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


VERSION 2. REPORTING AND VERIFICATION (MRV) MANUAL. 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. ACRONYMS AND ABBREVIATIONS ACR American Carbon Registry AD Activity Data AFOLU Agriculture. Markets and Communities Program FCPF Forest Carbon Partnership Facility REDD+ MEASUREMENT.0 v .

0 vi . Technology and Innovation MMU Minimum-mapping unit MRV Measurement.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. REPORTING AND VERIFICATION (MRV) MANUAL. 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. 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. VERSION 2. Land-use Change and Forestry GPS Global Positioning System IDEAM Colombian Institute for Hydrology. Reporting and Verification N20 Nitrogen oxide NAMA Nationally Appropriate Mitigation Strategies NASA National Aeronautics and Space Agency REDD+ MEASUREMENT.

accuracy. REPORTING AND VERIFICATION (MRV) MANUAL.0 vii . comparability. sustainable forest management and enhancement of forest carbon stocks. plus the role of conservation. completeness. 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. 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. VERSION 2.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.

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

3. Section 3. This chapter is relevant to the activities highlighted on the following page.0 REMOTE SENSING OF LAND COVER CHANGE Authors: Marc Steininger and Jennifer H ewson 5.1 INTRODUCTION This chapter focuses on the application of remote sensing-based approaches to forest cover and change monitoring. REDD+ MRV MANUAL: CHAPTER 5.0 – REMOTE SENSING OF LAND COVER CHANGE 82 . of this Manual outlines the sequence of steps required for generating a national greenhouse gas (GHG) inventory. Inventory and Reporting Steps. 5.

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

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

the extent over which a human activity occurs. or the extent over which a human activity occurs.g. Land use. documenting the area of each land use over time and with full coverage.. unless noted otherwise. See Chapter 3). The definition REDD+ MRV MANUAL: CHAPTER 5. These approaches. However. the coefficients quantifying the emissions/removals per unit area.g. EFs are coefficients that quantify the emissions/removals per unit area. AD are combined with Emissions Factors (EFs). yet no tracking of specific conversions among them Approach 2: Tracking of land-use conversions.1: Approaches Approach 1: Net area of each land use reported at different time period. In Australia.0 – REMOTE SENSING OF LAND COVER CHANGE 85 . the creation of a multi-temporal map of change in forest cover as well as some sub-categories enabled Approach 3 for those categories. as it is based on relationships between observed spectra in the images and the structural characteristics of the soil and vegetation covering land.2). While satellite-based remote sensing is a valuable tool for monitoring several parameters of land use. Throughout this chapter. listed in Box 5. plus the role of conservation. 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. some types of land-use categories (e. over a given period of time.Activity data (AD). The costs associated with these approaches could be significant and thus necessitate a sampling-based approach. Countries have some flexibility in developing their forest definition.1. can usually be inferred based on local context and a general knowledge of the area. sustainable forest management and enhancement of forest carbon stocks in developing countries (REDD+). are used to estimate AD for each Category: Box 5. by the EFs. satellite monitoring is more suited to detecting land cover. For example. it is acceptable to use a mix of the three approaches among regions or categories in a country.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. 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. The IPCC (2006) describes three overall approaches. While countries must report on land use.1 of the GPG-LULUCF. or regions (e. may be more effectively monitored through airborne or field-based data collection approaches. mountainous areas).. thus enabling Approach 2. Data existed on land-cover change. Forest definition A fundamental step in the development of a MRV system is the national definition of forest. yet they are constrained by certain criteria. existing data available for the Argentine Pampas were sufficient for either Approach 1 or 2. however. forest degradation). not to be confused with tiers. thus providing the data on differences in the carbon stocks before and after the cover change has occurred. provides the estimated associated GHG emissions for each category. definitions have been adapted from the 2006 IPCC Guidelines and the GPG-LULUCF 2003. Multiplying the AD. Forestry. documenting transformations between natural grasslands to pasture and cropland. 5. As illustrated in Chapter 3 (Figure 3. and Other Land Use (AFOLU) in Volume 4 (IPCC 2006. thus enabling Approach 1.2. are data on the area of a Category that potentially results in GHG emissions or removals. Agricultural census data. Case studies of countries that have used different approaches are provided in Annex 2A. for the representation of land use (see Chapter 3). existed for the entire region.

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

as this may result in the inclusion of forests that were cleared just before the start date. it will necessitate several iterations of image analysis followed by reviews from experts and stakeholders. and Step 5. and those strata with little difference in stocks may be combined. This could be the date of initiation of REDD+ activities. Generally.4.2 of the GFOI MGD (GFOI 2013).2 of the GOFC-GOLD SourceBook (GOFC-GOLD 2013) provide useful overview information on the range of approaches and methodologies available. Different potential strata can be assessed by merging with any existing data on carbon stocks.2. The creation of a forest benchmark map facilitates the estimation of change within forests. strata can be created based on expert opinions of forest types that should have different biomass levels. such as sub-national governments and REDD+ project implementers. or from very high resolution imagery. resulting in fewer total strata. Other characteristics often discernable from satellite data are major differences in canopy structures. where more deciduous forests may have significantly lower biomass levels. as this sets an important precedent for where REDD+ activities may or may not be implemented.5. Section 3. Step 5. It is important that a final benchmark map is communicated to stakeholders with any concerns addressed. e. 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. or Section 3. Stratification can also make use of ancillary GIS data. gaps can occur as a result of cloud cover.0 – REMOTE SENSING OF LAND COVER CHANGE 87 .. of this Manual. it will eventually need to be updated to define the extent of forest at the beginning of the implementation of REDD+ activities. soils. liana forests. and palm forests. as this is the main emissions source for many countries. if acquired for particular leaf-on and leaf-off months. Many approaches exist for estimating change within forests. such as data on climate.2 of the GOFC-GOLD SourceBook (GOFC-GOLD 2013). see Chapter 2). Section 5.5.are excluded from the national definition. A key consideration is the type of change to estimate. Section 5. The creation of a forest benchmark map will require a careful and iterative process. A conservative approach is to avoid using images from before the start date. Seasonal information could be obtained from coarser resolution data.. and elevation. In the latter case. At a minimum. 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. such as those of inundated forests. Satellite images for an exact date can be difficult to acquire due to cloud cover or gaps in data collection. etc.g. and thus it defines the extent within which REDD+ crediting may be possible and where monitoring is required. While the inclusion of the smallest clearances. Stratification is not required. Once the forest benchmark map is finalized. A decision should also be made on the minimum size of clearance to be estimated. Therefore. analysts will need to identify an appropriate methodology for analyzing multiple images. Even the creation of a few broad classes is very useful for these purposes. 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.e. although it is recommended as a method to reduce both field-survey costs and carbon-stock uncertainties. forest clear-cutting (removal of all trees) should be estimated. such as montane shrubs. smaller than a few hectares.4. as discussed in Chapter 4. such as monthly image composites from MODIS. or open-canopy vegetation. for example. image dates used for the benchmark map may vary. This may necessitate the acquisition of multiple images for the same area and. 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. The benchmark map should be based on satellite data from dates closest to the start date of a monitoring period. A forest benchmark map should have as few data gaps as possible. thus.2 of the GFOI MGD (GFOI 2013). i. satellite sensor problems. and these can be re-combined later if field data indicate otherwise. See. such as open woodland. This is a critical step for analysts to correctly train and assess classifications of forest for the benchmark period.2. Stratification based on satellite imagery can benefit from the inclusion of seasonal information. If such field data do not exist. 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. can REDD+ MRV MANUAL: CHAPTER 5. of this Manual.

However. Again. Table 5.1 presents an example of a land-cover change matrix with three broad categories (forest. an initial step could be to produce a forest benchmark map. Issues related to sub-classes are somewhat different for forest versus non-forest. a country will conduct a stratification for a national forest inventory.. 2000). not all of these must be estimated in the same process. as it can be assumed there are no transitions among these naturally- occurring vegetation types over the required reporting periods. while not requiring a very difficult process of spectral classification of all transitions among sub-classes. Table 5. and many different opinions within the research community. such as sampling of higher-resolution data or even targeted field surveys. with forest sub-classes that have significantly different carbon stocks. and may not reflect a significant source of emissions. 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. In contrast. This can be performed one time. degraded forest. Other classes It may be important to further stratify the six broad use classes into sub-classes where carbon stocks vary significantly. cropland. could lead to greater classification error.1 and 5.2 presents an example with greater thematic precision where the broad categories have been divided into sub-classes. the time and resources involved may be costly. For example. For example. For forests. Again.2 illustrate this below. strong justification in terms of improvement of emissions estimates is needed to justify attempts to include transitions among these classes. Tables 5. plantations and fallows may be difficult to distinguish. such as wildfire. 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. By combining the two. especially when one cannot be very selective about the season of the imagery used for analysis because of frequent cloud cover. and then use a different approach. broad deforestation class that occurs anywhere within this benchmark. depending on the season and stage of crop development at the acquisition times of available images. A step-wise process may be worth exploring that uses different methods and levels of detail for different monitoring aspects. a country could first produce an estimate of forest clear-cutting using one approach. deforestation can be attributed to different forest sub-classes. Including sub-classes may provide data that are very useful for REDD+ national strategies and management policies.0 – REMOTE SENSING OF LAND COVER CHANGE 88 . Spectral distinction of these uses is often difficult. Conversely. for degradation. as well as any important transitions among those. Other forms of degradation. This is an area where there are many options. The expected increase in the accuracy and overall usefulness of emissions estimates should be clearly assessed. Foody. may be considered. nonforest). transitions among different post- deforestation land uses do occur over short time periods. A second step could be to produce a map of a single. Combining approaches like these can provide all of the necessary estimates to complete a full land-cover change matrix. For example. 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. following the creation of the forest benchmark map. 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. countries should have strong justification for including sub-classes since this will require methodologies that consume more resources.g.potentially yield a more correct and higher rate. managed grassland. 1999. Mather. REDD+ MRV MANUAL: CHAPTER 5.

such as hectares.970 25 2. 4. T1 and T2 are the first and second time periods.2 100 Degraded 5 1.6 percent (adding values 0.000 4. and in (b) are percentages.1: Example of a land-use change matrix with few land-use classes and change categories. “Forest” in this table is non-degraded forest only.7 98. % Forest Forest forest T1 Forest 9.0 100 forest forest Sum T2 9.0 – REMOTE SENSING OF LAND COVER CHANGE 89 .010 4. 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 5.4 and 0.3 98.000 Degraded 0.2 in the first row of (b)).5 Table 5.4 0. In this example. 100. “Non-forest” includes all non-forest.5 1. T2 a) Forest Degraded Non-forest Sum T1 b) Forest Degraded Non.4 0.045 % T2 99. Values inside the matrix are areas and percent change for each category of persistence or change.945 2. both naturally-occurring and anthropogenic.3 100 T1 T1 Forest Forest Non.9 101.000 Forest 99.940 40 20 10.000 Non. Values in (a) are in absolute units. referred to in the IPCC as “Initial land-use class” and “Land use during reporting year. gross deforestation plus forest degradation is 0.” Values in Sum T1 and Sum T2 are total area and percent change for each class.

5.1 27.3 98.1 0.0 100 Pasture 3.4 0.3 87.1 100 Montane Forest 99.0 0. including the types and availability of different satellite data.4 100 Fallow 63.1 100 Degraded Lowland Forest 0. either via a shortening of fallow cycles or an increase in permanent pasture. the large areas of change from cropland to fallow (200) or pasture (100).6 9.0 – REMOTE SENSING OF LAND COVER CHANGE 90 . The 35. croplands.8 135. classification algorithms.4 0.3 0.8 100 Degraded Montane Forest 98. approach should be applied to monitoring different land cover classes. and pasture) and forest degradation also occurs in the lowlands. the majority of deforestation (to fallow. level of automation and analyst expertise. the majority of forest occurs in the lowlands. Natural Grassland. according to the national forest definition. Values inside the matrix are areas and percent change for each type of category.1 0.1 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.2 99.0 10.2: Example of a land-use change matrix with more precise land-use classes and change categories.2 0.2 0. non-degraded forest.4 0.3 0.1 percent increase in pasture indicates an increasing importance of this use. or sampling-based. Cropland.8 100 T1 Natural Grassland 99. Criteria include the type and resolution of satellite data and the degree to which a full coverage. b) T2 Degraded Degraded Lowland Montane Natural Lowland Montane Fallow Cropland Pasture % T1 Forest Forest Grassland Forest Forest Lowland Forest 99.6 100 % T2 99.9 0.7 99.3 100 Cropland 20. pre-processing. Each of these decisions is discussed below. T1 and T2 are the first and second time periods. A high degree of rotational land use is also indicated by. The appropriateness of different monitoring methodologies will need to be assessed. referred to in the IPCC as “Initial land-use class” and “Land use during reporting year. it is important to consider how consistency will be achieved and what methods will be used to effectively combine data from different time periods. In this example.0 70. Values in (a) are in absolute units. The 12. Where automation is not possible.5 80.1 Table 5.8 99.5 percent reduction in fallow indicates intensification of land use.4 96.5 0. such as hectares and in (b) are in percent. “Forest” here means intact. REDD+ MRV MANUAL: CHAPTER 5.3 OVERALL STEPS AND NEEDS Figure 5. for example. and Pasture represent non-forest classes.1 0.” Values in Sum T2 are total area and percent change for each class.4 0. Fallow.5 98.

Figure 5. whereas other dynamics. do change events occur in small patches of several hectares. 2) What are the appropriate scales and/or sampling approaches for monitoring? Once the categories and classes to monitor have been assessed. it is necessary to consider the appropriate scale and approach. 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. 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.0 – REMOTE SENSING OF LAND COVER CHANGE 91 . These latter land-use dynamics may require more costly data -collection REDD+ MRV MANUAL: CHAPTER 5. and thus where monitoring should be conducted (see Chapter 2). For example. Finally. For MRV. For example. or are they much larger? Different types of changes may also be most appropriately monitored with different sources of data. may require airborne or field-based monitoring. a country must determine the geographical extent of managed land.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. particularly some forms of degradation and post-deforestation land-use changes.

While older secondary forest can be difficult to distinguish from mature forest. or a case should be made that using a larger MMU does not significantly affect resulting area estimates. while using a larger MMU may not significantly affect area estimates for static areas of classes. possibly combining information from multiple dates into a single multi-date product. 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. 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. The MMU should be smaller than the minimum patch size included in the national forest definition. although it is only vaguely defined in the context of satellite monitoring. such as deciduous woodland. Question 6 includes merging of temporary sub-classes. 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. focusing on the categories identified in the KCA. some vegetation types. estimates of change can be very sensitive to the MMU. and what level of analyst expertise is required? 4) How much of the process can be automated. However. 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. By distinguishing mature forest versus fallow areas. either during creation of the forest benchmark map or during the stratification of forests in this map. Many differing views exist regarding optimal methods for monitoring land-use change and. therefore. Further.0 – REMOTE SENSING OF LAND COVER CHANGE 92 . a country should seek to produce the most accurate estimates possible for key categories while using an approach that is “replicable. most fallows younger than 10 years are distinguishable. and the great majority of fallow periods are shorter than 10 years.processes and thus necessitate a sampling approach. including data sources and sampling? Some of the most important considerations are whether to use optical data versus Radio Detection and Ranging (RADAR) data. and conduct assessments with national data. a country can minimize confusion between mature forest clearing and fallow cycles in later monitoring. This will avoid either: (i) the use of methods with little justification based on in-country testing. may require data from particular or multiple seasons within each year. 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. 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. and for those parts that are dependent on analyst interaction. A country should obtain opinions from a range of international experts. 3) What methodological aspects should be considered? A country should consider a range of methodological options. For question 4. and often some type of filtering to a defined minimum-mapping unit (MMU). again necessitating increased data and analysis demands. Question 7 should consider REDD+ MRV MANUAL: CHAPTER 5.

as well as an independent team of analysts to interpret validation data and conduct error calculations. Numerous text books are also available on remote sensing of land-cover. Links to selected internet resources are provided in Appendix 5B.2). sampling approach.4) and. repeated coverage offered by one or more satellite data sources. The amount and type of energy sensed is usually recorded in digital form. and the type representing the recording of the signal across a spectrum. this is the visible portion of the EM spectrum. Fundamental assumptions. or entire countries. the amount representing the strength of the signal. though not always valid. The sun is the source for visible and shortwave-infrared spectral regions of the earth. the instrument does not emit its own signal. Airborne remote sensing capacities are also of interest. by combining images of energy measured in different spectral bands and assigning a separate color for display. REDD+ MRV MANUAL: CHAPTER 5. “multi-spectral” images are produced.4 REMOTE SENSING OVERVIEW This section provides a summary of remote sensing fundamentals. Passive remotely-sensed data are acquired by a sensor that passively receives energy originating from another source. rather than being in direct contact with it.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.0 – REMOTE SENSING OF LAND COVER CHANGE 93 . the feature itself is the source for thermal-infrared regions (Figure 5. 5. 5. Passive remote sensing The majority of remotely-sensed data used for monitoring land use is passive. Satellite-based remote sensing is most common because of the full. Different spectral regions are represented by relatively narrow “spectral bands” (Figure 5. At the highest level. The human eye.4. 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. senses a relatively small portion of the total spectrum of energy emitted by the sun.various sampling schemes and the availability of very high-resolution satellite or aerial observations. thus enabling national monitoring for terrestrial-based applications. for example. and resources available.5.3). are that different land-cover types can be distinguished based on this recorded information and that land use can be inferred from land cover. as illustrated in Figure 5. as such systems could be applied over large regions. two broad types of remote sensing for monitoring land-cover exist: passive and active. depending on the type of equipment.

view angle and atmospheric properties. with so many channels and bands would be considered hyper-spectral. Shortwave energy is emitted by the sun. and in the example here they are divided by white lines. A single channel of a multi-spectral sensor is sensitive to energy only within a certain band.2: Optical satellite remote sensing. Most types of land-cover are a mixture of these features.4: Example of spectral resolution. In (a).Figure 5.0 – REMOTE SENSING OF LAND COVER CHANGE 94 . and a sensor with channels along these bands would have a high spectral resolution. passes through the atmosphere. A sensor. A) B) Figure 5. plus non-green vegetation and shadows caused by the geometry of terrain and vegetation. reflects off a surface. Both (a) and (b) represent the entire visible range of the electromagnetic spectrum. REDD+ MRV MANUAL: CHAPTER 5. passes again through the atmosphere and reaches a sensor on board a satellite. topography.3: Generalized spectral curves of fundamental features in remote sensing of land-cover. the bands cover a wide range of energy. The signal detected is dependent not only on the reflectance properties of the surface but also on the sun angle. the bands are narrow. and a sensor with such bands would be considered a broad-band sensor. Spectral bands are defined by a range of wavelengths. In (b). such as illustrated in (b). Figure 5.

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

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. per transmitted pulse. for example. 2002. Nelson et al. Both forms of LiDAR have been shown to be useful for estimating forest biomass via comparison with field data and modeling. these parameters can be used to model estimates of aboveground biomass (see. sub-canopy topography... Clark et al. Lim et al. Full-waveform systems record the entire waveform of a returning pulse. and the horizontal and vertical distribution of vegetation. usually between one and five. does not penetrate clouds. discrete waveform LiDAR (Lim et al. However. others have scanning abilities to collect data both along and across sampling lines. Applications of LiDAR in forestry have mainly focused on measurement of canopy height. 2003. Lefsky et al.. Full Discrete Waveform Return LiDAR LiDAR Figure 5. . While some LiDAR instruments collect data only along sampling lines. in contrast to RADAR. while discrete systems sample a discrete number of points.As with RADAR. LiDAR systems are also generally classified into full-waveform LiDAR and discrete LiDAR systems (see Figure 5.6). thus.6: Example of full waveform vs. 2003) REDD+ MRV MANUAL: CHAPTER 5. 2004.. enabling the creation of images.0 – REMOTE SENSING OF LAND COVER CHANGE 96 . LiDAR operates in the visible and near infra-red portions of the EM spectrum and. some of which is scattered back to the sensor by the target. 2009). LiDAR instruments emit a pulse of energy.

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

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

• 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. yielding a dataset with accuracies within 30m and. Pre-processing usually includes geometric registration and co-registration. Co-registration is a standard. is optional depending on the image analysis approach. thus. analysis. 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. although images have been geometrically registered. and enables the estimation of rates and patterns of land-cover change to be generated. Automation is increasingly available for processing numerous images. though useful. and is dependent on both the incident and reflected directions. and image analysis is the process of generating a land-cover class for all parts of an image. Co-registration among images should be reviewed and may require adjustments. co- registration may still be necessary. 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. These data have already been geometrically corrected using precision ground control points and Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) information. 25 http://landsat. However. but traditional analyst- driven methods are also sufficient. As previously outlined. The United States Geological Survey (USGS) has been reprocessing much of the Landsat archive. The summary of pre-processing. post-processing activities may include a number of steps.0 – REMOTE SENSING OF LAND COVER CHANGE 99 . Data transformation. the calculation of change rates and error estimates are required. Therefore.php REDD+ MRV MANUAL: CHAPTER 5. This is to understand the geographical area represented and is applied when importing the image into a GIS or image- analysis format for processing. and occasional data transformation. Atmospheric correction may be necessary depending on the image analysis approach that will be used. using examples of Landsat data analysis. geometric registration may have errors up to 100s of meters. Atmospheric correction is frequently performed in combination with a bi-directional reflectance distribution function (BRDF) correction.2 Image pre-processing. Therefore. 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.8). resulting in the creation of a L1T precision and terrain corrected product analysis. atmospheric correction. and post-processing steps below is based on optical data and approaches to classification. BRDF defines how light is reflected from a surface. Post-processing occurs after the image analysis step. eliminating the need for further geometric correction. Image pre-processing Geometric registration and co-registration Geometric registration is the process of mapping data in a geographical coordinate system. and post-processing Image pre-processing refers to any step that is applied to an image in preparation for the image analysis step.usgs. Finally.

These corrections can be applied to several partially-cloudy images of the same area. Some semi-automated methods apply constant signatures over multiple images or image dates. appearing as darker or brighter patches (Figure 5. 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. and these methods are highly dependent on careful atmospheric correction. 2008). all Landsat data. These usually assume constant atmospheric conditions across an image. is the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) tool (Masek et al. LEDAPS also generates water. in unitless values from zero to one. 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.9). cloud shadow and snow masks. although there is active research on accounting for variability within an image. BRDF characterizes how an object illuminated by a source. Figure 5. The images can then be combined to produce a single. such as the sun. and use a single correction algorithm for the entire image. methods that apply constant class signatures over images with variable atmospheric conditions should include atmospheric correction.. 26 http://landsat. Sun and sensor view angles also impact the effects of the atmosphere.Atmospherically-corrected images contain data representing surface reflectance.. While the corrections are applied to all the images and the resulting composite should therefore appear seamless. Several programs exist to perform atmospheric corrections over entire images. for example.8: Atmospheric effects on optical data. beginning in summer 2013. One local histogram matching could be applied to further reduce artifacts. “gap-filled” composite mosaic. will be available with top-of-atmosphere (TOA) reflectance corrections applied 26. developed by NASA. atmospheric artifacts may remain. cloud. as opposed to the digital numbers of the raw image data.0 – REMOTE SENSING OF LAND COVER CHANGE 100 . including those from the LDCM. additional algorithms based on. atmospheric pressure. Most atmospheric correction algorithms are applied to satellite images prior to mapping. a topography-dependent Rayleigh scattering correction and an aerosol optical thickness component based on Kaufman et al. LEDAPS uses information on water vapor. LOTRAN and 6S are the most common. appears brighter or darker depending on the angle of the source and the angle at which it is viewed by a satellite sensor. In addition. Conversely. (1997) to generate a surface reflectance value for each pixel. and requiring each atmospherically-corrected image to be classified separately and then combined. ozone. and several tools have been created to facilitate their application. Performing atmospheric correction depends on the image analysis approach used (Song et al. and this radiation can be blocked or scattered by clouds in addition to being affected by a “clear” atmosphere.php#3b REDD+ MRV MANUAL: CHAPTER 5. 2001). Alternatively. The electromagnetic radiation source is the sun.

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

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

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

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

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

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

The Kappa coefficient can also be calculated. . The user’s accuracy indicates how often a pixel was incorrectly assigned to a given class.6 Forest 94. The statistics include overall accuracy.930 4.5 percent. i. Foody.960 2. how often a pixel was incorrectly included in a class. Accuracy assessment Uncertainty is the error in a particular estimate.e.880 90 30 10.3: Example of an error matrix and resulting overall accuracy.000 = 94. or confusion matrix. In this hypothetical case. the columns contain verified land uses and the rows contain estimated uses from the classification. Pontius and Millones. Reference Forest Degraded Forest Non-forest Map total Forest 9. the producer’s accuracy for degraded forest is: 100 x 1.. The error matrix is generated by comparing the classification results with reference data.. An accuracy assessment allows errors in the map to be estimated and uncertainty quantified. Overall accuracy is the portion of the total number of correctly mapped pixels. In the example table. the results. 2002).000 16. In the example error matrix in Table 5. Performing an accuracy assessment of a thematic map represents a very important component of the process.0 – REMOTE SENSING OF LAND COVER CHANGE 107 . The user’s accuracy for the same class is: 100 x 1.1 REDD+ MRV MANUAL: CHAPTER 5.6 percent. This statistic is based on errors of omission.. how often a pixel was incorrectly omitted from the class.2 Forest 98. 1991). and thus accuracy assessment also refers to assessment of error or uncertainty. The main elements of an accuracy assessment are the error matrix. 2012. 2011.000 Reference total 9. and those off the diagonals are errors of omission and commission. The values along the diagonal are the number of correctly classified pixels. in this case the estimate of an area of change over time. and associated statistics (Congalton.8 Degraded Degraded Forest 92.040 4. i.000 Producer's accuracy (%) User's accuracy (%) Forest 99. thus providing additional explanation of.3.3 Table 5.000 Non-forest 10 60 3.890 / 2.040 = 92.890 40 2. but many articles highlight the limitations of this statistic (Olofsson et al. and validity to.000 Land-use map Degraded Forest 70 1. and the producer’s and user’s accuracy for each class in the product.e. This is based on errors of commission. the land use totals are the same as in the beginning time in Table 5. The producer’s accuracy indicates how often a pixel is correctly assigned to a specific class.3 forest 98. Accuracy equals one minus the error value.890 / 2.5 Non- Non-forest 98.

These include: i) ensuring the reference data are of a higher quality than the map data (for example. and iii) analysis.e. is estimated. use the information available in the matrix.0 – REMOTE SENSING OF LAND COVER CHANGE 108 . such as deforestation. (2013). An appropriate approach is the use of careful. They detail three separate steps that should be undertaken to complete an accuracy assessment including: i) the sampling design. Proportional sampling could then be focused in these strata to ensure each class is adequately represented in the validation analysis. together with the total area of each class identified in the map. Error matrices and accuracy assessments can also be extended to provide confidence interval (CI) information. 2010). cross-checked. ‘what is a suitable subset area to sample?’. and ii) Combining reference data sources (i. (2013) describe a process for creating CI bounds based on area-adjusted error matrices In addition. Multi-date accuracy assessments use the information available from two dates. which parts of the GHG inventory are contributing the greatest errors in GHG estimates and should be reviewed as part of the KCA.. to generate area-adjusted errors based on the proportional area of each class and errors identified in the matrix. While they differ from the reporting-table format of the IPCC. field and aerial surveys which can be particularly cost effective if resources are limited).. The examples in Table 5. Multiple interpreters can be used. It also allows one to assess where improvements are most needed. visual interpretation of a combination of very high resolution imagery.Both the land-use change and error matrices are common formats for reporting land-use change and errors. This will lead to stronger estimation of the overall uncertainties in estimated GHGs. Section 3.3 demonstrate error estimation in land-use cover for a single date. it is important to account for a rare class when developing a validation strategy (Stehman et al. i.2 from the UNFCC “Good Practice Guidance for LULUCF” (2003) for the propogation of errors.e. along with imagery used in the classification itself. However. ii) the response design.7 of the GFOI MGD (GFOI. and a stratification could be used to categorize areas of high change probability and low change probability. Olofsson et al. the data can easily be transferred. as this allows one to combine errors with errors in carbon stock for each stratum. This is especially important as it quantifies the confidence of a particular class. The GFOI MGD also provides two example approaches for performing an accuracy assessment and area estimation. such as degraded forest. which answers the question ‘are the maps and reference data in agreement?’. it can be combined with the error of the change in stocks per unit area estimated via field inventories. REDD+ MRV MANUAL: CHAPTER 5. This type of proportional sampling design ensures that adequate sampling occurs in sparser yet critical classes. (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 MRV programs should work with expert statisticians in developing strategies for validation sampling and combining information on uncertainties in AD with those in EFs. understanding that evaluation of the entire map is not possible. provides guidance on considerations for generating reference data and performing an accuracy assessment.2. thus providing very pertinent additional information. Error-adjusted area estimates. The sampling design could focus on areas of deforestation identified in the map. It is useful to calculate these statistics for the different strata in a study area. countries must estimate errors in land-use change estimates over time.. Accuracy assessment is very important. 2013). and the consistency of their interpretation can indicate confidence of the validation data set itself. Manual interpretation of an image by an analyst is generally considered higher quality than an automatic classification algorithm). Field surveys can be valuable for the classes that are the most difficult to interpret even with very high resolution imagery. Once the error for a particular area of change. such as those described in Olofsson et al. or AD. which answers the question. Olofsson et al. one stratified and one model-based approach. This can be done using Equation 5. similarly to the application described in Chapter 4. shown at the end of Chapter 3. which answers the question of ‘how to calculate accuracy and quantify uncertainty?’.

GOFC-GOLD. providing information on the distribution of aboveground biomass. deforestation. and the properties of the emitted radiation from active sensors are controlled and well known. in monitoring. and the increased use of field sampling to facilitate remote sensing product validation. A) B) Figure 5.5 EMERGING AREAS OF RESEARCH Several areas of particularly active research in support of REDD+ activities include: the mapping and monitoring of degradation. While the brightness variations in both are affected by terrain. In addition. they are able to penetrate clouds and are thus useful for monitoring in areas with persistent cloud cover. 1999. to map and monitor forest extent and characteristics. Multiple definitions of forest degradation exist. Asner et al. 2003. optical sensors. with estimates ranging from 20 to 50 percent of total land-use GHG emissions over large regions (see.1. forest fires (canopy and sub-canopy) and fuelwood collection. The IPCC’s definition of forest degradation is provided in section 5. Mapping and Monitoring of Degradation Forest degradation is a substantial contributor to GHG emissions from land-use change (Nepstad et al.. this is more noticeable in the latter REDD+ MRV MANUAL: CHAPTER 5. RADAR images are directly comparable over time.. Hirata et al. the use of other sources of remotely-sensed data. 2009. forest fires (canopy and sub-canopy) and fuelwood collection. First. Peru. Readers should also refer to the relevant sections of both the GOFC-GOLD SourceBook and the GFOI MGD. Figure 5.0 – REMOTE SENSING OF LAND COVER CHANGE 109 . while the PALSAR image in B) is cloud free. and degradation. GOFC-GOLD (2012) lists a range of human activities that result in forest degradation including selective logging..11 illustrates a detail of a Landsat image compared to a PALSAR satellite image for an area in San Martín.. such as RADAR. Houghton and Hacker. for example. 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. These different activities may require different monitoring approaches. Souza and Roberts. Conversely. 2005.5. 2005). such as RADAR. adding to the complexity of mapping and monitoring forest degradation.2. from Peru. Lambin et al. 2012. GOFC-GOLD (2012) presents a range of human activities that result in forest degradation including selective logging. Several characteristics make RADAR an attractive source of information for such applications. 1999. Stickler et al. for example. because RADAR sensors operate in longer wavelengths (generally 1cm to 1m) of the EM spectrum than. Mapping and monitoring of forest degradation remains challenging.11: Examples of Landsat and RADAR images. and a country should seek to understand the implications and applicability of different approaches. 2012). because the signals received by the sensor are less affected by atmospheric conditions. RADAR signals are also sensitive to the geometric properties of a forest. Ground observations in the Landsat data in A) are partly obscured by clouds.

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

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

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

ideally the sample should be well-distributed across the area to which the model will be applied. Field measurement plots can calibrate and validate prediction models. providing an equivalent range of spectral values within the area sampled via remote sensing. the Japanese Space Agency (JAXA) launched ALOS-PALSAR-2 in May 2014.0 – REMOTE SENSING OF LAND COVER CHANGE 113 . However. Therefore. Field measurements should also be collected as close as possible to the acquisition date of the remote sensing image. a sample of plots are measured and held in reserve. 2011).2 ha is recommended for use with airborne LiDAR to reduce model errors (Zolkos et al. and various kinds of LiDAR hold promise to more efficiently map biomass stocks. Additional evaluations are also needed to assess the utility of newer RADAR data sources in mountainous areas. In the latter case. though the range of advanced SAR processing techniques capitalizing on the availability of multi- polarimetric information is evolving. PALSAR and ENVISAT. This is of particular interest in the often vast and difficult-to-access landscapes contemplated under most national and jurisdictional REDD+ MRV. Finally. Use of Field Sampling To Facilitate Remote Sensing Product Validation Remote sensing 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. to optimize the design. and after model construction and application. from remotely-sensed data. Large. these data can be valuable in extrapolating field-based estimates over larger areas. reducing the impact of GPS inaccuracy (Mascaro et al. RADAR. The sample should achieve a roughly even distribution across the range of conditions of the remotely-sensed data. REDD+ MRV MANUAL: CHAPTER 5. This can be achieved via weighted random sampling. fixed-area plots are best suited for generating ground-reference data to compare with remotely-sensed data. SAR data generally yield less robust results than Landsat data for forest/non-forest classification in most studies. Although. over a wider spatial extent. Circular plot designs and a minimum plot size of 0. in particular high resolution satellite imagery.areas with a greater abundance of secondary forests may yield significantly lower accuracies. used to assess the accuracy of the predicted values at different scales. Sampling strategies differ however. for example. and this instrument again contains an L-band. However. which both provided fully polarimetric L-band data. and classification methods based on polarimetric decomposition are being developed. While these approaches hold promise. The same field measurement and QA/QC procedures described in Chapter 4 are also relevant in the collection of ground truth data. 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. they are still in the process of refinement. Uncertainty related to the long-term data continuity of space-borne RADAR systems could also prove a limiting factor for forest monitoring. 2013). Note that these preferred sample designs require that the remote sensing data already be in hand. as the objective is not to produce a field- based estimate but instead to model predictions. other unequal probability sampling approaches or a systematic sample approach. especially remote areas that are costly to access. such as the range of reflectance values. Larger plots facilitate the alignment of ground samples and remote sensing imagery... ground-reference data used to build prediction models for remotely-sensed data do not need to be strictly uniformly sampled. are no longer collecting.

Broadbent.C. and A.. M.wur. S. R.M. 2008a. B. Geneva. N. 2013. and D. Goetz. T. Townshend and R. Roy. Keller and J. A. Brodley.. Oliveira.A. Costa. Remote Sensing of Environment 114: 168-182. Status of land-cover classification accuracy assessment. 2010.A.. Evans. 1991. Sibley. E. M. REDD+ MRV MANUAL: CHAPTER 5. Pizurica and W. Huang. Sulla-Menashe. Science 310: 480–482. Pattern Recognition Letters 27: 294-300. GFOI.B.S. W.L. 2006. Telmer and T. Strahler. and X. Baccini. Borghys.P. Maximizing Land Cover Classification Accuracies Produced by Decision Trees at Continental to Global Scales. Laporte. Using ALOS/PALSAR and RADARSAT-2 to Map Land-cover and Seasonal Inundation in the Brazilian Pantanal. Available at http://www. Hansen M. D. 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.R. P.J. A sourcebook of methods and procedures for monitoring and reporting anthropogenic greenhouse gas emissions and removals associated with deforestation. S. E.6 REFERENCES Asner. gains and losses of carbon stocks in forests remaining forests. Global land cover classification at 1 km spatial resolution using a classification tree approach. Adusei. Clark. Clark. 2013. K. Benediktsson and J. Friedl. 2002.0 – REMOTE SENSING OF LAND COVER CHANGE 114 .A. 2004.gofcgold. Y. GOFC-GOLD. N. A. Sulla-Menashe. Hackler. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps.H. A. Altstatt.. D. Selective logging in the Brazilian Amazon.E.. Switzerland. 2005. Yvinec. Samanta and R.O. Hansen M. Friedl. J. Walker. Beck.N.E. 2010. Foody. Choosing a Forest Definition for the Clean Development Mechanism. Group on Earth Observations. A. Remote Sensing of Environment 37: 35-46. Transaction on Geosciences and Earth Observation 37(2): 969-977.C. Tan. J.J.R.O.P. 2006. Gislason. R. FAO.5. The Netherlands). Houghton.. J. A.G.A. Perneela.L. M. M. Friedl. 2006. Forests and Climate Change Working Paper 4. S. A method for integrating MODIS and Landsat data for systematic monitoring of forest cover and change in the Congo Basin. A.. G. Philips. MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets. M. Schneider. IEEE Journal of Selected Topics in applied Earth Observation and Remote Sensing 3: 560-570.. Sun. 2000. R. M. P. http://www. Justice and A. 2012. A Review of Assessing the Accuracy of Classifications of Remotely Sensed Data. D. Pattern Recognition Letters 27: 252–258.S DeFries. P. Wageningen University. Knapp. D.F. International Journal of Remote Sensing 21(6): 1331-1364. D. Lindquist. Random Forests for land cover classification.N.. Ramankutty. Small-footprint lidar estimation of sub-canopy elevation and tree height in a tropical rain forest landscape. GOFC-GOLD Report version COP19-2 (GOFC-GOLD Land Cover Project Office.C. M. C. Sveinsson. Nature Climate Change 2: 182-185. Silva. Roberts. Remote Sensing of Environment 91(1): 68-89. 1999. C. C. S. D. Congalton. Dubayah. Remote Sensing Environment 80: 185- 201. Remote Sensing Environment 112: 2495–251. T. G. Silva. A. Sohlberg. and forestation. Supervised feature-based classification of multi-channel SAR images.

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

M.J. R. M. H. Saatchi.F. 2009. C. A.T. G. Remote Sensing Environment 148: 45-57. J. Large-scale impoverishment of Amazonian forests by logging and fire. Hyde. Hill. 2011. Mather. 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. Lewis. Remote Sensing Environment 115: 2861–2873.. A. Mapping tropical forest biomass with radar and spaceborne LiDAR in Lopé National Park. P.. Brooks. Investigating RADAR-LiDAR synergy in a North Carolina pine forest. Estimating Quebec provincial forest resources using ICESat/GLAS. M. Rogan. Boudreau.M. Sonk´e. Leal. Edwards 2007. Meir.L. S. Woodhouse. J. T. Stehman. Meir. E. Williams. N. Schlesinger. Stow. Skole. M. 2005. Puyravaud. D. Williams.. Woodhouse. T. P. I.H. Miller. 1999. R. Levian and C.M. Rignot. Woodhouse P.S. Franklin. Campbell. 2009. Potter. E. Standardizing the calculation of the annual rate of deforestation. Incorporating the Downscaled Landsat TM Thermal Band in Land-cover Classification using Random Forest. G. and R. C. Nepstad.A Lefsky.S. T. M. Quantifying small-scale deforestation and forest degradation in African woodlands using RADAR imagery. P.S. Ryan. Ryan. Global Change Biology 18(1): 243-257. Mendoza. Nature 398: 505-508. Gobakken and G. C. S. Grace. J. Woodhouse and M. Pardo-Iguzquiza. T.T. S. P. Woodcock. Olofsson. Lewis. T. Foody.G. E. Mitchard.A. Congalton. A. 2012. Land-Cover Change Monitoring with Classification Trees Using Landsat TM and Ancillary Data.E. Moutinho. K.M. Naesset. Cochrane and V. Good practices for estimating area and assessing accuracy of land change. C. M. T. Verissimo.F. B. Stehman and C.E. E.L. L. Jeffery. Canadian Journal of Forestry Research 39: 862-881. Using satellite RADAR backscatter to predict aboveground woody biomass: A consistent relationship across four different African landscapes. Saatchi. Mitchard.R.. V. Photogrammetric Engineering & Remote Sensing 78(2): 129-137. and P.. Pal. E. T. 2003. L. M. Woollen.M. C. Collins. Lima. Remote Sensing Environment 129: 122-131. 1997. Foody. G. and Meir. An assessment of the effectiveness of decision tree methods for land cover classification. J. Abernethy. Rodriguez-Galiano. Geophysical Research Letters 36. C. Wulder 2014. and W. Ghimire.. E.A. Stahl.H. J.. M.J. Imhoff. Nobre. Random forest classifier for remote sensing classification. Johnson. 2003. K. Gregoire. Remote Sensing Environment 59: 167-179. S. P. S.M. Chica-Olmo.L. W. Margolis. Lewis. Alencar. Cassells. Emessiene. E. Remote Sensing Environment 110: 98–108 Nelson. M. International Journal of Remote Sensing 26(1): 217-222. I. Mapping deforestation and secondary growth in Rondonia. Brazil.Mitchard. Woodcock 2013. 2012. M. Nangendo. A. J. Ghee. Feldpausch and P. and P. I.. Herold. Fischer 2003. D. Lefebvre.E. 2012. Remote Sensing Environment 86: 554-565. G. Ribeiro. E. S. E.V. Photogrammetric Engineering & Remote Sensing 69(7): 793-804. I. Biogeoscience 9: 179-191. S. E. Forest Ecology and Management 177: 593-596. P. Rowland.V. Gabon: overcoming problems of high biomass and persistent cloud. Mitchard.G.. T. using imaging RADAR and thematic mapper data. S. B. P. S. Salas and D. Pal. Feldpausch. Olofsson. S. A. Saatchi. REDD+ MRV MANUAL: CHAPTER 5. C.0 – REMOTE SENSING OF LAND COVER CHANGE 116 . White..P. M. Making better use of accuracy data in land change studies: Estimating accuracy and area and quantifying uncertainty using stratified estimate.H. Nelson R.. B.

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

100m thermal roads. 5m panchromatic (2.5. encroachment. roads. Space 1. 185km 15m panchromatic 16-21 days Yes yes deforestation. canopy gaps. log ponds SPOT-6 Airbus Defense & 60km 6m multispectral Varies No Yes deforestation. NASA 30m multispectral encroachment. with SLC-off deforestation. 1b Space multispectral logging roads REDD+ MRV MANUAL: CHAPTER 5. 165km 15m panchromatic 16-21 days Yes Yes.NASA 60km 15m multispectral Varies No Partial (no SWIR deforestation. roads. log ponds ASTER Terra .5m encroachment.7m panchromatic 26 days Yes Yes skid trails. interpolated) roads. log ponds Landsat 8 OLI LDCM . 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-5 CNES 60km 20m multispectral Varies No Yes deforestation. log ponds CCD CBERS-2B – INPE 113km 20m multispectral 26 days Yes Yes deforestation. canopy gaps. NASA 30m multispectral gaps encroachment. HRC CBERS-2B – INPE 27km 2.5m panchromatic encroachment.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 . GeoEye nadir 1m panchromatic (1) illegal fishing vessels / logging vehicles Pleiades 1a and Airbus Defense & 20km 50cm panchromatic 2m Varies No Yes Skid trails.0 – REMOTE SENSING OF LAND COVER CHANGE 118 . illegal fishing vessels MODIS Terra / Aqua – 2330km 250m visible 4 times per Yes Yes fires. channels) encroachment. 60m thermal roads.3km @ 4m multispectral (4) Varies No Yes skid trails.

1km 0.7m SWIR (8) logging vehicles PALSAR-2 PALSAR-JAXA varies 1-3 m Spotlight 15 days Varies Yes deforestation.4m multispectral (4) illegal fishing vessels / logging vehicles RapidEye BlackBridge 77km 6.24m multispectral (8) illegal fishing vessels / 3. GeoEye (resampled to 0.5m multispectral 1-5 days No Yes Deforestation. 25 m quad pol standard roads. log ponds REDD+ MRV MANUAL: CHAPTER 5. skid trails WorldView-2 WorldView-2 Varies 0.quad pol roads.0 – REMOTE SENSING OF LAND COVER CHANGE 119 . Varies 0. DigitalGlobe 1.all mode No deforestation. DigitalGlobe 2. canopy gaps. log ponds 30 m quad pol 100 m Scan SAR WorldView-3 WorldView – 13. canopy gaps.65m multispectral (4) logging vehicles QuickBird QuickBird – Varies 0. canopy gaps. canopy gaps.4m panchromatic (1) Varies No Yes skid trails.GeoEye-1 GeoEye-1 .6m (1) panchromatic Varies No Yes skid trails. 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. 3-10 m Stripmap encroachment. roads. Constellation encroachment. 19 m dual pol No.5m) illegal fishing vessels / 1.5m panchromatic Varies No Yes skid trails.31m panchromatic (1) < 1 day No Yes skid trails. 100 m Scan SAR roads. DigitalGlobe 1.8m multispectral illegal fishing vessels / logging vehicles Radarsat-2 CSA varies 8 m quad pol fine 24 days varies Yes deforestation.

in fires calibration REDD+ MRV MANUAL: CHAPTER 5. 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).Future Missions Sensor Satellite – Swath Resolution Repeat Cycle Systematic Operational Monitoring Agency Width Acquisitions Status Applications SPOT-7 Airbus 60km 1.0m roads. deforestation. then (2B) VIIRS NPV – NASA 3000km 750m 2 times per day Yes recently launched.0 – REMOTE SENSING OF LAND COVER CHANGE 120 . Space 6.5m No Launched June 2014. Defense & panchromatic (1) in calibration encroachment.

0 – REMOTE SENSING OF LAND COVER CHANGE 121 Software EXELIS: ENVI http://www. http://www. Open source GIS and remote sensing software http://biodiversityinformatics.8 SELECTED RESOURCES Online guides and other materials United Nations Space Science and Technology: American Museum of Natural History (AMNH) Biodiversity Informatics Random forests Software REDD+ MRV MANUAL: CHAPTER 5.amnh.esa. NASA Earth Observatory GRASS GIS PCI Geomatics http://www.pcigeomatics.eo.aspx INTERGRAPH: ERDAS Imagine http://geospatial.php?section_id=17 European Space Agency Earthnet Systems for World Surveillance.amnh.html Biodiversity Informatics Facility IDRISI GIS and Image Processing Software http://www.clarklabs.cfm Quantum GIS http://www.qgis.

Map Data access USGS Earth Resources Observation and Science Center (EROS) (Range of Landsat and CBERS Rulequest data mining tools.berkeley. Ikonos.eo.1 RAT – Radar Tools http://radartools. as well as various ASTER and MODIS products) USGS LandsatLook Viewer http://landsatlook. ESA – polsarpro (Polarmetric SAR Data Processing and Educational Tool) http://earth.html R statistical language (Enables searching of both LandsatLook images & Level1 Landsat data) National Research Institute (INPE) of Brazil http://www.usgs.0 – REMOTE SENSING OF LAND COVER CHANGE 122 . and MODIS) RapidEye Catalog Global Land Survey (GLS) REDD+ MRV MANUAL: CHAPTER 5. See5 classification software http://www.http://www. SAR Tool Kit NEST – Next ESA SAR toolbox (Landsat Archive. as well as various MODIS products) Global Land Survey (GLS) 2005 products: Global Land-cover Facility http://www.stat. (Range of data sources including the Landsat archive and selected imagery for a range of instruments including Open source Alaska Satellite Facility .alaska. General remote sensing: Canada Centre of Remote Sensing http://www. NASA fundamentals of remote sensing http://gcmd. S. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Sulla-Menashe. Zutta. Proceedings of the National Academy of Sciences 14.0 – REMOTE SENSING OF LAND COVER CHANGE 123 . Goetz.L. Buermann. E. These may be useful for national stratification of field sampling of biomass in a MRV Two maps of global forest biomass. Sun.L.fas.doi. M. Petrova.R. Laporte.whrc.jpl. A.A. Brown. (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. produced by the University of Maryland and distributed by Google.nrcan. D.appspot. http://carbon.html Saatchi S. B.108(24):9899-904. remote-sensing/11740 USGS Change-tracking tool http://pubs. M.ersdac. Beck. Harris. A. based on a suite of satellite data inputs. http://earthenginepartners. Nature Climate Change http://dx. Hackler. W.nasa. Lewis.S. Baccini. S.A. Lefsky.1038/NCLIMATE1354 Pantropical National Level Carbon Stock Dataset http://www.S.spotimage. Silman and A. calibrated with plot data. R.html REDD+ MRV MANUAL: CHAPTER 5. Friedl. Mitchard. P.aster. 2011. Morel. S J. Samanta and R. Walker.gc.nasa. W. 2012. L. S. J. with a 1-km resolution. Tutorials The remote sensing tutorial: Federation of American Scientists (FAS) http://www.(Access to the BlackBridge RapidEye image archive) SPOT Catalog http://catalog. M. W. Dubayah. T. Benchmark map of forest carbon stocks in tropical regions across three (Access to the SPOT satellite archive) Earth Remote Sensing Data Analysis Center (ERSDAC) http://imsweb. REDD+ MRV MANUAL: CHAPTER 5.esa.1.html An introduction to remote sensing CSIRO http://www.html NOAA’s Satellite and Information Service: Learning About Satellites and Remote Sensing An introduction to radar remote sensing: Canada Centre of Remote Sensing Radar polarimetry: Canada Centre of Remote Sensing http://www.noaa.nrcan.cmis.0 – REMOTE SENSING OF LAND COVER CHANGE 124 .Introduction to remote sensing SAR-Land-Applications-Tutorial/sar_land_apps_1_theory.Virtual Hawaii ESA’s RADAR Tutorial ESA’s Synthetic Aperture radar: Land applications tutorial

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