 Thank you Elizabeth. Good morning, everyone, both here in the room and so forth. Thank you, Lonnie, for that reminder. So good morning and really good morning. Welcome. We appreciate everybody that could join us here in the room and and online today. We think we have a very interesting and should be insightful meeting today. So as many of you know, this committee is one of several standing committees under the Board of Earth Science and Resources and this committee is the locus of all the activities of the academies related to energy and mineral resources and that includes ranges from the full life cycle of their exploration, characterization, production and recovery use, recycling, reclamation, regulation and stewardship. So the whole range. So the committee has existed for about 30 years and has been fortunate to have the support from several federal agencies and those include the DOE Geothermal Technologies Office, the DOE Office of Fossil Energy and that's both the Oil and Gas and the Coal Technologies Office and the National Institute for Occupational Safety and Health Mining Program. As a non-profit organization with a congressional academies, a congressional charter, the academy serves in a unique function in the nation's dialogue which brings expert scientific and technical information and advice into the policy and decision-making arena. In addition to the annual support of the committee's work from those agencies I mentioned, the committee couldn't function without our members who are seated here at the table with me and they're all experts that are selected by the national academies to serve and we all, committee members all serve as a volunteer in addition to their other normal routine activities and businesses. So for this meeting today, so the committee works together on planning these meetings and today I'd particularly like to acknowledge our three moderators today, David Spears, Bridget Ailing and Joel Renner who have kindly spent many hours planning the meeting and have agreed to serve as our moderators today. So the committee works with the national academy staff to identify and main an awareness of specific energy and mineral resources issues, whether addressing the subsurface characterization and management of resources or surface manifestations and legacy of their extraction or training and education of the workforce associated with the resources. The committee hosts, we typically have two meetings a year and in addition we try to then build off those meetings with some webinars or other activities and then we also work on building other projects such workshops around tables or national academies, consensus studies. The meeting today grew out of interest the committee had to focus on some advances and challenges within the geothermal energy sector. And in discussions with the geothermal technology, the DOEs checked geothermal to office, we worked on the scoping of the meeting and agreed that the strength of our committee could offer was to convene a meeting around several challenges that different industry sectors, geothermal, oil and gas and mining have in common when working to characterize and manage the subsurface. Three things we identified will each be addressed by a separate panel of experts today who will present their perspectives from across these three sectors with the idea of sharing common experience, novel solutions and a view toward future advances in areas of progress. Sue Ham who directs DOEs, geothermal technology office will be kicking off the meeting with her keynote remarks about geothermal energy and research, giving us a baseline to see the value across the sectorial and the sectional conversations. I would and these meetings are one of the richness they have is the interaction and input from everyone who participates in the meeting today. So everybody that's here should be considered themselves not an observer but a participant and so the panels are designed to leave a very significant amount of time for discussion and input in the meeting. So I mentioned the national candidates function in terms of groups like this committee which is a convening body and our ad hoc committees that conduct consensus studies like some of you, there are some copies of the consensus studies out in the hallway and but the academy works as a service to the nation on with these and dates back to 1863 and it also recognizes the nation's top engineers, scientists and medical health professionals with the National Academy of Science, National Academy of Engineering and the National Academy of Medicine. We're fortunate that this meeting was scheduled to coincide with the end of the meeting of the National Academy of Engineers here in Washington and a number of members of the NAE and specifically the energy resources engineering section of the NAE have stayed an additional day so they could be with us and participate in our meeting. So I'd like to recognize all of the NAE members that are in attendance today and some of which are on our committee and on our board but so everybody that's an NAE member if you just raise your hand, welcome, special welcome to the NAE members. And co-chairing this meeting with me is the chair of NAE section 11 which is the energy resources engineering section, Dr. Lynn Orr and many of you know Lynn was at the Department of Energy a few years ago and I actually was at the Department of Energy a few years before that and so I think that one of the things and this is that I think is evidenced here today that people don't always have the perception with what you see in the media today but that two members of two different administrations that were controlled by two different political parties can sit up here together and have a very constructive and useful but I was going to add that that Elizabeth must have been worried. So with that I'll turn this over to Lynn to make some additional opening comments. I'm not sure I can follow that. So it is a pleasure to be here and as far as I can tell I must have at least three hats here. One is the chair of section 11. That's a temporary position. We rotate the leadership of section 11, the energy resources engineering and it's largely connected with the subsurface and the earth sciences and the way fluids move around in the subsurface and the way we use the subservice in lots of ways. We do have a responsibility to serve what Jim called a unique function which is to provide advice to the nation and the way that's done is through the National Research Council in a variety of ways. I personally have participated I think in most of the elements of that process, whether it comes from participating in studies or chairing a few of them, being a reviewer on many of the reports and then at the Department of Energy being on the other side of the process of receiving the reports and I guess that job was under secretary of science and energy at DOE and there were things that could be done under the sort of ages of the NRC that were hard for us to do within DOE and that provided an independent view of what went on and the potential research agendas for the nation that was extremely valuable. I'll give you one sort of humorous example. We had a deep interest in the whole energy efficiency area and at least part of that has to do with how we pesky humans make decisions about how we use energy but the Congress in its wisdom had prohibited us from doing anything that any sort of social science associated with this because I won't speculate on motives but in any case they thought that we weren't the right folks to do that. So of course we didn't do any social science. We did some decision science which was to try to understand how humans make decisions about energy but we were able to partially support a study of all of those aspects of energy use that could be done by the National Research Council in a way that was complied with what Congress wanted us to do and helped us think about how to organize the energy efficiency programs. That was one of many, many places where the advice from the independent groups that the NRC put together was valuable to those of us who were in office. So with that I'll just say that I've also spent most of my career thinking about how fluids move around in the subsurface so I'm personally interested in practically everything that we'll talk about today. I'll say thank you to all of you for participating in this program. Thank you to the members of section 11 of the National Academy of Engineering. There's about, I don't know, 10 or 11 who doesn't have us here and that's a reflection of our interest in the topic of the day. So thank you all for being here and let's get started. Okay. Thank you, Lynn. So with that we'll get underway and let me introduce Sue Ham. Sue is the Office Director of the Geothermal Technology Office which is in the Energy Efficiency and Renewable Energy Office at the Department of Energy. Before DOE she's had been in significant positions at the National Science Foundation and the Department of Homeland Security. So with that I'll turn it over to Sue. Her more complete and detailed bio is available in your packet. Good morning everyone. Is this loud enough? Okay. So I just want to say how thrilled I am to be here. Thank you for inviting me. Subsurface is obviously absolutely critical to geothermal and geothermal research. This is a fantastic opportunity for us to hear from all of you and really learn from others around. I'm personally excited to hear the panels, the questions, et cetera. So I just want to point out that we actually have two staff members from our office here. We have Zach Frone who is an ORISE fellow in our office and Doug Blankenship back there. Doug, can you wave? Thank you. Who is on loan to us from San Diego. We also have a number of our PIs that we fund in the audience and I'm sure they're going to identify themselves as the day goes on. I don't want to name them all. Okay. So I'm going to take, I don't know, 20 minutes or so to go through our program and why we're doing what we're doing. But with a long presentation it's really hard to know what the speaker really wants you to remember. So I want to start up front with, I'm going to stop in a couple of places and talk about sort of future R&D directions and that's really what I want this group to think about. We're going to go through a lot of stuff, but that's really the fundamental pieces that I want you to think about throughout the day and take home with you. Okay. So here are the four things I'm going to talk about. Why do we care about geothermal? I'm sure many in the room already know that. Let's just go over that. Why is the geothermal technology office exist? What is our mission? What are we doing? I'm going to spend some time talking about the results of our recent, relatively recent GeoVision report which came out in May. And then I'm going to go through some of our programs and talk just, it's not going to be all of our programs, but some of the major ones, maybe. All right. What am I doing around Remy? Towards you. Oh, okay. So why geothermal? So beneath our feet lies this vast resource, right? And it's relatively untapped. Geothermal is always on. It's secure and flexible. It provides base load power. It's also, you can also, it's also dispatchable in addition to being base load, creates thousands of energy sector jobs. And when you combine both power generation and direct use, heating and cooling, it's an everywhere solution. So that's why geothermal is really important for the United States. Trying one more time. All right, Remy, I give up. There we go. Thank you. All right. I'm just going to say next. We said we weren't going to do that, but I guess we're going to. So this is a map at about 10 kilometers depth of heat across the United States. It does not include Alaska and Hawaii. And what you can see is that for the most part, when you're talking about the highest of temperatures, that's the purple, you're really looking in the western half of the United States. There is, however, heat at depth everywhere. And that's what we're going to be looking to tap. Current installed capacity is actually an old slide. It's somewhere between 3.9 and 4 gigawatts now. Undiscovered hydrothermal. So we're going to talk a little bit about this when you have a hydrothermal system. That's where you have heat and permeability in the subsurface and a geothermal brine running around the subsurface. You have the complete system already for you. That's a hydrothermal system. If you have something like a geyser or a fumarole or hot springs, then you know that that's there. But if they don't have a surface expression, then that's what's called an undiscovered or hidden hydrothermal system. And the USGS has estimated that there's about 30 gigawatts of undiscovered hydrothermal resources in the United States. So it's left to be tapped. We're going to talk about enhanced geothermal systems. So I talked about the three pieces that make up a hydrothermal system. If you are missing either permeability in the subsurface or a geothermal brine in the subsurface or both, you can add those. And that's what's called an enhanced geothermal system. And the western USGS potential is 500 gigawatts. If you look at the entire US, you're up at 2.3 terawatts. Okay, can I try this again? All right, so based on that enormous resource, what is a geothermal technology office doing and why? So if you look at where geothermal sits, a couple of slides are going to be very government-like, and I somewhat apologize for that, but that's how we think. So geothermal sits inside the Office of Energy Efficiency and Renewable Energy. We sit inside the deputy assistant secretary's sector for renewable power. Within renewable power, there's geothermal, water, wind energy, solar energy, and then grid modernization. So we're part of that. And the three main priorities for EERE right now are energy affordability, energy integration, and energy storage. With the GTO programs, we work on all three, but really our focus tends to be on affordability and storage. Great, so let's look at what it takes. This is when you're talking about geothermal power. So right now I'm not talking about direct use as much. Really focusing on geothermal power generation and to get from pre-survey, which is all the way on the left, to operations and maintenance of a power plant on the right, what we're looking at here is risks and cost. So the color, the dark, the heavy black line with the color underneath is really looking at the risk. And you can see that the risk is very high in the upfront stages because you're thinking about the subsurface, and we don't know. It's really hard to see in the subsurface and understand what's there. And so the risk tends to be very high. Where our costs tend to increase is when you get to drilling, can you hit next? When you get to drilling in the test drilling and in the drilling phases, that's where our costs tend to skyrocket. And so what we do is we are looking to support research in these key areas, drilling, success probability, new technologies. What we're really looking to do is reduce the cost and the risk to help, it's really looking to reduce that early stage and make it accessible for industry to pick up and get these plants online. Okay. Wouldn't be a government presentation if we didn't actually have a budget slide. So one of the reasons that I put this in in particular is to look at the very large growth that we've seen since FY 15, we started about $55 million. Next. So for FY 20, the President's request was $28 million. The house mark next is $90 million. And the Senate mark is a whopping $115 million. If you look at the Senate mark at $115 compared to the FY 15 number of $55 million, that's a pretty large potential increase. We anticipate that the actual FY 20 budget will come out somewhere between $90 and $115, just still going to be a pretty big increase. That gives us a lot of room to do a lot of really great research. Next. Okay. The Office of Energy Efficiency and Renewable Energy has put out a number of vision studies. There was a sunshot vision, a wind vision, a hydro vision. And in the geothermal office, we really wanted to follow on. We really wanted to follow the other offices. We saw a need for a study of this magnitude and the scale to more accurately assess the impact of geothermal in America's energy future. And we really wanted to include a comprehensive look at how innovation and innovative technologies would impact it and how non-electric applications, the direct use would also impact that picture. And so what we did is we looked at, we sort of started with a fundamental question. If everything went right, given where things are now, given what the barriers to deployment are, given what exists now in technology and what could exist in technology, what level of deployment would be possible by 2050? And given that level of deployment, what would be the economic impacts, what would be the environmental impacts of those deployment levels on the United States? Next. So the geovision was not just our office. There were about 120 to 150 people involved in the creation of the geovision document. It's a product of years of rigorous research and analysis with contributions from a broad range of participants. And we really had industry, academia, national laboratories, federal agencies. You name it, we had them involve public utilities. And this was a really great document. And I highly recommend that anyone take a look at least if you don't want to read all 200 pages of it and I can understand why someone would not want to read all 200 pages of it. Read the executive summary. You can find it. I'm pretty sure we have the link in just a few slides. Go ahead. Next slide. So the main results from the geovision study showed, this is what the analysis showed, that optimized permitting, so just reminding the permitting, could cut the timeline. Remember I talked about those stages of our plant that you could cut the timeline in half with streamlined permitting and that would lead to a doubling of geothermal development to 13 gigawatts by about 2050 versus the business as usual. So what that says is that with absolutely no increases in technology, no advances in technology, nothing. If we just work on the streamlined permitting, we would double the size of the industry. Direct use could increase by orders of magnitude as well. There's currently 21 geothermal district heating installations in the United States and what the analysis shows is that with the appropriate advances, let's just leave it that way, we could get up to 17,500. In addition for backyard geothermal, which is what a lot of people think about when you're, so ground source heat pumps, geothermal heat pumps, it's what people think of when you say geothermal. Currently two million installations in the United States and the analysis shows you could go to 28 million by 2050. That's a huge number. The really, the big money number though is that with advanced technologies with technology improvements and the bringing online of enhanced geothermal systems as I talked about before where you're adding either water or permeability to the subsurface, definitely it could reach 60 gigawatts by 2050. So that's a pretty impressive number. So what do you need in order to get to that 60 gigawatts? Next slide, please. Technology innovation is therefore going to be absolutely essential. It improves our understanding of the subsurface, helps us reduce risk, and accelerates growth of domestic geothermal power. Next slide. So at the last chapter of the GeoVision document is our GeoVision roadmap. And what we looked at, remember I said that we started with okay if everything went right, well then the question becomes what all has to become right? And we have these four action areas, I'm just going to read them to you because I know they're a little hard to see on the slide. Action area one is really, that's really the advances in exploration and achieve key technology advancements. Action area two is optimizing regulatory processes. That's really where you're going to get that big jump to 13 gigawatts. Action area three, optimize revenue and market structures. Action area four, improve collaboration, education, and outreach. All of these four action areas are really aimed at getting to these three key objectives. Increase access to geothermal resources, reduce costs, and improve economics for geothermal projects, and improve education and outreach. So I said I was going to stop for just a couple of seconds in two places during the talk. This is one of them. So I would like to call your attention again to this document. This is the roadmap chapter five of the GeoVision document. It's, I don't know, 20 or 30 pages. This is really going to help define where the geothermal industry and where the Geothermal Technology Office needs to go over the next five to 10 years. So if you're looking for where are the big research gains to have, I would recommend taking a look at this. You can find the GeoVision document at www.energy.gov. So again, I really highly recommend the executive summary. You can skip the letter from the director. I pretty much said everything you need to hear from there. But the executive summary and the roadmap, which is chapter five. Next slide. I'm going to quickly go through the four action areas and just point out a number of the programs in the Geothermal Technology Office where we are already doing some of the work that you see in these action areas. And then after that, we're going to go through a couple of them in detail. So the first is the Frontier Observatory for Research in Geothermal Energy. We've got Play Fairway, Jim Falz is here. Efficient drilling for Geothermal Energy Edge. We've got CoLab, Waterless Stimulation, Zonal Isolation, Subsurface Stress and Law Circulation and Machine Learning Exploration. So a lot of these, this is really where the GTO intersects with the GeoVision. We do a little bit in action areas two, three and four, and we're going to see some of that. But this is really the key technology advances and that's really where our emphasis is. Next slide. Action area two is optimized in those regulatory processes. We're working with the Bureau of Land Management Department of Interior to try to figure out where can we help make some of that streamlining happen. We're working with Forest Service. A lot of the land that has geothermal resources on it is either BLM land or Forest Service land. We're looking to do some collaboration with state and local governments. In fact, I'm actually talking with the Connecticut State Government next week, I think, that we're going to do a webinar looking at GeoVision and the Northeast, right? Because if you think back to what that map looked like, all the high temperature was over in the west. And so how do you really think about using geothermal resources in the Northeast? And that can be a challenge. It's doable, but it is a challenge. And then we're really looking to collaborate with the Department of Defense. The military installations may be on some really interesting lands in terms of geothermal resources. And one of the things that the military really wants to do is they want to be able to have their own power if and when they need it. And so we're talking to them about how can we collaborate, how can we make sure that we are able to harness geothermal resources in a way that makes sense for military installations. Next slide. Optimizing revenue and market structures, one of the things that is just absolutely critically important when you think about geothermal. Geothermal is not the cheapest renewable energy out there. We all know that, right? Solar is at what? Two or three cents per kilowatt hour. And geothermal is at, you know, seven when you're talking about a hydrothermal system. But if you're talking about EGS, it's significantly more. So you really need to think about what is it that geothermal brings to the table. And when we talked about it at the very first slide, we talked about how it's secure and flexible and reliable, base load, et cetera. All those things have value. It can also help stabilize the grid. Things like that. And all those things have value. But right now, when people think about renewable power, they only think in terms of a levelized cost of electricity. So we're working with others in EERE on something called Beyond LCOE, which is looking at how do you value some of these other elements, these other benefits that any one of the power technologies could bring to bear. We look at critical materials, right? So some geothermal brines are pretty heavy in lithium or other critical materials that we want to try and harness. How do you bring those out in an economically viable way? We have something called Advanced Energy Storage Initiative. I'm going to have a slide on that a little later. So I'll move on. Okay, next slide. Obviously, we're identifying plans for regular updates on the geovision roadmap. We're working on a multi-year program plan, which is the next step after this roadmap, right? The roadmap is not specific to the Geothermal Technologies Office. It's really for the entire geothermal industry stakeholder base. And so our multi-year program plan, which Doug is working very actively on, is looking at what are the next steps for GTO in these areas. We also have a number of international collaborations. We have one with New Zealand. We have the Collaboration Geothermica, which is with the European Union Nations, and that is a lab call on our side to see how can our lab folks collaborate with folks in the European Union to share information, share data, share facilities, those kinds of things. I already talked about the military bases. Next slide. Okay, so I'm going to go a little bit more into detail in some of our programs. Next slide, please. Our biggest program is this Frontier Observatory for Research in Geothermal Energy. Can I get a show of hands? Who's heard of FORGE? Okay, so about two-thirds. Great. So the idea behind FORGE is that we're looking to design and test a breakthrough approach developing large-scale, economically sustainable EGS reservoir. So what does that actually mean? I talked in the beginning about hydrothermal being where these three things already exist, heat, water permeability. When you don't have all three of them, you need to create them. But it's actually not that simple. You just can't just go down to the subsurface, create some fractures, throw in some water, so the idea behind FORGE is we started with a number of potential sites. We're down to a final site. What can we do to make sure that we could make the technologies? How can we test various technologies and get a large EGS reservoir that is maintainable, sustainable over time? And so right now we are in the process of just about to launch phase three. Phases one and two are really getting the site ready to go, making sure we had the right site. And phase three is a five-year phase three. And on each year about half of the money will go for the team that is operating the site for characterization and drilling. And the other half is going to go for solicitations that go out to the entire geothermal stakeholder community for R&D. Next. Did somebody have a question or something? Okay. I'll just keep going. So one thing that's very important is understanding where we want to go with the Forge R&D. And so we did a Forge roadmap that was stakeholder based. The Science and Technology Policy Institute which is an FFRDC run out of the Institute for Defense Analysis supports the White House but also does many other studies with other parts of the government. And we worked with STIPI to figure out where do we want to go with Forge R&D. So the roadmap is really a detailed outline of the recommended R&D activities for the site in Utah. It's a plan for understanding the key mechanisms controlling EGS success. And it had input from EGS experts across academia, industry, government and national labs. And really what we were looking for were what are the critical research areas that would allow us to create an EGS reservoir and then repeat it across the country. So the three areas are stimulation planning and design, fracture control and reservoir management. And all three of these obviously are subsurface and really very critical to what we're going to be talking about today. And the website for the roadmap is listed here. It's another great thing to read. And I said I wanted to stop two places to talk about future R&D directions. This is the second of those places. So first is roadmap, second is the Forge roadmap. Geovision roadmap and the Forge roadmap. So I highly recommend that as you get a chance to take a look at those two documents. Okay, next slide. If you've got really expensive wells drilled on the Forge site and you want to test technologies and you don't know what's going to happen, one of your concerns is that you don't necessarily want to test dangerous technologies and I mean dangerous in the sense that they might break a well on your super expensive Forge site. And so what we have is something called wells of opportunity. This has gotten a lot of support from Congress already. And the idea here is where can we use abandoned wells, idle wells, less critical wells to test some of these technologies. And what you see here on this slide is going from the left, which are the least invasive to the right, which are the most invasive, highest risk types of activities that you could try in a well. Right now we have a request for information out looking for input on wells and sites that could be put to use for this. So as you're thinking, especially oil and gas folks, think about what kinds of wells might be useful for this. We will have money to bring these wells up to the standards that we might need. And the RFI is listed over here on the right. You do have to go into the exchange to actually give a response to this RFI and we hope to hear from any of you. Next slide. EGS CoLab. So the EGS CoLab is in a mine in South Dakota. How many people have heard of a home stake mine in South Dakota? Biggest, great, almost everybody. Biggest gold mine in the United States at one point, and now being used mainly for physics experiments. But they let us geothermal folks come in and drill some wells. So what we've done, this is a the idea behind EGS CoLab was to try things at a smaller range and test some of the things that we would be eventually doing at FORGE. So with these and Doug again is a co-PI on this. If you the idea is to go into the drift, the well drift with, I don't know, 50 or 100 meters and try to observe so you've got boreholes and they're full of monitoring. Try to observe fractures happening in real time. And that's what we were able to do with experiment one. We're just about to start experiment two. So experiment one was really just a straight fracture and experiment two will include some shear in it as well and the experiment two should start sometime Doug next in the next few months. Okay, thanks Doug. Alright, next slide. Play fair way analysis. So Jim can speak to this one exceedingly well. The play fair way analysis is an idea we adapted from oil and gas where you're taking various data sets and combining them to see where would it be more likely that you might find a geothermal resource. And we started with 11 sites, ended up doing drilling at these final phase three confirmation drilling and what we're finding is that these are working out pretty well. That we're trying to find so again they were looking for these hidden systems. I don't think I said that. So hidden systems, combining data sets like gravity and magnetic all sorts of stuff to try to see where might you find a resource. And what we've been able to find with confirmation drilling in phase three is that you actually do get elevated temperatures where some of these combined data sets are suggesting that you should see them. So it's very exciting. Okay, next slide. I mentioned when we were talking about about getting from the beginning to the end that drilling is really expensive. So obviously one of the things we want to do is put a lot of money into drilling research. Drilling can account for up to 50% of that cost to get to geothermal full plant. So we are now funding 10 projects for a total of $14.5 million and we're looking to reduce delays in drilling operations. We're looking for innovative drilling technologies and we also want to accelerate technology transfer from lab to the real world. So these are getting started and you have the awardees down here on the bottom right and we'll see. I'll have more information on this in six months to a year as to where we are with these. Next slide please. Machine learning for geothermal energy is a super exciting area for us. Machine learning is a huge buzzword for geothermal engineering. Drilling in oil and gas has been using machine learning for years and we really want to try and make this work for geothermal energy. One of the issues for geothermal is that we don't have the data. We just don't have as much subsurface data as the oil and gas industry does. And so we're trying to identify what would we actually need to understand? What would be the data acquisition targets? What could be some new signatures for detecting those hidden geothermal systems? We want to minimize power production, understanding what's going on at your plant and your reservoir monitoring and then we want to improve prediction and detection of trouble events, right? You want to be able to use a machine to see things that we as humans don't see as quickly or as well. So another 10 awards for $5.5 million in funding. These are actually small awards. They're phase one we will be doing. We'll be down selecting to a couple of them for a phase two and putting additional money into it. Something that's really important, obviously, and those who know the subsurface know this, right, is that you need to understand the state of stress. And that is not a simple thing to understand in the subsurface. That along with lost circulation, so I'm sure most of you know what lost circulation is, but it's the idea where you're trying to... What we're looking at is trying to minimize the impact of lost circulation events where your mud disappears while you're drilling. Because that we want to we want to continue to reduce the drilling costs while improving efficiency. And so we have up to $7 million in funding that's available. This is 2019 funding, and those proposals are now under review. Next slide. I talked about the fact that we were really focused on storage, and this is our assistant secretary, Daniel Simmons, talking about how important storage is, and I'm sure we all agree that storage is critically important. Within the geothermal technologies office, we're really looking at this in three separate ways. We're looking at deep direct use as a way to think about storage, bi-directional energy storage, and flexible generation. And with all of these we have six deep direct use awards that have been going since 2017. We will be putting out another funding opportunity announcement to look at the next stage of those types of awards. And then we have lab projects in bi-directional energy storage and flexible generation, and likely we will have additional awards into the future in this area. This is just a really critical piece to understand how can geothermal contribute to storage of energy. It's going to be a big deal as we move forward. Next slide. One of the things that I promised my team is that everywhere I go I will ask for reviewers. If you are interested in serving as a mayor reviewer for us, please contact us at this web email address. Actually, if anything you want to talk to us about you can use this address. You can always email me, but doe.geothermal at ee.doe.gov is looked at several times a day by our staff to make sure that we are answering anything that comes in from the community. And then our website is energy.gov slash eere slash geothermal. And I just want to say thank you so much for your kind attention during this. And I don't know if I'm taking questions OK. Thank you Sue. So just we will have some time for some questions and discussion. And I would just ask folks to just you can always if you have a card you can put it up so we know or you can, you know, motion that you want to ask a question please use your mics. And we want to make sure that we have questions we'll do we'll get to those as well. And I would also say that there are folks in this audience who are much bigger experts on particular areas than I am so I may be turning to them if anyone asks me a question I don't quite exactly know the answer to. Yes, I'm Christine the Economist. My question is just with 60 gigawatts that's a lot. So it seems like most of this is quite far from the market. What's the plan to get the power to where it would be used? OK. So there's sort of two pieces to that question. The first 60 gigawatts is a lot. Actually what I should have said is that analysis shows that that would be 8.5% of national generation by 2050. So right now geothermal provides 0.4% of national generation so going to 8.5 is quite a lot. OK. So the question of a lot of it is far away and it's very expensive. And so one of the ideas with EGS is can you do EGS in a way that is safe closer to population centers? You're not going to do it right smack in the city we've seen what happens when you do that. You can end up with a lot of induced seismicity. So the US geothermal industry is very careful and the researchers are very careful about following induced seismicity protocols and making sure that we are not putting anyone or anything at risk. So the question is how much transmission costs? Was it a million dollars a mile or something like that? So that does add to the cost of what the eventual LCOE will be. So it's true. But the idea is with EGS to get closer to population centers because you aren't necessarily required to be at the risk zone of the West United States. Right. Well we're demonstrating in the middle of Utah and it's not near anything. But the idea is when you demonstrate there, can you make sure that you don't have induced seismicity issues and then try to move a little closer to population centers? Great question. I had a quick question. You had a figure up there where you had risk on the Y axis and time on the X axis. I was just wondering how you were defining risk in that. Is it cost or? Well really it's how likely are you to hit a resource so that if you're in the exploration stage you don't exactly know, right? So if you're talking about oil and gas and you're looking at seismic reflection you can really see the resource when you're looking at seismic reflection in geothermal you can't see it. And so the risk comes in with how likely is it that you're going to be able to hit a resource exactly where you think it is. And we're trying to bring that down. We're trying to say okay let's narrow it. Jeffrey? I'll repeat your question. Go ahead. Oh there we go. Okay. Jeff Yaris with Case Western. I'm going to answer the first part of the question I think because you just said it which is well placement seems to be important. So how extensive are the subsurface earth models that you're building to reduce the uncertainty around well placement? The Forge project is doing a massive amount of that kind of modeling. We don't actually fund modeling just for random places. And so I think some of the projects do it and some of them don't. That's not an area where we particularly are focused with. It's not an area that we are currently focused on. How do you make those earth models better? So the well placement itself is it done based on very broad parameters or are you trying to You know, let me turn this to Jim. Jim, do you want to talk for just a minute about how you did it in your play fairway? There we go. Yeah. So there's a number of parameters that we take into account in multiple geologic geophysical parameters to model both the surface and the subsurface. And then we try to use the most sophisticated 3D modeling programs and so on such as EarthVision and there's others. And really to and then we we take, Sue mentioned the play fairway program. So that's not only estimating the probability of a resource on the surface but where in the subsurface that resource might be. So actually a play fairway that's three-dimensional. Okay. So I don't know Earth modeling has been a specific focus of a program but many of the programs like the play fairway program have gotten into some pretty sophisticated Earth modeling. If that makes sense. Thank you. There's another question up here. Ken Woodley, Condor Consulting. You mentioned one of your first point, your first point I think was enhanced permitting. Are those places that would be developed irregardless but you've cited a figure of a significant resource available if enhanced permitting worked out? And I just thought well okay is it a question of time because sometimes those we've well in the mining we've seen for pebble for instance. Very difficult getting all the stakeholders lined up and the government is now trying to expedite that but these can be thorny issues that go out for some time and so a guarantee that you're actually going to see this increased resource is sort of like if you get a third of it you're probably doing well and you comment about what you thought that would give us because it was the first point that you mentioned. Right it's a big point because to us it's a somewhat easy fix right it just relies on the government maybe changing some of its rules which exists for a reason, I'm not saying they don't but when you're doing geothermal you go back for NEPA testing you have to go back for NEPA documentation six or seven times in the process of getting to a project if you're doing oil and gas it's something like one of the concepts is if it's very expensive to go back for NEPA there's a lot of time involved and so if we can reduce the number of times that a geothermal project has to go back for NEPA you are then reducing the cost reducing the risk and increasing the probability of getting to a resource is that what you're Yeah no thank you good luck Actually there are some bills in congress that are putting forward categorical exclusions that are similar to oil and gas for geothermal we'll see if they go through but they're there Yes Well kind of a pardon me, Tom Craffer with USGS but before I worked with USGS I was in the state of Alaska and I know of one instance in the state of Alaska and this is kind of a continuation of the permitting theme where there was a company that was seeking to drill some geothermal exploration wells near Mount Spur which is an active volcano in Alaska and because of the lack of infrastructure they were having to rely upon a very expensive helicopter supported drilling operation and eventually they gave up because they were being required to deal with the drill cuttings as hazardous waste and they had to collect and bag and transport to an approved disposal facility all of those drill cuttings and to kill the project so I think there's maybe another component to the permitting there that ought to be considered and that's these kinds of issues in relation to the drilling Actually that really helps bring up one point which is that when you talk to BLM one of the concerns is that depending on whether you're in Alaska or Nevada or Illinois or depending on really where you are you can get a very different level of expertise in the BLM office and I'm assuming that BLM is working very hard to address so that if X is required in Nevada the same X would be required in Alaska and they're really trying to get to that In this case it wasn't on BLM land so I think there's the EPA can have a role in this as well That's very fair Yeah I guess I didn't catch how this EGS differentiates from previous efforts towards EGS Could you comment on that? Sure so there were a number of EGS demonstrations that were done in 2014 and those were sort of at the 5 megawatt or less level and what we're really looking for is more at the 50 megawatt level so this is bigger reservoir bigger team a lot more money right so that when you are talking about $35 million a year for 7 years approximately you can really get much farther than you could with a few million dollars and a random we've got a lot of experts working on this we have not only the team at Utah we also have the science and technology analysis team which is made up of a whole bunch of experts that are helping sort of guide that science as well and there's just a lot more concentration of the entire geothermal community on this one project to make it a success we have a lot of interest from some of our members on eastern demonstrations of EGS so I'm wondering what do you see as the timeframe for that and are you doing any work specifically on eastern areas like the West Virginia hotspot or anything So in terms of the eastern part of the United States I think where we're going to see enhanced geothermal systems technologies show up is in deep direct use type areas because in order to make deep direct use viable even on the east coast you're going to need some of those technologies as well and I would say the timeline for that we have of our you mentioned West Virginia we also have an upstate New York project on deep direct use so we have two of them already looking at those and sort of the feasibility and the question really becomes is it feasible and that's not even necessarily a technology question but more of an economic question I mean it combines right because if it's cheaper to get down and do these technologies that's going to bring down the cost and make it more feasible those feasibility studies should be coming in this fall and we'll know a little bit more then alright I've hit my time thank you all so much for your attention I really appreciate it thank you Susan so what we're going to do now we're going to transition directly into our first panel and we will so I'll turn it over to David Spears and he's going to go until 11am some of you have a little air on your agenda but we have a break at 11 and David I'll let you take it over okay good morning I'm David Spears I'm a member of the committee on our three sources and this is our first panel it's on managing uncertainty in the subsurface and we have three eminent panelists here on my far right is Ken witherley who is a consultant nowadays has his own consulting company Condor consulting but Ken has a long history in the mining industry he was worked for BHP for many years by the way the complete bios are in your handout beginning on page five but Ken worked for many years for BHP Minerals as a geophysicist and since then he's had a technology based consulting company so we have strong expertise from the mineral sector and then in the center we've got Colin Williams who's currently at the United States Geological Survey in Menlo Park where he directs the geology minerals energy and geophysics science center otherwise known as GMEG but before that you were chief of geothermal resource investigations the earthquake science center almost also in Menlo Park so Colin has a long and strong background in geothermal energy and then immediately to my right is Jeffrey Yaras who has currently is currently at Case Western Reserve University but has a long history in the oil and gas industry over 42 years with companies such as Amaco and Marathon and now Ken is basically an earth modeling specialist and so I think we've got a good round panel here with representation from oil and gas geothermal and mining and I think we're going to start with Colin right? You ready? So I'll get started while the slides are coming up try to save a little time I should mention that I promised Elizabeth I would have 10 slides I actually have 15 so we're at 50% uncertainty already but we're trying to bring that down and I'm going to talk fast anyway one of those just a title slide so it doesn't really count but I'm Colin Williams I'm the USGS David introduced me mostly I'm going to try to speak to the work that many other people have done at least touch on it both within the USGS and in particular outside the USGS highlight a few issues related to uncertainty and subsurface data and also just briefly mentioned that we are in the process of moving and I'm no longer officially in Menlo Park but in Moffitt Field which is 8 miles to the south where we're co-located with the NASA Ames Research Center see if this works next please I'll just do the Susan pioneered the way here so I'll have a very brief introduction talk a little bit about challenges in geothermal energy from the perspective of assessments resources assessments like we have done in USGS exploration and resource management of course where the common themes are in the context of natural hydrothermal systems and also enhanced or engineered geothermal systems or EGS as Susan introduced it and ideally just important topics and possible future directions for research next so geothermal systems happen in a wide variety of environments these are a couple of classic images one from Marshall Reed a former colleague illustrating what's conceived as a simple fault-hosted warm water circulation system this might be an idealization of basin boundary fault in the basin and range where Jim of course is the expert and would draw a much more complicated section and we'll get to that and the other from Henley and Ellis back in the early 80's kind of a cross section through a volcanic system magmatics hosted geothermal systems tend to be at least the natural hydrothermal ones tend to be the best place to find high temperature and large power potential but diverse environment of course you've got high temperatures and alteration focused in the center of the volcano but because of fluid pathways especially if there's a ring of faulting or groundwater flow down the volcanic edifice you can have hot water and hot springs way out on the distal end of course the common elements here are there's certain fundamental common elements fluid and heat flow being the main one trying to understand those fluid flow pathways typically there's a cap rocker seal because cold groundwater is the enemy of hydrothermal systems and sealing away that cold groundwater is shallow groundwater is important and of course being able to classify these systems in a geologic context with examples like the play fairway work being prominent is invaluable tool next can you go back yes surface manifestation so it's worth mentioning Susan touched on this a little bit that in geothermal of course it's not unlike oil and gas in that it helps a lot if there are surface manifestations and much of the early geothermal development and many of the largest geothermal fields developed both in the United States and around the world were found because there were hot springs or other you know steaming ground other thermal manifestations obvious hydrothermal alteration and that was the low hanging fruit just as if for example you look at a map of oil seeps in California and a map of all the oil fields that were discovered in the early part of the 20th century they pretty much overlay and just as the oil and gas and mineral resource communities have had to move beyond that to finding systems under cover or systems without obvious surface manifestations the same as true for geothermal that the frontier for exploration right now is in these blind hydrothermal systems where or at least ones for which the surface manifestations are indirect or complex indicator of the subsurface next then of course Susan already talked about this but in the context of EGS we want to be sure to note that EGS is a much broader resource it relates to being able to exploit the high temperatures that are ubiquitous at depth in the earth and it involves a broad range of stimulation potential mechanical thermal and or chemical and can be at a site like this where you're presumably creating a hydrothermal reservoir in a rock that was previously low in permeability and not part of an active hydrothermal system to working on the margins or within existing geothermal reservoirs and I think based on one of the questions that came in earlier that's another advantage that's worth pointing out to the DOE FORGE project is that many of the earlier EGS demonstration projects earlier this decade were involved working in or on the margins of existing geothermal fields and there are limits to what an operator will let you do in producing geothermal field and having a dedicated test site is a great step forward next. So some of the challenges and here I want to try to emphasize challenges that in some ways are at least partly common to natural hydrothermal systems and also EGS well of course you need to find high temperature right you've got to find predominantly high temperatures relatively shallow depths geothermal is where where those high temperatures come up in the near surface which of course within the high temperature halos when you can identify them typically the permeability is a much smaller volume and predicting where that high permeability is a real challenge even when you have a good the distribution of temperature and part of big reason for that is because most systems in the United States are fracture dominated in terms of the permeability which means there are dramatic variations spatially in permeability and predicting it is not just a geologic problem but a geomechanical problem of course and much of the exploration and the technologies to characterize the subsurface focus on cap rock and reservoir rock properties as do many other fields but one of them and again Susan alluded to this is because we're basically looking at water and hot water in the subsurface tends to look a lot like warm water we don't have the physical properties that necessarily come out of say looking at gas or petroleum versus water that you do in the petroleum field with EGS we're moving from an environment in which he transport is predominantly convective to one that is predominantly conductive so understanding the subsurface temperatures generally takes a somewhat different approach rock properties are also the key and the research projects are the way in which we're moving ahead and better understanding how factors like lithology pre-existing permeability and situ stress etc affect the viability of EGS projects and the other challenges of course since the EGS is broadly defined as a resource as opposed to narrowly focused in particular locations we have to be able to work consistently skills especially from the assessment perspective from the continental down to the local and be consistent it's been alluded to already but one thing that's definitely true with geothermal is their sparse data and limited resources and the important question obviously here is how will these challenges evolve and what carries across as we move from a geothermal community that's predominantly based on exploiting natural hydrothermal systems to one that's based on EGS next and here I'll just highlight that things are complicated no big surprise so this is a cross-section idealized based on a study of the Brady's geothermal field that my colleague Drew Seiler at the USGS worked on predominantly worked on this though I want to give credit with Jim Falls when Drew was at the University of Nevada Reno and it's highlighting in this great basin hydrothermal system context all the geologic factors that contribute to and affect fluid flow system there are many geological factors related to variations in mythology and structure and then of course the mechanical factors like the dilation on these active or recently active faults their slip tendency how that interacts with the thickness of various units what's the inherent orientation character permeability in these units Drew and his colleagues looked at a lot of these things in Brady's partly because Brady's has been operating for a while there are a number of wells in the subsurface they were able to make three comprehensive 3D geologic map and or should say model and analyze it in terms of synthetic wells that penetrate the surface and highlighted that dilation tendency in these faults in particular and coulomb shear stress change had a significant effect on which wells tended to be permeable but again we often in geothermal don't have access to the kind of data that allows for this sort of analysis which is absolutely key next this is a few slides I borrowed from Bill Cumming who's a consultant geophysicist quite successful in the geothermal regime he focuses primarily on magnetoteluric systems and in magmatic setting takes advantage of the fact that he and others that the alteration patterns of created as a halo around these hydrothermal systems in magmatic settings lead to very different resistivity characters and in particular this shows the idealized cap of lower temperature Smectite clay rich alteration above higher temperature reservoir where there's higher temperature Illite alteration and the higher resistivity associated with that reservoir next and the fact that you can see this in actual systems this is his section through geothermal field which is located at medicine lake volcano in Northeastern California and the reds highlight that low resistivity clay cap over the reservoir there which is very important because in these volcanic environments you get active shallow groundwater flow through the younger more permeable volcanic which are the very dark blues the Smectite fills the pore spaces and seals off the underlying hydrothermal reservoir which is again somewhat higher in resistivity with the higher temperature alteration the tricky thing here is that while the empty surveys do a great job in highlighting zones of potential it does turn out that in these environments it's individual faults and fracture systems that are again the conveyance of the high permeability for geothermal reservoirs and it takes additional techniques to actually target those next and then just reiterating for EGS we're trying to work across these larger scales and down to the local scale and there's a different challenge in the context of estimating in particular subsurface temperature and rock properties. Next. We need to develop better and more comprehensive 3D models of the subsurface this is just an example from a few years back in the vicinity of Biowawa, Nevada where through inversion of various geophysical data sets in particular gravity and magnetics combined with some drilling results from the nearby geothermal system geophysicists were able to tease out the 3D geometries on various important components there, Paleozoic rocks that have been intruded by much later Granitic rocks and then the intrusive characteristic of the northern Nevada Rift. In an EGS setting each of these would have different physical properties probably different temperatures at different depths because of their different thermal conductivity and being able to understand how to exploit this geology is very important but also often with these subsurface maps we don't tend to have a very good idea of what the uncertainty is. Next. And this question came up earlier and this is another one of Drew's slides but I tried to put from the related paper everybody's name on it because this is essentially a cross-section through the Fallon Forge 3D model that Doug Blankenship and Jim Falls and Bridget and many others worked on and it shows how the cross-section the geological understanding from the top to the bottom evolved from phase one to phase two B through a variety of efforts geophysical data acquisition a new well that was drilled and studied a lot of analysis of existing data in greater detail and you can see the lithology higher resolution the lithology is popping up in the subsurface the interesting thing and the reason I wanted to highlight this here is not to go into the details of how the model evolved but to note that one of the things that this group did that I think is in the geothermal context pretty pioneering is to note the uncertainties associated with their model and those are the little boxes in the lower right of each of the cross-sections which show the log of the distance from each model pixel to the relevant data sets that constrain those observations and you can see through the analysis and the new drilling how the data density improved tremendously and being able to carry these kinds of analyses into other settings and quantify them is very important. Next and the last one in the CGS context but perhaps most important is similar analysis something that we worked on a few years back is how to characterize your uncertainties in crustal temperature you often see in talking about EGS and geothermal you see maps like on the left of heat flow this is just a map of heat flow in the western United States of course reds are hot high heat flow blues are low heat flow and the right is a model based on this data set of the estimated temperature at 6 km and again you see that it looks a lot like the heat flow map but to get to that map on the right involves a lot of analysis it involves estimates of where there's predominantly conductive heat transport where convection or convection is influencing the shallow heat flow measurements what are the uncertainties on those heat flow measurements the spatial variation in the coverage of those heat flow measurements which are all the dots on the map you can see there's high densities in some locations very low densities and others estimating the thermal properties of the crust at depth with EGS we're obviously going to in many cases much greater depth than traditional hydrothermal systems. Next and by going through as well as we could what the quantitative uncertainties were we said that even in the best case at this kind of regional scale our estimates of temperature are only good on the order of plus or minus 20% but are highly variable depending on local crustal structure and data availability and quality and noted that we need to make progress in quantifying the uncertainties of these heat flow measurements there are a lot in the literature that do not have uncertainties associated with them we need additional measurements we need consistent scaling of models from local to regional and continental dimensions better constraints on thermal properties and the last one which is most relevant here is improved approaches to uncertainty analysis finally last slide I think so just to summarize I think as I look across having been involved with geothermal for a few decades now it's definitely made significant advances in integrating diverse data sets as Jim mentioned and I think it's worth mentioning that the 3D data sets that Jim refers to whether play fairway or Fallon Forge or a few other cases that we've been involved with DOE support has been key that many of the geothermal companies are operating on shoe string and some of them do this kind of analysis and some of them don't so developing these 3D models has been important we're still at the early stages of quantifying the uncertainties associated with those models and we need to learn a lot more especially as we move into the EGS context and I think this came up in Seuss talk and I'm glad it did but I think there's no substitute for drilling and when holes are drilled because many of these companies are operating on limited budgets they're not necessarily collecting all the data that they could or it isn't necessarily finding its way into the public domain and so projects like the one that Sue mentioned of working with holes of opportunity or Idle Wells is really important and I think leveraging what the industry is doing could be a cost effective role for government and that's it thank you Colin if you have questions please hold them till the end of the panel here write them down if you have to and now we're going to bring up Ken witherley good morning everybody I think one of my first men being from Vancouver not a lot of geothermal around there but I spent some time in Reno and I remember there was a spot about 150 miles east of Reno along the road there was a bathtub with warm water in it so I saw a manifestation of geothermal I believe I'm not sure if that bathtub is still there but it was useful the topic of the theme we're calling it managing uncertainty in the subsurface data ducker told us that if you if you can't measure it you can't manage it and I would say by and large the mining industry that I've worked for for almost a century doesn't really do a great job of defining uncertainty in part there's reasons for that one of the things that many mineral deposits unfortunately are quite unique and so what are brothers in the oil and gas sisters in the oil and gas and geothermal businesses probably have as more generic geological model whereas each mine is often very different and so special knowledge has to be built up around that actual deposit and when special knowledge is created it can be applied obviously to that deposit but companies tend to have very little motivation to share that information either within their own organizations or put that out into public in the exploration part of the business so that's the downstream in the exploration part of the business where I've spent most of my life I thought about this and there are some people certainly engaged in more rigorous mathematical approaches to trying to define uncertainty but by and large the industry has relied on experience it's basically the knowledge you develop working in mineral systems as we call them now is to try and make the best guess as to what's going on because we simply haven't over the years acquired sufficient petrophysical data calibration information to do a quantitative assessment of these things it's night and day compared to what the oil industry does in system characterization and we're too late now many of the sites which we would have liked to examine have simply passed this by so I'm going to show you some images and talk some about mostly focusing on the downstream side because I think that's where the money comes from and that's often where society sees the biggest impact so with Remy's blessing it's the top button right one on the right magic so scale dependent targeting this comes out of a presentation a colleague made that shows where we have our information and to some degree the certainty of that information we begin an exploration program really on the far left this could be this could be global state geological terrain and then as we move further and further down moving over to the right we put a lot more time and money into prediction and so we reduce uncertainty moving from the left to the right but it costs us a great deal of money to put that information into play and you can see the relative scales there and so I spent most of my time to the left of that red arrow but the mining business the extractive side spends most of their time at the consequences of the successful outcome of looking for deposits on the right hand side so the degree of uncertainty is reduced but it comes at a high cost I mean the very same things that Colin mentioned the drilling, Susan very expensive and we rely very very heavily on that information and in part as I said earlier most mining deposits are fairly unique and they require probably a lot more information to basically be able to figure out the proper metallurgy rock stability issues issues of fluid flow in and around the deposits all of this has to be monitored for the life of the mine and afterwards these things, these big holes that we create in the ground don't go away very very comfortably so some idea of the deposits, critical some ideas that were put out by a colleague he used to work for Newcrest and they've pioneered the development of large scale underground copper deposits in the state of New South Wales in Australia and he put some points out he's an SEG special lecturer about seven, eight years ago, Dan Woods' name so I've just put down some of his points and used some of his figures because I think showing these you get some idea where in the mining industry we're going to have to focus in getting better control of the uncertainty because as deposits go deeper it's probably going to have a larger footprint underground even if it is underground and there's going to be more challenges from how to manage that so mass mining he calls it these are constrained economically it's always running up against barriers because what we produce in our country is not typically unique we're always competing against sometimes the same companies who are mining here in the US are also mining overseas and it's up to them to make a return to their shareholders not just keep a good balance sheet for their domestic production where are we currently working with underground depths to 300 to 2 kilometers and that's in part it's a bit of an iffy thing it depends on somewhat on the local conditions some environments like in the Sudbury Basin they're pushing down 3 kilometers at Kid Creek in Canada and other places it's difficult to go beyond a kilometer just because of rock conditions future underground deposits I think we're thinking maybe or Dan was thinking maybe a 3 kilometer limit part of this has to do with gradients as well some of you have probably heard of the resolution deposit show of hands who you've heard of resolution there you go one of the things they're dealing with at resolution is ultra high temperatures probably requiring some sort of autonomous mining equipment but really having to design equipment at this point probably for one deposit in the whole world but it's a very very attractive deposit Rio Tinto is spending a huge amount of time and effort on autonomous mining trucks in the Pilbara the idea they will eventually be able to bring that technology to places like resolution open pits at one and a half kilometer deep hole that is definitely a pretty scary thought a number of you probably saw what happened at Bingham about three years ago where they had a it was not unexpected but still a fairly disruptive slump coming in from the north face and cost them quite a bit of money to fix that up the sorts of beasts that we're actually finding now and we'll probably see in the future these are cross-sectional views as if you were down in the earth these are very very large operations as Dan says we're going to have to get much better if we're going to reduce the overall footprint on the environment to knowing everything about what that cube is actually doing over time we have to have four dimensional models of these in this one in the bottom right I'll just expand on that notice the base scale 6.6 kilometers and that's just where the deposit is so they're taking a chunk of earth going down approximately three to four kilometers and have cube that's probably going to be 10 by 10 and all of that material by and large or it'll actually be changed in some way there'll be a lot of fluids flowing through there a lot of things happening that somebody's going to have to take accountability for so we need to reduce the uncertainty we need to be able to define what's going to happen before we actually start mining something like this on the geological side this is just the results of follow up on a workshop that I attended last month in Perth there's a very clever group of people at what they call the center for exploration targeting largely focused on the upside exploration side of things for 3D modeling and that's one of the things we probably need to see more interaction between the people the bright people on the exploration side communicating more with the miners that would be very very helpful for our business and you can see in this example when they're asked this is a questionnaire I've circled the one here and we're talking about mine dimensions of under a kilometer well that deposit in Mongolia is far bigger and if we have a 10 meter resolution we are talking very very large data sets very large data sets will be created for models of this so we're going to have to work hard to get control of this information otherwise there'll be blowback at a number of levels that will be ultimately probably damage the social license of the operation if we don't get it right and this is what these blowbacks can look like this was called unexpected subsidence of Ridgeway about 500 meters below that was a block caving operation my understanding is they knew this was going to happen they just didn't know when and unfortunately there was a loss of life attributable to this so this is the sort of consequences these mines of the future they potentially have obviously if it's an open pit you can guard the perimeter but these underground block caving operations are they going to be what the future is about for mining particularly in built up areas so we are uncertainty far better and define it that's it thank you thank you Ken and then we will finish up with Jeffrey Yaras I'm not sure whether the uncertainty here is greater with respect to advancement of the slides or whether or not I can get through my eight slides in 10 or 12 minutes so what's that there okay so I'm a little bit more in the weeds on the earth modeling side here looking at the uncertainty in fairly good detail around geological reservoirs and just as a matter of interest the scale of the models that we build today used to cover 5, 6 square kilometers those were fairly typical but today we're closer to 2,000 kilometer models or somewhere in the north of 2 million cubic cubic kilometers in size so the earth models have become quite large and the purpose of course is to really measure uncertainty that's what we really do that's our job and to move these models downstream to engineers to dynamic simulation or for drilling or for completion so with that in mind what I'm going to do is talk a little bit about the definition behind what I perceived to be uncertainty at least in the petroleum arena and then I'll talk a little bit about the sources of uncertainty in general and then I'll talk about our wishlist what we'd like to do with respect to our models and then actually what we do do in practice with some of the pitfalls that we have and I'll give you some examples of that so I'd like to start off with this definition that is kind of well known in our industry particularly from what I refer to as the three Johns John Harbaugh John Dubton and John Davis who really were some of the earlier pioneers in risk assessment and modeling of the subsurface and sadly John Harbaugh recently passed away so we owe him a debt of gratitude in the work that he and the others have done and their definition really was there's an inadequate awareness of the differences of the geological features as they are perceived and as they exist in reality and I would add to that that represented by the data available to describe them so we often don't have all the data that we would like and hence it's the idea of having a probabilistic model set in some kind of a probabilistic framework is really the way we approach this what we'd like to do theoretically of course is have all the possible outcomes of all possible values across all possible scenarios but that's really an impossibility we really can't do that so from a pragmatic point of view we put this into some kind of a probabilistic framework that we frame and we test uncertainty against that model and the real engine under the hood, the real work course here is the geostatistical methods that were developed very early on by Danny Krieg and George Matrone back in the late 50s and early 60s and that of course really originated in the mining industry and the petroleum industry has leveraged that as well and has taken it to slightly different directions in many cases than the mining industry has so we use something called conditional simulation and that's the geostatistical approach to a really reducing uncertainty so the reality too is that we live with the uncertainty assessment today that it's necessarily the models that we produce are necessarily constrained so we know we're not going to capture everything we're going to do the best job we can with the models that we create and we'll test against those models so that's basically the overview of that some of the sources of uncertainty that we have are the inherent randomness in nature there's a limited understanding of the phenomena we're trying to study we have measurement errors we have incomplete measurements of the phenomena that we're trying to measure and there are lots of interpretation differences so I really agree with Ken in the sense that our experience which is something I'm really worried about today in terms of losing that with the younger folks that are coming into the industry and not having the older folks available to impart that wisdom but the experience is still a huge important area and my focus has been more on the mathematical side so I've kind of got both of these things going on the conceptual models, our experience are hugely important and I see that something's wrong even though the math worked and the computer programmed in crash it doesn't mean that it's right so we've got to be very careful about that and I think one of the more important aspects of this is really to produce enough of what we would call realizations enough different models that we capture the unexpected contributions it's not the stuff that we expect to happen we kind of all get that what we expect to happen and the idea of capturing that entire probability distribution is really important so we want to get the extremes and that's something that we really lack and don't do well in our industry at least not well enough these unmeasurable phenomena are another thing if something we totally didn't expect at all to happen then it actually happens and how do we capture that in the conventional reservoirs plastic and carbonate reservoirs we've established a long history of experience so we kind of know where the uncertainties lie and the magnitude of those uncertainties so things like structure is really the first and the highest amount of uncertainty you want to add a little bit more reservoir to your reservoir quality rock to your model it's not going to be by increasing the spatial range on your variegram for example in the geostatistical world it's going to be the structure add more structure there's a lot of uncertainty around it and certainly room to do that and then that's followed by the stratigraphic relationships we have to pin down the stratigraphic relationships then generally the facies or rock types and then finally the petrophysical properties play an important role so we've kind of established that and that's generally the rule today with the unconventional things are kind of upside now we don't quite know really what facies are in these shale reservoirs we don't quite understand the rock types they're complex certainly we refer to them as shales but they're not really shales they're all sorts of things so what we kind of think is that the stratigraphy in terms of not only the chronostratigraphy that we're kind of used to or lithostratigraphic stratigraphy that they could override any of that so that's a huge part of the uncertainty that we need to capture and the rock types themselves whether they're clay, silica, carbonate, organics that's probably one of the higher levels of uncertainty that we have to deal with and then the geochemical or geomechanical processes that they go on so that part is really important and a little figure that you see in the bottom right hand corner I captured that from Gross from 2009 it just shows the overprint of the mechanical stratigraphy on top of the litho or chronostratigraphy so it's quite complex and it's something that we don't model very well today and it's really one of the sources of uncertainty that we have and have to do a better job capturing should I give up Remy here oh there we go I guess you did that so let's talk about what we do in practice we tend to really oversimplify our models that we build at least in the petroleum industry and the reason is complexity leads to increased CPU time which leads to increased dollars and everybody's under the hammer of time the drilling schedule is not going to change and you've got so much time to do your project so you cut back one of which is the structural or the stratigraphic complexity the other thing that we don't do, we should do is we run too few realizations in the probabilistic framework we don't capture the extremes and believe it or not those extremes become really important in trying to mitigate cost expenditures so we really want to be able to take the advantage of our computing systems today and our performance computing and parallel processing to try and run the realizations as much as we can to capture as much of the uncertainty as possible we run scenarios at the expense of realizations and make a subtle distinction between the two terms so a scenario is kind of your best guess and you could say well you know I think the high side the medium side and the low side would be these scenarios and they may be different but the real systems that are similar but you choose them and John Davis years ago showed that human beings tend to pick scenarios that are much closer to the mean and not the extremes so we kind of fall prey to that problem so we really want to run the stochastic realizations to try and help us capture those extremes and the other thing that we do is often tend to spend more time on variable sensitivity than we do in capturing the uncertainty and those two terms on a practical level on a day to day basis is often confused by our modelers they tend to say well I ran realizations I ran those things but what they're really doing is variable sensitivity so the variable sensitivity is not going to change from realization to realization right so you really need to do both and that's important if you run a single realization you have no idea where you are on that cumulative ability distribution function and to know where you are you really have to map it all out so we have a couple of examples there's what I call the hand-picked example and that's where someone says here are my three scenarios and for the reasons I just explained that's not really a good idea to do the other way people will work is they'll say it's what I call the one one and many they're going to run one realization for each of the structure and the facies and then they're going to do multiple realizations on the petrophysical properties which have the lowest degree of uncertainty and sensitivity in our model generally so that's kind of an issue but it saves time another way to do it is to run many structural realizations and then run one of each of the rock types and the petrophysical properties and again that saves time and it captures the greatest uncertainty which is the structure generally in conventional reservoirs but doesn't really capture the rest of the uncertainty that could end up being very important and then the last one is what I call the one one one which is you just choose one realization of each and you focus on variable sensitivity that's one of the things that's more common in the industry to do and I don't think that's really the right thing to do and the little example that I have off on the right there gives you an idea of why it's a problem right if I say I want to run a probabilistic sense and then for each one of those I want to run a hundred realizations of the stratigraphic relationships you're up to 10,000 realizations if I want to run a hundred realization of each of those 10,000 for the facies or rock types then I'm up to a million realizations and then finally the petrophysical properties each will increase that thing to over a hundred million realizations and really even now with the HBC the high performance computing systems that just isn't really going to cut it people aren't going to waste the time to do that but maybe that's something that we can accomplish in the future or we can mitigate that by finding simpler ways to capture the extremes on those realizations and cover all the properties that are necessary so that's kind of one of the issues that we face so this next one I like this example I think you'll appreciate this this came from the University of Houston with my students in my class last year looking at the uncertainty in a reservoir model this is the static model that's going to go to the reservoir engineer on the left you've got the probability distribution curve which gives you the plethora of realizations and ranked and then on the right where it was in video mode was showing all the different possible from the most conservative to the most extreme and if you really got a sense of it it was changing quite a bit so there's a great deal of uncertainty in terms of this reservoir model if you just pick p50 you might be in danger not understanding what could happen on the extreme so what we like to do with uncertainty is say well I'm not going to give you all 100 or 1000 realizations because my reservoir engineer would not be happy with that that's a lot of work to do and particularly when I tell them that each realization consists of something like 50 or 100 million nodes that just isn't going to happen but they might take three they might take that p10 they might take the p50 and they might take the p90 so you bracket the uncertainty and then of course we could discuss upscaling and whatnot which may also lead to some additional uncertainty but that's practice in the industry so we want to bracket that uncertainty and this way get a sense of what we could have and what we have to be concerned about and what are most likely cases the other thing that I think is important is really in this slide and that is all too often we tend to run these deterministic models which is on a far right so that's a that's creaking which is the geostatistical approach but it could be any kind of interpolation algorithm it gives you this static model which is a single result and it shows the distribution in this case of a porosity in this particular field and then I show you the recoverable hydrocarbons of the base of the figure but if you do a co-simulation you integrate more than one variable say two you know porosity impermeability or some secondary variable or seismic data for example a tertiary variable if you need to you get a much more detailed solution and you can see that in the middle figure it's more homogeneous it gives you solutions that increase in this case the recoverable hydrocarbons and the image on the right is even better because you've now done it in a stochastic matter so the first two images are static they give you single solutions but the last image involves a co-simulation as opposed to co-creaking and it looks at all possible realizations 200 or 1000 whatever you've run so those images that you see show the degree of change that takes place in the inner well space and then the resulting probability curves on the right kind of summarize that the blue curve would be the one with best recoverables based on the better measurement of uncertainty and the black curve would be the middle one which is co-simulation or co-creaking and then the little red dot that you see there is simply the result of co-creaking itself so we really encourage the idea of a stochastic framework in order to assess the uncertainty the last one I want to show again I'll point to the University of Houston where I was teaching last year and this is one of my students who just presented this at the International Association of Mathematical Geosciences a few weeks ago at Penn State and this is kind of one of those unexpected uncertainties it's kind of one of those things that you think you'd know but maybe you don't or you forget and that has to do with missing data right missing values and sparse data so one of the things that I think Susan alluded to earlier in her talk was that the petroleum industry has a lot more data well that may be true in many ways but when it comes to the unconventionals we're not spending the money on drilling the wells logging the wells or evaluating the wells so you have all these horizontal wells all over the place with almost no information and the little information that might be there is proprietary so you just don't get to evaluate it properly so how do you manage uncertainty in that context and what the figure here shows particularly on the right are the different patterns that come up with missing values your petrophysicist is going to know this but the modeler may not and the little patterns that you see there indicate where missing values occur and the green zone the green area that you see this little stair step pattern is really at the base of the well because we stack our tools right so at the bottom of the well you might get your gamma ray but you might not get your neutron and as a consequence you get this funny pattern then if you're looking to fill in that data because of an interesting part of the formation you're going to need to subset it and look for well to have that information so that you can then use machine learning an appropriate way of looking at with machine learning and filling or predicting or imputing those values the blue values are randomly random issues or outliers for example the yellow or gold strip that you see there is just a completely missing log it wasn't available to the public or it wasn't available because it was never run and then the blue area at the bottom is where the hole was drilled deep but it wasn't logged that deep so by categorizing these things and finding similar wells in your area that have this information you can begin to impute or predict those values properly on the left you can see the result of missing values often leads to sparse data so here's something out of the Permian on the left you can see the distribution of those wells and then on the top if you try to krig it you don't get a very or do some kind of a simple algorithm to interpolate the data you don't get a very good solution so it's it's pretty broad in the solution if you do a conditional simulation what might look noisier look more heterogeneous but it's really no better you really need to add a secondary variable what are you going to do if you don't have seismic how do you manage that right so one suggestion is that you do basin modeling and that's a volume based model uses finite volume modeling you distribute the properties you use it as a secondary variable and you fill in the data and then reduce the uncertainty it'd be just like using seismic if you had it but you don't have it so there's ways we can mitigate these things but it's important to think about that and how you're going to manage that so at the end of that if you look at on the left again you see the basin model is on the upper figure on the left there and the lower figure is the minimized uncertainty version where we integrated the well data with a physics based model in order to get the proper spatial distributions of the properties of interest alright so about done here so just a few pondering things that we can be thinking about I think that one of the solutions that we'll have in terms of uncertainty is really the use of machine learning and big data analytics of course high performance computing there's no question that that's going to play a very important role for the future and using it properly is important so we want to make sure we don't lose the physics and the insight of these folks when we do that you don't want to just throw stuff in a neural net and say here's the answer you really want to make sure that you've captured the proper physics and chemistry to get the right solutions looking at maximum diversity realizations maybe there's a way we can shortcut the number of realizations we have by looking at maximum diversity so we can pinpoint the end members of those distributions in terms of dynamic simulation those are generally based on best estimate models a single solution how do we integrate multiple realizations into the dynamic simulator so that we can not just choose two or three or even one how can we integrate hundreds and look at the uncertainty in dynamic simulation can we use proxy modeling that's another popular thing to look at where we use machine learning to run the first few time steps on a simulation and then use machine learning to complete it reduction of nonlinear responses from data sparsity, diverse realizations and upscaling is another thing to consider and then in terms of also with respect to simulation dynamic simulation in unconventional reservoirs modeling the impact of natural fractures that's a huge deal what is the impact of a hydraulic fracture into a naturally fractured system I think that's where everybody's really looking to try and figure out when and where it's important the use of non-darcy flow equations in our simulation particularly in unconventional reservoirs is it important, is it not important sometimes yes, sometimes no what's the effect of a stimulated reservoir volume how do we model that so that's an important piece the reduction of uncertainty through unstructured grids or grid list topologies I think that's something that we can be looking at and then finally the last thing I'll mention is well placement which is really what we're all about modeling earth models, building earth models well placement how do we pass that information to a completion and then ultimately from production just because you have a great frack job or a completion job doesn't mean you're going to get really good production if you drill it well in the wrong place the best completion job isn't going to help you very much so we want to have good well placement minimize the uncertainty for that and then there's a whole level of minimization of uncertainty with respect to the completion methods themselves in terms of what components are used and how much of or how little of those components are needed to have effective effective work so that's it, that's all I have to say and I'll turn it back over to the chair okay Jeffrey, thank you let's move immediately to questions from the audience Jeff and Anna Colin and Jeffrey you both spoke to the importance of natural fractures to in shales and in hydrothermal systems what is the technology applications for identifying that and what's the state of the art currently so we rely a lot at least in the work I'm familiar with we rely a lot on borehole image logs and in trying to understand the state of stress and from that and knowing something about the structure surface geology trying to predict where the open fractures would be out of the population of fractures when you look at the entire population of fractures tends to be quite large and quite diverse in their orientations and their character but how about pre-drill I mean drilling's expensive pre-drill that's where Jim comes in having the geologic mapping and the structure pinned down as well as you can in advance through the combination of geologic mapping geophysics I think has proven to be critical because it's those structures that define where the fault intersections are and the favorable permeability and I don't know if Jim wants to add anything to that I could say more I'll cover it with a few slides later but basically we've characterized what some of the most favorable settings are and then gotten into the devils and the details why those particular settings are more conducive for fluid flow and then had good people like with the USGS modeling slip and dilation tendency so I'll add my two cents let me know if I'm getting too down into the weeds there but there are really three components that we need to look at for the natural fractures and one as Colin just mentioned borehole images are critical for us to have so if we have a borehole image we know exactly where those fractures are in the borehole and we know the number of fracture sets which is important the other part of that is the use of seismic data where we can look at seismic attributes and if you peel down into the seismic and you look at the smallest type of fractures that you can model from some attribute you can think of those then as fracture swarms or the likelihood of a fracture occurring so that would be our secondary data so I mentioned in the presentation that we have this ability to do co-simulation so we want to use this idea of co-simulation where we know that the data are pinned at the borehole images we know that that exists we use the fracture of the seismic attributes as a secondary variable and then we use the stochastic geostatistics to propagate the fractures in space and there's a third variable that can be used you can integrate multiple tertiary or multiple variables into the co-simulation process which could be gravity or magnetics or fractured density maps or even a hand-drawn map to make sure we're incorporating experience but obviously the biggest issue we have unlike mapping fractures on the surface we can't see them so that is the issue so we want to do multiple realizations then of these fractures and we want to make sure that we have the fracture attributes distributed not just the presence of the fracture but stochastic versions of the fracture aperture and the fillings of the fracture so we actually have that software to do that in the industry and it's quite complex and the last thing I'll mention on that is that we have to be more attentive to the complex effective stress field it's not a single value every time you hit a natural fracture or your near a natural fracture the effective stress field changes so we've taken approach using something called material point method which allows us to map the effective stress which now becomes a component in terms of how you orient your horizontal wells and then the kind of fracture completion methods that you might use and then the impact of those hydraulic fractures on that natural fracture system so that's the workflow so to speak question from the far end table down there my quick comment is everything you said in addition to that when I worked basins for fractures the key thing that I found most effective like in the montan basins or whatever you're working is the anisonic stress states first thing you want to do is understand the anisonic stress states for the area that you're working in you hit a strike slip regime or you're in a first fault regime understanding the anisonic stress state that you're dealing with first and then look across the basin and look at the different fracture sets and things near the wearable and then looking for fractured intensities in different areas but the anisonic stress state for that region is probably a good starting point can I make a comment here sure go ahead John you're right on and I think that when we're actually building a model of subsurface that includes natural fractures we we would start from the basin scale because the physics would take place at that scale dictate what's going to happen at every other scale you'd run into trouble if you just try to do the local area without understanding the bigger picture so it's really important to take in that basin scale model and drill it down to your area and you can inherit that information at every scale so you're correct Lynn so that last question anticipated mine a little bit I wanted to try to back up and think about this the overall question of what are we trying to know and how can we reduce the uncertainty and it seems to me that maybe one way to think about that is to ask something about the length scales that matter for a particular process so I'll pick a simple example because the harder ones are too hard but suppose you're thinking about the temperature distribution then there's some length scale associated with the underlying regional gradient there's some length scale associated with conduction there's some length scale associated with convection maybe through fractures and so I wonder if in thinking about a general problem can we gain anything by stepping back and asking which length scales matter maybe they're all there all the variability is on scales that are so small that we can't ever resolve that that might be one of the things that could happen but it also could be that the combination of drilling and other remote sensing might give us information that helps band the uncertainties in a particular area you can think about variability and permeability where which really matters the extremes dominate all the flow stuff that's different from something that has a big conductive smoothing so adding some physics to that process and then the last piece of it seems to me to be about sensing length scales so asking ourselves what under what circumstances can we do some sort of measurement that has a length scale appropriate to whatever we're trying to do now maybe this is all obvious to all the experts in the field that probably is but so I wonder if you could just comment on whether there's a fruitful path here or we already followed that pathway as far as it'll go I think you're absolutely right one of the when I referred probably a bit incoherently to working across these multiple scales for mapping subsurface temperature in the upper crust if you take for example thermal conductivity I would kill for a three-dimensional geologic map of the United States to a depth of 10 kilometers at scales of even a few kilometers in spatial resolution one of the real challenges we have is we often have very good information in a few local places we have what we think is pretty good information at the scales of tens to hundreds of kilometers and what we think are relatively good generic crystal properties but that that middle scale the middle scale that matters and then in some cases the very local scale where we don't have a lot of data we need we need things that the appropriate length scale that we don't have and probably are attainable by bringing together multiple datasets and working on them with at least in my case the thermal model in mind there's not more much more that I can add to that but I would say that there are two things that I would just simply state one is first of all your question is totally relevant and right on and the fact that that's really where a lot of research is being done today to determine these appropriate scales and one of the most important scales that we're not really understanding very well at all is the effect at the nano scale so flow at the nano scale and the impact of how hydrocarbons make it out in our business how the hydrocarbons make it out of the source rock into the surrounding shale so to speak at that nano scale and we're not going to understand how to model hydraulic fracturing and its relationship to natural fractures until we get that scale down there are others but I think that's the key there is we're trying to figure out the best scales I'm afraid we haven't exhausted the topic but we've exhausted the time we had available Jim would you say we need to move on at this point? I think we have one or two more questions since we have some interest and then we'll take a break. I think this should be a pretty quick question again Tom Crawford with the USGS I know that on the mineral side of things there's a real issue with preservation and accessibility to data and more often than not it's not the first company that explores an occurrence that actually defines an economic deposit it's maybe multiple companies down the line and in the geothermal world I'm wondering if that's typically the case and if so what about the preservation of that data from the original companies and the accessibility to that information down the line I know it's a big problem on the mineral side I'm wondering about geothermal it is a big problem it varies somewhat by state some state as you know the federal government doesn't require samples or data to be provided but some states in Nevada as one example do a fairly good job but basically it's the same as the minerals world the one if I could say benefit in a way from the fact that the geothermal industry is a small industry is most people know each other and they know who has what in their garage or what data trades happened with what companies 10 or 20 or 30 years ago the DOE has done I think a fantastic job of in two ways one in capturing a lot of the gray literature and a lot of the data that's out there in the public domain but in obscure places and also something that they didn't used to do but they have done in recent years is require that anything that they're helping to fund make sure that the data from that experiment come into the public domain okay Jim yeah okay join me in thanking our first panel this morning okay we're gonna we'll take a break we're gonna stay on schedule which means to return by 1115 okay so we already have ten minutes thank you is going to is chairing our next panel so let Joel I'm gonna turn it over to you well we have three speakers for this session I'll briefly go through their biographies a bit before we get started uh Jim Falls is going to lead off the session Jim is director of the matter of your minds and geology structural geologist primarily structural deformation uh he also recently has been well not so recently more than ten years now he's been involved in geothermal energy so he's our first speaker uh Azadeh I is our second speaker she'll be talking about uh geomechanical aspects of things and that's mainly what she works on with the itasca consulting group been working primarily on uh fractured rock and its mechanical numerical or mechanical aspects of rock doing numerical modeling and developing technologies looking both geothermal oil and gas and mining and our third speaker was on the first panel so I won't uh say any more about him but if you want to find out more you of course read their bibliographies in our report Jim I'll turn it over to you right uh thanks very much I'm gonna talk a little bit about some of the strengths as we've been able to make primarily in Nevada looking at some of the geothermal systems they are modeling them and and really kind of trying to define where we have hidden blind geothermal resources of significance so I'm gonna start here I actually have a laser pointer my kids say this is so 90s dad you know they're 20 something but anyways but basically there's a shot of McGinnis Hills and about the actual sort of center of Nevada and it's now the largest producing geothermal system in Nevada 165 megawatts is the latest from ORMAT as of a few days ago in terms of what three relatively new power plants all built in the past five years are producing there and so this is an amazing system it's a blind system no hot springs or steam vents here at all and it's now been developed into the largest system in Nevada and so I'll talk a little bit about these kinds of systems in the region because we think there's many more of them and just to acknowledge the significant support we have received from the geothermal technologies office at DOE which has funded most of this research over the past several years see if this actually works there we go so just real quick a quick summary this region the Great Basin region produces nearly a gigawatt well at least in terms of nameplate capacity of power plants typical systems produce anywhere from 10 or even under to as much as a couple of hundred maybe a few hundred megawatts I think at one time at COSO produce that much and and just so we're all on the same page one megawatt is enough energy for about 750 to a thousand homes so as others have mentioned Colin and others this region has vast more potential than this I think some of the estimates have been 30 gigawatts I actually think it's much much more than that especially when you consider that most of the conventional resources are blind or hidden and this is what we've been doing the past several years this is just geothermal production in Nevada and you can see there's a significant climb in terms of production in Nevada in the past ten years and the opening up of several additional new power plants which are shown by the purple stars on this map here alright so we know that this region is under crustal extension and of course crustal extension brings warmer parts of the lithosphere up closer to the surface and so the primary reason for the significant geothermal activity in this region is high geothermal gradient not necessarily magnetism magnetism is really isolated to a couple of small parts of this region and not a factor at least we don't think it's a factor in most of the geothermal activity in the region so the crust is pulling part we have lots of faults lots of active faults nice conduits pathways to bring hot water back up quickly to the surface to essentially exploitable levels so the secret is bringing these pathways and using the geologic and geophysical techniques to figure out where they are at depth so this is a typical geothermal reservoir this is just a simple version if you will portrayal fractured rocks as mentioned earlier that fracture permeability is the key to geothermal activity in this region and so that's a simple view but the devil is in the details and this is a nice 3D model that Drew Seiler has generated for the Brady's geothermal system that was mentioned earlier Drew is with the USGS now and so lots of complexity and so on to understand sort of where those sweet spots are in terms of fluid flow at depth so basically you know we need geothermal production wells injection wells etc and of course these are always at much deeper levels than any of the geothermal aquifers and we want to re-inject of course the water into back into the system to keep fluid loss at a minimum so we have a great region undergoing crustal extension high geothermal gradient you think it would be easy but it's not necessarily easy to develop a geothermal system there are a number of challenges one that I'll focus on meaning we don't have any surface hot springs or steam vents so finding sufficient permeability at depth is key and it's actually more important than finding temperature we have high temperature in this region the key is where do we have permeability at depth so any challenge creates an opportunity and the opportunities we have is we have lots of great techniques to characterize the favorable settings for the known systems in this region on that for about 15 years now a lot of innovative geologic and geophysical techniques to apply to that both in terms of characterizing on the surface and then after you've characterized going after similar systems based on subsurface data maybe in other regions we have lots of great software techniques now to develop 3D conceptual models and this is all culminated for some of us anyways over about the past 5 years in a DOE funded project the play fairway project that I'll describe in a little bit of detail here so this is just to portray the main challenge these are just a couple of cross sections and most of the fluid flow at least the geothermal fluid flow in this region is controlled by faults the fluids come up along the faults but then they commonly hit permeable layers and leak out and maybe the hot springs are several kilometers away from the upwelling at depth so if you drill over here you're in for a surprise because the temperature is actually going to go down with depth so how do we sort of focus back into finding those sweet spots if you will where the geothermal upwellings are actually occurring and then this is actually what we think is the most common case the fluids come up along faults leak out into permeable layers on the surface and the adjacent basins and really never see the light of day so how do we find those systems that have absolutely no reflection at the surface we know that 40% of the known systems in this region are actually blind and just about all of those have been discovered accidentally through agricultural wells or mineral wells that have found hot water and then many of those have been later developed into geothermal power plants so we estimated about three quarters of the resources are blind and so then you have a significant sort of exploration risk in finding these systems and developing them and so again you need high quality geologic geophysical data in order to have a successful exploration program so this is what we started doing now about 15 years ago culminated in a major DOE funded project in the early part of this decade and we started characterizing the structural settings the sort of fault patterns that are conducive to geothermal activity in this region by looking at the known systems so there's about 450 known systems we were able to catalog about 250 of them and we found that these four fault patterns controlled nearly 90% of the systems fault terminations where faults break up into multiple splays fault stepovers fault intersections and areas that we call accommodation zones where there's multiple faults intersecting one another and dying out and terminating as well so most then of the known systems are on rather they're on young faults paternity faults that accommodate the regional extension but less conspicuous faults not the major range bounding faults and the higher temperature systems actually all of the higher temperature systems occur along or near faults they're less than 750,000 years old so we found similar findings in other parts of the world and fundamentally though you need detailed geologic maps and of course we have a lot of techniques now that help to sort of facilitate the production of geologic maps like LiDAR but we also need good subsurface data then to find these systems at depth so just real quick most of you probably know what LiDAR is but we need to produce good geologic maps and realize that only about 20% of this region is mapped in sufficient detail like this and then but today we can expedite that mapping if we have the funds to acquire LiDAR and so this is just a view of an area of forested area as you sort of climb up into the sierras and then we'll reveal what that area really looks like with LiDAR high resolution topographic imaging and you can see a little stuff over here in a normal fault system and this is the kind of area that we might explore for a blind geothermal system so LiDAR is very important in moving forward and oops we went too far back one more please there we go so and then of course subsurface geophysical methods to elucidate what we have in the subsurface and of course gravity is very important noting lateral contrast in the density of rocks faults at depth magnetic data is very important might indicate altered rocks at depth, geothermal systems commonly correspond to mag lows magnetotoleric data very important basically mapping electrical resistivity at depth and geothermal systems commonly correspond to areas of low resistivity due to the alteration at depth or fluid saturation and seismic reflection data really sort of mapping the subsurface at depth and noting some of those favorable fault patterns if you will at depth so that we can go after some of these blind geothermal systems the problem is where do we have these kind of data in a particular region or in this region and most of this region lacks that kind of detail data that we need but nonetheless we sort of started combining the geophysical methods and the geologic methods to develop a geothermal potential map part of the play fairway program this started in 2014 phase 1 was then producing a geothermal potential map there we go for that box in Nevada about 96,000 square kilometers one third of Nevada we then did detailed studies of some of the promising areas within that box and then finished just finishing up the third phase where we did drilling of two of the most promising areas so this is our Great Basin Transact Nevada Transact again a high heat flow abundant quaternary faults and we had the advantage in this particular box of having a number of benchmarks that we could gauge against including 14 geothermal power plants and 34 areas where we knew the temperature was 130C or above so these were some of the parameters we used the parameters shown in red were what we used in creating that initial geothermal potential map all of these parameters were ultimately used in some of the detailed study areas and you can see on a regional basis we lacked a lot of the critical geophysical data sets to really understand geothermal potential but of course we did the best we could and then on a more local scale than commonly we have to acquire these data sets shown in blue so a huge challenge is basically availability of data and then an additional challenge is once you have all those data sets how do you combine that into something meaningful to actually evaluate geothermal potential and I purposely made this small because I don't have time to talk about it but basically we combined a number of those parameters about 10 parameters into evaluating regional scale permeability, local scale permeability and something we called intermediate permeability and then we combined permeability with heat to develop a play fairway map of the region a term that we stole from the oil industry and the warm colors indicate areas of higher geothermal potential cool colors lower potential and then from this we studied five areas in detail one of which I'll describe now and that's a place called Southern Gabs Valley which is sort of in basically West Central Nevada oops there we go this is very sensitive there we go so this is Southeast Gabs Valley and this is a new discovery we found a new blind geothermal system there have been absolutely no geothermal exploration in this region a master's student of mine did much of this work Jason Craig he should get significant credit but first we noticed that there was a favorable structural setting in this area there were active fault zones and so we went in and looked at it in a little bit more detail and there was a branch in the area we found warm ag wells we then did geothermometry that was promising then did a two meter temperature probe into this area and then we spent the money on additional geophysics I should say most of the geophysics was done by the USGS crew headed by Jonathan Glenn so we did detailed gravity detailed mag and detailed magnetotelurics we found then a sort of gravity anomaly intersecting gravity gradients indicating intersecting faults at depth of mag low indicating altered rocks at depth and a low resistivity anomaly indicating either altered rocks, clay cap or possibly this saturated fluid flow at depth all of these anomalies are co-located indicating the potential for a blind geothermal system so then that led to the third phase where we drilled this about a year ago, June of 2018 and right where we expected to find the system we did 124C at about 150 meters depth it's about a two or three kilometer long system that's probably viable for something maybe on the order of a 20 or 30 megawatt power plant so we've confirmed if you will a blind geothermal system in this area and I would say our methodology is guarded lead verified we did do a second system another system and another location in Nevada a place called Granite Springs Valley so we feel pretty good about this especially since there was nothing here initially not even any indicator of geothermal activity at the surface not so different than finding hydrocarbons in most of the world so and I just want to compare phase one and phase two results real quick this is phase one it's a blob okay we knew there was a favorable structural setting and then phase two then when we did additional geologic work and especially the geophysical work the basin blossoms with data if you will in terms of understanding where the most favorable setting might be within that initial blob in other words we're vectoring into the most promising area the area that has the highest probability of a geothermal system we drilled it and we found one so this is what most of the region is like in terms of availability of data and this is where we want to get if eventually we can have detailed gravity MAG MT geologic work that covers most of the region alright so conclusion it's a great region lots of favorable settings for geothermal activity a favorable tectonic setting vast geothermal resources I think far greater than the 30 gigawatts that are estimated nearly 90% of the known systems occur in four types of fault patterns and the main challenges are finding sufficient permeability at depth in those blind systems but you can have a systematic workflow where you're combining geologic and geophysical data to essentially develop 3D geologic maps and then reduce the risk of where you ultimately drill plus you can combine that with a new generation of geothermal potential maps that are combining multiple parameters that play fairway analysis approach and we think there are many more McGinnis-Fills than to be developed in the region we're now beginning to apply machine learning techniques to take this to the next level but again the major challenge is availability of data through most of this region so thanks very much thank you Jim we'll hold the questions until after the three speakers has the date today I'm going to talk about fracture network engineering and the role of numerical modeling in engineering of subsurface problems basically we have come to an understanding that fracture network and pre-existing fractures are important to many of the industries that are important for us including hydraulic fracturing in hydrocarbon reservoirs enhanced geothermal systems block cave mining, radioactive based storage and also CO2 sequestration for example we now know that the hydraulic fracturing in shale gas is very complex these hydraulic fractures are not planar surfaces they interact with pre-existing fractures which results in very complex geometries and fracture network engineering is the engineering of rock masses such that we get the desired performance and this performance can be different from one industry to another with the extensive use of numerical modeling and also correlation with field observations such as micro seismic or seismic data so why fracture network engineering is so challenging in different industries to my view it comes from two different reasons or there are two contributing factors one is the complex mechanisms so we are dealing with very complicated mechanical behavior coupled usually with hydraulic behavior and sometimes with thermal and chemical responses and this gives rise to a very complex engineering behavior but then also we are dealing with uncertainty and this uncertainty is coming from different reasons one is the lack of access to subsurface the other one is the heterogeneity we are dealing with rock masses that are heterogeneous at different scales from micro to like kilometer scale and finally there is a significant presence of discontinuous surfaces such as pre-existing fractures so this picture here tries to show schematically what fracture is we tend to rely on field observation one of the most important one would be micro seismic data analyze these data and then feed them to different type of numerical modeling at different scales and use these numerical modeling to gain insight into field characteristics for example what are the status of stresses what is the fracture properties but also to gain a better understanding of the mechanical behavior for example how hydraulic fractures are formed and how they interact with pre-existing fractures and use this understanding both in terms of field characterization and also in terms of mechanical behavior and feed it back to the model in real time to improve our field characterization and also improve our design and operation so this graph here tends to show the project life cycle and how traction network engineering fits in usually the project life cycle starts with resource targeting site characterization coming up with the design criteria and it's through this design process that numerical modeling traditionally come into play and traditionally numerical modeling has been extensively used as a validation tool eventually we come up with an engineer design and some sort of field observation sometimes when there's a deviation in response, in expected response we again rely on numerical modeling to gain a better understanding of what went wrong but in the workflow that we are advocating for numerical modeling is ubiquitous through the life cycle of the project so we use numerical models at all including microscale, laboratory scale, borehole regional and reservoir scale to again use these information and numerical modeling to gain a better understanding of the data constraint uncertainty and then continuously update these models which are multi-scale but coupled meaning that all of these models eventually need to conform to one set of data and that by itself is very important in constraining the uncertainty and use this process and continuously feed to our site characterization to design and to operation so what are the characteristics of these numerical models first what we are advocating for is coupled geomechanical and micro seismic modeling basically we are using micro seismic modeling as a way to calibrate our geomechanical model but then also we are advocating for the use of this continuum modeling so basically in plain words that means that discrete fractures are important in these systems fractures and their individual results in mechanisms that cannot be captured with traditional continuum models such as finite element or finite difference and therefore explicit representation of these fractures into geomechanical model are important for subsurface engineering then these models must be physically based basically mechanical, thermal, hydraulic and chemical processes need to be formulated explicitly in these models and finally there is a coupled model that produces synthetic micro seismicity and by synthetic micro seismicity I mean being able to predict, slip on these fractures and translate that to individual events with magnitude their time, location and eventually coming up with a micro seismic cloud so here I have one example that we applied this approach to underground mining field study and here hydraulic fracturing was used as a way to precondition the mine basically coming up with a method to reduce induced seismicity as a result of mining operation so here what they do was that there were a number of boreholes and they inject into these boreholes sequentially in different stages with the purpose of creating hydraulic fracturing and thus changing the state of stress and then using this precondition formation and subsequently do the mining excavation so we started with a borehole model and we modeled the process of injection into each of those individual boreholes we captured the synthetic micro seismicity and then we used this synthetic micro seismicity calibrated the result against the field observation during the preconditioning for each of those preconditioning stages and the calibration process led to a much better understanding of in-situ stresses and also fracture properties we then built a global model of the mine with a better understanding of in-situ stresses and fracture properties and used that to predict the induced seismic events the graph on the left button side shows two lines in the first case so actually let's go to the dark blue one the dark blue one is the induced seismicity in a formation that wasn't preconditioned and we had fewer number of events the y-axis is the log of number of events but the magnitude went up to almost two and then the numerical modeling which was calibrated predicted that once we do the preconditioning we will have more number of events but with much smaller magnitude and this is exactly similar to the pattern that it was observed during the mining operation there are more number of events but with less magnitude when we carry out the preconditioning so this example shows how numerical modeling can be used for gaining insight into side characteristics including the fracture information and properties and also stresses and then how to reliably use these type of calibrated models for predictive analysis another example that shows the workflow was in the Fallon Forge project that we were involved for numerical modeling aspects of it so this was conceptual design these are horizontal wells and it was envisioned that stimulation would be carried out in stages sequentially the animation here shows the result of numerical modeling with explicit representation of fracture in the model and the region is basically the fractured fractures which have experience pressurization and the graph here shows fractures that experience sleep as a result of injection so the red color shows a sleep of about 2 centimeter and then this information which was obtained from a geomechanical model with discrete fracture representation was then translated to a micro seismic model and the micro seismic model was able to predict the micro seismic events, individual events on these fractures and eventually obviously their magnitude, location and time and eventually a micro seismic cloud that could be used for aiding the design for example where to put the production well so this concludes what I wanted to say today there are a few other things that I want to mention right now we have the capability of modeling heterogeneity and fracture in these models at all scale for example this is a laboratory scale of rock we can put a discrete fracture network and create a rock mass at a scale of meters we can increase this and go to a scale of a regional scale which is kilometer by kilometer and then we can get a portion of this regional scale model and create a reservoir scale model again the catch here is going through multi-scale modeling but the models that conform to one another so the capability for numerical modeling is there but then as our capabilities increases our ambitions also increase and computational time therefore increases this requires investing more in more advanced computational procedures such as use of parallel processing and also coming up with more efficient solvers there are also two criticism often associated with numerical modeling in general and especially with numerical modeling of fracture network the first one is that we don't have enough information about fractures and the answer to that would be obviously this calls for better side characterization but again I want to emphasize on the role that numerical modeling itself can play on constraining the data so these type of multi-scale modeling and explicit representation of fracture network are a very efficient and reliable tool for gaining a better understanding of the side characteristics including the stresses fractures their properties orientation and size the other criticism is that the range of variation in DFN response is very wide and my answer to that would be so is the range of variation in reality this doesn't mean that discrete elements modeling should not be used in fact they should be used and they should be used in a way we would be able to create a stochastic models and also run multi-scenario simulation so discrete elements models are useful when we do calibrate with respect to field observation such as micro seismic data and also when we are relying heavily on a stochastic model generation and multi-scenario testing and simulation and finally we are able to put these heterogeneities in these models but we don't want to put them with the field observation at all different scales so the amount of information that we put in a model which is one meter by one meter is very different from the number of fractures and the amount of details that we put in a regional model so we need some sort of research and some sort of approaches to be able to put these fracture network in different models at different scales by selecting the most important mechanism and representing those mechanism and those features in these models so that concludes my talk Thank you Azadeh. Now we will move on to Ken for his second appearance. The US Army during World War II wrote a book using a divining rod to find water I guess they were keen to try any approach to support the boys overseas so there you go. Anyways this is a good story because starting off with Colin and Jim I think we are looking at narrowing things down to issues of we know what we can do with this data if we have it and what sort of data do we need we don't have to wait for a Manhattan project to produce the sort of technologies that would be useful to us, they are actually there now it's a question of I would call it the social and political will to use them in a way which addresses the issues that we are facing going forward so I will give you a little bit of I may promise me this is one of these we call them now iconic images how many have seen this before this is Olympic Dam we have one, alright two so here the little bit of red at the top is one of the largest copper, gold uranium deposits in the world called Olympic Dam it's situated in the state of South Australia it's found back in the 70's it's now being mined by BHP but the important part from the point of view of this discussion are the images in color above and then with line work that show the seismic and electromagnetic patterns associated believe to be the root system of this giant ore deposit and this is what we call a footprint they are calling now the fingers of God coming out of Australia the arrogant Australians but anyways I told them I don't think it's necessarily the fingers of God but anyways they got naming rights so this is something which obviously we can take advantage of we're acquiring data sets like that here but I don't think any of the major deposits like for porphyries or in Carlin have actually had the same type of data acquired so one of the things isn't the invention of new technologies but the application of them in areas that are of interest to us and how do we calibrate moving over to terra firma here certainly we haven't been slack in terms of acquisition of regional scale data but even though there seems like a lot of dots these are I think about 70 kilometers apart which is really a semi-regional assessment this is a continental scale assessment and Carlin referred to kind of a sweet spot scale which is something I think we have to look at we need to tighten these things up and we can in order to come up with allow us to actually break apart individual basins look at specific geological formations that might host deposits we're not talking about actual target detection we're talking about the mineral system that these would be hosted in how do we characterize these and then hopefully with the suitable input from government academia the private sector can actually go in and do the hard slog and spend the big bucks with the drilling but they need to have it de-risked and they need data to de-risk it and this is I think where we're at this is helping there's been some really great projects using this data but you can see there's a large part of the US is actually doesn't have much coverage and notice a good part of Nevada as well as Arizona which are major mineral provinces in this country don't have any dots there are some dots out there but they're probably in the private sector and not available and some universities have probably taken some as well so something to be filled in also increase the resolution would be good Australia's busy with their program looking for more fingers of God they called the Aus lamp they started up much later then we did here and so they're probably working with advances in technology which are good getting longer frequencies which allow us to see deeper and they have a very systematic approach churning through probably in the next five years they will cover the entire continent Australia is a place with 80% 80% of that continent is covered by foreign material so they have just like Jim's comment about looking for the geothermal in Nevada a lot of under covered stuff that people don't even know exists because it's covered and so you need this sort of remote sensing to see it here in the states this came out last spring an initiative by the USGS and they called it the filling in the gaps which unfortunately a lot of those gaps and a lot of the existing data is circled there and that piece is sadly out of date I was actually starting in this business in the mid 70s when the major airborne magnetics initiative happened in this country totally by chance it was actually a radiometric survey triggered by the first oil crisis and they were going to fly the entire country with radiometrics then some trouble maker at the USGS put up their hand and said there's a magnetics to the survey they said as long as it doesn't interfere with the radiometrics we'll get magnetics data and that today is about I think 80% of our coverage in this country is data that was generated in that period mid to late 70s we should be able to do better here's a vintage map this is a qualitative ranking but I think the important thing is red is good 4 and 5 are pretty grotty you can quickly tell that most of the data we're working with in the continental US is pretty grotty but what do we map well we actually map the planet really well other than our own other than our own country you show this at an international conference and it's like heads are bowed down this is not something to be really proud of and the very few places that have high resolution aren't really probably really focused on minerals which is fine I mean if you do this sort of national mapping program it isn't just for one constituency it needs to be for everybody and it's a legacy that will go forward for the next 20 to 30 years just as the first data set did this is an investment in our future there's another initiative to look more from the USGS looking more at specifics in certain areas which is good increase the resolution image here the idea of you got to get more layers add gravity to the mix as well as electromagnetic and magnetic data so we have largely the technology we need it's really getting out there and putting it into platforms that allow us to acquire the data cost effectively and quickly here's my little cartoon of just or sketch at the back of the envelope if we were to do the sort of surveys that I think Colin Jim and I are quite familiar with through the geothermal and the mineral side of things somebody else in the room if we were to cover the entire continental US which is about according to my wiki site about 3100 square kilometers and if we were to use a line spacing for a survey of about 500 meters we have about 20 million line kilometers and it would cost us about just a little over a billion dollars which when I went on again another wiki site is about the cost for the last two missions to Mars so for what we've done in small part to explore the rest of our solar system we could actually contribute an enormous wealth of information to our country that has not been done and I think Remi you can probably see if there's one more but I think that's it there thank you I'll now open it up to questions I'll start can drone get these data are you thinking that direction drones should certainly be part of the story the ones that are in commercial use as opposed to military still have issues of payload size and endurance size so if we were to commit to using the sorts of ones the military have access to they could replicate the platforms that we use commercially for doing geophysical surveys but those when they cost five to six million dollars each those particular ones whereas would be fairly cost prohibitive for at least in the commercial world because I was thinking surely they'd be cheaper than flying people to get this right? it's not so much people it's you want a multi-sensor platform and these have existed but if you want to put a magnetic sensor a gravitational sensor a lidar system a hyperspectral system I mean these have to be scoped out but the commercial world actually has built systems like that one was used by BHP and Chile they called it the Griffin and it was a four-engine aircraft but the commercial world just didn't have enough use for this and it was decommissioned if I may add to that it's a really good question and a very logical application there's a major regulatory hurdle right now because the FAA requires for domestic applications of these drones that the pilots on the ground have to remain in sight of the drone you need to address that issue yeah it's like Susan's thing about the regulatory issues there are things that would have to be looked at and of course large part two who's doing it if it's some paparazzi you worry about loosening the things up flying around Hollywood but if it's for a program that the government is supporting and you probably would get one-off permissions Colin? Thank you just a couple of quick comments there are similar anomalies in the Great Basin like the ones in Australia in a rough way and it's not clear in a lot of cases how they relate to mineralization or geothermal fluid upwelling and some of the work we're trying to do in places like the Gaps Valley basically we're figuring that we need spacings of these stations on the order of 10 kilometers or maybe even less to get the resolution and so while the U.S. array data just to underline your point the U.S. array data is really useful at this 70 kilometer spacing it just doesn't give us the spatial resolution we need for the crystal scale features the other thing just to add on the drone issue it could also be true in the context of traditional aircraft and geophysical surveys in working with drones we've had some great success in collaboration with NASA who of course everything they send to other worlds is essentially a drone of one sort or another these days is payload directed flight where you can actually have onboard computing the analysis of the data it comes in and alter the configuration the spatial configuration of the survey depending on whether it sees anomalies or not one of the challenges we have it's not a great challenge but we could do better is that when you buy a survey you basically have to tell the company in advance the spacing and the layout of the survey and whether we use human or robotic pilots being able to have that real time information come back and scale things would be a significant advance too Tom One more question this for Jim I'm wondering do you know if like hyperspectral remote sensing has been applied to trying to tease out perhaps very minute changes in mineralogy alteration chemistry to help delineate geothermal systems yeah it has there's a couple groups have worked on that including some folks from UNR and it has netted some positive results borate minerals for example and clays and so on indicative of recent geothermal or hydrothermal alteration so yeah but it's just a few kind of local studies but I was going to maybe can I add one thing to the previous discussion in drones and geophysical surveys and all that are absolutely essential but let's not also lose sight of fundamental geologic mapping and LiDAR of course can facilitate that but you still need feet on the ground and so on and in a place like Nevada or much of the Great Basin where it's primarily federal land it's very difficult to get funding for detailed LiDAR studies and then just in general geologic mapping if we take Nevada about 25% of the states well done in terms of geologic mapping about half of that has been accomplished by the USGS and about half of that by Nevada Bureau of Mines and Geology so there's so much more left to do there in addition to the geophysical surveys and you need both to really understand where these resources may lie Next question for Azada about the fracture modeling, natural fracture modeling and induced fractures so one of the things that we've noticed in the research that we do is when we have we look at the density of natural fractures that occur in a borehole image and then you compare that to the stochastic models that we build if we kept that same density of fractures it's huge, it's absolutely huge and even if we kind of do some cleverness to coarsen that a bit it's even far too big or far too the density of fractures and the distribution of their length types is far too much to really be managed by a typical dynamic simulator so what we have found and this is kind of leveraging some of our experiences in the past is that just like when we build the high resolution earth models that geologists are so fond of and we find that going to the dynamic simulator we need to upscale them it's not the same if you build those earth models at a coarser scale to begin with and then just feed it to the engineers, there was a thought in the past that if we just use the scale of the geophysics the seismic, well that's just fine because it's coarser than it is based on the well board data which is much more highly resolved in the vertical domain so what we have found is that coarser it's important to build an engineering-scaled natural fracture model that is upscaled from a more dense model because otherwise you're just going to be waiting a very long time for those simulators to actually process the information and there's some cleverness with respect to building those upscaled models and this is an excellent area where machine learning could come into play where you could do some testing against the dynamic simulator with a highly resolved set of fractures on a small area and then upscaling it to an efficient set which gives you similar results I'm wondering if you've done anything in that area or you've experienced the same thing. We're dealing with the exact same issue even for a very sparse fracture density putting that density into a reservoir scale model would result in a lot of fractures. In one case I actually took this slide out a 10 by 10 meter fracture. If I exactly represent the density that we got from the field it would result in 13,000 fractures that in a reservoir scale model would mean million. We are even incapable of creating that model with the current CPU of a desktop. So we didn't take a machine learning approach but we do we exactly follow this line that goes back to my last point on this slide. Coming up with the research methodology that we get these fractured densities and then we create fracture network at different scale is an area of research and what we do right now is that we try to come up with depending on the scale we try to come up with the most dominant fracture characteristics so for example we tend to simplify the smallest fracture in the system disregard them knowing that it's an exponential law and replace that amount of try to keep to match the permeability but then replace the porosity of those fractures with a matrix permeability so we create models of different size and put pressure boundary condition and match for example hydraulic properties and mechanical properties but then again if we go to thermal that can completely change the story because then one fracture so that goes back to the point of understanding this simplification methodology depends on the problem that you're solving for a mining project it might be different compared to EGS project. Our approach is a scientific approach we don't do machine learning we try to build samples and take part of the smaller fractures out and replace them with the equivalent properties. Yeah so it's interesting I think it's relevant to point out that there's some good research to be done on finding the effective scale of the fractures for these different domains whether it's thermal or whether it's petroleum or whatever hydrology find the effective scale that makes sense and that in part addresses what Linda was talking about earlier with respect to finding the right scales of these particular features. Yes I want to pick up on what you're saying it's very very accurate first of all let me say to Elizabeth I have found this the most stimulating meeting I've been in many years it's very great that I'm delighted but the underlying philosophy of Itasca has always been to try to use the mine or dam foundation or whatever as a laboratory and we went there with the idea of actually they have offices in Australia and Chile where the action was but one of the most fascinating areas that we've been working in has been block caving and if you know anything about block caving it's an attempt of huge gravitational forces with the rock to fracture and the process of extraction of the mine causes major redistribution of stresses etc but recently and this goes to the question we were doing it largely by seat of the pants picking some boronite oscillations and saying what size of fracture should we use in trying to see if we violated the principle of minimal potential energy and so on but a recent study on a block caving showed some behaviors that were quite unanticipated for example that a tunnel converges when you unload it now why was that? because in the earlier stage of the loading of the block caving process put high normal stresses on fractures and later those were released and so they could slide when they couldn't slide before and so this notion of using a mine as a laboratory to me is a very powerful one where there's a great deal of you know I don't know where we'll lead to with getting ideas for practicing engineers underground how they do it but and then I look at the amount of funding that we get for looking at these things on the earth compared to what is for other planets and somehow we have to make the case that we have a major set of problems here that are very important thank you for the fracture modeling I just saw last week something that really caught my eye I think I can share with you who it was but I have to check to be sure I remember right but the thing about fractured systems is that they very arbitrarily connect things at odd distant scales and the model I'm thinking of that's what it focuses on and so swarms connected to swarms and different length scales and it gets away from the need to specifically grid around the fracture system which is impossible as you see there's too many but rather interconnect them and I've also seen another effort that I think is related the author is a cunha that one I do remember I mean stuff is out there that's going a little bit different direction very interesting going back to the pure point about fractures randomly connecting so this is the nature of fractures the fracture size distribution is often exponential having very large fractures at a scale of kilometers and going all the way down to meters and again this goes back to the concept of understanding these fracture systems and being able to put them in the model by respecting their characteristics the dominant one and one of them is this fracture size distribution which naturally gives rise to the type of mechanisms that you are mentioning so part of it is understanding what their characteristic is in nature and being able to reflect that in the numerical modeling and we are able to and again this also goes back to Jeff's comment that then we can look at these models through numerical modeling look at what fracture connect which region to what region so basically come up with the connectivity map of the fracture system that we have put in the model so regarding the role of natural fractures there are actually good news now so some data emerging from some of the field experiment especially the hydraulic fracture test site in Permian basin so the data released last year and this year shows although we have this kind of foot spacing natural fractures only about a couple percent actually participate in the flow so that give us a spacing of like a 10 centimeters so we are seeing the same thing in our colab experiment we have 400 or 600 natural fractures locked in the system but only the active ones are in the spacing of 10 centimeters so that is kind of a sweet spot for the current model so it is we don't necessarily need to model all the natural fractures most of them are sealed and they are not permeable in the kind of scale we care we still have time to talk about this so to some extent to overcome some of the issues of modeling natural fractures the approach the research that we have taken is to go to a particle based system so we are not using grids at all and that is the work originally that was done a number of years ago by Crenshaw at Stanford and then eventually after that from Shavasthava Crenshaw used a finite element based model and then Shavasthava turned it into a geostatistical based model that has growth at the point etc so if you read that stuff you can understand it and then the gridding comes later so you can use unstructured grids to get very very fine cells within the fracture domains and then the matrix goes to coarser cells and you can capture that information we found that works pretty well but it doesn't negate the need to develop the engineering scale that is critical for the natural fractures so I am still a believer that you build the natural fracture model with high density I don't like the idea of building fractures based on grids I think a particle based system is better and then you coarsen it to an engineering scale using unstructured grids that would be kind of the approach that we have been taking and it works pretty well I am not familiar with the way that you are mentioning to creating the grid I completely understand the structured versus unstructured mesh I just don't know how that would conform to a discrete fracture modeling the way that we do that we use the log data and create fracture network initially so independent of our finite element or finite difference underlying mesh we create a statistical model this is a Poissonian distribution of fractures in a space with fractures in 3D in a 3D model and then we create an unstructured mesh that would conform to the geometry of the fracture now during that fracture generation once we create that 3D fracture model then we take those upscale modeling approaches so we first create a very dense fracture network and then we say that which one of them we can represent with a background continuum model out of the model and which features we want to keep and then reduce the number of fractures that we have in the system by respecting mechanical properties, hydraulic properties and then once we reduce the number of fractures brought it to a manageable size we create a mesh which is unstructured and conformed to the exact geometry of the fracture I have a final comment perhaps the last one before we break for lunch but just in response to your ideal data acquisition for the US the Australian acquisition survey that was done in late 2000s and about 2008 and it was called the Australian Airborne Geophysical Survey AWACS and it cost them 2.6 million to fly the whole continent at 75 km spacing for the airborne radiometrics and MAG and now they estimate that would be about 200 million so I guess my comment was maybe it's not as bad as in the billions for the US you know I mean that's similar size countries so we could get away with cheaper and we don't have to worry about not going to Mars so ok well please join me in thanking Joe ok we have looks like lunch is ready so we will adjourn and reconvene at 120 so we purposely schedule a full almost full hour for lunch so folks have a chance to interact informally so please take advantage of that time to network she might have we're buying on time she kind of has to keep going all the way through so oh no we have a break at the end of the panel before we do the discussion I I ok Bridget if you want to folks please grab a seat and we'll we'll let Bridget get started with our third panel hi everyone I'm Bridget Ailing and it's my pleasure to welcome three panelists this afternoon for our third panel we have some really interesting talks looking at aspects of dynamic permeability first up on my right here we have Ping Cheng Fu who's an earth scientist at Lawrence Livermore National Lab he received his PhD from University of California Davis and has been studying dynamic permeability for about 10 years so very happy to have him next we have Christine Eleg Economedes who's a professor and the Hugh Roy and Lily Kranz Cullen distinguished university chair at the University of Houston previously she was a professor at Texas A&M for 10 years and worked on slummajay before that for 20 years so a wealth of experience she has more than 120 technical papers and two patents at the end we have Kuchuk who retired from slummajay in 2017 he was the chief reservoir engineer for slummajay and was a consulting professor at Sanford University in the late 1980s he's worked on many projects including work for BP and has also many papers published one thing I noted that he was also the Nobel laureate physicist I can peep the gold medal awardee which sounds pretty darn impressive so it's a pleasure to have you here so we'll get started first with and we'll hold questions again until after the three panelists have presented good afternoon so on behalf of the EGS CoLab experiment team so I'm going to share some of the observations and insights we obtained from the ongoing field experiment so the system motivation we have it's a project sponsored by the geothermal technologies office so we often see this kind of lab experiment the scale is from millimeter to you know half a meter or foot scale and occasionally we have the fortune to carry out a real field experiment such as you'd have forged sites but there's a vast scale gap in terms of scale between these two and if we're talking about geothermal thermal processes hydro and mechanical process they have different scaling rules so it's not very straightforward to actually extrapolate what you get from the lab to the field so the first motivation for this field experiment is to bridge the vast scale gap we have several older magnitudes of scale gap between these two kinds of experiments and we also want to validate the EGS codes we developed in the past couple of decades in the environment that's relevant to the field operation so for this to satisfy these objectives we have made several strategic choices so the first one is we carry out the experiment in a mine that's 1,500 meters deep so the stress state is relevant and the temperature there is not very hot it's about 35C so that's not ideal but it's very friendly to operate and to make a measurement so we partially compensate by circulating chilled water about 10 degree C so we have a 25 degree C of temperature difference and we also made a very heavy investment in characterization and monitoring and we are carrying out this experiment using a community based management approach so we have a fairly large team the experiment involves 9 national labs and the 6 universities and a few private entities it's led by Tim Nipsey of Lawrence Berkeley lab and Dr. Blankenship who is in the audience so here we have a list of names involved we have more than 100 people involved and about 15 to 20 people are heavily involved so we have at least two experiments the first one is concluding that's what we have been doing the past two years it's about hydraulic fracturing the second one is starting it's mainly to investigate shear stimulation so both experiments are in the Sanford underground research facility in South Dakota that used to be a gold mine you can see the equipment along the drift here so here this is the kind of conceptual model we have about the experiment so the idea is this long line that's the drift we wanted to create a couple of hydraulic fractures and we drilled 8 wells we used to stimulate and circulate the fluid and there are 6 wells in different angles trying to provide a very geophysical coverage of the experiment so the blue one this is injection and the red one was the production well so before the experiment we spent a lot of time and effort trying to characterize the site, the testbed so we retrieved a continuous course from the 8 wells about 500 meters and we blocked all the fractures with all kinds of properties we built a discrete fracture network that's almost deterministic and this is the inversion results from PNNL about the electrical resistivity of the rock body it provides very useful information about the fabric of the rock in that testbed and we deployed a lot of sensors so we have literally 5 miles of cables and we have fiber we have acoustic sensors temperature sensors in the 6 monitoring wells so we have done a few simulations this is the fracture we're using now for circulation so we did this carefully did this in 4 or 5 little episodes we'll check a little bit see what happens so this figure shows this is the fiber measured temperature in one of the wells so every time a hydraulic fracture hit that well, hit the fiber you'll see a bright spot so corresponding to the injection history and we have continuous fiber in all the wells so this one happened to be hit by the hydraulic fracture and we also have cameras we'll put cameras in the production well so we actually captured a precious footage showing the actual intersection of the hydraulic fracture with this well bore and we did some analysis to see the trajectory of the hydraulic fracture so on the right you see a lot of dots that's apparently micro seismic so you see we can see several planar features and we believe those are newly created hydraulic fractures because there was no we have a good map of that natural fracture there was no natural fracture in that orientation so if you take a closer look at the micro seismic you actually see two planar features one here and one there from the results of LBL and Oak Ridge National Lab and these two fractures correspond to these two temperature anomalies we received we saw so it's just very strong evidence to show where the fractures are we have multiple ones here and also showing the black circle here that's the natural fracture we know we saw that natural fracture big natural fracture permeable in all the wells intersected so they have very consistent orientation so we have very high confidence in the resisting of the natural fracture so when we inject fluid from this well we get fluid both from this natural fracture intersection with the production well and the hydraulic fracture intersection so starting in October last year we did some circulation test the first one we did about one month of circulation test so this is the data from the first four days so we kept constant injection rate 400 milliliters per minute basically just drink a soda in one minute that kind of rate so the pressure was stable the injection pressure so that's what we know as the fractures limiting behaviors once the fracture is open you don't need much more pressure to push the fluid in so everything was happy this is what we had expected and we were getting 90% of the fluid recovery from different wells just noisy data and from the DTS the temperature we solved one single or hydraulic fracture showing up in this map so it was a very happy moment and we achieved what we wanted then but you know real experiment never rarely does what you hope then the pump broke so this is the dynamic permeability evolution part that broke we put on new pump and that pump worked with a higher rate then we start to see these micro seismic events these red dots and turns out we were actually slowly propagating this fracture in 24 hours we solved consistent micro seismic and at some point we start to see temperature signal in another well so this signal is where the fracture so later we learned that the fracture was propagating then when we went back to the lower flow rate then the pressure start to increase continuously at a very high rate that's not acceptable because it becomes unsafe then we did a little bit stimulation try to already open the fracture then we come back the fracture was still increasing and we put a chiller on the fracture water at like a 10 to 15 degrees C the pressure drops a little bit thermal mechanical effect but it starts to increase again so during this time we did seven tracer test and I'm picking three recovering curves you can see from three channel each different color represents a different channel you can see we have very different breakthrough curves at these three moments so it's a very dynamic system it's very stressful and confusing time because why the pressure was increasing and we cannot control that it turned out it might be it was likely something to do with geochemistry and when we finished this we pulled out the tool we saw the mineral deposition on the tool so it just hypothesis now is the mineral precipitation was blocking the fracture so we later resolved that problem since April this year the circulation test has been largely continuously carried out for now six months as we speak now the circulation test is going on so the system behaved reasonably well and we resolved the water source problem so the pressure was still going up and we see this just that it goes this kind of continuous very slow increase continuous increase we thought through some modeling work we believe that's poor elasticity so the simple way to look at this is to open the fracture we need the pressure to be higher than the total in situ stress in this kind of elevated pressure the pressure is going to diffuse into the rock matrix and cause the pressure to increase like this and through the poor elasticity the poor pressure increase will cause the total stress to increase and you got this feedback to cause the pressure to continue to go up so that's the part we can explain there are still other parts we couldn't couldn't explain so every once in a while the pressure will suddenly just drop a couple hundred MPa sometimes they increase a little bit we don't fully understand those we're still trying to do modeling interpretation of the data the recovery rate from different wells, different channels they evolve all the time although the total recovery ratio we have is somewhere between 80 and 90 percent so I'm going to show you a specific window we saw something very interesting so this window is from April to May and in this time window from this well there's a very significant signal in one of the DTS along one well and in May we didn't do anything we're still pumping and this signal kind of disappeared we know at this location because we have such a good mapping of the fracture we know there's a natural fracture naturally open fracture there at that depth this is from the televiewer log this is from the core we know there's a fracture but after about a month of injection that fracture used to be conducting fluid then you lost it at the same time the recovery rate from that well went down and the flow rate in other channels went up so we know that fracture was for some reason closing it's actually very nice to have this kind of well characterized testbed to definitive tell what's going on in the experiment so we have learned a lot in this uh experiment and I'm a modeler most of my time spent at the lab do geomechanics and flow modeling so we were involved in the design of the testbed we thought we have cubic law, we have cup holder model HPC enabled modeling this is going to be a slam dunk but in this two years experience I'm very very humbled so one thing we'll learn is this kind of deep intermediate scale experiment is a very effective way to learn about subsurface and also to test and verify tools it's so nice that when you see something you actually know what's going on for certain so in some other kind of experiment you might have 10 different explanations for the phenomenon and you don't have a sufficient constraint of the physics so the investment we made in site characterization and monitoring paid off handsomely so but there are still behavior and processes about fractured permeability that we don't fully understand so after we learn this many people look at the data it might become predictable and exploitable as we learn more so we're working very hard to get data available to the public so we're releasing the raw data process of data so here all these little things these are the notes people took during real time experiment this is this is one week of data so we're making that available so more people can look at the data to help understand more so that's what I have for sharing today thank you good afternoon so I'm going to be showing you other people's data and we'll focus here on the source rock plays the unconventional so just a couple of quotes to start with first of all impermeable rocks impermeable rocks do not exist in nature according to irison et al. 1994 fascinating observation and very well defended and secondly Roberto Aguilera told me just last week all rocks are naturally fractured so there you go we'll start with eagle ford source rock and looking at some fractures that are circling yellow in case you have trouble seeing them and they've both with bedding planes at zero degrees to the drilling and at 90 degrees to the drilling and then these same cores were put under 500 psi confining pressure and there's no trouble to see the fractures once that's happened so I think why this is relevant is because when we hydraulically fracture we are changing the stress and potentially doing this kind of thing to the rock now we're switching over to eagle ford and looking at the micro seismic surveys micro seismic was acquired during the hydraulic fracturing of wells P2 and P3 so let's see if I can figure out how to make that work over here, nope I don't see how to make your arrow work anyway you can see the P2 and P3 well trajectories there and you can see that these wells have been drilled more or less on minimum stress so that the hydraulic fractures are perpendicular to the wells and these micro seismic events are primarily sheer from what I understand and also in the lower picture we can see that they're pretty limited in height growth to see where exactly this is in the map on the right you can see the eagle ford trend you can see the wells that are green as oil the wells that are red as gas and then the star if you look towards the right side of that trend you can see where these wells actually are and this particular study was one of the field sites and it's absolutely fascinating the observations that they were able to make they drilled some wells especially for data gathering and so you can see those P2 and P3 wells again on the map view of kind of an idealized well trajectory map on the left and you can also see these observation wells S3 and then sidetracks from S3 labeled 1, 2 and 3 and you can see that S3 paralleled P3 and I guess it's on the order 70 feet distance seems like that's what I've been told and then some of the wells were drilled above P3 so trying to get an idea of how the fractures are by drilling through them after the hydraulic fracture treatment work they also cored through S3 and so the upper core that you see we're seeing fractures that resulted from the hydraulic fracturing we also see the core spread out so you see the images of that as flattened and then you can see the borehole seismic image and it's very interesting because you're seeing swarms of fractures 100 feet away from the well that was generating those hydraulic fractures so it's really interesting to try to imagine what's making that happen on the right we see that in some of the fractured areas they did see propellant the upper image shows propellant on the face of a fracture in the core and the lower one showing embedment and then the core image at the bottom is showing the distinction between unpropped and propped fractures so the propped ones of course are the main hydraulic fractures the thing is if we are getting swarms like this that are more or less parallel to the main fracture this is actually not doing that much for the enhancement it's kind of widening the hydraulic fracture but it's not taking you far into the shale necessarily so it's very interesting to see this kind of data and imagine what it might be telling us this is a different example it's up in the DuVernay shale shale acquired these data on the left once again micro seismic on the right an interpretation that helps you appreciate the fracture envelopes and it's very interesting what they did here the well that follows minimum stress so you can see it coming sort of northwest southeast is perpendicular to the created fractures as you can see in the micro seismic and the one that went north south you can see the fractures are at an angle so the minimum stress direction is not varying aerially here and we can see that the fractures in the north south well appear to extend a lot farther than the ones in the well drilled on minimum stress well they did something else they used a polymerized thick fracturing fluid for the minimum stress direction well and the off azimuth or north south well they used slick water so their observation is really counterintuitive because the on minimum stress well has much better productivity than the north south well despite what you would expect from what you see there and I want to point out a couple more things to you the first well was the one in the minimum stress direction and the second well was fractured right after the first one so there was elevated stress when they fractured the north south well maybe it's my imagination but I would say that the fractures in the north part of that well are avoiding the fracture area of the elevated stress due to the first well and I have seen that on other micro seismic images so the fractures will they're lazy just like people are they're going to go where the stress is lower and so they've gone to the left and there is an actual there's another development to the southwest where you see that marking maximum stress and there the fractures actually may be avoiding elevated stress to the right so I think this is really interesting case and I wonder whether the well on minimum stress might be producing some of the fractures in the north south well and that's why it looks like that well produces more because it is really counterintuitive to see that and on the right you can see these trends how the micro seismic company interprets the fracture intensities just an outcrop here illustrating brittle and ductile this is in the Woodford shale as we go to drill these horizontal wells the operators will look for brittle and ductile even while drilling you can get this without even logging the horizontal well and the fractures the hydraulic fracturing is much easier in the brittle and ductile rock I don't think they really do it but they could be altering their stage by stage hydraulic fracturing according to whether the stage is dominated by brittle or ductile but I don't think it's what they do another thing that's interesting I like this picture I think the red cylinder is representing the idea of a horizontal well we're looking at an outcrop and the possibility that the fracture starts up is the blue arrow and then makes a T and T's and T's and shifts and does all kinds of stuff according to those layer interfaces and this would be very counterproductive to the height growth of your fracture because it's going to be hard to prop it around those sharp bends and even if you do manage to do that in the subsurface your maximum stress most likely is working like crazy to crush that propane so it really enlightens for me how we should understand production data in light of the possibility of this sort of behavior salt as well it's a tracer and in the shale gas wells it's been observed that a small fraction of the water injected for the hydraulic fracturing actually comes back as flowback most of it stays in the formation and it may go in fresh but it comes out essentially saturated with salt so 10 times saltier than the ocean even and so there's a lot of salt being contacted by the hydraulic fracturing fluid on the left we can see a Marcellus core with very clear fractures that are salt so natural fractures are nice to us in conventional formations but in these source rock plays they're healed if they were working for us we wouldn't need to be fracturing the way we do hydraulically but the hydraulic fracturing is altering stress and causing these shear events that the seismic sees and when they're connected those could be tensile as well so a lot of activity in this material that I think is very different from the tight rocks where traditional hard rock fracturing got its start also the graph on the right is the reduction data and looking at the salinity versus time with an interpretation that when the salinity increases a lot more before it levels out that might be related to more natural fracture or secondary fracture generation then when it levels out at a lower salinity value so I think this is pretty interesting possibility some work out of UT looking at modeling fractures and as we get into these DFN models it's very important not just to model where the fractures may have gone and even use the micro seismic to guide that but as well where did the proper go because there's a strong possibility that unpropped secondary fractures are not going to stay open as you produce because the lowered pressure due to production will actually aggravate the stresses against those fractures and tend to close them on the right a simulated comparison for the fractures modeled on the left and then for the left picture the left simulation picture assuming that the unpropped ones are closing and on the right assuming they all stay open more or less with the same width a big difference there in my world I'm a conceptual model we might consider these wells more or less what you see here map views of the wells and you can see here again kind of helping you appreciate what the Shell wells in the Duverne may have been doing so the north well north south well is like the one on the top and the on minimum stress well is like the one on the top and the north south well is like the one in the lower picture and let's see if Lynn this is for you he was interested in what people are thinking about how to improve the recovery because the oil recovery is very low just primary recovery 5 to 10 percent but operators are trying already cyclic injection and production Huff and Pop or I'm just going to call it HAP, H-A-P typically using produce gas that's stranded gas anyway but we've also been looking at completions that would enable simultaneous well alternating injection and production so the idea here is that every other fracture you complete the well in such a way that every other fracture is either on injection or on production and try to produce simultaneous injections and productions between these created hydraulic fractures and the third case there the single well alternating production you don't do it simultaneously but rather you inject the injection fractures then you switch the flow controllers and inject and produce the producing fractures and why do we go to this level of complication why don't we just do well to well well people have tried it and I wouldn't have done this so here's a bucket water flood experiment where they saw the wells already there and you see those horizontal wells but they're all also hydraulically fractured imagine the fractures that I drew in the simplified view on the previous slide so the minimum stress direction is north south the fractures are going to connect between the injector and the wells on the east and west of that injector and sure enough they did I think this should be expected and what we are doing with these wells is not amenable to well to well unless we start deliberately creating the fractures so they're not intersecting so this is a modeling case looking at the Huff and Puff and this is in the Eagle Ford Nexon Energy did this one the swap well as I mentioned the idea is this is a picture that helps you appreciate what I'm talking about zooming in on the fractures we see every other fracture red if it's on injection green if it's on production and the advantage actually of swap over the simultaneous the simultaneous could be subject to the risk of things like a cement connection and if you had let's say we have 50 of these displacements going on but one cement connect that could take over the well the risk is very strong in this kind of a configuration and you're putting a lot of jewelry in the well so you don't want a risk and we find the swap to be almost like self-healing that is far less vulnerable to a short circuit but anyway yeah this is my last slide this is just a comparison of the three so Lynn looking at possibilities of far better recovery and even the Huff and Puff already is improving what's being done and here just looking at NPV so I hope this I hope you find this interesting the way the fractures work in these source plays is quite a bit different I think from what we hope to do with things like geothermal wells thanks very much good afternoon I'm going to talk about permobility of fracture systems in respect to modeling rather of modeling but I'm going to not include everything what goes to model I'm going to refer to permeability with respect to fracture modeling so basically we have three types basic fracture system basement which we have a bunch of reservoirs in Vietnam Yemen, few other places sandstone few in Texas but most of the fracture reservoirs they are in carbonate systems especially in Middle East I put some of them I'm also lucky that I worked with most of them as Christine said in all carbonate fracture reservoirs they are faulted and fractured okay but in terms of permobility fracture permobility as you see Christine I didn't put zero approximately zero okay so basically you have matrix almost no permobility what permobility comes from the fracture aperture more aperture you have better permobility you have also that on the density of fracturing that is really one of the most important parameters in basement reservoir density and the aperture sandstone reservoirs this is really one of the best defined permobility especially core permobility most of the aeropermobility knowledge comes from sandstone because a lot of reservoirs in texas you know sandstone and these as I developed it very early 60s even much earlier so most of our knowledge about permobility comes from sandstone but sandstone permobility is homogenous very little anisotropy meaning the very little fractional permobility and because of lack of diagnosis fracture conductivity varies with aperture but not a lot actually so sandstone first of all there are few of them fractures fractured and it is actually most of things is well defined when you go to carbon reservoirs as I said almost all carbon melted and fractures great geologist Nelson said if you don't have one that means you have not discovered when you have a fault that implies you have fractures in fact with seismics we don't see only 60% of the faults because rest they don't have no throw you will not see seismic although it is improving then fracture reservoirs as you know there are two types one continues the fracture reservoirs meaning part of the reservoirs fracture communicating each other I don't think there are very few reservoirs very few of them they are completely continuous but most the reservoirs are discreet the fracture reservoirs and three examples left you see you have a fault clearly you see the fault bunch of fractures again that's another fault maybe you know you can see the throw and then many fractures because of the faulting on the right you see this some these are outcrops of course these are two is they are the example for continuous fracture reservoirs as I said there are very few of them in the nature so most of the fracture parameters normally exhibit power load distribution actually this is a little bit good news although highly heterogeneous but at least we know that they have power load distributions basically whether these hundreds of fractures may have high conductivity while thousands of millions of them have low in fact you see that curve is fracture spacing very few fractures has you know above the 10 feet or 100 feet but thousands of them has very small spacing so that is what I consider fractures are basically macro fractures they are primary fractures they are open they provide the most of the conductivity second one is secondary fractures they are short their aperture is much smaller and they enhance the permeability third one is really the tertiary fractures they are much smaller some of them they are even not open but still they they increase the permeability affect the permeability when we look at these pictures we have okay we have three kinds of fractures but that is not enough if we don't know the orientation of these fractures if we don't know the density of those fractures length of them and the height of them knowing only the permeability is small part of the you know the game basically these three or four parameters are essential to do the correct you know the fracture natural fractures rather were modeling of course permeability is very important you have to have it other ways anything doesn't mean it doesn't mean that you cannot produce but what I'm saying that these are other parameters orientation density spacing length in both vertical and horizontal are very important parameters to characterize the you know whether geothermal reservoirs whether you know the carbon reservoirs in fact I'm not familiar with geothermal fields in Turkey most of them they are carbonates are fractured heavily so basically carbonate reservoirs you have core permeability of course you can get a core okay but the one again complexity most of the carbonate cores have walks you know in these figures you see the you know the walks you know you can take any part of this you know the core if you measure permeability is going to be different is going to be different in horizontal direction is going to different with vertical directions now it's very important especially from image logging getting the the walks and then do some statistic or just statistic well to well correlations get some idea of distribution of these walks because they enhance permeability considerably okay the the hygienic is significantly effective fracture conductivity particularly mineralization and scaling and of course these two you know mineralization and scaling is reduced to permeability that makes actual fracture system much more heterogeneous because some of them mineralize some of them open many faults and thousands of primary fractures and millions of secondary tissue fractures with different apertures conductivity length orientation size spacing exhibit power load distribution and that is little bit helps in the modeling we have to expect that variability power load distribution as I said micro fractures and walks considerably enhance mass experimentability and it becomes scale dependent one thing that I might add micro fractures and walks in the unconventional especially the carbonate shale they have out of macro fractures and my personal belief that because of those fractures and micro fractures we are able to produce this none none of Darcy reservoirs so basically again when it comes to model I try to say you know first you have to model basic open fractures with the sizable length explicitly you cannot upscale that you cannot upscale this you know the one you know 10 Darcy with one macro or one you know mini Darcy you cannot upscale that you have to respect these features because they are going to dominate the behavior of the system so forget the upscaling fault and major fractures then then what we do with millions of small fractures not one or two or three so actually first we have to do upscaling in this case if their permeabilities are not very high second you know we have to use image works you know to and small scaling you know the pressure transient test and maybe well test calibrate them and then what happens you will end up with effective permeability in the what we see basically effective permeability of fracture reservoirs is the matrix permeability permeability of the wags permeability of macro fractures and secondary fracture fractures and this gives you after the upscaling effective permeability no not now our system is having major faults explicitly in the model finite difference as you said may not work very well in that case then upscale those smaller features and then you have a matrix so okay one thing that I put this scale for shape formation things are much more complicated so as you see you start with the matrix or macro fractures etc you have nano scale diffusion gas this is I'm talking about gas not the oil oil does not have that much absorbed gas but the gas formation shape formation have a lot of absorbed gas probably there's no gas in the absorption state state then the pores itself because pores is low so basically we start with diffusion from the nano pores then it goes to you know macro scale which is the normal where the the core itself and then there the flow is mixture flow not necessarily darsian flow then we go to macro scale which is you know you have many fractures macro fractures etc and these are like a dual porosity then where you have to do upscaling then you have a macroscopic scale explices the other fractures and finally you have hydraulic fractures so as I said you know compared to carbonate natural fracture reservoirs this is much more complex due to diffusion due to desorption these are non-linear things it's very difficult to solve with many many parameters finally I'll come to last slide in the oil industry many years I mean maybe 50 years we use the sugar cube model whether numerical, whether analytical it really doesn't make any sense and many many years we thought that this sugar cube model is a you know epoximate model for fracture system it is not basic and simple models for fracture reservoirs but there are few reservoirs they can be applied where you have very very small matrix you don't have very large explicit fractures you don't have faults and in this simple model we have two parameters lambda omega which is one is related to matrix parameter the other is porosity when I put those pictures you cannot really represent this complex system with these two parameters thank you very much let's open it up to a question for the three panelists first question well done I'm doing a real world field test my question to you is how well can you assimilate even that in-situ test you have a pretty good idea of the parameters how well can you assimilate and depending on that answer what is the prognosis of assimilating other real world field tests where we don't even have half the data that you have so there are many aspects of the observations some part of the data are easier to assimilate we're very proud that we predicted the fracture propagation direction based on the thermal gradient which it causes the stress gradient we predict how where the fracture is going to propagate so the shape was not exactly right but the direction was right the fracture initiation so some of the unexpected results come from the sudden pressure changes and also a challenge also it's a very important merit of the experiment because now you not only need to match the pressure data you have temperature data you have tracer data so it's actually very hard to come up with a model that can cover all these aspects but that's the value of this data set where we have many teams and modelers working very hard trying to do better model and understand more it's an ongoing process so how about the prognosis for assimilating when we don't even have that data it's hard to but now after seeing this much data in the real foresight at least now we know what to look for it's more inspirational so in the forge if we see interesting pressure behavior then we know there are many possibilities then we knew 2-3 years ago Sulfide on your equipment did anybody analyze that for gold? No Now the questions Lynn So I'm struggling to figure out how to ask this question so I like Fickery's picture of the complexity of the transport in shale because it has multiple length scales and it has multiple transport scales so he talked about adsorption and diffusion and there's some local permeability through the fractures and then transport through the bigger fractures to a well so I'm trying to think about how that would translate to a geothermal system and maybe one in which we're trying to induce a set of fractures that would cover thermal energy effectively allow transport through the system now it seems to me that not only do we want permeability you need to get fluids in and out but you also need to have some surface area because if you only have one big honking fracture then it doesn't take very long to get that one the heat unloaded from that and so there's a combined problem where how do we generate the kind of fracture system that you want for that kind of rock and then of course how do we work out what the transport properties are that seems to me it's fundamentally multi-scale in lots of ways and quite challenging to put together all the pieces that go with that and of course I'm not an expert in geothermal systems so I've probably left out two-thirds of the important stuff so I'd like the folks who are expert on oil and gas recovery fractures to say what I forgot there and then maybe one of the geothermal types could respond to the overall question okay so basically if you forget the absorption etcetera I think rest of my slide applies very well to geothermal system because geothermal system especially if you have a carbonate rocks they have all those complexities so you have a one poor scale as you mentioned second the rocks micro fractures small fractures etcetera big fractures and the faults and faults are big scales so basically we have many scales I think the problem in geothermal as well how we are going to reduce this scale at a manageable level probably we need more work some of similar experiments we can reduce this scale otherwise we cannot simulate and of course we can without simulating we can generate electricity for sure but to understand and also to enhance I'll tell you something interesting you know when you are producing who cares about fractures if you have good productivity nobody cares really nobody cares it took 40 years for Aramco and earlier Chevron etcetera to realize that cover field which is the largest field is fracture reservoir 40 years they produce 5 no 50 60 billion barrel of oil from that reservoir we didn't know that it was fractured what happened when they start putting horizontal wells faults and fractures coming constantly then of course the 80s 90s 3D seismic was not good 93 95 they did really very good 3D seismic all of a sudden it becomes fractures Gawar is today is a fracture reservoir so what I am saying that so when you are basically what happens when you are start injecting anything then fractures hits you so today like some of the big fracture reservoirs many others their water cut is 95% many reservoirs in fracture reservoirs in Oman is 95 to 100% they are still producing basically what happens because when you are happy you are good PI you don't invest but then you inject water water comes to well one week later then the problem starts but now your cash flow is not allowed to get out of data add a different twist on that because the way you ask the question implied perhaps an interest in also how do you make the fractures because he is talking a lot about fractures that are there and they are conductive but in your enhanced geothermal system you might be actually trying to create the fracture system and I think a couple of things I would point out just by way of analogy the the DuVernay case that I showed you they fractured one well with the usual viscousified fracturing fluid which over time was the obvious practice for creating massive fracture extent the other they did was slick water and historically in the Barnett the discovery that finally led to all the success first with natural gas and then with oil was slick water which was counterintuitive because that's not the way you carry propant but it turned out you don't need much thinking propant as Nathan Mann said because the rock permeability is so low at that point that the crack is mostly what you want and it turned out that the slick water was working and now the other point is the material because I think this source rock material is different from sandstone a lot of people make the mistake of saying hydraulic fracturing enhances permeability normally no but maybe in shale yes but it's not really the matrix permeability you're enhancing you're creating permeability with a secondary fracture network so I don't know the materials for the geothermal systems probably aren't shale like but to the extent that they might be or that you might want to reassess what kind of rock you're looking for these material properties really matter and how you go about creating fractures could really matter through a crack or a pour when you have you know the relative permeability is not right so it's easier to put oil through that it's common, it's just a basic physical principle now what I have observed is in a Brazilian field in the sandstone reservoir I've seen lots of wells build in the same reservoir and the interesting thing about it I had a chance to examine how the wells were performing in different spots in the sandstone and where you had high density fractures where your density fractures were high and your sandstone permeability was high you got a much better well so that was a turn on thing so all of these so your porosity enhances the rock and the fracture density enhances the rock for flow so all of these things you have to keep in mind and but the main thing I think your question is relative permeability is putting water in and getting it back sometimes it's a little bit difficult in smaller pore spaces can I make a comment from the geothermal side of things and talking about conventional resources see if I can articulate this in terms of understanding really those favorable settings I think we're still sort of in the infancy I mean we're decades behind the oil and gas industry I mean just think of structural and stratigraphic traps that basic science was done decades ago and we really just begun in the last decade to understand geothermal systems in that way and so you can think of for conventional systems of course and we need that source to be connected to our reservoir so source rock, reservoir rock it isn't too different but you need a deep connection of fluids so how do you do that with natural permeability and there's lots of combinations that you can do that and you add it up for like the box we did in Nevada it's 10-12% of the area but not every one of those favorable settings is going to be a system because we still don't understand the hydrology behind it and how recharge occurs in these systems so what I'm getting at is there's a huge potential I think in conventional resources not that I'm not minimizing EGS in any way because that's critical as well but we're at the infancy in really understanding the connections between what's to say deeper levels in the crust not the deep crust but deeper levels and upper levels and if we can get to the point that the oil and gas industry has gotten in understanding connections between source rocks and reservoir rocks and permeability and so on and porosity with geothermal I think the sky's the limit just for conventional resources so there's a lot of work to be done and once we establish the reservoir the detailed modeling and so on is absolutely critical as well but we still need that sort of broad scale modeling So this question to the geothermal folks here and what I want to bring up is an important point that Christine made and it has to do with the horizontal layering in the rock that you're fracturing so there's two points that I think she made and I think it's important to recognize this is that there's dissipation of the frack energy along these horizontal planes often and of course with prop and getting them around corners is just not the easiest thing to do unless you can make them walk or something like but the question is how much horizontal layering do you have is that an issue for you because there is if there's horizontal layering the thought of dissipating the energy of the frack along these horizontal beds is critical now we have mechanisms to avoid that we've got diverters and things like that but you'll run into the same problem we have today in the oil business which is don't want to spend the money to do that right so that's the question how much horizontal layering do you have in these reservoirs geothermal reservoirs that you're fracking into say one more thing one of the things that I observed in these basins is the fact that a lot of the things that are deep into basins that we use in the unconventional world are turbidites they're just turbidites so you've got this uniform turbidite these flows and these fans or whatever is coming in pretty fine grain material like in the Delaware Basin in the Texas there so the stuff is very uniform you can take a log over here in this side of the basin or another one over on this side and you can match the continuity across there correlation is just fine all the way through that turbidite system because the pulses that have come out there sequentially and covered a large area out of New Mexico now here's the trick though I think that you see is you want to use the fundamental geology of depositional environments now we hadn't talked a lot about that specifically I'd use that term but the first layer I put down is a depositional environment layer so I see if I can see these fans that point bars and things out there coming out and match the sinuosity of these things into the basin there's a lot of work but if you have that layer down pat and you look at the net and in turbidites we know what we have lots of bomber sequence and layers and so forth in this thing so what I find out is that you want to examine where your fundamental basic permeability and porosity is in the conventional sense in this network and then where you find the best conventional permeability and porosity and you can impose the best fracture density on that man mate that's a good shot for you so you've got to stack these things in your mind maybe that's something that the machine learning can help with in geothermal though we don't typically have these nice sedimentary sequences you know they're commonly in volcanic terrain and you have ignimbrates or other volcanic units or the host rock for the actual reservoir and not the source rock per se but so or you have actual fault zones that are probably the primary reservoir for geothermal fluids so I mean there's a couple of examples where geothermal fluids like in the Salton Trophy Imperial Valley Southern California and it looks like there you have such huge flow rates because the rate of extension is so high and a pull apart in the San Andreas four centimeters per year you get enough flow rate in the outflow in the sedimentary sequences to provide a consistent reservoir that you can put a number of straws down into and sustain a system for decades but that's relatively rare otherwise it's pretty extratigraphy in the Salton Trophy in some other places too but that's relatively unusual usually not but those rocks because the temperatures are so high I would tend to maybe call them as metamorphose rather than diagenesis very high temperatures, very high salinity things have changed a lot from the sedimentary body but even if you get metamorphic layers you can have horizontal layering metamorphic shifts and the like so you can still have this issue of dissipation of energy along those horizontal points perhaps I'll draw this panel session to a close so please join me in thanking the three panelists and we'll reconvene in 15 minutes so we have on the schedule to take a very brief break and then we're going to have our wrap up discussion Bridget has agreed to lead that so let's come back promptly at 2.45 if we could which is 11 minutes so we can have time to have a good conversation there's probably a continuation of this but with a few more key themes I think one thing I'd like to touch on which we didn't get a chance to argue about but I will mention that came before the component of fracture resolution so I think that the observations that Colab was looking at the potential closure of that fracture that's the problem so thinking about I think we put the phone about that some of the shear fractures were closing through time and I didn't know as I didn't get a backup that I would be able to talk about it but we need that fracture to be extended to the term it could easily be that there's a fracture so I think yeah so I think there's more to yeah so I think there's more to come there so what I plan for this is just to summarize a few of the things that I'd like to talk about for the next few days that probably opened up a number of observations opportunities that we had with the observations that were located around here so which were the questions that we had for the Colab sessions which used to be fine in the few days before it used to be fine so if you're there an hour and a half before yeah you'll be good if you just get an Uber from there mine is 5.30 oh you're playing it 5.30 but you'll be fine the Reagan is 20 minutes at 3.30 you'll be out the door yeah so what I'm from college what what what what what what what what what what what what what what what what what what what what what what what what