One of the major challenges for the REDD+ mechanism, is the provision of reliable estimates of emissions reductions on which performance-based payments can be made. General guidance from the UNFCCC states that all estimates should be transparent, consistent, as accurate as possible and should reduce uncertainties, as far as national capabilities and capacities permit, and are subject to an independent review. Remote sensing technologies are a central part of REDD+ monitoring systems because many existing and upcoming missions provide open access to temporally and spatially explicit information on land cover and use. Dedicated research and development (R&D) implemented through the Global Forest Observations Initiative (GFOI, www.gfoi.org) has been instrumental in developing technical solutions to integrate remote sensing and ground based observations for national forest and GHG monitoring. In 2016, the Land Cover Project Office of the Global Observation of Forest Cover & Land Dynamics (GOFC-GOLD) took the lead of the GFOI R&D Component with the support of the European Space Agency.

GFOI works on the understanding that joining forces of international partners can better support REDD+ countries, and moreover provides the breadth of expertise required to address all issues. In addition, GFOI partners aim to provide a more coordinated and consistent package of support where efforts are not replicated but built upon systematically and complement each other. A number of available tools and resources for REDD+ monitoring which were developed through the GFOI include the GOFC-GOLD sourcebook of the science related to estimating emissions and removals for REDD+, and the GFOI Methods and Guidance Document (MGD): which provides step-by-step advice on the design and implementation of MRV systems for REDD+ countries.

Furthermore, GFOI’s R&D priorities are identified through Country Needs Assessments (CNAs) undertaken by the Initiative, and also through dialogue with partner organizations who are actively supporting countries in REDD+ monitoring.

Priority research topics related to utilizing data from new satellite missions in REDD+ monitoring systems will be discussed in the presentation including;

• Integration of newer data sources, i.e. Sentinel data with existing Landsat time series. In the context of near-real time data analysis, the integration of multiple data sources can provide benefits in terms of increasing the temporal density of data, however research gaps remain and these will be discussed. The cloud based platform, SEPAL, of FAO which facilitates the access to these data sources will be included.
• Use of high resolution data from commercial sources for targeted sampling in stratified area change estimation (for long-term monitoring). The key challenge here is the estimation of uncertainties from these approaches. GFOI has already produced guidance, which describes experiences with and options for improving methods used to estimate uncertainties in activity data, which is available through the REDDcompass: https://www.reddcompass.org/mgd-material.
• Describing best practices related to estimating uncertainties from emissions factors / biomass. Identifying and estimating the main sources of uncertainty, accounting for uncertainties due to varying sampling methods, and the integration of biomass maps, and plot data with National Forest Inventory data. Approaches for combining uncertainties (for example from AD and EF), including the use of Monte-Carlo will be discussed.
• Biomass mapping from space. A series of upcoming space-based missions (e.g, ICESAT-2, GEDI and NISAR and Biomass) provide new opportunities for large-scale biomass mapping. The potential for this data to meet users’ needs depends on the users’ definition of biomass, timeframe of interest, and area of interest. The potential for these data and approaches to both integrate and validate biomass maps with ground data sources will be discussed.
• Early warning (EW) forest monitoring systems. Research related to time-series analysis in near-real time is central to this, and the upscaling of demonstration case studies is a priority topic. Additionally, a number of research questions were identified that proponents of EW systems need to address in order to make best use of the available data sources to inform decision making and action. Understanding drivers of change, and whether and how to respond to the change detected are research questions being considered by the GFOI.

Achievement of international goals and national commitments related to forest conservation and management, climate change mitigation, and sustainable development requires credible, accurate, and reliable quantification and monitoring of forest biomass and carbon stocks. Of most prominence is the requirement of the Paris Agreement on Climate Change to transparently report progress on the implementation of National Determined Contributions (NDCs), which often requires dedicated improvements in the quality of national GHG inventories. In this context, there is an urgent need to obtain quality and transparent estimates of forest biomass stocks and changes that can be used by many stakeholders (government agencies, civil society, forest managers and the private sector) involved in the design and implementation of forest and climate policies. At the same time, there are rapid technological developments that can support the gathering of biomass related information, which include the series of upcoming space-based missions (e.g, ICESAT-2, GEDI and NISAR and Biomass). Many of these will open source data targeted at large area biomass monitoring, with this creating unprecedented opportunities to fill the existing data and information gaps for these applications.
In the presentation, we identify key policy drivers and forest management processes underpinning the need for a long-term collaborative effort among forest biomass data producers and users. Experience to date has shown disconnection between these two groups, with biomass datasets not being able to serve all users’ needs in terms of definition and type of biomass measurement, geographic area, uncertainty requirements and time frame (most recent versus long-term trend). Particular focus will be on:
1. The outcomes and requirements of the Paris Climate Agreement and the evolving REDD+ programmes (i.e. from the Worldbank FCPF, Green Climate Fund)
2. New estimation guidance and data requirements derivate form the 2019 IPCC Refinement of the Good Practice guidelines that includes updated Tier defaults and a section on the use of biomass maps
3. The potential strength of upcoming biomass missions to ensure continuity for long-term data provision, which one-off research missions cannot provide.
4. The integral role and efforts of international coordination bodies (e.g., GFOI, CEOS and GOFC-GOLD)
5. New approaches to better link space-based data with national forest inventories, ground research networks and biomass mapping efforts
We will conclude with an outlook and next steps with focus on the need for improving partnerships of data producers and countries for joint initiatives and data exchange, generating more practical country examples on integrating space-based data for improving national GHG inventories and developing approaches for continuous improvement and sustainability for national forest monitoring.

The need for accurate information to characterize the evolution of forest cover at the tropical scale is widely recognized particularly to assess carbon losses from processes of disturbances such as deforestation and forest degradation (Achard & House 2015). In particular, the contribution of the degradation is a key element for REDD+ activities and is presently mostly ignored in national reporting due to lack of reliable information at such scale.

Recently Vancutsem et al. (2019) produced a tropical evergreen forest change product at 30 m resolution, which delineates forest cover changes from the early 1980s until the year 2017. The use of the full Landsat historical time-series at high temporal and spatial resolution allows to monitor more accurately deforestation and degradation processes and to better assess carbon losses from disturbances in tropical forest cover compared to previous estimates. A degradation event is defined here as temporary absence of forest cover (visible within a Landsat pixel during a maximum of one three years duration) and includes impacts of fires and logging activities.

We quantified the annual losses in above-ground carbon stock associated to degradation and deforestation in tropical evergreen forests over the period 2010-2014 by combining the annual disturbances in forest cover derived from the Landsat archive with a pan-tropical map of aboveground live woody biomass density (AGB) at 500 m resolution valid for year 2008 (Baccini et al, 2012). We obtained the pre-disturbance carbon value of each pixel of our map by multiplying the biomass density by the area of a Landsat pixel and using a biomass:carbon ratio of 0.5. The carbon loss due to degradation or direct deforestation is accounted as full carbon loss within a pixel. As in the case of logging activities the detected pixels include areas where large trees are logged and hence have much higher biomass densities than derived from the pan-tropical map, our approach is conservative by underestimating the biomass loss from degradation. To avoid double counting of carbon removal, deforestation happening after degradation is not accounted as carbon loss in this study.

Our results can then be compared with estimates of previous studies that cover different periods and forest domains: (i) Tyukavina et al (2018) provide estimates of carbon loss from deforestation for the period 2000-2012 for all forests (evergreen and deciduous) discriminating natural forests from managed forests, (ii) Achard et al (2014) provide estimates of carbon loss from deforestation for the period 2000-2010 for the humid domain only, and Baccini et al (2017) provide estimates of carbon loss from deforestation and degradation for the period 2003-2014 for both evergreen and deciduous forests.

We produced first estimates of annual loss in above-ground carbon stock associated to degradation and deforestation in tropical evergreen forests (using the 2008 AGB map at 500m) at 458 TgC / year and 323 TgC / year respectively for the period 2010-2014. These estimates show a lower contribution of degradation to the total carbon loss than recently reported from a coarser resolution study (Baccini et al 2017), i.e. 59% from our study versus 69% for period 2003-2014 from this previous study. Our estimate of overall annual carbon loss from tropical evergreen forest changes (782 TgC/year) is at least 50% higher than previously reported for similar study areas (table 1): 528 TgC/year for ‘natural forests’ of the humid domain (Tuykavina et al, 2018) and 424 TgC/year for humid tropical forests (Achard et al 2014) for the 2000-2010 decade where the deforestation rate was 30% higher than during the 2010-2014 period. Moreover, although our estimate of carbon losses from tropical forest changes is 9.5% lower than reported by Baccini et al (2017) at 862 TgC/year, this latter includes all deciduous forests with 110 TgC/year contribution only from southern African woodlands (McNicol et al 2018) and covers the period 2003-2014 where the deforestation rate was 9% higher than for post-2010.

Further work will be completed and presented at the symposium. First, we intend to use two other recently published maps of aboveground live woody biomass (AGB): (i) AGB map at 30 m resolution valid for year 2000 (Zarin et al. 2016) and (ii) AGB map at 100 m resolution valid for year 2010 (Santoro et al. 2018). The use of these maps in combination with the tropical evergreen forest change product at 30 m resolution will allow to quantify carbon losses over a longer time period (from year 2000) and/or at finer spatial resolution (30m or 100m). The uncertainties of these carbon loss estimates will be assessed through the computation of standard errors.

Despite the unprecedented figures of deforestation and degradation rates derived from the new tropical evergreen forest change product at 30 m resolution (Vancutsem et al, submitted), the contribution of degradation to overall carbon losses is likely to be underestimated, in particular for sub-pixel level forest cover disturbances which are visible over very short-duration. In order to better estimate this contribution, the recently developed ΔrNBR approach (Langner et al, 2018) will be applied to the Landsat archive over two large regions (Mato Grosso state of Brazil and Indonesia). The ΔrNBR approach allows detecting even subtle changes in the forest canopy cover (with impacts at sub-pixel level) as for example caused by selective logging events. These sub-pixel canopy cover disturbances will be assessed on a bi-annual basis. It is expected to obtain more exhaustive assessments of degraded areas in these two regions and consequently more accurate estimates of the contribution of degradation to carbon loss.

In a further step, we plan to use Sentinel-2 data for monitoring the location and extent of logging activities and burnt areas. Mapping forest types and monitoring forest conditions (e.g. intact versus degraded) at tropical scale with finer spatial resolution from Sentinel-2 (10 m) is expected to further improve the estimates of carbon losses from forest degradation. Initial assessments are expected to be carried out in the two test regions.

References
Achard F et al. (2014) Determination of tropical deforestation rates and related carbon losses from 1990 to 2010. Glob. Chang. Biol. 564 20, 2540–2554
Achard F, House JI (2015) Reporting carbon losses from tropical deforestation with Pantropical biomass maps. Environ. Res. Lett. 10, 101002
Baccini A et al. (2012) Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nat. Clim. Change 2, 182-185
Baccini A et al. (2017) Tropical forests are a net carbon source based on aboveground measurements of gain and loss. Science 358, 230–234
Harris NL et al (2012) Baseline map of carbon emissions from deforestation in tropical regions. Science 336, 1573–6
Federici S et al. (2015) New estimates of CO2 forest emissions and removals: 1990–2015 Forest Ecol. Manage. 352, 89–98
Langner A, Achard F, Grassi G (2014) Can recent pan-tropical biomass maps be used to derive alternative Tier 1 values for reporting REDD+ activities under UNFCCC? Environ. Res. Lett. 9, 124008
Langner A et al. (2018) Towards Operational Monitoring of Forest Canopy Disturbance in Evergreen Rain Forests: A Test Case in Continental Southeast Asia. Remote Sensing 10, 544
McNicol IM at al. (2018) Carbon losses from deforestation and widespread degradation offset by extensive growth in African woodlands. Nat. Commun. 9, 3045
Santoro M et al. (2018) GlobBiomass - global datasets of forest biomass. PANGAEA, https://doi.org/10.1594/PANGAEA.894711
Tuykavina A et al (2018) Corrigendum: Aboveground carbon loss in natural and managed tropical forests from 2000 to 2012. Environ. Res. Lett. 13, 109501
Vancutsem C et al. (2019) Pantropical analysis of the Landsat archive reveals unprecedented deforestation rates in evergreen forests. Submitted to Nat. Commun.
Zarin DJ et al. (2016) Can carbon emissions from tropical deforestation drop by 50% in 5 years? Glob. Chang. Biol. 22, 1336–1347

Table 1: Comparison of estimates of annual losses of Above Ground Biomass from tropical deforestation and degradation (in Tetragrams of Carbon per year)
Table 1 is provided as jpg image

Within the frame of the REDD+ mechanism, developing countries obtain and receive results-based finance only if emissions reductions are measured, reported and verified (MRV) relative to forest reference emission levels. For this purpose, multi-temporal maps of aboveground and belowground biomass generated from Earth observation and in situ data are recommended. Global biomass maps are increasingly available and will certainly be improved in terms of both quality and resolution, for example with the ESA 7th Earth Explorer BIOMASS mission. Alternatively, the standard method applied in practice consists of using widely available Earth observation optical data to assess the annual change in forest cover (Hansen et al, 2013) together with a priori knowledge of the values of forest carbon per unit area in different forests to estimate carbon loss. Crucially, these estimates of forest cover and forest cover change, and the carbon emissions that are effectively taken into account, are highly dependent on the definition of a forest.

In this study, we aim to assess the impacts of the forest definitions adopted by each African country involved in the REDD+ and/or UN-REDD programmes on national carbon emission estimations. To do so, we estimate the proportion of national carbon stocks and tree cover loss that are found in areas considered to be non-forest areas. These non-forest areas are defined with respect to a threshold on the percentage of tree cover adopted by each country. Using the percent tree cover and tree cover loss maps from Hansen et al. (2013) and the biomass map from Bouvet et al. (2018), derived from remote sensing data and fully adapted to this study, we quantitatively show that in many countries, a large proportion of carbon stocks are found in non-forest areas, where a large amount of tree cover loss can also occur. These new results can be used as guidance for adjusting national forest definitions in order to improve the REDD+ MRV efficiency, and for choosing relevant national forest definitions for countries that did not select any forest definition yet. We further found that under the REDD+ framework, some partner countries have proposed activities related to only reducing deforestation, even when a large proportion of their carbon stocks are stored outside forests. This situation may represent a limitation of the efficiency of the REDD+ mechanism, and could be avoided if these countries choose a lower tree cover threshold for their definition of forests and/or if they are engaged in other activities.

Hansen et al., 2013. High-resolution global maps of 21st-century forest cover change. Science, 342(6160), 850-853.

Bouvet et al., 2018. An above-ground biomass map of African savannahs and woodlands at 25m resolution derived from ALOS PALSAR. Remote Sensing of Environment, 206, 156-173.

Developing countries wishing to implement REDD+ need to develop a national forest reference emission level as part of the process. Reference levels are expressed as tonnes of CO2 equivalent per year for a reference period. Reference level serves as a benchmark for assessing a country performance in implementing REDD+ activities. According to IPCC guidance, emission level is calculated based on the multiplication of activity data which the total area affected by changes during the reference period considered and emission factor which is expressed in tonnes of CO2 equivalent per unit area. Several countries in the Congo basin have expressed their commitment in the REDD+ process and are in the process of submitting their forest reference emission levels. However, the level of deforestation currently observed in those countries tend to be relatively low compared to other regions. This makes it particularly challenging to define suitable methods to estimate activity data with a low level of uncertainty. Design-based activity data estimation often rely on a probability sample of reference data that are obtained from the visual interpretation of satellite imagery. Several estimation methods exist either based on using a forest change map in a stratified estimation or a model assisted approach. Both methods were implemented for the Cameroun and Republic of Congo. Initial results for the Republic of Congo indicate that the two methods lead to equivalent results, but the level of uncertainty still appear relatively high despite following the guidelines suggested in the relevant literature. This is primarily due to the difficulty in appropriately characterising the changed areas in the maps used. This is because deforestation tend to be related to small patches scattered over the entire study areas. Potential improvements are being implemented in relation to (i) improving the quality of the change strata and (ii) combining the stratified with model assisted estimation.

The Forestry Department of the United Nations Food and Agriculture Organization (FAO) has introduced the System for Earth Observation Data Access, Processing and Analysis for Land Monitoring (SEPAL) in 2013 with the support of Norway. SEPAL emerged from concrete and practical needs based on years of experience and expertise in capacity building in developing countries using remote sensing in forestry applications. SEPAL is a cloud-based computing platform for fast access and processing of remotely sensed data sources designed which was built in the first place to assist national forest monitoring and reporting for the Reducing Emissions from Deforestation and Forest Degradation, Forest Conservation, Sustainable Management of Forests and Enhancement of Carbon Stocks in Developing Countries (REDD+) mechanism.
FAO, as a member of the UN-REDD consortium (jointly with UNDP and UN Environment), is responsible for assisting countries in development of robust national forest monitoring systems (NFMS) and operational satellite land monitoring systems (SLMS) for to follow the implementation of national REDD+ policies and measures, and to monitor, measure, verify and report (MRV) the results of these efforts. SEPAL has succesfully supported countries to establish and maintain a NFMS SLMS capable of producing the information required to make consequential decisions about forest management; decisions that promote sustainable forest management and can potentially mitigate the effects of global climate change. Specifically, SEPAL addresses the existing challenges countries face when developing forest monitoring systems due to difficulties accessing and processing remotely sensed data; a key source of information for monitoring forest area and change over large, often remote forest areas. After the Paris agreement, the use of SEPAL in-country is moreover expanded to also monitoring the implementation of the NDCs.
SEPAL functionality depends on four main factors, (i) a powerful, cloud-based computing platform for big data processing and storage, (ii) a direct link to earth observation data repositories, (iii) an open-source set of software tools capable of efficient data processing and (iv) project funding allocated to the purchase of high resolution datasets for participating countries. The computing platform enables national/local technicians and the national governments to process data quickly without locally maintained high performance computing infrastructures. The direct link to data repositories (e.g. Landsat, Sentinels) allows fast access to satellite imagery and other earth observation data for processing. The software tools, initially based on the FAO’s Open Foris Geospatial Toolkit (www.openforis.org) perform powerful image processing, are completely customizable and function similarly ‘on the cloud’ or on the desktop. The Bfast algorithm of WUR (NL) was deployed on SEPAL in order to help advance countries in their monitoring of actual changes and degradation. Some practical examples will be shown (e.g. Ethiopia).
SEPAL has been deployed initially in selected UN-REDD participating countries. Longer term, plans for SEPAL include deployment in all countries participating in UN-REDD, FCPF or the REDD+ mechanism as well as countries with NDCs. The project is led and run by a Project Team in the Forestry Department of FAO, with the assistance of an Advisory Group with representatives from key user and provider institutions such as the Global Forest Observations Initiative (GFOI), World Bank Forest Carbon Partnership Facility, space agencies and user countries. The SEPAL system is actually being deployed as a country system in Ecuador. Other applications of the SEPAL cloud has been started, namely in the context of SAFE (Safe Access to Fuelwood and Energy), as well as other monitoring options (e.g. near real-time monitoring). Finally, given there is interest to test the platform also for marine applications, this presentation will touch upon more possible applications and the way forward towards tackling climate change.