Research progress and prospects of total water consumption prediction

Abstract

[Objective] Total water consumption control serves as a key indicator of the most stringent water resource management system. Faced with continuously fluctuating water demand, traditional single prediction models fail to effectively capture such dynamic adjustments. [Methods] To enhance the accuracy of total water consumption prediction, based on current research progress in this field, the development trends of existing prediction method for total water consumption and their applicability are systematically reviewed from the dimensions of statistical patterns, water consumption mechanisms, water consumption quotas, and model construction. [Results] The results show that China's early prediction result of medium-and long-term total water consumption are generally higher than the actual data. Water consumption predictions based on statistical patterns rely heavily on historical data, demonstrating better reference value for short-term predictions. The prediction method based on water consumption mechanism is suitable for experimental sites or at relatively small-scale areas, but cannot be directly applied to macro-planning. The prediction method based on water consumption quota is applicable to medium-and long-term, regional predictions for different water users. The model construction-based prediction method, while requiring high-quality data and strong dependence, demonstrates high prediction accuracy and strong adaptability. [Conclusion] Early predictions of long-term total water consumption suffer from limited accuracy. Furthermore, the water consumption trend in China demonstrates nonlinear evolution characteristics, and the existing prediction method have limitations in applicability and accuracy. Therefore, future research should focus on comprehensive integration of multiple method or multilayer model construction to enhance scientific validity and practical applicability.

Keywords

total water consumption / water consumption prediction / applicability / research progress / water resources / nonlinearity / artificial intelligence / machine learning