These are the USGS flow and river level (height) readings at Trenton for the Delaware River floods of 2004 and 2005.

Notes:

1. The scales and time frames for '04 and '05 are slightly different. (3 days for '04, 4 days for '05.)

The USGS website doesn't allow setting arbitrary scales (just number of days),
and I didn't anticipate that there would be another major, even bigger flood in less than 6 months.
Otherwise, would have saved a greater variety of plots (archived stream flow data doesn't seem to be available).

The last major flood here happened in 1955; river crested at 28 feet, with flow of 329,000 cubic feet per second;
see http://www.capitalcentury.com/1955.html


2. The plots for '04 are just photographs of print-outs. The gage height plot was stitched together
from two separate plots. Ugly. Sorry about that, but see (1.) for reason.

The gauges came in handy, especially upstream gauges (not shown here), in predicting the
magnitude and time of the flood crest. Same as scanning the Accuweather maps for forecasting the next
day's weather, this turned out to be much more reliable than the rumors or inaccurate
information spread by the media. (Do you believe what you hear on TV?)

I'm not a hydrologist, but I figured a few things out by examining the data.
The shapes of the flow and level curves should match, if the river channel
has a parabolic cross-section for all flow levels (the volume, to a first approximation,
would be proportional to the square of the depth; plotting the volume
with a logarithmic scale should give the same curve shape as the depth). So, can tell when
the river has spilled over its banks (or if there is a narrowing of the channel at a certain flow volume) by
comparing the shapes of the depth and volume curves. Also, by the same token, the
upstream gauge volume is a better predictor, than the upstream depth, of the downstream crest depth.

The red and green lines on the plots denote the height and flow, respectively, of the river at flood stage.
-