Water data does not flow freely through governance systems. It accumulates in agencies, is defended by institutions, and becomes a source of power for those who control access to it. See The Fortress of the Weak for the related pattern of mandate defence as institutional survival strategy. See Fragmentation as Inherited Design for the structural condition that makes data territoriality so difficult to resolve.

Water data collected with public resources becomes the de facto property of the agency that hosts it — and whoever controls access to the data controls a form of power that no mandate explicitly granted them.

The mechanism of capture is mundane. An agency is given the task of collecting rainfall records, river discharge measurements, or groundwater levels. The data accumulates. The agency develops the systems to store and process it. Over time, the boundary between custodianship and ownership quietly dissolves. Access becomes a negotiation. Requests from other agencies are delayed, qualified, or priced. The data that was collected to understand the system becomes a resource the agency uses to sustain itself within it.

This is not always cynical. Agencies that host data invest real resources in its collection, quality control, and storage. The claim to ownership is not invented from nothing. But the effect is the same regardless of the motivation: public data becomes institutional property, and the public interest in its circulation is subordinated to the institutional interest in its control.

The second force is more explicitly political. Water data is not neutral. A discharge record reveals how much water actually reached a downstream user compared to what the allocation scheme promised. A groundwater level time series shows whether abstraction is sustainable or whether the aquifer is being mined. An infrastructure dataset shows which embankments have not been maintained and where the next failure is most likely. Sharing this data openly is not an administrative act — it is a political one. It hands rivals, regulators, and affected communities the evidence they need to challenge existing arrangements. The agency that shares honestly may be sharing the case against itself.

The third force is the one least acknowledged in international water debates. The donors and global north institutions that push hardest for open data in developing country water sectors are frequently the same institutions whose own data sharing practices are anything but open. Commercial satellite data is licensed and priced. Proprietary flood models are sold, not shared. Research datasets are restricted pending publication. The demand for openness travels in one direction — from funder to recipient — and the recent history of data sharing within the donor community's own world is quietly set aside when the conversation turns to transparency elsewhere. The open data agenda is not wrong. But it is selectively applied, and that selectivity is noticed by the people it is applied to.

Treat data sharing as a political negotiation, not an administrative request. Address institutional capture by establishing clear legal frameworks that distinguish custodianship from ownership — the agency that hosts public data does not own it, and that distinction must be written into mandate, not assumed. Address strategic concealment by creating graduated sharing arrangements that allow sensitive allocation and infrastructure data to be shared within defined governance boundaries before it is shared openly — full transparency is a destination, not a starting condition. And address the hypocrisy of the demand directly: donors and international partners who advocate for open data must apply the same standard to their own models, datasets, and research outputs, or accept that the advocacy will be read as an instrument of power rather than a principle.

Linked patterns: The Fortress of the Weak — data territoriality is mandate defence by other means. Fragmentation as Inherited Design — fragmented institutions fragment data as a structural consequence. Governance Cannot Reflect on Itself — the agencies most resistant to data sharing are the least able to evaluate what that resistance costs. The Incomplete Attribution — restricted data makes honest causal attribution after disasters structurally impossible. The Two-Faced Map — the strategic concealment of water data and the management of national narrative are driven by the same political logic.