Innovations in Water Quality, The Real-Time Revolution

[00:00:00] Max Herzog: Welcome everyone to the first edition of the Water Data Forum of 2022, my name is Max Herzog I'm a program manager with Cleveland Water Alliance and really excited to welcome you to  a great panel discussion here today, dealing with topics in water data, particularly focused on, real time water quality and water monitoring.

[00:00:26]  This is the first edition of a series here that we're running. We started last year, with Cleveland Water Alliance in partnership with the Water Environment Federation, or WEF, and the Midwest, Midwest Big Data Innovation Hub really focused on Making a broad range of topics in the field of water data more accessible and available to folks across the country and the world.

[00:00:50]  So really excited to have you here with us today. If you enjoy what you hear today, we're going to be doing these sessions about every other month, every 3rd month this year, so we'll definitely want to get you tuned into future sessions as well. As we go through today, we're going to have kind of an open panel discussion kind of format.

[00:01:12] Well, where I'll serve a few questions to our panelists, but we should have, you know, 10, 15, maybe 20 minutes at the end for open questions and answers with panelists. With this format, you won't be able to unmute and ask your question directly, so what we prefer instead is if you go to that navigation bar at the bottom or top of your screen and drop those questions in the Q&A throughout, you can enter those questions whenever they pop up, and we'll get through as many of those as we can as we get to the end of the presentation today.

[00:01:44]  As I said, I'm Max Herzog, a program manager with Cleveland Water Alliance. And I'm really excited to facilitate this panel today on innovations and water quality with a focus on what we call the real time revolution in this transition to much higher frequency collection of water data.  It's my great honor to facilitate a fantastic panel today with speakers from across a range of occupations and focus areas in the water sector.

[00:02:18]  And dealing with water data and then a real diversity of different approaches.  First we have Ting Lu, business practice leader in digital solutions with clean water services, up in the Pacific Northwest. We have Adam Hobson hydrologist and application development manager at In-situ, a private sector sensor manufacturer.

[00:02:45] And, we have Ed Verhamme, principal and senior engineer at Limnotech Water resource consulting firm, based out of Ann Arbor. So we've got a real range of folks here thinking about the development of sensors and their application, getting those sensors out into the field. And, you know, with Ting, really that focus on how these can inform watershed management and utility management.

[00:03:10]  So really excited to dive into this conversation today.  With that, I'm going to stop my screen share so you can see our great panelists' faces, and we'll dive right into our conversation today. So I want to get things started, and sort of by way of introducing our panelists. So I want to just dive into this kind of getting a high level understanding of this topic we're trying to address here.

[00:03:40] And so I want to do this by framing a couple of, this question here to our panelists.  How do you see the transition to real time water monitoring, changing the process of collecting data? And transforming it into useful information in your field or your area of the country, knowing that you each come from a pretty different perspective in terms of the water sector, and you also have your own geographic interests as well.

[00:04:05]  So I think I'll serve this question up first to Ed if you wouldn't mind commenting on how you're seeing this transition from, you know, lower frequency sampling to real time monitoring, you know, affecting your work and your field.

[00:04:23] Ed Verhamme: Thanks, Max.  Hi, everyone. I'm here in Ann Arbor, Michigan. I think the best way to introduce that topic is to just talk about my background here.

[00:04:35] I sort of chose this one. This is Lake Erie behind me. Cleveland is off in the background and their water intake is here up front. And, you know, I tend to explore the relationship between data and timeframes to make decisions. And so we're all sort of familiar with it. Sending out field staff and collecting water samples.

[00:04:56] I think that the technology here and this sort of real time. We're not necessarily replacing those people's trips, but we're trying to find ways to monitor these systems when we're not there and how do we expand that monitoring and have people not be the limitation to do that. So I've always seen Technology as sort of a yes and? You know, that we need to keep growing the space as opposed to doing this, these swaps and just recognizing where the skill sets fly with sending people into the field. It's a little bit different than how we set up and program and maintain technology. So there's a tech transfer, we're sort of asking some of our field biologists to become IT specialists in some cases. And so I really like to talk about that base skill set. And I've also just to talk about the future.

[00:05:55] I also want to see our universities and our educational institutions respond to this new need for this cross training. I see very few institutions arming sort of the quote field biologists with some of these sort of advanced or just technology related skill sets. So that again, I'd My perspective is mostly on large water systems, great lakes, and I think the unique aspect is, that drinking the water and that decision, of what's happening with the water right now. So thanks, Max, for that opportunity to introduce.

[00:06:33] Max Herzog: Thank you so much, Ed. Yeah, and I think that's a great transition over to Ting's perspective,  with clean water services. So I'm wondering, Ting, if you could speak a little bit to your work and how you're seeing this transition to real time monitoring really impacting that.

[00:06:48] Ting Lu: Yeah, thanks Max. And hello, everyone. So I want to talk about my career and also how this interweaved with the sensor development and the water quality monitoring. So I still remember I started testing TSS and nutrient sensors about 13 years ago. When I was a process engineer in Ohio. So back then it was to get operators and businesses to trust this technology so that you don't have to manually grab samples.

[00:07:15] So what I see the impact of this real time water quality sensor is to adopt with the adoption of all the sensors and the treatment plan was able to do a lot of more advanced controls for a region or nutrient control. And that actually, it's a big part of how we can manage what is shared as a whole and also with the real time sensors at the freeze up the operations time to analyze the data and transform perform process control rather than spending time to do a lot of manual collections there.

[00:07:50] So then about six years ago I started my career at the Clean Water Services in Oregon. So we're Clean Water Services already have a great foundation of sensor networks. The opportunities are more about how we identify more cost effective sensors. How do we look for more technologies on a large watershed scale to help us better understand flooding conditions, water quality and other things.

[00:08:16] So we have the IOT team within Clean Water Services that is looking for different ways to do this. And so one of the projects we partnered is actually with Ed and the Limnotech colleague and developed our IOT DIY sensor stations, that's able to monitor water quality and a flow on a continuous basis. Again, it's getting to use the information to guide us in understanding water quality improvement, calibrating water quality models, inform our decisions or investment on stream restoration project water qualities on a holistic basis.

[00:08:56]  So it's like the years and the eyes for our system that we are able to track down industry discharges, stream conditions and what are quality conditions. So it's been transformative.

[00:09:12] Max Herzog: Absolutely. Yeah. Thanks so much for sharing that, Ting. I think it's so interesting to hear your perspective really ranging across, you know, all the different types of monitoring from the very established sensor brands to these more open source DIY approaches. I think that dovetails well into Adam's approach and perspective, coming from in situ, a sensor manufacturer, but one really focused on trying to bring the cost down on some of these base level monitoring components. I'm wondering, Adam, if you could speak to how you're seeing this transformation impacting your work.

[00:09:44] Adam Hobson: Absolutely. Yeah. Thanks, Max. And hello, everyone. Thanks for having us on this discussion. It's really exciting to be here. So again, from my perspective, as kind of a hydrologist and application development manager for a, again, an equipment manufacturer who's actually creating the sensors and the telemetry on this stuff.

[00:10:01] I think we have looked really at this idea of real time or almost what we always refer to as near real time data or this idea of water monitoring really pertains mostly to water quality.  Just to be kind of clear on that. And the use for that could be a lot of different things. Ed mentioned some of those things. Ting mentioned some of those. 

[00:10:19] There's also a lot of public safety concerns here that are really critical now to what we have. And the data that we're talking about here is often being collected at high frequency. We're talking data that could be, you know I think traditionally data used to be connected. You know, maybe it was only once a week or maybe once a day.

[00:10:37] But now we're talking every hour, every 15 minutes or even faster. And then that data is ultimately then being transmitted up to again to anywhere in the world and to make decisions.  And that's a really big part of it. I said, for instance, you were building, you know, water quality sensors for also level sensors and flow sensors, but also the telemetry units as well for that.

[00:11:03] And also the data management software because one of the big challenges we have here is we get more data coming in. We have to manage it. You can't just assume it's all going to be, you can use the same techniques as used when you had less data coming in. So I think, you know, with real time monitoring with the more frequent data you get, we can actually better understand details of the hydrologic system or any type of water system, which is really critical.

[00:11:40] We were not missing events, not missing data that may be there, you know, whether it's a water treatment plant discharge or an intake or something like that. This real time revolution is allowing us to better manage our water resources, is what it really comes down to.

[00:12:02] But I do also want to echo that, I think what Ed said, which is very important that this is still a need here for manual data collection for people to be in the field for us to lay eyes on things. We have to be able to continue to do that. This is not fully, it's not fully automated.

[00:12:25] It, I don't know if it ever will be to be honest, but it does by having more real time data again, we get more information, more data, we can make better, faster and more effective decisions that ultimately save costs on how we manage our resources and protect the environment and our precious resources.

[00:12:46] Max Herzog: Absolutely. Thanks for sharing that perspective, Adam, you know, I think you touched on a couple of really Key things, you know, talking about how this real time revolution isn't just about the sensors. It's about the way we're pulling data off those sensors and then managing those data. And I know folks are going to touch on that a little more as we move forward.

[00:13:04]  And then talking about some of the unique challenges of monitoring in the water space and talking about, you know, near real time, maybe not to be a sexy sounding, near real time revolution, but the fact that we don't really need data every 10 seconds, when we're monitoring water in most cases and and a lot of the challenges are just making sure that these sensors continue to work in the field because of the harsh conditions that they're experiencing.

[00:13:30]  I think that's a really great transition to talking about our next question, which is really about the challenges and barriers that impedes some of the speed of this transition and really determines the scope to which it can have impact. So I'm wondering. You know, we can follow the same order, I think, for this cadence here.

[00:13:51] If Ed, you could speak first to some of the challenges or barriers associated with transitioning to real time or neo real time,  as it were, and what opportunities you see,  around recent innovations, whether it's in technology, support infrastructure, support programming,  that folks are engaging with to try and address some of these challenges.

[00:14:14] Ed Verhamme: Yeah Thanks, Max. One of the biggest challenges I've seen with this technology adoption or these new options we have is we're not starting from the foundation of the problem or the decision or the phenomenon we're monitoring. We're often jumping to the end of our previous iteration of this phenomenon and saying, well, because of this phenomenon, we monitor this In with people or grab samples.

[00:14:43] And now we're trying to add in this right to the end. I think that the technology we need to go back to that original problem and decision and evaluate time scales, how much data do we need to save? You know, I think a lot of the problems. I see people throw up as problems with real time sensors are because they're growing from a different data foundation that's not back to the original decision here. So I think we really need to peel that back and some of the opportunities are that, you know, technology is way more accessible and available to the environmental field than it ever has been. And I haven't seen us do that. Launching the problems into the technology space to have them respond.

[00:15:30] I've seen us generally limits what we think technology can do and plan procurements and like, well, that's what I can buy, and that's how I'm going to set up procurement. So I think a barrier is putting more challenges to the tech industry to respond to the types of Either price points or capabilities as a specification.

[00:15:52] One thing I wanted to also point out is who's left behind. I think we have this idea that technology is this universal, level or, economics and social issues, I think social media and others have shown it's turn into a different, a very strong differentiator, of equity and justice and opportunity if the only ones that overcome these barriers are already set up to do so, you know, financially training and otherwise.

[00:16:24] And so there's always who are the marginalized communities that are dealing with our water that we're leaving behind as we push forward with these new things. So I don't, you know, it's that's a challenge for us as you know, and part of it is just being open about it. Any new procurement processes, any specifications, allow others to copy and share those.

[00:16:47] So I always, I promote people to sort of publish all these internal evolutions of why you chose AB, you know, C technology, because others can benefit from that. So I'm just really watching you know, even locally, we marched on the Lake Erie shoreline and the capabilities of technology and treatment and knowledge change every 10 miles with a different water intake, a different set of operators, a different budget and county.

[00:17:16] And so I don't know how technology can help have this universal layer.  So that's a challenge. 

[00:17:23] Max Herzog: Absolutely. And one that kind of, in a lot of ways. Moves beyond the capacity of technology to address but also is one that technology can exacerbate as we have access to new tools and then that access isn't universal.

[00:17:39]   I think maybe from here we can transition over to Ting to hear from your perspective, you know, what are the challenges or barriers you're experiencing,  in this kind of transition and how are you seeing folks trying to address those?

[00:17:55] Ting Lu: Yeah, I think I really love what I talked about. So from a utility perspective, I see one of the important factors is data quality, which is not what I'm saying about the technology issue.

[00:18:09] You know, we all know technology is mature today. But its data quality could be the drift could be because of a calibration that is required in order to make sure the sensor is performing well. So, you know, now we look at we have a different skill sets requirements that we can offset some of the graph samples with the real time sensors, but if you're deploying hundreds of sensors on watershed, then you're going to require different needs with the maintenance needs, with the calibration needs, and also the tools that you are able to calibrate and that kind of making sure the quality is what you are getting to truly assessing the water quality there.

[00:18:53] So we do have some tools and then we are  having this real time water quality program that has folks that are dedicated to do the maintenance and work. So what I see the technology could improve is more just a kind of a bio falling resistance and things. And then, just that it's going to reduce a lot of the calibration frequency and requirements.

[00:19:21] And then there is also the engineering design that plays a big part. You know, when you talk with the vendor, the technology is saying, Oh this is what  we get compared to the grab sample depends on where you install your sensors. So sometimes the source system is really rough with the high pH and racks and things.

[00:19:39] So to overcome that, our team has done an engineering design. This is with a 3d printer to print the design and apprentice the guard that actually Keeps the sensors that are clean, relatively clean. So you don't get caught on a lot of this tracking thing. So that's something that I see inspires a lot of innovation just beyond just a typical we talk about the hardware sensors and things.

[00:20:05] The other thing I just want to Talk about is what Adam talks about is that it's not just about the sensors and data, but it's about the telemetry data that you can get to real time and people can actually see the data in real time. So that's actually also another one for utilities, when you look at the whole sensor package.

[00:20:27] It's not just about the sensor cost. Typically now I see the service really as a service so vendors are providing us the monthly service or charge that you pay beyond, besides your initial investment on the technologies. So the way we look at this is that it depends on the fit for purpose.

[00:20:50] Some of the things we use the commercial sensor package all the way to the platform and the telemetry and the platform and the visualization. But we do have the internal capability that we have built over the last few years that we are able to kind of utilize this as a modular approach and do some of this in house.

[00:21:09] So that's really helping with the cost effectiveness and the scalability to adopt a large scale sensor network.

[00:21:20] Max Herzog: Absolutely. And I'm sure that’s something that we see a lot of the larger capacity utilities doing, but something that might be beyond the capacity of some smaller utilities and really exciting to hear about the ways that you're working to address this kind of in the field resilience components.

[00:21:35] I'm sure that that ties in as well to some of the challenges that Adam is seeing.  So maybe Adam, you can share your perspective now.

[00:21:42] Adam Hobson: Absolutely, and again, I really appreciate Ed and Ting's perspective on this because this is, it's a lot of what's driving what it's it to do, to address a lot of these concerns.

[00:21:55] One of the big barriers that we've got, as I think Ed again alluded to, but I think Ed's is a little bit bigger picture there, but really it's cost. A lot of what comes down to this is who can afford a lot of the stuff. And it may be perceived, it may be real, it may just be hard dollars, but it also might just be what's the investment in the time,  to deploy sensors, to deploy networks, to process data, whatever that might be in any way that we can make that, to reduce that cost, is going to allow for better adoption of that.

[00:22:30]  Another part is knowledge. And like the knowledge, the expertise, that's actually required to deploy and maintain all these sensors and telemetry units and do that. Again, there may be a, you know, if I hesitate to use the word underserved community, but it might be where maybe the knowledge isn't there.

[00:22:56]  Where do I need to monitor this thing? Well, maybe we don't know.  Maybe we need to put more sensors out, but how can we do that? Is that something that's going to be affordable? Maybe we have a citizen science group that's out there that really wants to, that they're not hydrologists, they're not electrical engineers, they're not, you know, they don't have PhDs and things like that to do.

[00:23:15] They're just concerned citizens. Well, to create a series of sensors, telemetry, instrumentation that actually can be used by them and that's real. That's not a, it's not a toy. It's it's not just for curiosity. It actually has real value and it's easy to go. So I think really the opportunity that we're seeing on a lot of this is developing this guy that they did the sensors and technology that's affordable, reliable and ultimately easy to use is what we have to, we have to have.

[00:23:48] So again,  you know, I think you may be a hydrologist looking at, you know, water quality on say Lake Erie, but you don't also have to be an electrical engineer how to, you know, and with an enormous budget to set up a telemetry unit, because all you really want to do is you know that there's a harmful algal bloom that happens every year and you want to know when it's going to happen, or is the severity changing, or where it's happened.

[00:24:18] There's so much more that we can do with that. So, you know, one of the things that in situ does, I think very well, to be honest, one of the reasons I actually joined the company is there's this idea of your experience anywhere is your expectation, expectation everywhere.

[00:24:39] So if in everyone, some everyday life, if you can do something really easily, and this is a, you know, with new technologies and this sort of thing, if you can do that, you want that to be in everything that you do. And I think the environmental monitoring field, water monitoring in particular, is finally starting to move that way.

[00:25:00]  you know, if many people can get online and order something, you know, or places, you know, purchase something, well, maybe I want to be able to do that and have that similar experience with my data.  There's a lot more happening. Ting made a very interesting point about data as a service.

[00:25:21]  I think that's a very strong trend. That I think is one solution to help overcome some of these challenges where you're seeing this now as a full service provided to an entity. It's not just about the sensors and the telemetry. It's actually selling or providing the data.

[00:25:45] That's really what people want. It's the data and that information. It's, Hey, your sensor may be, you know, X, y, z, what, whatever it is, and they need to work, they need to be maintained, they need to provide reliable quality data. But ultimately, it's just the data that they're, that people are, are looking for.

[00:26:03] You gotta make it easy to use and affordable, and that's the only way it's going to work.

[00:26:09] Max Herzog: Yeah, I really appreciate that perspective, Adam, and thinking about, you know, the broad diversity of potential users here. I'm really glad to hear you bring up citizen science.  You know, we have some experience at C. W. A. Working with these volunteer monitoring groups across the Lake Erie Basin. It's certainly been our experience that, you know, there's a much lower tolerance for complexity and or, you know, having to engage with the technology just to get it to work. Right. And there's so many folks that want to start these programs that want to know their local monitor water quality, but just don't have the capacity to troubleshoot and experiment, even in the sort of ways that Ting group is doing with with the DIY sensors, where I think that open source has a really important role in this space, but it's not necessarily accessible to.

[00:27:03] You know, the community at large.  So I'm glad to hear you bring that up. And I feel like you're really starting to talk to  you know, folks, folks here have addressed some of these kind of core concerns around water monitoring,  and the transition to real time, you know, how do these sensors exist out in the field and remain functioning, how do these sensors pull the data out, in a reliable rapid way so that it's accessible for folks need to make decisions.

[00:27:33] And then what are the challenges then with managing that data and actually transforming it into useful information.  So for the last piece of the facilitated part of this discussion, I'm wondering if we can start to transition to, you know, let's imagine we're addressing these challenges, right? Let's imagine we figured out how to make sure these solutions are affordable, are reliable, and are delivering data in a way that becomes actionable information.

[00:27:58] What do you see as the future there? You know, what's the potential of water monitoring in that future? And what role really does real time sensing play?  I think with this, let's mix up the response order a little bit and actually come right back to Adam to hear from your perspective.

[00:28:17] What does that future look like, from an in situ perspective?

[00:28:20] Adam Hobson:  It's a great way to be thinking of it. I think one thing is, while we can say there's an ideal state that we're going to get to. Yes, we can keep trying to get there. We never will. There's always going to be new challenges we're going to be facing.

[00:28:34] So it's going to be constant innovation, everything from, as I think you know, Ting was alluding to about, you know, just there's fouling concerns, you know, in a municipal system might be ragging or something like that. That happens. Whereas in, you know, more natural systems, it's just bio fouling that may be happening. There's going to continue to be innovations around that sort of thing.  So I don't think we'll ever actually get to a complete state, but it's going to get a lot better, a lot more affordable, all of that. The other thing is, I think really just different sensing technology.

[00:29:14]  We're going to be able to sense better things with, in a more remote way. You know, we have sensors, you know, we have laboratory instrumentation, laboratory procedures that can detect all, you know, whatever we may be looking at. But to then take that technology and move it from the lab into the field, in a,  you know, an affordable way.

[00:29:34]  And in an easy to use way, stuff is going to be a major, kind of point as we move forward. But ultimately, once we have some of that, I think we're going to see a more, you know, a broader distribution of sensors, sensors in more places, things will be monitored more.

[00:29:56]  And, you know, instead of one instrumentation, one instrument being put on, you know, on a lake, whatever size it might be, you're, you could have 10.  And now look at the variability that may be happening. Maybe you can actually put one at each of the water intakes that someone may have or along the coast.

[00:30:18] Because by the way, this area of, you know, the lake or river or ocean, whatever it might be, or the groundwater, is actually seeing, you know, there's more discharge or something like that to this location, But just two miles away it's different. We need to know that, that's the way we can make better decisions.

[00:30:40] But the other thing I think we're going to see is, we talked about this a little bit before, but we're going to have more data. Any way we look at it, there is going to be a lot of data. I get into the conversation a lot with many of our customers and just kind of concepts that we're thinking of this idea of machine learning.

[00:30:57]  And now we sometimes call it big data. And I think one thing I've learned is that, well, actually in the environmental field, it's not really big data quite yet, but it will be. I really do. I think that thing, once we have more sensors, once we have that out there, you're going to start seeing this idea of, and I'm always hesitant to use the word AI and it's not necessarily that it's just the idea of machine learning of looking at trends, looking at what is really going, and can we use that to start making predictions. 

[00:31:30] So if we start going that route, I think that, you know, real time monitoring its role, that it is going to be expected. It is not an option. It is what is standard on what you have. Yeah, certainly. Can you do it? Can you know how to measure your parameters without it?

[00:31:50] Yeah, of course you can, but I would say it's almost like the auto industry. You know, there's so many safety features now on cars, with, you know, blind spot warnings and that sort of thing. It's just expected. I think  very soon, in some cases we're already there. Real time monitoring is just, it's going to be expected.

[00:32:11] It's going to be, I want that, It’s not,  what, you don't have that? That's not on your system? People will question that. A true water monitoring system will just  have to have it.

[00:32:25] Max Herzog: Absolutely, yeah, thanks for sharing that perspective, Adam.  I'm wondering, Ting, if you, you know, from the utility perspective, have something to add to build on, you know, a different, a different kind of lens to put on this.

[00:32:42] Ting Lu: Yeah definitely.  Max, I love what Adam just talked about. It's very inspiring and actually I want to reinforce it. I do see this continue to be a future technology here. I remember 15 years ago I saw, you know, this sensor is the future technology for watershed management and wastewater and we're still here.

[00:33:04] And then we definitely expanded the adoption of more sensors. And I want to just add on to what Adam talked about on machine learning and artificial intelligence. I do see that there's a future where we can leverage sensor data and real time data that will be able to apply more of these data analysis tools, like a weather forecast, that you can predict flooding conditions along with your historical data.

[00:33:35]  And also the water quality and level data. You can predict the plant influence flow so operators can be more proactively in deciding what to bring another clarifier online. How do you divert flow? How do you change the SRT?  So that's one thing we talk about: we have the years and eyes on the system.

[00:33:56] So now we're adding more smart brains for the system to automate and to manage itself.  The other thing I see with the remote sensing and the real time, this decision making, I think it's the integration of all these different technologies. So the real time, the water quality sensor, in situ sensors are one of the area.

[00:34:19] But this can be integrated with other remote sensing technologies like drones and satellite imagery that there are technologies you can look at on a watershed scale. Obviously, it's not real time, but they can augment each other to look at the stream restoration, vegetation health, and other things together.

[00:34:39] Along with the artificial intelligence and machine learning, you can really predict as well as the pinpoint some of the water quality issues there. And then there are fingerprinting technologies like,  you can use molecular tools to track down, whether it's dog poop or human, or the things that, whether it's for S-S-O, C-S-O, were just better informed investment.

[00:35:04] We all see the pandemic has been here in the wastewater. Based epidemiology has played a huge role to inform public health, and I see it's all these technologies. How do we understand them better? There are limitations, there are applications, so we can apply those at the system level together. And lastly, just, you know, from a utility, you're looking at one region, right, as far as the Tualatin River.

[00:35:34] And then I really,  see how much you have done with the Cleveland and Lake Erie areas. It's always a good tool having this sharing best practices and understand and learning innovations from each other. I think ultimately with the real time water quality information and platform that we can all see the water quality improvement and learn from each other and at the national level.

[00:36:03] Max Herzog: Absolutely. And I think it's really exciting to see the role that Anchor institutions like utilities can have in accelerating the adoption of these technologies and really cementing some of these networks, for information sharing really excited to see, you know, the growth of participation in groups like this one in the water data forum with Interested in learning about these technologies and the growth of groups nationally, internationally, like the like swan, the smart water network, really focused on propagating the adoption of these technologies.

[00:36:39] Thanks, Ting. With that, maybe, Ed, would you like to share your perspective on what you see as the future of water monitoring and, again, the role that real time sensing or near real time sensing may play in that? 

[00:36:55] Ed Verhamme: Yeah, I you know just thinking about a lot of the discussion, there's a few,  I'd say cross sector lessons I'd like to point out.

[00:37:02] I think our water industry, Should be having these data density cost discussions with the healthcare industry, the automotive manufacturing they've wrestled with some of these things already, and they've really thought about, you know, quality when to save data part cost production and I think we've only had a very few examples of that in our water industry timeline. So I think we're always going through this sequence of producing new stuff or changing. At a certain pace. That's maybe a lot faster in other industries. So I think transferring any lessons learned,  you know, another one is I, I'd say like law enforcement, for example, the amount of data streams that they're having to white jet in chest with video recordings, all these laptops connected devices.

[00:37:52] I think they've already wrestled with some of this. So sort of this technologist, I've seen most agencies just have a technologist on staff. And so I think we've seen some water industries do that. I also wanted to just point out some unique, I'd say more community driven,  future for this. And, you know, there's cryptocurrency backed incentives to have more environmental data to reward people for providing network coverage or collecting sensor data.

[00:38:21] There's a Kickstarter campaign to seed it. Thousands and thousands of new devices that are monitoring weather and water. Some have been extremely successful. I think we have this very enterprise view of some of these systems we're talking about. But there's a spectrum all the way down to community science and empowering people with these personal or household neighborhood type devices.

[00:38:45] You know, like Ring Doorbell is an example. It's a huge tool for law enforcement. Where are some equivalent relationships with these consumer products? That we can extract value and have some differentiators of this community data streams that are just constantly being added to and when do they cross over into this enterprise barrier.

[00:39:04] So I, you know, we're going to have to talk about data streams and importance and mining it and value. I certainly like to have that discussion. And I think that the final takeaway is the adoption is going to be driven by leaders, people, and our ability to communicate, and network effectively.

[00:39:29] The only time I've seen this timeframe move faster. Is when a neighbor sees someone do something, they meet, talk about experiences, they plan a similar path, and it goes from one right into the other, you know, and just with our city of Toledo example, getting real time monitoring for the city of Toledo,  with real time data about the lake within less than a year, every water treatment plant on Lake Erie had adopted that same monitoring scheme, that same data management, that same decision.

[00:40:01] And it was a really nice example. So, the more we can do to facilitate that data transfer the better. And again, I'll just say that cross sector one seems so important to me because we're not in we're not in isolated silos dealing with these problems.

[00:40:20] Max Herzog: Absolutely. I really appreciate you sharing that and I think it kind of highlights the need for folks like those in the room here today who are focused on developing these use cases so we can prove the value of this for different types of stakeholders across the scale of our communities, so folks can see how this can work and the value that folks can get out of it.

[00:40:45]  Really appreciate, you know, our panelists sharing their perspectives on these different facets of the conversation about real time monitoring, we have about 15 minutes left. So I'm going to go straight into some questions from our attendees here today. We've got a first question here,  about what to look at when trying to analyze the benefits of moving from diesel and gas powered marine vessels to all electric or hybrid, or negative trends being observed without positive trends. Let me try and dig into this a little bit. Real time high frequency may not be necessary for this. Okay. Yeah. So the question is what parameters, processes or analyses should this attendee be looking at if they're trying to quantify the benefit of moving from

[00:41:41] Diesel or gas powered marine vessels to all electric hybrids or, negative trends being observed without positive changes is real time even necessary for this kind of monitoring, but it's hard, you know, kind of trying to say it's hard to gauge without knowing what's possible. I know this is maybe a little outside of looking at here, but I'm curious if folks have thoughts on this.

[00:42:08]Ed Verhamme: Yeah, Max, I can take that one. I just want to point out, I didn't answer a question, so there's a question hidden under answered if you wanted to see that one too, but I think that one just brings out the, you know, the sort of decision here, and I think there's some questions embedded to that that are more maybe carbon footprint related or sort of some other output.

[00:42:27] I think an easy one to answer is like an oil and water sensor. Like, obviously, if you're making that transition, you're assuming that, so I imagine any gas engine in water is always leaking gas into the water. So it's an easy one to monitor.

[00:42:44] Adam Hobson: I'll just add to that. Yeah, you're looking at, specifically and that's going to be refined oil. And that is, there are sensors out there that can absolutely do that. But I think this is also where there's some knowledge that's involved. We can't, you need to understand exactly what you're looking for. And you know, we get, I know one thing that we see a lot is people say, I just need to look for oil and water.

[00:43:10] What they forget is that there's a lot of different types of oil and there's a lot of different sensing technologies in there and how they actually work. So this is, again, I come back to the idea of knowledge. You do need to have some knowledge and some understanding and an appreciation that there may not be a catchall for what you're looking for.

[00:43:28] You may need to look for surrogates. You may need to understand how things are actually related. You may need a different parameter that may be indicative of the thing that you're actually looking for.  In this case again, yeah, as Ed said, you're looking at a kind of gasoline leaking effect or diesel fuel leaking, but then again, gasoline versus diesel.

[00:43:52] Is a different signature from a chemical standpoint. And the sensing technology can be different. So instead of one sensor, you might need two or three or something like that to actually,  to take a look at that. Whether or not real time data is needed for that. I think that's a, that's a question of, and just understanding the process.

[00:44:13] And I think there were some comments made about this earlier that you do have to, you know, not everything is just gonna be real time. And, we're going to collect real rapid data. Absolutely not. It may not be necessary. You may only need one sample.  That's fine. You may just need a synoptic check of, you know, a bunch of different areas at one time, once a week.

[00:44:34] That's fine. But it depends on the phenomena that you're actually looking for. So it's also understanding that system.

[00:44:42] Max Herzog: Thanks, Adam. Thanks for thoughts on that question. Another one here, in the Q&A data could just be a warning flag, but to use it effectively, you need a lot of it and good quality and this implies some cleaning of the data.  So the question is, do you know of a product that works for most folks in this area? Even SEEQ requires skill sets beyond that of a lot of utilities.  I guess maybe since a Ed and Adam served that, I wonder if Ting has thoughts on this.

[00:45:16] Ting Lu: Yeah, with the large volume of data, it's definitely, it's the next skill gap or technology that you are looking for is data cleaning or data closure.

[00:45:28] So there are ways that you could do some of the statistical analysis or even the analysis that looks into data gaps and data cleaning. The peaks or things that there are definitely tools or analysis you could do for that. There are softwares that is associated with that. So we don't use a kind of a one software to do that.

[00:45:52] It's really depends on the process and the way we use them, whether the data scientists or the tools to do the data cleaning here.

[00:46:05] Ed Verhamme: Max, I'll just add, you know, just trying to clarify, quality is different than talking about accuracy or precision here, and I think people tend to confuse that quality, you know, I tend to think of it in terms of quality of information to make a decision, and sometimes the accuracy and precision very widely needed to make that, so I would just always watch that I've, you know, just thinking about optical sensors for a moment you know, I don't think you should ever have to calibrate an optical sensor.

[00:46:39] Ed Verhamme: You're only verifying how dirty the window is or something. So there's some thoughts towards trends or information that's not tied to this absolute or decimal places of accuracy and it's stable forever. So I think I want to always keep pushing people away from this scientific interpretation of that phenomenon and just think about the decision and quality in that terms.

[00:47:08] Max Herzog:  Yeah, and I'm trying to keep up with the chat. There's great conversation going on here and just want to mention that someone did chime in that they've been using the tools in Aquarius for cleaning data. And that they're actually writing an SOP to standardize it currently.

[00:47:21] So that may be a good resource to look at as well. 

[00:47:27] Adam Hobson: Yeah, if you can afford Aquarius

[00:47:32] Adam Hobson: I will put a little plug in there. There are lots of, there was, I think data cleaning is a very important thing. And I think also just again, as Ed's pointed out, what's the purpose of your data?  Why do you have that?

[00:47:43] Why are you going? Why do you need your data? What accuracy, what precision do you actually need to make a decision? And that's very important to know. And so why, how well do you need to clean it? What is good data? What's quality data?  You have to kind of answer that. And there's a lot of tools out there, that again, maybe overkill for what you actually need, or maybe you don't need to actually, you know, write some crazy computer code to search all these massive records or whatever.

[00:48:14]  But there are other advantages here that we can look at where, it may not just be good data. Again, it's a lot of this is also looking for trends that may indicate that there, your data is off. And a lot of that can be built into, you know, honestly, some of the sensors and that telemetry type systems that are actually just kind of backend type user stuff that may give you a warning that says, Hey, I'm sending my data via telemetry, but as it's coming in, I'm already starting to notice that you're seeing a trend one way or another. Maybe, you know, what can you do? That's where a lot of this machine learning concept really comes in.

[00:48:56]  And the question though, is whether that's who takes on that responsibility.  Is that done by the manufacturer, the provider of the software, or is that a customized thing that you, that the user ultimately wants to have? To be able to turn every single dial, to make that decision. That's a question that you'd have out there.

[00:49:15]  There's a lot of good tools out there, again, I'll  put a little plug in here. I'd be remiss if I didn't, but, in situ does offer a great data visualization and data management program for that.  Again, it won't meet everybody's needs, but it's going to meet a lot of people's needs.

[00:49:34] And so I encourage you to check that out. It's called hydro view. It's part of our, which goes in our telemetry and remote data, remote management system, you can take a look at it on our website. I'll put my contact info in the chat. Also, if people want to reach out and have questions on it.

[00:49:50] Max Herzog: Thanks for sharing that Adam. One more question here in the chat that I think touches on this a bit. Great advances have been made in data driven methods and machine learning that can complement Water quality monitoring. And as folks in the panel have talked about. So these methods, you know, from this person's perspective can potentially reduce the number of sensors and the frequency of measurements we need to get a full picture of the system.

[00:50:15] So the question is, do you all think we're going to rely on more sensors, as the sort of data driven methods and machine learning tools advance, or are we actually going to need fewer sensors, with the help of these sorts of methods. I'm wondering if one or two folks have thoughts to share on that before we wrap up here.

[00:50:37] Adam Hobson: I'll jump in real quick. I think we're going to see more. I think you need to have more sensors to understand that phenomenon. Now, I think it's the way the system has to work because the system is always going to be changing. So you need to understand that it needs to continue to adapt.

[00:50:58] So if you want a good forecast model, by using machine learning, you're going to need more sensors. And by more I mean, I'm thinking by the geographic diverse or distribution. If you want to have a better sense of what's going on, or spatial distribution, I should say, because that could be horizontal or vertical, depending on what you're looking at.

[00:51:24]  Yeah, my sense is we're going to see more. I know in the future, could you have less because we've already established our model? You could argue that. However, I think we've already seen through history that, you know, we thought we had all the stream flow data for the U S all figured out. We made all these models and assign water rights all based on them. And guess what? It changed. And that created a big problem.  so I think we've learned now that you got to keep monitoring to monitor actually for the change in the system.

[00:51:54] Max Herzog: Yeah. Ting, do you have thoughts to share on this? 

[00:51:58] Ting Lu: Yeah, I'm thinking about Adam's response here. I agree with your one way design of the sensor network. You need to have the key strategic locations that have real time data. When you build this machine learning model and algorithm and the how the machine learning is going to take a while to train the model and going to continue to improve.

[00:52:21] So you need the data, but I don't know whether we continue to need to expand the sensor network or maybe some of the locations we could move things around because that's the beauty of the sensor stations. You don't have to have it in one place. So you can move things around and to validate the model once your model is established. So that's my take on here.

[00:52:43] Max Herzog: Absolutely.  We have time, I think, to just take one last question here.  We have one here. When using real time sensors in the Great Lakes, how can a local agency reduce costs for purchasing sensors when there are many interferences,  you know, or contaminants of concern? You know, they list CDOM, DOC, nutrients, chlorophyll, but each needs their own sensor to collect the data of interest.  So do folks have thoughts on that? Maybe we can start with Ed, if you have thoughts. Yeah, I think the easy

[00:53:16] Ed Verhamme:  Yeah, I think the easy answer is to collect all the needs of that community and approach either the manufacturer or an integrator with that need and help them figure that out.

[00:53:30] There's a lot of these cost things are solved with volume and pulling apart the specification and accuracy. So, you know, if in that example, what we really want to monitor is blue green algae, that's the one thing we spent the most money on and then let's just look for some value sensors for the other parameters that are going to help but aren't the primary.

[00:53:50] So I think it's really that specification process and this is extremely routine. Every car, every piece of like manufacturing piece goes through a very sequenced process and evaluates a lot of these. And we rarely do that,  just a multi-step process to go final to a purchase decision. We often just jump to procurement,  I need to buy five of those.

[00:54:15] Max Herzog: Absolutely. Oh yeah, Adam. Yeah, 

[00:54:18] Adam Hobson: I would say Ed makes a really good point. I think, in those conversations, You need to engage the manufacturer, the provider, don't assume that you or others may have the answer, you know, you're looking at sensors, you're trying to figure out where it's going to make sure you engage those, the manufacturers of that. How can I do this in a better way? And if they can't help you, look around.

[00:54:53] Max Herzog: Absolutely.  Well with that, we're getting towards the end of our time here. So I want to really thank our panelists. Thank you, Ting, Ed and Adam for sharing your perspectives, and really your expertise on this set of issues around real time monitoring, I feel like you've really helped us map out a lot of what's possible, what's probable and what the challenges are in this space right now. And I think this is really going to continue to be a hot conversation in the coming years, you know, across the swath of the water sector, whether we're talking about, you know, utilities or surface or groundwater management.

[00:55:35]  So I really want to thank you again for taking the time today.  I do want to close on a quick plug for the next session, in the water data forum series.  We will be reconvening this kind of format of a panel discussion with some experts, in May, and we'll be focusing on cyber and water.

[00:55:57] The question of driving digital security across the water sector. So really thinking about, you know, as we have these sensors and in some cases, maybe even connecting these sensors to controls and, you know, really building out these cyber physical systems.  How do we secure those? You know, we've certainly been seeing a lot of concern around cyber attack and seeing it specifically in the water sector. So how do we start to address those sets of concerns in the water data sphere? So hope to see a lot of you back for that session, as well as the sessions that will be convening throughout the rest of the year, and we'll be doing that session in May.

[00:56:41] So with that, I'd like to thank our panelists one last time, and also, you know, thank all of you in the audience for taking the time over this lunch break, at least here on the East Coast time slot. And hope you enjoyed the conversation and please do feel free to follow up with the panelists that drop there, their contact information if you have any additional questions or reach out to us at Cleveland Water Alliance if you have some interest in continuing this conversation.

[00:57:14] So with that, I'll thank folks one last time and bid you all a good rest of your day. Have a good one, y'all. 

[00:57:22] Ting Lu: Thanks, everyone. Thanks Max

[00:57:24] Adam Hobson: Thanks, Max. Thanks, everyone for joining.

[00:57:26] Ed Verhamme: Thank you. Bye, Max. Bye now.

‍