Looking at the big bend at mile 277, where the boundary crosses from river left to right. This does not involve the Reservation. Using the Google satellite map brings up the brown "flats" as shown on map 50. The brown river is flowing close to the right bank. If we use this as a discussion platform for "to the Colorado River", then we can see that
1. historically, the boundary went to something like this river-like stream (even though the channel of 1880 might have been further to the south), and
2. the representation should surely coincide to that current (the new) channel.
In other words, to be true to the EO of the Reservation, we need to bring the boundary to the river, as defined on that Google map by the wet brown and dry brown. Even though there have been times when water covered the dry brown flat. That is, we can define a river resembling the 1880 river, and a south bank resembling the 1975 bank. I.e., to meet the EO description requires that the lake spread be ignored, wet or dry, as a phenomenon different from the daily fluctuation of the river from dam releases.
Looking at various places along the shore, there are sandbars/islands where the map shows the blue reservoir; implication being that at certain points, the boundary needs to be checked against the photo. But same thought applies, that the south bank, TO the river, is stable, and the sand bars are more similar to daily fluctuation.
The Google photos show the slumping of the silt banks, and some sand bars on the maps have eroded (mi 262). However, a comparison of the relevant 40 miles on the maps and on the photos indicates there is very little change. Drawing the boundary on the river then is not only symbolically correct, but usually physically accurate in that it matches the silt deposit.
It is still possible that the blue between reservoir max. line and the old river channel (which may have shifted due to silt deposit) should not be just blue, but blue-green-brown.