REGENESIS Technical Perspectives Series Part 3 – Use of Modelling Beyond Design

Hello and welcome everyone. My name is Dane Menke. I am the digital marketing manager here at Regenesis and LandScience. Before we get started, I have just a few administrative items to cover. Since we’re trying to keep this under an hour, today’s presentation will be conducted with the audience audio settings on mute. This will minimize unwanted background noise from the large number of participants joining us today. If the webinar or audio quality grades, please try refreshing your browser. If that does not fix the issue, please disconnect and repeat the original login steps to rejoin the webcast.

If you have a question, we encourage you to ask it using the question feature located on the webinar panel. We’ll collect your questions and do our best to answer them at the end of the presentation. If we don’t address your question within the time permitting, we’ll make an effort to follow up with you after the webinar. We are recording this webinar and a link to the recording will be emailed to you once it is available. In order to continue to sponsor events that are of value and worthy of your time, we will be sending out a brief survey following the webinar to get your feedback.

Today’s presentation is the third in our Regenesis Technical Perspectives webinar series on the application of advanced modeling for in-situ groundwater treatment using colloidal activated carbon. This module explores how modeling can be used as a support tool throughout a project assisting in communication, data interpretation, and performance optimization. With that, I’d like to introduce our presenter for the series.

We’re pleased to have with us Dr. Jeremy Birnstingl, Vice President of Environmental Technology at Regenesis. Dr. Birnstingl serves as a senior Regenesis technical resource on key remediation projects involving advanced in situ technologies worldwide. He’s the author of the PlumeForce software used for design and modeling of the Regenesis activated carbon-based technologies. Dr. Birnstingl received a bachelor’s of science in environmental biology from the University of Essex and a PhD in environmental chemistry from the University of Lancaster. He is a fellow of the Royal Society of Chemistry in the United Kingdom and a chartered environmentalist. He has over 30 years experience in the commercial and academic environmental sectors, including 20 years with Regenesis. His creative and scientific insight have been recognized through three commercial patents.

All right, so that concludes our introduction and now I will hand things over to Dr. Jeremy Birnstingl to get us started.

Thank you very much for that introduction, Dane. So this webinar series is about the use of modeling to expand our understanding of in situ activated carbon remediation projects. Modeling can offer a window into the governing processes behind the limited snapshots available to us through points groundwater samples and this information can be leveraged to improve design performance tracking and projection and engineering management and control. This is the final webinar in a series of three and in it we’re going to look at how modeling can be used beyond the design stage of a project and how it can assist with communication support expectations management performance tracking and the maintenance and of a plume stop barrier but first a quick recap.

The first webinar in this series of three set the scene. We looked at the basics. What is plume stop? Well it’s tiny particles of activated carbon suspended like an ink that can paint the subsurface and essentially stop plumes. What is a model? Well a model is a flight simulator that can be used to quickly and easily explore and interrogate different treatment scenarios. And how do we use them together? Well, any fate and transport model can be used for exploring Plumestop to some extent, but sophisticated models can be used for formal quantitative design, and indeed for many other purposes too, as this webinar series presents.

The second webinar explored some deeper dives into strange and curious dark waters. We looked at emergent phenomena and how simple processes can combine into complex processes that might not be obvious. We explored the impacts of hidden mass in the subsurface, how groundwater contamination is only a fraction of the total contaminant mass, and how hidden compartments strongly influence what we see in groundwater over time. And we introduced kinetic equilibria, static patterns in non-static systems that can resemble stall even when everything is cruising. We saw how modelling can expose these processes and provide a window into what is really going on so that we can understand and manage the systems appropriately.

And so, without further ado, let’s move on with webinar three, the use of modeling beyond design. In this webinar, we’re going to cover the use of modeling to communicate expectations, the use of modeling to assist in calibrated performance tracking, and how modeling can be used to support remedial performance optimization and installation maintenance. We’ll then wrap up the webinar series with an overview of what we’ve learned and the new perspectives that we’ve gained. Much of the previous two webinars used VOCs, chlorinated ethenes, as examples, and this was useful to demonstrate the impact of degradation, transformations, and the relative kinetics and rates of these. But in this module we’re going to focus much more on PFAS, and this is because their recalcitrant nature highlights the interplay abiotic principles, and in particular advection, diffusion, and sorption, which we’re going to explore in the present webinar.

And it’s also because to date, spring 2023, Regenesis has undergone some 40 or so commercial PFAS remediation projects around the world using plume stop injectable activated carbon, including some large airport and industrial installations. And modeling has proved to be a particularly useful tool in the support of these, in this webinar I’d like to share why. But first a few words on the model that we’ll be using to generate the examples. PlumeForce is modeling software for in-situ groundwater remediation. It was developed by Regenesis and we use it to support our clients’ projects. It was designed by me starting in about 2016 when it became clear that existing off-the-shelf fate and transport models were limited in their ability to accommodate activated carbon and indeed other remediation projects in the calculations that they would do.

Plume Force is a multi-phase finite difference model that accommodates dynamic sorption, destruction, and competitive interactions between target species and between target and non-target species, and what the software does is it predicts groundwater concentrations at any point in space and time in the model domain, and this has valuable uses at different stages of a project. It can be used for refinements of our conceptual understanding, refinements of the conceptual site model, in other words, highlighting information gaps or the driving processes that lie behind what we see. It can be used in the remediation design process for optimizing reagent selection, dose and arrangement of application, and for communicating what we might expect to see following a treatment and when we might expect to see it. And we can use it as a performance yardstick for project tracking and management, and we’ll get into this in the present webinar.

From the first version in 2016, Plume Force has progressively been refined and extended as we’ve been able to compare model predictions with actual field performance and use this comparison to hone the model as a working tool. As of spring 2023, when I’m giving this webinar, the present functionality is that it can handle the complexities of multiple reagents, the spatial arrangements and timing of their additions, and the reagent consumption and longevity. It considers the interplay of multiple sorbing species. It handles contaminant mixtures and parent-daughter cascades. It considers natural and hidden competitors that might not be from contaminants of concern or even in our analytical suite, but will nevertheless compete with our target species for sorption sites on the plume stop activated carbon.

Plume force also considers the interplay of multiple phases or compartments. It considers the aqueous phase that we principally measure in groundwater samples from monitoring wells. It looks at the aqueous phase in the lower permeability zones, zones such as the silts or clays that may harbor mass that can back diffuse out at us. And it considers the mass sorbed to the natural soil organics. And of course, it considers the mass that is sorbed onto the activated carbon of plume stop. Such multiple phase consideration is important because in many cases, relatively little contaminant mass is present in the groundwater, yet this is all we can generally measure. Like an iceberg, most of the mass is hidden. And we looked at the significance of this in the previous webinar.

So we’ve got a cool tool, let’s apply it. We’ll start with expectations management. Expectations do not have to be wishful thinking. There doesn’t have to be a gulf of uncertainty between what we’re hoping or expecting to see and what will likely transpire. We can reduce that uncertainty. We can use modeling to lay down a visual guide of what we’d expect based on the design and the monitoring well positions, and this isn’t about placing bets. Don’t place bets. Some monitoring data will be on track and some won’t be, but in each case, we can learn from what we see and use it to the advantage of the project.

The PlumeForce modeling software produces graphical data outputs in formats that correspond to typical project monitoring graphs. For example, we can look at how the concentrations of a mix of contaminants at a given monitoring well might change at different points in time, such as different monitoring rounds, for example. This is particularly useful in the early stages of a project or pilot study when there have not been enough sampling rounds to generate a time series graph. But we can also generate these familiar time series graphs, predicted concentration trends in a selected well, that is.

So in the examples that we can see on the screen, the concentration trends of multiple contaminant species overlaying forming quite a crowded graphic, but we can also single out individual species of concern within a competing soup of mixed contaminants when we need this clarity. It’s also possible to plot measured site data as an overlay on the model data to see how actual performance compares with the modeled expectation and we’ll come to that in a bit, but in the meantime notice how the dynamics for some of the species in the graphs on the right. And this is because the modeled well, in this case, is down gradients of the barrier. We’re seeing matrix effects come into play.

The residual concentrations that we can see, especially on the right, are due to mass that’s coming into the groundwater from the hidden compartments. The curious dynamics that we see in a couple of the species due to kinetic equilibria with the loss processes coupled with a bit of competitive sorption. So more on this coming up, but it’s important to note that the monitoring well position makes a difference to what we see, and in particular it makes the difference if the well is located within the plume stop barrier or gradient of the barrier. So why is that?

Well to explain this we’re going to need a deep dive, So fair warning, the next slides introduce some complex concepts. Once we’ve explored these and the mechanisms will come up for air and see what they actually mean in real life. So down we go. So with respect to expectations management, the first point we’re going to explore concerns in anticipated performance difference between wells in a plume stock barrier and wells a little distance down gradient. I’m not going to go over what a time, we covered that in the first webinar, and you can view this on the Regenesis website, Regenesis.com. But for starters, if you haven’t seen the first webinar, you can think of a plume stop barrier as rather like an injectable brita filter in the subsurface.

Anyway, a well located within the zone of plume stop application, an in barrier well, that is, will be subject to treatment influence and plume stop carbon will be present around the well. But a well-located down gradients of the barrier will also be subject to the influence of the barrier, just up gradient. But in this case, there is no carbon immediately surrounding the well. And this leads to important differences in the data patterns that we’d expect. Much of this is driven by matrix diffusion and results from back diffusing and desorbing mass as the system re-equilibrates and responds to the treatment.

In the barrier, this equilibrator The matrix flux is captured by the plume stop carbon, but down gradient there is no carbon to capture it, as the carbon is up gradient in the plume stop barrier. And this difference leads to different emergent phenomena, and we introduced this concept in the second webinar of the series. Anyway, rather than attempting to unwrap these complex points in the present slide, I’m simply going to list them here, and we’ll explore them with some graphical assistance from the plume force model outputs.

For now, the point to note is that monitoring wells within a barrier and monitoring wells downgradient from a barrier tell us different things, and the data they yield are complementary. Both well sets will respond to a plume stop application, but the data patterns that emerge will differ. So now we’ll take a look at why, and then we’ll explore the practical implications. This is one of the block diagrams generated by PlumeForce that we explored in the last webinar. We’re looking at mass on the vertical axis, micrograms per litre of aquifer, and time on the horizontal axis. In this case, the total pithola mass in the zone we’re looking at does not change at all.

The sum of the mass in all compartments remains constant at a little over 2.5 micrograms per litre of aquifer material. And this is micrograms per litre of aquifer material, not to be confused with an aqueous concentration. It’s the combination of mass in all of the different phases, soil, water, the lot. The colours show how this mass is distributed. Blue is the mass in the dissolved phase, the mass we can measure as concentration in a monitoring well. Brown is the mass sorbed onto or into natural soil organic carbon, the fraction of organic carbon or FOC. Grey is the mass in the lower transmissivity storage units such as silts or clay lenses and finally black is of course the mass on the plume stop carbon.

We can see that initially for the first 10 days or so of the model period that is, the PFOA was distributed between the dissolved phase and the groundwater, the soiled mass on the FOC and the stored mass in the low K units. The mass in the groundwater is only about 10% of the total mass in this unit. Like the submerged part of an iceberg, the balance of the mass is in the hidden compartments. It’s sorbed to the soil, brown, or soaked into the silts and clays, grey. But plume stock was applied here at the 10-day point. Immediately, the blue water phase mass disappears and black carbon phase mass appears, and this is because the dissolved mass has sorbed onto the carbon.

Concentrations in the groundwater are now non-detect. The mass is still there, but it’s switched compartments from the water to the plume stop. We can also see that the mass on the carbon increases with time. The black compartment grows. The total system mass remains the same as we saw to begin with, but the brown mass on the natural soil organic carbon and the gray mass in the low K units slowly transfers to the plume stop carbon. And this is because the groundwater concentration has dropped right down, therefore matrix mass from these compartments desorbs or back diffuses into the aqueous phase as equilibrium seeks to reestablish, but as soon as it enters the aqueous phase it instead partitions onto the plume stop carbon.

The mass goes to the plume stop carbon, therefore progressively increases, but the total mass in the system stays exactly the same. Most of this process is hidden from view. The groundwater that we can measure in monitoring wells remains non-detect or low, but the situation in the subsurface is changing. Importantly, as the sorptive mass on the carbon increases, as it loads up with back diffusing and dissolving mass, Competitive interactions between sorbing contaminants can lead to degrees of displacement for the weaker sorbing species that initially were captured, as we’ll see shortly. But in the zone downgradient of the barrier things are a little different.

Initially the distribution of mass is the same, blue, brown and grey, just as on the left. Plumestop application was here at about 10 day point as before, but this time nothing immediately changes, not for another 10 days or so. And this is because our well is 10 foot down gradient and it takes this time for the barrier influence to arrive with migrating groundwater. When the influence does arrive, we see a rapid concentration drop in groundwater, the blue mass, but this time it doesn’t drop to detect. We also see no black plume stop carbon mass and this is because we’re not in the barrier so there’s no plume stop carbon to capture the back diffusing and dissolving mass so it shows as blue in the groundwater.

Because the groundwater contaminant concentration has dropped the initial equilibrium has again shifted just as we saw in the barrier but this time with no plume top carbon to capture it, the re-equilibrating mass enters the groundwater and our monitoring well reports a concentration. And we can see this as the thin blue line along the top of the block diagram. We see the blue water mass initially drops to a certain point, but then declines only slowly. Well, why is that? Well, it drops to a semi-stable low concentration. And this is another example of a kinetic equilibrium, an emergent phenomenon, which we introduced in the second webinar of this series. In this case, the system has reached a balance point between the steady input of Pifoa mass from desorption and back diffusion and the ongoing dilution of this with the clean water that is coming out of the barrier.

And we know that the water coming out of the barrier is clean in this simulation from the block diagram on the left, no blue mass. But although the groundwater concentration is low and declining only slowly, the total system mass is declining much faster. A year after the plume stop application, the right of the diagram, we can see the total system mass is declined by about 90% as we can see from the sum of all of the compartments. The brown and grey mass is depleted via the moving groundwater even though the groundwater concentration has remained low. The downgradient zone is therefore cleaning up. Is the low groundwater concentration still high enough to be a concern?

Well, maybe yes, maybe no. That depends on the site and on the concentration. But if it is a concern, it could be addressed with a dusting of plume stop or a second further barrier down gradient. Either could be far smaller than the principal barrier as they only need to deal with matrix mass, not the influxing mother load. I call this the windbreak plus air freshener approach. This in fact helps me make another point of clarification in what we’re seeing here. It’s important to note that in this we’ve just been looking at the mass already in the barrier zone and the downgradient zone. Influx into the barrier from upgradient is occurring, but that hasn’t reached the mid and downgradient zones we’re looking at. It’s still being captured by the first part of maybe in 10 or 20 years it’ll get as far as the frames in the present example.

Anyway, deep dive over, we can breathe again. We’ve looked at the principles for those who’d like to know the details but now we’ll take a look at what it actually means in practice on a regular project. So let’s take an example of a well that’s in a barrier. Let’s see what patterns and trends we might expect from a groundwater taken from wells that are within a barrier and then we’ll take a look at down gradient examples but first an in barrier example. What we’re looking at here are modeled data for an in barrier well on an actual United States project. We can see how the concentrations of all the PFAS species drop sharply on carbon application but we can also see how the least XA, are not captured to the same degree. And we can see how their modelled concentrations in groundwater creep up over time. The creep up is due to competitive sorption and indeed the incomplete capture to begin with is due to the same.

The lighter species are more sensitive to competition and so as back diffusing and desorbing PFAS mass loads up the carbon as we saw in the deep dive, These lighter species are partially displaced. Well, is it a problem? Again, that depends on the project objectives. If these species need to be kept below a certain threshold, it just requires a heavier plume stock dose. But on this particular site, the targets were based on PFOS and PFOA individually, plus the sum of the wider set of PFAS species. And we were at compliance in this scenario and indeed for perspective the local drinking water standards for PF-HXA is 400 ,000 nanograms per litre and so from the get-go we’re less than one percent of this. PF-PEA isn’t listed in the regulatory drinking schedule either so the model performance is therefore at compliance.

So this is the model data but in practice this is what the actual performance look like. So here we have the first five months performance monitoring data overlaid onto the modeled trend lines. The actual site performance data of the markers joined by solid lines. The actual performance is on track for the target PFOS and PFOA, and the heavier species performance is very much on par, to use a golfing analogy. But notably, the reductions of PFPA and PFHXA are performing significantly better than modelled. And this is encouraging. It suggests the impacts of the hidden mass may be less than we actually modelled. So perhaps our estimates and contingency for these unknowns may therefore have been conservative.

This promises well for barrier longevity. Any spare capacity in the estimate translates directly to increased longevity of the barrier. And this is welcome. None of the carbon is wasted. We may even be able to recalibrate our projected longevity of the barrier a little. What about down gradients of a barrier? In our deep dive we looked at what we’d expect to see downgradient of a barrier using the earlier block diagrams combining the various soil and water mass compartments. These diagrams might be informative but they’re rather theoretically abstract compared to graphs we might see during the course of a normal project.

This graph therefore is another output of the PlumeForce modelling software presenting just one contaminant species out of the wider mix that was modelled, it’s PFOS in this case.The graph is directly related to what we looked at in the block diagram deep dive, but in this case it only shows the concentration of PFOS in the groundwater phase, what we’d measure from a well, in other words. Here, we can now clearly see the three phases that we touched on in the multi-compartment block diagram deep dive analysis. We can see a lag phase, followed by a fast decline phase, and then that is followed by a slower decline phase. We’ll compare these model data with measured data in a moment.

But first, and at risk of laboring the point and indeed of getting wet again, Here’s a quick recap. In the barrier, the dissolved phase is quickly captured on the carbon, leaving us with clean water within the barrier. This clean water creates a back diffusion and desorption gradient from the soil into the water as the system tries to re-equilibrate. The back diffusing and desorbing mass is then itself captured by the barrier carbon. At this stage, this is all matrix dynamics. Incoming contaminant mass flowing into the from up gradient is being held back by the front section of the barrier which is absorbing the principal advective barrage. Downgraded to the barrier there is first a time lag while we wait for the clean water from the newly installed barrier to arrive.

We then see a fast concentration drop on arrival of the clean water front and the spectrum of particle velocities in the different k units of the aquifer softens the arrival curve of this front at a given well. In words it’s a relatively abrupt decline close up to the barrier, progressively more gradual at increasing distance from the barrier. The clean water from the barrier creates a back diffusion and desorption gradient from the soil phases but we’re downgrading to the barrier now so there is no carbon here to capture this equilibrating mass. Instead the back diffusing and desorbing mass reaches a kinetic equilibrium with the clean water flowing out of the barrier and this kinetic equilibrium results in a semi-steady low concentration that slowly declines as the matrix mass depletes and with it the input rate. So we see these three phases.

Communicating that these will occur and why can be helpful information in advance. What should I expect to see in my downgradient well after a plume stock barrier has been installed? When will the concentrations decline and by how much? and when might we reach target concentration X, Y or Z. In this particular graph, we’re looking at a comparison of modeled and measured data from a site in the Midwest. The actual site data, triangle points and solid line are plotted as an overlay on the modeled data, the dotted line. This site has fairly stable conditions and we have a well populated data set. The graph therefore doesn’t bounce around too much And we can see that the fit between the model data and the actual data is good. Overlying modeled and measured data can be a very helpful exercise, whether the fit is good or bad.

More on this shortly. But before we get to that, you might also have noticed that compared to the earlier model plot that we looked at, the lag is longer, the principal decline not as steep, and the semi-plateau height of the secondary decline phase is higher. And this is because this particular well is further downgradient from the barrier than the one we looked at before and this greater distance naturally results in a longer lag time before the clean water arrives. The principal decline phase is less abrupt as the marathon runners, to use an analogy, are further from the starting gun and so not all particles arrive at once owing to the different k-zone velocities within the subsurface. And the secondary decline plateau is higher because the greater distance from the barrier means there has been a greater distance over which the sum of back-diffusing and dissolving mass can accumulate.

I call this the tributary effect as it’s rather like a river slowly growing with distance as many tributaries feed into it. Here’s the tributary effect shown in another block diagram output window of plume force. We still have mass per litre of aquifer material on the vertical axis as before, but this time the horizontal axis is distance from the barrier. The figure therefore shows the profile of mass versus distance at a given time interval. At time t, one year in this case, the front has migrated 365 feet because I’ve set in the model a groundwater velocity at 1 foot per day. We can see that depending on how far along this axis we are from left to right and how far a monitoring well might be, the higher the residual mass in the solution phase will be as a result of the longer path over which equilibrating matrix mass is fed in.

It’s also clear to see that the lag will be longer. For any well towards the right of the figure, the influence of the barrier put in a year ago would only just be arriving. Okay, modelling of 360 feet of heterogeneous aquifer is inappropriate on a practical level, But here, the exercise is one of understanding rather than prediction. We can also see how in the zone to the left, closest to the barrier, the total matrix mass has already declined by some 90% and is continuing to clean up. The rest will follow in due course and we can model and we can follow that. But importantly, we can see from the modelling what is going on. So one of the key take-homes from this modeling exercise is to understand that the position of a monitoring well in relation to a barrier makes a difference to what to expect. So these are three examples that I’m going to show you now. Same barrier, three wells, each at increasing distance downgradient.

This is the first. It’s at the barrier periphery, zero distance downgradient. The PFOA reduction is immediate and the concentration is reduced to almost nothing. Back to fusion and desorption are occurring from the hidden compartments, but this is being captured in another hidden compartment, namely the barrier carbon. So what we’re looking at here is the actual performance of the barrier. There is no interference from matrix mass between the barrier and the monitoring well, because there is no distance between the barrier and the monitoring well in which this can occur. It’s always good to have a monitoring well at this location for confirmation purposes, even if the formal project compliance point is downgradient.

So now we’re at the second well of the set of three, and this time we’re 10 feet downgradient from the barrier, and we can immediately see a difference. There is a slight lag, this may or may not be noticeable depending on the monitoring frequency, but more importantly we can see the kinetic equilibrium plateau as the downgradient The gradient equilibratory back diffusion impacts the clean water exiting the barrier. The back diffusion rate decreases with time as the matrix reservoir depletes and the plateau declines with this. The comparison of data from a downgradient well, such as this, with an in-barrier well enables us to calibrate the back diffusion parameters in the model, which in turn enables improved forward projection.

So now we’re at the third monitoring well, this time 30 feet down gradient from the barrier. The response delay is longer. We’re only seeing things start to change three months or more after the barrier went in. And this itself tells us not to spend monitoring pocket money on high frequency sampling immediately after application if the monitoring well is at a distance. And I’ve seen this happen. Monthly sampling for the first quarter, reducing to quarterly sampling thereafter, and then the influence of the barrier arrives at some time in the second quarter, and this makes it a tense six months before any results are seen, despite all of the sampling. This is a good example of where it’s helpful to manage expectations.

When do we expect to see a monitoring result? Well, this can be estimated in advance using modelling. So let’s compare these three graphs side by side. Now we can compare the expected performance patterns at increasing distances down gradients of a barrier, three different monitoring wells for example. In the second and third graph we can clearly see the two decline phases, the initial rapid decline and the slower gradual decline. At one year in, 365 days, the concentrations in the respective monitoring wells are approximately 0, 200 and 400 nanograms litre as we move further from the barrier. And these represent reductions of 187 and 75% respectively. But the 87% reductions and the 75% reductions are not the barrier performance.

The barrier performance was close to 100% reduction as we can see in the first graph. Rather, these are the barrier performance with recontamination from existing mass down gradient the barrier overlaid. This is to be expected and this is an expectation to manage. We’ve put a fence through a buffalo herd. No buffalo are crossing the fence but the buffalo already on the other side will take time to disperse. Is this a problem? What if this is a compliance well and the original reduction at this point has not reached target? Well, several answers to that.

First, the concentrations are declining just more slowly, the modelling shows us why and we can estimate or calibrate the rate to determine when compliance will be reached. And second, we can easily expedite the process using a light air freshener plume stock dose in the downgradient zone. This is a lighter application as it only needs to address back diffusion, not the far greater advective flux the principal barrier addresses. And for large plumes, of course, the design might even comprise a sequence of barriers and it typically would. But also we could consider, at the design stage, whether the barrier needs to be closer to the compliance point in the first place or vice versa. And this is another example of where modelling can help with the design.

The concentration at which the initial decline phase transitions to the secondary decline phase is a point of kinetic equilibrium, which will be determined by a combination of factors. whether this concentration is significant or not in the contents of the project will depend on the setting and modeling will help us estimate this in advance. So there are solutions in other words but importantly we can see that the declining low-level plateau concentration in a downgradient wells is nothing to do with the carbon dose or the barrier integrity. It’s therefore not fixed by fattening the barrier for example. The barrier effectiveness and integrity is complete as we can see from the graph on the left, the PFOA reduction is 100%.

We can therefore understand the basis of what we’re seeing and therefore know how to address it. And once more, modeling helps us understand better, communicate better, and design better. Let’s move on to performance tracking. Modeling as a performance yardstick. When a design is developed using modeling, or indeed if a model is built using design parameters, we will already have from this a set of expectations of how things will unfold if our input estimates are appropriate. These expectations could be considered as par for the course, expected performance that is, to any non-golfers, and this gives us a measure of what good performance looks like.

We can then compare actual site data to this as the monitoring results come in to see how we’re doing and the PlumeForce modeling software simplifies this by allowing us to overlay measured data on the model data outputs to quickly expose any discrepancies and this helps us in all cases. If the fit is good we can relax and put our attention somewhere else. We know performance at point is on track. But if there is a discrepancy it’s quickly exposed. We can determine if it is significant and therefore a flag that this area requires attention. And from the scale and shape of the discrepancy, is it sudden, gradual or erratic, we can gain insight into what might be the underlying cause.

Highlighting areas of concern and gaining insight into the cause informs an appropriate course of any necessary intervention and corrective action, and it is in fact a formal component of the commissioning of PFAS barriers that Regenesis guarantees under our plume shield warranty. So in the example we’re looking at here, we’re looking at data from another site in the US Midwest. The dotted line is the plume force model data output, and the solid line and data points are measured site concentrations of this well. I’ve also included an insert showing the numeric values. On this site we have a relatively noise-free groundwater data set again, in other words the measured values are not jumping up and down with rainfall events, compound monitoring errors and the like, and we also have a reasonable number of data points from quarterly sampling over a two-year period. This gives us a well-populated stable data set that makes for a good webinar example.

On this On occasion, the performance was certainly on par. You can see this for the first 400 days. But a nearby zone was not, and we’ll come to that in a moment. There was therefore a supplementary plume storm application to adjust the nearby zone. This kind of post-design commissioning adjustment is analogous to tweaking extraction rates on individual wells of pump and treat systems to fine tune, draw down, and capture, for example. You build the principle design with a model and then you fine-tune as you compare it to measured data. The initial drop that we can see at about 546 days is likely due to some carbon bleed from the nearby supplementary application, depositing downgradient in the locality of this particular well and capturing some of the back-diffusing mass.

And this is in fact an example of that air freshener approach to reducing the secondary declining plateau phase, as I mentioned before, albeit not intentional on this occasion. But nevertheless, it takes performance at this point from a 98% to 100% reduction, despite the back diffusion. And I think it’s quite helpful to see how this can work. But here is a downgradient well at a different point along the barrier. This one shows a deflection from PAR. The initial timing and performance was on track, we can see in the first data points, but we then saw a rise in concentration from about 180 to 400 days. Without going into details, this divergence pattern resembles what we’d expect from localized dose deficiency rather than from breakthrough.

And again, modeling tells us why, but that’s for another time. We therefore performed the aforesaid localized supplementary commissioning application in the upgradient barrier zone to bring the performance back on track and we can see that this was successful as the performance is back on par from the 546 day point to the last data point that we measured on this occasion or in this data set. So being able to track performance against qualitative and quantitative expectations is very helpful to a project for our clients, for their clients or even for us the data jungle can be bewildering. Modeling gives us a map. Where are we? What direction are we going in? And do we have to course correct? And using this, we can perform periodic project health checks. What would we expect to see at this stage? What are we seeing at this stage? Are we on course? And what are the key indicators to watch?

Addressing these four questions in a project performance review can cover a lot of ground. So we’re getting close to time so let’s wrap up this webinar and then I’ll move on to closing the webinar series. In this third webinar we’ve explored the use of modeling beyond the design stage. We’ve looked at how modeling can assist with expectations management for the designer as well as the client in fact and we can use this to monitoring programs to reduce uncertainty and to generally streamline project communication and delivery. We’ve seen how and why the position of a monitoring well relative to a barrier makes a difference and what these differences mean and we’ve touched on how the combination of in barrier and downgradient well data helps with calibration and therefore future performance projection.

And we’ve looked at the use of modeling in performance tracking. Is my project on track and how can we keep it there? What do the data patterns that we’re seeing mean? On large complex sites especially, it is super useful to be able to quickly screen performance data to determine which wells are tracking to expectations and which we might want to watch or put some attention to and indeed what might be behind any divergence that we might see. So let’s close the webinar series. But before we do, let’s take a moment look back at the ground we’ve covered and how we can now see this in perspective.

This was the series and congratulations if you’ve stood the course and attended them all. If you haven’t and would like to catch up with any of the previous webinars, they’re all three available to stream at Regenesis.com. Way back in the beginning, we introduced plume stop using a simple concept of contaminant capture and we described the plume stop barrier as analogous to a britter filter in the ground. But through the series we’ve moved beyond this to introduce a more powerful way to understand what plume stop is actually doing. Shifting beyond a simple capture paradigm to one of understanding plume stop as a means of engineering the subsurface KD is really what we’ve done. And through this the modeling dimensions are unlocked, and we can build new tools to better understand processes, to make predictions, and to gain insight and perspective into what a plume stop application will do.

And now the predictions we have are quantitative. We can put numbers to them. For example, we can determine the plume stop increases to KD and therefore retardation factor by typically two to three orders of magnitude for conventional contaminants, and by typically three to five orders of magnitude for PFAS. And these numbers, these changes are in competitive settings. So how can we understand this in terms that are easier to relate to? Well, let’s take a freeway example. You might be driving at, say, 70 miles an hour. Let’s use this as the reference velocity, r equals 1. So this is the velocity of groundwater in our analogy.

Against this, what does a retardation factor of 100 to 1000, two to three orders of magnitude look like? Well, that is a tortoise to your convertible, the velocity of a conventional contaminant in a plume stop barrier relative to groundwater. And what about a retardation factor in the 1000 to 10 ,000 range, three to five orders of magnitude. Well, that is a garden snail and actually quite a fast one, in fact, as snails go. This would be the velocity of a PFAS species in a plume stop barrier relative to groundwater. So let’s say that it takes groundwater 30 days to transit a plume stop barrier. This is by definition a retardation factor of one, i.e. the groundwater velocity. A 30-day transit would be a 12-foot barrier, say, that’s 12 feet parallel to flow that is, and 150 feet per year groundwater seepage velocity. So how far could you drive on a freeway in 30 days at 70 miles an hour 24-7? And how long would it take a tortoise or a snail to walk or slide this far. That is numeric perspective.

Even a weakly-solving species would be a pedestrian to your corvette. So that’s what we’re talking about in terms of barrier longevity and the slowing of contaminant migration with plume stop. The clue is in the name plume and stop. This is a simple numeric perspective that we can derive from modelling at the most basic level. So finally, a question that has come up at both previous webinars is, can I get a copy of the PlumeForce modeling software? The answer is no, it’s Regenesis Internal only at this time. And we use it in the design process and in the wider technical support that we offer through a project. So let me take a moment to touch on that.

As many of those attending this webinar will be aware, Regenesis provides technical support free of charge on projects that use PlumeStop or indeed any other Regenesis projects. We have a large technical support team that works behind the scenes and the experience that we have within this team is formidable. It’s drawn from more than 30 ,000 project sites over the years in some 25 countries on six continents over approaching 30 years. We complete more than 100 project designs per month and we’ve applied Plume Stop commercially on some 500 project sites over the last decade. Forty of these have been commercial PFAS projects. Overall we have several hundred years of combined experience on in situ project design and review using reagents. The Plume Force software is just part of this. It’s a tool built from experience that combines with experience. And it’s a useful tool.

I’ve maintained from the beginning of this series that modeling can help us understand better, design better, and communicate better, and how these points combine to ensure the delivery of a successful project. I trust that this webinar series has clarified some of the hows, whys, and wherefores of this. Thank you.