date: 2021-06-18T10:14:15Z
pdf:unmappedUnicodeCharsPerPage: 17
pdf:PDFVersion: 1.7
pdf:docinfo:title: Remote Sensing Based Yield Estimation of Rice (Oryza Sativa L.) Using Gradient Boosted Regression in India
xmp:CreatorTool: LaTeX with hyperref
Keywords: yield estimation; high resolution; Remote Sensing; MODIS; Leaf Area Index (LAI); machine learning; Gradient Boosted Regression (GBR); India
access_permission:modify_annotations: true
access_permission:can_print_degraded: true
subject: Accurate and spatially explicit yield information is required to ensure farmers? income and food security at local and national levels. Current approaches based on crop cutting experiments are expensive and usually too late for timely income stabilization measures like crop insurances. We, therefore, utilized a Gradient Boosted Regression (GBR), a machine learning technique, to estimate rice yields at ~500 m spatial resolution for rice-producing areas in India with potential application for near real-time estimates. We used resampled intermediate resolution (~5 km) images of the Moderate Resolution Imaging Spectroradiometer (MODIS) Leaf Area Index (LAI) and observed yields at the district level in India for calibrating GBR models. These GBRs were then used to downscale district yields to 500 m resolution. Downscaled yields were re-aggregated for validation against out-of-sample district yields not used for model training and an additional independent data set of block-level (below district-level) yields. Our downscaled and re-aggregated yields agree well with reported district-level observations from 2003 to 2015 (r = 0.85 & MAE = 0.15 t/ha). The model performance improved further when estimating separate models for different rice cropping densities (up to r = 0.93). An additional out-of-sample validation for the years 2016 and 2017, proved successful with r = 0.84 and r = 0.77, respectively. Simulated yield accuracy was higher in water-limited, rainfed agricultural systems. We conclude that this downscaling approach of rice yield estimation using GBR is feasible across India and may complement current approaches for timely rice yield estimation required by insurance companies and government agencies.
dc:creator: Ponraj Arumugam, Abel Chemura, Bernhard Schauberger and Christoph Gornott
dcterms:created: 2021-06-18T09:58:07Z
Last-Modified: 2021-06-18T10:14:15Z
dcterms:modified: 2021-06-18T10:14:15Z
dc:format: application/pdf; version=1.7
title: Remote Sensing Based Yield Estimation of Rice (Oryza Sativa L.) Using Gradient Boosted Regression in India
Last-Save-Date: 2021-06-18T10:14:15Z
pdf:docinfo:creator_tool: LaTeX with hyperref
access_permission:fill_in_form: true
pdf:docinfo:keywords: yield estimation; high resolution; Remote Sensing; MODIS; Leaf Area Index (LAI); machine learning; Gradient Boosted Regression (GBR); India
pdf:docinfo:modified: 2021-06-18T10:14:15Z
meta:save-date: 2021-06-18T10:14:15Z
pdf:encrypted: false
dc:title: Remote Sensing Based Yield Estimation of Rice (Oryza Sativa L.) Using Gradient Boosted Regression in India
modified: 2021-06-18T10:14:15Z
cp:subject: Accurate and spatially explicit yield information is required to ensure farmers? income and food security at local and national levels. Current approaches based on crop cutting experiments are expensive and usually too late for timely income stabilization measures like crop insurances. We, therefore, utilized a Gradient Boosted Regression (GBR), a machine learning technique, to estimate rice yields at ~500 m spatial resolution for rice-producing areas in India with potential application for near real-time estimates. We used resampled intermediate resolution (~5 km) images of the Moderate Resolution Imaging Spectroradiometer (MODIS) Leaf Area Index (LAI) and observed yields at the district level in India for calibrating GBR models. These GBRs were then used to downscale district yields to 500 m resolution. Downscaled yields were re-aggregated for validation against out-of-sample district yields not used for model training and an additional independent data set of block-level (below district-level) yields. Our downscaled and re-aggregated yields agree well with reported district-level observations from 2003 to 2015 (r = 0.85 & MAE = 0.15 t/ha). The model performance improved further when estimating separate models for different rice cropping densities (up to r = 0.93). An additional out-of-sample validation for the years 2016 and 2017, proved successful with r = 0.84 and r = 0.77, respectively. Simulated yield accuracy was higher in water-limited, rainfed agricultural systems. We conclude that this downscaling approach of rice yield estimation using GBR is feasible across India and may complement current approaches for timely rice yield estimation required by insurance companies and government agencies.
pdf:docinfo:subject: Accurate and spatially explicit yield information is required to ensure farmers? income and food security at local and national levels. Current approaches based on crop cutting experiments are expensive and usually too late for timely income stabilization measures like crop insurances. We, therefore, utilized a Gradient Boosted Regression (GBR), a machine learning technique, to estimate rice yields at ~500 m spatial resolution for rice-producing areas in India with potential application for near real-time estimates. We used resampled intermediate resolution (~5 km) images of the Moderate Resolution Imaging Spectroradiometer (MODIS) Leaf Area Index (LAI) and observed yields at the district level in India for calibrating GBR models. These GBRs were then used to downscale district yields to 500 m resolution. Downscaled yields were re-aggregated for validation against out-of-sample district yields not used for model training and an additional independent data set of block-level (below district-level) yields. Our downscaled and re-aggregated yields agree well with reported district-level observations from 2003 to 2015 (r = 0.85 & MAE = 0.15 t/ha). The model performance improved further when estimating separate models for different rice cropping densities (up to r = 0.93). An additional out-of-sample validation for the years 2016 and 2017, proved successful with r = 0.84 and r = 0.77, respectively. Simulated yield accuracy was higher in water-limited, rainfed agricultural systems. We conclude that this downscaling approach of rice yield estimation using GBR is feasible across India and may complement current approaches for timely rice yield estimation required by insurance companies and government agencies.
Content-Type: application/pdf
pdf:docinfo:creator: Ponraj Arumugam, Abel Chemura, Bernhard Schauberger and Christoph Gornott
X-Parsed-By: org.apache.tika.parser.DefaultParser
creator: Ponraj Arumugam, Abel Chemura, Bernhard Schauberger and Christoph Gornott
meta:author: Ponraj Arumugam, Abel Chemura, Bernhard Schauberger and Christoph Gornott
dc:subject: yield estimation; high resolution; Remote Sensing; MODIS; Leaf Area Index (LAI); machine learning; Gradient Boosted Regression (GBR); India
meta:creation-date: 2021-06-18T09:58:07Z
created: 2021-06-18T09:58:07Z
access_permission:extract_for_accessibility: true
access_permission:assemble_document: true
xmpTPg:NPages: 18
Creation-Date: 2021-06-18T09:58:07Z
pdf:charsPerPage: 3903
access_permission:extract_content: true
access_permission:can_print: true
meta:keyword: yield estimation; high resolution; Remote Sensing; MODIS; Leaf Area Index (LAI); machine learning; Gradient Boosted Regression (GBR); India
Author: Ponraj Arumugam, Abel Chemura, Bernhard Schauberger and Christoph Gornott
producer: pdfTeX-1.40.21
access_permission:can_modify: true
pdf:docinfo:producer: pdfTeX-1.40.21
pdf:docinfo:created: 2021-06-18T09:58:07Z