Incorporating interaction terms in multivariate linear regression for post-event flood waste estimation.

Area of Science:

• Environmental Science
• Disaster Management
• Statistical Modeling

Background:

• Conventional flood waste models assume independent input variables, which may not reflect real-world conditions.
• Accurate flood waste estimation is crucial for effective disaster response and mitigation strategies.

Purpose of the Study:

• To evaluate the effectiveness of including interaction terms in flood waste modeling.
• To identify optimal strategies for flood waste mitigation.
• To provide a plausible explanation for the modeling results.

Main Methods:

• Statistical analysis of ninety flood cases in South Korea.
• Application of multivariate linear regression with interaction terms.
• Selection of input variables from a national disaster information system.

Main Results:

• Incorporating interaction terms enhanced the accuracy of flood waste estimation models.
• Mitigating damage to rivers and croplands was identified as the most effective way to reduce flood waste.
• Interaction terms explained the nonlinear response of waste generation and compensated for single-term estimation errors.

Conclusions:

• Including interaction terms is an effective method to improve flood waste estimation models without extensive additional data collection.
• The findings provide practical insights for flood waste mitigation strategies.
• Field observations corroborated the nonlinear and interactive patterns observed in the regression analysis.