We discuss the accuracy and performance of the adaptive neuro-fuzzy inference system (ANFIS) in training and prediction of dissolved oxygen (DO) concentrations. The model was used to analyze historical data generated through continuous monitoring of water quality parameters at several stations on the Johor River to predict DO concentrations. Four water quality parameters were selected for ANFIS modeling, including temperature, pH, nitrate (NO3) concentration, and ammoniacal nitrogen concentration (NH3-NL). Sensitivity analysis was performed to evaluate the effects of the input parameters. The inputs with the greatest effect were those related to oxygen content (NO3) or oxygen demand (NH3-NL). Temperature was the parameter with the least effect, whereas pH provided the lowest contribution to the proposed model. To evaluate the performance of the model, three statistical indices were used: the coefficient of determination (R (2)), the mean absolute prediction error, and the correlation coefficient. The performance of the ANFIS model was compared with an artificial neural network model. The ANFIS model was capable of providing greater accuracy, particularly in the case of extreme events.
Global warming has led to sea levels raise (SLRs) and Malaysia is no exception to this problem. Especially for low-lying coastal areas including the Kuala Kedah area which is active in agricultural and fisheries activities. Farmers have had to bear up to 75 % of yield losses due to seawater breaches since 2016. Therefore, this study is designed to assess the impact of seawater encroachment on water quality through spatial technology approaches and hydrodynamic modeling related to the growth of paddy trees. The study was conducted during two different paddy cultivation seasons namely Season 1-2019 and Season 2-2019 which take place in the southwest and northeast monsoon in Kuala Kedah, Malaysia. The study involved three phases, which are the assessment of salinity and pH concentration levels, the assessment of the health of paddy crops through multispectral image analysis involving three plant indices (VI), namely Normalized Difference Vegetation Index (NDVI), Blue Normalized Difference Vegetation Index (BNDVI) and Normalized Difference Red Edge (NDRE), and finally, the assessment of the impact of SLR through the numerical method in MIKE 21 for hydrodynamic modeling considering two conditions that are without mitigation factor (K1) and with existing mitigating factor (K2). According to the findings, the salinity concentration trend is decreasing across the growth stage during Season 1-2019, whereas it is the contrary during Season 2-2019. It was discovered that during the study period for both tidal events, 73 % of the 44 sampling points in Season 1-2019, as opposed to just 3 % in Season 2-2019, were categorized as Class 4 and Class 5. Even though there were fluctuations throughout the observation, the pH reading is still within the allowed range of 6.5 to 9.0 for the estuary area. Following that, the ANOVA analysis proved that salinity concentration a statistically significant difference with tidal variations and pH levels. Moreover, the multispectral image analysis findings revealed that the VI value was correlated with both the yield and the health of the rice crop, with R-square values of 0.842 compared to 0.706 and 0.575 for NDVI and BNDVI values, respectively. It confirmed that NDRE granted a more accurate and reliable measurements. Additionally, the hydrodynamic simulation results demonstrated that, if the mitigation factors were considered in the modeling, overflow seawater to the mainland could be reduced by up to 20 %, reducing the impact of coastal flooding on the local area as well as the nearby rice cultivation area. Ultimately, these three elements-water quality, vegetation index, and hydrodynamic modeling-can assist in identifying the underlying cause of the problem and develop short and long-term solutions.