 Okay, very good. Well, thanks very much for having me on. Pleasure to be here talking to you all. So I'm going to talk a bit, as Liz has just introduced, about sensors for water. At the moment, the company is selling sensors for nutrients and also pH. But through the NOC and the University of Southampton, we've developed a very large number of different water sensors. So if there's something particular that I don't talk about today that you want to measure do get in touch. That's one of the most enjoyable bits is talking to stakeholders who want to measure something that currently can't be measured well and solving that problem for them. So I'm a dual roles. So I'm now a PI at the Oceanography Center in in Southampton. I'm an honorary professor at the University of Southampton, and I'm also the CTO and director of a new company clearwater sensors limited which is where you see all the branding for this presentation. So it's very exciting, having been in academia for a while, and then going into into industry really interesting lots of tales to tell so if anybody's fancying having a go at that. And please also do get in contact. Hopefully this is a bit of a success story for the NERC. So the idea for doing work in this area first happened around 2002 when I joined the NOC after my, after my PhD, and NERC and NOC did a lot of pump priming around that we did a few early small grants for a few years and really proved the concept I guess the first I was concerned wouldn't work was putting micro sensors and micro fluidics into natural systems into natural waters. I thought things would grow on them and they'd never work but actually that hasn't proven to be true right from the beginning even without very much firefowling protection. So with with some initial results under our Bell EPS RC and NERC work together and co funded the ruggedized micro systems technology grant back in 2000 and well grant was written in 2005 I think that 2007 we restarted in earnest. And that was 2.2 million quite a big investment 70% 30% sorry 70% EPS RC 30% NERC. And there's been a number of big programs including oceans 2025 that followed on since. And we've been getting good results and, and the whole thing has grown really nicely started looking at commercialization when we had functional systems in 2011. And I think that's an idea about some of the challenges associated with commercialization the startup with the license was operational 2019 actually we started selling first units last year, at the end of last year. We've been not milk have funded this the tune of around 10 million pounds. There's been a lot of other funding actually the European Union's been really important, especially because they typically fund a high technology readiness levels. And in that pre commercialization speech where you've got some piece where you've got something working, but it needs to be taken through still getting European Union funding at the moment. And I think we'll have to think about what's going to happen. If, if that doesn't continue, but at the moment it certainly seems to be stable and we've got a few more grants. We've also had EPS RC funding as I said, a couple of grants the ruggedized micro systems technology grant and the mission grant. We've also worked with BBS RC on things like detecting pathogens for shell fisheries harmful alcohol blooms and the like. So, you know, there's a lot of money been spent a lot of people have worked on this for many years what's the results where we have commercialized so we have a list of about 20 different measurement technologies that can be deployed from environments including subglacial lakes all the way up to the energetic surface ocean. What's commercialized is this technology which is lab on a chip or micro fluid technology to produce analyzers that use reagents to analyze water. That's important because it gives very high quality measurements very similar to the laboratory high quality laboratory measurements that people run on samples. And this technology is quite mature because of the length of time and the number of people have worked on this we've done over 200 deployments learned from over 200 deployments has been a lot of interest development. 90 journal publications, some of these devices have been out for over a year, for example one under the under the ice in the Arctic. It's very robust 6000 meter depth rated, which is, you know, world best. They also have world leading metrology performance into, in terms of sensitivity and accuracy. And they've been proven all over the, all over the globe. So why are we doing this well, many of you will know but water chemistry is is really central to understanding many aspects of our natural environment in mitigating and managing climate change and environment environmental degradation but also for issues such as water supply and food generation and food industry. So there's, there's a number of drivers why this information is needed. And I think that's perhaps a space where I am and as a technologist keen to work with the note community to really understand how we can make best use of that and make impact in these these various application areas. And why did we focus on nutrients and the carbonate system to start with well it's you know the nutrients and carbonate system is central to most of the state of the art by geochemical and ecological system models in in oceans and in catchments and river systems. And they were poorly measured. There's a lot of variability in in those parameters and doing a grab sample once a week doesn't really catch that you know this title and storm driven variations. And of course the biology comes to play so there's massive variation as biology comes to play. And that's where we focused our efforts because there was a technology gap and they were very important, really the nutrients carbon system drives all of this activity in this environment. As it happens measuring those things then enables you to support a number of other applications including, you know, the sustainable development goals in various tiers because you know directly you can see how measuring water chemistry feeds into clean water and life below water, but you know through supporting another bunch of sustainable development goals you can see how you also have impacts on good health, gender equality and work and economic growth for example. So that's, that's become apparent wasn't our initial driver but that's that's certainly where we've, where we've landed. In terms of the ocean spaces where we started. You can see that the international community is also agreed that these are are important so this is in order of priority and this is in terms of what the ocean community really wants to measure oxygens right the top and then it's nutrients and in organic carbon and and that's exactly why we started measuring in in that order. I've listened and greatly enjoyed some of the other webinars that have been given by the the construction digital environment series and I was really struck actually by a couple. One by Prof Savage on the law of averages, exactly what we see, you know, a lot of parameterizations and understanding is based on averages. And that isn't very good because there's a lot of variability in the system so that means that most of the time actually those simple models based on averages are actually wrong and they need more data to be able to constrain the variability in the non linearity in the systems and that's where measurements have value. It's also a non linear system the cost functions and non linear so you know things like fines for regulatory exceeding is a step function or a delta direct function. You can stock health in agriculture is the same deal you know if you can, everything looks fine and then you have a harmful alcohol blue, and you can lose a lot of stock and a lot of value, I mean a single single incidents can be up in the sort of hundreds of millions of dollars for some some companies, and so you know the potential benefits of understanding water quality is an economic driver as well as an environmental driver. So the dimensions equally a costly so you know cessation of activity or doing remediation is also very expensive, which means it's, it's often worth making the measurements. However, people don't make measurements because of the logistical and challenges with current current systems and that's hopefully where our technology can come to play and we've done this a couple of times so we've looked at both the decreasing the cost of getting your data. By investing in in some technology you can decrease the cost of getting the data you're already getting so we've looked at this with people like see age and the environment agency. But also it can give you a lot more value because you get more better data than you would get by by doing something alone or not measuring and relying on proxies or inferred models. I think that's the space where we're very keen to talk to people to try and develop those numbers and and some of the theory around that but certainly our initial work certainly shows that it's, it's worth measuring and the data is, is valuable in that space. So current methods in a number of environments deep sea to rivers to tanks, you know, we're all taking manual samples, which is is okay. But it has relatively low temporal resolution resolution and actually when you work out the per sample cost is expensive compared to other methods. And because it requires lab analysis you can have storage and preservation problems. There are some water samplers which give you, you know, slightly higher resolution. But again it still requires lab analysis and preservation can still be an issue. Online analyzers have really made an impact, particularly in things like the water industry, and they're high resolution that they've generally like sensors, but the infrastructure required means that you've got to have significant investment. You need to build a shared and supply mains power and typically ethnic, and somebody needs to go and maintain, maintain them, and that and moving those around a catchment for examples really prohibitive. And so you only get data from a very few locations in contrast, you know, having an in situ sense you can be a very high resolution data. And because you haven't got the infrastructure you can move that around the catchment and you can have many more of those systems, turning out, turning out the data. Okay, so that's why we've gone into this technology so the technology is in situ so it can, it can go into the environment submerged based on chemical analyzers based on microphones you get really frequent measurements with lab grade or better as you'll measure it measurements which really improves the data and your control compliance understanding in the environment. They've got a long lifetime and they're automated which means that the actual cost per measurement comes down dramatically and we see at least an order of magnitude sometimes two or three orders of magnitude, savings in cost per sample. They're very robust, they can they can go in a number of environments love, as I said we've done this a lot right we've learned a lot from over 200 deployments is some figures of performance of merit. I'm not going to go through these in detail but you know it's it's low, low energy, so you can have you know several thousands measurements from a battery. The limits of detection and precision are actually better than you'll get in a standard scalar or lab grade analyzer for these for these parameters. So it's, you know, the results are very good. One of the ways in which we do that is we have on board standards so you will have a preserved standard on or to on board plus a blank and that enables in situ calibration gives you really good. Really good results there's a user swappable reagent canister which means that they're relatively easy to use you don't need an animal chemist to turn these things around and recharge them like plugging in a new printer cartridge. Typically 2000 measurements per canister but we can do bigger ones if people want to go longer. It has quite an onboard quite a sophisticated onboard microcontroller, which means that you know the the digital part of the digital can interact with this very well there's an RS 232 interface, which you know which we plugged into modems which enables this to go wireless and get connected to to the internet. With a with a gsm modem or a really remote them and we've got solutions for about data that into flow either into data centers or to the cloud or to organizations bespoke data systems with all the good things like labeling data, traceability of parts and of calibration so that that data can can be of the best standard. A little bit of under the bonnet, because I know a lot of people like to look under the bonnet of things, what's inside the sensors well this this is a kind of schematic here on the left of of what's inside the device. And broadly there's a network of channels connected to the reagents the environment and the standards and blanks and some syringe pumps. So actually it's a very high precision syringe pump so to move, you know, a few microliters of fluid around a piece of plastic, you actually need a very precise and coordinated set of pumps and so we do that with this. We have a syringe pump that has many barrels all connected to a single plate, which means they all drive exactly the same rate and you don't get any differential dilution problems. And then we have optical cells so we can measure for essence luminescence or absorbance but actually all of the ones that commercially available at the moment measure absorbance, and that is done with with optics. I'm going to bring up to another slide in a minute to show that a bit more in detail, but the electron you can see the structure of the device on the, on the right hand side you can see the little on the bottom there some channels they're cutting a piece of plastic that's what it looks like before we stick the lid on in the middle is how we polish up those channels to get them smooth with with some solvent polishing. So that the everything operates as it well it should nothing gets stuck in it the optical performance is very good. And then you can see the electronics in the top right there. The chip is the black piece of plastics actually on a stand in that image in the top right. The chip with the fluidics in it is actually at the base of that unit and the electronics are stuck onto the top of it and the vows are stuck on the top and the pump stuck on the top. And that forms the inside of a pressure case. So the end plate of the pressure case is actually the micro fluid chip and inside the pressure cases will be electronics and what have you we protect that from the water by filling it with oil. And that means that the pressure communicate into that side of the of the device and that's where we get a very high pressure tolerance it does mean all of our electronics are running at ambient pressure. Okay, so under the bonnet a bit more. The one of the key things is the optics so we have a patent describing how we use tinted or dark materials to suppress the stray and scattered like you'd otherwise get measuring absorbance in a micro channel in a clear plastic you can see that in the top image is like bouncing around or all over the place. And I think I can do this. Can you see my mouse if I do that. Yeah, okay so you can see in the top image there is there's there's lights all over this is where the detector would sit and there's light all over that and that's come from the LED bounced around the device. And there's a very large illumination spot and most of that hasn't been through the fluid and that's what a lot of systems do and they just measure the try and blank off some of that and then measure as best they can the light that's been through the liquid. If you make it in a darker or tinted material you get this lovely little tiny spots of light that's only gone through the liquid. And that's what really enables us to make some really high performance measurements. And the nice thing about that is there's no optics, you know no lenses no expensive bits involved and we can make multiple measurement cells of different lengths to give you very high dynamic range. So actually our nitrate analyzer has got the highest dynamic range in the business we can measure up to the millimolar with a very short channel and on the same device we've got some long channels that maybe get that very very low. And if people want to go lower, we can make longer channels. So do talk to us if that's something you're interested in. Okay. So we spent a lot of time on some of the value that you're getting is around assay optimization, really important to say that you can't just apply assays off the bench in an in situ device and so we've done a lot of work on that looking at things like interferences the saliency effects. I'm not going to go through all of the detail on that slide but it's been very important to fit the assay to the device and to the application and to the sample that you're working with. And that's one of the ways in which we get very good performance. There are some you know physical limitations that you're dealing with things like the interaction with surfaces and carry over in pumps fluidics. I think we've been through that enough times now that we've minimized the effects of those, those things. In terms of the device optimization. Again, awful lot of work learned from so many different deployments with colleagues some here present and others in the community. So looking at things what breaks when you when you try this, the electronics are really important and actually we've got very, very good reliability now in our electronics can survive these temperature and pressure extremes and very robust subsystems like pumps and the micro fluid chip itself has to be very robust has be very what the lid has to be very well bonded on for it to be able to withstand the rigors of being deployed saying the surface of the southern ocean. Okay, and you know methods for making sure that's all happened. We've we've got a very good quality management system both at the NOC and now in the company has been 17 years of you know testing and problem solving and 200 deployments and that perhaps sets us apart from some of the other new entrance into the market you know it's just the number of times that this has been tested as meant has been an awful lot of improvements. I've been lucky enough to have a large team of engineers working hand in glove with analytical chemists that's been really important as has been having really good engagement from environmental scientists and stakeholders throughout this development has been really important and you know it's been good everybody's been really motivated to do this work and you know the teams have been fantastic. Okay, so getting to some of the sensors and some of the deployments. So this is my was PhD student then postdoc and now is in the company Alex beaten trekking around with a an early version of the nutrient sensors sensors on his back and believe this is in Peru. Here you can see a device and it's deployed with a solar panel and I think that's a Campbell scientific logo and a and a modem. I think that's in a glacial setting here's some work with Bristol and with with Liz in the past, it's like sure. I'm looking at nutrient nitrate levels in the couple of glacial systems and on the right here was the work that we did in the macro nutrient macro nutrient cycles program. Measuring chemistry in the even catchments and here you know there was a battery a solar panel and a YSI storm logger which meant that the data was coming off live of nutrients every half an hour. On to the internet and straight into then data sensors so you know this. This is the possibility that you can have low infrastructure, high quality data or water chemistry streaming straight into your desktop with this type of apparatus. Here's some some actual data, so you can see here that we get these this is a test deployment in right outside the NOC in off the pontoon this is this is phosphate is two months of data here, and this has been published by Max grand. And you can see these are these are discrete measurements matched to in red matched to the gray sensor measurements from the from the lab on chip. And you can see there's a great deal of variability that is not captured by the discrete samples and actually there's quite a lot of there is actually quite a lot of noise and uncertainty with the discrete samples itself so very important. proof that the is worth deploying a sensor and you can see you can start to see some of the trends so you can see. You know association with this drop in salinity and increase in oxygen with an increase in phosphate and so it was a rainfall event. And so you start to see the natural system responding in high frequency and you can see how an average value here would not have much. Again here just showing how phosphate here in in blue is moving up and down with the title signal so again making a know if you don't get your sample acquisition time. To the right point in the title cycle you're going to get a very different answer and start inferring things are just on on true. And you can see here on the right looking at the title variations that you also get these modal shifts so if you were to to do some modeled and this is very common right to do some models correlation between either salinity or title height and the nutrients, which is what most people certainly a lot of people do those. Those relationships change over time because of things like spring blooms different source waters different inputs. And so you have to be careful. Again another reason why it's important to make make measurements and reduce that uncertainty. So we also done some integration with vehicles, it's really lovely way to get data straight into the digital system most of these vehicles will contain a modem of some sort this is a Kongsberg sea glider that we've we've already demonstrated in fact we've had nitrate, phosphate and pH sensors in the payload by the back of the sea glider making measurements for extended times. It's fairly low power requirement but because we are low power, we're able to do this for multi month deployments, and we've done this over 10 sides, 10 times there's a couple of papers is a Vincent paper in 2018 and just how it's a virtual paper anti virtual just release to paper with with me on as well. Looking at phosphate. So yeah there's those those two papers. And the nice thing about getting a nutrients on to a glider is you can make these measurements and redirect from from shore. So you know desktop sciences, it's really there. And this is really the only way you can currently measure phosphate and citricate on a glider, and actually the sensors have gone and can go on things like. surface vehicles or moorings or whatever you whatever you want really it's only, but this is the only way you can make phosphate and citricate measurements on a glider currently. Here's some of the data so you can see here changes over this is time along the on this access and then this is depth, and you can really start to see the draw down of nitrate in the surface waters as the spring room develops there. This is in April 2015. Yes, this is one big, we're also integrating these sensors currently on to see Explorer from our smart and on to the auto north UK developed and produced autonomous surface vehicle. Okay, so I'm going to round off with, you know what was the opportunities for partnership, but this fantastic technology out now and available. But what could we do together. Well, I think I'm really interested in what I've recently learned from the webinars here and take that to the next stage looking at the expected value of information in various applications. And thinking about where these, where these measurements really do add value to understanding and decision making. We're getting a lot of interest in in science and in in other applications. We are getting a lot of interest from aquaculture, you know, whether that's fish health, proprietary, you know, regulatory compliance or looking at harmful alcohol bloom prediction, which as we said has a really major effect on the economy of our culture. We're looking at processes and water quality. We haven't had so much yet from agriculture. We're just starting to see that now looking at efficiency and environmental stewardship. So really keen to look at applications. So anybody who's interested in that please, please let us know. As I said at the beginning in my research roles and a little bit in the company would be very keen to try and understand if there are new sensors or assays that need to be developed. If there's something else that you want to measure in water, water chemistry particularly. And then how can we get these systems better integrated with existing or future sensing networks and data systems really on on theme, hopefully from the constructing digital environment call. And also integration with models. It tends to be a little bit of a has been historically and perhaps less so now that gap between the modeling community and the observationists and the operational data gatherers and we see a lot of value in making sure that the models inform the observation strategy and that the data helps constrain uncertainty in models to it the best of its ability and so then it does need to be that dialogue between the observationist technologies and modulus going forward. Okay, that is my, my last slide. So, thank you very much for listening and I'll be very happy to take any questions that you might have. Fantastic, thank you very much Matt that was a really great overview and yeah some some questions coming in in the chat now. So I'm going to start with a few from from Steven and he'd like to know a little bit more detail about the biofouling, and if you could comment on the battery life when you've got an install running with one of the wireless modems. Okay. So by fouling. We do see by fouling. And, but what we do is we all we do is we place a filter at the front end of the sample inlet. So we generally allow the sensors to foul a little bit. We might put some, if that's a problem for the vehicle or the, for the way in which it's being used for other systems we might, we might put some anti fouling coat coatings or copper one to the device but for the instrument itself really doesn't care if it's unless the fouling is so heavy actually perturbs the chemical environment and all we do is we stick a typically a point 45 micron syringe filter on the front end because this is micro fluidics the actual volume of sample that we're pulling in is very small and you don't need a large capacity filter to just pull out the stuff that would otherwise go in and clog the samples. There is a couple of exceptions where we have seen that the existence of a thick biofilm actually perturbs the chemistry. So we did see that when we had a clear filter holder in the surface ocean and obviously some respiration was happening on it perturbed you could quite clearly see that in the value of the pH that we were measuring. But by having a dark filter case and managing that a bit better we've managed to, I think eliminate that kind of problem we've seen it also once in nitrate measurements and again it's just about where you place that filter in your system and how much sample you draw through to flush it and manage those background effects but usually there's enough flow in the ocean that actually the biofilm doesn't cause a chemical participation and that point four micron point four five micron typically filter. They can they can handle filtering you know many many liters of water, and typically we're taking in a total of two liters over a you know several thousands measurements so two to five litres. So the capacities are about right. We did see it once where the filter got physically blocked. And that was in a very very turbid river in the US when we were doing the nutrient challenge. The paper coming out actually in front is that goes through some of these some of these issues and in deployments that co what they said industry partnership paper with me both my knockouts clear water roles. Battery life. So it really depends on the system so there are some very good micro power modems out there. We tend to deploy probably a bit over eight with a fairly a large battery in in things like catchments because it's, it's cheap and you can so we would use a small automotive battery and a small know half meter squared or less solar cell. So the panel. And that's more than enough, even in winter to to manage running things like a storm logger and and sensors and actually hung up other sensors. I don't have the power consumption figures for the storm logger and all the rest of the tip of my tongue but you know when operating the sensors are typically drawing somewhere around one to two watts depending on what they're doing and you can put them in the sleep mode where it's significantly less than that so if you're measuring, you know, for 10 minutes every hour getting hourly measurements and catchment you can really account even relatively small. Thank you. Another question from Steven which I think also ties in one that that Michael Stankers put in the in the chat. Can you talk about the in situ calibration process so Michael's asked if the filter itself would affect the water chemistry by stuff going on the filter and you've talked about that one example where we have that. But I wonder if the the in situ calibration would take into account of that and can you tell us a little bit more about the in situ calibration process. The in situ calibration is really for the instrument itself and we don't currently supply the caliber and the front end of the filter the filter is in the environment. And we could do that. But like I say that the fouling and that issue really hasn't proven to be the case in the vast majority of the environments unless we do something wrong. And dealing with that filter effect is it is really about good practice in where the filters placed and all of those kind of things the calibration happens immediately after the filter so we're carrying on board standard standards and the blank. And we can also do things like reagent blank so you can just pump reagent through the obstacle so the nice thing about the way that the system set up you've got active valves. You can just route through to different areas of the pumps you've got quite a lot of control in terms of how you program the device and which fluids get measured wet. But typically we will, we will measure a blank, which just gives us an idea of how much light is coming through our obstacle cell that will then run through with a reagent blanks as the same thing with the reagent in that gives you an idea with your reagents and then we will measure a usually a high standard or a high and low standards to give a full calibration in situ. We've got we've got valves within the device and they're literally just switching the inlet to our analytical system between the environment, a blank or one or one or one or more standards and those standards. We will analyze before the deployment typically and we recommend that they're analyzed after the deployment to see if they shift. But we do include in there, you know mild fire sites to make them last as long as possible and we've had the last, you know, over a year, depending on which standard it is. And so typically we'll put a little bit of chloroform actually into things like the nutrient standards is all most of this is described in the papers that come out of the NOC. The one that doesn't need the standard is the pH because the diet self is self calibrated effectively once once you've stable enough that once you've determined the constants for that diet you don't need a caliber so we don't currently carry any calibration on board and yet we get very very good long term accuracy because the diet is just so stable. Thank you. So, a question, a couple more technical questions then we'll zoom out a little bit more. And so Matt Fry asks, he's very interested in the expected value of the information that you can capture. Is this something you've considered in rivers for example how many sensors are required to capture and or enable prediction of dynamics across the system is the age old question in the digital environment sphere and wonder if you could comment on that. Yeah, so I utterly agree with Doug Hubbard wasn't it, who was who gave the talk on EVI when he explained the curves and you get a lot of value early on, you know, measuring something is so much better than the not measuring. If you're already measuring then it's about constraining the uncertainty. And that's going to vary in catchments I would have thought, you know, you get a lot of value by by instrumenting key reaches and junctions. And then once you've done that, then I suspect it's likely diminishing returns but you will continue to get value spotting things like unexpected sources, and potentially sink so those things tend not to be such a big deal like in catchment systems. So in some areas groundwater is important so we have done some work looking at putting devices into boreholes or on the end of pumps analyzing groundwater. But it's, I'm afraid it's going to be a bit of it depends on so because it depends on the characteristics of the catchment and the natural system. So you're looking at and whether really nutrients are the dominant factor and whether measuring nutrients gives you a better handle on that system. So in some river systems nitrate is already high and not limiting, and it's all about phosphate and so you know the expected value of nitrogen measurement might not be so high, but phosphate might be the thing in other areas, it's nitrogen that's high and is close to the regulatory limit, you've got to be really really on the ball on on nitrogen so yeah, it's going to depend and I think that's why it's so interesting to use that expected value of information framework to actually try and put a number on the value of the information you're providing and to tailor that that deployment strategy to get the data that your want. Thank you. And that leads us nicely on to to question from from Edward Darling of a red list revival and he's very interested in the technicalities but he'd like to ask about the thoughts on reaching stakeholders who are causing issues for climate marine life, life on land and clean water so how do you relate the measurements that you have to the sustainable development goals. Yeah, I think our tactic so far is to be to work with with really big players who are either able to solve problems or to to change their mode of operation to cause less problems. So I think it is challenging because there isn't always a financial incentive for them to do so. But increasingly, even some of the most resistant organizations have corporate responsibility policies and the like and that can be helpful, but actually the best, the best comes from if you can save the money. So, I'm very, I'm very hopeful that subsequent to what's been happening in the UK around water quality maybe there'll be a higher level of testing and enforcement, which I think will be useful in achieving change. But for some people it's, it's, it's a, it's a money issue, you know, for example, optimizing agriculture. There's an awful lot of loss and money loss if you do it the wrong way without, without being able to measure. And so I think, yeah, the expected value of that information is going to be high when you have, you know, potential catastrophic stock losses finds for damaging the environment and the ability to not put very many fishing upon that could perhaps take more if you if you knew what the water quality issues actually were rather than what you predicted. Then regulations really important so it's been interesting to see the regulation coming in around building a nitrogen. So, you know, planning permission is now tied to a very crude paper based model of how much nitrogen a horse or sheep person and a bath and a shower generates. And you have to then mitigate that again with a paper based model, or based on averages to whether that's going to actually in your in your environment and in the location of that development solve the nitrogen problem. So I think that's important regulation that's come in but I think there's an opportunity there to do so much better in terms of seeing if those paper based models are correct and whether it is actually resulting in, in managing nitrogen effectively in the catchments. So it's going to be very specific to each, each area, I think, but our tactic has been to really to go for those organizations that have the biggest impact one way or another. That leads us quite nice to a question from from Ron, and he'd like to hear your thoughts on sensor and model integration so you know can we can we predict conditions based on the data that you can capture. Yeah, so I think models are getting much better at analyzing where the uncertainty is in their, their own predictions. And that's a natural place for us to target a measurement campaign. But I also think that data gathering in areas of high variability is also a fairly obvious place to look to try and improve and constrain models. So what is really nice is if the models can inform where, where and when are the best places to measure to focus that you know you always going to have a limited measurement effort. And whether that's a Aussie or, you know, simulation experiment where you look at different measurement strategies or where you just, you just look at where the error or variability is greatest. I think both can be all of those things can be can be valid. I think it's to be honest though for us it's still quite an early stages we're not having, we're not having many of those good conversations with models trying to constrain that error so anybody wants to get in touch about that great. Can you speak a little bit in the talk about how you're making sure that the data that goes into online streams is, you know, conforms to fair protocols and so on. Can you talk a little bit more about how you've approached capturing and integrating data into kind of cloud platforms and you know if you've got views on on the best way to do that to ensure the data is usable to the widest possible set of users. So for sort of open access data we still really go through the data centers, and we're working with them to, to make sure that the sensors are reporting all the correct metadata so that that comes through and we've been working for example with the EDC on using things like unique identifier so every sensor has a unique identifier and then, you know we're in systems where actually communication costs quite a lot so we only want to send a very small unique identifier over the why say I'm here and this is who I am that's what the sensor will tell the data system and then the data system can go alright, and they can look up on, you know, a web based or a data center held repository, and then associate those measurements with the last good calibration, where that is for what it is what it can measure what these interferences are and all of that kind of good stuff. So that's been our tactic to date. I mean it's the Wild West in cloud domain. There are a few, you know, providers that are trying to come up with similar standards. There's a lot of commercially commercially driven there's a lot of commercial offers there in in cloud based computing and so I think, again, we are in learning mode as much as other people are but there are, you know some customers have already set up their own cloud based repositories that they mandate that they want us to use. But in science, it tends to be data data driven into the data sensors at the moment, but we all know that the data sensors are on a journey of discovery like the rest of us into the cloud computing as well. That's very positive to hear that. Yeah, you're working directly with the data centers. That's, yeah, I think making making the most of the resources that we already have is quite important for for all of us in sensor design. I mean, the nightmare for them is that we, you know, we generate, you know, large data sets and they're going to be doing this ongoing for the future and that there has to be some manual step and personnel involved. So they're they're really working with us hard to try and make it automated ingestion of data. So all of the data flows naturally that the metadata flows naturally and that's that sort of manual work and personal involvement is is limited to setting up the system well rather than actually manually typing in numbers. That's good to hear. Anyone interested in that side of things do come along to our next webinar where Scott and Shannon are going to be talking about what they've done from a kind of whole system design perspective. They're in the US so so don't have access to our data centers, but that will be that'll be really interesting if you're interested in that sort of things. And a couple of kind of zoom out questions to to finish with, and wondering if you could comment first on the challenges of working in a truly cross disciplinary team, you know, you've talked about how you have analytical chemist you have engineers you've worked from everyone from glaciologists to, you know, autosub people like there's there's a lot of different languages that you've had to learn if you could comment on that. And then also if you had any nuggets of advice for anyone moving from academia to industry and in the spin out realm, if there's anything that you've learned that you could share. I'll go for the nuggets first is all hard work. There's no easy ticket right it's it's it's all really hard work and do your homework and and be prepared to put put the time in I think is. It's probably my overall but it's a lot of fun I mean it's great to see things grow and people to grow with it and take on the responsibility that you once had fantastic results to come out and I have been very lucky to work with some really fantastic people. I think the key one of the things though is to try and give people responsibility so they can, they can develop and and have ownership of these things and hopefully my team would say that's what what we try and do. Working with lots of different people, I just really love it. I think it's, you know, it's one of the most interesting things about the role and I quite like it when they start talking to different language that I can't understand and to do a bit of homework and I think it's understanding that there are those differences and there's no better or worse. But I quite like the common language of mathematics and of uncertainty and of, you know, science, which despite the different disciplines, we all share so I think it sometimes it's great to go back to fundamentals with people. But you know, I'm very, despite my interest in chemistry and analytics and science I'm quite person focused I think that's really important in dealing with both large interdisciplinary multidisciplinary teams. And we're talking to stakeholders you've got to see what's motivating people from multiple aspects.