Tip:
Highlight text to annotate it
X
[0:00] [silence] Hydro Seminar SeriesSession 4 “SPARROW”
and Lightning Talks Transcript
Al Rea: [0:40] Next we have a couple of lightning talks. These are short five‐minute talks
by speakers who have topics of interest that we've chosen. The first talk today is going
to be Dr. Bill Samuels. He was the guest speaker on our first seminar of this series. He told
us then about the ICWater Application for modeling contaminant transport.
[1:08] With the news ‐‐ the big spill in Colorado in the news the last month or
so ‐‐ we asked Bill to give us a quick update on how that application's been used
in response to that incident. Bill, could you just briefly re‐introduce yourself and
go ahead with your presentation?
William Samuels: [1:32] Thank you, Al. I am Bill Samuels. I'm with the Leidos Center for
Water Science and Engineering in Alexandria, Virginia. We have developed a Toxic Spill
modeling tool under contract with the Defense Threat Reduction Agency. It's called ICWater.
[1:51] I'm going to show how it was used to do some forecasting of the downstream transport
of the Gold King Mine spill. This spill event occurred in early August of 2015. It was a
result of an EPA investigation into the mine to relieve the mine of the water that had
been accumulating there for some period of time.
[2:21] As a result of the excavation of the mine, the entry point had collapsed and a
large pulse of contaminated metal‐rich mine wastewater spilled into the north fork of
Cement Creek which is a tributary to the Animas River.
[2:43] It was very quick release, about three million gallons of water spilled in about
one hour. The graphics on this slide show an overview map of the location of the Gold
King Mine in the vicinity of many other abandoned mines in this region.
[3:04] On the right hand side zoom in on this area showing the location of north fork of
Cement Creek and some of the downstream tributaries.
[3:16] One of the first things we did was, in ICWater what we call a quick trace which
is just a time of travel calculation we went down stream 48 hours and you can see then
we color‐coded this downstream trace by four different time intervals over that 48‐hour
period.
[3:38] The graphic in the left shows the hydrograph from a gage on Cement Creek near Silverton
Colorado. You can see the pulse of the spill clearly shows up in that particular gage on
August 5th. We did some comparisons between the modeling results and some observations
that were reported.
[4:06] This was a report that was published in "The Denver Post" on the time of travel
of the spill at a number of different locations downstream. We superimposed those locations
on the ICWater map and compared the forecast time of arrival with the observations. You
can see the graphic or the table there in the middle of this graphic.
[4:33] Fairly good agreement between observations and forecast. We are off, a little bit slow
initially in predicting the time of arrival but as it moved downstream the results seem
to converge a little bit better with the observations.
[4:57] Another analysis that we did was to forecast the time of arrival at Lake Powell.
This was an eight‐day travel time period. There were about 10 USGS real time gages reporting
flow in its path of about 400 miles downstream from the spill site to the confluence with
Lake Powell.
[5:25] As sampling was done and measurements were being made of various metals by EPA,
that data was being made available to us as well as many others. We did some initial comparisons
between manganese levels that were observed. You can see that in the upper graph there,
there are actually number of different metals that are being reported there.
[5:53] Our model predicted a peak concentration of manganese of about a little over 7 milligrams
per liter, what was observed at this location which was on the Animas River just below Silverton
was about 6.65 milligrams per liter so fairly close agreement.
[6:15] There was some disagreement in the peak time of arrival between the model and
what was observed. We're looking into improving some of the flow velocity relationships in
ICWater.
[6:30] Just a quick summary of what we've done, we used a quick trace tool to compute
initial time of travel as mentioned before it was about an eight‐day travel time to
Lake Powell. There are hundreds of abandoned mines in the vicinity of this particular mine
and some of the improvements that we'll be looking at is the flow velocity relationships
in ICWater.
[6:56] Just a quick comparison between these spills and the one that occurred in West Virginia
back in January 2014. The Elk River spill was 10,000 gallons of organic solvent. This
was three million gallons of acidic metal‐rich mine wastewater. Elk River was a tank leak,
Gold King Mine was a mine breach. The major effect on The Elk River was the contamination
of a major water supply of the city of Charleston, West Virginia.
[7:28] For Gold King it was effects on drinking water availability on several municipalities
downstream as well its effects on irrigation and the use of the river for recreation. On
tracing for Elk River, we were looking at about 250 miles downstream trace and here
we looked at about 400 miles of the river in terms of the downstream trace.
[7:55] That concludes my presentation and thank you for allowing me to show these results.
Al: [8:04] Thanks a lot, Bill. That was very interesting. Nice to get an update on ICWater
and see how that's working.
[8:16] Our next lightning presentation is by Elizabeth Zeiler. Let her introduce herself.
She's going to talk about making the case for integrating 3DEP and NHD. Elizabeth, can
you share your screen and go ahead with your presentation?
Elizabeth Zeiler: [8:40] Good morning. My name is Liz Zeiler. I'm the NHD Steward for
New Mexico. Today I'm going to talk to you about the need to integrate two USGS products
‐‐ the National Elevation Dataset and NED data.
[8:57] As mentioned by Jeff [inaudible 8:59] in his 2013 concept paper, this is somewhat
of a chicken or egg problem. Historically hydrologic data was mapped by cartographers
first. With the advent of new technologies and especially LiDAR, we find that this new
dataset can help us to inform...use the elevation dataset to, the LiDAR dataset to inform the
NHD.
[9:27] This image of our project area here and we have a stippled area in this rectangular
area. This is Santa Fe County and down here this area, the purple areas are quality level‐two
data...do not represent quality level‐two data nor are they publicly available data.
The hachured area here and here are data that will be processed in 2016.
[9:52] You can see that we're a long way from completing the LiDAR processing in New Mexico.
[10:01] The project area covers some 3,000 square miles and includes Santa Fe County
and the LiDAR was collected and based on a nominal pulse spacing every 710 centimeters.
This results in a one meter DEM. The image on the right shows the same surface reconditioned
to reflect surface water flow. This example here there are over 180,000 vectors modeled
from this associate DEM.
[10:26] Clearly, we need to determine how to efficiently get at the value of the data.
When you look at the branch complexity here on the right, stream orders range from 1 to
10. Clearly not all of these streams are modeled with precipitation or soil moisture in mind.
Three quarters and a quarter of the streams are classes 3 or greater.
[10:52] This presents a conundrum of sort. In this example of the DEM was reconditioned
and you can clearly see that the channel flows accurately down this culvert here. The NHD
data flows outside of the channel here. Here we see more discrepancies in the data. The
reconditioned surface flows properly down this channel.
[11:18] However, when you get to this point here, it appears that the reconditioned surface
may not be routed correctly and perhaps this barrier needs to be removed.
[11:30] If we are to use NHD dataset for modeling, it might result in lives and property being
damaged. This is somewhat of a conundrum. We have this very accurate elevation dataset
that we can use to improve the NHD, and here you can see that the geometry align here.
Over here, it's not so obvious where the alignment occurs. Over here we have a different issue.
[12:06] Here, we have the reconditioned surface that flows outside of this levy. However,
the NHD looks like it runs properly down the channel and actually is perpendicular to the
levy. A similar problem occurs here where the reconditioned surface goes around the
levy and we actually need to remove this barrier here so that the channel flow is proper.
[12:30] In closing, I'd like to say that there is a need to integrate the 3D elevation dataset
and the National Hydrologic Dataset. That we should make use of the more accurate elevation
data and that we've seen and expressed need from the hydrology community where this data
as reflected in the HRBS Study which is the Hydrologic Requirements and Benefits Study
as well as the LiDAR requirements study.
[12:59] We need more research and better tools to integrate the two datasets. Last of all,
I'd like to thank my two partners, Santa Fe County and Bohannan Huston. Thanks.
Al: [13:11] Thanks, Liz. That's very interesting presentation. That will conclude our presentations
for today. I appreciate everyone's participation and attendance today in our seminar. Thank
you.