Brazil's National Interconnected System (SIN) is now one of the world's largest wind markets, with a fleet that has nearly doubled since 2020 and is overwhelmingly concentrated in the Northeast and South subsystems. As wind penetration deepens, fast variations in aggregate generation - wind ramp events - are emerging as a first-order operational constraint, driving reserve sizing, flexible dispatch and electrically-driven curtailment. Yet a country-wide, plant-level characterization of ramp frequency, severity and seasonality based on official operator data has not been published for Brazil.

This paper closes that gap with the first open ramp event catalog for the SIN, built from four years (2022-2025) of 30-min plant-level generation and capacity records published by the Brazilian National System Operator (ONS). The catalog is produced by a streaming detection pipeline that scans every consecutive 30-min pair per plant and registers a ramp whenever the absolute change in active power exceeds 20% of the local installed capacity. Each event is tagged with timestamps, magnitude, ramp rate, percentage of capacity, host state and plant; names are normalized across heterogeneous spellings and overlapping events are deduplicated.

Results are aggregated by year, state, season and hour of day. The central finding is a sharp geographic concentration: the Northeast states (Bahia, Rio Grande do Norte, Ceara and Piaui), together with Rio Grande do Sul, account for the overwhelming majority of detected events, with per-plant ramp frequency peaking in the late afternoon and early night - a diurnal signature consistent with the local sea-breeze regime. Counter to the intuition that wind-rich regions also experience the strongest ramps, frequency and severity decouple: peak severities occur in the transition seasons rather than in the windiest months, and the most ramp-prone region is not the most severe one. We term this the Northeast Frequency-Severity Decoupling.

A second structural finding concerns the temporal trend: both the absolute number of events and the per-plant rate grow year on year, yet the per-plant ramp rate stays roughly stable, indicating that the signal is being inflated by fleet expansion - not by wind variability - and will keep growing on the current capacity-addition trajectory regardless of meteorology. These results suggest that ramp-aware planning anchored to national aggregates will systematically underserve the regions where operational stress is most acute. Targeted reserve, storage and ramp-product design in the Northeast addresses the binding constraint more effectively than system-wide averaging.

The pattern is likely to recur in any large, geographically concentrated wind market as penetration deepens. The work is based on field data (operational records published by ONS) processed through original analysis developed as part of a research project.