This is a first approximation of the now-submerged land which was exposed at the last glacial maximum. Bathymetry is based on the 2-minute global dataset. Rebound, sedimentation, and scouring are ignored in this version.

Our DEM is in a Mercator projection, so cell sizes are 2 minutes square at the equator, and 2 minutes by a smaller delta-latitude elsewhere. The scale varies with the cosine of latitude, but determining the scale as a function of the Mercator Y value is challenging. (See Norris and POSC to see why Mercator is among the hairiest map projections.) Eventually, we created the necessary background grids to calculate contributing area and slope.

Because cell edge and cell area vary from 2469 meters and 6.1 square kilometers at Cape Flattery to 1773 meters and 3.1 square kilometers at Anchorage, the DEM gives us only a rough approximation of stream locations and slopes. We are defining a stream cell as any cell with one or more cells flowing into it. Slopes are calculated from a cell center to its lowest neighbor, accounting for direction and edge length. Perhaps we can look at detailed local datasets to adjust these numbers. So every cell with any contributing area and with a calculated slope less than 4% is considered potential salmon habitat. For our first run of calculations, this includes cells which are upstream from very steep cells. Calculations are clipped to the Pacific watershed. Results are shown below.

Red (mostly covered with white) indicates areas with a depth of 125 meters or less: a first approximation of the land exposed by sea level drop at the height of the glacial advance. Most (?) of this area would have been ice free. Potential salmon-bearing cells below present sea level are shown in white. Potential salmon-bearing cells above present sea level are shown in blue. The cell counts are 38984 and 70160, respectively.