Lidar Surface Water in Stormwater Network; Updated DC Watersheds

Video Transcript
Download Video
Right-click and save to download

Detailed Description

National Hydrography Advisory Call:

This episode features speakers Roger Barlow (USGS Physical Scientist) and George Onyullo (Environmental Protection Specialist, D.C. Department of Energy and Environment) Roger discusses his work with The Department of Energy and Environment and the D.C. government for integrating Lidar-derived Surface Water with Stormwater Network and Updated Watersheds for Washington D.C. into the NHD and WBD.

Hydrography for the Nation:

High-quality hydrography data are critical to a broad range of government and private applications. Resource management, infrastructure planning, environmental monitoring, fisheries management, and disaster mitigation all depend on up-to-date, accurate, and high-quality hydrographic data. The U.S. Geological Survey National Geospatial Program National Hydrography Advisory Call has initiated a series of virtual seminars to highlight the uses of hydrographic data. These presentations are intended to share success stories from users who have solved real world problems using hydrography data, provide information about the National Hydrography Dataset and related products. The USGS manages surface water and hydrologic unit mapping for the Nation as geospatial datasets. These include the National Hydrography Dataset (NHD), Watershed Boundary Dataset (WBD), and NHDPlus High Resolution (NHDPlus HR). Hydrography data are integral to a myriad of mission critical activities undertaken and managed by government entities (Federal, State, regional, county, local, Tribal), nonprofit organizations, and private companies.

For more detailed information on national hydrography products visit


Date Taken:

Length: 00:51:27

Location Taken: Reston, VA, US


Al Rea: So this is Al Rea, and this is the USGS hydrography advisory call. Today, we're going to have Roger Barlow talk about a project that he's been involved with for developing a lidar drive to NHD and WBD for the District of Columbia. Roger, do you want to start giving a little introduction? It's still downloading.
Roger Barlow: Okay.
Al Rea: So as soon as it gets finished downloading, I will share it.
Roger Barlow: Okay. Thank you, Al. I am a geospatial liaison for the National Geospatial Program in the mid-Atlantic states, specifically New Jersey, Maryland, Delaware, District of Columbia and the Chesapeake Bay watershed, and Chesapeake Bay watershed drives a lot of science and decisions and ways that we look to analyze various environmental factors, so Chesapeake Bay is one of the primary ecosystem study areas for USGS and has been received this year an extra $2 million, which was a surprise, so that has been definitely a favorable involvement for me in cross-cuts with other of my colleagues in adjacent areas. Before we get started, I do want to mention all the partners that help ...
Attendee: Trav, did you want to hear this?
Attendee: You got it?
Attendee: I have it up. I just got to put it on speaker is all.
Al Rea: Sorry. I'm going to mute everybody. Hold on just a second, Roger.
Roger Barlow: Yep.
Al Rea: Roger, now I need you to unmute yourself. I'm looking for you in the list. Ah, I found you. Okay.
Roger Barlow: Okay, can you hear me?
Al Rea: Yes, I can hear you, Roger, and you should be able to see my slide, or your slide now, shared.
Roger Barlow: Indeed, I can.
Al Rea: Okay, so just tell me when you want to advance slide.
Roger Barlow: Okay. Thank you. So the first thing I want to do is thank all the people who are involved in the project, and it was a tremendous effort on a number of fronts. So first of all, I want to mention my partner, George Onyullo and Young Tsuei at the DC Department of Environmental Protection who funded this project, and George and I had a similar vision from the moment we got the stewardship, NHD stewardship agreement signed back in 2014, 2015, and this was part of that, so we're very excited to have this up and available as part of the NHD and the WBD as of late last summer. Also, I want to thank Dan Barker of DC Water which provided us with stormwater network and infalls and outfalls, one of the versions of it, anyway, and I want to thank Quantum Spatial, who was our contractor who did the work to put this into a form that could be ingested, and I want to thank Dave Anderson of NGTOC at USGS, who managed to squeeze and shoehorn the thinned version into the NHD, and I want to thank Kim Jones for doing the WBD plein air work of getting this topographic and stormwater data into the Watershed Boundary Dataset. So let's proceed to the next slide, please. So here is what we're looking at in 2018. This is the 1:24k representation of the District of Columbia which resides north of the Potomac River, and as you see, it's pretty minimalistic, and this goes to what I term the urban desert effect whereby there is really a lot less hydrography shown that it exists, and that's because of impervious surfaces in all the large cities in the mid-Atlantic, and since I have relationships with DC and Baltimore and Philadelphia and the New Jersey suburbs and New York and the New Jersey suburbs, that's a considerable part of the area that I do a lot of coordination and partnership with. So this was, when you look at the NHD, there were obvious gaps in the urban areas. It was just a lot of content, so that is a problem that, from the USGS standpoint, I thought we should try and address. From the Department of Energy and Environment and the DC government, there was a need to establish a better mechanism to calculate total maximum daily loads as it relates to Chesapeake Bay requirements. Can we go to the next slide, please? So I've already addressed some of these, and of course, in some of the early NHD data, the stream data is discontinuous because it goes underground, and there's just no connectivity between the stormwater system and the surface water, so that provides a lack of comprehensive understanding and linkage between the two systems, and stormwater is much more of a thing now than it was 5 to 10 years ago. A lot of people are really starting to pay attention to it. There's been some massive civil projects, DC having one of them where they hollow out underground chambers that can hold hundreds of millions of gallons of combined sewer overflow to be processed later, so this is something that is really starting to affect how people calculate overall nutrients and sediment and impacts of sewage overflow into the surface water system. Also kind of pivot-y, with the State of Maryland, from which all the water the District of Columbia flows, was an issue, and we'd like to have those watersheds, particularly the Anacostia and Rock Creek, be all one with continuous and consistent representation into the district. Can we go to the next slide, please? So we're looking at using lidar to create elevation-derived hydrography, and we are going to combine that with the DC stormwater network and the infalls and the outfalls and various culverts and tunnels and pipes to create a connected network integrated between the surface water and the stormwater. Additionally, work will be done by the Maryland Department of Environment that created the enhancement of perennial drainage for the Upper Rock Creek watershed which was outside the funded project area but will used in the upcoming StreamStats project. And we created two content levels: a thinned network which would be held in the National Hydrographic data set and an unthinned version, the full content of which you'll see an example of further on, that will be delivered to the district government and which they will maintain both versions, one internally, one externally for the NHD. Next slide, please. So here is the project area, 8 HUC12, and it goes considerably into Maryland, as you can see if you stare at the map long enough. Next slide, please. So this outlines, in the dark area, the Anacostia watershed, which is the bulk of the water following into DC, but it also highlights the impervious area in the white area, and that gets fairly considerable within DC. When you look at the District of Columbia, there is one unimpervious area called Rock Creek Park which goes from the Potomac River up to the northern point of the District of Columbia. That's all National Park Service land. DC is a little bit of a different character in the East and has almost 30 percent of its land owned by the Federal government. Next slide, please. This is another illustration in dark mustard color. You can see the impervious area, and that's the area where the stormwater system is doing all of the work for getting water off the land. Next slide, please. And here is another representation using land cover and showing the impervious areas and the heavily built-up areas. There's considerable residential areas up in the northwest and northeast and southeast District of Columbia, and DC has increased its population in the past few years, and we're well over 700,000 people now. That is up from less than 600,000 back in the early 2000s, so a lot of inflow and redevelopment going on. Next slide, please. So this is an overall production flow provided by our partners, commercial partner, Quantum Spatial. Next slide, please. And so in terms of preparation, we had a lot of data to assemble and provide to QSI, including the existing lidar, the infalls and the outfalls, which were reconciled in 2017. There were three different versions, one held by the Department of Environmental Energy and Environment and DC Water and DC Octo, so those had to be reconciled and made into a single version that everybody agreed on, and that took 4 or 5 months just by itself. There is considerable checking for connectivity, for pipes and tunnels and culverts within the district. Fortunately, DOEE had a staff member who is really familiar with all of them and could immediately go and check any of them. They also made a tool using oblique imagery and orthoimagery and lidar and any other data available, including the stormwater network, that they assembled to try and work out some of those connectivity issues just from the office. In the end, there were still things to resolve, particularly flow direction in the stormwater network. And a joint funding agreement was signed, a task order created, and in 2017 and '18, work got underway at QSI. Next slide, please. So one of the things I have to stress here is the amount of field work that went into going out and ascertaining all the information, and this goes specifically to my partner, George Onyullo, and his assistant, Young Tsuei, going out in the middle of streets, putting up traffic cones and pulling manhole covers to look at flow direction, particularly in the stormwater aggregation aspect of the project, but you can see the field annotations here as to how things were really connected from manhole to manhole or not, as the case may be. Next slide. And this is just another look at that. Next slide, please. And yet another version of surface water being shown flowing into one of the manhole culverts. It was a considerable amount of work in cold winter conditions, in some cases, so I heard about that. Next slide, please. Another kind of a standard slide from QSI showing flow accumulation. Next slide, please. And the attribution of feature codes. Next slide, please. And District of Columbia does have some waterbodies. Most of them are reservoirs, and they are being captured from orthoimagery, which DC flies every other year, and in fact, they're flying lidar now every other year as well, and they're getting ready to collect another QL2 version this spring. Next slide, please. So here is the final NHD project view, so in the red, you can see the 24,000 scale representation prior to the enhancement of the integration project. You can see the surface water enhancement from lidar in blue. In white, you can see the thinned stormwater network which is in the NHD, and in black, you can see the full content stormwater system. This is a tremendous contrast to the second slide we showed with about your 10 different streams, and this will be used by the DC government for a number of things, including trying to figure out some of the stormwater flooding issues when combining the stormwater with the surface water in lower-elevation areas. Stormwater flooding is easily the most frequent type of flooding in the DC when compared to river flooding or tidal flooding. In addition, the extra stormwater calculations will be used to refine the TMDL calculations by the Department of Energy and Environment for District of Columbia, so those are the two major use cases, and I think there will be some other restoration and best-management practice applications further down the line. Next slide, please. Another version showing the entire project and the enhancement outside the District of Columbia which was less. We didn't have the stormwater network outside of the district, so that's just a comparison view. Next slide, please. So this project relates to the silver jackets, and for those of you who aren't familiar with silver jackets, that is a euphemism for flood mitigation that is a nationwide network by state or jurisdiction that coordinates a number of Federal and State agencies together to look at flood mitigation projects, and so Watts Branch in northeast District of Columbia is a stream that flows, overflows in most streams above the 25-year event, and there's a particular culvert which blocks things up, but without the overall stormwater network involved, you don't understand the entire inflow in and around that particular blockage which is kind of right in the upper middle area. It's a big, long culvert of about 100 yards, and so there's other constraints as well, but understanding where the stormwater comes and adds to the overall flooding network will help understanding what measures to take to help with that. Unfortunately, the biggest constraint is the I-295 and railroad conveyance which the railroad has now shown any interest in enlarging, and that's kind of that northeast to southwest linear aspect near the left side of the watershed, but there's other things that can be done upstream, just an example of how the stormwater can be applied to some of the stormwater flooding aspects of local government. Next slide, please. So this is another example of what the 24,000 scale looked like prior to enhancement, and this is the Upper Rock Creek watershed in Montgomery County, adjacent to District of Columbia. Next slide, please. And you can see the difference as additional perennial streams and probably some intermittent streams, plus some water bodies added makes to the overall network, and my thanks to Frank Ciano at Maryland Department of the Environment. He is the NHD and WBD steward for the State of Maryland. Next slide, please. So what we run into with this high level of content is, the NHD is not built for the heavy level of content, and so we went with the thinned network for a couple of reasons, so, one, so it would fit into the NHD, and we also gained favor with DC Water because they were not entirely comfortable with the full resolution stormwater system being publicly available, and so when we came up with the thinned network, that really eased their concerns and helped us move forward with the project. WBD also had some limitations because of, there was no standards for the 14-digit and the 16-digit watersheds, and those had to be kind of drafted as we went through them for the Urban Environment Group. Another barrier was the limit of 99 subunits. There were a number of small stormwater sheds in the District of Columbia, and so we had to aggregate them to come within the 99 subunits to aggregate up to the 14-digit HUCs. Next slide, please. And so here is an example of how that all fit together. If I remember correctly, and if George is able to speak, he can straighten me out, but as I recall, we did the 14-digit HUCs first and then the 16 using the stormwater, and the 16s were only done in the District of Columbia so that the stormwater and the MS4 outlines were tallied up against that, and then we used the network of the actual full-content stormwater pipes to help delineate some of the 16-digit HUCs, and then the 12-digit HUCs were revised, and you can see the yellow lines there as they were once represented and then revised to the red lines there. Next slide, please. And so here is a list of the benefits of the project, and the one thing I do want to stress again is, this is a field-verified and hydroenforced model, so we know the connectivity is as it is depicted on there, and that is really important because previous versions of stormwater, and I'm thinking particularly of my colleague's entry some years ago in Los Angeles, was provided, but it was never field-checked, and so this is a little bit better, and then in the St. Louis Stormwater Project, which was done by StreamStats folks, their line work and network was never submitted to the NHD, nor were their watershed polygons submitted to the WBD, so this kind of puts it all together, and I know that my colleague, George, is pretty pleased with this. And so we're looking to apply this to other cities, and I made conversation with the City of Philadelphia over the last couple of weeks to possibly bring this process to there, to them. One other aspect, I have looked at culverts across the nation, and there is a Northeastern or a national culvert database at the University of Massachusetts, but there's no common attribute schema, and I'd love to have one developed, been trying to float one of about nine attributes, just the basic stuff that everybody would use, and everybody kind of nods their head, yes, but there's been no real conversation on that. The new specifications for elevation-derived hydrography that USGS is putting out helps address some of the culvert issues as a feature in NHD, so we'll see if that comes to pass. Next slide, please. So this is not just a project in and of itself, and we currently have the StreamStats agreement with the DC government that hopefully will be signed this week, and so we'll be using the project that was developed here to feed that project along with the lidar, and then we'll have even better chance to calculate some of the TMDLs coming out of the stormwater system. Also, Frank Ciano at MDE has graciously offered to enhance two of the Northeast and Acostia watershed, the only two 12-digit HUCs that are missing, so hopefully, he can add to the overall network on that remaining gap in content. Also, in the future, NHDPlus, we'd like to put the stormwater system through that type of a rigorous enhancement and see how that works, and then we go to best-management practices, BMPs, for stormwater, understanding the transportation and what enhancements need to be made to that to keep that from flooding. Also, there has been talk of adding flow regime, the perennial, intermittent and ephemeral coding to some of the streams, so that is something that the Department of Energy and Environment is interested in doing down the road, and the next project, question mark of, is pertaining to the potential for Philadelphia being a follow-on project here. Next slide, please. So here, we have the groundbreaking aspect of the project, the first field-validated stormwater content conflated to the NHD, the first integrated field-checking stormwater, infall and outfall to surface water in NHD, and the first WBD entry using not just surface topography, but some of the other stormwater data that helps form the watersheds. And at the end, this project was nominated for the Caffritz Award for civil works in DC, and my colleague, George, has been helping our nomination process by creating a text for submission to put us into the final evaluation for who might win that award. Next slide, please. That's it. This is from the top of the Kennedy Center, looking upstream and the Potomac River. That's the Georgetown Waterfront there, and at this point, I'll take any questions or comments. Stunned silence.
AttendeeCan you hear me?
Roger BarlowI can.
AttendeeThis is Sean Von from Minnesota. I have a few questions and some real quick background here so you understand my questions, 15 years of advocacy for deriving hydrography from lidar, and know all the pros and cons and nuances. One of the things I was hoping to get out of this presentation is a little bit more about explicitly how lidar was used. I saw a tremendous amount of field input on-site drive attribution as well as hand delineation. Could you go back to some of the slides, and one in particular that was more related to flow network and talk to us a bit really about the absolute role that lidar products played in this? And then lastly, hold off on this, but talk a little bit about lidar quality level. We'll hold off on that one first, or cover the first one.
Roger BarlowOkay. So Lidar was certainly important for the water bodies, and the Department of Energy and Environment actually did a wetlands update while we were going on with the project and used those polygons as well, but lidar was important for all the surface water and that linkage between the surface water and the stormwater system. Certainly, QSI developed the enhanced stream content from QL2 and from Montgomery County. At the time, we had just the QL3 lidar, but putting the streams in the bottom of the channels was absolutely important for making the alignment in the system connect and work together.
AttendeeAnd so, was that standard flow accumulation threshold work, that you're just putting a flow line into what I would describe as a water conveyance feature identified within lidar-derived products?
Roger BarlowWe certainly use the D8 flow accumulation as part of the derivation of that, and I think Mischa Hey or other folks from QSI can certainly address that in [Indistinct].
Mischa HeyHi, Roger. This is actually Mischa Hey, if you want me to touch on that a little bit.
Roger BarlowSure, please.
Mischa HeySo, yeah. During this process for the flows, the surface water was primarily a flow accumulation, flow direction process, a lot of hydroenforcement required to get that to reach false obstructions, as many of you are probably aware. One component of how this really, the lidar really supported this, is integration and particularly the development of watershed boundaries and fitting criteria for the urban site where we have the CSO. The CSO is a many to many inlet-to-outlet relationship, and so the lidar allowed us to essentially draw micro [Indistinct] for every inlet so we could then push that flow accumulation through the subsurface network because unlike a surface network, the water enters constantly throughout the path of that, and the subsurface network, it only enters the system at inlets, so [Indistinct] one other component where the lidar made this kind of possible.
AttendeeWell, that helps a lot for me. Now, talk a little bit about quality level, quality level three, certainly some of our older lidar data holdings across the nation ...
>> [Indistinct]
AttendeeSir, I didn't ... Would quality level one have improved your project had it been available?
Mischa HeyI believe, I'd have to double check. We didn't have a consistent quality level throughout, but most of this was, I believe, done with quality level one, and then we had a little bit of quality level two. The difference there and the elevation model really depends on how much above-ground obstruction you have, so we're typically working off 1 or 1/2-meters DEMs. On a clear open desert, quality level two is still going to support that. As you get into a more occluded ground, then you would need to start thinking about quality level one and QL0. Does that answer your question?
AttendeeYeah, it helps a lot, and just coming from Minnesota perspective, we're on a plan to acquire new lidar data statewide and advocating for quality level one for the reasons that you have indicated, but also looking for quality level one to have a longer shelf life and provide value of benefits for a longer shelf term over the course of the next 10 years of application.
Roger BarlowJust to add onto that, it was quality level two for the entire area except for Montgomery County, which was quality level three. What I can say is, last week, quality level one data was delivered to Fairfax County, and that was funded by their stormwater folks, so that is clearly desirable from the stormwater standpoint.
AttendeeGreat, thanks for the clarification. Thanks, guys.
Al ReaOkay, if anyone else has more questions, you need to go to yourself in the participants panel and unmute yourself. Some of you may have trouble if you didn't give it your participant ID number when you joined. I can turn off mute all here and see if it works. I'll do that.
AttendeeHello? Hello?
Al ReaYes, we can hear you.
AttendeeHello? This is Anita from Washington, and I have a question. Can you hear me?
Roger Barlow Yes.
Al ReaHi, Anita. Go ahead.
AttendeeOh, okay. I've been struggling a little bit with this same issue. We've got a project in King County, which stormwater project, and it drains to Puget Sound, and it's my understanding that you're putting in the thinned network into NHD. I'm a little bit ... I'm wondering kind of of how you chose what to put into NHD, and how are you coding those stormwaters, like the white stormwater networks? Are those pipes, or ...
Roger BarlowYeah, the white is all pipe.
AttendeeOkay, and those are in NHD, or you're going to maintain those in NHD?
Roger BarlowYes, our steward will maintain those as changes come about in concert with DC Water, the water utility.
AttendeeAnd what was the ... Can you explain just a little bit about your thought process about how you chose which ones, what quantity of stormwater pipes to put into the national dataset?
Roger BarlowI'm going to ask Mischa for ... Let's go with if Mischa can comment on that.
Mischa HeyYeah, I can. This is an interesting and very [Indistinct] project. So the NHD data structure has limitations to it, and it was designed to meet a particular need, and the resolution of data that we're trying to push into that structure, it breaks the system, so there are two goals for this. One was to achieve roughly comparable [Indistinct] density to the surface water area. That was a semi-arbitrary but necessary step in order to not rake the inclusion rules, such as features being too close to each other, too many features, short line segments and things along those lines. The selection process for prioritization of what is included, I actually ... I generated something to map unique outlet combinations for every inlet, so as I said before, it's a many to many relationship, and then we pushed the flow accumulation actually through the subsurface network, and so that was our initial prioritization of, these seem to be the major conveyance pathways for subsurface flow, and then we added and filled in other stuff to approximate the feature density of the surface flow features outside the area. That relatively clear?
AttendeeYep, that sounds good.
Roger BarlowAnd from a cartographic standpoint, this provides a fairly good representation of the stormwater network in a thinned and digestible way without going onto the full-on content, which NHD just can't digest.
Al ReaOkay. Do we have other questions? Again, you may need to unmute yourself.
AttendeeHi, this is Jason from Missouri.
Al ReaHi, Jason.
AttendeeI just got a question on how you plan to keep this stuff updated with all the field work that you did. You'd be ashamed to waste all that if somebody came in and redeveloped a bunch of stuff and didn't provide any updates or anything, or is this a passive, you're going to collect the lidar from the next 2 years to update the system, or just curious how you plan to keep it updated.
Roger BarlowSo the stewardship by the Department of Energy and Environment is ongoing with DC Water, which is the utility, and so when they have upgraded projects or something new, that will be added to the network, and they meet regularly and the general GIS council, so that dialogue between those two agencies is certainly on a quarterly basis, if not more. So we're depending on them to keep it updated. That's for their own benefit, and they certainly understand that better than we do.
AttendeeOkay, thank you.
Al ReaMore questions?
Attendee There's some questions in chat.
Al ReaOh, okay. Let's see. The contact at UMass for the national culvert database?
Roger BarlowYeah, I don't have that. I visited it once, and I found it online, but it's haphazard and sporadic. It's by no means comprehensive, that I can tell you. It's volunteer, but it's probably the most comprehensive of any other place in the nation that I know of, and I've talked with colleagues at NRCS about this, and they're not aware of anything better than that in terms of overall collection, but I've advocated a culvert database on a statewide basis because it's ... In Maryland, there is an excellent culvert database which is a combination of FEMA region three and the Maryland State Highway Administration combined together, and that's a good start, but by no means is that all the culverts in Maryland.
Roger BarlowAnd so there's a lot of local government culvert activity. There's certainly private culverts, and there's probably some in other administrative layers as well. It's not easy, and that's why the attribution schema is, to me, an important thing to figure out. If we can figure out who owns the culverts and maintains them, that would be a step forward for any number of agencies, and we did try to get the DC Department of Transportation to take ahold of this, and they nodded their head in the affirmative, and then nothing happened, so it's not easy to get somebody to take ownership of this, and by the lack of overall data on the statewide or even a county basis ... I did do a canvass of several counties around Baltimore, and I found out who to contact for the culverts, but nobody was ready to provide the metadata for the culverts, and I did ask for that. That was, to me, the first step, is getting the metadata for culverts in various counties.
AttendeeAnd others, this is Sean again from Minnesota. I support what's being said about the difficulty in acquiring culverts. It's an effort we've tried to do here for nearly a decade in Minnesota. Many folks on this call know how valuable it is to have a culvert inventory for what we describe here as DEM hydromodification, where you are essentially removing digital dams, those locations where culverts are in a lidar-derived Digital Elevation Model. One thing that hurts us in those efforts to acquire those data or find funding to do culvert inventory is the assumption that culverts have been inventoried. It's so common for people to say, "Oh, well, Minnesota had a culvert inventory because our DOT has inventoried culverts." The inventory of culverts is taken on as an asset to an entity, and folks only inventory what is their asset, so you can have a county inventorying culverts, but they will park their truck over a road that belongs to a township and not inventory the culvert because it's not their asset. Culverts are in the millions upon millions, and lidar cannot, absolutely cannot, produce a DEM that routes water through it for modeling purposes, and collectively, as a nation, we just don't drive this hold enough, to help get decision-makers to realize that culvert inventory is necessary in DEM hydromodification or conditioning, whatever you want to call it, absolutely must be done if you're going to use it for modeling purposes, and so I like presentations like we've seen here today because it helps build that groundswell of understanding of where we need to go to have lidar-derived hydrography.
Roger Barlow[Indistinct] agree, and by the way, Collin McCormick of NRCS forwarded me the Minnesota schema, so my knowledge, or at least initial knowledge, of your work there on this topic is known to me.
AttendeeYeah, that's great, and we've established a Arc Collector culvert inventory tool because we need our field staff out there just inventorying even a culvert underneath a state trail can absolutely derail the routing of water in a DEM, so the importance of inventorying every culvert really becomes critical, and it's a need that needs to be in that conversation for funding and project scope and definition.
Roger BarlowAnd for me, particularly in the urban areas where culverts are the norm rather than the exception.
AttendeeYeah, absolutely, and we can't overlook things like beaver dams and dams and water control structures. Anything that impedes the flow of water on the landscape will be interpreted in a lidar-derived product as what we call a digital dam and must be dealt with. Otherwise, you can't route water [Indistinct] water. You can route water up a hill. You can route water backwards. We buy lidar data to replicate the connectivity of Earth's surface in the hydraulic realm, but we cannot to incorporate and achieve funding for modifying the DEM to provide the products for which we marketed lidar for. It's a really, really difficult business model that we have in this country, and that's why in Minnesota, we acquired a statewide lidar collect 10 years ago, and we still don't have a modified DEM where we can route water across the state. It's just a difficult topic.
Roger BarlowThank you.
Roger BarlowThanks. I agree, and I think as a group, the NHD advisory folks should probably list this as a portion of the hydrographic infrastructure, at the very least, and do an inventory of what available data can be provided by various stewards and possibly work with the UMass folks to enhance what they have or figure out a better way, if there is one because as we get into finer scales or larger scales of data, nothing is going to flow unless we understand where the culverts are?
AttendeeAnd I think for the purpose of this conversation, folks who aren't immersed in this topic may not understand that culvert inventory is actually only part of the equation. In most cases, you would have to create a breach line of sorts, a hydro break line, I should say, that represents the location and the positional accuracy of the culvert to be utilized in the DEM, and so we have the cost of the culvert inventory. Then we have the cost of what we call here in Minnesota creating the digital dam breach line, and those two things have to go hand-in-hand to create a product like we see on the screen before us.
Roger BarlowYeah, I would have that as one of the key attributes for attribution schema for culverts.
Al ReaYeah, this is all really good discussion. There's a couple more questions in the chat that I'll try to get to. So there's a question about channel heads, and was a particular flow accumulation threshold used, or what other, or some other method to identify channel heads the beginning of a synthetic stream, if you will?
Roger BarlowMischa?
Mischa HeySo headwaters are, once again, really complicated, so we start with typically a 2-acre threshold. That yields a lot of commission errors. We've started to rely pretty heavily on direct detection of the channel character within the landscape, looking at topographic openness and geomorphic classification. It's currently a bit of a blend between those two. Regardless of that, our friends at the University of Maryland have actually been spending a lot of time looking at direct channel detection through topographic openness, geomorphic character of the landscape and then gone out to the field to verify headwaters with significant effort of hiking, and it's a relatively beautiful country, so I don't feel too bad for them, but what they're finding is neither flow accumulation nor channel detection is really perfectly aligning with the true initiation point, so when that stream leaves the ground or becomes an actual stream still remains a bit of an uncertainty, and so this is our best guess based on what we can tell from the lidar-derived landscape, but we do not actually expect the headwaters to be perfectly accurate at their initiation point by any means.
Mischa HeyDoes that answer that question?
AttendeeYeah, thank you.
Al ReaYeah, there would be lots more discussion, I think, on that point as well.
Roger BarlowBest available, based on the data.
Al ReaYeah. So I think we're running up against our time limit for the hour here. I really appreciate, Roger, your presentation and everyone who's participated in the questions and discussion. It's been a really good discussion, I think. So thanks again for everybody's coming today. We'll have another call next month. Usually, it's the fourth Tuesday of the month, and we'll send out an announcement. And I think next month, there may be a conflict with the AWRA conference, so it may be early, so I'll send out a calendar invitation soon for the next call. Thanks. Thanks again, everybody. We'll talk to you next time.