Well today our topic is dilution as an attenuation process sometimes.

So Dave this can be pretty controversial idea but

can you use dilution as a natural attenuation process.

>> Well I think the answer is it depends.

To backup a second, I think we need to start this with

a view of groundwater flow, so we'll start with the basics, then show some

correlations between groundwater flow and plume links at the end.

>> Okay, well, let's start with our domain, this is what it looks like, we're

primarily talking about unconsolidated settlements like we see here.

Our highly fractured rock that can be analyzed using Darcy's law.

But to understand the groundwater flow through this material we'll need several

things, so what's happening here?

>> Well the first is understanding of what hydraulic head is.

This is basically the distance that water and

the monitoring well will rise above an arbitrary data.

Often times we use sea level as the statum.

Okay, so Well 1 is higher than Well 2 in this picture, right?

>> Exactly.

>> What about confined aquifers?

How does that work?

>> Well it's pretty much the same way, except that you

have to make sure that you use the water level in the monitoring well,

not the actual location at the top of the aquifer.

>> Got it, okay, so then as you see here, by taking the difference between the two

points, both the difference in the water elevation divided by the distance between

the two points, you get this hydraulic gradient.

Which we also call the i, which reflects the amount of energy that's able

to force water through the open pores of the aquifer.

And you can do a few more things if you know this hydraulic head, right Dave?

>> Well yeah, if you have three of these points you can get the groundwater flow

direction.

Most of the time you'll construct a potentially metric surface map,

a map of these hydraulic heads.

And get the gradient off that map and

most importantly you use a hydraulic gradient to get a ground water velocity.

>> Okay but to do that you'll need this hydraulic conductivity or what we call K.

So here's a brief summary of this, that you can see on the left

you have smaller grain sizes, you have smaller pores,

you have more frictional resistance that occurs in a lower hydraulic conductivity.

And, on the righthand side, the larger pores is the opposite,

less friction resistance, higher hydraulic conductivity, and a higher, sort of,

flow through that aquifer, mm-hm.

>> Yeah, and this is the kind of information you can get from slug tests,

or pump tests, or even looking up charts that match the geologic material to

the k value >> It gets a little bit more complicated

if you have a lot of different sizes and materials that can reduce your K

compared to aquifers with more uniform seeing greens.

>> Got it, okay so

then you can sort of put together a year of your hydrolic connectivity,

hydrolic radiant with the help of, what I call the big guy, Henry Darcy.

Sort of the original founder maybe of our discipline, hydrogeology and so

here is just a picture of Darcy himself.

He is in De Joan France.

Okay if I assign some sort of non standard homework to the group out here?

>> Well, I guess so.

>> I've done this a number of times but to everybody out there.

Your homework is to go to Dijon France and take a picture of you in front of

the statue of Henry Darcy and then email it to me, I'd like to see.

So this is all,

sort of home mentioned this guy sort of figured all this stuff out.

But what he did, he did this lab setup that you're going to talk about, right?

>> Okay, what this is is the lab setup piece.

Basically we've got two reservoirs of water here on the left and on the right.

These are filled with water,

and in between is a sand in a tube with a specified cross sectional area of A.

And you're basically measuring then the flow that's coming out

into the second reservoir based on that change of heads across that distance,

that delta L that's given there.

Okay and so if you look at the graph on the left,

he's got this flow over an area that too the cross sectional area.

On the x axis is this hydraulic gradient and the key thing in this experiment,

he got this straight line between those two measurements on the right.

And so, at the end of the day you end up with the Darcy formula down below.

You want to go through that?

>> Yeah, it's just basically saying that flow is

proportional to that hydraulic gradient, based on this K value.

That K is the definition of that proportionality, right?

>> Right, I don't see porosity in there, so what's with that?

>> Well, we'll come up to that in a little bit.

But that is a very important point and so we'll save a whole lecture for basically

talking about the distinction between darcy velocity and seepage velocity.

For now, this shows how to calculate darcy velocity, and that doesn't mold porosity.

It's more or less built into that K term.

For seepage velocity, you do divide the Darcy velocity by the effective velocity.

More on that in the next lecture.

>> Okay, so that's a short primer on groundwater flow and velocity, but do you

think groundwater velocity is correlated to something like MNA processes?

Do you think that higher groundwater flow will give you longer contaminate plumes?

>> Well let's find out, let's go to one of the most important early studies of MNA.

This is the Lawrence Livermore study.

They're, in this case, looking at several hundred b text plumes

from gas stations in California in this case.

So looking at this graph that's in the lower right hand corner we have

the groundwater gradient on the X axis and on the Y axis we have the plume lengths.

They're going from shorter plumes to bigger plumes there and

everything's given in units of feet in this case.

So each one of those individual points is an individual site where B text was

released right?

>> What do you think the R squared on that graph is?

>> Well I'm going to guess it's not too good.

So their conclusions were that these individuals were combinations of variable

such as groundwater depth of range, they didn't have relationship to plume-lengths.

It really couldn't be predicted by any consideration of hydrogeologic

information.

So there may not be any strong controlling variables that are not measured.

So there maybe something else going on, they talk about

bioremediation as being one of these potentials that's going in there.

So idea is that maybe spider gradation is so

strong it's wiping out that velocity signal.

So let's go to another study done by the same group for

a different set of contaminants.

What's this one?

>> Now we're looking at chlorinated solvents.

It does look like we have a correlation here.

We got our same variables here but in this case,

the plume length is on the x-axis and log scale in this case.

And the groundwater velocity is on the y axis.

In this case you're able to draw a line

with the best fit regression lines through these things is significant and

with a pretty reasonable R squared, so the conclusions in this case, Chuck?

>> That they said that these plume lines are possibly chlorinated with this maximum

chlorinated concentrations of the source.

And this mean groundwater velocity, I think maybe the idea's are less

biodegradation in each corner and soft in sides, and so that's why you do see

the same pact of, sort of groundwater velocity on the lengths of these things.

>> So more biodegradation relative to the tech site's in this case.

>> That's right.

>> And you also did your own sort of plume [INAUDIBLE] study I think at Air Force

sites, right?

>> Yeah, we had about 23 of them, let's look at this date to hear this is

a biochlorinateds solvent database >> Here are the different

lengths of the different plumes.

We've got plume length on the y axis, the log of seepage velocity in the x axis for

the different compounds, PCE, TCE and

some of the R squareds are better than the others.

But generally we did see this correlation and

we did say that they were moderately correlated with seepage velocity.

So the groundwater flow did also show up to be important

Variable here in this particular study.

>> Okay, so a good confirmatory study in this case, so is that it for this lecture?

>> Well no I maybe just got one more and you know one thing that I'm real

interested in, what we do in our shop GSI is really think about graphics and

we've been real acolytes of this particular book called The Visual Display,

Quantitive Information.

We sort of teach people about it but they really talk about data density in here,

this is by Edward Tufty from Yale.

He talks about you want graphics where you have a lot of information at once so

people can see these correlations in here.

So what we did is actually try to apply all of our data from this study and

to come up with we'll call the mega graphic and it looks something like this.

What do you think?