This study aimed to assess suitable water level sensor types and implement the automated monitoring of water levels within a Class A pan evaporation system using the Internet of Things (IoT). Both analogue and ultrasonic water level sensors underwent testing in controlled laboratory conditions for performance analysis. The results showed that the analogue water level sensor exhibited suboptimal output sensor responses compared to the ultrasonic sensor, primarily due to its susceptibility to variations in solution types and immersion depths. In contrast, ultrasonic sensors demonstrated strong performance with acceptable error rates, as evidenced by the Mean Absolute Error (MAE) of 1.03, Root Mean Squared Error (RMSE) of 1.42, and Coefficient of Determination (R²) of 0.94 during laboratory testing. However, the ultrasonic sensor's performance was somewhat reduced during field testing, exhibiting accuracy levels ranging from 6.7% to 51.2% within a greenhouse environment during rock melon cultivation. These discoveries highlight the feasibility of using ultrasonic sensors with environmental calibration to automate real-time evaporation measurements towards precision irrigation practices.
