Potential Corrosivity of Untreated Groundwater in the United States
- Corrosive groundwater, if untreated, can dissolve lead and other metals from pipes.
- National maps have been prepared to identify the occurrence of potentially corrosive groundwater in the U.S.
- These findings have the greatest implication for the 44 million people dependent on domestic wells for drinking water.
Location Taken: USGS Menlo Park, CA, US
[ Silence ]
Okay, we’ll go ahead
and get started here.
Again, good evening. Welcome to the
USGS to the monthly public lecture.
I’m William Seelig. I’m with
Science Information Services.
I’m glad to see everyone
here for tonight’s lecture.
Just a quick reminder
about next month.
Our May lecture is Underwater
Secrets of the Hayward Fault Zone,
Integrated 3D Imaging to
Understand Earthquake Hazards.
You can pick up a flier
on the back table there.
Tonight’s lecture is – the title is
Potential Corrosivity of
Untreated Groundwater in the United
States, presented by Ken Belitz.
And a little bit about Ken.
He is currently the chief for
groundwater studies for
the USGS NAWQA project,
which is a component of the
national water quality program.
Ken received his Ph.D. in hydrogeology
from Stanford University in 1985.
Then, after graduation, he worked for
USGS California District for five years.
Was also a professor for seven years
and then returned to the USGS in 1998.
Ken has developed groundwater models.
He’s led the Santa Ana
NAWQA study unit and
the California Groundwater
Ambient Monitoring Assessment,
which is the GAMA program.
And he was also part of the team
that developed the science and
implementation plans for the third
decade of NAWQA’s activities.
Ken and his co-authors have
published more than 100 science
publications and reports.
And, again, he’s happy to be back here
at Menlo Park for this lecture,
and we’re happy to have him here.
So please give a
warm welcome to Ken.
[ Applause ]
[ Silence ]
- Thank you.
As I said, I’m happy to be here
because it’s 45 degrees and raining
in Massachusetts presently.
As Will noted, the title of
my talk is Potential Corrosivity
of Untreated Groundwater
in the United States.
For those of you who are not familiar
with the National Water Quality
Assessment Project, it’s a project
that was initiated in 1991.
The United States Congress asked
the U.S. Geological Survey to assess
the quality of the nation’s water.
Is it getting better or worse?
And what are the factors
affecting water quality?
The work that I’m presenting today
is part of that larger project.
The motivation for our work
was the occurrence of lead in
drinking water in Flint, Michigan,
which raised questions about
the quality of drinking water
elsewhere in the nation.
This headline is a little bit
more than a year old.
USA Today noted that
high lead levels were found
in more than 2,000 water
systems across the U.S.
And that headline was based upon some
EPA data that is publicly available.
Well, as it turns out, media attention
grew because there are comparable
problems in many places,
including large cities such as
Washington, D.C., Philadelphia,
and other water systems.
Given the occurrence of
these types of problems,
congressional representatives – media –
approached the USGS asking questions
like, well, where else in
the nation might this occur?
As the person responsible for
groundwater quality investigations,
we turned to look at groundwater.
And to start, it is important to note
that groundwater provides about 45%
of the nation’s drinking supply.
For those of you who live here on
the peninsula, most of your water
is coming from surface water sources –
not so much groundwater.
A particular area of concern
are private wells because
I’m getting error
Private wells are unregulated
and largely unmonitored.
The number of people on public – on
private wells is about 44 million people,
so the potential exposure
is possibly widespread.
Well, given a problem of this magnitude,
given the nature of groundwater,
what can the USGS do to
inform the public on this issue?
The U.S. Geological Survey
is an information agency.
It’s not a regulatory agency.
We don’t deliver water.
We don’t set the rules.
Well, the USGS, however,
does have a large data set.
We have the ability to model.
And we also have research
capabilities which allows us to
characterize source water quality.
Source water is water
before it gets delivered.
After water is extracted
from the ground, it can be treated,
it can be blended,
or otherwise changed.
Similarly, surface water
can undergo treatments.
Source waters can include groundwater.
They can include surface water.
When we look at groundwater,
the issue largely is how does
water quality vary over space.
When it comes to surface water,
there’s a question of how it varies
over space but also over time.
So, again, we’re going to
focus on groundwater.
The data sets that we’ll draw upon are in
our National Water Information System,
and the two largest sources of data into
that system are our National Water
Quality Assessment, or NAWQA,
program, initiated in 1991,
and it’s still ongoing
into the present.
We also have a
Cooperative Water Program
where our Water Science Centers
work with local or state agencies to
engage in local or state-scale projects.
When we collect those types of data,
they’re brought into
our online data system.
Well, so that’s a little bit of background.
So how can source water quality
affect the occurrence of lead
in household water supplies?
So, again, source water gets changed
before it may come out the tap.
There have been a large
number of previous studies
which clearly show that source
water can interact with the pipes.
So lead generally does not come out of
the ground with the groundwater.
If lead is present,
it’s because the groundwater
has interacted with
the pipes in the system.
For example, there was a
Virginia study by Pieper and others.
They sampled more than
2,100 private systems,
and they found that nearly 20%
of what they call first-flush water
exceeded 15 micrograms per liter,
which is the EPA action level.
By “first-flush,” they mean, you turn
on that tap, and the water comes out.
If that first flush is in the morning
when the water has been sitting in the
pipes overnight, the concentrations can
be a lot higher than later in the day.
And if you let that water run, the
concentration is going to also drop.
When a source water moves
through pipes, if that source water
contains calcium carbonate,
it can deposit that calcium carbonate
on the inside
of the pipes.
So many of you have lived in places
in the country with hard water.
So you get build-up on your fixtures.
You get build-up on the
inside of your dishwasher.
Your pipes can get filled up.
Generally, that’s considered
a nuisance or a problem.
But it does have the
benefit of lining the pipes.
If the pipes are lined, then that water is
not going to interact with the pipe.
So it’s a protective layer.
In addition, the same minerals
that are deposited on the
inside of the pipes
can also sequester lead.
So lead carbonate can be sequestered
along with calcium carbonate.
So in addition to armoring the pipes,
a water that tends to precipitate can also
precipitate out lead minerals, and hence,
the water out the tap might not have it.
With that said, if the water
changes in that system,
you can also flush some of those
previously deposited minerals out.
If the pipes are not lined,
and the water is in contact
with those pipes, you can
simply get dissolution.
We’ve all seen it. You put
salt in water. It disappears.
If you put a piece of, you know, metal in
that water, you may not think that the
lead or other metals have gotten into the
water, but small amounts have dissolved.
Another way that lead can get into the
water is if lead pipe is in contact with
copper, or if copper pipe
is in contact with lead solder.
When you have two different metals,
you have a difference in electrical
potential, and you set up an electron
flow, and you also set up an ion flow.
When you have copper and lead,
the lead sends electrons to the copper,
and it also dissolves
off lead ions.
So it’s just like your car battery.
So that’s galvanic corrosion.
So if lead is present in the system,
you can get galvanic corrosion.
These three plots are taken
from a paper by Hu and others.
There’s a group that operates
out of Virginia that have done
a lot of the work on lead
in drinking water systems.
And what this plots shows
on the X axis is time.
On the Y axis,
They took three
different source waters.
They labeled then here
Water A, B, and C.
And they simply allowed that
water to be in contact with lead pipe
over a period of several weeks.
Each week, they extracted a sample.
They made a measurement
of lead concentrations.
And those red lines that you see all
have concentrations on the order of
300 micrograms per liter.
The EPA action level is 15.
So if you have an aggressive water,
simply allowing that water to be
in contact with the pipe
will put lead into the system.
They then connected those lead pipes
to copper and then did the same thing.
And what they found is
that the concentration
increased by a
factor of 3 to 15.
And that increase in
lead concentrations when
copper is in contact with
lead is galvanic corrosion.
If you have a home that
was built in the 1920s or previous,
you might have
some lead pipes.
If you bought a home that was built as
recently as a few years ago, there’s a
good chance there’s copper pipes.
And copper pipes often have lead solder.
It wasn’t until the 1990s when solder
was really eliminated from use.
So that’s just a –
kind of a background.
We can get lead in water, not because
it was in the original source water,
because the source water
interacts with the piping system.
The quality of the water that
extracts those metals is corrosion.
So it doesn’t have to be like an acid,
you know, where you’re pouring acid
on something and you
just watch it dissolve.
That term “corrosion” is just used
generally to mean dissolution.
In order to address the question,
how can source water affect
the occurrence of lead in
household water supply,
we utilized our data from our
National Water Information Systems.
One of the advantages of being a long-
lasting agency that keeps records –
these were data that we collected.
And we limited our – what’s that?
- I was just looking at the date.
- Gotcha. [laughs]
We have data available from
27,000 locations, and we use
two different indices of corrosion.
One of these is called the
Langelier Saturation Index, or LSI.
The Langelier Saturation Index is
essentially a calcite saturation index.
It’s a measure of the likelihood
of calcium carbonate, or scale,
to form on the
inside of a pipe.
Another measure is the Potential to
Promote Galvanic Corrosion, or PPGC.
As a government agency,
we have to use acronyms. [laughter]
And the Potential to Promote
Galvanic Corrosion is a function of
the chloride-to-sulfate mass ratio.
That’s because chloride is an electrolyte,
and that helps carry that charge.
The sulfate tends to suppress that.
In addition to the chloride-to-sulfate
mass ratio, the alkalinity of the water
tends to suppress
And alkalinity is another way
of looking at calcium carbonate.
It’s important to note that the corrosivity
of the source water is one of many
factors that can affect the occurrence
of lead in household water supply.
If you’ve got PVC pipes, and your
water is corrosive, you won’t have lead.
If the water is moving quickly through
the system, you might not have lead.
And similarly, the precipitation
of minerals on the inside of the pipe
is not just a question
of calcium carbonate.
There are other solids
that can precipitate out.
So there’s more complexities
than this might imply.
It’s important to note
that the USGS typically samples
at the wellhead,
not at the tap.
We sample public supply wells.
So that picture on the left is one of our
employees at a large public supply well.
We sample domestic wells,
and we sample observation wells.
Public supply wells in most parts of
the country are relatively deep.
Here in California, a public supply well
will generally draw water from
depths of 200 to 500 feet
from beneath the surface.
There aren’t a lot of domestic wells
in this part of the world because
it’s so urbanized. In other places,
domestic wells are present.
And typically, a domestic well
will be drilled to a depth of about 100,
perhaps 200, feet. Sometimes they’re
deeper, but that’s a typical depth.
Observation wells are drilled so
that we can make observations.
Typically, they’re drilled shallow
because there aren’t supply wells,
or they’re drilled really deep
because there aren’t supply wells.
After water comes out of a
public supply well, or if it comes out
of a domestic well,
it can be treated.
The picture on the left
is a large treatment facility.
The picture in the middle
is a household treatment system
of one of our
And after that water comes out of the
ground, and before it comes out the tap,
it can be treated.
So the source water
is not equal to the tap water.
So, again, USGS evaluates the
quality of untreated groundwater.
Let’s look at our first image that
is a map of what we’ve done.
And the question with regard
to the Langelier Saturation Index,
or LSI, is, to what extent
might pipes be protected?
We had sufficient data in our
data set to look at 22,000 locations.
The orange circles are groundwater
samples which are potentially corrosive.
They have a Langelier Saturation Index
that’s more negative than negative 0.5.
The blue X’s are groundwater samples
which are potentially scaling and have a
Langelier Saturation Index, or LSI,
greater than positive 0.5.
Because of the complexities in
actual mineral formation,
values of the Langelier
Saturation Index between
negative 0.5 and positive 0.5
we’ll call indeterminate.
You really can’t say that minerals are
going to scale or not in that range.
On a national basis, 32% of the sites
are corrosive, 63% are indeterminate –
we don’t know,
and 5% are scale-forming.
So on a national basis, we generally
do not have scale-forming water.
If you look at all those orange
locations – if you look at New England,
you look at the mid-Atlantic, you look at
the southeast, you see a lot of orange.
If you look at California, that orange
field that you see there is the eastern
San Joaquin Valley, the eastern
Sacramento Valley, the Sierras.
You look at the Pacific Northwest,
there’s a couple of valleys
where we have
those orange values.
Now, it’s important to note that this map
is very much controlled – or the look of
this map is very much controlled
by which dots you put on top.
If you put the blue dots on top,
or if you put the gray dots on top,
it changes the
way it looks.
So we like to quantify this so it’s not just
dependent upon how we plot the image.
So what we’re going to do
is take all those data
for each state
and summarize it.
And the first thing we’ll do to
summarize it is simply calculate,
for each state, what is the average
value of the Langelier Saturation Index.
If the average value is more negative
than 0.5, we’ll color that state salmon.
If the average value is greater than 0.5,
we’ll color that state blue.
There are no blue states.
If the average Langelier Saturation
Index is between negative 0.5
and positive 0.5, we’ll color it gray.
So we’ve characterized each state
on the basis of its average value.
What we’ve also done is shown in
pie charts what proportion of the sites
or samples in a state are corrosive,
indeterminate, and potentially scaling.
And if you look at all those
you see a lot of orange
on all those pie charts.
So the pie charts are a visual way of
summarizing the information as well.
The Pacific Northwest, we saw two
states that are now salmon-colored.
You’ll notice California comes
out in that indeterminate category.
So even though there are some locations
where the water is potentially corrosive,
on a statewide basis, for the sample set
that we have, the state is indeterminate.
You’ll notice Hawaii,
which is basalts – volcanic rocks –
is also potentially corrosive.
Let’s take a closer look at California
since this is a California audience.
We’ve taken the state,
and we’ve divided it up into
what I’ll call
We’ll start in the northwest,
and we’ll rotate clockwise.
The Klamath Mountains –
not a lot of folks live up there.
The Modoc Plateau and Cascades –
again, relatively unpopulated.
The Sierra Nevada –
a large number of vacation homes,
a large amount
of tourist industry.
The Basin and Range –
relatively lightly populated.
The Desert, which tends to
be populated up against
the mountain fronts that
separate it from the L.A. Basin.
Winding back up, the San Diego
mountains and other areas in that
southwestern corner of the state.
The Transverse and Selected Peninsular
Ranges includes the L.A. Basin,
Orange County, the Inland Empire –
that’s the set of mountain ranges
that tend to run west to east,
whereas most of the mountains in
the state tend to run north to south.
The Southern Coast Ranges, the Central
Valley, the Northern Coast Ranges.
The map in the middle
shows our point values.
The orange circles are
values more negative than 0.5.
Langelier Saturation Index
suggests potentially corrosive.
The gray are indeterminate.
And then you can see some blue values.
There are not very many there –
mostly in the L.A. Basin.
Well, again, it’s difficult to
sort of synthesize that data up.
So what we’re going to do is,
rather than summarize our data
on a statewide scale,
we’ll summarize it on a province scale.
So what we see is the Northern
Coast Ranges, the Modoc Plateau,
the Klamath Mountains, the Sierras,
and the Basin Range, on the average,
have Langelier Saturation Indices
indicative of potentially corrosive water.
Southern Coast Ranges, Central Valley,
southern California tend not to.
So if you look at the Central
Valley, it’s colored gray.
There’s a slice
of orange in there.
But when you look at those values,
you’ll see the eastern San Joaquin,
the eastern Sacramento Valley,
if we were to subdivide that
would likely show up as corrosive.
The west side,
not so much.
So that was the Langelier Saturation
Index, a measure of, will the pipes
tend to have a protective layer of scale,
or otherwise known as a nuisance layer,
since people generally
don’t like that.
We’re now going to look at the Potential
to Promote Galvanic Corrosion – PPGC.
Again, if you’ve got lead and copper in
a system – copper pipes, lead solder –
you have the potential
for corrosion of the lead.
We’re going to evaluate our
samples with respect to the
ratio and the alkalinity.
We’re going to code a sample as
either high, moderate, or low.
The high values are shown in brown.
The moderate values are
shown in blue-green circles.
The low values are shown in blue X’s.
The blue X’s here are plotted on top.
First of all, you should see,
there’s a large field of blue-green
circles across the entire domain.
If we look at the 27,000 locations
where we have data, 67% are moderate.
And I really want to
emphasize something here.
In the previous plot – the LSI –
the Langelier Saturation Index, that
middle category was indeterminate.
We couldn’t say one way or the other.
But moderate doesn’t mean good.
Moderate means that there is
a potential for galvanic corrosion.
It’s just not high.
8% of the sites are high.
26% are low.
Look at New England.
Look at the mid-Atlantic.
Look at the southeast,
The same places where you
saw the Langelier Saturation Index
having potentially corrosive values,
we see the Potential to Promote
Galvanic Corrosion likewise
showing that as being relatively high.
So, again, we’re going to
summarize that data.
It’s hard to take
all those point scales.
Now, we’re dealing with
categories rather than numbers.
We can take the Langelier
Saturation Index, just take an
average value, but now we’ve
got low, moderate, and high.
So we’re going to code
our states as follows.
If more than half the wells in a state are
low, we’ll categorize that state as low.
There were only six states – the northern
Midwest – well, the northern Great
Plains and the Rocky Mountain states.
So you see those six gray-colored states.
If a state is not low,
then it’s either moderate or high.
If more than 25% of
the sites in a state are high,
we’ll characterize that
state as a high state.
So if you look at New England,
you look at the Pacific – if you look at
the southeast, you’ll see that
those states get categorized as high.
Most of the country is
categorized as moderate.
So, again, that’s not
necessarily a good thing.
When we’ve shown these maps,
people tend to see that orange color –
god forbid, we should have made it red,
and they really would have seen it.
So we see a similar pattern northeast,
Again, let’s take a
look at California.
Again, we have our provinces on the west.
If you look at that middle image,
you just see nearly every value in
the state categorized as moderate.
Every single one of our provinces, at a
province level, categorized as moderate.
If you look at the Sierras,
you’ll see there’s a brown slice in there.
So it doesn’t mean that all parts
of the Sierras are moderate.
There is groundwater in the Sierras that,
if you put that water into a system that
had copper pipe, lead solder,
you could leach lead from that system.
So we have two different indices –
one measuring the potential for scaling,
the other for the potential to
promote galvanic corrosion.
We’re going to combine
these two to get just an overall picture
of what the
nation looks like.
We’ve got three
categories for the LSI –
Langelier Saturation Index – potentially
corrosive, indeterminate, or scaling.
For Potential to Promote
Galvanic Corrosion – PPGC –
we have high,
moderate, and low.
Let’s start with this
indeterminate category on the LSI.
It’s indeterminate. We really can’t say
one way or the other what’s happening.
Well, there are no states
that are high on PPGC
There are no states that are low and –
for PPGC and indeterminate.
So, since we can’t make a
statement based on LSI,
then we’ll inherit that classification.
So if a state is moderate on PPGC,
we’ll call it moderate if
it’s indeterminate on LSI.
It turns out, since we had
no states that are scaling,
so those are all not applicable.
There are no combinations there.
We go up one level from
indeterminate to potentially
corrosive on the LSI index –
we’re still in the moderate class.
that as high.
Moving over to the left,
a state is potentially corrosive,
a state is high PPGC,
we’ll call that very high.
So it’s just kind of qualitative terms.
And the color scheme we’ll use
is kind of that darker orange for very
high, that medium orange for high, and a
green for moderate, and then a low – it’s
indeterminate and it’s low on the PPGC.
So this is the sum of all that,
so 27,000 data points.
Make those calculations for Langelier
Saturation Index, Potential to Promote
Galvanic Corrosion, get those state
summaries, combine the summaries.
And then what we see – if we look again
at New England, we see states coded
as very high prevalence for
the potential to have corrosion.
Southeast – if we look at prevalent,
then that area grows, and it covers
the entire Atlantic seaboard,
the Pacific Northwest, and Hawaii.
Six states get categorized out as low.
They were indeterminate on the LSI.
They were low.
So therefore they stay low.
Half of the country is
in that moderate category.
Well, what does this mean?
Well, let’s go to some numbers.
If you look at those states with the
darker orange, 8 million people have
domestic wells that they depend
upon for drinking water supply.
If we look at the states
colored in that lighter orange,
that’s 16 million people
dependent on domestic wells.
24 million people live in states where
the groundwater, at a state scale,
is either high or very high
prevalence for corrosion.
If we look at those green states,
the moderate ones,
it’s an additional
18 million people.
If you look at those six states
that have a low potential,
there are 1 million people on
domestic wells in those states.
So of the 44 million people –
the numbers don’t add up
because there’s round-off error –
of the 44 million people nationwide
that depend on domestic wells for their
source of water, only 1 million are
not living in states where we have
some kind of prevalence for corrosion.
The results that I’ve just shared
with you have been published
in a set of USGS reports.
There’s a Scientific Investigations report.
some data sets.
Those data sets have all been released
and made available to the public.
We also have a website
where I’ve given you this link.
You can go to that site, download
the report, download the data.
These data have been downloaded.
For example, Bloomberg downloaded
these data, along with other
data sets, and published it.
This work has been picked up
by the Boston Globe, CNBC –
it’s gotten a fair
amount of coverage.
Equally important, we have frequently
asked questions on this website.
My colleague, Mike Woodside,
has put this site together.
He’s done a really nice job.
And pretty much, if you have a question,
you might find an answer to it there.
And if you don’t have an answer
to a question you’re looking for,
there’s a long list of agencies –
state, local, federal, and industry trade
groups, all of whom are providing links.
So I encourage you
to take a look there.
That’s work that we’ve done.
The next thing I’ll do is sort of
move forward to show you
where we’re heading.
The first thing I’ll do this evening
is share with you a case study that
comes from the Virginia
Household Water Quality Program.
Erin Ling, who works with that program,
has kindly shared data with us.
She and her colleagues
have sampled households –
6,000 households in
81 counties in Virginia.
And when they sample, they sample
what they call first-flush water.
They go in the morning when that
water has not been running,
obtain a sample.
They then let that water run for a while,
and then they sample it again.
So let’s do the following.
We’re going to take that data
that they kindly shared with us,
and on a county basis, we’ll ask the
question, what percentage of the samples
have high concentrations of lead.
And by high we mean
more than 15 micrograms per liter –
the EPA action level.
The X axis divides the data into six
categories based on those percentages.
They are at 10% increments.
The Y axis is the number of
counties that fall into the category.
So let’s start with zero.
If a county falls into that zero category,
it means that there are no samples in that
county with a high concentration of lead.
There were two of 81 counties
that fall into that.
The most common category
is the 10 to 20% range.
Go to a county, collect the samples,
calculate the percentage of samples
with high concentrations of lead.
If it’s between 10 and 20%
of the samples, and this county
falls into that category,
37 of 81 categories –
of counties fall into that category.
There’s a county out there that has 44%
of the samples of high concentrations
of lead in first-draw water.
So that’s a fairly high number.
Well, that’s sort of the
bad news side of this.
There’s some good news in this,
and the some good news in this is,
if you let the water
run five, 10 minutes.
What I’ve done is taken
the plot from the previous slide.
I’ve pushed it up to the top here.
There’s only two counties where
there were no samples
with high concentrations.
There were 30-plus counties
in the 10 to 20% range.
If you let the water run,
51 counties have water samples
without detectable lead using
the methods that they have.
On the previous axis, the categories
were every 10%. Now it’s every 2%.
So 15 counties are in the zero to 2% range,
11 in the 2 to 4 range,
one county in that 8 to 10% range.
So in a way, it’s good news.
You’ve lowered the lead
concentrations by letting the water run.
But probably, if you’re
one of those households,
that part of that 8 to 10%,
and you’re still having water
in your glass with lead concentrations
above 15 micrograms per liter.
So it’s sort of a good news and
sort of a bad news type of story.
Now what we’re going to do is take
this data set and ask the question,
how do the indices that the USGS
is using to map have relevance?
Now, in order to do that, we’re going
to look at the state of Virginia,
which is shown in that blue
highlight on top of the other states.
And those color codes are
So just as we can identify
hydrogeologic provinces in California,
we can do the
Shown in that blue color –
blue-green color is the Valley & Ridge.
Shown in that darker gold
is the Piedmont Crystalline.
These are hard rocks.
Water largely comes out of fractures.
And then, in the lighter gold color
is the North Atlantic Coastal Plain.
Recall that the Langelier
Saturation Index – the LSI –
is a measure of the likelihood
of depositing calcite scale.
So a sample can be categorized
as potentially corrosive,
The Potential to Promote
Galvanic Corrosion – PPGC –
is an indicator of whether or not
when lead and copper are in contact
we’ll get dissolution
of that lead.
A value – a sample can – the value
of a sample can be characterized
as high, moderate, or low.
And, again, I’ll point out
that that moderate value, green,
is not the same as indeterminate.
So let’s start with the
I’ve drawn a nice orange box
so to match those colors.
That pie chart under the LSI has a
number posted on it – negative 2.3.
That is the average value for the
Langelier Saturation Index for all wells
in the state of Virginia
in the Piedmont Crystalline.
And there’s more than
1,000 samples here,
to give you some
sense of the data set.
The pie chart shows what proportion
of those samples are more negative
than 0.5. The gray shows what
proportion are indeterminate.
And I don’t know that there’s
even a blue slice in there for scaling.
The PPGC is categorical.
The blue slice is the low values.
A state or province can be
categorized as low if
more than half the values
in that state or province are low.
Well, that’s not more than half.
That’s a little blue slice.
That orange value is, you know, pretty
large – not quite – probably 30, 40%.
Moderate is more than half,
but we can see that the high slice
is more than 25%, so this province,
Piedmont Crystalline, is high for LSI,
it’s high for PPGC, so when we
combine it, it’s a very high province.
Let’s take a look now at the
North Atlantic Coastal Plain.
The average value for
all the wells in Virginia,
in the Piedmont Crystalline,
is negative 1.0.
That’s less than negative 0.5,
so therefore, it’s potentially corrosive.
The pie slice is –
for high values is not 25%.
The pie slice for low
is not more than 50%.
So it’s moderate with respect to PPGC.
We bring it together, it’s high.
Finally, let’s look at
the Valley & Ridge.
The average value in the Valley & Ridge
is negative 0.55 if you round to the
hundredth place, and it’s negative 0.5
if you round to the tenths place.
So it’s pretty close to the border.
But strictly speaking, more than half –
the average value is more negative
than 0.5. It’s potentially corrosive.
Not more than half the values are low,
so it’s not low, less than 25% high,
so it’s also moderate. So it has
a high combined index, but let’s
just say it’s borderline. We’re
awfully close to negative 0.5
So we’re going to
color code that with
our blue-green color rather
than the orange color.
Let’s take a look at these
average values for a second.
Negative 2.3, negative 1, negative 0.5 –
a clear gradient in the average value.
Look at the size of
that orange slice.
There’s a clear gradient
in how big that slice is.
The Piedmont Crystalline is the
most corrosive, the Valley & Ridge
least corrosive, North Atlantic
Coastal Plain intermediate in Virginia.
Now, this is a complicated slide,
so please bear with me
as I walk you through it.
We see in that blue-green
color the Valley & Ridge.
We see, in the gold color,
the Piedmont Crystalline.
We see in the light – the light gold
the North Atlantic Coastal Plain.
We have data for 81 counties,
but two of those counties have
fewer than 10 data points. So we’re
just going to kind of set those aside.
We’re going to look at what
proportion of the samples
in a county have concentrations
above 15 micrograms per liter.
Dark green is zero to 10%.
Red is more than 40%.
Where you don’t see any of these
five colors is where we don’t have
sufficient data. So you can see
that orange color coming through.
That’s a county where
we don’t have sufficient data.
So let’s start by looking at
the Valley & Ridge.
You can see a little bit of the green
coming through where there’s no data.
You can see a little bit of the blue-green
coming through where there’s no data.
Where you see white, there are
other hydrogeologic provinces,
but they’re – the picture starts
to get really complicated
when you start looking
at really small pieces.
The counties in Virginia tend to
run with the grain of the topography.
And so you can see the
line of counties that fall in
the Valley & Ridge along that arc.
And what you see is that all of the
counties in the Valley & Ridge are in the
two categories zero to 10 and 10 to 20.
The average LSI
was negative 0.5.
The one high value – the one high-value
county sits in the Piedmont Crystalline.
That’s the county with 44%
of its samples in first-flush water
above 50 micrograms per liter.
One orange county sort of
sits on both sides of the
Piedmont Crystalline and
the North Atlantic.
The remaining orange counties
sit within the Piedmont Crystalline.
If you look at the yellow counties,
they largely sit in the
Piedmont Crystalline, but they also occur
in the North Atlantic Coastal Plain.
The Piedmont Crystalline had a value –
I think it was negative 2.3.
North Atlantic Coastal Plain was
somewhere around negative 1.
The indices that we’ve been using
to map nationally, in fact,
are correlating in this qualitative
way with our county-scale data.
Well, why is that important?
It’s important because there isn’t data
for lead in tap water typically
available in the United States.
Virginia is an unusual state that
they have such a proactive program
in reaching out to
domestic well owners.
Public supply wells are
regulated in the United States.
As much as people might not like
regulation, it means that those
purveyors need to make
measurements on their water samples.
Sometimes the system fails, as we’ve
seen, but by and large, it is protective.
So this map is showing the potential for
corrosivity is what we have at this time.
And we’ve seen, in Virginia, which,
by the way, is prevalent, not very high.
It says high, not very high.
So there are states with potentially
more corrosive water than Virginia.
So what are
the next steps?
We would like to – in fact,
we have developed alternative indices
that directly address the potential
occurrence of lead in the water.
And how do
we do that?
I put the word “titrate” in quotes.
Titration is where you
incrementally add small amounts
of something to a solution.
So we want to, quote,
titrate lead into each of
Now, those groundwater
samples only are on our computer.
Those water samples
have been long gone.
Instead, we’re running
geochemical computational codes.
David Parkhurst, one of the
authors on this upcoming paper,
was the author of an important
program called PHREEQC.
He’s been working with the
senior author, Bryant Jurgens.
And we’re mathematically
titrating lead into our 23,000 samples.
And we’re asking
the following question.
How much lead goes into solution
before it begins to precipitate?
And we’re going to call that the
Lead Solubility Potential.
So if you can put a lot of lead into
that water, and nothing precipitates out,
that water has a great
capacity to dissolve lead.
Otherwise, if it didn’t have the ability
to dissolve lead, it would precipitate.
So those are
So here’s an image of what that
looks like. It’s a preliminary image.
This work is now in review.
So the red values are sort of
the top 10% of all the 20,000-plus
samples that we have.
Those are samples where you have to
add 300 micrograms per liter or more
before you see any lead precipitate out.
So that’s 20 times the
EPA action level of 15.
The green values that you see
are samples where you would
add somewhere between
15 and 300 micrograms per liter
before you see the lead
even begin to precipitate.
The blue values are those samples
where lead begins to precipitate
before you have to add 15.
So this is a more direct indication.
Well, lo and behold,
you see New England
is no longer red,
but there’s a lot of green there.
You look at the mid-Atlantic,
you see a lot of red.
You look down the rest of the –
into the southeast, you see a lot
of green and start
picking up red again.
Pacific Northwest, we see green.
You see green in the Sierras.
You see green in
the eastern Central Valley.
So Langelier Saturation Index
we’re seeing something comparable.
Well, the next step includes
looking at where in the U.S.
are people dependent on
domestic wells for drinking supply.
Typically, these numbers
are accumulated by agencies
like the U.S. Geological
Survey at the county scale.
We’ve taken on this as a project
using census data and road networks
to look at where the people are
who depend on domestic wells.
The lead values that you see are
locations where there are
more than 100 people per square
kilometer with domestic wells.
The blue values are places where there’s
one person per kilometer squared.
The gray values, less than 1.
And then the white space,
there are no people, basically,
according to the Census Bureau.
Or they’re too far from
the road to likely be located.
Well, let’s compare for a
moment this map of where
the people are with the
Langelier Saturation Index.
If we look at that mid-Atlantic up
towards New York, you see a large
number of places where there’s a large
number of people on domestic wells.
If you look at that LSI –
Langelier Saturation Index map,
the water is
If you look in Colorado, you’ll see
there’s a bunch of blues and greens,
and that’s a place also where
there’s some corrosive water.
If you look in California,
you’ll see some orange and red.
That’s largely the
eastern San Joaquin Valley,
the eastern Sacrament Valley,
the low Sierras.
If you look at the
Langelier Saturation Index –
the LSI, you find
potentially corrosive water.
So we’re beginning to
put these data sets together.
If we’re going to make maps better
than the states, what’s the motivation
for that? You saw it in California
that the provinces vary.
You saw it in Virginia that it
matters where you are.
The U.S. Geological Survey has mapped
what they call principal aquifers.
A principal aquifer is a large area
of geological materials sharing age
and lithology, which is
important for water supply –
or drinking water.
There are about 65 principal
aquifers mapped by the USGS,
and shown in these various colors other
than the beige are 20 principal aquifers,
which account for more than 90% of
the water use in the United States.
When you look at those beige areas, it
includes about 40, 45 principal aquifers,
but most of that beige area has not
been identified as a principal aquifer.
If we’re going to provide
a context for mapping,
we need to
characterize those areas.
So we’ve also done that, and we’re
in the process of writing that up.
So we can take that framework data,
as I call it, the hydrogeologic
framework – I’ve turned it on its side.
I’ve shown depth – that third dimension.
And I’ve shown depth
because those hydrogeologic units
do not occur uniformly
as you go downward.
They can have an outcrop area, and then
they can plunge into the subsurface.
So they have a three-dimensional shape,
and we’re mapping that.
We have our population data.
We have our new indices.
There are additional factors
that are also important,
things like the age
of the housing stock.
We also had additional data available.
The states have responsibility
their water supply.
That regulation requirement
really stems from EPA.
The states regulate the water supply.
They require purveyors
to sample those wells.
They then tabulate those data.
And nearly all 50 states have
now shared their data with us.
We will be able to bring
into this analysis at least
100,000 additional data
points to do this work.
So we have additional factors,
additional data, framework, population,
new indices, and the goal is to
bring it all together to identify risk.
So let’s summarize.
Corrosive source water,
if untreated, can dissolve
lead from pipes and other
components in water
The USGS has prepared
maps of potential corrosivity
of untreated groundwater
in the U.S.
Untreated groundwater is
not the same as tap water.
The lead in people’s glasses as it
comes out the tap is not generally
because that lead was present in the
groundwater but because the – rather
because the groundwater, in contact with
their plumbing systems, extracted lead.
We used the Langelier
Saturation Index – LSI.
It’s a measure
of calcite saturation.
We have a new way to look at something
called Potential to Promote Galvanic
Corrosion – PPGC – based on chloride-
to-sulfate mass ratio and alkalinity.
We have a combined index
based on joint classification.
Potentially corrosive groundwater
occurs in all 50 states.
25 states were classified as
high or very high prevalence.
24 million people depend on
private wells in these states.
We published these data in
Scientific Investigations Reports,
data releases – we have a website.
I encourage you to visit that website.
Our work is ongoing.
Our work is ongoing with
respect to corrosivity,
but the National Water Quality
Assessment Program has a
broader mandate than corrosivity.
We’re looking at a large number
of water quality issues as well.
And with that, I’d be very
happy to take questions.
[ Applause ]
- All right.
Thank you, Ken.
Again, any questions for the audience,
there are two mics toward the back.
For the benefit of our streaming
audience, please line up behind there,
or raise your hand and I’ll –
I can bring you this hand mic as well.
- I wonder if you can tell us
something about lead testing kits.
How do you get them?
How reliable are they?
Should the public actually use them?
Do they give reliable data?
- I have not looked into
the question of those kits.
The frequently asked questions
website points towards a number of
trade groups that probably,
if you follow those links,
you would be able to
find something reliable.
[ Silence ]
- From a public health standpoint,
are there public health departments
that are reporting birth defects or
learning disabilities associated
with the areas that you
have identified high lead
or high potential for corrosivity
in the groundwater?
- Lead is a known neurotoxin which
interferes with brain function, and there
are strong correlations between lead
levels in children’s blood and their IQs.
Lead is – lead in water is
not the only vector by which
children come into
contact with lead.
- But have public health
departments in different states
looked at your data and …
- Yes. The CDC has an environmental
health unit, and these data have been
loaded into that computer system.
- I have two questions.
Could you tell us about the
current situation in Flint?
And also, for the homeowner,
is there a way in which a type of
filter could be installed in the
main line to get rid of lead before it –
the contaminated water
comes into the house?
- There are two questions.
One question has to do with Flint.
- Flint is a surface water story.
It’s a story of changing water supply.
They changed from one river source
to another, and the treatment
systems they put into place
for the new water source were
inappropriate for the water source.
And then, of course,
there’s a lot of political, slash,
associated with Flint.
It’s not a – you know, it’s not a
groundwater source type issue.
With regard to filtering,
there are ways to filter supply,
and I would encourage you
to contact professionals.
- Contact who?
- Professionals. There are
people who can treat your water.
Each case, you know,
will potentially be different.
I don’t think there’s
a generic solution.
Let me add one thing.
You can overtreat your water.
One of my colleagues –
it’s kind of typical.
He had – you know, he had hard water.
He decided to soften his water.
And he figured, if a little bit of softening
is good, a lot of softening is better.
[laughter] And he had a
new home with
beautiful copper piping
throughout the whole house.
And then he
noticed pinhole leaks.
He had so treated his water that it
aggressively attacked his copper pipes.
So treatment is something
- So in Flint today,
is the situation cured?
Or is the water
- I refer you to the newspapers.
I think the newspapers have a position
on that without me having to weigh in.
- We’ve had a refrigerator for years
which has the purification canister.
It seems to – when I read
the information, it looks like
it takes a lot of stuff out.
Does it do anything for lead?
- A lot of filters
are charcoal filters.
And charcoal …
- I have no idea. I never …
- And charcoal filters are good
at grabbing organic compounds.
Getting at smaller molecules – smaller
atoms. Smaller ions is a bigger challenge.
And so filters that filter metals
in solution are not the same filters
as you’ve kind of found
with a charcoal filter.
There’s usually more involved
treatment involved in removing ions
from water rather
- I guess what I’m asking,
is this thing doing any good?
I use it for – when I make coffee,
I take the water out …
- Much of the water in –
well, you know, it depends on
where you live in California.
If you live in southern California,
the Colorado River is an
important source of water supply.
The Colorado River runs
across agricultural fields.
And there are pesticides routinely
at very, very low concentrations –
Carbon filters are very good at removing
organic compounds from water.
You know, Hetch Hetchy is
a relatively pristine water supply.
So the need for filtering
Hetch Hetchy water is not as great
as the need for filtering
Colorado River water.
It is required by law for
a public purveyor to deliver a
water supply that meets standards.
But of course, laws can be broken, right?
It’s against the law to speed, and yet
people blow by me on the freeway.
So, you know,
no laws are perfect.
But public purveyors are required to
deliver water that meets standards.
- From what you said, then, it sounds
like the – having hard water is good.
It tends to prevent the …
- It will keep lead out of your water
if nothing mobilizes that lead scale.
So if somebody starts knocking on
the pipes, and bits of scale wind up
in your glass, and you don’t filter
those particulates, then you could
ingest lead from the particulates.
- Hi. I was wondering if the increase
in fracking across the United States
is creating havoc with your research
and data, especially in the Montana
and North Dakota regions.
And, well, I guess Oklahoma
and even here in our state,
where we are disturbing the aquifers
deeply and bringing up radioactive
particles and everything else.
I’m sure that it makes it
very complicated for your
research on everything.
- Fracking is a – kind of a catch-all term.
So your fracking is not just
the breaking up of the – of the rock.
- Well, the injection –
the injection of [inaudible].
- Right. I mean, there’s all kinds of
things associated with fracking.
- You know, you have to build a road.
You have to build paths.
You have to dispose of fluids.
You have to move the new petroleum.
So there’s a large number of
potentially altering activities.
And so you mentioned a lot of
things which have impact.
With regard to water quality,
in most places so far, the depths
at which they fracture the rock
are deeper than the depths
from which they
extract drinking supply.
And it takes time for water
to move from deep to shallow.
And the effects on – the direct effects
of sort of the breaking of the rock,
the injection, on that water
quality will take time to see.
Certainly longer than the
amount of time it takes for, say,
a spill, which can then affect the supply.
So depending on how you try to
define the term “fracking,” and what its
impacts might be, it’s a complicated –
it’s more than one question,
maybe is one way of putting it.
So there’s more than
one answer there.
The samples that we’ve collected,
you know, date back to the 1990s.
And we have some studies,
and we’re starting to publish –
we have – we will have
sampled about eight or 10 locations
in the United States which
are areas of active fracking.
And we are evaluating the
quality of water in domestic wells
as a function of how far they
are from those fracking operations.
And those results
will be forthcoming.
- Any other questions from
the audience? Oh. Over here.
- I was just curious –
noticed these southeastern states –
Alabama, Georgia, and South Carolina –
having, you know, a higher potential …
- What do you – what’s behind –
I may have missed it, but …
- Well, those few states as well as
some of those
New England states
are getting their water from the
North Atlantic Coastal Plain.
Virginia is kind of a funny example
because the North Atlantic Coastal Plain
in Virginia doesn’t have
particularly corrosive water.
But other parts of the North Atlantic
Coastal Plain, or the South Atlantic
Coastal Plain, are host to
very corrosive water.
What’s sort of buried under
New Jersey particularly has
very corrosive groundwater over –
under a very large portion of its state.
And it has to do with
the mineralogy mostly.
There’s not a lot of buffering
capacity in those sediments.
So the water stays acidic, in essence,
after it enters the ground.
- You talked about, in 1990 is when lead
soldering of copper pipes was changed.
Was that because there was
government regulation that
came into place at that time?
- Mm-hmm. Yes.
- And are the states doing – did the
states do something different?
Did they react strongly?
Do you know anything about that?
- I don’t.
- Okay. Last call.
Any other questions from the audience?
Okay, once again, I want to
thank everyone for coming.
[ Applause ]
Please thank Ken Belitz. Thank you.
Have a good night.
[ Applause ]