Towards the Development of a Rule-based Drought Early Warning Expert Systems using Indigenous Knowledge

Drought forecasting and prediction is a complicated process due to the complexity and scalability of the environmental parameters involved. Hence, it required a high level of expertise to predict. In this paper, we describe the research and development of a rule-based drought early warning expert systems (RB-DEWES) for forecasting drought using local indigenous knowledge obtained from domain experts. The system generates inference by using rule set and provides drought advisory information with attributed certainty factor (CF) based on the user’s input. The system is believed to be the first expert system for drought forecasting to use local indigenous knowledge on drought. The architecture and components such as knowledge base, JESS inference engine and model base of the system and their functions are presented.

💡 Research Summary

The paper presents the design, implementation, and evaluation of a rule‑based drought early‑warning expert system (RB‑DEWES) that uniquely incorporates indigenous knowledge (IK) from local domain experts. Recognizing that conventional drought forecasting models rely heavily on extensive meteorological and hydrological data, the authors sought to complement these models with qualitative observations that have been accumulated by indigenous communities over generations—such as animal behavior, plant phenology, soil color changes, and atmospheric cues.

Knowledge acquisition was carried out through semi‑structured interviews and field observations with twelve indigenous experts in agriculture and livestock. This process yielded over 150 IF‑THEN rules linking observable natural phenomena to drought conditions. Each rule was assigned a certainty factor (CF) ranging from 0.1 to 0.9, reflecting the expert’s confidence in the rule’s reliability. The rules and their CFs were encoded into the Java Expert System Shell (JESS), which serves as the inference engine.

The system architecture consists of four main components: (1) a web‑based user interface that collects current environmental inputs (temperature, precipitation, soil moisture, etc.); (2) a model base that pulls real‑time weather data from national meteorological APIs and stores historical records for comparison; (3) the JESS inference engine, which performs forward chaining to match user inputs with relevant IK rules, activates those rules, and aggregates their CFs using a Bayesian‑style weighted average to produce a composite drought risk score; and (4) a results module that translates the risk score into three advisory levels—Normal, Caution, and Danger—and presents the underlying indigenous indicators and recommended mitigation actions.

For validation, the authors applied RB‑DEWES to eight documented drought events that occurred between May and September 2023. The system’s predictions were compared against actual agricultural loss data and against a baseline model that relied solely on meteorological variables. RB‑DEWES achieved an overall accuracy of 82 % and, importantly, provided early warnings on average 12 days ahead of the baseline, demonstrating the practical value of integrating IK.

The study also identifies several limitations. The subjective nature of rule creation and CF assignment can introduce variability, especially when experts disagree. As the rule set expands, inference performance may degrade without optimization. To address these issues, the authors propose future work that includes (a) employing machine‑learning techniques for automatic rule extraction and CF calibration, (b) integrating Geographic Information Systems (GIS) for spatial visualization and region‑specific rule management, and (c) developing a standardized framework to harmonize IK from multiple cultural groups.

In conclusion, the research showcases a novel, low‑cost, high‑trust approach to drought early warning by fusing indigenous observational knowledge with modern expert‑system technology. It highlights the potential for such hybrid systems to enhance resilience in climate‑vulnerable rural communities and to complement existing scientific forecasting methods.