Event abstraction is the process of summarizing or simplifying a complex sequence of events into higher-level descriptions. In the domain of process mining and process modeling, it enables us to create more understandable and manageable models by reducing the level of detail and complexity. Here are ways event abstraction can improve readability and manageability of complex process models:

1. Reducing Complexity: Large-scale logs that involve thousands of events can be made easier to understand through event abstraction by simplifying the represented processes. Reduction of these details reduces the complexity of the derived models and makes it easier to focus on primary processes and their interactions.

2. Highlighting Major Events: By abstracting events, major events that significantly contribute to the process model's understanding can be easily separated from minor or exceptional events. This makes it easier for a viewer to identify key processes point and their interactions.

3. Grouping Related Events: Event abstraction supports the grouping of related events under more general categories. This allows the viewer to understand the relationships and commonalities among different events and helps to provide a more structured and compact view of the process model.

4. Improving Scalability: Abstraction helps improve the scalability of process models by allowing the developer to selectively expose the details of the process. This means that models can be tailored to different audiences or objectives, showing all the details when necessary or just the big picture when that's all that is required.

5. Maintaining Consistency: By abstracting the events into more general classes or categories, the event logs can maintain consistency. Abstraction can help maintain a standard level of detail across different processes, making it easier to compare and manage multiple process models.

6. Enabling Hierarchical Analysis: Event abstraction allows for hierarchical analysis where the viewer can navigate from a high, abstract level to lower levels with more detailed events. This facilitates a more manageable approach to the analysis of complex process models.

In conclusion, through event abstraction, complex process models can be simplified, making them more readable and manageable. This makes the understanding, communication, analysis, and improvement of these models significantly easier.