尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證When is the road to Lieutenant Island (Wellfleet, MA) under water? A tide forecast that corrects NOAA's predictions for sea-level rise and weather.
翻譯準備中
取自儲存庫 README
正在載入完整 README。 以下為已儲存的摘錄。
When is the road to Lieutenant Island (Wellfleet, MA) underwater?
Live at https://aliabuckner1.github.io/lieutenant-island/. The short link https://ltisland.netlify.app forwards there.檢視完整 README ↗When is the road to Lieutenant Island (Wellfleet, MA) underwater? Live at https://ltislandtides.com. The old address, https://aliabuckner1.github.io/lieutenant-island/, and the short link https://ltisland.netlify.app both forward there.
Residents of Lieutenant Island use printed tide calendars, built from NOAA's predictions for Wellfleet, to know when the road on and off the island will be flooded. Those predictions have grown less reliable. They're pinned to 1983–2001 sea level, and the sea at Boston is about 6½ inches higher now. They leave out wind and air pressure. And Wellfleet has no tide gauge of its own, so its numbers are Boston's, scaled and shifted. Today 94% of high tides come in above the prediction, by about 8 inches on average.
This site corrects for that. It adds today's sea level and a weather model (National Weather Service wind, Open-Meteo pressure) built from more than 3,000 past high tides, leans on NOAA's live Boston tide gauge for the next day or so, and compares the result with the road's height from USGS airborne lidar. Islanders can also report what they see at the road. Reports are recorded in a Google Sheet and can help pin down the road's real height and check the forecast against what actually happens at Wellfleet, so the model can be improved over time.
Replaying about 900 high tides from 2024–26 with the weather forecasts available at the time, against what Boston's tide gauge measured:
| Forecast | Typical error |
|---|---|
| NOAA's prediction (what the calendars use) | about 9 in |
| This site, a day ahead | about 4 in |
| This site, a week ahead | about 5 in |
The page runs entirely in the browser. On every load it fetches NOAA tide predictions, the live Boston tide gauge, the NWS wind
forecast and Open-Meteo pressure, then compares the result with the road's low point: 9.9 ft above MLLW, measured from the road on 15 Sep 2026
(nine tape-measure depths through one tide, data/road_measurements.csv). It had been 10.4 ft from lidar, good to about ±6 in.
No server and no build step. Between NOAA's high and low times the curve is Boston's, stretched to fit, with a quicker fall off the top: depths
measured at the road every minute or two (data/road_measurements.csv) fit Boston's shape on the rising tide but run about 14 min
ahead of it on the falling tide, so the curve after high water is pulled 14 min earlier. The USGS harbor gauge two miles away
(2017–22, scripts/wellfleet_shape.py, data/wellfleet_shape/results.txt) shows the same quicker fall, at about 8 min there.
| Folder | What it is |
|---|---|
apps-script/ | The Google Apps Script that receives road reports and adds each one to a Google Sheet. Setup steps are at the top of the file. |
site/ | The live site. index.html, tide-data.js (data fetching and the forecast model), base.css. Edit these directly. |
scripts/ | Python analysis: tide and weather history, the road profile, and backtest_fade.py, which replays 2024–26 tides to check the forecast. |
data/ | Downloaded data and results (see data/backtest/results.txt). |
archive/ | Not deployed. Earlier versions of the live page; design-options/, the mockups the current layout was chosen from; original-analysis/, the first one-page analysis; and field-card/, the September 2026 card for logging when water reached the road. |
GitHub Actions publishes the site/ folder to GitHub Pages on every push to main that changes it
(.github/workflows/pages.yml). It can also be run by hand from the Actions tab.
python3 scripts/fetch_backtest_data.py
python3 scripts/backtest_fade.py | tee data/backtest/results.txt