An overview of the model integration process: From pre-integration assessment to testing.

Area of Science:

• Environmental science
• Computational modeling
• Software engineering

Background:

• Model integration is crucial for complex environmental science research but faces challenges due to diverse tools, methodologies, and assumptions.
• Limited reuse of existing frameworks and sharing of science components hinder the development of integrated modeling systems.

Purpose of the Study:

• To enhance understanding of the model integration process through a comprehensive literature review.
• To present improved strategies for constructing robust integrated modeling systems.
• To identify key factors promoting reuse and interoperability in scientific software.

Main Methods:

• Literature review to analyze existing model integration processes.
• Characterization of the model integration process into five distinct phases.
• Comparison of integration practices across different research groups.

Main Results:

• Identified five phases of model integration: pre-integration assessment, model preparation, simulation orchestration, data interoperability, and testing.
• Observed a general lack of cross-team framework reuse and science component sharing.
• Highlighted the need for strategies to increase reuse and interoperability.

Conclusions:

• Effective model integration requires addressing challenges in interoperability and component sharing.
• Adoption of specific strategies, features, and standards can foster greater reuse of scientific software.
• Facilitating complex workflows is essential for leveraging current environmental science advancements.