5.2 Implications and topics for future research
5.2.1 Forest Remote Sensing
The results of this dissertation reflect advances in forest remote sensing over the previous century. Consistent data time series, free access to imagery, and cloud computing allow for complex analysis over the Amazon Ecoregion. The procedures used here were adapted from previous research, including
spectral mixture analysis and calculation of NDFI (Adams et al., 1995; Souza et al., 2005), structural break detection (Kennedy et al., 2010; Verbesselt et al., 2010; Zhu and Woodcock, 2014), large area forest monitoring on GEE (Hansen et al., 2013), area estimation through statistical inference
(McRoberts, 2011; Olofsson et al., 2014; Stehman, 2013), sample
interpretation (Stehman, 1999; Tyukavina et al., 2017a), and land cover classification (Friedl et al., 2002; Pal, 2005). Access to these tools and data have facilitated methodologies, such as CODED, that are designed to detect subtle changes in forest condition.
While designed in Rondônia and applied in the Amazon Ecoregion, it is likely that CODED is applicable to other forest ecosystems. CODED has already been implemented in Guatemala (Bullock et al., In Press) and the Republic of Georgia (Chen et al., Unpublished), and tested in Uruguay, Paraguay, Costa Rica, and Laos. While the Amazon Ecoregion is relatively ecologically homogenous and therefore could use a single set a parameters for
the change detection, expanding to diverse landscapes will require adjustments to the algorithm and different parameterization.
Currently, we are using a set of endmembers that were calculated for the Amazon (Souza et al., 2005). However, these endmembers will not accurately characterize other landscapes if there are drastic ecological differences. Future research should evaluate approaches to calculating endmembers. A possible approach is to use a multi-endmember model that allows for different endmembers for each observation, although that has yet to be tested with time series data (Roberts et al., 1998). The sensitivity of CODED to disturbance should also be compared for differing ecological and climatic conditions.
Spectral mixture analysis improves the sensitivity of CODED to partial canopy damage. However, it is likely that detection of smaller disturbance patches can be achieved using higher resolution data, such as Sentinel 2. The spectral bands are slightly different for Landsat and Sentinel 2, indicating that the current implementation could not be directly
implemented on Sentinel 2 data. Additionally, CODED requires multiple years of data for the training period before change can be detected and, since Sentinel 2-A was first launched in 2015, could only be used for monitoring disturbance after 2016.
The current implementation of CODED relies on spectral and temporal metrics to detect change. However, it is possible that disturbance can be more accurately detected by incorporating spatial information. The LandTrendr forest monitoring algorithm uses spatial segmentation to find congruent patches of disturbance (Kennedy et al., 2010). The resulting change maps contain clusters that represent specific change events. I believe a similar approach could be advantageous for change monitoring with CODED. The resulting change map would contain less “salt and pepper” noise, and the shape of the patches (e.g. patch size or roundness) could be useful for classifying disturbance type.
During this dissertation I compared change maps created using CODED to field data on dates and locations of fire and selective logging presented in Longo et al. (2016) and Rappaport et al. (2018). Whiles overall the maps corresponded well to the field data, there were notable examples of errors of omission in the CODED change maps. The statistical estimators used in this dissertation accounted for these errors when reporting areas. However, these data will be useful for calibrating the CODED methodology and reducing errors of omission in the future. By reducing errors of omission the dataset will be more useful for spatial analysis. Additionally, as errors of omission increase uncertainty in the area estimates, a more accurate
For Chapter 3 of this dissertation I utilized a secondary change detection algorithm to locate possible errors of omission. The approach was based on an offline version of CUSUM and therefore detected change in a fundamentally different way than CODED. However, both change detection approaches utilize regression models, which requires a semi-consistent data time series. It is likely that neither approach would effectively detect
disturbance in regions with minimal cloud-free data. In such locations, it may be preferable to detect change using image composites or direct classification of multi-date imagery. Given the minimal computational requirements for performing change detection, I believe future methodology development should focus on ensemble algorithms that can effectively detect change under varying levels of data availability.
For this dissertation, I utilized unbiased statistical estimators to calculate area and variance. Consequently, the area estimates reported in this dissertation were generally consistent with “Good Practice” guidelines for reporting areas of disturbance (Penman et al., 2003). However, there is still a need for improvements on research designs that can efficiently
estimate areas with high precision. For example, in this dissertation I did not differentiate degradation and natural disturbances. However, with the
availability of medium-to-high resolution Sentinel 2 and Planet data, I believe this distinction can be determined in the future. This will be
necessary for programs such as REDD+, in which distinguishing
anthropogenic and natural disturbance is essential. I believe that activity data on degradation can be achieved using the CODED methodology to produce a change map, then deriving a sample and assigning disturbance type labels by comparing the samples to high-quality reference data. The activity data (or areas of specific disturbance types) could then be estimated with a post-stratified estimator (Olofsson et al., 2013b).
The results of this dissertation were the first attempt to estimate areas for the entire study region. The area estimates are considered unbiased and provide an estimate of uncertainty. However, sample-based estimates of area do not provide information about spatial patterns. It is likely that D/ND and deforestation are strongly correlated with human activity and development, and possible that they correspond to certain forest types or ecological
conditions. Comparison of multiple spatial variables (such as disturbance and population) is challenging using statistical inference, and therefore it would be useful to analyze trends directly with the mapped disturbance dataset. This analysis could give insight as to predictors of future disturbance, which would be useful for planning conservation initiatives.
5.2.2 Conservation and Restoration
In the 20th century, increasing ecological destruction and land degradation
has been met with efforts to conserve tropical forests ("World Conservation Strategy", 1980; Watson et al., 2014). Conservation in the Amazon has received particular international attention due to its’ importance in global natural systems and the extensive indigenous communities (Nepstad et al., 2006; Schwartzman and Zimmerman, 2005). The disturbance dataset produced in Chapter 3 would be useful for evaluating the effectiveness of protected areas across the Amazon. In particular, it can be used to determine whether disturbances in protected areas are being driven by land conversion through deforestation or D/ND.
The CODED methodology developed in Chapter 2 would be applicable for international carbon offset programs that promote conservation.
Currently, there are few countries that report emissions from degradation for REDD+ or other greenhouse gas reporting mechanisms (Milbank et al.,
2018). This research shows that it is possible to estimate area and emissions from D/ND that are unbiased and provide an estimate of uncertainty.
This dissertation found that the rates of deforestation and D/ND in the Amazon have changed since 1995. D/ND appears to be increasing due to droughts, resulting in a similar aboveground carbon loss as deforestation. While previous conservation initiatives have focused on large-scale
deforestation (Godar et al., 2014), our results stress the importance of prioritizing efforts to prevent or mitigate D/ND. Ultimately, restoration priorities are project-specific and deforestation and D/ND result in different ecological changes. Regardless, the effects of D/ND on Earth’s natural systems cannot be ignored when planning for long-term conservation initiatives.
While this dissertation details the extent of disturbed forests in the Amazon, those forests offer opportunities for restoration and resilience. Restoration of degraded forests can restore ecosystem services and sequester carbon from the atmosphere. Before this is realized, however, it is essential for disturbed forests to be mapped and quantified. The disturbance dataset produced in Chapter 3 can be used for targeting forests for restoration through projects such as Initiative 20 x 20 (Vergara et al., 2016).
5.2.3 The Carbon Cycle
The tropical forest carbon flux is considered one of the largest sources of uncertainties in the global carbon cycle (Houghton et al., 2001). While the Amazon is only part of the global tropical forest ecosystems, it contributes approximately 11-66% of the global carbon emissions from land cover change (Chapter 4). This dissertation supports previous estimates of emissions due to deforestation, but suggests that there is an unaccounted for source of emissions due to D/ND. In addition to the previously reported emissions from
deforestation, this dissertation suggests that there are an additional 0.12 Pg C/year emitted due to D/ND, equivalent to 13% global emissions from land use change.
The carbon model used in this dissertation was limited due to a lack of data to parameterize the model. Future research about the carbon dynamics of D/ND would greatly benefit from extensive field campaigns to disturbed forests. Ideally, forest conditions would be examined before and after a disturbance. Specifically, there is a need for information on emissions due to D/ND for the major forest carbon pools. It is possible that carbon loss from a disturbance is correlated with spectral change in corresponding Landsat images, which is recorded in CODED as the change magnitude. It is also possible that regeneration is modeled by the slope coefficient in the post- disturbance regression model. Put simply, there is value in comparing the change metrics calculated in CODED to field data on forest change.
5.2.4 The Amazon
This dissertation found that the estimated area of disturbance in the Amazon is 44-60% higher than comparable assessments. Consequently, the Amazon is closer to the theorized “tipping point” for hydrological collapse than
previously realized (Lovejoy and Nobre, 2018). The extreme droughts in 2005, 2010, and 2016 may be early indications of a drastic hydrologic response to disturbance and climate change. It is likely, however, that D/ND does not
have the same widespread effect as deforestation. More research is needed on the regional effect of D/ND on weather patterns, the hydrologic cycle, and biodiversity.
Aalders, E., Espejo, A., Kopperud, T., Reed, P., Shut, V., 2015. Verification of Interim REDD+ Performance Indicators under the Guyana-Norway REDD+ Partnership.
Achard, F., Beuchle, R., Mayaux, P., Stibig, H.J., Bodart, C., Brink, A., Carboni, S., Desclée, B., Donnay, F., Eva, H.D., Lupi, A., 2014.
Determination of tropical deforestation rates and related carbon losses from 1990 to 2010. Global Change Biology 20, 2540–2554.
Adams, J.B., Sabol, D.E., Kapos, V., Almeida Filho, R., Roberts, D.A., Smith, M.O., Gillespie, A.R., 1995. Classification of multispectral images based on fractions of endmembers: Application to land-cover change in the Brazilian Amazon. Remote Sensing of Environment 52, 137–154. doi:10.1016/0034-4257(94)00098-8
Alves, A.S., Pereira, J.L., Sousa, C.L., Soares, J. V., Yamaguchi, F., 1998. Classification of the Deforested Area in Central Rondônia, using TM Imagery. Simpósio Brasileiro de Sensoriamento Remoto 9 1421–1432. Alves, D., Soares, J.V., Amaral, S., Mello, E., Almeida, S., da Silva, O.F.,
Silveira, A., 1997. Biomass of primary and secondary vegetation in Rondônia, Western Brazilian Amazon. Global Change Biology 3, 451– 461.
Aragão, L.E.O.C., Anderson, L.O., Fonseca, M.G., Rosan, T.M., Vedovato, L.B., Wagner, F.H., Silva, C.V.J., Silva Junior, C.H.L., Arai, E., Aguiar, A.P., Barlow, J., Berenguer, E., Deeter, M.N., Domingues, L.G., Gatti, L., Gloor, M., Malhi, Y., Marengo, J.A., Miller, J.B., Phillips, O.L., Saatchi, S., 2018. 21st Century drought-related fires counteract the decline of Amazon deforestation carbon emissions. Nature Communications 9, 1– 12. doi:10.1038/s41467-017-02771-y
Arévalo, P., Olofsson, P., Woodcock, C.E., 2019. Continuous monitoring of land change activities and post-disturbance dynamics from Landsat time series: A test methodology for REDD+ reporting. Remote Sensing of Environment. doi:10.1016/j.rse.2019.01.013
dynamics and drivers in different forest types in Latin America: Three decades of studies (1980–2010). Global Environmental Change 46, 139– 147. doi:10.1016/j.gloenvcha.2017.09.002
Asner, G.P., Kellner, J.R., Kennedy-Bowdoin, T., Knapp, D.E., Anderson, C., Martin, R.E., 2013. Forest Canopy Gap Distributions in the Southern Peruvian Amazon. Public Library of Science One 8, 1–10.
Asner, G.P., Knapp, D.E., Broadbent, E.N., Oliveira, P.J.C., KELLER, M., Silva, J.N., 2005. Selective logging in the Brazilian Amazon. Science 310, 480–482. doi:DOI 10.1126/science.1118051
Asner, G.P., Powell, G.V.N., Mascaro, J., Knapp, D.E., Clark, J.K., Jacobson, J., Kennedy-Bowdoin, T., Balaji, A., Paez-Acosta, G., Victoria, E., Secada, L., Valqui, M., Hughes, R.F., 2010. High-resolution forest carbon stocks and emissions in the Amazon. Proceedings of the National Academy of Sciences 107, 16738–16742. doi:10.1073/pnas.1004875107
Avitabile, V., Herold, M., Heuvelink, G.B.M., Simon, L., Phillips, O.L., Asner, G.P., Armston, J., Peter, S., Banin, L., Bayol, N., Berry, N.J., 2016. An integrated pan-tropical biomass map using multiple reference datasets. Global Change Biology 22, 1406–1420. doi:10.1111/gcb.13139
B. Adams, J., O. Smith, M., E. Johnson, P., 1986. Spectral Mixture Modeling: A New Analysis of Rock and Soil Types at the Viking Lander 1 Site. Journal of Geophysical Research 91.
Baccini, A., Goetz, S.J., Walker, W.S., Laporte, N.T., Sun, M., Sulla-Menashe, D., Hackler, J., Beck, P.S.A., Dubayah, R., Friedl, M.A., Samanta, S., Houghton, R.A., 2012. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nature Climate Change 2, 182–185. doi:10.1038/nclimate1354
Baccini, A., Walker, W., Carvalho, L., Farina, M., Sulla-Menashe, D.,
Houghton, R.A., 2017. Tropical forests are a net carbon source based on aboveground measurements of gain and loss. Science 358, 230–234. doi:10.1126/science.aam5962
Barlow, J., Lennox, G.D., Ferreira, J., Berenguer, E., Lees, A.C., Nally, R. Mac, Thomson, J.R., Ferraz, S.F.D.B., Louzada, J., Oliveira, V.H.F., Parry, L., Ribeiro De Castro Solar, R., Vieira, I.C.G., Aragaõ, L.E.O.C., Begotti, R.A., Braga, R.F., Cardoso, T.M., Jr, R.C.D.O., Souza, C.M.,
Moura, N.G., Nunes, S.S., Siqueira, J.V., Pardini, R., Silveira, J.M., Vaz- De-Mello, F.Z., Veiga, R.C.S., Venturieri, A., Gardner, T.A., 2016.
Anthropogenic disturbance in tropical forests can double biodiversity loss from deforestation. Nature 535, 144–147. doi:10.1038/nature18326
Barlow, J., Peres, C.A., Lagan, B.O., Haugaasen, T., 2003. Large tree mortality and the decline of forest biomass following Amazonian
wildfires. Ecology Letters 6, 6–8. doi:10.1046/j.1461-0248.2003.00394.x Barona, E., Ramankutty, N., Hyman, G., Coomes, O.T., 2010. The role of
pasture and soybean in deforestation of the Brazilian Amazon.
Environmental Research Letters 5. doi:10.1088/1748-9326/5/2/024002 Berenguer, E., Ferreira, J., Gardner, T.A., Aragão, L.E.O.C., De Camargo,
P.B., Cerri, C.E., Durigan, M., De Oliveira, R.C., Vieira, I.C.G., Barlow, J., 2014. A large-scale field assessment of carbon stocks in human- modified tropical forests. Global Change Biology 20, 3713–3726. doi:10.1111/gcb.12627
Brando, P.M., Balch, J.K., Nepstad, D.C., Morton, D.C., Putz, F.E., Coe, M.T., Silverio, D., Macedo, M.N., Davidson, E.A., Nobrega, C.C., Alencar, A., Soares-Filho, B.S., 2014. Abrupt increases in Amazonian tree mortality due to drought-fire interactions. Proceedings of the National Academy of Sciences 111, 6347–6352. doi:10.1073/pnas.1305499111
Brazil, M. of A., 1989. Alteração da Cobertura Vegetal Natural do Estado de Rondônia: Relatório Técnico.
Brondizio, E.S., Moran, E.F., 2012. Level-dependent deforestation trajectories in the Brazilian Amazon from 1970 to 2001. Population and Environment 34, 69–85. doi:10.1007/s11111-011-0159-8
Brooks, E.B., Wynne, R.H., Thomas, V.A., Blinn, C.E., Coulston, J.W., 2014. On-the-fly massively multitemporal change detection using statistical quality control charts and landsat data. IEEE Transactions on
Geoscience and Remote Sensing 52, 3316–3332. doi:10.1109/TGRS.2013.2272545
Browder, J.O., Godfrey, B.J., 1997. Rainforest cities: Urbanization, development, and globalization of the Brazilian Amazon. Columbia University Press.
Constancy of Regression Relationships over Time. Journal of the Royal Statistical Society 37, 149–192.
Bullock, E., Nolte, C., Segovia, A., Woodcock, C., 2019. Ongoing Forest Loss in Guatemala’s Protected Areas. Remote Sensing of Ecology and
Bullock, E.L., Woodcock, C.E., Holden, C.E., 2019. Improved change
monitoring using an ensemble of time series algorithms. Remote Sensing of Environment 0–1. doi:10.1016/j.rse.2019.04.018
Bullock, E.L., Woodcock, C.E., Olofsson, P., 2018. Monitoring tropical forest degradation using spectral unmixing and Landsat time series analysis. Remote Sensing of Environment. doi:10.1016/j.rse.2018.11.011
Cameron, A.D., 2002. Importance of early selective thinning in the
development of long-term stand stability and improved log quality: A review. Forestry 75, 25–35. doi:10.1093/forestry/75.1.25
Chen, S., Olofsson, P., Woodcock, C., Bullock, E., n.d. Carbon Dynamics of Forest Degradation in the Republic of Georgia.
Chomentowski, W., Fernandes, C., Lisboa, A., 2005. Conservation units: a new deforestation frontier in the Amazonian state of Rondônia,. Environmental Conservation 32, 149–155.
Cochran, W.G., 1977. Sampling techniques. New York: John Wiley and Sons. Cochran, W.G., 1977. Sampling Techniques. 1977, New York: John Wiley and
Cochrane, M.A., Alencar, A., Schulze, M.D., Souza, C.M., Nepstad, D.C., Lefebvre, P., Davidson, E.A., Saldarriaga, J.G., West, D.C., Sanford, R.L., Saldarriaga, J.G., Clark, K., Uhl, C., Herrera, R., Turcq, B., Meggers, B.J., Cochrane, M.A., Schulze, M.D., Holdsworth, A.R., Uhl, C., Peterson, D.L., Ryan, K.C., 1999. Positive feedbacks in the fire dynamic of closed canopy tropical forests. Science (New York, N.Y.) 284, 1832–5.
Cochrane, M.A., Schulze, M.D., 1999. Fire as a recurrent event in tropical forests of the eastern Amazon: Effects on forest structure, biomass, and species composition. Biotropica 31, 2–16. doi:10.1111/j.1744-
Connette, G., Oswald, P., Songer, M., Leimgruber, P., 2016. Mapping Distinct Forest Types Improves Overall Forest Identification Based on Multi- Spectral Landsat Imagery for Myanmar ’s Tanintharyi Region. Remote Sensing 8. doi:10.3390/rs8110882
Dejesusparada, N., Sonnenburg, C.R., 1978. Overview of Brazilian remote sensing activities.
Duffy, P.B., Brando, P., Asner, G.P., Field, C.B., 2015. Projections of future meteorological drought and wet periods in the Amazon. Proceedings of the National Academy of Sciences 112, 13172–13177.
Edwards, D.P., Larsen, T.H., Docherty, T.D.S., Ansell, F.A., Hsu, W.W., Derhé, M.A., Hamer, K.C., Wilcove, D.S., 2011. Degraded lands worth protecting: the biological importance of Southeast Asia’s repeatedly logged forests. Proceedings. Biological sciences / The Royal Society 278, 82–90. doi:10.1098/rspb.2010.1062
Eiten, G., 1972. The Cerrado Vegetation of Brazil. The Botanical Review 38, 201–341.
EMBRAPA, 2014. Sustainable Landscape Brazil.
Environment, B.M. of the, 2018. Brazil’s submission of a Forest Reference Emission Level (FREL) for reducing emissions from deforestation in the Amazonia biome for REDD + results-based payments under the
Environment, B.M. of the, 2017. Brazil’s Forest Reference Emission Level for Reducing Emissions from Deforestation in the Cerrado biome for Results- based Payments for REDD+ under the United Nations Framework
Convention on Climate Change. FREL Cerrado.
Environment, B.M. of the, 2004. Plan for the prevention and control of deforestation in the legal Amazon. Report of the Presidency of the Republic, Civil House.
Erfanian, A., Wang, G., Fomenko, L., 2017. Unprecedented drought over tropical South America in 2016: Significantly under-predicted by tropical SST. Scientific Reports 7, 5811. doi:10.1038/s41598-017-05373-2
Eugenio Y. Arima, Paulo Barreto b, Elis Araújo b, Britaldo Soares-Filho, 2014. Public policies can reduce tropical deforestation: Lessons and
challenges from Brazil. Land Use Policy 41, 465–473.
FAO, 2017. From reference levels to results reporting: REDD+ under the UNFCCC, Forests and Climate Change Working Paper.
Fearnside, P.M., 1990. The Rate and Extent of Deforestation in Brazilian Amazonia, Environmental Conservation.
Fearnside, P.M., 1982. Deforestation in the Brazilian Amazon Region with what intensity is it occurring? Acta Amazonica 12, 579–590.
Franke, J., Navratil, P., Keuck, V., Peterson, K., Siegert, F., 2012. Monitoring fire and selective logging activities in tropical peat swamp forests. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 5, 1811–1820. doi:10.1109/JSTARS.2012.2202638
Franklin, S.E., Wulder, M.A., 2008. Progress in Physical Geography Remote sensing methods in medium spatial resolution satellite data land cover classification of large areas. Progress in Physical Geography 2, 173–205. doi:10.1191/0309133302pp332ra
Friedl, M.A., McIver, D.K., Hodges, J.C.F., Zhang, X.Y., Muchoney, D., Strahler, A.H., Woodcock, C.E., Gopal, S., Schneider, A., Cooper, A., others, 2002. Global land cover mapping from MODIS: algorithms and early results. Remote Sensing of Environment 83, 287–302.
Fund, I.U. for C. of N. and W.W., 1980. World conservation strategy: Living resource conservation for sustainable development. Gland, Switzerland: IUCN.
Gatti, L. V., Gloor, M., Miller, J.B., Doughty, C.E., Malhi, Y., Domingues, L.G., Basso, L.S., Martinewski, A., Correia, C.S.C., Borges, V.F., Freitas, S., Braz, R., Anderson, L.O., Rocha, H., Grace, J., Phillips, O.L., Lloyd, J., 2014. Drought sensitivity of Amazonian carbon balance revealed by
atmospheric measurements. Nature 506, 76–80. doi:10.1038/nature12957 GFOI, 2013. 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: Pub: Group on Earth Observations Observations Initiative 2, 226.
Godar, J., Gardner, T.A., Tizado, E.J., Pacheco, P., 2014. Actor-specific contributions to the deforestation slowdown in the Brazilian Amazon.
Proceedings of the National Academy of Sciences 111, 15591–15596. doi:10.1073/pnas.1322825111
Goetz, S.J., Baccini, A., Laporte, N.T., Johns, T., Walker, W., Kellndorfer, J., Houghton, R.A., Sun, M., 2009. Mapping and monitoring carbon stocks with satellite observations: a comparison of methods. Carbon balance and management 4, 2. doi:10.1186/1750-0680-4-2
Goetz, S.J., Hansen, M., Houghton, R.A., Walker, W., Laporte, N., Busch, J., Hansen, M., Laporte, N., Busch, J., 2014. Measurement and Monitoring for REDD+: The Needs, Current Technological Capabilities, and Future Potential. Climate and Forest Paper Series 49.
Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., Moore, R., 2017. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment. doi:10.1016/j.rse.2017.06.031 Hansen, M.C., Krylov, A., Tyukavina, A., Potapov, P. V, Turubanova, S.,
Zutta, B., Ifo, S., Margono, B., Stolle, F., Moore, R., 2016. Humid tropical forest disturbance alerts using Landsat data. Environmental Research Letters 11, 34008. doi:10.1088/1748-9326/11/3/034008
Hansen, M.C., Potapov, P., Tyukavina, A., 2019. Comment on “Tropical forests are a net carbon source based on aboveground measurements of gain and loss.” Science 363, eaar3629.
Hansen, M.C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L., Justice, C.O., Townshend, J.R.G., 2013. High-resolution global maps of 21st-century forest cover change. Science 342, 850–853. doi:10.1126/science.1244693
Harris, N.L., Brown, S., Hagen, S.C., Saatchi, S.S., Petrova, S., Salas, W., Hansen, M.C., Potapov, P. V., Lotsch, A., 2012. Baseline map of carbon