Hourglass Climate's Dr. Grace Andrews and Kristi Weighman on the launch of the Framework for Ecotoxicological Modeling of mCDR

In this episode, hosts Anna Madlener and Wil Burns are joined by Dr. Grace Andrews and Kristi Weighman of Hourglass Climate — a leading nonprofit researching ocean-based carbon dioxide removal (oCDR, also known as mCDR ) methods like ocean alkalinity enhancement (OAE) — to discuss the launch of the Framework for Ecotoxicological Modeling of mCDR (FEMM). This project explores how ecotoxicological modeling and existing statistical approaches can be applied to OAE and oCDR projects, improving the field’s understanding of these potential climate solutions’ environmental risks. Dr. Andrews and Weighman offer insight into their process building and receiving feedback on FEMM, the framework’s regulatory potential, and how FEMM can be applied across oCDR research. 

Dr. Grace Andrews, Founder and Executive Director of Hourglass Climate, now in her tenth year of working in the CDR field, last appeared on Plan Sea in 2024 to discuss Hourglass’ role in advancing monitoring, verification, and reporting (MRV) for OAE. In this episode, she’s joined by Kristi Weighman, an Hourglass scientist with expertise in ecotoxicology. Together, Grace and Kristi discuss how they recognized a critical gap in oCDR research — the lack of tools to monitor and model environmental risk — and developed a first-of-its-kind framework to fill that gap. 

Grace explains how our understanding of oCDR’s environmental safety has lagged behind scientific developments in the field. In order to advance these projects in a responsible way, Grace believes that the field needs more rigorous, standardized approaches for modeling and measuring environmental risks. 

FEMM aims to address this gap through combining established statistical approaches with emerging modeling techniques, borrowing existing protocols from the ecotoxicology space and applying them to the nuances of oCDR. The framework begins with a screening-level assessment that uses highly conservative assumptions to determine whether a project’s risks can be ruled out. Projects with identified risk may need to redesign aspects of their approach before moving on to more realistic assessment tools. The modeling relies on species sensitivity distributions (SSD) and calculations based on predicted environmental concentration (PEC) and predicted no effect concentration (PNEC). While this SSD approach has been applied to other environmental stressors, this is the first time it’s been applied to oCDR. Grace and Kristi also highlight examples of specific mCDR stressors and conditions where data may be too sparse to fully apply this approach today, and outline research priorities that will enable a standardized approach for these over time.

Looking ahead, Grace and Kristi share their optimism about FEMM’s utility for researchers and broader oCDR stakeholders. They hope the tool will enable users to identify potential risk in their proposed projects, integrate cross-disciplinary data, and foster greater regulatory dialogue. 

Plan Sea is a semi-weekly podcast exploring ocean-based climate solutions, brought to you by the Carbon to Sea Initiative and the American University Institute for Responsible Carbon Removal.

To listen to Dr. Grace Andrew’s first Plan Sea podcast appearance,

Plan Sea is a semi-weekly podcast exploring ocean-based climate solutions, brought to you by the Carbon to Sea Initiative & the American University Institute for Responsible Carbon Removal.

Transcript

00:13 - Introduction to FEMM and Hourglass Climate

Wil (00:13): Hello, welcome to a new episode of Plan Sea: Ocean Interventions to Address Climate Change podcast. As always, I'm your host, Wil Burns, co-executive director of the Institute for Responsible Carbon Removal at American University. And I'm joined by my co-host, Anna Madlener, senior manager for MRV at the Carbon to Sea Initiative. Today, we're welcoming Dr. Grace Andrews, Executive Director of Hourglass Climate and her colleague and ecotoxicologist, Dr. Kristi Weighman. Hourglass Climate is a nonprofit organization that conducts research on marine carbon dioxide removal methods. Grace was previously on Plan Sea to introduce the organization's work, so make sure to check out that episode, which we'll also link in the show notes. Today, Grace and Kristi join us to present and discuss a new publication that they've been working on, entitled the Framework for Ecotoxicological Modeling of mCDR, or for short, FEMM. Ecotoxicity is a common metric in fieldwork planning in many areas, as it evaluates how organisms respond to exposure to a substance and is often used by regulators under pertinent statutes and regulations. But it hasn't yet been broadly or easily adaptable to marine CDR. The framework that Hourglass is developing is intended to guide environmental risk assessment approaches for proposed MCDR projects, notably those that alter seawater chemistry, such as ocean alkalinity enhancement and direct ocean capture. Anna, what are you looking forward to today? 

Anna (1:55): A number of things. I think. Broadly speaking, I'm very curious to hear more about how Hourglass developed this framework and how they navigated some of the data gaps that still exist around ecotoxicity or ecotoxicology. I think in the marine CDR world in particular, we're in this sort of awkward phase where we still lack empirical data to really draw definitive conclusions or rules, but at the same time, we're really in need of protocols or some level of regulatory guidance. And tools like this framework, I think, are incredibly important for aligning research early and providing these guardrails while still allowing integration of new knowledge as that comes along and adjusting those guardrails. So that's an incredibly difficult task and can be quite time intensive, quite difficult to rally folks around it and building consensus. Yeah, I'm curious to hear more about that process and where they've landed, because it'll be, I think, ultimately applicable to other areas as well. 

Wil (2:58): Yeah, I agree with all of that. And beyond that, no surprise, I'm interested in how they think this may be integrated into pertinent regulatory systems, either at the national level and at the international level. And I'm interested in how they proposed to assess long-term exposures, right, which are more difficult in some ways to assess through the kind of research that we currently do in the context of MCDR. So it should be very interesting. 

Anna (3:36): All right, shall we bring them in? 

Wil (3:39): Let's do it. 

Anna (3:42): Hello, hello. 

Grace (3:43): Hi Anna, hey Wil, good to see you. 

Kristi (3:45): Hi. 

Will (3:48): Okay, at the outset, Grace and Kristi, could you briefly introduce yourselves? 

Grace (3:53): Yeah, happy to. Great to see you guys and thank you so much for having us on Plan Sea. Really, really happy to share what we've been working on. But I'm Grace Andrews. I'm the founder and executive director of Hourglass Climate. We are a nonprofit working on the responsible development of R&D around marine carbon dioxide removal strategies. But me personally, I'm a geochemist and I have spent the last 10 years working on carbon dioxide removal, both at land-based technologies like enhanced rock weathering, but more recently ocean alkalinity enhancement. And then the only thing else I'll say about myself is I realized the other day as the new year ticked over that 2026 marks my 10-year anniversary of working on carbon dioxide removal. That's been 10 years since I pivoted from sort of geologic time scale problems. So I feel like this year at some point I need to pop a bottle of champagne or something. 

Anna (4:52): Have you seen these, you know, there's this trend on social media to post about 2016 right now. I don't know, my social media is quite full of it. So your 10 year 2016 throwback would include, I started working on rocks. Yeah, yeah. 

Wil (5:08): Wow. 

Grace (5:08): I'm not on social media. 

Anna (5:09): Good for you. No need to be. 

Wil (5:13): You're not missing a thing, trust me. 

Kristi (5:16): Hi, I'm Kristi I am an ecotoxicologist at Hourglass Climate, so I am not coming from an mCDR background. I'm not quite even at a two-year anniversary in this space. I'm newer to it. I really come from aquatic toxicology and my background is largely in mechanistic modeling. So until I came to Hourglass, my work was really focusing on metals in particular, but just generally how substances affect organisms and ecosystems and how I can translate that into models. So now at Hourglass I'm working to try to bring that skill set and that perspective as we're trying to build out these assessments. 

6:06 - Why Ecotoxicology Matters for Marine CDR

Wil (6:06): All right, great, thanks for that. So let's set the table a little bit in our discussion of FEMM. Let's start off briefly, what is FEMM and what was your motivation in developing this? 

Kristi (6:20): So maybe I added a little bit with that accidentally in my intro, but FEMM is a framework for evaluating environmental risk from marine carbon dioxide removal projects. And as to the motivation, I think that, well, a few things drew me to it. I guess I can't speak to Grace's motivation in sort of like starting the project, but I think as an ecotoxicologist, there just seems to be a real potential for overlap between the disciplines. It seems like fundamentally trying to address a lot of the same concerns. So I think that there's a lot that can be borrowed from the ecotox discipline and really trying to close the gap between the amount of environmental data being collected, but a relative limitation of like specific guidance on how that data should be interpreted. So we have a lot of monitoring protocols, but relatively little shared understanding of what those results mean in terms of ecological safety. 

Wil (7:27): Anything to add there, Grace? 

Grace (7:28): Yeah, I can add. I mean, so FEMM, which I'll say it stands for the Framework for Ecotoxological Modeling of mCDR. FEMM is something I have wanted to do since I founded Hourglass two years ago. It was in the original grant proposal that I wrote to get funding to found Hourglass. And for me, the motivation there was coming from the observation that as an industry, we talk about the two pillars of knowledge that we need being around carbon removal, efficacy, and efficiency, and environmental safety. Those are the two things we always talk about. But I observed that, in general, we spend a lot more time talking about, how do we quantify carbon removal? We spend a little more resources thinking about that. That's actually the topic that gets the vast majority of attention. The environmental impact side, I think is frankly kind of lagging behind and so we  really identified this need that if we're actually going to move the industry forward and sort of get past where we are right now, which is the research scale or, basically doing such small field trials that that’s really our protective mechanism for the environment. If we're actually going to advance these technologies, if we're going to think about getting to a place where we can have climate scale impact, we need to get a lot more serious about how we're quantifying environmental impacts. We need to be a lot rigorous, we need a lot of standardized, and we need to be moving this forward in a sort of structured way. The world of environmental risk is wide, and this is something Kristy can speak to at depth. But there is so much precedent out there from other industries. We should not be reinventing the wheel here for mCDR. We should be jumping on the decades and decades of work that has gone into thinking about how we do this well and really just figure out how we can sort of adopt that knowledge for MCDR. So that was really…I didn't know how to do that. Again, I'm a geochemist, but I knew I felt it could be done and I just needed to find the right team to help accomplish this goal. So I was really excited when I found Kristi. She's fantastic. 

Kristi (9:57): And for me, I thought the project was really exciting just in that nobody goes into environmental sciences because they're indifferent to the environment. But in ecotoxicology, I think that sometimes it can be sort of a bummer when you don't get to... There's not a lot of avenues to work on something like climate change mitigation. So, there's a lot of opportunities where you could really contribute to something like that, at least that I encountered. So yeah, it was really exciting to be invited into a different context, different mindset, a slightly different set of problems, but deeply familiar in other ways. 

Wil (10:41): Yeah. I think you're both right here. If this is going to be regulatory ready, any of these approaches that we're talking about, right, that this kind of assessment is just a critical component, right? That anybody that's seeking permits, whether it's for research or for deployment, is going to have to run this gauntlet and they're going to need guidance. So I think this could be really, really important in the field and I'm glad you're doing it. So tell me a bit about the process in building FEMM and kind of a sub-question in this, what was the response that you received during the public consultation period for this? 

Kristi (1:27): So, it's really a multi-phase process. So, I mean, of course it started just internally and that was actually quite a long phase of conversations between me and people who know a lot more about mCDR than I do and me being able to bring in more ecotox context and going back and forth in those resources and sort of sniffing out what we thought was just going to be of the most value in the space. From there, it kind of moved to just a really limited circulation with some experts in the fields from different perspectives. So really trying to get eyes on it from NGOs, from regulatory bodies, from academia, across all the disciplines that sort of fold into this. And after, their extremely helpful input we moved to the public comment phase, which was a preliminary draft of the document. We've gotten a lot of feedback that I think has been generally really encouraging. I think there's a lot of interest around this, but of course people also raise some really interesting concerns or like aspects of the documents that could be stronger places where we could really follow through on some of these ideas more. And that's led to a lot of, I would say, improvements in the document and the new version. 

12:55 - Applying FEMM to Real MCDR Technologies

Wil (12:55): So when we look at FEMM and in the context of mCDR, what technologies is it applicable to? Is it really a focus on ocean alkalinity enhancement or will it be more universally applicable? 

Kristi (13:12): So it's definitely immediately ready for certain applications in the context of ocean alkalinity enhancement. So I think that's why throughout the doc, you'll see a lot of mention of ocean alkalinity enhancement because these things are really being lifted directly from the ecotox space. The easiest way to apply them as they are is for things like metals, but it's also a really immediate priority for us to figure out the most effective way of adapting these fundamental elements of these approaches to more novel stressors like the carbonic chemistry perturbations that are associated with a lot of other MCDR technologies, so things like direct ocean capture. We're really thinking about how we can expand it out from there. So it's ready for some uses in spaces already, and then we're hoping. In the near term to have made headway on even widening that use. With that said, because it is ready for use for toxicants, anything with metals, really, anything with pesticides, those are already really well-worn tracks in the ecotoxicology space. So we have a ton of reference points and actually a lot more data than we do on the more novel things. So I think that any other MCDR technologies that have concerns around more traditional toxicants like metals or pesticides, this is ready to go to address those concerns. 

Grace (14:47): Yeah, and I would add to put some names on this specifically. I mean, of course, ocean iron fertilization, right? There's a lot of metal in that. So that's a great example of another mCDR strategy for whom FEMM is immediately applicable. Kristi’s reference to pesticides, I mean I think that's helpful one, because there are CDR strategies out there that grapple with that. So, terrestrial biomass sinking is one, for example, and pesticides, as Kristi said, are a traditional toxicant. So that is another strategy that FEMM could step in and be immediately actionable around. And then I would even say, we're obviously focused on the marine world over here. But one of the really cool things about FEMM is that it is tech-agnostic, right? It is a structure for how to think about ecotox risk. And so it is inherently tech-agnostic, which means, although again, we focused on the marine world because we think that's where it has the highest potential to really move the field forward, you could actually use the exact same structure to think about risk for strategies like river alkalinity enhancement. It doesn't actually matter. So we think that this is something that we got interested in for ocean alkalinity enhancement, but it actually has so much potential to move environmental risk assessment forward for the whole swath of open system CDR approaches. 

Wil (16:25): Yeah, that's interesting. And who's the intended audience for this? 

Grace (16:32): Yeah, great question. I mean, FEMM is…one of the ways we're hoping it can drive impact is it has use cases for project developers, startups, anyone who's actually trying to build projects that can also include academics, of course. But anyone who is trying to a project, it can be helpful for regulators who are trying to get their heads around what is mCDR, what is ocean alkalinity enhancement, and how should I think about assessing permits? It has use cases, huge use cases in MRV, right? So for protocol developers who are thinking about how do we build rigorous protocols that again balance both that CDR but also environmental impacts. So there are a bunch of different people that we're hoping and that we believe can use this document. And so it has use cases on both, sort of all sides of when you're building a project, when you're scoping a potential project and you're thinking about where do I build this, how do I cite a project in a way that's likely to be safe. Again, it has use cases when you're going through the permitting process, when you actually want to quantify the potential risk of a project that you're proposing. And then on the back end, when you've implemented a project and you're sitting on a mountain of environmental field data and you are asking the question, OK, was the project safe? We collect all this data, what do we do with it? How do we say it was actually executed safely and buyers should feel comfortable moving forward with buying credits from this particular project? It's got applicability there, too. 

Wil (18:14): You know, we do a lot of work at the Institute with community engagement. And I was curious, given the fact that this is pretty complicated at times, abstruse science, do you contemplate that it might be possible to develop a version of FEMM, or to translate results from FEMM, in a way that community stakeholders might be able to find it useful?

Kristi (18:49): Yeah, I think that element is actually really important to seeing use. You know, I think that the community buy-in element is really critical in actually seeing these technologies be deployed. So we have really tried, especially me I think, because I can sometimes be a little pedantic, to interrogate the way these things are presented. And so even…it's not ready for just anybody to immediately understand all of these. I won't oversell that, but we tried really hard to think about the structure of this in a more intuitive way. So you see in the document a lot of flowchart pathways and kind of trying to always reduce it down to what is the fundamental ask here? Like, what is the slate as bare as it can be? And I think that maybe not all the nuts and bolts of the models are going to be of interest to communities or going to useful when you're trying to get buy-in, but I think that the outputs and like those decision points are things that can be really effectively communicated and even the uncertainty around those decision point is something that continues to be communicated so that people can have a better sense of what's being proposed for their communities. I think it'll help to have those more clear checkpoints to foster that dialogue. 

Wil (20:21): Yeah, I think that'll be…if you can pull that off, that'll a valuable part of the deliverables. 

Anna (20:29): Maybe somewhat in that vein, I wanted to circle back before we dive into the details of what's actually in the document, of course. I want to just circle back to something you said earlier, Grace, which was that of course environmental safety and efficacy are sort of these two big buckets that we've talked a lot about in the mCDR community within the bucket of environmental safety. What is FEMM trying to do and what isn't it trying to do? I'm wondering if you want to share a bit more about the stage or the contribution to environmental safety assessments that we're really talking about here and how that will sort of live in that realm of environmental safety. 

Grace (21:10): Yeah, yeah, it's a great question. So FEMM is definitely not trying to do it all. I think when you think about environmental safety and environmental risk assessments, there is a lot that goes into that. I can speak from experience having been through the permitting process for multiple mCDR projects, when you're actually working with regulators on environmental safety, they're asking questions about everything from sort of the unique stressors of your to the noise pollution from the boats that are going to be on the water to do the monitoring, right, and everything in between. So where FEMM is trying to step in is around the aspects of the environmental impacts that are unique to mCDR, right? So the noise, pollution from a boat, whether that's a boat out there doing dredging or a boat up there doing monitoring for an mCDR project, understanding the risk and the impact of that is the same. What FEMM is trying to do… and all that to say, that's great because there is a lot we can build on for that aspect of environmental risk assessment. What FEMM is trying to do is address the unique parts of marine CDR and and that includes certainly for things again like OAE and direct ocean capture, that carbonate chemistry perturbations, but it can also include the addition of metals that have traditionally not been a big focus for other industries. We're sort of adding unique minerals to the ocean for a lot of these strategies that have compositions that you don't commonly see elsewhere. So it's grappling with that. It's also just grappling with adding these stressors to the ocean in kind of unique circumstances. So it is trying to deal with the parts of environmental risk assessment that would be sort of a step outside the normal boundaries. And so the goal here is that if a project planner were to do a holistic risk assessment for a project, they would go about their business for most of that risk assessment. And then FEMM would step in to just sort of tackle that novel part. So it would be like a piece of that overall assessment portfolio. 

23:29 - How the FEMM Framework Works

Anna (23:29): Yeah, I'm excited to dive into how the framework works and maybe circle back to this point. I think environmental safety, environmental impact are often discussed in the context of monitoring and sort of, what should our monitoring plan be? What should all the parameters be that projects should be monitoring? And what you're trying to propose with this framework is coming in a step before that, right? You're saying here's a framework to assess your actual risk beforehand and use that as a tool to guide potential monitoring that will also be inclusive of all the other environmental stressors that you've talked about that are not specific to mCDR. So I’m wondering if you want to build on that thread a little bit and share more about how the framework itself works. What does it consist of? 

Kristi (24:14): Yeah. So as you said, this is something that can really come into play before a project is implemented, but Grace touched on this earlier and I do want to loop back to it. It is also just as easily applied in a post-appointment assessment to make sure that things were executed, they did play out as you thought they would. So it's sort of cyclical in that way. But in very broad strokes, FEMM is a tiered approach. There's basically two big bucket steps in FEMM. So the first thing in the document you'll see is a screening level assessment. And in that step of the assessment of the screening level, we're really working with initially very conservative, arguably hyper-conservative assumptions, where it’s possible to basically identify if risk can confidently be ruled out. This is really a filtering process. If a project can establish low risk under those really conservative assumptions, then they can stay in this screening step where the approaches are relatively simple in that they don't require modeling background or expertise. We kind of want to, as much as possible, avoid directing people to more high-effort things where we can. So that's what that screening step is for under hyper-conservative assumptions. If there's no risk, we don't need to make it harder than it is. But where the ball is on the line, where there's a potential for risk flagged under those conservative assumptions, we are eally trying to direct users to more refined tools with a heavy emphasis on mechanistic modeling that will allow for more realistic assessment, still conservative, but closer to reality than the sort of hyper-conservative approach you take on the screening level. 

Anna (26:14): And can you say a little bit more about what you're screening for f in this first stage, what is the first metric that users would look at and what would they compare it against? 

Kristi (26:25): Sure. So at both steps in this, whether you're working with a more refined thing or the screening level, the framework is ultimately always going to come down to the same question. And  I'm happy to talk about how these values are derived, but you're always at a fundamental level comparing two values and that's your PEC, or predicted environmental concentration, and your PNEC or your predicted no-effect concentration. At the end of these assessments, the question is going to be, is the predicted environmental concentration higher than the predicted no-effect concentration, because that's going to be indicative of risk. And basically, the difference between the steps is the approaches you're using to arrive at estimates of either of those values. 

Anna (27:18): I just wanted to make sure that I understood this right, the underlying metric of the PECs and PNECs - are these species sensitivity distributions or SSDs, is that correct? 

Kristi (27:30): That's the case for the predicted no-effect concentration. So yeah, that's where that's coming from. 

Anna (27:39): Do you want to share a bit more about these species sensitivity distributions? Because they form quite a foundational element of the framework, right? And that's the metric that we don't have for marine CDR, if I understood that correctly. 

Kristi (27:52): Yeah, yeah, I think that's a good place to go back to. So SSDs are species sensitivity distributions, and it's basically just a type of statistical model that describes how sensitive a wide range of species are to a given stressor or toxicant. So you have an actual statistical distribution of what the range of sensitivities look like for different species. So when you're fitting that distribution, each point in there represents the concentration of that stressor that a particular species is sensitive at. And if you get diverse enough taxa in one of these distributions, it starts to describe basically the range sensitivities in a community as a whole. So from those distributions, you can derive predicted no-effect concentrations for the assemblage of species or thinking really on the ecosystem level. So in that way, it's very much a summary statistic that is trying to think not on the level of individual species, but thinking about communities of species. 

Anna (29:09): And therefore, they're a good tool for field research design because they, in simple terms, would sort of give an indication of the level at which a field effort, for example, would be considered safe, using specific stressors. And if the predicted environmental condition of that actual field effort would be higher, there is a risk. And so for marine CDR, these underlying data to derive these SSDs for the predicted no effect concentration don't exist. Is that sort of the right simple summary? 

Kristi (29:46): Yeah, yeah, I think so. So these things are buried in a lot of guidance documents and stuff, but it's now more of the nuts and bolts being brought to light. I don't think that this approach is being applied in the mCDR space, but as we've kind of been touching on, it has been applied very much to stressors that are relevant to that space. So for some of the stressors that are going to be popping up for mCDR, it'll be more ready to go. Some of them, the development of this is going to be contingent on getting the data first. Can't get anywhere in a statistical fit without data. 

Grace (30:29): I'll add though, one of the cool things about this approach when Kristi really first started working on it that got me excited was that I could see immediately that the underlying data that you need to sort of fuel this modeling approach was starting to be created in the mCDR space. So you do see a lot of companies out there doing sort of standard lab-based ecotox testing. So actually zeroing in on some key species and doing exposure testing following, you know, your sort of standard EPA protocols and things like that. That is the data that you need to make this whole thing work. And so it's something that project developers were already doing. And I think one of the big advances that FEMM is creating is it's taking these disparate data sets where a few live over here, and a few over there, and maybe you have an academic lab group doing some of the work. And it creates one framework that pulls them all together to make conclusions at the ecosystem level. So it's building on the work that's already being done and creating more out of it, which is cool. 

Anna (31:54): I want to follow up on what Grace said around laboratory experiments, because that was something that as a non-expert on ecotoxicology, obviously, that stood out to me. One of the, I guess, shortcomings…not shortcomings but one of the challenges with the lab-based work that you mentioned, Grace, that at least I've heard about is this difficulty of actually mimicking the dissolution of feedstock in the water fast enough in a context, therefore either over underestimating the actual concentration or effect on species because you lack the physical conditions of the ocean in the real world where temperature, salinity, dilution, currents, etc. quite quickly, potentially quickly, hopefully quickly, reduce the actual concentration of your feedstock in that area. Obviously, the framework suggests to basically identify a risk when the predicted environmental concentration is higher than the no-effect concentration. I'm wondering how you're handling, or you're thinking about, this challenge from the modeling perspective and if there's, I guess, a bias in the model as well if it's based on the laboratory data that can over- or underestimate this concentration, because it doesn't mimic these physical conditions. Does that make sense? 

Kristi (33:16): Sure, sure, sure. So, I mean, I think largely what you're driving at, while I think there are some unique logistical challenges and sometimes feedstock experiments, is that, is the universal tension in environmental risk assessment, which is that we're always trying to translate stuff from a controlled lab environment into the environment, and that's just a fundamental problem, you know? But how this really is generally addressed is one, measuring stuff more, really in the controlled environment and in the actual environment, natural environment, really taking care to quantify the concentration of your stressor in dissolution or different water quality parameters so that those systems are better defined and you can more easily compare them to one another. So that type of direction on data collection is also something you'll see throughout the document. We really want to encourage people, when they're doing experiments, to take measurements that are going to add value, both in terms of the SSD construction stuff and hopefully more precise mechanistic modeling stuff down the road, because when you have a shortcoming in that information, it just really limits the utility of those data sets. I think that the other thing I would quickly note is there is no way to eliminate uncertainty in modeling. And I think the benefit of this type of a lens is that you can quantify not knowing on some level. I mean, that doesn't resolve the question in its entirety, but it can help identify blind spots, right? Where do we have less knowledge about things? Where do we have a shortcoming of data? So I think that although quantum modeling stuff will never resolve uncertainty in its entirety, especially when you're jumping scales from the lab to the ocean, it can help us pinpoint where the uncertainty is really living. 

35:43 - Uncertainty, Modeling Limits, and Scaling Risk

Wil (35:43): So chronic marine species sensitivity distributions have really largely been developed only for the impact from a handful of metals. And that's obviously relevant for mCDR approaches that are seeking to use a mineral feedstock that may have impurities such as olivine. That seems relatively, and I put that in quotes, straightforward. But you note that fewer SSDs have been developed for carbonate chemistry stressors, right, such as total alkalinity or pH or PCO2 changes. And a lot of those are going to be terribly relevant in the regulatory context, right? Can you say more about why this is challenging to date and what you envision and practice in this regard? 

Kristi (36:35): Sure, so not to start out like catastrophic, but I would take it farther than fewer have been made, essentially none have been made. There has been…I've seen one paper doing some of this with pH stuff, but this is not, this is somehow just not moved. When I say somehow, I do have an idea of how, because we've tried to play this through. So in the document there is now, in the newer version, a case study of a preliminary version of this, but in executing, moving the SSDs over, you do encounter just some fundamental problems. The first of which I would say is that these carbonate chemistry parameters are just intrinsically interdependent. They move with each other. So that just complicates the math in terms of, it's not like something like nickel, where you can deal with responses and how that plays out in isolation. The second thing is that these aren't toxicants in the way that nickel is. They can certainly be stressors, but because of that they don't follow this standard pattern of like. More of this or a higher value of this means there's more stress. Often you'll see more of a u-shape in that distribution because there's an ideal range and higher or lower resource. So that just changes how it plays out a little. The good news is there are actually a lot of people who have been working on defining those shapes in the mCDR context, but also there are sort of analogous things in ecotox that deal with things like essential trace metals that set that up. So we're kind of borrowing from both of those schools of thought to try to translate this. 

Wil (38:35): Yeah, I mean, I know from a standpoint of U.S. regulations, right, when you're trying to get an NPDES, right, you're trying to get a permit under the Clean Water Act, these are really critical things to be able to assess, right? So I think that's a really important aspect of what you're working on. And it's striking to me that that's been under-emphasized to date, given how many of these technologies are going to, at least if they're being permitted in the United States, going to have to be able to grapple with those issues if they're going to get permitted. 

Anna (39:12): I fully agree with that and related, as we mentioned a number of times now, the idea of the framework is that the risk is not present or the framework gives an okay if the no effect concentration is higher than the predicted concentration. What happens when it's the opposite, and how likely do you think that that is for projects today?

Kristi (39:38): So, if the predicted environmental concentration is higher than the predicted no-effect concentration, that does indicate potential for risk. So if that's occurring in that more conservative screening step, then that may be able to be resolved through the refinement in the more advanced assessments, but if that is persistently being flagged, even upon refinement, when you've already arrived at pretty realistic estimates of those values, that may signal the need for some kind of project redesign. So that could be including things like dose timing or some kind of operational strategy adjustment. So a benefit, even if I think that's not the answer anybody will want in their assessment, but I think another value out of this is because it is the sort of stepwise approach working with the species' sensitivity distributions. It's structured in a way that I think can help you identify where that risk is coming from and what's driving that. So I think that can also be informative if some kind of redesign is called for. I think at the current scale, odds of that are quite low, at the pilot scale, because we're talking about relatively small deployments. I think that's unlikely. I don't want to be too, meet you with too much certainty there, but if we're really talking about scaling the use of these technologies as to seriously be talking about climate change mitigation, it's going to have to be bigger. So, then I think that risk becomes a lot more real. And I think we have to, I think that's sort of a reason why this is a natural moment to be really interrogating our approaches here and having these types of assessments in place seriously. 

Anna (41:43): I think we'll also encounter, I mean, generally in this world, we'll encounter some philosophical questions around comparing that risk against the sort of larger risk of climate change itself, right? So I bet that that will come up. Have you guys considered or are you planning to maybe pilot this with some of the projects that are currently ongoing? 

Grace (42:03): Yeah, I mean, if you look in the doc, I think one of the nice things we've done is we have already provided some case studies where we applied FEMM to real-world project data from some of the field trials that Hourglass is involved in. So that's in there already. One, I think it’s just quite helpful for illustrating in real terms what the principles are. But it also demonstrates in real terms the utility, even at the field trial scale. But yeah, I mean, I think in terms of where we go next, one of the things we want to do is take…FEMM is itself a guidance document. What we want do next is actually start building out some of the tools that will bring FEMM to life. So we're actually interested in our next project being like a software development project that will make FEMM just open and a lot easier for anyone to actually implement. So taking these ideas and having software that actually builds your SSD for you using vetted data, you input your project conditions and it can generate that stuff for you. That's always been one of our goals ism how do we take the big wide and complex world of ecotoxicology and make it accessible for everybody, right? Including, you know, these startups that don't happen to have a Kristi on their team, right? How do we actually get these startups to a place where they can implement them, even though they don't necessarily have that expert on staff? How did we get this to a place where even again,  different representatives from regulatory bodies who may not have a background in alkalinity and carbonate chemistry and all that all that stuff, how do we get them to a place where they can also understand the conversion of principles that they are familiar with, again, some of these existing ecotox principles, and its applicability to things like alkalinity. So, yeah, that's where we're hoping to go next. 

Anna (44:13): Okay, so on that note, I was really, really struck by the amount of examples of regulatory bodies actually already working with model-based concepts in ecotoxicology, not related to marine CDR, of course, but by the amount of sort of parallel learnings that you were able to draw from other areas. Do you have any intuition around where other places in the mCDR world could be, where such parallel industry learnings or so, could be applied? I'm thinking of course, maybe to model-based quantification efforts. Did anything stand out to you in this process where you thought, oh, this should really be shared with the quantification folks? 

Kristi (44:58): So I can take a first stab at that and then Grace, maybe if you have a different perspective, because you're looking at very different things than me. So I just think that the problems or the concerns associated with mCDR, it's such an interdisciplinary thing. You are operating at the intersection of geochemistry and ocean sciences and ecotox and it’s just a lot. It's a lot of stuff to bring these to bear, and it's wholly logical that no single field has all the answers for these projects. So I think borrowing proven concepts from these other fields, as we've kind of been saying, avoids unnecessary reinvention and the pitfalls that come with inventing something, right? That is probably not the most direct path up the mountain. And I think that FEMM is a good illustration of that in action. Now we still have inventing, adapting, we have to do to bring this to what we would like to see eventually, but we are now standing on the shoulders of several decades worth of ecotox work that we didn't have to do. And hopefully that can inform…we can avoid those pitfalls by mirroring that process in the things we need to adapt. Now it's really hard to foresee, I think, from once that type of thing will arise, but I think any discipline that operates kind of adjacent to the mCDR space, I think it's worth really trying to create those opportunities and foster dialogues and just get other sets of eyes on things, because you don't know what you don’t know is always the trick of it, right? Somebody else may easily be like, Oh, have you thought about it can be really beneficial. 

Anna (46:44) One thing that really stood out to me was, in that second section where you present and review all of the different models. I mean first of all, there's so many models to be used, but what stood out for me was how often you were, for example, saying that these models have been used in other contexts and the Environmental Protection Agency, for criteria for what makes this model fit for purpose, or what model performance criteria have to be met, how validation should work. And that's top of mind for me right now in terms of quantification and something that's been really hard to wrestle our heads around, I think, with respect to quantification. And we don't have to go into the details of that, but it got me thinking about ways to look to these worlds because as Grace mentioned at the top, we don't always have to reinvent the wheel and that is something we fall into quite often. One last sort of nerding out question in this regard though, since you also talked about software development, maybe this is totally in the wrong direction, but one of the models that you outlined or described was dynamic fate and transport models, which to me sounded quite similar to when in the quantification world, we talk about near field models or trying to simulate how the alkalinity  or the decarbonized water dilutes away from the source of addition. But so far I've only really heard those models in the context of trying to simulate the transport in order to simulate air-sea gas exchange. Are they the same models used for ecotox? And if so, are they even used in sort of unison right now? Or if not, is there a pathway to do that and couple these approaches for field trial planning for quantification in the future? 

Grace (48:37): I'll take a first stab at that answer because I think it's a funny anecdote. So when we first started working on FEMM and Kristi was writing the first draft of that section on the fate and transport models, we went around and around in circles where I was like, what are these models? Like, I need to understand this in real terms. And basically where we came out on the other end of that was, for example, one of the other projects Hourglass has going on is we're building what we geochemists call a sediment reaction transport model, which predicts alkalinity fluxes from sediments. And we were talking about, Kristi was talking about these fate and transport models, and I was like, is this like a radio project? Is that the same model? And we were just talking past each other for weeks as we were trying to understand whether we were even talking about the same model and like surprise the answer is yes. It turns out that the model we were building at Hourglass is exactly the kind of model that you would use to do this refinement in FEMM, we just had different names for it right? And so this is really well exemplified, really nicely exemplified. 

Anna (49:52): In the same organization. Now extrapolate that to hundreds…

Grace (49:59): If that doesn't tell you about the nature of interdisciplinary science, I don't know what that is. 

Will (50:06): Right? In a bunch of different languages 

Grace (50:08): In a bunch of different languages. So the short answer is, yes, the kinds of models that we are already building in mCDR for other purposes are exactly the kinds of models you would want to use to do refinement around your exposures. 

Anna (50:28): We'll have a follow-up podcast on that. 

Wil (50:33): Early in the document, you outline certain steps along the FEMM pathway that might not be ready for prime time, can't be implemented yet due to the emerging nature of the mCDR field. Can you say more about where this might be the case and how the framework still remains important and usable even in these early stage efforts that we find ourselves in? 

Kristi (51:02): So I think that really the next step, the main thing we want to get at is adapting these approaches to the carbonate chemistry system. I think there is potential to do that and  during and subsequent to the public comment period, we’ve had opportunities to have a number of conversations with people who work closer to that space. And we're gonna keep ironing out the details of that, but it is a significant lift, but I think we're really getting somewhere there. I think that will widen things up quite a bit. But I think now, for the things that it's ready for, the main thing is getting it in a format where people can use it. It doesn't matter how right it is, if nobody can figure out how to turn the thing on. I think that goes back to the web tool that Grace was talking about earlier. I think directing people to resources, but also building resources that are intuitive for the space, and present things in the context of what we're trying to apply as part of that. 

Wil (52:12): And how will we know beyond getting everybody up to speed and actually starting to use the tool, how would we know when it is ready for use, right? Can it be operationalized? If there's some things that have high levels of uncertainty and mean that we can't draw meaningful conclusions yet, how will we know when we've reached that point? 

Kristi (52:38): I think that it depends very much on the type of uncertainty you're talking about. If we're talking uncertainty around an actual model fit, because we've made a real point of including things in here that have at least some regulatory implementations. There are generally guidelines that I think we can really just borrow from because they've already been hashed out to evaluate the NISA fit criteria and things like that, they have those kinds of rubrics in terms of actually developing methodologies. I think that is  an inherently bigger lift. So I think as we've taken a first kind of dip into the pool with the case studies, showing people what that looks like when we're playing around with this, but calling out the shortcomings, calling out uncertainties, I think that's going to continue to be an iterative process and we can theorize. We'll play it through, but I think that in continuing to be public and get that feedback and show the validations, show the pitfalls, and just have it be an open dialogue. I am quite proud of the fact that all of the public comments are available, going to be available with the document so you can see what wasn't clear to people or where people let us know we could do something better and the document changed a lot in response to that. I think it's stronger and I think it's just going to be iterative. 

Grace (54:09) Yeah, I was going to say,, the biggest thing  in terms of making this document fully actionable is the carbonate chemistry work. As Kristi said, this newest version, the final version that we're releasing, Kristi and many other people, I have to say that she worked with during the public consultation period,  made huge strides in developing the framework that could be applied to carbonate chemistry. FEMM will come out next week and then like the week after this is where Kristi is going with her time. This is a top priority for her and many other people in the field, so this is something that I am very optimistic is actually a near term problem that we'll be able to solve, right? This is not way off in the future. This is 2026. And so I think that's an amazing step that's going to come quickly and will hugely leap the field forward. And in the interim, I think the benefit of FEMM is that as people continue to do ecotox work for their particular project, now FEMM gives them the guidance in terms of smart experimental design to make those experimental outputs applicable to, for example, that carbonate chemistry modeling that will be coming shortly. So I think this is something that everyone should be reading right now in terms of like, how can we use the work that we're already doing and make smart decisions and ensure that when we do do experiments we're getting the maximum benefit out of it, so that we can integrate those results into these kind of modeling approaches later this year?

56:01 - Regulatory Pathways, Tools, and Future Research Priorities

Wil (56:01): So you kind of contemplate that there'll be ongoing dialogue with key stakeholders as this process moves forward. 

Grace (56:11): Yeah, absolutely. I mean, we are building this tool for other people to use. Of course, with any smart product development, you're talking to your users along the way to make sure that you're creating something that's going to fit their needs. That's what we've been doing. There are a lot of quote unquote users, right? There are a lot of stakeholders who should and have provided input and we're going keep that conversation going. 

Wil (56:38): Well, that's a good segue to the last question that I have, which is, one of the key stakeholders is obviously, or at least potentially, the regulatory community, right? And we've talked a bit about some of the regulatory implications of what's being done. I'm curious how you contemplate that the regulatory environment may take this up, and have you seen some early stage efforts to do so? 

Grace (57:06): The public comment period was anonymous, but what I can share is we did get feedback from government representatives during that process. So that is obviously an important first step in terms of making sure that this is a tool that is useful for that stakeholder group. 

Kristi (57:33): Yeah, and I think that because the sourcing of the approaches owes so much to regulatory frameworks, actually, that's a lot of what we're citing all the time. A lot of this stuff is really in the regulatory lexicon already. That's what's being used a lot of the time to arrive at regulatory thresholds anyhow. So I think that eases both the interpretability of it, of the thing that's being produced for someone who's trying to make a regulatory decision, and just getting a common language going a little bit, or thinking about this. It's nice in that way. This document does owe quite a bit to regulatory guidance documents. 

Anna (58:19): I think the document is awesome. I read it on Friday, I had a lot of other things to do, but I ended up reading the whole document. Now, I said this in our introduction, it's kind of like a common challenge right now in various areas of Marine CDR, maybe even CDR in general, that we're in this awkward phase where we don't really have enough data yet to draw super definitive conclusions. But enough is happening to require some level of guard railing and frameworks and common language and guidance, et cetera. So this is obviously one of those projects and tools that are going to be, I think, massively helpful and helping gather new data and adapting to that new data and putting that data into context. So kudos. We have so many other topics for which we have to do a similar type of work. And because we have so much to do in this world of marine CDR and OAE and docs, et cetera, wondering if you want to close on the research priorities that you identified and also end with in your document, and all those projects that are going to refine the framework and deliver the data that we need. 

Grace (59:35): Kristi, do you want to speak to what you want to be doing for the next three years of your life, if everything goes according to plan? 

Kristi (59:42): I would say, I think probably the community has more data than it thinks it does in some ways. I think that because we don't have this common lexicon, we're not looking at it in the same way. We're missing each other, we're talking by each other, like me and Grace were talking about earlier. I have a strong suspicion, but grounded suspicion, that there are quite a few data sets that can be useful in developing these approaches, I think both in terms of the carbonate chemistry thing in isolation and also just the mixture elements that come into this and all the complexity that goes with that. I think a lot of the lab data and data from field trials, things that are being enacted right now in keeping with MRV protocols, I think those have more utility than people may give them credit for up top because I know they weren't designed for this, that doesn't mean they can't be useful. Of course, there's going to be gaps, and I think kind of the first labor is finding out what we already have so we know what we need to have. So I think it's gonna be a lot of literature review, a lot more conversations, and identifying where the pain points really are in bringing these things forward. Because I think there's more out there than you might see at a first glance. 

Wil (1:01:07): That seems like a good inspiring place to stop. I'd like to once again thank Grace and Kristi and as always thanks to our listeners. If you enjoyed this episode, please leave a comment or review and share this episode with others. If you want to suggest a specific topic for an upcoming pod, please feel free to reach out to us either through LinkedIn or via our email address, which is plansea@carbontosea.org. And with that, Anna and I say thank you and hear you next time. 

Anna (1:01:41): Thank you for joining us. That was great. 

Grace: Thank you.

Kristi:  Thank you