Reconciliation begins by aligning boundaries and definitions before comparing values.
Apparent conflicts often come from different definitions of diversion, delivery, use, storage, geography, or time rather than simple error.
That distinction matters because irrigation decisions cross several boundaries: a water record becomes a physical delivery, a delivery becomes a hydraulic condition, and that condition becomes a root-zone response. A number at one boundary cannot automatically prove performance at the next.
The mechanisms to keep visible.
Water-year and calendar-year records can shift totals across periods.
Facility, service-area, basin, and statewide aggregation answer different questions.
Revisions, estimates, rounding, and missing reports produce version differences.
Turn the pain point into a defined operating question.
Create a data dictionary and reconciliation table before choosing a preferred value.
Start with the decision the analysis must support. Then identify the smallest set of inputs capable of changing that decision. Keep source facts, calculations, model outputs, estimates, and operator observations distinct throughout the workflow.
Make the result reviewable.
Preserve original values, units, revisions, transformations, exclusions, and reasons for selection.
What this analysis cannot establish by itself.
Unresolved differences should remain visible and may limit the claim.
SIM treats an honest limitation as part of the result. Where evidence is incomplete, the correct status is inconclusive or unverified—not a more confident sentence.