Agricultural cropland extent and areas of South Asia derived using Landsat satellite 30-m time-series big-data using random forest machine learning algorithms on the Google Earth Engine cloud

Abstract

The South Asia (India, Pakistan, Bangladesh, Nepal, Sri Lanka and Bhutan) has a staggering 900 million people (~43% of the population) who face food insecurity or severe food insecurity as per United Nations, Food and Agriculture Organization's (FAO) the Food Insecurity Experience Scale (FIES). The existing coarse-resolution (≥250-m) cropland maps lack precision in geo-location of individual farms and have low map accuracies. This also results in uncertainties in cropland areas calculated from such products. Thereby, the overarching goal of this study was to develop a high spatial resolution (30-m or better) baseline cropland extent product of South Asia for the year 2015 using Landsat satellite time-series big-data and machine learning algorithms (MLAs) on the Google Earth Engine (GEE) cloud computing platform. To eliminate the impact of clouds, 10 time-composited Landsat bands (blue, green, red, NIR, SWIR1, SWIR2, Thermal, EVI, NDVI, NDWI) were derived for each of the three timeperiods over 12 months (monsoon: Days of the Year (DOY) 151-300; winter: DOY 301-365 plus 1-60; and summer: DOY 61-150), taking the every 8-day data from Landsat-8 and 7 for the years 2013-2015, for a total of 30-bands plus global digital elevation model (GDEM) derived slope band. This 31-band mega-file big data-cube was composed for each of the five agro-ecological zones (AEZ's) of South Asia and formed a baseline data for image classification and analysis. Knowledgebase for the Random Forest (RF) MLAs were developed using spatially well spread-out reference training data (N = 2179) in five AEZs. The classification was performed on GEE for each of the five AEZs using well-established knowledge-base and RF MLAs on the cloud. Map accuracies were measured using independent validation data (N = 1185). The survey showed that the South Asia cropland product had a producer's accuracy of 89.9% (errors of omissions of 10.1%), user's accuracy of 95.3% (errors of commission of 4.7%) and an overall accuracy of 88.7%. The National and sub-national (districts) areas computed from this cropland extent product explained 80-96% variability when compared with the National statistics of the South Asian Countries. The full-resolution imagery can be viewed at full-resolution, by zooming-in to any location in South

Monitoring ecosystem restoration of multiple surface coal mine sites in China via LANDSAT images using the Google Earth Engine

“…Different levels of processing products can be used, including raw data, TOA, SR, and composite products. Based on the GEE platform, researchers have achieved a great deal of forest monitoring, urban sprawl monitoring, and ecosystem monitoring at regional and even global scale (Dong et al, 2016; Gumma et al, 2019; Hansen et al, 2013; Kennedy, Yang, & Cohen, 2010; Liu et al, 2018; Phalke et al, 2020). Benefiting from researchers' contributions and improvements in the computing codes, an increasing number of functions are implemented on the GEE platform, greatly improving the accuracy and speed of data processing (Ermida, Soares, Mantas, Göttsche, & Trigo, 2020).…”

Synergistic Use of Radar and Optical Satellite Data for Improved Monsoon Cropland Mapping in India

“…The results presented in this work are important, as this will provide the first high-resolution (10 m) monsoon cropland map generation, and can also be transferred to other agro-ecoregions. This method shows an improvement over existing methods that are primarily focused on non-monsoon/winter cropland mapping at 30 m or coarse resolution [20,24,27,85].…”

Current and near-term advances in Earth observation for ecological applications

“…Thousands of environmentally focused papers have been published using Landsat data, including the first National Land Cover Dataset in the 1990s for the conterminous USA (Vogelmann et al 2001 ) and continental scale maps of land cover change (Townsend and Walsh 2001 ; Hansen et al 2014 ) and the forestry map of Canada (Wulder et al 2003 ). Notable ecological studies have addressed forest health (e.g., monitoring woolly adelgid outbreaks in eastern hemlock; Royle and Lathrop 1997 ), forest survival after wildfire (Kushla and Ripple 1998 ; Miller and Yool 2002 ; Karlson et al 2015 ), mapping the distribution of semiarid vegetation and environmental controls on species abundance patterns (Smith et al 1990a , 1990b ), and innumerable other ecosystem applications from agriculture (Leslie et al 2017 ; Gumma et al 2020 ) to wetlands (Johnson and Barson 1993 ; Tang et al 2004 ; Schneider et al 2009 ; Halabisky et al, 2016 , biodiversity hotspots (Gould 2000 ; Helmer et al 2002 ; Brandt et al 2013 ; Cavender-Bares et al 2016 ), alpine ecosystems (Dozier 1989 ; Bolton et al 2018 ; Gianinetto et al 2019 ), the arctic (Stow et al 2004 ; Huang et al 2017 ; Griffin et al 2018 ), and to dry lands (Qi et al 2000 ; Langley et al 2001 ; Bradley and Mustard 2005 ; Chen et al 2005 ; Sohn and Qi 2005 ) and landscape structure (Saunders et al 2002 ) among other applications.…”