegal logging in the Brazilian Amazon are a good example of how painstaking data gathering and permit analysis can generate strong leads and direct field investigations to concessions with a high probability of illegality (see Case Study 1).

By this stage, it may alternatively be clear that no permits have been issued in the area of interest. In such cases moving on to the mapping and fieldwork stages may provide more answers. However, research should also be broadened to encompass other companies and operations, whether concessionaires or sawmills. Examining the routes out of the area – typically roads but often also rivers – can lead to operators with licenses nearby, who may be laundering timber from areas without any authorisation.

Often where there is no right to harvest, the picture at the point of harvest is complex and opaque. The process of harvesting itself may appear sporadic or disorganised. Yet in many instances the timber that is being harvested will be aggregated at sawmills or downstream facilities that are being run by companies in a more organised operation. This type of operation has been identified in both Peru (see Case Study 6) and Brazil (see Case Study 8). As such, looking at ostensibly legitimate logging or processing operations and working back may provide more answers than looking at the point of harvest.

Eye in the Sky: Comparing permits with data from satellites

The next stage in the investigation is to compare data found in the permits with other, non-permit data. This can identify where the provisions that have been identified through permit analysis have been complied with. Boundary maps, cutting plans, and prohibited areas that were found during that process become of critical importance here. They can be overlaid with other spatial data and satellite imagery and used to directly detect some types of illegal logging or help direct the field work required to document other types. In Sarawak, for example, maps included in Environmental Impact Assessments have been compared with satellite imagery to demonstrate logging outside concession boundaries and other offences (see Case Study 2).

Example of illegal logging detected via Landsat imagery
Example of illegal logging detected via Landsat imagery

Until recently, analysing land cover change to detect logging or forest conversion required ownership and knowledge of Geographic Information System (GIS) software and the purchase and processing of expensive satellite imagery. However, rapid advances in processing of imagery and the development of online GIS platforms, have made the technology more accessible and easier to use. They are increasingly making even high-resolution satellite imagery freely available in user-friendly formats.

Google Earth, which is free to download, hosts satellite imagery in varying resolutions. Most areas are covered with a resolution of approximately 15 meters per pixel (from Landsat satellites), which is sufficient to determine clearance and the spread of logging roads associated with selective logging into virgin forests. However, some areas display imagery at a resolution of 60cm, which enables the identification of very small areas of clearance and can be used to document cutting in river buffers or clearance along logging roads in excess of legal limits. Google Earth also hosts historical imagery, that allows changes in cover to be identified over time. This satellite imagery is sourced periodically by Google from third-parties. It is now relatively easy for NGOs to search for, identify and obtain additional high-resolution imagery from the same providers directly (see High-res Imagery).

Users can upload both concession boundaries and other contextual spatial data to Google Earth. This enables the analysis of forest cover changes within concession boundaries, but can also show whether concessions fall within protected areas, community territories, or other zones where harvesting is prohibited.

In 2013 the World Resources Institute relaunched Global Forest Watch (GFW), an interactive online forest monitoring and alert system. GFW hosts a range of geographical datasets that can be used to analyse and identify illegal logging, including forest change, forest cover and forest use data. The latter includes concession maps (including names of licensees) for logging and plantation concessions in many forested countries, though the data is known to be incomplete. The data should be treated with caution, since some boundaries are not drawn precisely and some information is likely out of date.

As with Google Earth, GFW allows users to upload their own spatial data and carry out analysis. Unlike Google Earth, however, much of the analysis on GFW is automated. It allows users to view and quantify forest cover loss (identified automatically from Landsat imagery) within a user-defined area over time and create alerts for future loss. In 2016, GFW made a new dataset available that also provides raw satellite imagery. This imagery is more recent and much more regularly updated than that available in Google Earth, and in some cases also of a higher resolution. Comparing land change in satellite imagery over time with permit dates can present a prima facie case that logging took place before the correct permits were obtained.

In many cases, concession maps will not be available in the course of an investigation. In these instances, Google Earth and GFW can be equally valuable in refining the location in which illegal logging is taking place, and quantifying the scale of it. While this may not move closer to identifying the perpetrators, it can provide clues as to whether the activity is on an industrial – or small-scale, and identify if it is taking place in areas where concessions are not legally allowed to be issued. It can also assist in guiding next steps, particularly locations for fieldwork.

Satellite imagery, and especially ‘forest loss’ maps extracted from it automatically (as GFW does), should be treated with caution. At lower resolutions it is not possible to determine if clearing is occurring in forests, or other types of vegetation such as farmland, scrub or even plantations. Automated analysis may not necessarily show clearing, and selective logging may not be visible in lower-resolution imagery, particularly if it is low-intensity or in forest which is already disturbed. It is also not possible to determine if commercial timber is being produced from disturbances seen, and if so at what volumes, let alone to determine who is doing the cutting. Mapping and satellite imagery analysis is useful for building the data, filling in parts of the picture and, particularly, guiding fieldwork where the questions it raises an be answered.