You are on page 1of 53




This publication was produced for review by the United States Agency for
International Development. It was prepared by Tetra Tech.

This publication was produced for review by the United States Agency for International Development by Tetra Tech ARD,
through a Task Order under the Prosperity, Livelihoods, and Conserving Ecosystems (PLACE) Indefinite Quantity
Contract Core Task Order (USAID Contract No. EPP-I-00-06-00008-00, Order Number AID-OAA-TO-11-00022).

Tetra Tech
159 Bank Street, Suite 300
Burlington, Vermont 05401 USA
Telephone: (802) 658-3890
Fax: (802) 658-4247

Tetra Tech Contacts:
Ian Deshmukh, Senior Technical Advisor/Manager

Forest Carbon, Markets and Communities (FCMC) Program
1611 North Kent Street
Suite 805
Arlington, Virginia 22209 USA
Telephone: (703) 592-6388
Fax: (866) 795-6462

Stephen Kelleher, Chief of Party

Olaf Zerbock, USAID Contracting Officer’s Representative

• Jennifer Hewson, Conservation International
• Marc Steininger, FCMC & Conservation International
• Stelios Pesmajoglou, Greenhouse Gas Management Institute

Contributing authors:
• Angel Parra, Consultant; GHG inventory & LULUCF sector expert
• Gordon Smith, Greenhouse Gas Management Institute
• David Shoch, TerraCarbon, LLC
• John Musinsky, National Ecological Observatory Network
• Fred Stolle, World Resources Institute
• Kemen Austin, World Resources Institute
• Irene Angeletti, Greenhouse Gas Management Institute

The US Agency for International Development (USAID) has launched the Forest Carbon, Markets and Communities
(FCMC) Program to provide its missions, partner governments, local and international stakeholders with assistance in
developing and implementing REDD+ initiatives. FCMC services include analysis, evaluation, tools and guidance for
program design support; training materials; and meeting and workshop development and facilitation that support US
Government contributions to international REDD+ architecture.

Please cite this report as:
Hewson, J., M.K. Steininger and S. Pesmajoglou, eds. 2014. REDD+ Measurement, Reporting and Verification (MRV) Manual,
Version 2.0. USAID-supported Forest Carbon, Markets and Communities Program. Washington, DC, USA.





The author’s views expressed in this publication do not necessarily reflect the
views of the United States Agency for International Development or the United
States Government.


TABLE OF CONTENTS....................................................................................... III
ACRONYMS AND ABBREVIATIONS ...................................................................... V
ACKNOWLEDGEMENTS ................................................................................... VIII
1.0 INTRODUCTION....................................................................................... 1
1.1 PURPOSE, SCOPE AND STRUCTURE ...................................................................................... 1
1.2 BACKGROUND ............................................................................................................................... 4
1.3 REFERENCES...................................................................................................................................... 8
2.0 INSTITUTIONAL ARRANGEMENTS.............................................................. 9
2.1 INTRODUCTION ............................................................................................................................ 9
2.2 ELEMENTS OF A MRV SYSTEM FOR REDD+ .......................................................................13
2.4 STEPS IN ESTABLISHING INSTITUTIONAL ARRANGEMENTS.....................................19
2.5 EXAMPLES ........................................................................................................................................22
2.6 EPA NATIONAL SYSTEM TEMPLATES ...................................................................................26
2.7 REFERENCES....................................................................................................................................30
3.1 INTRODUCTION ..........................................................................................................................31
3.2 IPCC GUIDANCE...........................................................................................................................34
3.3 INVENTORY AND REPORTING STEPS .................................................................................38
3.4 DEFINITIONS OF CARBON POOLS AND LAND USES ..................................................41
3.6 REFERENCES....................................................................................................................................49
4.0 FIELD-BASED INVENTORIES .................................................................... 50
4.1 INTRODUCTION ..........................................................................................................................50
4.2 CARBON POOLS AND THEIR MEASUREMENT ................................................................54
4.3 CONCEPTS AND CONSIDERATIONS IN INVENTORY DESIGN ................................ 57
4.4 THE FOREST CARBON INVENTORY TEAM .......................................................................66
4.5 FIELD WORK AND ANALYSIS .................................................................................................67
4.6 CALCULATING CARBON STOCKS FROM FIELD DATA ...............................................69


4.7 DATA CHECKING ........................................................................................................................75
4.8 CONSOLIDATING INVENTORY DATASETS ......................................................................76
4.9 THE GAIN-LOSS METHOD ........................................................................................................78
4.10 REFERENCES .................................................................................................................................79
4.11 SELECTED RESOURCES ............................................................................................................81
5.0 REMOTE SENSING OF LAND COVER CHANGE ........................................... 82
5.1 INTRODUCTION ..........................................................................................................................82
5.2 LAND USES AND CATEGORIES IN THE UNFCCC ..........................................................84
5.3 OVERALL STEPS AND NEEDS...................................................................................................90
5.4 REMOTE SENSING OVERVIEW ................................................................................................93
5.5 EMERGING AREAS OF RESEARCH....................................................................................... 109
5.6 REFERENCES................................................................................................................................. 114
5.8 SELECTED RESOURCES............................................................................................................ 121
6.1 INTRODUCTION ....................................................................................................................... 125
6.2 REPORTING.................................................................................................................................. 127
6.3 VERIFICATION ............................................................................................................................ 141
6.4 REFERENCES................................................................................................................................. 151
7.0 THEMATIC REVIEWS ............................................................................. 152
7.1 HISTORY OF REDD+ UNDER THE UNFCCC .................................................................. 152
7.2 COMMUNITY-BASED MONITORING ................................................................................ 165
7.3 NEAR-REAL TIME MONITORING AND ALERT SYSTEMS ........................................... 181


Markets and Communities Program FCPF Forest Carbon Partnership Facility REDD+ MEASUREMENT. VERSION 2.0 v . ACRONYMS AND ABBREVIATIONS ACR American Carbon Registry AD Activity Data AFOLU Agriculture. Forestry and Other Land Use AGB Aboveground biomass BCEFs Biomass conversion and expansion factors BRDF Bi-directional reflectance distribution function BURs Biennial Update Reports CH4 Methane CI Conservation International CMP Conference of the Parties serving as the Meeting of the Parties to the Kyoto Protocol CO2 Carbon dioxide COP Conference of the Parties CV Coefficient of Variation DBH Diameter at Breast Height DEM Digital Elevation Model DTs Decision Trees EFDB Emissions Factor Database EFs Emissions Factors EM Electromagnetic EOS Earth Observation System EPA Environmental Protection Agency FAO Food and Agriculture Organization FAS Fire Alert System FCMC Forest Carbon. REPORTING AND VERIFICATION (MRV) MANUAL.

FIRMS Fire Information and Resource Management System FREL Forest Reference Emission Level FRL Forest Reference Level FSI Forest Survey of India FUNCATE Foundation of Space Science. Meteorology and Environmental Studies ILUA Integrated Land Use Assessment INPE Brazilian National Space Research Institute IPCC Intergovernmental Panel on Climate Change KCA Key Category Analysis LDCM Landsat Data Continuity Mission LEDS Low Emission Development Strategies LiDAR Light Detection and Ranging LUC Land-use Change MADS Colombian Ministry for Sustainable Development MCT Brazilian Ministry of Science. Reporting and Verification N20 Nitrogen oxide NAMA Nationally Appropriate Mitigation Strategies NASA National Aeronautics and Space Agency REDD+ MEASUREMENT. Technology and Innovation MMU Minimum-mapping unit MRV Measurement. REPORTING AND VERIFICATION (MRV) MANUAL.0 vi . VERSION 2. Land-use Change and Forestry GPS Global Positioning System IDEAM Colombian Institute for Hydrology. Applications and Technology GEF Global Environmental Facility GFIMS Global Fire Information Management System GFOI MGD Global Forest Observation Initiative Methods and Guidance Documentation GFW Global Forest Watch GHG Greenhouse gas GHGMI Greenhouse Gas Management Institute GIS Geographic Information System GLAS Geoscience Laser Altimeter System GOFC-GOLD Global Observation of Forest and Land Cover Dynamics GPG-LULUCF Good Practice Guidance for Land Use.

and consistency TOA Top-of-atmosphere UMD University of Maryland UNDP United Nations Development Programme UNEP United Nations Environment Programme UNFCCC United Nations Framework Convention on Climate Change USAID United States Agency for International Development USGS United States Geological Survey VCS Verified Carbon Standard WGs Working Groups WMO World Meteorological Organization WRI World Resources Institute REDD+ MEASUREMENT.0 vii . completeness. VERSION 2. plus the role of conservation. comparability. REPORTING AND VERIFICATION (MRV) MANUAL. sustainable forest management and enhancement of forest carbon stocks.NCs National Communications NFMS National Forest Monitoring System NGGIP National Greenhouse Gas Inventories Program NGO Non-governmental organization NNs Neural Networks NRT Near-real Time PCA Principal components analysis PRODES Projeto De Estimativa De Desflorestamento da Amazoni (Brazilian Amazon deforestation monitoring program) QA/QC Quality Assurance and Quality Control QUICC Quarterly Indicator of Cover Change RADAR Radio Detection and Ranging REDD+ Reducing emissions from deforestation and forest degradation. SBSTA Subsidiary Body on Scientific and Technical Advice SES Social and Environmental Soundness SINA Colombian National Environmental System SLR Side Looking RADAR SRTM Shuttle Radar Topography Mission TACCC IPCC principles of transparency. accuracy.

and Charles Scott of the US Forest Service. The authors also thank those who have made graphics available for use in this Manual. Axel Penndorf of BlackBridge. Karyn Tabor. Chris Potter of the US National Aeronautics and Space Administration’s Ames Research Center. SilvaCarbon Consultant. Mario Chacon and Johnson Cerda of Conservation International. Maggie Roth. REPORTING AND VERIFICATION (MRV) MANUAL. VERSION 2. Andrew Lister. Megan McGroddy. Rishi Das. Michael Gillenwater of the Greenhouse Gas Management Institute. including Ned Horning of the American Museum of Natural History. REDD+ MEASUREMENT. Deborah Lawrence of the University of Virginia. and members of the USAID Climate Change office. Christine Dragisic of the US State Department. and Leif Kindberg of FCMC. Asim Banskota of the University of Minnesota. Frank Martin Seifert of the European Space Agency's Centre for Earth Observation ESRIN. the International Panel on Climate Change and the Landsat program.0 viii . Additional figures are from the websites of the United Nations Framework Convention on Climate Change. and Jamie Eaton of TerraCarbon LLC. Ronald McRoberts. Brice Mora of the Global Observation of Forest Cover and Land Dynamics. including Colin Silver.ACKNOWLEDGEMENTS The authors thank the various colleagues in our organizations who have commented on the text. Carly Green and Jim Penman of the Global Forest Observations Initiative. John Rogan of Boston University.

1 INTRODUCTION This chapter focuses on the application of remote sensing-based approaches to forest cover and change monitoring. of this Manual outlines the sequence of steps required for generating a national greenhouse gas (GHG) inventory.0 REMOTE SENSING OF LAND COVER CHANGE Authors: Marc Steininger and Jennifer H ewson 5.3. Section 3. Inventory and Reporting Steps.0 – REMOTE SENSING OF LAND COVER CHANGE 82 . 5. This chapter is relevant to the activities highlighted on the following page. REDD+ MRV MANUAL: CHAPTER 5.

if using the gain-loss method. STEP 1: Estimate the land areas in each land-use category.0 – REMOTE SENSING OF LAND COVER CHANGE 83 . STEP 7: Perform verification and implement quality control checks. STEP 2: Conduct key category analysis (KCA) for the relevant categories. assess significant non-CO2 gases and carbon pools and prioritize such pools in terms of methodological choice. pool or non-CO2 gas. ensuring that the requirements in terms of emission and removal factors are met. REDD+ MRV MANUAL: CHAPTER 5. STEP 5: Quantify emissions and removals. through stratification and other methods. for representing areas in the GPG-LULUCF. appropriate to the tier level identified. for the time period required.STEP 0: Establish Institutional Arrangements. STEP 3: Design a forest carbon inventory to generate Emissions Factors (EFs). AD represents the extent over which a human activity occurs. EFs represent coefficients that quantify the emissions/removals per unit area. Within the categories designated as key. STEP 6: Report emissions and removals estimates. and worksheets where appropriate. Document and archive information used to produce the national emissions and removals estimates following specific instructions under each land-use category. STEP 4: Generate Activity Data (AD). carbon pool and non-CO2 source. including expert peer review of the emission estimates following specific guidance under each land-use category. using the reporting tables. Emissions and removals estimates represent the product of the AD by the associated EFs. estimating the uncertainty in each estimate.

where estimates of changes between uses is a specific input to emissions estimates. see Chapter 2). However. Wetlands. The information in this chapter summarizes remote sensing issues for a NFMS. it provides limited information on specific approaches to remote sensing of land use. 5. • Overall steps and needs for consideration in developing a satellite-based forest monitoring system. Stratification should seek to identify forest types with potentially significantly different levels of biomass to assist field sampling strategies (see Chapter 4). 2012). Finally. Optical satellite remote sensing is the most heavily used type of remote sensing for this application. 4. Land-Use Change and Forestry (GPG-LULUCF) is a key resource for countries.2 LAND USES AND CATEGORIES IN THE UNFCCC LULUCF.In the context of the activities highlighted above.8. particularly in the stratification of forests for field sampling. this chapter discusses: • The context of land uses within the United Nations Framework Convention on Climate Change (UNFCCC). while the Guidance from 2006 uses Stock-Difference (IPCC. and thus in this Manual we use the former term throughout. refers to land-use change or persistence among the six broad uses defined by the IPCC: Forest Land. the Global Forest Observations Initiative (GFOI) Methods and Guidance Document (MGD) (GFOI. even though the 2006 Guidelines are more up-to-date and use the latter. and spectral variations due to very different canopy structures. Settlements. and the Forestry and Forest Products Research Institute’s REDD-plus CookBook (Hirata et al. Vol. Links to additional resources for training on remote sensing are provided in Section 5. Additional forest strata could be of interest for national management and planning purposes and such stratification activities can be facilitated through the use of remotely-sensed data. 2003). especially satellite-based approaches. or persistence within. these six broad uses are called Categories. 2006). Remote sensing. the remote sensing techniques and issues discussed in this chapter are also relevant to application in the Stock-Change method. However.0 – REMOTE SENSING OF LAND COVER CHANGE 84 . REDD+ MRV MANUAL: CHAPTER 5. 23 The GPG-LULUCF (IPCC. the IPCC recognizes two methods to estimate carbon emissions: the stock change method 23 and the gain-loss method (IPCC. This chapter emphasizes optical satellite remote sensing of deforestation. This chapter discusses remote sensing of land-use change in the context of the gain-loss method. as noted in Chapter 3. For land use. other valuable resources include the Global Observation of Forest and Land Cover Dynamics (GOFC-GOLD) Sourcebook (GOFC-GOLD. 2013). 2003) uses the term Stock-Change. As mentioned in Chapter 1. Another important use of remote sensing in a National Forest Monitoring System (NFMS) is to produce a forest benchmark map. This is needed to define the national forest area at the beginning of a reporting period and within which carbon stocks and forest changes will be monitored. There has been no decision for non-Annex regarding use of the 2006 Guidelines. and deforestation represents the largest source of GHG) emissions from the land-use sector in most tropical- forest countries. Possible types of land-use change among.. 2013). • A brief review of remote sensing for forest monitoring. remotely- sensed data represent a key input to a stratification of forest types and can be used to characterize the seasonality of leaf cover. Sub- categories can be defined within a category to more precisely define changes and emission sources. and • An overview of emerging areas of remote sensing-based research for forest monitoring. Grassland. within the context of the United Nations Framework Convention on Climate Change (UNFCCC). Cropland. inundation. The Intergovernmental Panel on Climate Change (IPCC) Good Practice Guidance for Land Use. and Other Land (IPCC 2006. provides the most practical option for monitoring land cover change over large areas.

thus enabling Approach 2. mountainous areas). are data on the area of a Category that potentially results in GHG emissions or removals.1.. documenting the area of each land use over time and with full coverage. Multiplying the AD. It is important to consider the characteristics of land-use parameters that will be monitored and the cost implications of a full-coverage mapping versus a sampling-based method.1 Definition of national forest and other classes The 2006 IPCC Guidelines for National Greenhouse Gas Inventories consolidated both LULUCF and Agriculture into the Agriculture. These approaches. unless noted otherwise. Throughout this chapter. it is acceptable to use a mix of the three approaches among regions or categories in a country. provides the estimated associated GHG emissions for each category. Land use. The costs associated with these approaches could be significant and thus necessitate a sampling-based approach. While countries must report on land use. by the EFs. Forest definition A fundamental step in the development of a MRV system is the national definition of forest.g. listed in Box 5.0 – REMOTE SENSING OF LAND COVER CHANGE 85 .g. existed for the entire region. EFs are coefficients that quantify the emissions/removals per unit area. 5. yet on a non-mapped manner Approach 3: Tracking of land-use conversion on a mapped manner Approach 3 is most informative and applicable to a mechanism for reducing emissions from deforestation and forest degradation. over a given period of time.1 of the GPG-LULUCF. As illustrated in Chapter 3 (Figure 3.2). sustainable forest management and enhancement of forest carbon stocks in developing countries (REDD+). or regions (e. can usually be inferred based on local context and a general knowledge of the area. plus the role of conservation. Countries have some flexibility in developing their forest definition. may be more effectively monitored through airborne or field-based data collection approaches.. thus providing the data on differences in the carbon stocks before and after the cover change has occurred. existing data available for the Argentine Pampas were sufficient for either Approach 1 or 2. AD are combined with Emissions Factors (EFs). the extent over which a human activity occurs. are used to estimate AD for each Category: Box 5. However. the coefficients quantifying the emissions/removals per unit area. In Australia. as it is based on relationships between observed spectra in the images and the structural characteristics of the soil and vegetation covering land.2.Activity data (AD). definitions have been adapted from the 2006 IPCC Guidelines and the GPG-LULUCF 2003. Data existed on land-cover change. Agricultural census data. thus enabling Approach 1. not to be confused with tiers. yet they are constrained by certain criteria. forest degradation). documenting transformations between natural grasslands to pasture and cropland. Forestry.1: Approaches Approach 1: Net area of each land use reported at different time period. See Chapter 3). The definition REDD+ MRV MANUAL: CHAPTER 5. and Other Land Use (AFOLU) in Volume 4 (IPCC 2006. the creation of a multi-temporal map of change in forest cover as well as some sub-categories enabled Approach 3 for those categories. satellite monitoring is more suited to detecting land cover. For example. The IPCC (2006) describes three overall approaches. While satellite-based remote sensing is a valuable tool for monitoring several parameters of land use. or the extent over which a human activity occurs. however. for the representation of land use (see Chapter 3). yet no tracking of specific conversions among them Approach 2: Tracking of land-use conversions. Case studies of countries that have used different approaches are provided in Annex 2A. some types of land-use categories (e.

and use as much field information.05 ha to 1 ha.. selective logging may occur in a site defined as forest. as this defines the geographical extent for conducting the other two activities. However. there are carbon stock losses in the ‘forest remaining forest’ class. they are part of a non-forest use. since a country would not be required to estimate rates of the appearance of new fallows and their re-clearance when reporting changes in forest area 24. forest degradation could be defined as “a direct human-induced long-term loss (persisting for X years or more) of at least Y percent of forest carbon stocks [and forest values] since time T and not qualifying as deforestation or an elected activity under Article 3. and Minimum patch size of 0. these areas have urban and agricultural uses (i. if agricultural fallows 24 Examples of national forest definitions are available at http://cdm. The forest benchmark map should be created first. stratification and change estimation The development of forest mapping for REDD+ activities can be considered as a trio of components: creation of a forest benchmark map. While in terms of structure these are young. as much of their agricultural land is in some stage of fallow.’ and this subclass should be sampled to estimate carbon stock change. Therefore. If the site is still under a forest use ( then the site remains forest. Box 5.0 – REMOTE SENSING OF LAND COVER CHANGE 86 .html REDD+ MRV MANUAL: CHAPTER 5. and estimation of changes within the map extent. it would still be defined as forest despite the structural change.2. thus. as tree crown cover is defined by the periphery of the crown. A site is defined as forest if it meets the above criteria and if its main use is assumed to be forest-related.unfccc. The GFOI MGD provides a further discussion of forest degradation in Chapter 2.2: Definition of forest criteria Potential to reach a minimum canopy height at maturity of 2m to 5m. For example. Considering agricultural fallows as non-forest greatly facilitates reporting on deforestation and associated GHG emissions. Forest benchmark. in addition to the change in physical structure.2.must be developed based on both the physical structure of the present and potential vegetation as well as how the land is used.4 of the Kyoto Protocol” (IPCC 2003). It is also preferable to conduct these three activities one at a time. attempting to do all in a single process can lead to slow and overly-complicated processes.. another site that has been logged and did cross below this threshold could be classified as deforestation. The physical criteria for forest.2. This will likely necessitate a subclass for ‘intact to degraded forest. If the tree cover was not reduced enough to pass below the threshold of the forest definition. the use of the land must also change. and are expected to be re-cleared after that period. Agricultural fallow is a particularly important example for many tropical countries. while an urban park or agricultural fallow may meet the physical criteria of forest. stratification of forests within the map extent. The development of the forest benchmark map should reflect the national forest definition. In this case. forest concession subjected to some selective-logging cycle).e. and the range countries can select for their definition. but has undergone degradation.e. For example. they belong to a non-forest category. Tree-crown cover is not the same as leaf cover. very high resolution imagery and expert knowledge as possible to facilitate remote sensing imagery interpretation. Minimum tree-crown cover of 10 percent to 30 percent. regrowing “forests. Conversely. According to the IPCC Report Definitions and Methodological Options to Inventory Emissions from Direct Human- induced Degradation of Forests and Devegetation of Other Vegetation Types.” they are part of an agricultural cycle with a defined temporal period. For example. aerial photos. while leaf cover is the proportion of leaf versus canopy-gaps. non-forest uses). are included below in Box 5.

A conservative approach is to avoid using images from before the start date.0 – REMOTE SENSING OF LAND COVER CHANGE 87 . such as open woodland.2. Many approaches exist for estimating change within forests. such as monthly image composites from MODIS. as discussed in Chapter 4. Different potential strata can be assessed by merging with any existing data on carbon stocks. although it is recommended as a method to reduce both field-survey costs and carbon-stock uncertainties. The benchmark map should be based on satellite data from dates closest to the start date of a monitoring period. any areas defined as unmanaged forest (see Chapter 3) can be omitted in order to provide a final definition of where REDD+ can be implemented and where monitoring must be conducted. and those strata with little difference in stocks may be combined. as this may result in the inclusion of forests that were cleared just before the start date. strata can be created based on expert opinions of forest types that should have different biomass levels. Seasonal information could be obtained from coarser resolution data.. such as those of inundated forests. A decision should also be made on the minimum size of clearance to be estimated. aerial photos or very high resolution imagery could be obtained over sample areas to allow analysts to train themselves in the interpretation of fallow versus mature forest in different parts of a country. etc.g. satellite sensor problems. and Step 5. of this Manual. and thus it defines the extent within which REDD+ crediting may be possible and where monitoring is required.2 of the GFOI MGD (GFOI 2013). Stratification is not required.are excluded from the national definition. liana forests. smaller than a few hectares. such as sub-national governments and REDD+ project implementers. Therefore.2. such as montane shrubs. and elevation. It could also be the start date of a historical analysis for use in estimating a Forest Reference Emissions Level or Forest Reference Level (FREL/FRL.2 of the GOFC-GOLD SourceBook (GOFC-GOLD 2013) provide useful overview information on the range of approaches and methodologies available. Stratification of forests within the benchmark area should be conducted in coordination with the field-based inventory team (See Chapter 4) and with the guidance of an expert statistician. see Chapter 2).e. Step 5. can REDD+ MRV MANUAL: CHAPTER 5. In the latter case. of this Manual.5. See. A key consideration is the type of change to estimate. and palm forests. Stratification based on satellite imagery can benefit from the inclusion of seasonal information. soils. This could be the date of initiation of REDD+ activities. if acquired for particular leaf-on and leaf-off months.2 of the GFOI MGD (GFOI 2013). Satellite images for an exact date can be difficult to acquire due to cloud cover or gaps in data collection. A forest benchmark map should have as few data gaps as possible. Section 3. This is a critical step for analysts to correctly train and assess classifications of forest for the benchmark period. The creation of a forest benchmark map will require a careful and iterative process. as this is the main emissions source for many countries. such as data on climate.4. At a minimum. or from very high resolution imagery. Stratification can also make use of ancillary GIS data. Other characteristics often discernable from satellite data are major differences in canopy structures. Other characteristics that analysts should pay special attention to include gradations to lower-stature vegetation that fall below a threshold in the national forest definition. gaps can occur as a result of cloud cover. analysts will need to identify an appropriate methodology for analyzing multiple images. it will eventually need to be updated to define the extent of forest at the beginning of the implementation of REDD+ activities. e. The creation of a forest benchmark map facilitates the estimation of change within forests.4. Once the forest benchmark map is finalized. If such field data do not exist..2 of the GOFC-GOLD SourceBook (GOFC-GOLD 2013). forest clear-cutting (removal of all trees) should be estimated. where more deciduous forests may have significantly lower biomass levels. This may necessitate the acquisition of multiple images for the same area and. Section 5. Section 5. Even the creation of a few broad classes is very useful for these purposes. or Section 3. image dates used for the benchmark map may vary. i. as this sets an important precedent for where REDD+ activities may or may not be implemented. Generally. resulting in fewer total strata. or open-canopy vegetation. for example. and these can be re-combined later if field data indicate otherwise. thus. While the inclusion of the smallest clearances. It is important that a final benchmark map is communicated to stakeholders with any concerns addressed. it will necessitate several iterations of image analysis followed by reviews from experts and stakeholders.5.

and then use a different approach. Issues related to sub-classes are somewhat different for forest versus non-forest. deforestation can be attributed to different forest sub-classes. For example. Again.g. cropland. In contrast. Conversely. could lead to greater classification error. plantations and fallows may be difficult to distinguish.potentially yield a more correct and higher rate. Spectral distinction of these uses is often difficult. the time and resources involved may be costly. 2000). and many different opinions within the research community. broad deforestation class that occurs anywhere within this benchmark. while not requiring a very difficult process of spectral classification of all transitions among sub-classes. Tables 5. For forests.1 presents an example of a land-cover change matrix with three broad categories (forest. This should be assessed as part of both the national Key Category Analysis (KCA) discussed in Chapter 3 and the forest stratification process discussed in Chapter 4. especially when one cannot be very selective about the season of the imagery used for analysis because of frequent cloud cover. managed grassland. This can be performed one time. such as wildfire. A step-wise process may be worth exploring that uses different methods and levels of detail for different monitoring aspects. A second step could be to produce a map of a single. countries should have strong justification for including sub-classes since this will require methodologies that consume more resources. a country could first produce an estimate of forest clear-cutting using one approach. Table 5. Combining approaches like these can provide all of the necessary estimates to complete a full land-cover change matrix. depending on the season and stage of crop development at the acquisition times of available images. This is an area where there are many options. strong justification in terms of improvement of emissions estimates is needed to justify attempts to include transitions among these classes. may be considered. nonforest). as it can be assumed there are no transitions among these naturally- occurring vegetation types over the required reporting periods. as well as any important transitions among those.0 – REMOTE SENSING OF LAND COVER CHANGE 88 . The expected increase in the accuracy and overall usefulness of emissions estimates should be clearly assessed. For example. Other forms of degradation. 1999. and may not reflect a significant source of emissions. a country will conduct a stratification for a national forest inventory. Table 5. For example. transitions among different post- deforestation land uses do occur over short time periods. Degradation from selective logging is more difficult to detect and often only detectable for more intensive forms and with data obtained soon after the logging event. A third step could be to use samples of airborne or other very-high resolution data to estimate the proportions of sub-classes of non-forest following deforestation.2 presents an example with greater thematic precision where the broad categories have been divided into sub-classes. with forest sub-classes that have significantly different carbon stocks. By combining the two.1 and 5. Foody. degraded forest. Including sub-classes may provide data that are very useful for REDD+ national strategies and management policies. not all of these must be estimated in the same process. Again. and countries should consider the trade-off that usually exists between thematic precision of a land-use change study and the accuracy of the change estimates (e.. Other classes It may be important to further stratify the six broad use classes into sub-classes where carbon stocks vary significantly. following the creation of the forest benchmark map. for degradation. Mather. an initial step could be to produce a forest benchmark map. such as sampling of higher-resolution data or even targeted field surveys.2 illustrate this below. However. REDD+ MRV MANUAL: CHAPTER 5.

Values inside the matrix are areas and percent change for each category of persistence or change.970 25 2.000 Non. “Non-forest” includes all non-forest.9 101. gross deforestation plus forest degradation is 0. Values in (a) are in absolute units.1: Example of a land-use change matrix with few land-use classes and change categories. T2 a) Forest Degraded Non-forest Sum T1 b) Forest Degraded Non.045 % T2 99.000 Forest 99. and in (b) are percentages. 100. T1 and T2 are the first and second time periods. In this example.3 100 T1 T1 Forest Forest Non.3 98. both naturally-occurring and anthropogenic.000 Degraded 0.2 in the first row of (b)). referred to in the IPCC as “Initial land-use class” and “Land use during reporting year.4 0. % Forest Forest forest T1 Forest 9. such as hectares.4 and 0. 4.000 4.2 100 Degraded 5 1. a) T2 Degraded Degraded Lowland Montane Natural Lowland Montane Fallow Cropland Pasture Sum T1 Forest Forest Grassland Forest Forest Lowland Forest 7945 35 3 5 7 7995 Montane Forest 1995 5 2 3 2005 Degraded Lowland Forest 5 1500 2 6 12 1525 Degraded Montane Forest 470 1 4 475 T1 Natural Grassland 993 3 4 1000 Fallow 350 50 150 550 Cropland 200 700 100 1000 Pasture 50 1400 1450 Sum T2 7950 1995 1535 475 993 608 771 1673 REDD+ MRV MANUAL: CHAPTER 5. “Forest” in this table is non-degraded forest only.5 Table 5.7 98.5 1.0 100 forest forest Sum T2 9.0 – REMOTE SENSING OF LAND COVER CHANGE 89 .945 2.” Values in Sum T1 and Sum T2 are total area and percent change for each class.940 40 20 10.6 percent (adding values 0.4 0.010 4.

2: Example of a land-use change matrix with more precise land-use classes and change categories.4 0.0 0.4 0. for example.1 100 Degraded Lowland Forest 0.6 100 % T2 99.” Values in Sum T2 are total area and percent change for each class.2 0.6 9. the large areas of change from cropland to fallow (200) or pasture (100).5 percent reduction in fallow indicates intensification of land use. approach should be applied to monitoring different land cover classes.1 100 Montane Forest 99.4 96. non-degraded forest.1 0.7 99.5 0. Fallow. it is important to consider how consistency will be achieved and what methods will be used to effectively combine data from different time periods.3 87.3 100 Cropland 20.8 99. the majority of forest occurs in the lowlands.1 0. In this example. A high degree of rotational land use is also indicated by. The appropriateness of different monitoring methodologies will need to be assessed.4 0.5 80.8 135. Natural Grassland.3 OVERALL STEPS AND NEEDS Figure 5.2 99. b) T2 Degraded Degraded Lowland Montane Natural Lowland Montane Fallow Cropland Pasture % T1 Forest Forest Grassland Forest Forest Lowland Forest 99.4 0.3 98. referred to in the IPCC as “Initial land-use class” and “Land use during reporting year.0 – REMOTE SENSING OF LAND COVER CHANGE 90 . Cropland.1 percent increase in pasture indicates an increasing importance of this use.0 10. according to the national forest definition. “Forest” here means intact.8 100 Degraded Montane Forest 98. Values in (a) are in absolute units.1 illustrates the key decisions a country should consider when developing an effective and efficient satellite-based forest monitoring system after identifying which categories and sub-categories to monitor and the scale of monitoring. and pasture) and forest degradation also occurs in the lowlands.5 98.3 0. croplands.1 27. either via a shortening of fallow cycles or an increase in permanent pasture.1 Table 5. The 35. REDD+ MRV MANUAL: CHAPTER 5. the majority of deforestation (to fallow. Where automation is not possible. or sampling-based. classification algorithms. Criteria include the type and resolution of satellite data and the degree to which a full coverage.2 0. 5. The 12.1 0.9 0. Values inside the matrix are areas and percent change for each type of category. such as hectares and in (b) are in percent. level of automation and analyst expertise. T1 and T2 are the first and second time periods.3 0.0 70. pre-processing.4 100 Fallow 63.0 100 Pasture 3. Each of these decisions is discussed below. and Pasture represent non-forest classes.8 100 T1 Natural Grassland 99. including the types and availability of different satellite data.

whereas other dynamics. Finally. it is necessary to consider the appropriate scale and approach.1: Key considerations in the development of a NFMS 1) What categories are most important to monitor? A KCA involves identifying the major land-use-based sources of GHG emissions. 2) What are the appropriate scales and/or sampling approaches for monitoring? Once the categories and classes to monitor have been assessed. particularly some forms of degradation and post-deforestation land-use changes. the process should extend to defining the types of land-cover changes that are major GHG contributors and to aligning these definitions with the land-cover change categories defined by the GPG-LULUCF. For example. some land-use dynamics may be very appropriate for satellite-based monitoring. A KCA should be performed as part of the development of a REDD+ strategy within the national development-planning context. For example. or are they much larger? Different types of changes may also be most appropriately monitored with different sources of data. a country must determine the geographical extent of managed land. may require airborne or field-based monitoring. and thus where monitoring should be conducted (see Chapter 2).Figure 5. These latter land-use dynamics may require more costly data -collection REDD+ MRV MANUAL: CHAPTER 5.0 – REMOTE SENSING OF LAND COVER CHANGE 91 . For MRV. do change events occur in small patches of several hectares.

estimates of change can be very sensitive to the MMU. Some of the main questions to consider within the methodological options are: 1) What types of satellite data are most appropriate for monitoring the classes identified? 2) What type of classification approach should be used? 3) What types of pre-processing are needed for the particular method of image analysis being considered. a country should seek to produce the most accurate estimates possible for key categories while using an approach that is “replicable. how can consistency and reliability be assured? 5) How should data from different time periods be combined to produce change estimates? 6) What post-classification processing steps should be applied? 7) What validation approach should be used. some vegetation types. Question 7 should consider REDD+ MRV MANUAL: CHAPTER 5. most fallows younger than 10 years are distinguishable. This will avoid either: (i) the use of methods with little justification based on in-country testing. although it is only vaguely defined in the context of satellite monitoring. may require data from particular or multiple seasons within each year. However. therefore. including data sources and sampling? Some of the most important considerations are whether to use optical data versus Radio Detection and Ranging (RADAR) data. either during creation of the forest benchmark map or during the stratification of forests in this map. and for those parts that are dependent on analyst interaction.processes and thus necessitate a sampling approach. or a case should be made that using a larger MMU does not significantly affect resulting area estimates. While older secondary forest can be difficult to distinguish from mature forest. and what level of analyst expertise is required? 4) How much of the process can be automated. Question 5 includes both the approach to processing the satellite imagery from multiple dates and the approach to estimating change rates from completed land-use change maps or sample estimates. a country should seek its own cadre of experts with strong fundamental backgrounds in remote sensing to fully access and understand the relevant literature and options. such as deciduous woodland. a country can minimize confusion between mature forest clearing and fallow cycles in later monitoring.0 – REMOTE SENSING OF LAND COVER CHANGE 92 . and conduct assessments with national data. and the great majority of fallow periods are shorter than 10 years. The MMU should be smaller than the minimum patch size included in the national forest definition. By distinguishing mature forest versus fallow areas. and often some type of filtering to a defined minimum-mapping unit (MMU). Question 6 includes merging of temporary sub-classes. For question 4. Many differing views exist regarding optimal methods for monitoring land-use change and. A country should obtain opinions from a range of international experts. what spatial resolution is needed and whether the data source has an appropriate archive and acquisition strategy.” This is a fundamental requirement of the GPG-LULUCF. again necessitating increased data and analysis demands. focusing on the categories identified in the KCA. Further. possibly combining information from multiple dates into a single multi-date product. or (ii) spending too much time investigating issues that can be well-informed by existing literature or are not significant to the potential accuracy of the final emissions estimates. while using a larger MMU may not significantly affect area estimates for static areas of classes. 3) What methodological aspects should be considered? A country should consider a range of methodological options.

the feature itself is the source for thermal-infrared regions (Figure 5. as well as an independent team of analysts to interpret validation data and conduct error calculations.3). or entire countries. the amount representing the strength of the signal. The human eye. The sun is the source for visible and shortwave-infrared spectral regions of the earth. Passive remotely-sensed data are acquired by a sensor that passively receives energy originating from another source. by combining images of energy measured in different spectral bands and assigning a separate color for display. The amount and type of energy sensed is usually recorded in digital form. this is the visible portion of the EM spectrum. though not always valid. 5. 5. Different spectral regions are represented by relatively narrow “spectral bands” (Figure 5. Satellite-based remote sensing is most common because of the full.4. rather than being in direct contact with it. the instrument does not emit its own signal. Passive remote sensing The majority of remotely-sensed data used for monitoring land use is passive. “multi-spectral” images are produced. and the type representing the recording of the signal across a spectrum. as such systems could be applied over large regions.1 Types and characteristics of remote sensing data Remote sensing is the process of sensing energy emitted or reflected at some wavelength along the electromagnetic (EM) spectrum by an object. are that different land-cover types can be distinguished based on this recorded information and that land use can be inferred from land cover.various sampling schemes and the availability of very high-resolution satellite or aerial observations. thus enabling national monitoring for terrestrial-based applications. depending on the type of equipment.0 – REMOTE SENSING OF LAND COVER CHANGE 93 . REDD+ MRV MANUAL: CHAPTER 5.5. Numerous text books are also available on remote sensing of land-cover. At the highest level. repeated coverage offered by one or more satellite data sources. sampling approach. The portion of the sun’s energy across these spectral regions reflected by the land surface is often indicative of the structural and chemical characteristics of the surface features (Figure 5. and resources available. Links to selected internet resources are provided in Appendix 5B. for example.4 REMOTE SENSING OVERVIEW This section provides a summary of remote sensing fundamentals. senses a relatively small portion of the total spectrum of energy emitted by the sun. two broad types of remote sensing for monitoring land-cover exist: passive and active.4) and. as illustrated in Figure 5.2). Airborne remote sensing capacities are also of interest. Fundamental assumptions.

3: Generalized spectral curves of fundamental features in remote sensing of land-cover.4: Example of spectral resolution. the bands cover a wide range of energy. and in the example here they are divided by white lines. passes again through the atmosphere and reaches a sensor on board a satellite. The signal detected is dependent not only on the reflectance properties of the surface but also on the sun angle. such as illustrated in (b). and a sensor with such bands would be considered a broad-band sensor. A) B) Figure 5. Spectral bands are defined by a range of wavelengths.Figure 5. Both (a) and (b) represent the entire visible range of the electromagnetic spectrum. and a sensor with channels along these bands would have a high spectral resolution. passes through the atmosphere. Shortwave energy is emitted by the sun.0 – REMOTE SENSING OF LAND COVER CHANGE 94 . REDD+ MRV MANUAL: CHAPTER 5. topography. reflects off a surface. In (a). plus non-green vegetation and shadows caused by the geometry of terrain and vegetation. view angle and atmospheric properties. Most types of land-cover are a mixture of these features. A single channel of a multi-spectral sensor is sensitive to energy only within a certain band.2: Optical satellite remote sensing. A sensor. with so many channels and bands would be considered hyper-spectral. Figure 5. In (b). the bands are narrow.

Brightness levels. 2012). and data from other satellite sources. active fires. RADAR is further discussed in Section 5.. for example.Figure 5. In addition to the visible and near and shortwave-infrared regions. Such color-composite images aid visualization and interpretation of the land-cover. Until recently. versus optical sensors which measure reflected energy that is largely a function of canopy architecture. as well as modeling various ecosystem processes and vegetation-climate interactions. it facilitates the detection of clouds. and soil background.5. RADAR data are also sensitive to canopy and soil moisture. RADAR data provide information related to the density of leaves in the canopy.. Several recent satellites carry RADAR sensors that collect data in multiple bands and in different polarizations. and are extremely influenced by topography. 2011. The resulting images did not have the dimensionality that multi-spectral images have. In forest environments. One major advantage of RADAR systems is their ability to penetrate clouds due to the longer wavelengths. and thus yielded limited potential for classification of land-cover types. RADAR information is primarily related to structural features at the wavelength scale of the energy being sensed. RADAR data. the microwave portion of the EM spectrum. RADAR and Light Detection and Ranging (LiDAR) are the most commonly used active remote sensing techniques for terrestrial applications. an instrument sends out a signal at certain wavelengths and measures the return time and strength of the back-scattered signal. while rarely used for distinguishing types of land-cover.5: Image data combined from three sensor channels to produce a multi-spectral image. and urban heat islands. or branches and tree trunks. Because of their sensitivity to the geometric properties of forests. RADAR data have potential for relating to forest biomass. Active remote sensing In active remote sensing. all RADAR sensors on board satellites collected measurements in only one wavelength band and one polarization. shown as grey tones. Baccini et al. passive remote sensing systems also acquire data in the thermal region. Thermal energy is emitted by the land surface itself and. represent values in the individual channels. leaf pigments.0 – REMOTE SENSING OF LAND COVER CHANGE 95 . depending on the wavelength used. were used to produce two recent maps of global forest biomass (Saatchi et al. thus extending their utility for classification of land-cover types. REDD+ MRV MANUAL: CHAPTER 5.

Applications of LiDAR in forestry have mainly focused on measurement of canopy height. . The distance between the sensor and the target is then calculated from the elapsed time for the LiDAR signal to make a complete round trip.6). and the horizontal and vertical distribution of vegetation. Lim et al.As with RADAR. enabling the creation of images.0 – REMOTE SENSING OF LAND COVER CHANGE 96 . LiDAR instruments emit a pulse of energy. Both forms of LiDAR have been shown to be useful for estimating forest biomass via comparison with field data and modeling.. 2002. 2004.. some of which is scattered back to the sensor by the target. sub-canopy topography. LiDAR systems are also generally classified into full-waveform LiDAR and discrete LiDAR systems (see Figure 5. 2009).. usually between one and five. Lefsky et al.6: Example of full waveform vs. Clark et al. per transmitted pulse. However. 2003. others have scanning abilities to collect data both along and across sampling lines.. LiDAR operates in the visible and near infra-red portions of the EM spectrum and. 2003) REDD+ MRV MANUAL: CHAPTER 5. Nelson et al. thus. Full Discrete Waveform Return LiDAR LiDAR Figure 5. Full-waveform systems record the entire waveform of a returning pulse. discrete waveform LiDAR (Lim et al. in contrast to RADAR. While some LiDAR instruments collect data only along sampling lines. does not penetrate clouds. while discrete systems sample a discrete number of points. for example. these parameters can be used to model estimates of aboveground biomass (see.

For example. some countries are investigating the utility of other sources of.7 illustrates the archive history of the Landsat satellite series. including many of the high resolution sensors such as RapidEye. KOMPSTAT-2 or RESOURCESAT-1 are pointable.The majority of LiDAR remote sensing to-date has been airborne-based.3. at least half the size of the scale of changes). has implemented a national-level. Conversely. Quickbird. The Landsat series is the most common data source for monitoring land-cover change. To facilitate consistent monitoring and ease of logistics. Guyana. While this can result in the same area being repeatedly imaged at much higher frequency. REDD+ MRV MANUAL: CHAPTER 5. has proven particularly useful for assessing and addressing the impacts of forest degradation. Box 5. annual monitoring activity using wall-to-wall RapidEye imagery. persistent cloud cover often reduces the frequency of acquiring useable images. Spatial resolution is important as the resolution must be fine enough to detect the changes of interest (i. has traditionally only been used over small areas because of cost and frequency of availability. It uses RapidEye imagery from the dry and wet season to better differentiate between seasonal biomass coverage. Cloud. especially in areas where degradation pervades. especially via sampling. as such data will not detect smaller-scale changes. cost of procuring the imagery and image archive length. However. have defined orbits that dictate how frequently the satellite will return to view the same location on the earth and acquire a new image. Data used for land-cover monitoring have resolutions ranging from 5m to 30m. the Geoscience Laser Altimeter System (GLAS) on board the Ice. The re-visit time for Landsat is 16 days.e. Data archive length is another important consideration for developing historical analyses. it is preferable to work with a single source of data throughout a study period when possible.. meaning they can be tilted to view a location that is at an angle to their defined orbit. for example. RapidEye. higher resolution imagery for monitoring purposes. Figure 5. GeoEye-1 & -2. is becoming more practical. Temporal resolution refers to the frequency with which data are collected. However. data collection strategy. less than 10m. SPOT HRV series. Resolution and other considerations In addition to the type and spectral characteristics of different images. such acquisitions occur for short periods and some require tasking. so the most that an area could be monitored is every 16 days. Data with coarser spatial resolutions are used for global studies and are not generally suitable for land-cover monitoring. with a ground resolution of 70m.0 – REMOTE SENSING OF LAND COVER CHANGE 97 . IKONOS. the use of these data over large areas. as data extend back to 1972 for the multi-spectral scanner (MSS) and 1982 for Thematic Mapper. WorldView-2. For Guyana.7m to 1km. CBERS HRC. The linear samples from GLAS were inputs to the two global biomass studies noted above. the DMC constellation. Many satellites. Examples of the use of high resolution imagery in NFMSs While many countries are generating forest benchmark maps and deforestation monitoring products using Landsat imagery. for example. Publically available data have spatial resolutions ranging from 0. high resolution data. The use of such data for land-cover monitoring yields an inherent bias in such derived estimates.5m. Mexico utilizes Landsat in its MRV system and has also incorporated RapidEye. even for periods as brief as the past decade. as high resolution data continue to become available and more affordable. a ground area of one hectare would be represented by 11 pixels at this resolution. represents a ground area of 900 m2 (30m by 30m). While this has historically limited the practicality of using such satellites in a monitoring capacity. and land Elevation Satellite (ICESAT) was satellite-based. GLAS provided full- form LiDAR information for linear tracks along the satellite path. one LiDAR instrument. However. Other satellites. such as Landsat. A 30m resolution observation. or pixel. Box 5.3 highlights two examples where very high-resolution imagery is being used extensively in the development of NFMSs. consideration must be given to: spatial and temporal resolution. with five MSI bands of 6. Though the ICESAT satellite is no longer operational.

4). as well as cost policies. the Sentinel data products available to all users. where data are collected.usgs. for example.0 – REMOTE SENSING OF LAND COVER CHANGE 98 . with five satellites in its current constellation. with 13 bands located in the VIS. are being constructed. From http://landsat. The Sentinel-2 series of multi-spectral imaging satellites. phenological. key data characteristics to consider are: • What geographical.6. will provide additional data options for consideration in the development of NFMS. The Landsat Data Continuity Mission (LDCM). ensures that future National Aeronautic and Space Agency (NASA) satellites will continue to provide a long-term data record. including the successful launch of Landsat 8 in February 2013.Further. roughly the same area is acquired each time the satellite returns to view the same location on the earth. Landsat 1 – 3 carried only the MSS instrument. will be launched between May-July 2015 and will offer a revisit time of 10 days. Figure 5. relative to the scale of the land- cover changes to monitor? REDD+ MRV MANUAL: CHAPTER 5. the example of Sentinel-2 The European Space Agency (ESA) is responsible for the space component of the Copernicus programme. Box 5. thus providing another source of data for an extended period.7: Landsat archive timeline. The upcoming launch of the Sentinel-2 series.4: Additional future data sources for consideration in the development of NFMSs. 20m and 60m.php Finally. is listed in Section 5. carrys the OLI and TIRS instruments. Landsat 7 carries the ETM+ instrument but since 2003 has experienced data gaps due to mechanical failure. NIR and SWIR wavelengths and spatial resolutions of 10m. and changes among them? • What is the spatial resolution of the data and how appropriate is it. with 2A scheduled for launch in 2015. has a follow-on mission planned for 2019-2020. Landsat 4 – 5 carried both the MSS & TM instruments. as the Landsat satellites have a defined under which a series of dedicated satellites. and how do these relate to the potential for distinguishing the land-cover types of interest. Landsat’s no-cost data policy allows flexibility in data use. will provide an additional source of imagery for consideration in the development of NFMS (see Box 5. a satellite program’s future sensor launch and data acquisition strategy. the Sentinels. are important considerations when planning a monitoring program. a 5-day revisit time will be possible. When the second unit (Sentinel-2B) is launched in 2016. A table of current and future satellite data options. meaning image pairs from multiple dates mostly overlap. and bands within them. Based on the free. The first satellite. The interoperability between these sensors and the Landsat missions will further enhance the revisit time and improve overall data availability. Sentinel-2A. including characteristics. In summary. RapidEye. Costs of other data sources are also trending downwards. and atmospheric (especially persistent cloud cover) conditions exist? • What are the spectral regions. full and open data policy adopted for the Copernicus programme. Landsat 8.

yielding a dataset with accuracies within 30m and. Atmospheric correction The atmosphere has several effects on visible and infrared energy as it passes through the atmosphere from the sun to the land and back to a satellite or airborne sensor (Figure 5. and image analysis is the process of generating a land-cover class for all parts of an image. and post-processing steps below is based on optical data and approaches to classification. Post-processing occurs after the image analysis step.4. Finally. the calculation of change rates and error estimates are required. Atmospheric correction is frequently performed in combination with a bi-directional reflectance distribution function (BRDF) correction. post-processing activities may include a number of steps.php REDD+ MRV MANUAL: CHAPTER 5. Co-registration among images should be reviewed and may require adjustments. analysis.0 – REMOTE SENSING OF LAND COVER CHANGE 99 . Automation is increasingly available for processing numerous images. Data and enables the estimation of rates and patterns of land-cover change to be generated. • What is the temporal resolution in terms of potential frequency of acquisition of non-cloudy observations compared to the desired frequency of monitoring? • What is the longevity of the image archive length – does this meet the historical mapping needs? • What are the cost implications of these data in terms of purchase and analysis? • What are the future satellite development and launch commitments? 5. The United States Geological Survey (USGS) has been reprocessing much of the Landsat archive. is optional depending on the image analysis approach. This is to understand the geographical area represented and is applied when importing the image into a GIS or image- analysis format for processing. As previously outlined.2 Image pre-processing. BRDF defines how light is reflected from a surface. Image pre-processing Geometric registration and co-registration Geometric registration is the process of mapping data in a geographical coordinate system. The summary of pre-processing. resulting in the creation of a L1T precision and terrain corrected product 25. Therefore. Therefore. These data have already been geometrically corrected using precision ground control points and Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) information. and post-processing Image pre-processing refers to any step that is applied to an image in preparation for the image analysis step. Atmospheric correction may be necessary depending on the image analysis approach that will be used. geometric registration may have errors up to 100s of meters. it does not mean that images of the same area acquired from different dates will overlay well enough to avoid errors in change estimates resulting from poor co-registration. but traditional analyst- driven methods are also sufficient. However. although images have been geometrically registered. co- registration may still be necessary. 25 http://landsat. atmospheric correction. and occasional data transformation. analysis. Pre-processing usually includes geometric registration and co-registration.usgs. thus. using examples of Landsat data analysis.8). Co-registration is a standard. eliminating the need for further geometric correction. simple process that takes a modest amount of time and involves the identification of one image to use as the base image to which the remaining images will be co-registered. and is dependent on both the incident and reflected directions. though useful.

is the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) tool (Masek et al. While the corrections are applied to all the images and the resulting composite should therefore appear seamless. LOTRAN and 6S are the most common. and these methods are highly dependent on careful atmospheric correction. LEDAPS uses information on water vapor. These usually assume constant atmospheric conditions across an image. as opposed to the digital numbers of the raw image data. local histogram matching could be applied to further reduce artifacts. cloud. The electromagnetic radiation source is the sun. 26 http://landsat. such as the sun. Several programs exist to perform atmospheric corrections over entire images. including those from the LDCM. and use a single correction algorithm for the entire image. BRDF characterizes how an object illuminated by a source. Some semi-automated methods apply constant signatures over multiple images or image dates. These corrections can be applied to several partially-cloudy images of the same area... appears brighter or darker depending on the angle of the source and the angle at which it is viewed by a satellite sensor. Sun and sensor view angles also impact the effects of the atmosphere. beginning in summer 2013. additional algorithms based on. Most atmospheric correction algorithms are applied to satellite images prior to mapping. and several tools have been created to facilitate their application. and this radiation can be blocked or scattered by clouds in addition to being affected by a “clear” atmosphere. “gap-filled” composite mosaic.Atmospherically-corrected images contain data representing surface reflectance.9).8: Atmospheric effects on optical data. will be available with top-of-atmosphere (TOA) reflectance corrections applied 26. 2008). 2001). in unitless values from zero to one. and requiring each atmospherically-corrected image to be classified separately and then combined. for example.0 – REMOTE SENSING OF LAND COVER CHANGE 100 .php#3b REDD+ MRV MANUAL: CHAPTER 5. appearing as darker or brighter patches (Figure 5. Conversely. atmospheric artifacts may remain. methods that apply constant class signatures over images with variable atmospheric conditions should include atmospheric correction. A “clear” atmosphere still causes scattering and absorption of the radiation as it is transmitted from the sun to the earth and back to the satellite. although there is active research on accounting for variability within an image. a topography-dependent Rayleigh scattering correction and an aerosol optical thickness component based on Kaufman et al. developed by NASA. LEDAPS also generates One example. atmospheric pressure. The images can then be combined to produce a single.usgs. cloud shadow and snow masks. all Landsat data. Figure 5. Alternatively. (1997) to generate a surface reflectance value for each pixel. Approaches to classification that involve the creation of sub-classes for each type of land use and change can yield accurate maps without atmospheric correction because sub-classes can account for different atmospheric conditions. Performing atmospheric correction depends on the image analysis approach used (Song et al. In addition. ozone.

note the orange-tone artifacts that appear to the left of the remaining cloud gaps in the upper-left of the image. REDD+ MRV MANUAL: CHAPTER 5. the artifacts in E) are no longer visible. green and blue colors. but with no histogram matching between them applied. E) a mosaic of two atmospherically corrected images. green and blue bands are assigned to the red.A) B) C) D) E) F) Figure 5. D) the same. allowing visual exploration of the infrared data.9: Example of atmospheric correction of Landsat data from San Martin. but with a Gaussian stretch applied to the data histogram. A) an unstretched “true-color composite” where the red. C) the near- infrared. a common assignment for a “false-color composite”. middle-infrared and red bands assigned to the red. thus requiring additional images.but with histogram matching applied.0 – REMOTE SENSING OF LAND COVER CHANGE 101 . F) mosaic of the same two images . some of the cloud gaps remain since they were also cloud in the second image. Note that while the linear gaps have been filled. The black lines are SLC-off data gaps. B) the same image. Peru. but post-atmospheric correction and with a cloud/cloud-shadow mask applied using LEDAPS. green and blue colors in the display.

Some of the common algorithms in remote sensing software packages are. output classifications are evaluated and. at a minimum. In vegetated lands. field analysis. 20 percent soil and 50 percent shadow. Classification Land-cover classification produces a thematic representation of land by categorizing pixels based on their spectral signatures. Depending on the classification approach used. and woody vegetation or litter. training data are modified and a new iteration of the classification is run.10 illustrates the supervised classification approach. estimating the proportions of each component for each pixel. for example. Often.0 – REMOTE SENSING OF LAND COVER CHANGE 102 . based on these. and this may suggest a need to merge or add additional sub-classes. Principal components analysis (PCA) is one example of a transformation technique involving ordination. An output could estimate. These can be visualized as “fractional images” and used as inputs to classifications. Figure 5. Spectral mixture analysis (SMA) is another data transformation approach. these are termed “image end members. Statistics of the pixel data within these areas are calculated and. unidentified pixels. SMA involves defining the spectral reflectance of each main feature representing the landscape under study and. When applied in SMA. SMA does not add to the information content.” However. Minimum Distance. such as various forms of ordination. in increasing complexity: Parallelepiped. based on conspicuous errors. Two broad types of classification exist: supervised and unsupervised. and Mahalanobis Distance. for example. Maximum Likelihood. Definition of the pure features may be via laboratory analysis. that a particular pixel represents an observation of a piece of land that is 30 percent sunlit leaf.Data transformation Some analysis methods include data transformation techniques. The “Tassled Cap” transformation has been used extensively in the classification of vegetation types. these include the means. these techniques may help to produce more accurate or efficient land-use classification. These techniques alter the information to facilitate interpretation. The level of statistical separability among the classes can be evaluated. variances and co-variance matrix of the spectral data. the analyst identifies “training sites” and creates training data by delineating areas known to be of each class. Soil theoretically could be split into multiple soil types with different reflectance properties. and water could be included. soil. as it explains the data in terms of physical features. as this is an important feature of most spectral images because of the geometry of the vegetated canopies. prior to classification. Based on these statistics. the resulting SMA analyses are relevant only to that image. Like PCA. various algorithms can be used to estimate the most likely class of the remaining. Shadow is also generally included.” since they are located at the outer ends of the multi-dimensional distribution of the spectral data. and other types of data ordination. these are sunlit leaf. if image end members are used. SMA utilizes estimated spectral reflectances of a set of “pure” features that aim to represent the observed surfaces. yielding a classified image. REDD+ MRV MANUAL: CHAPTER 5. or literature. In supervised classification. End members can also be defined by simply selecting the extreme pixels in the multi- dimensional data. SMA can be a useful approach in understanding the spectral data contained in the image data. they are referred to as “spectral end members. defining the “spectral signature” of each class.

The final tree is often composed of hundreds of splits and terminal nodes representing the land-cover classes contained in the training data. 2003. Rogan et al. Spectral images (A) from two dates can be combined and viewed to more easily observe spectral changes (B) indicative of land-use change. 1999. Training sites can be drawn (C) based on field and aerial data as well as an analyst’s knowledge of an area and expertise in interpretation. from Liberia. Pal and Mather. Both DTs and NNs have become favored over REDD+ MRV MANUAL: CHAPTER 5. Instead... and available ancillary field or other data. Numerous studies have used DTs to generate robust classification results in many regions (Friedl et al. DTs operate by iteratively seeking a binary split in the data in each of the bands. An assumed advantage of supervised over unsupervised classification is that the analyst directs the process based on a priori knowledge of the area being classified. though NNs can run more slowly than DTs. algorithms identify spectrally similar pixels and then assign them to a user-specified number of groups.. “Boosting” and “pruning” processes can be applied to DTs in order to improve the efficiency and reduce the number of final splits.0 – REMOTE SENSING OF LAND COVER CHANGE 103 .. 2008b). Many variants of NNs exist. Additional iterations are typically run to further split groups that overlap different land-cover types. The final product (D) is often filtered to eliminate small and spurious errors. the resulting DT is a set of rules that is applied to the rest of the pixels to produce a classified image. Two such algorithms are Decision Trees (DTs) and Neural Networks (NNs). In unsupervised classification. an assumed advantage of the unsupervised approach is the algorithm evaluates the distribution of the data itself.A) B) C) D) Figure 5. 2003. The ISODATA algorithm is common in most software packages. no training process is applied. the location of the pixels. The split is one that optimizes accuracy at that stage in the development of the DT. NNs attempt to mimic the human learning process to associate a class with the image data. Hansen et al. The output of an unsupervised classification is then reviewed by an analyst. based on the data in the training sites identified by the analyst. Supervised algorithms that explore the distribution of the data while still allowing the analyst to direct the process via training are increasingly used. 2000. Conversely. These are the basis for class statistics used to classify the rest of the image.10: Example of a supervised classification of two dates of images in a single process. Hansen et al. and each group is labeled to a class based on the analyst’s visual interpretation of the spectral data.

Another type could use the average. Longépé et al.. randomly selects some. or the analyst while developing the training data sites. This enables a direct estimation of change and persistence from the multi-temporal imagery. such as Random Forests 27. or further from. a five-by-five pixel window. for example. or to mosaics of images of the same area and time period. the variance of the pixels within a certain window around the center pixel. and generally perform efficiently. such as “linear trend in red reflectance” or “maximum middle-infrared reflectance recorded since the initial date” (e. accuracies for land-use classes such as agriculture and grassland tend to be lower. many studies with analyst interaction have produced accurate estimates of national forest cover. Evans et al. 2010 or Jiang et al. Harper et al. These are powerful because short-lived signals of land-use change are more likely to be captured. Friedl et al. Replicability and analyst interaction versus automation In the case of estimating deforestation. could be particularly useful in areas where the spectral separation of vegetation types is limited. 2008. 2010. 2012). Any of these methods can be expanded to become part of a contextual classification. All of the above approaches are examples of “per-pixel” classification. both of which may contain errors. and segmentation algorithms allow an analyst to specify the relative size and shape of the segments.g. and all available data are employed. Rodriguez-Galiano et al.g.0 – REMOTE SENSING OF LAND COVER CHANGE 104 . for example.. providing additional information that can be utilized by the classification algorithm. Segmentation generally represents an intermediate step prior to classification. One type of contextual classification is textural classification. the central pixel being classified. These approaches may be based on the seasonal signal of different types of vegetation and estimate changes based on where anomalies in these seasonal signals are detected (see.. Most recently.stat.. 2008b). Lindquist et Weighting can also be used to apply different weights to pixels that are closer to.. which may be critical in cloudy areas. another contextual approach.. Image segmentation. This process also includes a single classification step that yields a two-date classification.maximum likelihood and other classification algorithms. but not all.. Random Forest ensemble classification methods have been successfully applied to land-cover and land-cover change classification (Pal. In contextual classification approaches. Other approaches mine all available data. is used to inform the classification. such as the entire Landsat archive. They can also be applied to multi-temporal image data. 2011). 2006. of the pixels within the window. e. in which multiple decision trees are constructed and an output class is selected based on the majority of votes from all the trees. Gislason et al. or some other metric..berkeley. In this approach. These errors would be compounded when the two maps are combined during post-processing. 2008a. which can be a problem with DTs. Some form of “direct change estimation” process is usually recommended for change estimation.. Hansen et al. and these estimates are generally derived from local rather than national studies. meaning the pixel is classified solely based on its spectral characteristics. 27 http://www. of the variables to build the resulting tree and identifies resulting splits based only on this subset of variables. and generate many multi-temporal metrics.. unlike DTs. This approach. generally 70 to 80 percent. i.e.g.. Other approaches under exploration include learning classification techniques. Some recent semi-automated approaches use much more of the data archive than a single image from a start date and another one from an end date. images from the beginning and end of a study period.. 2012). rather than the classification of two individual images and two single-date classification outputs. 2007. rather than the pixel level. is a statistical method that groups contiguous pixels into areas (segments) that are relatively homogeneous. The resulting segmented image can then be classified at the segment level. Such methods do not suffer from overfitting. 2005.. Each of the above approaches can be applied to a single image at a time. a pixel is classified based on its own spectral characteristics as well as those of surrounding pixels. Random Forests.htm REDD+ MRV MANUAL: CHAPTER 5. Accuracies have often been reported over 90 percent (e.

and bright non-forest areas such as urban areas and exposed sand and soil. Difficult areas may require more direct analyst interaction to obtain accurate results. While rapid estimation of cover and change based on these rules could be achieved for much of the country. although in some cases it has also included the classification step. the replicability of the methodology. may only yield accurate results over certain parts of the study area where the cover types are most easily distinguished with the spectral data. and merging can be accomplished by recoding the values of all the sub-classes to a value that represents the final class. (2008b) sampled an existing vegetation map to generate training points. There are a wide range of options to apply automation in the classification step itself. In doing so. A related approach is to use traditional interpretation methods to identify training sites for classes. the recently published Deforestation Atlas of the Democratic Republic of the Congo was produced by an entirely automated approach (Hansen et al. the following step should be applied to the two classifications to create a two-date change map. A large set of training sites could be built for the entire country. in the form of thresholds applied to reflectance data. as the spectral variations in the new data are accounted for each time these new data are combined with the training site locations. or to classes that are spectrally distinct. If the classification methodology included the creation of sub-classes merged into the final desired classes. the validity of the results would be very dependent on precise correction and normalization of the images in the pre-processing step. like areas with modest topography and cloud cover.. and the cost.g. Once it is confirmed that this set can be used to produce an accurate map. there are well-published approaches that use automation for a series of pre-processing steps. the remaining areas or classes would need to be estimated via other approaches. This is often not the case and should therefore be tested.In recent years there has been valuable research on automated methods for processing satellite data. a country could choose to monitor certain parts of the country that are more appropriate for automated monitoring. classification algorithms could be rule-based. The values of the classifications from both dates should be recoded to form the basis for a final class map that records categories of change and persistence. then the actual change estimation is conducted using a set of rules or digital classification assisted by analyst interpretation (Souza et al. This has mostly been in the pre-processing steps. such as dark or clear water. Such rules.. if relatively few rules are used. However. or derived data in some other units. In this case. Masek et al. Another approach could be to automate the estimation of the most conspicuous changes (e. or for various strata within it. Further. With REDD+ MRV MANUAL: CHAPTER 5. While the results are encouraging. and the post-processing steps required will vary depending on the classification approach and attributes desired in the final map product used to calculate areas for categories. Post-processing Post-processing refers to any step conducted after the classification step. must be valid. For example. Conversely.. note that the previous section recommends directly estimating change from multi-temporal images. As the scientific community itself uses a broad range of approaches. for example. the assumption that accurate results can be achieved over large areas. 2008a). this should be done first. as is typically done in a supervised classification approach. Hansen et al. countries should seek an optimal balance between the accuracy of the final estimates. 2008). the same training sites could be applied to new data in later years to calculate new class spectral signatures to be used with the new imagery. 2005. The approach could be automated once a national training data set is defined. For example. If a two-date classification was not conducted. further testing should be conducted in other regions. using few rules. this indicates there is no single best answer and countries should evaluate options themselves. Each class in the digital output file of the classification has an assigned number. Alternatively. Countries could potentially automate all but the final steps of a methodology to estimate change. especially mountainous regions or areas with more deciduous vegetation.0 – REMOTE SENSING OF LAND COVER CHANGE 105 . clear-cutting of forest) while applying a less-automated method to estimate the less conspicuous ones. snow.. Another example of automating the classification step could be to automate the process of collecting data on training sites.

some filtering of the product is usually desirable. patches of cells with the same value are identified. This may warrant another stratification so that percent rates are not extrapolated into very different areas and thus not well-represented by those where data exist. each forest stratum experiencing change between the eight year difference would have an entry. in years (Puyravaud.0 – REMOTE SENSING OF LAND COVER CHANGE 106 . such as a three-by-three window. Filtering is generally performed for two reasons. many studies that report changes over five or ten years use images that are within one or two years of each target date. and rates of change calculated for each area with a given difference. 2003). In this case. therefore. and can be used in sequence. where the center cell is re-assigned to the most common cell value within the window.” In this. if the analysis was not for a single-year change detection. rates can be converted to absolute values by combining with the forest area at the beginning of the reporting period. while another is from 1999 and 2011. As mentioned above. After recoding to the final class values. This not only removes the speckled pattern but also smoothes jagged edges. which may be the great majority of the study area. data on change can be combined with those on the forest strata. In this case. then a correction must be applied. and for each forest stratum. For example. Calculating change rates Several factors must be carefully addressed when calculating change rates. in this case to a ten year period. After extrapolating the percent rates. this is useful because the final product can have a defined MMU that meets a country’s national definition of forest. which may or may not be desired. is representative of the entire study area. this is subtle. A common type is a local majority filter. to eliminate patches smaller than a desired MMU or the minimum patch size in a national class definition. the previous recoding step would not be required. Data gaps. Third. the source images may not be from the exact beginning and end of the time period reported. to eliminate small errors associated with micro-topography and other very local effects that produce a speckled pattern of mis-classified raster cells and. and patches smaller than a user-defined size eliminated. Two broad types of filters are commonly used. This is calculated based on areas at the start and end date: Annual change rate = [ (area t2 / area t1) ˄ (1 / (date t2 – date t1)) ] – 1 where t1 and t2 are the beginning and end of the time period of the study. the total rate should be reported for the entire time period. The second type of filter is a “sieve filter. If reporting in units of percent. must also be addressed. These areas should be defined and recorded as having an eight and twelve- year difference in dates. in order to report change for each forest stratum. “Local filters. not in per-year units. REDD+ MRV MANUAL: CHAPTER 5. errors present in each of the single-date classifications would be compounded in the merged classification output. especially where cloud cover limits the coverage of optical images. For example. but they are common artifacts that should be removed. one part of the analysis may be based on images from 2001 and 2009. the study areas should be divided into areas where the image pairs representing the start and end time have different lengths of time separating them. In the above example of 2000 – 2010. for a 2000 – 2010 study. Some classification methods are more prone to producing these errors than others.” are based on the class values around (within a three-by-three window) a center cell. If reporting in absolute units. if the aim is to report in units of percent per year. the output classifications will already have values representing classes of change and persistence and. and each forest stratum experiencing change between the twelve year difference would have an entry. First. This allows for a temporal extrapolation for each area in each stratum. from cloud cover or other reasons. extrapolation is needed. Further. second. one should consider if the sampled area. For small windows. yielding a two-date classification. Note that we use the term cell instead of pixel when referring to classification outputs rather than spectral images. In this step. First.

This is based on errors of commission. i.6 Forest 94. the producer’s accuracy for degraded forest is: 100 x 1.000 16. in this case the estimate of an area of change over time..0 – REMOTE SENSING OF LAND COVER CHANGE 107 .3.8 Degraded Degraded Forest 92. An accuracy assessment allows errors in the map to be estimated and uncertainty quantified. Foody. 2012.6 percent.3 Table 5.890 / 2. and those off the diagonals are errors of omission and commission. and associated statistics (Congalton.. The error matrix is generated by comparing the classification results with reference data. how often a pixel was incorrectly omitted from the class.040 4. In the example error matrix in Table 5.040 = 92. but many articles highlight the limitations of this statistic (Olofsson et al.e. or confusion matrix. Pontius and Millones.3: Example of an error matrix and resulting overall accuracy. The main elements of an accuracy assessment are the error matrix.2 Forest 98.000 Non-forest 10 60 3. In this hypothetical case.e.1 REDD+ MRV MANUAL: CHAPTER 5. In the example table. Reference Forest Degraded Forest Non-forest Map total Forest 9. The Kappa coefficient can also be calculated. thus providing additional explanation of. i.. Performing an accuracy assessment of a thematic map represents a very important component of the process.3 forest 98.930 4.000 Producer's accuracy (%) User's accuracy (%) Forest 99. Overall accuracy is the portion of the total number of correctly mapped pixels. the results. The producer’s accuracy indicates how often a pixel is correctly assigned to a specific class. .000 = 94. The statistics include overall accuracy. Accuracy assessment Uncertainty is the error in a particular estimate. 1991).880 90 30 10. The user’s accuracy indicates how often a pixel was incorrectly assigned to a given class.000 Land-use map Degraded Forest 70 1. This statistic is based on errors of omission.960 2. and the producer’s and user’s accuracy for each class in the product.5 Non- Non-forest 98. 2011.5 percent. and validity to. and thus accuracy assessment also refers to assessment of error or uncertainty. The user’s accuracy for the same class is: 100 x 1. 2002). the land use totals are the same as in the beginning time in Table 5. the columns contain verified land uses and the rows contain estimated uses from the classification.890 / 2.000 Reference total 9. how often a pixel was incorrectly included in a class. The values along the diagonal are the number of correctly classified pixels. Accuracy equals one minus the error value.890 40 2.

2010). similarly to the application described in Chapter 4. The sampling design could focus on areas of deforestation identified in the map. along with imagery used in the classification itself.. which answers the question of ‘how to calculate accuracy and quantify uncertainty?’. The GFOI MGD also provides two example approaches for performing an accuracy assessment and area estimation. This type of proportional sampling design ensures that adequate sampling occurs in sparser yet critical classes. Olofsson et al. which parts of the GHG inventory are contributing the greatest errors in GHG estimates and should be reviewed as part of the KCA. as this allows one to combine errors with errors in carbon stock for each stratum. it can be combined with the error of the change in stocks per unit area estimated via field inventories. together with the total area of each class identified in the map. which answers the question. and iii) analysis. and MRV programs should work with expert statisticians in developing strategies for validation sampling and combining information on uncertainties in AD with those in EFs. Multiple interpreters can be used. While they differ from the reporting-table format of the IPCC. thus providing very pertinent additional information. Olofsson et al. and a stratification could be used to categorize areas of high change probability and low change probability. Error matrices and accuracy assessments can also be extended to provide confidence interval (CI) information. or AD. is estimated.. Field surveys can be valuable for the classes that are the most difficult to interpret even with very high resolution imagery. (2013) describe a process for creating CI bounds based on area-adjusted error matrices In addition. which answers the question ‘are the maps and reference data in agreement?’.e. 2013). This can be done using Equation 5.7 of the GFOI MGD (GFOI. such as degraded forest.3 demonstrate error estimation in land-use cover for a single date. i. Once the error for a particular area of change.Both the land-use change and error matrices are common formats for reporting land-use change and errors. shown at the end of Chapter 3. Manual interpretation of an image by an analyst is generally considered higher quality than an automatic classification algorithm). An appropriate approach is the use of careful.e. one stratified and one model-based approach. (2013). Accuracy assessment is very important. This will lead to stronger estimation of the overall uncertainties in estimated GHGs. Section 3. These include: i) ensuring the reference data are of a higher quality than the map data (for example. field and aerial surveys which can be particularly cost effective if resources are limited). ii) the response design.. Multi-date accuracy assessments use the information available from two dates. It also allows one to assess where improvements are most needed. understanding that evaluation of the entire map is not possible. They detail three separate steps that should be undertaken to complete an accuracy assessment including: i) the sampling design. use the information available in the matrix.2.2 from the UNFCC “Good Practice Guidance for LULUCF” (2003) for the propogation of errors. such as those described in Olofsson et al. However. It is useful to calculate these statistics for the different strata in a study area. This is especially important as it quantifies the confidence of a particular class. it is important to account for a rare class when developing a validation strategy (Stehman et al. visual interpretation of a combination of very high resolution imagery. the data can easily be transferred. Proportional sampling could then be focused in these strata to ensure each class is adequately represented in the validation analysis. cross-checked. to generate area-adjusted errors based on the proportional area of each class and errors identified in the matrix. Error-adjusted area estimates. The examples in Table 5. ‘what is a suitable subset area to sample?’. (2014) also provide a full review of good practice recommendations for producing transparent and “scientifically rigorous” accuracy estimates and estimates of area based on change between time one and time two. and the consistency of their interpretation can indicate confidence of the validation data set itself. and ii) Combining reference data sources (i. REDD+ MRV MANUAL: CHAPTER 5.0 – REMOTE SENSING OF LAND COVER CHANGE 108 . such as deforestation. provides guidance on considerations for generating reference data and performing an accuracy assessment. countries must estimate errors in land-use change estimates over time.

11 illustrates a detail of a Landsat image compared to a PALSAR satellite image for an area in San Martín. Hirata et al. Ground observations in the Landsat data in A) are partly obscured by clouds. A) B) Figure 5.. optical sensors. Souza and Roberts. such as RADAR. GOFC-GOLD (2012) lists a range of human activities that result in forest degradation including selective logging. While the brightness variations in both are affected by terrain. 1999. Mapping and monitoring of forest degradation remains challenging. 2012. RADAR images are directly comparable over time. with estimates ranging from 20 to 50 percent of total land-use GHG emissions over large regions (see. because the signals received by the sensor are less affected by atmospheric conditions. forest fires (canopy and sub-canopy) and fuelwood collection. GOFC-GOLD (2012) presents a range of human activities that result in forest degradation including selective logging. adding to the complexity of mapping and monitoring forest degradation. Lambin et al. GOFC-GOLD.5 EMERGING AREAS OF RESEARCH Several areas of particularly active research in support of REDD+ activities include: the mapping and monitoring of degradation.. Readers should also refer to the relevant sections of both the GOFC-GOLD SourceBook and the GFOI MGD. while the PALSAR image in B) is cloud free. The use of other sources of remotely-sensed data in mapping and monitoring forest A second area of research in support of REDD+ activities is the use of other sources of remotely-sensed data. forest fires (canopy and sub-canopy) and fuelwood collection. deforestation.1. Stickler et al. providing information on the distribution of aboveground biomass. Figure 5. because RADAR sensors operate in longer wavelengths (generally 1cm to 1m) of the EM spectrum than. this is more noticeable in the latter REDD+ MRV MANUAL: CHAPTER 5. 2005. and the properties of the emitted radiation from active sensors are controlled and well known.. Houghton and Hacker. Mapping and Monitoring of Degradation Forest degradation is a substantial contributor to GHG emissions from land-use change (Nepstad et al. 2005).11: Examples of Landsat and RADAR images. 2003. In addition. Multiple definitions of forest degradation exist. from Peru. the use of other sources of remotely-sensed data. and a country should seek to understand the implications and applicability of different approaches. for example. such as RADAR.2. RADAR signals are also sensitive to the geometric properties of a forest. Peru. and degradation. to map and monitor forest extent and characteristics. 2012).5. 2009. Conversely. These different activities may require different monitoring approaches. they are able to penetrate clouds and are thus useful for monitoring in areas with persistent cloud cover.0 – REMOTE SENSING OF LAND COVER CHANGE 109 . The IPCC’s definition of forest degradation is provided in section 5. 1999. and the increased use of field sampling to facilitate remote sensing product validation. Several characteristics make RADAR an attractive source of information for such applications. First. in monitoring.. for example. Asner et al.

but smooth for longer wavelength RADAR such as P-band RADAR. small objects such as leaves and twigs are considered rough for small wavelength RADAR. and these are strongly affected by their moisture. the electric field of the resulting radiation has a preferred orientation. and polarimetry. including RADARSAT-2. The intensity of the signal that is scattered back to the receiver from this transmitted energy is recorded as the returned signal. Water bodies tend to be relatively smooth. resulting P-band images are important for measuring vegetation biomass and aboveground carbon stocks. For soil types of similar roughness. ENVISAT and ALOS/PALSAR acquire data with all four polarizations (“quad-pol”) or two polarizations (“dual-pol”). • HV: horizontal transmission and vertical reception.12. and the distance is calculated based on the time elapsed for the RADAR signal to make a complete round trip. with most of the energy being reflected away from the RADAR. while trees and other vegetation are rough. The difference in intensity of RADAR returns from two surfaces of equal roughness is an indication of the difference in their dielectric properties. several additional characteristics distinguish RADAR instruments and data from their optical counterparts and are useful to understand. Linear is the most common polarization used in RADAR remote sensing where a radiated electric field is oriented either horizontally (horizontal polarization) or vertically (vertical polarization) with respect to the direction of propagation. As with optical sensors. The next transmitted pulse illuminates the next strip of terrain along the swath. For example. while other satellites. geometric shape and dielectric properties of an object also affect the information received by the RADAR sensor. Synthetic Aperture RADAR (SAR). the surface with a higher moisture content will appear brighter. and a two-dimensional image is created (each pulse defines one line). directed by an antenna. this is described below. as shown in Figure 5. and thus have a bright appearance in a RADAR image.0 – REMOTE SENSING OF LAND COVER CHANGE 110 . such as an X-band (λ= 3 cm) may only penetrate the upper layer of a forest canopy. A sensor that can transmit either horizontally (H) or vertically (V) polarized waves and receive both will result in the following four polarized images: • HH: horizontal transmission and horizontal reception. A key characteristic of RADAR data not found in most optical data is polarization. causing backscatter. RADAR sensors exploit different wavelength bands. Apart from polarization. interferometry. The surface roughness. A shorter wavelength band.5 Polarization refers to the orientation of the electric field with respect to the direction of propagation. The below discussion includes a selection of SAR applications that are relevant to REDD+ activities. For example. the brightness of areas covered by bare soil may vary depending on the roughness and moisture content of the soil. Thus. side-looking observation. In active remote sensing. whereas a P-band (λ= 23 cm) may penetrate leaves and small branches thus providing information about both the big branches and stems of the trees. that illuminate a strip of the earth’s surface (swath). and • VH: vertical transmission and horizontal reception. • VV: vertical transmission and vertical reception. RADAR measures the distance between an object on the ground and the sensor based on the strength of radio waves that are transmitted as pulses of microwave beams.The following provides a brief introduction to a selection of active remote sensing concepts. REDD+ MRV MANUAL: CHAPTER 5. These include phase. Additional Such concepts are key to understanding the basic characteristics of active remote sensing data. these are outlined in Box 5. Surface roughness is a relative term that is dependent on the RADAR wavelength. Some space-borne satellites including RADARSAT-1 and ERS-1/2 have only single polarization (RADARSAT-1 with HH and ERS-1/2 with VV).

while some geometric and radiometric distortions due to terrain relief may persist. such as Synthetic Aperture RADAR (SAR) systems. SAR polarimetry is a relevant RADAR concept. Speckle can be reduced by averaging the backscatter response within a pixel. these processed and derived products are generally more appropriate for use in mapping.. and shadows that require full. If two SAR images have been acquired over the same area from very close antenna positions. Additional characteristics of RADAR Phase describes the relationship of the lead. REDD+ MRV MANUAL: CHAPTER 5. And. Data from such systems are generally processed by data distributors to Single Look Complex level data. Since the antenna positions are precisely known. Finally. unlike many optical sensors which acquire imagery at nadir (i. correction. termed Side Looking RADAR (SLR). The technique is known as SAR Interferometry. containing both amplitude and phase information. 360 degrees represents one complete cycle and. The path difference is geometrically related to the distance between two antennas and the terrain height. a RADAR image contains information about the intensity of the signal and the phase. Another relevant RADAR concept is SAR interferometry. As mentioned in the introduction. SAR speckle causes pixel-to-pixel variation in intensities even over a homogenous area. The different polarization bands may contain unique and additional information about the surface object. a signal that reflects off a tree trunk to the ground surface is likely to show distinctive polarization shifts from signals that return directly off the soil. are able to achieve a relatively high resolution without the use of a large antenna. introduces a range of geometric distortions including foreshortening.e. and is expressed in degrees. For example.5. Surface objects that scatter are vertically oriented and show high backscatter in vertically polarized imagery and low backscatter in horizontally polarized imagery. layover. of an electromagnetic wave with respect to a reference wave of the same wavelength. a wave that is lagging one quarter of a wavelength behind the reference has a phase of 90 degrees. and a range of derived products which are usually geocoded. and radiometrically corrected. or by applying smoothing filters. or partial. Among the distortions that may persist is speckle.5. therefore. or lag. This across-track capability. more parameters can be measured from polarimetric RADAR compared to single-channel RADAR. Some RADAR remote sensing systems. As previously discussed.12: Horizontal and vertical polarizations Box 5. the different path lengths from these positions to the object on the earth’s surface cause the differences in phase. this grainy ‘salt and pepper’ texture degrades the quality of the image and complicates interpretation. the observed phase differences can be used to infer three- dimensional information about the terrain height.0 – REMOTE SENSING OF LAND COVER CHANGE 111 . observing a location directly below the sensor). though this can effectively reduce resolution. Such unique information is important for discriminating different land-cover types. ortho- rectified. Most RADAR sensors are also side-looking instruments.

In addition to polarimetric information. For example. Similarly. several limitations exist and the history of SAR data usage for land-cover classification remains relatively recent. and complex REDD+ MRV MANUAL: CHAPTER 5. Walker et al. the first long wavelength (L-band. (2010) assessed the ability of PALSAR and LANDSAT data to classify and map forest cover in the Xingu River headwaters in southeastern Amazonia. compared to bare soil.5) related to the structure and complexity of the observed objects. measuring and monitoring aboveground biomass (AGB) and scaling-up ground-based AGB measurements are increasing. (2012) used a combination of PALSAR. and RADARSAT-2 has led to further land-cover classification research using SAR imagery.. Rignot et al. These data are sensitive to the geometric properties of the forest and directly related to measurements of AGB. However. Mitchard et al. has yielded improved estimates of AGB with little or no saturation. characterized by small scale deforestation and degradation. with an emphasis on characterizing several clearing practices and forest regeneration characteristics. They also mapped forest patches and fragmentation and found these data helpful in delineating areas with different degrees of forest disturbance. the fusion of spatial and textural information derived from various SAR polarizations has been shown to improve classification results (Borghys et al. 2009). 2006..Applications of SAR Applications on the use of SAR data for forest mapping. One major limitation of SAR data utilization and analysis is the difficulty involved in interpreting RADAR backscatter as compared to optical spectral data (Saatchi et al. (2012) generated biomass maps and changes in carbon stocks with known uncertainties using PALSAR imagery in a region in central Mozambique yielding maps with sufficient accuracy to enable changes in forest carbon stocks of as little as 12 tons per hectare over 3 years. the Brazilian Amazon deforestation monitoring program. while the Landsat TM yielded a more accurate deforestation extent classification.0 – REMOTE SENSING OF LAND COVER CHANGE 112 . The presence of topographic effect and speckle complicates both visual and digital analysis of RADAR images. data from PALSAR. ENVISAT. the combined use of both Landsat and RADAR imagery further improved the mapping accuracy. Mitchard et al. Saatchi et al... respectively. 2006). this sensitivity appears to saturate at biomass levels of around 100 tons ha-1 (Imhoff et al. and ground-based data to map AGB in Gabon’s Lopé National Park. The recent systematic availability of fully polarimetric SAR data from the ALOS-PALSAR. 25 cm) SAR satellite with the capability to collect cross-polarized responses. Recently. 2009). Multiple studies have tested the potential of combined RADAR channels of different frequencies and polarization for deforestation monitoring. They also found a high degree of spatial similarity among maps derived from PALSAR. producing overall accuracies of 92 and 94 percent with PALSAR and Landsat.. with 95 percent confidence to be detected. 2000) and approximately 81 percent of global forests are above this saturation limit (Nelson et al. 2000). While these results highlight the potential for space-borne imaging RADAR for estimating forest area and biomass. Mitchard et al. (1997) compared SIR-C data with Landsat TM in a test site in Rondonia. polarimetric interferometric SAR (PolInSAR) provides interferometric information (see Box 5. In addition. Landsat. (2011) used L-band synthetic aperture RADAR backscatter data from 1996 and JERS-1 and PALSAR data from 2007 to produce biomass maps of a forest–savanna ecotone region in central Cameroon. They found that the RADAR data detected changes in a broad AGB class in forest–savanna transition areas with an accuracy of 95 percent. Substantial improvements in land-cover change classification can be achieved by combining polarimetric and polarimetric interferometric information (Shimoni et al. for the forest versus non-forest classifications. such as trees. Ryan et al. SAR data are also being evaluated for scaling up ground-based AGB and monitoring changes over large scales (Lu. 2007). Brazil. up to 250-300 tons per hectare based on the sensor’s cross-polarized ability to exploit the strong response of three dimensional objects. and existing data from Projeto De Estimativa De Desflorestamento da Amazonia (PRODES).. space-borne LiDAR (ICESAT GLAS). (1997) used C-SIR data to map land-cover types and monitor deforestation in the tropics.

This can be achieved via weighted random sampling. REDD+ MRV MANUAL: CHAPTER 5.0 – REMOTE SENSING OF LAND COVER CHANGE 113 . other unequal probability sampling approaches or a systematic sample approach. as the objective is not to produce a field- based estimate but instead to model predictions. In the latter case. This is of particular interest in the often vast and difficult-to-access landscapes contemplated under most national and jurisdictional REDD+ MRV. for example.. RADAR. these data can be valuable in extrapolating field-based estimates over larger areas. in particular high resolution satellite imagery.2 ha is recommended for use with airborne LiDAR to reduce model errors (Zolkos et al. especially remote areas that are costly to access. the Japanese Space Agency (JAXA) launched ALOS-PALSAR-2 in May 2014. to optimize the design. are no longer collecting. Sampling strategies differ however. such as the range of reflectance values. Uncertainty related to the long-term data continuity of space-borne RADAR systems could also prove a limiting factor for forest monitoring. The same field measurement and QA/QC procedures described in Chapter 4 are also relevant in the collection of ground truth data. ideally the sample should be well-distributed across the area to which the model will be applied. 2011). Note that these preferred sample designs require that the remote sensing data already be in hand. SAR data generally yield less robust results than Landsat data for forest/non-forest classification in most studies. However. and after model construction and application. from remotely-sensed data. Therefore. Although. While these approaches hold promise. ground-reference data used to build prediction models for remotely-sensed data do not need to be strictly uniformly sampled. and various kinds of LiDAR hold promise to more efficiently map biomass stocks. Finally. The sample should achieve a roughly even distribution across the range of conditions of the remotely-sensed data. Field measurement plots can calibrate and validate prediction models. PALSAR and ENVISAT. This approach to assess model uncertainty is recommended because it is straightforward and captures the overall result of many potential sources of error that would otherwise have to be independently quantified and then propagated. used to assess the accuracy of the predicted values at different scales. 2013). and classification methods based on polarimetric decomposition are being developed. Field measurements should also be collected as close as possible to the acquisition date of the remote sensing image. they are still in the process of refinement.. reducing the impact of GPS inaccuracy (Mascaro et al. providing an equivalent range of spectral values within the area sampled via remote sensing. and this instrument again contains an L-band. Larger plots facilitate the alignment of ground samples and remote sensing imagery. fixed-area plots are best suited for generating ground-reference data to compare with remotely-sensed data. over a wider spatial extent. Large. Circular plot designs and a minimum plot size of 0. a sample of plots are measured and held in reserve. However. which both provided fully polarimetric L-band data. and any RADAR or LiDAR-based estimates of biomass will need to be calibrated using field-plot data. where significant relationships can be calibrated. These data likely will not provide more accurate biomass data for plot sites than those derived from field data. though the range of advanced SAR processing techniques capitalizing on the availability of multi- polarimetric information is evolving. Use of Field Sampling To Facilitate Remote Sensing Product Validation Remote sensing data.areas with a greater abundance of secondary forests may yield significantly lower accuracies. Additional evaluations are also needed to assess the utility of newer RADAR data sources in mountainous areas.

2006.6 REFERENCES Asner. Y. G. R.C. Walker.L. A method for integrating MODIS and Landsat data for systematic monitoring of forest cover and change in the Congo Basin. Lindquist. Status of land-cover classification accuracy assessment.5. Justice and A.. Clark. Foody. REDD+ MRV MANUAL: CHAPTER 5. Supervised feature-based classification of multi-channel SAR images.A. The Netherlands). G. A. M. R. 2006.L.P. A. C. Sohlberg. Goetz. Forests and Climate Change Working Paper 4.G.G. Friedl.A. Sulla-Menashe. S. Remote Sensing Environment 112: 2495–251. Remote Sensing Environment 80: 185- 201. S. Transaction on Geosciences and Earth Observation 37(2): 969-977. Small-footprint lidar estimation of sub-canopy elevation and tree height in a tropical rain forest landscape. Using ALOS/PALSAR and RADARSAT-2 to Map Land-cover and Seasonal Inundation in the Brazilian Pantanal. A. Switzerland. Evans. T. gains and losses of carbon stocks in forests remaining forests. M.H. 2013. GOFC-GOLD Report version COP19-2 (GOFC-GOLD Land Cover Project Office. Beck. E.S DeFries. N..gofcgold.. Yvinec. MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets. Samanta and R. W. T.P.C. A Review of Assessing the Accuracy of Classifications of Remotely Sensed Data. M. Telmer and T. D.R. Strahler.E. and A. 2000. GOFC-GOLD. Friedl. J. Remote Sensing of Environment 114: 168-182. A sourcebook of methods and procedures for monitoring and reporting anthropogenic greenhouse gas emissions and removals associated with deforestation. Benediktsson and J. Nature Climate Change 2: 182-185. 2005. Houghton. Philips. C.A. Silva.. Roy. Maximizing Land Cover Classification Accuracies Produced by Decision Trees at Continental to Global Scales.php. Wageningen University. Altstatt. 2006. and forestation. 2002. Science 310: 480–482.. Borghys. IEEE Journal of Selected Topics in applied Earth Observation and Remote Sensing 3: 560-570. Remote Sensing of Environment 37: 35-46. Choosing a Forest Definition for the Clean Development Mechanism. Keller and J.fao. Congalton. D. D. Broadbent. Sibley. Selective logging in the Brazilian Amazon. Schneider. Gislason.J.. Knapp. P. Huang. 1999. International Journal of Remote Sensing 21(6): 1331-1364. 1991. Friedl.E. Sun. Brodley. and Clark. E. M. Ramankutty. J. Laporte. D.S. Tan. 2010. D. Sulla-Menashe. S. Costa. J..A. Pattern Recognition Letters 27: 294-300.O. Sveinsson. M. R. J. P. Baccini. N.M.. Roberts. Pizurica and W. 2008a. Hansen M.B. Remote Sensing of Environment 91(1): 68-89. FAO. Silva. Global land cover classification at 1 km spatial resolution using a classification tree approach. Pattern Recognition Letters 27: 252–258. and D.0 – REMOTE SENSING OF LAND COVER CHANGE 114 . M. M.J. S. Integrating remote-sensing and ground-based observations for estimation of emissions and removals of greenhouse gases in forests: Methods and Guidance from the Global Forest Observations Initiative. 2013. Group on Earth Observations.R. Available at http://www. P. A. A. Townshend and R. K. Adusei. GFOI. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. 2004. Dubayah.N. A.wur.N.C. D. B.O. Geneva.. http://www.. Random Forests for land cover classification. Hansen M. 2012. Hackler.F. 2010. B. A. Perneela. Oliveira.

Jiang. Treitz. Wagner. Hyer.G. R. Miwa. Kutler. Muller-Landau. Uryu. May. Justice.C. J. Remote Sensing Environment 112: 2914–2926. L. Miwa. Buendia K. LiDAR remote sensing of forest structure. M. Published: IGES Japan. 2002.. 1999. P.R. Loveland. Fifty years of deforestation and forest fragmentation in Madagascar. Land-Use Change and Forestry.1371/journal. BioScience 52(1) 19-30. Sato. Krug. M. D. Computer Processing of Remotely-Sensed Images. Ren.J. Lepers. C. G. D. M. M. R.V. Good Practice Guidance for Land Use. P. Buendia.. W. T. D. eds. B. Steininger. Toriyama. Tanabe. K. Harper. 2003. Wulder.Hansen.J.. Y. The potential and challenge of remote sensing-based biomass estimation. Environmental Conservation 34: 325-333.F. and K. Y. Chicester. T. Ngara. Harding. K. L. J. Hansen. D.. (2008). S. Annual Review Environment Resource 28: 205–41. Longépé. eds. The suitability of decadal image data sets for mapping tropical forest cover change in the Democratic Republic of Congo: implications for the global land survey.J.S. Mascaro. 2007. 2012.W. T. T.. N. P. PLoS ONE 7(9): e45889. Transaction on Geosciences and Earth Observation 38: 1458–1463. C. Lawrence. G.O. A Simple Semi-Automatic Approach for Land Cover Classification from Multispectral Remote Sensing Imagery. Lim. Asner. Hirata Y. J. Definitions and methodological options to inventory emissions from direct human-induced degradation of forests and devegetation of other vegetation types.J. Lindquist E. L.. Detto. Lu. Flood. 2011. Rakwatin. Xu and H. Prepared by the National Greenhouse Gas Inventories Programme. Geist. M.). Humid tropical forest clearing from 2000 to 2005 quantified using multi-temporal and multi-resolution remotely sensed data. and C. Tucker.A. Remote Sensing Environment 115: 3770–3774. IPCC-IGES.L. Tanabe and F. M. K. M. Forestry and Forest Products Research Institute Japan. Mather. Johnson. G. J. Wiley & Sons. D. Kanagawa (eds. England. Zhu. 2006. IPCC. Potapov. Pipatti. Lidar Remote Sensing for Ecosystem Studies. W. Parker and D. P. H. Takao.K. F. REDD Research and Development Center. T. ISBN 978-4-905304-15-9. REDD+ MRV MANUAL: CHAPTER 5. R.pone. E. P. Harcombe. Townshend. Lefsky.. Steininger.. DeFries.P.. WB. P. 2003. J. W. Cohen. J.0 – REMOTE SENSING OF LAND COVER CHANGE 115 .S. IPCC. Stolle. Shimada. Dimiceli. Published: IGES. 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Stehman.C. Gytarsky. BioSAR™: an inexpensive airborne VHF multiband SAR system for vegetation biomass measurement. St-Onge and M. Pittman. 2000.J.. Assessment of ALOS PALSAR 50 m orthorectified FBD Data for Regional Land-cover Classification by Support Vector Machines. North American Forest Disturbance Mapped from a Decadal Landsat Record.C.J. Japan. Ngara and K. J. International Journal of Remote Sensing 27: 1297-1328. S. Huang.P.. Yulianto. Zhuang. H. 2006... G. M. and E. M.V. In: Penman. Imhoff. Huang.. Roy and C. Y. 2012.. Masek. P. 2003. Cohen.. 2011. X.. Hall. REDD-plus Cookbook. 2003. D. Eggleston. Dynamics of land-use and land-cover change in tropical regions. G. 2008b. Evaluating uncertainty in mapping forest carbon with airborne LiDAR. and Nelson. O. Wolfe.K. Hawkins.G. Isoguchi.R. M. PNAS 105: 9439-9444. And H. IPCC GPG-LULUCF. K. IEEE Transaction on Geosciences and Earth Observation 49: 2135-2050. and F. doi:10. Progress In Physical Geography 27(1): 88-106. Holford.. Carroll. 156pp. M.. Kruger.0045889 Lambin. 2008. International Journal of Remote Sensing 29: 7269-7275. F.

Stow. G. M. R. Cochrane and V. and Meir. Nelson R. Using satellite RADAR backscatter to predict aboveground woody biomass: A consistent relationship across four different African landscapes. and P. T. International Journal of Remote Sensing 26(1): 217-222.H. M. B. Large-scale impoverishment of Amazonian forests by logging and fire. Ryan. Skole. Miller. G. S. 2003.. A. J. P.L. S. B. Williams. Canadian Journal of Forestry Research 39: 862-881.A Lefsky. Williams. T. Mitchard. I. Potter.. Estimating Quebec provincial forest resources using ICESat/GLAS.M. Jeffery. G. P. Olofsson.T. P. Ghee.V. Stehman. V. Nobre. M. Levian and C. 2003. and P.. P. White.G. Ghimire. Remote Sensing Environment 110: 98–108 Nelson. Imhoff.H.S. M. S. T. REDD+ MRV MANUAL: CHAPTER 5. Woodcock 2013. Good practices for estimating area and assessing accuracy of land change. J. C. Ryan. Quantifying small-scale deforestation and forest degradation in African woodlands using RADAR imagery. M. Verissimo. Biogeoscience 9: 179-191. A. E. A. Nepstad. C. Cassells. T. W. I. Hyde. I. Rogan. Foody. C. L. Photogrammetric Engineering & Remote Sensing 69(7): 793-804. Mapping tropical forest biomass with radar and spaceborne LiDAR in Lopé National Park. I. L. M. P. Mapping deforestation and secondary growth in Rondonia.E.. Gobakken and G. M.. Making better use of accuracy data in land change studies: Estimating accuracy and area and quantifying uncertainty using stratified estimate. K. T.H. E. A. Lefebvre. Abernethy. Brooks.. Nature 398: 505-508. M. Meir. Gabon: overcoming problems of high biomass and persistent cloud. C. E. Brazil. Alencar. Gregoire.L. S. Stahl. Naesset. Lewis. Fischer 2003.A. Global Change Biology 18(1): 243-257. Campbell. Chica-Olmo. Saatchi..L. and W.R. Sonk´e. 1999. E. Land-Cover Change Monitoring with Classification Trees Using Landsat TM and Ancillary Data. P.A. Feldpausch and P.M. C. Margolis. N. Standardizing the calculation of the annual rate of deforestation. E.M. Leal. 2005. 2012. Pal. Remote Sensing Environment 59: 167-179. B. Remote Sensing Environment 115: 2861–2873. Remote Sensing Environment 86: 554-565. Remote Sensing Environment 148: 45-57. Remote Sensing Environment 129: 122-131. S. D.V. Lewis.S. Investigating RADAR-LiDAR synergy in a North Carolina pine forest.E. Moutinho. Measuring biomass changes due to woody encroachment and deforestation/degradation in a forest- savanna boundary region of central Africa using multi-temporal L-band RADAR backscatter. T. Rodriguez-Galiano. H. Grace.. S.P. J.M. 2009. 2009. R. 2011.. Forest Ecology and Management 177: 593-596. and R. Woodhouse P.. Lima. Salas and D. using imaging RADAR and thematic mapper data. Nangendo.T. Johnson. Herold. 2012. Mitchard. S. Feldpausch. E.G. Ribeiro.E. M. Edwards 2007.M. Franklin. Schlesinger. Stehman and C. Pal. K. J. S. J..F. T. Wulder 2014.0 – REMOTE SENSING OF LAND COVER CHANGE 116 . J. M. Woodhouse.Mitchard. Woodcock. Woollen. Random forest classifier for remote sensing classification. P. Incorporating the Downscaled Landsat TM Thermal Band in Land-cover Classification using Random Forest. Photogrammetric Engineering & Remote Sensing 78(2): 129-137. 2012.J. G. Woodhouse. Hill. Rignot. An assessment of the effectiveness of decision tree methods for land cover classification. Olofsson. Foody. S. Mendoza. Puyravaud. Geophysical Research Letters 36. Mather. Congalton.F. M. E. Woodhouse and M. C. Mitchard. Saatchi. Collins. Pardo-Iguzquiza. 1997. A. Lewis. Meir. T. E. E. C. Saatchi.J.S. D. S. Rowland. Emessiene. E. Boudreau.

M.. C.E. Kirsch and D. Large-Area Classification and Mapping of Forest and Land-cover in the Brazilian Amazon: A Comparative Analysis of ALOS/PALSAR and Landsat Data Sources. 2001. D. J. S. K. Seto. Walker. D. International Journal of Remote Sensing 26(3): 425-429.C. Mitchard. Brazzaville.M. A meta-analysis of terrestrial aboveground biomass estimation using lidar remote sensing. Lefsky. J. 2009. M. Saatchi. Perneel and M. D. Roberts. Leicester. R. Buermann. Shimoni. Moghaddam.S. Song. Estimation of crown and stem water content and biomass of boreal forest using polarimetric SAR imagery. W.0 – REMOTE SENSING OF LAND COVER CHANGE 117 . Brown. Sibley. C. P.O . M. C. Rodrigues. Kellndorfer.. M. Benchmark map of forest carbon stocks in tropical regions across three continents. L. M. Souza.. Sulla-Menashe and M. W. E. Reporting and Verification (M & MRV) in the context of REDD+ Activities.. Global Change Biology 15: 2803–2824.S. C. International Journal of Applied Earth observation and Geoinformation 11: 169-180. S. M. Republic of Congo. W. G.A. UK. Saatchi. C. National Forest Monitoring Systems: Monitoring and Measurement. Delabie. Claudia.T. S. S. B.V. S. Woodcock. J. S. Macomber. 2001. Acheroy. Nepstad. and D.S. Dubayah. White. M. Walker.Saatchi.. Fusion of Pol SAR and PolInSAR data for land-cover classification. W. 2011. 2000.P. Jr. 2005. K. The potential ecological costs and cobenefits of REDD: a critical review and case study from the Amazon region. Examining Fragmentation and Loss of Primary Forest in the Southern Bahian Atlantic Forest with RADAR Imagery.A.. Newell. 2005. Olofsson. July 20-23. N. Combining spectral and spatial information to map canopy damages from selective logging and forest fires. Mcgrath. J. Woodcock. Herold 2010. Lenney and S.L. Remote Sensing Environment 128: 289-298. K. D.R.. Davidson. H. Alves. Classification and Change Detection Using Landsat TM Data: When and How to Correct Atmospheric Effects? Remote Sensing Environment 75: 230-244. Accuracy 2010 Symposium.A. D. Lewis.S. Harris. Saatchi. M.A. UN-REDD Programme. Morel. Souza. Designing a reference validation database for accuracy assessment of land cover. Mapping forest degradation in the Amazon region with Ikonos images. A. Heremans. Hagen. 26-27 October 2012.G. Stickler.S. S..C. and M. Friedl.J. C. Goetz and R. D.. Soares and D. S. Nepstad. 2013. Musinsky. PNAS 108(24): 9899-9904. Conservation Biology 15(4): 867- 875. C. Zutta. Alger. and J. Salas.. Soares-Filho and E. Borghys.T. Mapping deforestation and land use in Amazon rainforest by using SIR-C imagery. Agosti. S. Petrova. Coe.M. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 3: 594-604. Cochrane. B. Silman and A. Stehman. IEEE Transaction on Geosciences and Earth Observation 38(2): 697-709. 2010. S.C. Remote Sensing Environment 59: 191–202.. 2009. M. Zolkos. Roberts and M. S. REDD+ MRV MANUAL: CHAPTER 5. 1997. Remote Sensing Environment 98: 329-343.L.

encroachment. large-scale NASA 500m multispectral day deforestation 1km thermal (diurnal/ nocturnal) AWIFS Resource Sat-1 730km 56m 5 days Yes Yes large-scale deforestation IKONOS IKONOS – 11. log ponds SPOT-6 Airbus Defense & 60km 6m multispectral Varies No Yes deforestation. canopy gaps.NASA 60km 15m multispectral Varies No Partial (no SWIR deforestation. illegal fishing vessels MODIS Terra / Aqua – 2330km 250m visible 4 times per Yes Yes fires. roads.3km @ 4m multispectral (4) Varies No Yes skid trails. GeoEye nadir 1m panchromatic (1) illegal fishing vessels / logging vehicles Pleiades 1a and Airbus Defense & 20km 50cm panchromatic 2m Varies No Yes Skid trails. with SLC-off deforestation.7m panchromatic 26 days Yes Yes skid trails. 60m thermal roads. HRC CBERS-2B – INPE 27km 2. roads.0 – REMOTE SENSING OF LAND COVER CHANGE 118 . log ponds ASTER Terra . 185km 15m panchromatic 16-21 days Yes yes deforestation. NASA 30m multispectral encroachment. 5m panchromatic (2. log ponds SPOT-5 CNES 60km 20m multispectral Varies No Yes deforestation.5m encroachment. 100m thermal roads. channels) encroachment. 165km 15m panchromatic 16-21 days Yes Yes. canopy gaps. 1b Space multispectral logging roads REDD+ MRV MANUAL: CHAPTER 5. interpolated) roads.5.7 COMMON SATELLITE DATA SOURCES FOR LAND-USE MONITORING Sensor Satellite – Swath Resolution Repeat Systematic Operational Monitoring Agency Width Cycle Acquisitions Status Applications Landsat 7 ETM+ Landsat 7 .5m panchromatic encroachment. log ponds CCD CBERS-2B – INPE 113km 20m multispectral 26 days Yes Yes deforestation. Space 1. log ponds Landsat 8 OLI LDCM . NASA 30m multispectral gaps encroachment.

skid trails WorldView-2 WorldView-2 Varies 0. Constellation encroachment.31m panchromatic (1) < 1 day No Yes skid trails. 3-10 m Stripmap encroachment.6m (1) panchromatic Varies No Yes skid trails.1km 0.4m multispectral (4) illegal fishing vessels / logging vehicles RapidEye BlackBridge 77km 6. 19 m dual pol No. log ponds 30 m quad pol 100 m Scan SAR WorldView-3 WorldView – 13. GeoEye (resampled to 0. DigitalGlobe 1. DigitalGlobe 1.8m multispectral illegal fishing vessels / logging vehicles Radarsat-2 CSA varies 8 m quad pol fine 24 days varies Yes deforestation.5m panchromatic Varies No Yes skid trails.65m multispectral (4) logging vehicles QuickBird QuickBird – Varies 0. canopy gaps. 100 m Scan SAR roads. canopy gaps.quad pol roads.4m panchromatic (1) Varies No Yes skid trails.0 – REMOTE SENSING OF LAND COVER CHANGE 119 .24m multispectral (8) illegal fishing vessels / 3. log ponds REDD+ MRV MANUAL: CHAPTER 5. canopy gaps.5m multispectral 1-5 days No Yes Deforestation.all mode No deforestation.7m SWIR (8) logging vehicles PALSAR-2 PALSAR-JAXA varies 1-3 m Spotlight 15 days Varies Yes deforestation. DigitalGlobe 2. roads. canopy gaps. 25 m quad pol standard roads.5m) illegal fishing vessels / 1.GeoEye-1 GeoEye-1 . Varies 0. log ponds 100 m wide ASAR ENVISAT-ESA varies 30 m polarization mode 36 days varies No deforestation 150 m Wide Swath mode 1 km Global Monitoring mode PALSAR PALSAR-JAXA varies 9 m Single pol 45 days Yes.

0m roads.Future Missions Sensor Satellite – Swath Resolution Repeat Cycle Systematic Operational Monitoring Agency Width Acquisitions Status Applications SPOT-7 Airbus 60km 1. Defense & panchromatic (1) in calibration encroachment. Space 6. log ponds multispectral (4) Sentinel-2 ESA 290km 10/20m/60m 5 day (once both 2A Anticipated launch deforestation A/B & 2A are launched) summer 2014 (2A). then (2B) VIIRS NPV – NASA 3000km 750m 2 times per day Yes recently launched.0 – REMOTE SENSING OF LAND COVER CHANGE 120 . in fires calibration REDD+ MRV MANUAL: CHAPTER 5. deforestation.5m No Launched June 2014. Software EXELIS: ENVI Systems for World Surveillance.php?section_id=17 European Space Agency Earthnet American Museum of Natural History (AMNH) Biodiversity Informatics Facility.aspx INTERGRAPH: ERDAS Imagine NASA Earth Observatory Random forests Software REDD+ MRV MANUAL: CHAPTER IDRISI GIS and Image Processing Software Open source GIS and remote sensing software http://biodiversityinformatics.0 – REMOTE SENSING OF LAND COVER CHANGE 121 .eo.amnh.clarklabs.cfm Quantum GIS http://www. SELECTED RESOURCES Online guides and other materials United Nations Space Science and Technology: http://www. Biodiversity Informatics Facility http://biodiversityinformatics.exelisvis.unoosa.aspx PCI Geomatics http://www.5.php?section_id=33&content_id=138 GRASS GIS http://grass.

SAR Tool Kit – REMOTE SENSING OF LAND COVER CHANGE 122 .htm Rulequest data mining NEST – Next ESA SAR toolbox Data access USGS Earth Resources Observation and Science Center (EROS) REDD+ MRV MANUAL: CHAPTER 5. RAT – Radar Tools (Landsat Archive.usgs.Map Ready.berkeley.berlios.inpe. and MODIS) RapidEye Catalog (Enables searching of both LandsatLook images & Level1 Landsat data) National Research Institute (INPE) of Brazil http://www. Ikonos. as well as various MODIS products) Global Land Survey (GLS) 2005 products: Global Land-cover Facility (Range of Landsat and CBERS imagery.html R statistical language ESA – polsarpro (Polarmetric SAR Data Processing and Educational Tool) (Range of data sources including the Landsat archive and selected imagery for a range of instruments including ASTER.r-project. Global Land Survey (GLS) as well as various ASTER and MODIS products) USGS LandsatLook Viewer Open source Alaska Satellite Facility . See5 classification software http://www.rulequest.http://www.alaska.

html Saatchi S. remote-sensing/11740 USGS Change-tracking tool NASA fundamentals of remote sensing http://gcmd.appspot.A. These may be useful for national stratification of field sampling of biomass in a MRV system.L. N. M. W. Nature Climate Change http://dx. Baccini. Mitchard. Hagen. Houghton. W.doi.1038/NCLIMATE1354 Pantropical National Level Carbon Stock Dataset http://www. Goetz. White.T. Hackler.nasa. Brown. P. produced by the University of Maryland and distributed by Google.(Access to the BlackBridge RapidEye image archive) SPOT Catalog http://catalog. S. Sulla-Menashe. Buermann. M. Salas. Zutta. B.L.gc. Silman and A. Lewis.or. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps.0 – REMOTE SENSING OF LAND COVER CHANGE 123 .R. Laporte.108(24):9899-904.usgs.fas.A. Petrova.whrc.aster. http://earthenginepartners. S. D. Dubayah. Beck. Morel.nrcan. Tutorials The remote sensing tutorial: Federation of American Scientists (FAS) http://www.S. Two maps of global forest biomass. J.jpl.spotimage. Lefsky. based on a suite of satellite data inputs. W. Harris.ersdac.nasa. L. Proceedings of the National Academy of Sciences 14. calibrated with plot data. http://carbon. E. N. with a 1-km T.html General remote sensing: Canada Centre of Remote Sensing http://www. M. A. 2012. Sun.S.html REDD+ MRV MANUAL: CHAPTER 5. Samanta and R. R. M.aspx?language=UK (Access to the SPOT satellite archive) Earth Remote Sensing Data Analysis Center (ERSDAC) (Access to the ASTER imagery archive) Global tree-cover loss and biomass data Global Landsat-based estimates of tree-cover loss from 2000 to 2012. Benchmark map of forest carbon stocks in tropical regions across three continents. S SAR-Land-Applications-Tutorial/sar_land_apps_1_theory.Virtual Hawaii ESA’s Synthetic Aperture radar: Land applications tutorial An introduction to radar remote sensing: Canada Centre of Remote Sensing http://www.nrcan.gc.gc.0 – REMOTE SENSING OF LAND COVER CHANGE 124 .gov/NOAASIS/ml/education.1.noaa.csiro.esa.nrcan.Introduction to remote sensing .html NOAA’s Satellite and Information Service: Learning About Satellites and Remote Sensing ESA’s RADAR Tutorial http://earth.html An introduction to remote sensing CSIRO http://www.pdf REDD+ MRV MANUAL: CHAPTER Radar polarimetry: Canada Centre of Remote Sensing http://www.

U.S.usaid. Agency for International Development 1300 Pennsylvania Avenue. NW . DC 20523 Tel: (202) 712-0000 Fax: (202) 216-3524 www.