Map Asia 2009

Abstract
Emissions from deforestation and forest degradation in developing countries constitute about 20 percent of the total global of greenhouse gases annually. These large emissions are not included today under the United Nations Framework Convention on Climate Change (UNFCCC) or its Kyoto Protocol, but will likely be included in the global post-2012 climate regime.

The method of choice for creation of data that is usable for applying for REDD funds is obviously remote sensing. Only using satellite data it is possible to monitor large areas of forest and their respective changes in countries such as Indonesia. Spatial data that is stratified and defines the extent of forest areas has to be produced repetitively and then combined with biomass carbon measurements.

In West-Kalimantan (Indonesia) the German Development Service is currently supporting the Provincial and one District Forestry Service in preparation for REDD. As the actual measurements and the application for funds will possibly later be carried out on district level, the preparedness activities focus on capacity building, support and aggregation of data later created in the districts. As pilot district Kapuas Hulu is involved in these activities as it is a pioneer in the use of GIS technology in Kalimantan and part of the Heart of Borneo initiative.

Introduction
About 20 percent of the global emissions of greenhouse gases are resulting from deforestation and forest degradation in developing countries. It can therefore be stated that these activities are an important driver of current climate change. In spite of this fact these emissions are not yet included under the United Nations Framework Convention on Climate Change (UNFCCC) or its Kyoto Protocol. In order to avoid the most serious impacts of climate change, humans will have to significantly reduce the amount of greenhouse gas emissions including those that result from the reduction of forests.

Reducing emissions from deforestation and forest degradation (REDD) describes a mechanism for supporting this aim. It is likely that REDD will be included in one way or another in future international climate change agreements, but the details will yet have to be decided upon. The basic idea of REDD is to reward individuals, communities, projects and countries that reduce greenhouse gas (GHG) emissions from forests. The key elements of REDD will undoubtly be carbon accounting, baselines, emission reduction strategies, monitoring/verification and sale of emission reductions. It is very obvious that a number of these key elements can only be handled using geographic information technology and remote sensing data.

Although satellite-based monitoring methods for REDD still have some limitations, the existing approaches have already been demonstrated to be practical for determining baseline deforestation rates against which future rates of change can be assessed, while satellite monitoring of changes in carbon stocks has advanced substantially in recent years. A wide range of optical, radar and lidar satellite sensors with different resolutions are available to support tropical forest monitoring efforts. A major information resource to help identify methods for monitoring and carbon stock assessment is a “sourcebook”, compiled by GOFC-GOLD, a working group on earth observations for REDD. In the mentioned book, the role of satellite remote sensing as a tool for monitoring forest cover changes is strongly emphasized.

The German Development Service (DED) started in the beginning of 2009 to support the province of West-Kalimantan in its preparation for REDD. A major part of this program is aimed at capacity building in regard to spatial information technologies, especially current remote sensing methodologies.

REDD basics


Fig 1: Base building blocks of REDD (From: Cortez and Stephen 2009)

The details of how the REDD mechanism will be included in future international climate change agreements have yet to be decided upon, but there are base building blocks that will be included without doubt. Figure 1 describes these building blocks and how they depend on each other.

Carbon Accounting
A REDD project or strategy has to quantify explicitly its benefit to receive compensation. Carbon accounting includes the use of remote sensing data to measure and monitor land cover change as well as field inventories to measure the amount of carbon that is stored in a forest. Combining these two sources of data allows calculating annual carbon emission rates which are the basis for defining baselines.

Baselines
The definition of baselines or reference levels is difficult but also very important for REDD. Actually there are three different things that are referred to as baselines in this context: historical baseline, BAU baseline and crediting baseline.


Fig 2: BAU and crediting baselines (From: Angelsen 2008)

The historical baseline describes the rate of deforestation and degradation and the resulting CO2 emissions over the past years. The Business As Usual (BAU) baseline refers to a scenario how emissions would develop without REDD activity, while the crediting baseline is the benchmark for rewarding projects or countries if emissions are below that level.

Emissions Reductions Strategies
In order to actually receive incentives a reduction of emissions has to occur. Strategies used in the context of REDD must be credible, verifiable and they additional, i.e. they would not have been carried out without REDD financing.

Monitoring and Evaluation
Reductions of emissions resulting from a REDD project have to be verified by an independent third party. When verified, the emission reductions will be certified and afterwards sold, traded or retired. To ensure that emission reductions are permanent both satellite and on-the-ground monitoring will need to be continued.

Sale of Credits
When the emission reductions from a project have been verified and certified they can be sold either through a carbon market, a private buyer or in exchange for donations from a public fund.

REDD in the political discussion
The United Nations Framework Convention on Climate Change (UNFCCC) set an overall framework for intergovernmental efforts to tackle climate change that was put into force in1994. Every year its Conference of the Parties (COP) meets to review the implementation of the Convention, adopt decisions and negotiate new commitments.

The first concrete protocol negotiated by the UNFCCC was the well known Kyoto Protocol in 1997 and entered into force in February 2005. The Kyoto Protocol set up an International Emissions Trading (IET) market and specific emissions targets for countries. The establishment of rules how to achieve these targets were developed in Marrakesh in 2001, including rules for emissions from land use, land use change and forestry (LULUCF). According to these rules developing countries can claim credits from afforestation and reforestation but not from avoided deforestation.

The REDD concept was first introduced at COP-9 (2003) and at COP-11 (2005) an official proposal on REDD was submitted. The submission was welcomed by most parties and launched a two-year process to design an effective REDD mechanism. Finally in 2007, COP-13 in Bali called for the inclusion of REDD as part of a post-2012 mitigation strategy. Besides this, an outcome of COP-13 was the Bali Action Plan (Bali Roadmap) which is to conclude at Cop-15 in December 2009. Only then final decisions on many open issues regarding REDD will be reached.

Preparation for REDD in West-Kalimantan
Indonesia ratified the UNFCCC in 1994 and the Kyoto Protocol in 2004. Interest in REDD is high, Indonesia is one of the biggest emitters of greenhouse gases and still has extensive areas of forest. The German development cooperation decided to focus its activities on Kalimantan (Indonesian Borneo). Regulation P.30/Menhut-II/2009 issued by the Indonesian Ministry of Forestry defines the implementation procedures of Reducing Emissions from Deforestation and Forest Degradation. With Regulation P.36/Menhut-II/2009 the procedures on business licensing for the utilization of carbon absorption and storing in production forests and protected forests are defined.


Fig 3: Location map – Kalimantan Barat

One of the major activities of the DED in the Climate Change and Forestry program is to support preparedness for REDD in West-Kalimantan. This includes measures in Community Forestry Management and GIS/Remote Sensing. The activities aim at empowering the District Forestry Services to implement REDD measures and projects. For this aim DED supports the provincial Forestry Service in setting up a GIS Service Centre that will train and support employees of district forestry services in regard to forest monitoring, carbon accounting and land-use planning. These activities are carried out in close cooperation with the Forestry Service of Kapuas Hulu. Kapuas Hulu acts as a pilot district for REDD preparedness – trainings are carried out and evaluated, pilot areas are identified and data about these pilot areas is collected.

Besides training and support the project also aims at setting up information systems about forest resources on both district and provincial level. These information systems will encompass spatial as well as non-spatial (tabular) data and will not only support the upcoming implementation of REDD but also the general workflows of the forestry services.

Technical Challenges
The technical challenges in regard to REDD are manifold. The aspects can be grouped into those problems related to setting up reference levels (historical baselines) and those in regard to the actual monitoring and carbon accounting.

Definition of historical baselines
One of the most critical elements of an REDD scheme is the definition of reference levels. Most REDD submissions to the UNFCCC suggest using historical baselines but in many countries this data is not available. The already mentioned GOFC-GOLD Sourcebook describes methods for establishing credible reference levels using existing scientific and technical tools. It is technically possible to monitor forest area change from 1990 onwards while estimation of forest carbon stocks can be obtained from a variety of sources.

The GOFC-GOLD Sourcebook suggests that a minimum requirement for establishing a historical reference scenario is the usage of mid-resolution (Landsat-type) remote sensing data for years 1990, 2000 and 2005, depending also on availability and cloud cover. It is also suggested to use a hybrid approach combining automated digital segmentation and/or classification with visual interpretation and/or validation of the results. This will allow assessing gross deforestation.

In this context it should also be considered that certainly a number of stakeholders, both governmental and non-governmental, are likely to have carried out mapping of forest and the deforestation process. Part of the project will be to identify these stakeholders and evaluate existing data and analyses.

Forest monitoring and carbon accounting
The implementation of a monitoring approach first requires the selection of a forest definition and the designation of initial forest area for which satellite data will be acquired. The actual selection of satellite imagery follows and depends on intended resolution of analysis, available funds but also on limitations of availability of usable data. Especially in Kalimantan cloud cover is frequent and extensive resulting in the probable need of several scenes of imagery to cover a specific area. Besides optical sensors a number of sources of radar data (SAR) are now available that are able to complement and enhance analyses. The use of SAR data seems especially promising in Kalimantan as the problem resulting from extensive cloud coverage can be solved. First evaluations have been carried out regarding the suitability of satellite data that is freely available. Since beginning of 2009 Landsat 7 data is available at no cost but suffers strongly from the SLC-failure and in some areas in Kalimantan also from almost permanent cloud cover. Fig. 4 shows a collection of Landsat7 scenes of the years 2008/2009 that cover all of West-Kalimantan and that were because of relatively little cloud cover. It is obvious that the usability of this data is quite limited due to mentioned problems.


Fig 4: Selected Landsat7 scenes (band combination 5,4,2) for 2008/2009

Besides the open question of suitability of available data, different methods for analysis still have to be evaluated. It seems likely that a combination of automated methods using ground truth and visual interpretation is the most viable approach. Even more challenging will be the monitoring of forest degradation as the spectral signatures of original and degraded forest are often almost indistinguishable.

In regard to the estimation of carbon stocks current guidelines define three general approaches or “tiers” with increasing levels of data requirements and analytical complexity. Tier 1 requires no new data collection but uses default values for forest biomass and its annual increment. Tier 2 is similar to tier 1 but it improves its approach by using country-specific data (e.g. from existing inventories) and by resolving forest biomass at finer scales by delineation of more detailed strata. Tier 3 requires the highest level of effort, using actual inventories or parameterized models in combination with plot data. The already mentioned Indonesian REDD regulation states that approaches from tier 2 should be used, working towards tier 3.

All approaches require some amount of stratification, i.e. the division of a heterogeneous landscape into distinct sub-sections based on a grouping factor, in our case stock of carbon. Approaches to stratification require spatial information on forest cover within a country. A usual approach is to overlay a map of forest cover (created from satellite data) with ecological zone maps and then refine the results with indicators like elevation and distance to roads. An alternate approach is the application of regularly-timed inventories that are carried out only on the areas where land cover change (deforestation or degradation) actually occurred.

Outlook
Activities pertaining to the implementation of REDD have just newly started and as a final regulation is not yet finished by UNFCCC many questions still remain open. On the other hand, the complexity of the subject necessitates that preparation starts early, keeping in mind that most countries that are eligible for application of REDD funds data infrastructures and processing capacities are rather limited. Capacity building generally, as well as in the described activities in West Kalimantan, has to start from the basics and if a Training of Training (TOT) approach is chosen the trainers have to be empowered first, before the potential implementers of REDD will benefit. To tackle REDD projects thorough knowledge of GIS technology, Remote Sensing methodology as well as of the mechanisms of carbon measurement is necessary. The activities of the Forestry Service in West-Kalimantan that are supported by the German Development Service target all of these aspects.

Literature
Angelsen, A. (ed.) 2008: Moving ahead with REDD: Issues, options and implications. Cortez, R. and Stephen, P. (ed.) (2009): Introductory Course on Reducing Emissions from Deforestation and Forest Degradation (REDD): A Participant Resource Manual.

GOFC-GOLD (2008): Reducing greenhouse gas emissions from deforestation and degradation in developing countries: a sourcebook of methods and procedures for monitoring, measuring and reporting, GOFC-GOLD Report version COP13-2.