Marine Layer Depth

How this is calculated

The marine layer is the cool, humid Pacific air mass trapped beneath a temperature inversion. Its depth is the height of the inversion base — the top of the cold air. This page measures the cold air directly and fuses five independent estimates:

1. Terrain ladder (primary, ~hourly): surface stations from sea level to 2,598′ on the coastal ranges. For each station we compute potential temperature (θ = T + 9.8°C per km of elevation) and compare it and the dewpoint against the coastal sea-level reference. Inside the marine layer θ and dewpoint match the coast (the layer is well mixed); above the inversion the air is several degrees warmer in θ and dramatically drier. The algorithm finds the cut height that best separates “marine” from “non-marine” stations, which brackets the inversion base.
2. Oakland radiosonde: the most recent 00Z/12Z balloon sounding, scanned upward for the first sustained temperature increase (the inversion). Direct measurement, but up to ~12–24 h old, weighted down with age.
3. NWS forecaster-stated depth: the Monterey forecast discussion frequently states a measured depth (forecasters see profiler and aircraft data). Parsed automatically, age-weighted.
4. Model profiles (HRRR + GFS): hourly virtual temperature/moisture profiles over a point ~20 km offshore. The inversion base is detected from the potential-temperature jump between levels, refined by the model’s own boundary-layer height. This supplies the forecast curve, bias-corrected by the live observed estimate (correction decays over ~18 h).
5. Stratus ceilings (constraint only): a broken/overcast coastal deck pins the inversion above the cloud base; used to trim the low side of the estimate.

Known limitations:

• No station rungs exist between 888′ and 1,760′ on the immediate Peninsula coast, so brackets there are wide; the fusion leans on models/sounding inside that gap.
• Model levels are ~700′ apart below 3,000′, so raw model depths are quantized; the boundary-layer-height refinement only partly fixes this.
• East Bay stations can sit in a different air mass than the coast; they get reduced weight and are dropped first when the ladder is self-contradictory.
• Nocturnal cold pools can mimic marine air thermally; the dewpoint-match test usually, but not always, screens them out.
• Mt Tam’s sensor reports dewpoint only, so it is classified on moisture alone at reduced weight.