IJA_2026v16n5

International Journal of Aquaculture, 2026, Vol.16, No.5, 378-396 http://www.aquapublisher.com/index.php/ija 389 operation of digital platforms, network communication, data storage, software upgrades, and remote services also require continuous investment. IoT-related reviews emphasize that connectivity in remote ponds often relies on low-power wide-area networks, satellite networks, or mobile networks, whose deployment and service fees may further increase costs; at the same time, ensuring stable power supply for devices may require additional energy systems. AIoT assessments further indicate that data processing infrastructure, as well as energy-intensive computing and communication, may also bring certain environmental and economic burdens (Tina et al., 2025). In addition, the cost-effectiveness of digital monitoring equipment varies across different aquaculture scenarios. High-density aquaculture, high-value aquatic product farming, and large-scale ecological aquaculture have stronger demands for environmental monitoring, risk warning, and process traceability, so digital investment is more likely to be transformed into economic returns. However, in low-density, extensive, or seasonal aquaculture scenarios, equipment utilization may be insufficient and the investment payback period may be relatively long. For ecological pond aquaculture, if profit margins are limited and ponds are numerous and spatially dispersed, the accumulated costs of sensors, networks, power supply, and maintenance services may slow down promotion and application even when the technical benefits are recognized. Therefore, the promotion of digital monitoring systems should not simply pursue complete equipment configuration, but should select appropriate technical models and equipment combinations according to aquaculture scale, species value, management needs, and economic affordability. 5.2 Insufficient data accuracy and system stability The application effect of digital monitoring largely depends on data accuracy, continuity, and system stability. However, in ecological pond aquaculture environments, monitoring equipment faces complex operating conditions. Factors such as turbid water, algal proliferation, sediment disturbance, temperature changes, biofouling, and long-term immersion may all affect sensor detection results. Many studies have shown that although low-cost sensors help reduce the threshold for application, they are also susceptible to drift, biofouling, and environmental interference. For example, if monitoring probes for indicators such as dissolved oxygen, pH, and ammonia nitrogen are not cleaned and calibrated in time, data drift or errors may occur, resulting in inaccurate system alarms. Analyses of smart sensor performance in aquaculture have indicated that electrochemical dissolved oxygen and nutrient sensors may show measurable deviations over time due to biofilm accumulation and electrode aging, thus requiring frequent recalibration to avoid delayed or incorrect management responses (Zamnuri et al., 2024). Insufficient system stability is also a practical issue in the promotion of digital monitoring. Pond aquaculture is mostly located in rural or suburban areas, where network signals may be unstable in some regions, easily causing data transmission delays, platform disconnection, or failure of remote control. Under special weather conditions such as thunderstorms, high temperatures, and humidity, outdoor equipment may also experience line aging, abnormal power supply, terminal damage, and other problems. If the system cannot maintain stable operation, the continuity and early-warning function of digital monitoring will be affected. To improve data integrity, many system studies have adopted pre- and post-calibration procedures, outlier rejection algorithms, cross-validation with professional instruments, or regression-based correction of low-cost sensor data (Lin et al., 2021). However, even with these measures, IoT and AIoT applications in aquaculture still face problems such as data loss caused by equipment or network failures, unstable connectivity, and uncertainty in predictive models, all of which may affect the reliability of real-time decision-making (Tina et al., 2025). At the same time, some digital monitoring systems still have insufficient data analysis capabilities. At present, some systems mainly remain at the stage of data collection, indicator display, and simple threshold alarms, and their ability to make differentiated judgments under different cultured species, seasons, and ecological pond conditions is limited. For example, the same dissolved oxygen level may represent different degrees of risk under different water temperatures, stocking densities, and cultured species. If the system only issues alarms based on fixed thresholds, false alarms or missed alarms may occur. In ecological pond applications, management decisions

RkJQdWJsaXNoZXIy MjQ4ODYzNA==