Open Data, AI-Readiness, and Building a Learning Data System

By: Robyn Grimm and Mitch Russo

The theme “Data Drives Us” recently took center stage at the California Department of Water Resources (DWR) Environmental Science Workshop, highlighting the vital role of high-quality, accessible data in advancing sustainable water management across California. As the state faces growing water challenges, collaborative efforts between the California Water Data Consortium, DWR, and other partners are laying the groundwork for a future rooted in transparency, innovation, and collaboration.

Open and FAIR Data in Decision-Making and AI Readiness

At the workshop, Mitch Russo from DWR’s Division of Planning highlighted important approaches DWR and other agencies developed to ensure open data adheres to FAIR principles (Findable, Accessible, Interoperable, Reusable) and continuously evolving AI-readiness standards. These agency efforts are helping make water and ecological data a more reliable and accessible resource for decisionmakers and cutting-edge analytical tools.

Key Features of Open, FAIR, and AI-Ready Data

• Data Quality: Prioritizes completeness, consistency, timeliness, and integrity while addressing potential biases.
• Accessibility: Publishes data in versatile formats with transparent licensing and secure delivery options.
• Documentation: Provides robust metadata and data dictionaries to ensure seamless use and integration.

DWR also developed internal documents such as the Open Data Handbook and the Open Data Publishing Guide to support these efforts and formed the Open Data Working Group to foster dialog around open data topics. The publications help empower to implement best practices, ensuring California’s water and ecological data are both user-friendly and innovative.

Community Feedback to Build a Learning, Trusted Data System

Achieving open, trustworthy, and AI-ready water and ecological data also depends on a foundation of open communication. California’s data systems must be designed for continuous improvement, collaboration, and responsiveness to user feedback. This means rethinking the traditional data life cycle, which historically moves from collection to publication with limited opportunities for iterative refinement or user engagement.

A learning data system integrates community feedback and rigorous quality checks throughout the data process, ensuring relevance, reliability, trust, and alignment with end-user needs.

Six Key Elements of a Learning Data System

1. Dedicated Capacity and Ongoing Opportunities for Dialogue: Allocates resources for engagement with involved parties and regular discussions to build alignment, transparency, and trust.
2. Well-Documented and Contextualized Data: Ensures clear metadata and documentation enhances usability and relevance for diverse audiences.
3. Educational and Training Resources: Connects scientists and end-users through workshops, training sessions, and accessible materials to bridge knowledge gaps.
4. Feedback Mechanisms: Establishes intuitive systems for users to flag issues, provide insights, and suggest improvements, supported by an actively maintained contact management system.
5. Issue Prioritization Systems: Systematically evaluates and addresses flagged issues to drive timely and meaningful improvements.
6. Funding for Research and Data Refinement: Supports ongoing innovation and adaptation of data systems to emerging challenges and quality standards.

By embedding these principles, state agencies and other data providers can ensure data systems remain adaptive and inclusive, integrating insights from the communities most affected by data-driven decisions. This dynamic feedback loop improves data quality and usability, making open data an invaluable tool for sustainable water management and scientific progress.

OpenET as an Example of a Learning Data System

Before joining the Consortium, Robyn Grimm worked at the Environmental Defense Fund and helped launch OpenET, a platform designed to provide open-source evapotranspiration (ET) data. Before its public release, the OpenET team developed a contact management system to help collect and respond to community feedback, ensuring data accuracy and addressing concerns in real time. The team provided end users with clear instructions on how to flag issues to help them understand why certain data in particular locations may be raising concerns.

Following the platform’s public launch, the team received a lot of enthusiastic support from people who were excited about the promise of OpenET. However, they also received some negative criticism from users flagging inconsistencies between the ET data and their findings in the field. Initially, this negative feedback was difficult to hear, but it quickly became an invaluable resource. The OpenET science team systematically analyzed flagged issues, identified patterns, and—where concerns were valid—experimented with ways to improve the ET models and overall quality of data.

This experience demonstrated the power of open science and data collaboration – accelerating scientific advancements and continuously refining data quality. Just as OpenET’s community feedback mechanisms strengthened data reliability, implementing these learning data system elements across the state will ensure California’s water data is continuously refined to meet real-world needs. More importantly, the dialogue between scientists and end users fostered trust, reinforcing the credibility of these data. This principle of open, ongoing exchange is central to the Consortium’s mission and approach to California’s water and ecological data systems.

Conclusion

As California continues to navigate diverse water sustainability challenges, the importance of open, FAIR, and AI-ready data cannot be overstated. Initiatives like DWR’s push for enhanced data quality and accessibility, combined with the principles of a learning data system, represent significant strides toward a more transparent and collaborative future. By prioritizing community feedback, robust documentation, and ongoing refinement, California’s water data systems are better positioned to support decisionmakers, researchers, and the communities that rely on them.

Collaboration among agencies, researchers, and resource managers will be key to ensuring California’s water and ecological data remains a trusted, actionable resource for the future.

Special thanks to Division of Planning Program Manager Mitch Russo at the California Department of Water Resources for his insightful presentation and permission to include content from his work in this blog