Regional Mapping and Spatial Distribution Analysis of Canopy Palms in an Amazon Forest Using Deep Learning and VHR Images

Mapping plant species at the regional scale to provide information for ecologists and forest managers is a challenge for the remote sensing community. Here, we use a deep learning algorithm called U-net and very high-resolution multispectral images (0.5 m) from GeoEye satellite to identify, segment and map canopy palms over ∼3000 km 2 of Amazonian forest. The map was used to analyse the spatial distribution of canopy palm trees and its relation to human disturbance and edaphic conditions. The overall accuracy of the map was 95.5% and the F1-score was 0.7. Canopy palm trees covered 6.4% of the forest canopy and were distributed in more than two million patches that can represent one or more individuals. The density of canopy palms is affected by human disturbance. The post-disturbance density in secondary forests seems to be related to the type of disturbance, being higher in abandoned pasture areas and lower in forests that have been cut once and abandoned. Additionally, analysis of palm trees’ distribution shows that their abundance is controlled naturally by local soil water content, avoiding both flooded and waterlogged areas near rivers and dry areas on the top of the hills. They show two preferential habitats, in the low elevation above the large rivers, and in the slope directly below the hill tops. Overall, their distribution over the region indicates a relatively pristine landscape, albeit within a forest that is critically endangered because of its location between two deforestation fronts and because of illegal cutting. New tree species distribution data, such as the map of all adult canopy palms produced in this work, are urgently needed to support Amazon species inventory and to understand their distribution and diversity.

[1]  M. Tabarelli,et al.  Seed Dispersal of the Palm Attalea oleifera in a Remnant of the Brazilian Atlantic Forest1 , 2004 .

[2]  N. Smith.  Palms and People in the Amazon , 2014 .

[3]  Weijia Li,et al.  Deep Learning Based Oil Palm Tree Detection and Counting for High-Resolution Remote Sensing Images , 2016, Remote. Sens..

[4]  A. Lima,et al.  Gradients within gradients: The mesoscale distribution patterns of palms in a central Amazonian forest , 2009 .

[5]  W. Laurance.  Conserving the hottest of the hotspots , 2009 .

[6]  K. Ickes,et al.  Alternative successional pathways in the Amazon Basin , 2001 .

[7]  H. Tuomisto,et al.  Palm distribution patterns in Amazonian rainforests: What is the role of topographic variation? , 2004 .

[8]  M. Keller,et al.  Seeing the forest beyond the trees , 2015 .

[9]  Weijia Li,et al.  Large-Scale Oil Palm Tree Detection from High-Resolution Satellite Images Using Two-Stage Convolutional Neural Networks , 2018, Remote. Sens..

[10]  David Kenfack,et al.  An estimate of the number of tropical tree species , 2015, Proceedings of the National Academy of Sciences.

[11]  C. M. Peters,et al.  Oligarchic forests of economic plants in amazonia: utilization and conservation of an important tropical resource. , 1989, Conservation biology : the journal of the Society for Conservation Biology.

[12]  F. Kahn.  The distribution of palms as a function of local topography in Amazonian terra-firme forests , 1987, Experientia.

[13]  M. Hansen,et al.  The last frontiers of wilderness: Tracking loss of intact forest landscapes from 2000 to 2013 , 2017, Science Advances.

[14]  Nguyen Duy Hung,et al.  Deep learning based banana plant detection and counting using high-resolution red-green-blue (RGB) images collected from unmanned aerial vehicle (UAV) , 2019, PloS one.

[15]  J. Fraser,et al.  Crop Diversity on Anthropogenic Dark Earths in Central Amazonia , 2011 .

[16]  C. Barbosa,et al.  LBA-ECO LC-07 Wetland Extent, Vegetation, and Inundation: Lowland Amazon Basin , 2015 .

[17]  G. Shepard,et al.  Secondary forests on anthropogenic soils in Brazilian Amazonia conserve agrobiodiversity , 2010, Biodiversity and Conservation.

[18]  Philip G Brodrick,et al.  Uncovering Ecological Patterns with Convolutional Neural Networks. , 2019, Trends in ecology & evolution.

[19]  Lindsay A. Turnbull,et al.  Identification of 100 fundamental ecological questions , 2013 .

[20]  J. Svenning,et al.  On the role of microenvironmental heterogeneity in the ecology and diversification of neotropical rain-forest palms (Arecaceae) , 2008, The Botanical Review.

[21]  G. B. Williamson,et al.  Seed Rain and Advance Regeneration in Secondary Succession in the Brazilian Amazon , 2011 .

[22]  Maria Knadel,et al.  Environment versus dispersal in the assembly of western Amazonian palm communities , 2012 .

[23]  Yuliya Tarabalka,et al.  Mapping Atlantic rainforest degradation and regeneration history with indicator species using convolutional network , 2020, PloS one.

[24]  William Farfan-Rios,et al.  Ancient human disturbances may be skewing our understanding of Amazonian forests , 2017, Proceedings of the National Academy of Sciences.

[25]  S. Hubbell,et al.  Dynamics of the forest communities at Pasoh and Barro Colorado: comparing two 50-ha plots. , 1999, Philosophical transactions of the Royal Society of London. Series B, Biological sciences.

[26]  William I. Woods,et al.  The domestication of Amazonia before European conquest , 2015, Proceedings of the Royal Society B: Biological Sciences.

[27]  W. Denevan.  A Bluff Model of Riverine Settlement in Prehistoric Amazonia , 1996 .

[28]  P. Hietz,et al.  Climatic and edaphic controls over tropical forest diversity and vegetation carbon storage , 2020, Scientific Reports.

[29]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[30]  Tomlinson Pb.  The uniqueness of palms , 2006 .

[31]  D. Clark,et al.  Edaphic and Human Effects on Landscape‐Scale Distributions of Tropical Rain Forest Palms , 1995 .

[32]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[33]  Florentin Wörgötter,et al.  Large Scale Palm Tree Detection in High Resolution Satellite Images Using U-Net , 2019, Remote. Sens..

[34]  L. Aragão,et al.  Life cycle of bamboo in the southwestern Amazon and its relation to fire events , 2018, Biogeosciences.

[35]  Stephen P. Hubbell,et al.  Density-dependent adult recruitment in a low-density tropical tree , 2018, Proceedings of the National Academy of Sciences.

[36]  O. Phillips,et al.  Using the U‐net convolutional network to map forest types and disturbance in the Atlantic rainforest with very high resolution images , 2019, Remote Sensing in Ecology and Conservation.

[37]  Dr. Francis Kahn,et al.  Palms in Forest Ecosystems of Amazonia , 1992, Ecological Studies.

[38]  Kalle Ruokolainen,et al.  Modelling responses of western Amazonian palms to soil nutrients , 2017 .

[39]  Helmi Zulhaidi Mohd Shafri,et al.  Young and mature oil palm tree detection and counting using convolutional neural network deep learning method , 2019, International Journal of Remote Sensing.

[40]  J. Terborgh,et al.  Markedly divergent estimates of Amazon forest carbon density from ground plots and satellites , 2014, Global ecology and biogeography : a journal of macroecology.

[41]  Maciel Zortea,et al.  Oil-Palm Tree Detection in Aerial Images Combining Deep Learning Classifiers , 2018, IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium.

[42]  L. R. Dice.  Measures of the Amount of Ecologic Association Between Species , 1945 .

[43]  T. Libourel,et al.  Viability of the Babassu Palm Eco-socio-system in Brazil: The Challenges of Coviability , 2019, Coviability of Social and Ecological Systems: Reconnecting Mankind to the Biosphere in an Era of Global Change.

[44]  J. Terborgh,et al.  Hyperdominance in the Amazonian Tree Flora , 2013, Science.

[45]  Jordan M. Malof,et al.  Large-Scale Semantic Classification: Outcome of the First Year of Inria Aerial Image Labeling Benchmark , 2018, IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium.

[46]  D. A. King,et al.  Height-diameter allometry of tropical forest trees , 2010 .

[47]  Juliana Loureiro Almeida Campos,et al.  Population structure and fruit availability of the babassu palm ( Attalea speciosa Mart. ex Spreng) in human-dominated landscapes of the Northeast Region of Brazil , 2017 .

[48]  W. Baker,et al.  Tropical rain forest evolution: palms as a model group , 2013, BMC Biology.

[49]  Francisco Herrera,et al.  Deep-learning Versus OBIA for Scattered Shrub Detection with Google Earth Imagery: Ziziphus lotus as Case Study , 2017, Remote. Sens..

[50]  Peijun Li,et al.  Estimation of the Distribution of Tabebuia guayacan (Bignoniaceae) Using High-Resolution Remote Sensing Imagery , 2011, Sensors.

[51]  S. Ferrier,et al.  Wilderness areas halve the extinction risk of terrestrial biodiversity , 2019, Nature.

[52]  H. Balslev,et al.  Geographical ecology of the palms (Arecaceae): determinants of diversity and distributions across spatial scales. , 2011, Annals of botany.

[53]  Gerardo Flores,et al.  Identifying and Quantifying the Abundance of Economically Important Palms in Tropical Moist Forest Using UAV Imagery , 2019, Remote. Sens..

[54]  R. Mittermeier,et al.  Biodiversity hotspots for conservation priorities , 2000, Nature.

[55]  Julien Michel,et al.  Orfeo ToolBox: open source processing of remote sensing images , 2017, Open Geospatial Data, Software and Standards.

[56]  C. A. Mücher,et al.  Environmental science: Agree on biodiversity metrics to track from space , 2015, Nature.

[57]  F. Forest,et al.  Origin and global diversification patterns of tropical rain forests: inferences from a complete genus-level phylogeny of palms , 2011, BMC Biology.

[58]  Anis Koubaa,et al.  Deep-Learning-based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images , 2020, ArXiv.

[59]  Millenium Ecosystem Assessment.  Ecosystems and human well-being: synthesis , 2005 .