Skip to main content Skip to navigation
Environmental Hydrodynamics Three-dimensional flow in tidal channels

Skagit Bay ’08 data

Please email Steve Henderson (email address on home page) before downloading data.


We measured bathymetry during a single low tide on Thur 9 October 2008 by pushing around a golf cart with GPS mounted (see picture right) and by holding a surveying pole over the side of a boat. Data were collected in an area near Craft Island, some parts of which (in the area immediately NW of Craft Is) were not covered by the Raubeheimer et al. group’s survey. We collected GPS waypoints every 2 seconds (in an .txt file with collumns for lat, lon, and m elevation in NAVD8 and also in a matlab file).

Known problems with data: This is pretty fresh off press, we’ll probably find and post problems as we go. Please email Steve Henderson (contact info on home page) if you find anything wrong. There’s a potential constant offset of a cm or two in all cart measurements (we only measured the elevation of the sensor above ground this accurately), and in the end I just chose an offset which lined our observations up best with the Raubeheimer et al. observations. The vertical accuracy claimed by the instrument was <2cm. I pushed a cart over a section of flats, then back over a few 10’s of m of the tire tracks in the opposite direction an hour or so latter, and got an rms error between the two passes of 1.5cm. We didn’t allow for cart tilt, which could be significant on steep-sided channels. We did measure heading, pitch, and roll of the cart every second, so corrections could be made – if you’d like to make such corrections, contact Steve Henderson. The last hundred or two waypoints were measured by planting a survey pole over the side of the boat (these are the sparsely-spaced measurements at the end of the file, giving cross-sections of the channel). Sometimes the pole dug a few (<5) cm into the bed.

Currents and Pressure

2MHz Aquadoppp current meters were mounted on pipes at fixed locations (some of our instruments, which were in channels, appear to be on the flats in this image, possibly a consequence of channel movement or inaccuracies in the GoogleEarth image). The instrument locations are also plotted on the bathymetry (WSU and Raubenheimer et al. surveys), and lat-lon’s are listed in a matlab file. Some instruments were moved from one pipe to another during the deployment.

There is one matlab data file for each pipe (Files) that had an Aquadopp attached at some point during the experiment. The files all run over the same time period: yearday 256.97 to 272.80 UTC, (12th Sept to 28th Sept 2008). During times when no instrument was present on a pipe, the data are NaNs. The sampling interval was 10s. Files all contain the following variables:

vel(ii,jj,kk)= iith velocity component in ENU coorodinates (1=E/W, 2=N/S 3= UP/DOWN) at jjth time, kkth rangebin. Velocities above the surface are NaNs. Units are m/s. At recorded (10s) sampling rate, measured velocities are dominated by noise (noise is uncorrelated in time, almost uncorrelated betwenn rangebins). You’ll need to average/low pass over time/rangebins to get an accurate signal.

bs(ii,jj,kk) =backscatter counts from beam ii, at jjth time in kkth rangebin.

location(ii): ii=1: Latitude, ii=2: Longitude, ii=3: Best guess at instrument elevation in NADV88 (the errors here are likely around 5cm but could be up to 10cm). The instrument sensors were approximately 10+/-5 cm above the bed in all cases.

t(jj) = jjth time. In yearday 2008 (UTC). This is the same for all files. (Note January 1st = yearday 1).

z(kk) = Vertical distance to centre of kkth rangebin (in m).

p_atm= Atmospheric pressure from weather station (sampling interval was 60s so data was spline interpolated to the same sampling times as the Aquadopps).

p_raw= Raw pressure.

p_c= pressure corrected for atmospheric fluctuations and instrument offset (all pressures in dbar).

Note = Note of any known problems with data. So far: Pipe 7: Low backscatter and hence likely bad velocity measurements from yd261.5 to yd262.8. Possibly owing to transducers being covered by seaweed.

Questions/comments? Email to:, C.C.

Pressure (old, we recommend using currents and pressure data above)

Water pressure (measured every 10s at 8 locations by aquadopps) minus atmospheric pressure is in the matlab file p.mat. Pressure gauges were about 10cm above the bed (we will add measured elevations above bed soon). The file p.mat contains the following variables: t(ii)=time ii’th measurement was made (yearday of 2008). p(ii,jj)=pressure (in m of water) at time ii, for instrument jj. lat(jj)=latitude (degrees north) of instrument jj. lon(jj)=longitude (degrees WEST) of instrument jj. offset(jj)=my quick best guess of the offset that should be added to p(:,jj) to estimate sea level (so the offsets effectively cancel pressure differentces resulting from elevation differences between in the instruments. I have yet to zero things to some absolute geoid).

Known problems with data: This is pretty fresh off press, we’ll probably find and post problems as we go. Please email Steve Henderson (contact info on home page) if you find anything wrong. We do know that many of the low tide pressure values are junk, because the instruments came out of the water. We’ll work on flagging them latter. Comparing measurements around high tide, I suspect erorrs (when instruments are in water) are typically about 2cm.