Cognitive computing models for estimation of reference evapotranspiration: A review
Introduction
Economic advancement and increasing global population expand the demand for agricultural production. The global food requirement will increase about 60% by the year 2050. Globally, it is assessed that agricultural activity consumes about 70% of the gross water (Provenzano & Sinobas, 2014). The global irrigated land occupies only 16% of cultivatable area (Playán et al., 2014). The arid and semi-arid regions occupy 36% of the global land (Safriel et al., 2005) and it is predicted that drought hazard will further increase due to global warming (Alcamo et al., 2007, Arnell et al., 2011). Water productivity (WP) is the ratio between crop yield and total water use (Pereira et al., 2002). Water consumed by plants is less than 65% of supplied water (Chartzoulakis & Bertaki, 2015). Irrigation automation is challenging due to the spatial and temporal variability of soil, weather and plant data (Dabach et al., 2011, Soulis and Elmaloglou, 2018). The evolution of first generation irrigation technology started with multi-client electronic hydrants for utilization on the dispensation network. The second generation irrigation technology was variable frequency pump. The microirrigation method was the third generation in irrigation technology wherein WP was increased but marginally installed due to high initial investment. The sub–surface drip irrigation (SDI) was the fourth generation in irrigation technology that was invented to solve the issues of surface drip irrigation, specifically for the elimination of emitter clogging issue. The fifth generation in irrigation technology was deficit irrigation that was invented to supply reduced amount of water without affecting the yield based on crop growth stage (Levidow et al., 2014, Kang et al., 2017). The cognitive computing-based intelligent irrigation system is the current trend to provide economical and efficient models for agricultural water management (Haider et al., 2007, Torres-Rua et al., 2012, Behmann et al., 2014, Kamilaris et al., 2017, Niu et al., 2017, Chlingaryan et al., 2018, Kamilaris and Prenafeta-Boldú, 2018).
The rest of the paper structure is described as follows. Section 2 exhibits the most widely used empirical method for the estimation of ET0. The crop water requirements prediction model is highlighted in Section 3. The cognitive computing methods used for ET0 prediction are discussed in Section 4. The summary of the review is described in Section 5. The conclusion part is described in Section 6. Finally, cognitive computing-based irrigation system (CCIS) is outlined in Section 7 as future work.
Section snippets
Reference evapotranspiration (ET0) estimation model
The ET0 is an important metric to understand crop water requirements and obtain a satisfactory yield (Temesgen et al., 2005). The ET0 plays a vital role in the computation of irrigation water requirements, estimation of crop evapotranspiration (ETc) and to design soil water balance models. Weather data, such as air temperature (AT), wind speed (WS), solar radiation (SR), sunshine hours (SS), relative humidity (RH), rainfall (RF), air pressure (AP) and vapour pressure (VP) are used as key input
Crop water requirements prediction model
Crop water requirements play a crucial role in irrigation systems for the determination of water productivity (WP), wherein ETc is determined using ET0 and crop coefficient (Kc). The Kc indicates the actual plant transpiration. The crop water requirements prediction model based on crop growth stage and weather data are highlighted in Eq. (2) (Allen et al., 2005).
Cognitive computing models for estimation of ET0
Cognitive computing is a domain of computer science that imitates the phenomenon of human brain (Gocić et al., 2015). Few aspects, such as consciousness and cognition, are key features of cognitive computing techniques. Cognitive computing methods exploit resistance for uncertainty and imprecision, ensure conformity and robustness and offer economical solutions (Keskin & Terzi, 2006). Cognitive computing techniques have been used for different agricultural applications to build intelligent and
Performance metrics for cognitive computing models
The key performance factors in cognitive computing models are local minima, convergence rate, over fitting and generalization potential. The review exhibited that MLFF-NN and GP models got blocked in local minima easily and that ANFIS + CSA and ENN models can be used to overcome this issue and predict ET0. The ANFIS + CSA, MLFF-NN and SONN models delay in convergence rate and NARX and ENN models can be used to address this issue and predict ET0. The MLFF-NN and GEP models are prone to over
Conclusion
The review of cognitive computing models for ET0 prediction was conducted to determine the most promising approach for irrigation automation. The survey firstly outlines the neural network approaches for reference evapotranspiration (ET0) prediction. The statistical performance analysis of neural network models exhibited that second order neural network (SONN) is the most promising method for ET0 prediction, considering the balance between RMSE and R2values.
Secondly, SVM approaches are outlined
Future irrigation system
Cognitive computing-based intelligent irrigation system (CCIS) can be designed, which offers a balance between high water productivity and economy. The CCIS approach preserves the benefit of low initial investment of traditional surface irrigation system with higher WP through the employment of supervised, unsupervised, semi-supervised and reinforcement learning techniques. The ET0 prediction plays a crucial role in CCIS to predict crop water requirements accurately.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors express their gratefulness to the Rev. Fr. Augustine Paimpallil, CMI, Principal, Christ Institute, Bengaluru and Daisy Agnes, English Professor, for the manuscript’s proofreading, language improvement and style improvement.
References (64)
- et al.
The implications of climate policy for the impacts of climate change on global water resources
Global Environmental Change
(2011) - et al.
A review on neural networks with random weights
Neurocomputing
(2018) - et al.
Sustainable water management in agriculture under climate change
Agriculture and Agricultural Science Procedia
(2015) - et al.
Support vector machine enhanced empirical reference evapotranspiration estimation with limited meteorological parameters
Computers and Electronics in Agriculture
(2020) - et al.
Machine learning approaches for crop yield prediction and nitrogen status estimation in precision agriculture: A review
Computers and Electronics in Agriculture
(2018) - et al.
Soft computing approaches for forecasting reference evapotranspiration
Computers and Electronics in Agriculture
(2015) - et al.
Monthly pan evaporation modeling using linear genetic programming
Journal of Hydrology
(2013) - et al.
A review on the practice of big data analysis in agriculture
Computers and Electronics in Agriculture
(2017) - et al.
Deep learning in agriculture: A survey
Computers and Electronics in Agriculture
(2018) - et al.
Improving agricultural water productivity to ensure food security in China under changing environment: From research to practice
Agricultural Water Management
(2017)
Pan evaporation modeling using least square support vector machine, multivariate adaptive regression splines and M5 model tree
Journal of Hydrology
Improving water-efficient irrigation: Prospects and difficulties of innovative practices
Agricultural Water Management
Support vector machines in remote sensing: A review
ISPRS Journal of Photogrammetry and Remote Sensing
Irrigation management under water scarcity
Agricultural Water Management
Advances in artificial neural networks and machine learning
Neurocomputing
Neural networks: An overview of early research, current frameworks and new challenges
Neurocomputing
A wavelet–linear genetic programming model for sodium (Na+) concentration forecasting in rivers
Journal of Hydrology
A survey on applications and variants of the cuckoo search algorithm
Applied Soft Computing
Optimum soil water content sensors placement for surface drip irrigation scheduling in layered soils
Computers and Electronics in Agriculture
Forecasting groundwater levels using nonlinear autoregressive networks with exogenous input (NARX)
Journal of Hydrology
Gene-expression programming for sediment transport in sewer pipe systems
Journal of Pipeline Systems Engineering and Practice
Performance evaluation of ANN and ANFIS models for estimating garlic crop evapotranspiration
Journal of Irrigation and Drainage Engineering
Evapotranspiration modeling using second-order neural networks
Journal of Hydrologic Engineering
Future long-term changes in global water resources driven by socio-economic and climatic changes
Hydrological Sciences Journal
Modelling reference evapotranspiration by combining neuro-fuzzy and evolutionary strategies
Acta Geophysica
Crop evapotranspiration-Guidelines for computing crop water requirements-FAO Irrigation and drainage paper 56
Fao, Rome
FAO-56 dual crop coefficient method for estimating evaporation from soil and application extensions
Journal of irrigation and drainage engineering
FAO-24 reference evapotranspiration factors
Journal of Irrigation and Drainage Engineering
A review of advanced machine learning methods for the detection of biotic stress in precision crop protection
Precision Agriculture
Evaporation Research: Review and Interpretation
Journal of Irrigation and Drainage Engineering
Daily reference evapotranspiration estimation based on least squares support vector machines
Computer and Computing Technologies in Agriculture V
Numerical investigation of irrigation scheduling based on soil water status
Irrigation Science
Cited by (25)
Short-term daily reference evapotranspiration forecasting using temperature-based deep learning models in different climate zones in China
2023, Agricultural Water ManagementA review of the Artificial Intelligence (AI) based techniques for estimating reference evapotranspiration: Current trends and future perspectives
2023, Computers and Electronics in AgricultureDesign data decomposition-based reference evapotranspiration forecasting model: A soft feature filter based deep learning driven approach
2023, Engineering Applications of Artificial IntelligenceForecasting weekly reference evapotranspiration using Auto Encoder Decoder Bidirectional LSTM model hybridized with a Boruta-CatBoost input optimizer
2022, Computers and Electronics in AgricultureCitation Excerpt :Nonetheless, the ETo computations using the FAO-56 PM approach served as the sole basis for estimating this input–output relationship. The topic has sparked a flurry of studies into the development of various forms of ML algorithms (Faskari et al., 2022; Naganna et al., 2020), as seen by the recent reviews in this domain lately (Chia et al., 2020; Dong et al., 2021; Hebbalaguppae Krishnashetty et al., 2021; Kumar et al., 2012; Mokari et al., 2022). Some examples of these computing models include adaptive neuro-fuzzy inference (Tao et al., 2018), artificial neural network (Sanikhani et al., 2018), gene expression programming (Muhammad et al., 2021), multivariate adaptive regression spline (Adnan et al., 2021), hybrid dynamic evolving neural fuzzy inference system (Ye et al., 2022), extreme learning machine (Granata, 2019), support vector machine (Ferreira et al., 2019), XGBoost algorithm (Han et al., 2019), gradient boosting decision tree (Zhao et al., 2021), and several version of hybrid ML algorithms tuned using nature-inspired algorithms (Roy et al., 2021; Ruiming and Shijie, 2020; Tikhamarine et al., 2019).