Big earth observation time series analysis for monitoring Brazilian agriculture

Abstract This paper presents innovative methods for using satellite image time series to produce land use and land cover classification over large areas in Brazil from 2001 to 2016. We used Moderate Resolution Imaging Spectroradiometer (MODIS) time series data to classify natural and human-transformed land areas in the state of Mato Grosso, Brazil’s agricultural frontier. Our hypothesis is that building high-dimensional spaces using all values of the time series, coupled with advanced statistical learning methods, is a robust and efficient approach for land cover classification of large data sets. We used the full depth of satellite image time series to create large dimensional spaces for statistical classification. The data consist of MODIS MOD13Q1 time series with 23 samples per year per pixel, and 4 bands (Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), near-infrared (nir) and mid-infrared (mir)). By taking a series of labelled time series, we fed a 92-dimensional attribute space into a support vector machine model. Using a 5-fold cross validation, we obtained an overall accuracy of 94% for discriminating among nine land cover classes: forest, cerrado, pasture, soybean-fallow, fallow-cotton, soybean-cotton, soybean-corn, soybean-millet, and soybean-sunflower. Producer and user accuracies for all classes were close to or better than 90%. The results highlight important trends in agricultural intensification in Mato Grosso. Double crop systems are now the most common production system in the state, sparing land from agricultural production. Pasture expansion and intensification has been less studied than crop expansion, although it has a stronger impact on deforestation and greenhouse gas (GHG) emissions. Our results point to a significant increase in the stocking rate in Mato Grosso and to the possible abandonment of pasture areas opened in the state’s frontier. The detailed land cover maps contribute to an assessment of the interplay between production and protection in the Brazilian Amazon and Cerrado biomes.

[1]  Damien Arvor,et al.  Classification of MODIS EVI time series for crop mapping in the state of Mato Grosso, Brazil , 2011 .

[2]  Damien Arvor,et al.  Mapping and spatial analysis of the soybean agricultural frontier in Mato Grosso, Brazil, using remote sensing data , 2013 .

[3]  Michael Stonebraker,et al.  SciDB: A Database Management System for Applications with Complex Analytics , 2013, Computing in Science & Engineering.

[4]  Mariana Belgiu,et al.  Random forest in remote sensing: A review of applications and future directions , 2016 .

[5]  Curtis E. Woodcock,et al.  From imagery to ecology: leveraging time series of all available Landsat observations to map and monitor ecosystem state and dynamics , 2016 .

[6]  Cesar Guerreiro Diniz,et al.  High spatial resolution land use and land cover mapping of the Brazilian Legal Amazon in 2008 using Landsat-5/TM and MODIS data , 2016 .

[7]  Victor Maus.  Time-Weighted Dynamic Time Warping for Satellite Image Time Series Analysis [R package dtwSat version 0.2.6] , 2020 .

[8]  Daniela M. Witten,et al.  An Introduction to Statistical Learning: with Applications in R , 2013 .

[9]  Gregory S. Biging,et al.  Automated mapping of soybean and corn using phenology , 2016 .

[10]  Per Jönsson,et al.  TIMESAT - a program for analyzing time-series of satellite sensor data , 2004, Comput. Geosci..

[11]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[12]  Jungho Im,et al.  Support vector machines in remote sensing: A review , 2011 .

[13]  R. DeFries,et al.  Decoupling of deforestation and soy production in the southern Amazon during the late 2000s , 2012, Proceedings of the National Academy of Sciences.

[14]  Zhiqiang Yang,et al.  Detecting trends in forest disturbance and recovery using yearly Landsat time series: 1. LandTrendr — Temporal segmentation algorithms , 2010 .

[15]  Luciano Vieira Dutra,et al.  Mapping croplands, cropping patterns, and crop types using MODIS time-series data , 2018, Int. J. Appl. Earth Obs. Geoinformation.

[16]  J. Pekel,et al.  High-resolution mapping of global surface water and its long-term changes , 2016, Nature.

[17]  D. Skole,et al.  Land Use and Land Cover Change , 2014 .

[18]  Laerte Guimarães Ferreira,et al.  Monitoring the brazilian pasturelands: A new mapping approach based on the landsat 8 spectral and temporal domains , 2017, Int. J. Appl. Earth Obs. Geoinformation.

[19]  G. Berndes,et al.  ff ects of Governance on Availability of Land for Agriculture and 2 Conservation in Brazil , 2015 .

[20]  D. Arvor,et al.  Transition in environmental governance in the Brazilian Amazon: emergence of a new pattern of socio-economic development and deforestation , 2016 .

[21]  Damien Arvor,et al.  Analyzing the agricultural transition in Mato Grosso, Brazil, using satellite-derived indices , 2012 .

[22]  G. Berndes,et al.  Effects of Governance on Availability of Land for Agriculture and Conservation in Brazil. , 2015, Environmental science & technology.

[23]  Scott L. Powell,et al.  Bringing an ecological view of change to Landsat‐based remote sensing , 2014 .

[24]  Gilberto Câmara,et al.  Web Services for Big Earth Observation data , 2017, GEOINFO.

[25]  H. K. Gibbs,et al.  Brazil's Soy Moratorium , 2015, Science.

[26]  Alan H. Strahler,et al.  Global land cover mapping from MODIS: algorithms and early results , 2002 .

[27]  Daniel Alves Aguiar,et al.  Studies on the Rapid Expansion of Sugarcane for Ethanol Production in São Paulo State (Brazil) Using Landsat Data , 2010, Remote. Sens..

[28]  Yang Shao,et al.  Comparison of support vector machine, neural network, and CART algorithms for the land-cover classification using limited training data points , 2012 .

[29]  C. Justice,et al.  High-Resolution Global Maps of 21st-Century Forest Cover Change , 2013, Science.

[30]  Trevor S. Wiens,et al.  Three way k-fold cross-validation of resource selection functions , 2008 .

[31]  Raul Queiroz Feitosa,et al.  Campo Verde Database: Seeking to Improve Agricultural Remote Sensing of Tropical Areas , 2018, IEEE Geoscience and Remote Sensing Letters.

[32]  P. Hostert,et al.  Mapping farmland abandonment and recultivation across Europe using MODIS NDVI time series. , 2015 .

[33]  T. Garnett.  Where are the best opportunities for reducing greenhouse gas emissions in the food system (including the food chain) , 2011 .

[34]  Eric F. Lambin,et al.  Land-Use and Land-Cover Change , 2006 .

[35]  J. Kastens,et al.  Soy moratorium impacts on soybean and deforestation dynamics in Mato Grosso, Brazil , 2017, PloS one.

[36]  Gérard Dedieu,et al.  Assessing the robustness of Random Forests to map land cover with high resolution satellite image time series over large areas , 2016 .

[37]  Damien Arvor,et al.  Combining socioeconomic development with environmental governance in the Brazilian Amazon: the Mato Grosso agricultural frontier at a tipping point , 2018, Environment, Development and Sustainability.

[38]  J. Mustard,et al.  Wavelet analysis of MODIS time series to detect expansion and intensification of row-crop agriculture in Brazil , 2008 .

[39]  Marcos Adami,et al.  Recent cropping frequency, expansion, and abandonment in Mato Grosso, Brazil had selective land characteristics , 2014 .

[40]  Jude H. Kastens,et al.  Classifying multiyear agricultural land use data from Mato Grosso using time-series MODIS vegetation index data , 2013 .

[41]  Vladimir Vapnik,et al.  Support-vector networks , 2004, Machine Learning.

[42]  F. Achard,et al.  Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s , 2010, Proceedings of the National Academy of Sciences.

[43]  Eric Armijo,et al.  Slowing Amazon deforestation through public policy and interventions in beef and soy supply chains , 2014, Science.

[44]  Gilberto Câmara,et al.  A Time-Weighted Dynamic Time Warping Method for Land-Use and Land-Cover Mapping , 2016, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[45]  Daniel Alves Aguiar,et al.  Remote Sensing Time Series to Evaluate Direct Land Use Change of Recent Expanded Sugarcane Crop in Brazil , 2011 .

[46]  Rob J Hyndman,et al.  Detecting trend and seasonal changes in satellite image time series , 2010 .

[47]  J. Assunção,et al.  Deforestation slowdown in the Brazilian Amazon: prices or policies? , 2015, Environment and Development Economics.

[48]  Gilberto Câmara,et al.  Big earth observation data analytics: matching requirements to system architectures , 2016, BigSpatial '16.

[49]  J. Mustard,et al.  Socioeconomic development and agricultural intensification in Mato Grosso , 2013, Philosophical Transactions of the Royal Society B: Biological Sciences.

[50]  Joanne C. White,et al.  Optical remotely sensed time series data for land cover classification: A review , 2016 .