Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning

Object-based analysis is widely used for extracting information from satellite data using machine learning, offering reduced sensitivity to fine-scale variability, noise, and computational cost compared to pixel-based methods. However, segmentation algorithms for center-pivot fields often treat fiel...

Full description

Saved in:
Bibliographic Details
Main Authors: Ting Li, Oliver Miguel Lopez Valencia, Matthew F. McCabe
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10963748/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849727960650612736
author Ting Li
Oliver Miguel Lopez Valencia
Matthew F. McCabe
author_facet Ting Li
Oliver Miguel Lopez Valencia
Matthew F. McCabe
author_sort Ting Li
collection DOAJ
description Object-based analysis is widely used for extracting information from satellite data using machine learning, offering reduced sensitivity to fine-scale variability, noise, and computational cost compared to pixel-based methods. However, segmentation algorithms for center-pivot fields often treat fields as single units, neglecting that a field can be subdivided into different sections caused by varied management practices, such as differing planting and harvesting dates, crop types, and rotations. This variability is particularly prevalent in hot, arid regions, such as Saudi Arabia, where precise water and crop management are crucial for sustaining agricultural productivity. However, such subfield division reduces the accuracy of object-based agroinformatics insights and the effectiveness of large-scale analyses. A machine learning-based approach combining Kmeans clustering and cosine similarity was developed to quantify subfield divisions using temporal features derived from Sentinel-2 normalized difference vegetation index (NDVI) time series. The performance of discrete wavelet transformation(DWT) and Savitzky&#x2013;Golay filtering was compared for processing the NDVI time series. When evaluated against a reference dataset, the approach achieved a maximum accuracy of 93.38&#x0025; with DWT level 1 decomposition using the &#x201C;haar&#x201D; wavelet function. These parameters were applied to map the nationwide center-pivot subfield division dynamics across Saudi Arabia from 2019 to 2023. Results revealed that approximately 20&#x0025; of center-pivot fields exhibited subfield divisions, ranging from 5740 fields (2083 km<inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula>) in 2020 to 7342 fields (2770 km<inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula>) in 2023. Larger fields were more prone to subfield divisions, with a median acreage of 40 ha compared to 20 ha for undivided fields. Dominant management strategies included half-to-half and 5:3:2 divisions. This approach enhances object-based agroinformatics products and facilitates more accurate food security assessments.
format Article
id doaj-art-35cb5730f7644f7a9554db28807ef3df
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-35cb5730f7644f7a9554db28807ef3df2025-08-20T03:09:42ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118116861170210.1109/JSTARS.2025.356007110963748Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine LearningTing Li0https://orcid.org/0000-0003-4170-8489Oliver Miguel Lopez Valencia1https://orcid.org/0000-0002-1559-5970Matthew F. McCabe2https://orcid.org/0000-0002-1279-5272Hydrology, Agriculture and Land Observation (HALO) Laboratory, Climate and Livability Initiative, Division of Biological and Environmental Sciences and Engineering, King Abdullah University of Science and Technology (KAUST) Thuwal, Saudi ArabiaHydrology, Agriculture and Land Observation (HALO) Laboratory, Climate and Livability Initiative, Division of Biological and Environmental Sciences and Engineering, King Abdullah University of Science and Technology (KAUST) Thuwal, Saudi ArabiaHydrology, Agriculture and Land Observation (HALO) Laboratory, Climate and Livability Initiative, Division of Biological and Environmental Sciences and Engineering, King Abdullah University of Science and Technology (KAUST) Thuwal, Saudi ArabiaObject-based analysis is widely used for extracting information from satellite data using machine learning, offering reduced sensitivity to fine-scale variability, noise, and computational cost compared to pixel-based methods. However, segmentation algorithms for center-pivot fields often treat fields as single units, neglecting that a field can be subdivided into different sections caused by varied management practices, such as differing planting and harvesting dates, crop types, and rotations. This variability is particularly prevalent in hot, arid regions, such as Saudi Arabia, where precise water and crop management are crucial for sustaining agricultural productivity. However, such subfield division reduces the accuracy of object-based agroinformatics insights and the effectiveness of large-scale analyses. A machine learning-based approach combining Kmeans clustering and cosine similarity was developed to quantify subfield divisions using temporal features derived from Sentinel-2 normalized difference vegetation index (NDVI) time series. The performance of discrete wavelet transformation(DWT) and Savitzky&#x2013;Golay filtering was compared for processing the NDVI time series. When evaluated against a reference dataset, the approach achieved a maximum accuracy of 93.38&#x0025; with DWT level 1 decomposition using the &#x201C;haar&#x201D; wavelet function. These parameters were applied to map the nationwide center-pivot subfield division dynamics across Saudi Arabia from 2019 to 2023. Results revealed that approximately 20&#x0025; of center-pivot fields exhibited subfield divisions, ranging from 5740 fields (2083 km<inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula>) in 2020 to 7342 fields (2770 km<inline-formula><tex-math notation="LaTeX">$^{2}$</tex-math></inline-formula>) in 2023. Larger fields were more prone to subfield divisions, with a median acreage of 40 ha compared to 20 ha for undivided fields. Dominant management strategies included half-to-half and 5:3:2 divisions. This approach enhances object-based agroinformatics products and facilitates more accurate food security assessments.https://ieeexplore.ieee.org/document/10963748/Agroinformaticscosine similaritydiscrete wavelet transformationKmeans clusteringSavizky–Golay filter (SG)subfield division
spellingShingle Ting Li
Oliver Miguel Lopez Valencia
Matthew F. McCabe
Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Agroinformatics
cosine similarity
discrete wavelet transformation
Kmeans clustering
Savizky–Golay filter (SG)
subfield division
title Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
title_full Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
title_fullStr Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
title_full_unstemmed Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
title_short Mapping Nationwide Subfield Division Dynamics in Saudi Arabia Using Temporal Patterns of Sentinel-2 NDVI and Machine Learning
title_sort mapping nationwide subfield division dynamics in saudi arabia using temporal patterns of sentinel 2 ndvi and machine learning
topic Agroinformatics
cosine similarity
discrete wavelet transformation
Kmeans clustering
Savizky–Golay filter (SG)
subfield division
url https://ieeexplore.ieee.org/document/10963748/
work_keys_str_mv AT tingli mappingnationwidesubfielddivisiondynamicsinsaudiarabiausingtemporalpatternsofsentinel2ndviandmachinelearning
AT olivermiguellopezvalencia mappingnationwidesubfielddivisiondynamicsinsaudiarabiausingtemporalpatternsofsentinel2ndviandmachinelearning
AT matthewfmccabe mappingnationwidesubfielddivisiondynamicsinsaudiarabiausingtemporalpatternsofsentinel2ndviandmachinelearning