Sobri Harun

Reference evapotranspiration (ET_o) is a crucial component in agro-meteorological processes, where accurate estimation is imperative for planning and managing irrigation practices. ET_o and rice irrigation requirements were first estimated using FAO Penman–Monteith (FAO-PM56) and the water balance model, respectively, and the obtained results were used as reference values in the machine learning algorithms. Two machine learning algorithms, named Support Vector Regression (SVR) and Random Forest (RF), were applied to predict ET_o and rice irrigation requirements using only climatic data (rainfall, temperature, relative humidity, and wind speed). The novelty of this paper is the application of machine learning techniques as an alternative to traditional methods and software solutions for estimating ET_o and irrigation demand. Emphasis is placed on the advantages of data-driven models, such as eliminating the need for internal framework factors, facilitating realistic calculations, and providing more straightforward solutions for handling multiple variables. A comparative study is conducted to evaluate the performance of SVR and RF. Comparison results indicated that the SVR model outperforms the RF model in predicting ET_o and irrigation requirements, with the minimum error observed in both the training and testing phases. All coefficient determination (R ²) values are generally higher than 0.90 except for RF during the testing phase in ET_o prediction, where the R ² value is 0.82. Alternatively, the excellent performance of SVR over RF for ET_o and irrigation requirements prediction in the training phase was also demonstrated by the mean absolute error (MAE) and root mean squared error (RMSE). The results have demonstrated that machine learning techniques, particularly the SVR model, can be confidently utilized as alternative methods for predicting ET_o and irrigation requirements, even with limited input data, with high accuracy.

Seasonal variations in the hydrochemistry of shallow groundwater can be due to the interactions between internal geochemical processes and external factors such as rainfall and human activities. This study applied seasonal and multivariate statistical analyses to understand the hydrochemical controls on shallow groundwater. The study area was divided into three sub-basins, or clusters (i.e., Birnin kebbi, Sokoto and Gusau). Fifteen shallow groundwater samples were derived from each cluster, totalling ninety shallow groundwater samples for dry and wet seasons. Physical parameters, including Temperature, Electrical Conductivity (EC), Total Dissolved Solids (TDS), Dissolved Oxygen (DO), and pH, were analysed in situ using handheld metres. However, chemical parameters (Ca²⁺, Mg²⁺, Na⁺, K⁺, Fe³⁺, Cu²⁺, Zn²⁺, CO₃^-, HCO₃^-, Cl^-, SO₄^2-, PO₄^3-, NH₃ and NO₃^-) were analysed in the laboratory. Subsequently, statistics were applied to study the impact of seasonality and groundwater evolution. Results of the Kruskal-Wallis test revealed that seasonality exerts a considerable influence on shallow groundwater through a significant difference in Temperature, EC, DO, TDS, HCO₃^-, Cl^-, NH₃ and PO₄^3-. Pearson’s correlation analysis revealed strong relationships between hydrochemical elements, which suggest natural and anthropogenic influences on shallow groundwater evolution. Correlation results were concurrent with principal component analysis (PCA), hierarchical clustering analysis (HCA), and Piper and Gibbs models. Therefore, this study inferred that seasonality and rock weathering are the primary mechanisms controlling shallow aquifers' hydrochemistry in a semiarid Sokoto Basin. The seasonal and multivariate statistics provide a framework for more accurate shallow groundwater quality analysis while considering multiple groundwater quality parameters under different environmental conditions. It is hoped that the results of this study will inspire other researchers to use a similar method, especially those in semiarid environments. Seasonal and multivariate statistical analyses provide a user-friendly tool for monitoring shallow groundwater quality monitoring systems in global semiarid environments.

Reliable projections of evapotranspiration (ET) are important for agricultural and water resources development, planning, and management. However, ET projections using well established empirical models suffer from uncertainty due to their dependency on many climatic variables. This study aimed to develop temperature-based empirical ET models using Gene Expression Programming (GEP) for the reliable estimation and projection of ET in peninsular Malaysia within the context of global warming. The efficiency of the GEP-generated equation was compared to the existing methods. Finally, the GEP ET formulas were used to project ET from the downscaled and projected temperature of nine global climate models (GCMs) for four Representative Concentration Pathways (RCPs), namely, RCP 2.6, 4.5, 6.0, and 8.5, at ten locations of peninsular Malaysia. The results revealed improved performance of GEP models in all standard statistics. Downscaled temperatures revealed a rise in minimum and maximum temperatures in the range of 2.47–3.30 °C and 2.79–3.24 °C, respectively, during 2010–2099. The ET projections in peninsular Malaysia showed changes from −4.35 to 7.06% for RCP2.6, −1.99 to 16.76% for RCP4.5, −1.66 to 22.14% for RCP6.0 and −0.91 to 39.7% for RCP8.5 during 2010−2099. A higher rise in ET was projected over the northern peninsula than in the other parts.