Permeability and the Emriver Stream Table

Got the students to do some groundwater stuff in lab today. One of the steps was to estimate the permeability of a column of sand. I used the falling-head permeameter rig that I had videotaped and demo'd earlier.

But instead of a third column of sand, I tossed some of the plastic media that comes with the Emriver setup. If you've seen these stream tables in action, you know that groundwater flow is pretty important in determining bank stability and, ultimately, what the open channel is doing.

Here's the rig:


Here's a scatter plot of all the measurements made by me and the students. For the most part, my measurements are the series that don't vary too much, while the student measurements are all over the place (well, within about 20% anyway)


This ins't a really good way to display univariate data - it's hard to determine any kind of tendency or something like that. So if I collect all the calculated permeability values and display them in a box plot, you can see that the permeability of the plastic media is about 0.3 cm/sec.


So the permeability of this material is really pretty high. You could decrease permeability a little bit more by compacting the media, but there's typically a lot of water that's flowing through the media as "groundwater." So if you're sketching up a plan for Emriver experiments, accounting for groundwater flow is probably a helpful step - especially if you're trying to produce steep gradients in a stream bed. The extra groundwater flow will tend to encourage movement of the plastic media in directions you may not expect at first.

Thurs-Demo: The one with Permeability

This week's demo is all about permeability. A permeable material is one that allows water to move through it easily. I was puttering around the lab last semester, trying to figure out how to get my students to measure the permeability of sand samples. I had some plastic tubes (designed to protect fluorescent lights) lying around - you can never have too many plastic tubes, if you're a sedimentologist.

I cut one of the narrower tubes down to about 50 cm. Then I plugged one end with some florist's foam and secured it with a few T-pins:

I added some sand and then put another hunk of foam on top to keep the sand in place and act as a baffle when adding water. I propped the tubes up with a ring stand and suspended them in a beaker - to act as a constant free surface.


Here's a diagram of the setup:


It's a "Falling-Head" Permeameter, which means that the height of the water column (hydraulic head) falls from one level to another. The permeability (k) of the material is proportional to the time (t₂-t₁) it takes for water to fall a certain distance.

Since my permeameter is designed for relatively permeable materials, the cross-sectional area of the head (a) equals the cross-sectional area of the soil sample (A) and this term simplifies to one. Students measure the change in head (h₁ and h₂), the length of the soil column (L) and the time it takes for the head to drop.

Some preliminary results showed I was within an order of magnitude (medium sand typically has a permeability of around 0.1 cm/s). For about $3 worth of random hardware, I can live with that kind of accuracy.
One interesting trend is that permeability for the samples (same type of sand for each plot) drops quickly then levels off after about 10 or 15 trials. I suspect this is because the water is consolidating the sediment a little bit, and the florist's foam tends to shed little green fuzz which probably clogs the pores in the sand (it may also increase the matric potential of the column, allowing more water to adhere to the sand).

And, it wouldn't be a true Thurs-Demo without a video:

Falling-Head Permeameter from Matt Kuchta on Vimeo.