If browser security blocks the API, save the JSON response from the NOAA links and select it here. The dashboard will parse it without a server.
NOAA data is fetched directly from your browser. If a source fails, open its API request below to see NOAA’s error message. A working NOAA connection is required for automatic plotting.
USGS 01578310 · Susquehanna River at Conowingo · discharge (cfs). The lower chart shows measured discharge and signed current residual on separate normalized scales, for timing comparison only. No causation is implied.
Requesting USGS discharge…
Correlation compares discharge at time t − lag with observed-minus-predicted flood-axis velocity at time t. This exploratory statistic is sensitive to tidal periodicity, autocorrelation, wind, sampling, and the selected lag. Do not interpret a high correlation as proof of dam influence. NOAA prediction depth and observed bin may differ. Open USGS API ↗
NOAA NDBC station 44063 (Annapolis-area buoy). Wind direction is the direction from which wind blows. The model uses northward and eastward wind components, plus lagged dam discharge, to explain the signed CB1102 current residual. This is exploratory association, not a validated forecast.
Requesting NDBC 44063 wind observations…
R² is the fraction of variation explained in this same short sample; a higher in-sample R² does not establish causation or better predictions. Wind components are signed (wind blowing toward north/east). Models include an intercept; no wind or discharge values are fabricated. NDBC 44063 station ↗
This compares observed flow and harmonic prediction at CB1102. The residual is the difference in velocity projected onto the predicted flood/ebb axis; it is not a complete two-dimensional wind-driven current vector. Currents at the start line or Thomas Point may differ from CB1102. The observation bin and prediction depth must be comparable before treating the residual as a precise physical measurement. Station outages and browser cross-origin restrictions may prevent loading; use the NOAA source links above to diagnose.
Auto-refresh: Every 10 minutes while this page remains open. The file itself needs no installation or account, but requires internet access. If you close the browser, it stops refreshing. Historical observations from the supplied NOAA export are bundled, so the measured series remains visible offline; new measurements and predictions require internet access. The timestamp is displayed in the chosen time zone.
Every 10 minutes while this page is open, check NOAA currents/predictions, USGS Conowingo discharge, and NDBC 44063 wind. New records are merged into an IndexedDB history in this browser; historical forecast scores and decay calibration recalculate when new observations arrive. No server or account required.
Opening browser history…
Important: Storage is tied to this browser and file origin; opening the dashboard at a different file path may use a different storage area. Back up the history before moving or replacing the HTML file. Some browsers restrict storage for local files or block live API requests. Data collection stops when the page is closed. NOAA is still the preferred forecast unless historical validation qualifies an alternative.
Official Chesapeake Bay Operational Forecast System (CBOFS) guidance is a separate physical-model forecast, not the CB1102 astronomical tide prediction. The dashboard discovers published NOAA CBOFS datasets, checks for browser-accessible numerical point services, and attempts a verified east/north velocity extraction when supported. It rechecks every 30 minutes while open. Browser cross-origin restrictions and NOAA dataset availability may prevent retrieval; failures are shown explicitly. Manual import remains available as a fallback. The dashboard does not pretend that map imagery is station-level velocity data.
Import CBOFS point time series (CSV) extracted at or near CB1102 and a comparable depth. Columns: time_utc,velocity_flood_kt (positive floodward, negative ebbward, aligned to NOAA CB1102's 30° flood axis). Alternatively time_utc,eastward_ms,northward_ms for vector currents; the dashboard projects them onto the 30° axis. Times must be ISO 8601 with Z or UTC offset. Do not import a different site/depth as if it were CB1102.
NOAA dataset discovery not yet checked.
Not checked yet.
NOAA catalog: Browse published CBOFS files ↗. CORS restrictions on a local HTML file cannot be bypassed by JavaScript. This tool does not use third-party proxy services or claim that grid-relative ROMS velocities are geographic east/north currents.
Extraction settings: 38.9800° N, 76.3900° W (CB1102 vicinity); request eastward/northward velocity components at approximately 0–3 m below surface, when available. NOAA model grid cells and depth layers may not coincide with the CB1102 instrument. The comparison remains experimental.
Imported point forecasts are retained in the same browser's IndexedDB history when available. Historical errors compare model values to observed signed currents within 20 minutes of matching timestamps. CBOFS is displayed as a separate experimental comparison; it is not automatically promoted to the preferred model without archived, issue-time-correct forecasts and robust out-of-sample testing.
Uses the previous 24 hours of observed current anomalies and wind/dam conditions. Future wind and discharge are held at their latest 3-hour average (persistence assumption). Historical backtests use only data available at each forecast origin. Experimental—not a navigational forecast.
Searches historical forecasts made with the same rolling 24-hour model, using only information available at each forecast origin. Compares predicted corrections near today's value with the observed correction at the target time. This is a historical test, not a claim of forecast skill.
Measured minus NOAA tidal current, signed along the flood/ebb axis. Negative = more ebbward than NOAA. This is an observation, not a forecast.
Safety rule: Until the adjusted model consistently outperforms NOAA on independent historical data, NOAA remains the primary planning forecast. A correction larger than the tidal prediction can reverse flood/ebb and is flagged as experimental.
A recent observed ebb/flood departure may fade with time. Candidate half-lives are selected using only earlier completed forecasts at each historical issue time. Each horizon is assessed separately. NOAA is automatically preferred unless an alternative demonstrates sufficient out-of-sample improvement.
All available historical forecasts—not just forecasts resembling today. Every row compares a forecast reconstructed at the issue time with a later measured current. Positive correction = floodward; negative = ebbward.
Confidence refers only to the quantity and consistency of historical tests. It is not a probability that the current forecast is correct. Tests may overlap in time and are not statistically independent.