Advances and integrations in wireless communications and sensor technology have promoted developments in wireless sensor networks (WSNs) and their applications in various fields. In large-scale WSNs used for environment monitoring, the sensors are often unevenly distributed, and issues of node failure, power loss, faulty data transmissions, and gaps in sensor coverage are common. These issues, referred to as sparsity and non-uniformity, may affect the accuracy of anomaly event detection. This study thus applied the catastrophe theory to establish potential functions for the environmental data collected from WSNs to make up for the missing and faulty data caused by sparsity and non-uniformity before conducting prediction and analysis.