Skip to content
DraconiticResearch
Search Sources API
184,584 studies31 sourcesUpdated 35 min ago
Peer-reviewedCC BYJournal of Agriculture and Food Research

Potato crop coefficient (

  • Mohammed Diykh1
  • Mumtaz Ali2
  • Aitazaz A. Farooque3
  • Saad Javed Cheema1
  • Mehdi Jamei1
  • Zoe Li4
  • Junye Wang5
  • Arnold W. Schumann6
  • Zaher Mundher Yaseen7
7 affiliations
  1. 1Canadian Centre for Climate Change and Adaptation, University of Prince Edward Island, St Peter's Bay, PE, Canada
  2. 2School of Business, Law, Humanities, and Pathways, University of Southern Queensland, QLD, 4300, Australia; Scientific Research Centre, Al-Ayen Iraqi University, Thi-Qar, Iraq; Corresponding author. School of Business, Law, Humanities, and Pathways, University of Southern Queensland, QLD 4300, Australia.
  3. 3Canadian Centre for Climate Change and Adaptation, University of Prince Edward Island, St Peter's Bay, PE, Canada; Faculty of Sustainable Design Engineering, University of Prince Edward Island, Charlottetown, PE, Canada; Corresponding author. Canadian Centre for Climate Change and Adaptation, University of Prince Edward Island, St Peters Bay, PE, Canada.
  4. 4Department of Civil Engineering, McMaster University, Hamilton, ON, Canada
  5. 5Faculty of Science and Technology, Athabasca University, Athabasca, AB, Canada
  6. 6Citrus Research and Education Centre, University of Florida, FL, USA
  7. 7Civil and Environmental Engineering Department, King Fahd University of Petroleum & Minerals, Dhahran, 31261, Saudi Arabia
Read full textPDFDOICite

Abstract

Accurate prediction of Potato crop coefficient ( K c ) is an essential factor for effective irrigation management, as it directly impacts the percentage of crop water use. Capturing dynamic patterns from agro-meteorological data for K c prediction presents a significant challenges due to the complex structure of the agro-meteorological data. This paper proposes a probabilistic dynamic graph deep learning (PDGDL) model for K c prediction in the Province of Prince Edward Island (PEI), Canada. The proposed PDGDL model employs an encoder and decoder to learn the dynamic graph representations, which involves three layers: causal layer, Reasoning Layer, and output layer. The PDGDL model learns the influence of variables from historical records to the future period to form the final K c prediction. Extensive experiments are conducted on two-year experimental data (2023-2024) collected in PEI from three fields representing three types of soils. Numerous statistical and visual metrics are applied to evaluate the proposed model against seven benchmark models, Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Networks (CNN) based a Bidirectional LSTM (CBiLSTM), BiLSTM, CNN based on GRU (CGRU), and graph deep learning models, Graph neural network (GNN), and Dynamic graph neural network (DGNN). The results demonstrated the ability of the proposed PDGDL model to predict K c in different soil types. It achieved the Root Mean Square Errors (RMSE) of 0.62, 0.63, 0.61, Mean Absolute Errors (MAE) of 0.312, 0.320, 0.301, and the coefficients of determination (R 2 ) of 0.995, 0.984, 0.972 for three fields. These improvements showed the ability of the proposed PDGDL model to effectively handle complex and non-linear patterns of K c , delivering accurate predictive results compared to other graph deep learning models, and offering helpful support for decision making in the agriculture sector.

Related

  1. TRP-Net: A Text-Guided Dual-Modal Network for Potato Leaf Lesion Detection Under Complex Imaging Conditions
    Plants (Basel, Switzerland)Peer-reviewedCC BYFull textPlant biologyEurope PMC
  2. Evaluation of the antifungal activity of Moringa oleifera leaf and stem bark extracts against Fusarium species causing potato wilt
    Gemechis Kinati, Asefa Keneni, Fekadu Kuma +1
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Plant biologyCrossref
  3. Genomic and metabolomic characterization of Bacillus velezensis CMC-6 reveals biocontrol potential against potato Fusarium Wilt
    Chao Zhang, Yuanzheng Zhao, Dong Wang +5
    BMC GenomicsPeer-reviewedCC BY-NC-ND 4.0Plant biologyCrossref
  4. Hybrid attention-based deep learning framework for potato leaf blight disease detection using CNN and fractal net
    Achin Jain, Rupal Gupta, Arvind Panwar +5
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Plant biologyCrossref
  5. Multivariate analysis and genetic diversity for tuber yield and quality traits in potato (Solanum tuberosum L.)
    Nikhil Thakur, Salej Sood, Manisha Kumari +7
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textPlant biologyCrossref

Cite

BibTeXRIS
BibTeX
@article{diykh2026potato,
  title = {{Potato crop coefficient (}},
  author = {Mohammed Diykh and Mumtaz Ali and Aitazaz A. Farooque and Saad Javed Cheema and Mehdi Jamei and Zoe Li and Junye Wang and Arnold W. Schumann and Zaher Mundher Yaseen},
  journal = {Journal of Agriculture and Food Research},
  year = {2026},
  month = oct,
  doi = {10.1016/j.jafr.2026.103292},
  url = {https://doi.org/10.1016/j.jafr.2026.103292},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Potato crop coefficient (
AU  - Mohammed Diykh
AU  - Mumtaz Ali
AU  - Aitazaz A. Farooque
AU  - Saad Javed Cheema
AU  - Mehdi Jamei
AU  - Zoe Li
AU  - Junye Wang
AU  - Arnold W. Schumann
AU  - Zaher Mundher Yaseen
PY  - 2026
DA  - 2026/10/30
JO  - Journal of Agriculture and Food Research
DO  - 10.1016/j.jafr.2026.103292
UR  - https://doi.org/10.1016/j.jafr.2026.103292
AB  - Accurate prediction of Potato crop coefficient ( K c ) is an essential factor for effective irrigation management, as it directly impacts the percentage of crop water use. Capturing dynamic patterns from agro-meteorological data for K c prediction presents a significant challenges due to the complex structure of the agro-meteorological data. This paper proposes a probabilistic dynamic graph deep learning (PDGDL) model for K c prediction in the Province of Prince Edward Island (PEI), Canada. The proposed PDGDL model employs an encoder and decoder to learn the dynamic graph representations, which involves three layers: causal layer, Reasoning Layer, and output layer. The PDGDL model learns the influence of variables from historical records to the future period to form the final K c prediction. Extensive experiments are conducted on two-year experimental data (2023-2024) collected in PEI from three fields representing three types of soils. Numerous statistical and visual metrics are applied to evaluate the proposed model against seven benchmark models, Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Networks (CNN) based a Bidirectional LSTM (CBiLSTM), BiLSTM, CNN based on GRU (CGRU), and graph deep learning models, Graph neural network (GNN), and Dynamic graph neural network (DGNN). The results demonstrated the ability of the proposed PDGDL model to predict K c in different soil types. It achieved the Root Mean Square Errors (RMSE) of 0.62, 0.63, 0.61, Mean Absolute Errors (MAE) of 0.312, 0.320, 0.301, and the coefficients of determination (R 2 ) of 0.995, 0.984, 0.972 for three fields. These improvements showed the ability of the proposed PDGDL model to effectively handle complex and non-linear patterns of K c , delivering accurate predictive results compared to other graph deep learning models, and offering helpful support for decision making in the agriculture sector.
SN  - 2666-1543
ER  -