 So, welcome to everybody to this special session entitled Faults, Rivers and Topography in Memory of Patience Kaui. So I'm Hugh Sinclair and I'll give a brief introduction before passing on to my co-conveners. As academics, our metrics of achievement are usually presented as publications, citations and awards, but natural abilities to enthuse, challenge and inspire colleagues and students can get overlooked by this need to quantify and rank. Patience Kaui was exceptional on both counts. In this session, we will hear about her outstanding scientific achievements, but we will also gain insight into how her personality was able to drive the science and inspire younger and particularly female colleagues. Patience started as an undergraduate at Durham University and then went on to master's and PhD at Le Monde Docherty under the supervision of Chris Schultz who will be opening our presentations. After a short postdoc in Nice in France, she received NERC and Royal Society Fellowships at Edinburgh University from 1993 and progressed to Professor of Geodynamics in 2011. She then moved to Bergen as Professor of Earth System Dynamics. Her CV is available online where you can see her many achievements, including being the recipient of the Koch Medal from the Geological Society of London in 2016. An award that I'm sure she appreciated as a mark of recognition for her contributions, which would also have challenged her natural reticence to be seen as part of any establishment, a rebellious streak always questioned convention. Besides being a good friend and colleague, my abiding memory of Patience is a tenacious questioner of meetings. Once she got a teeth into a problem she would not let go until fully satisfied she pushed the problem and the poor speaker usually to their limits. We hope that this symposium will demonstrate the important role Patience has had in building our understanding of the interaction between fault growth and surface processes. And I'll now pass on to my colleagues to say a few words before starting the session. So hello everyone. Yeah, thanks for joining this symposium. So my name is Anna Leunghurtz. I was doing a PhD with Patience in Bergen. And yeah, she has just been an amazing supervisor for me. And I think, yeah, in particular, what I really liked about her is that she was a very good listener. So she was, yeah, I hear this from many young scientists that the thoughts of the students, she considered as equally important as her own thoughts. So in that sense, she's really a role model for young scientists. And I hope she will be a role model for all the professors in our field as well. So I'll give the word to Mikael, I guess. Hi, and good afternoon. Good morning. My name is Mikael Rathal, and I'm at the University of Edinburgh. I came in 2005 in Edinburgh to work as a postdoc on a project with Patience. That was the Apennines project that we're going to hear a lot about today. And this were challenging years as a young PhD graduate. They were really stimulating years, and they were also some of the most fun years I had, like probably some of my best research years in my career. And Patience was just an incredible mentor, and we will probably hear more about that. And yeah, I wouldn't be here without her. So passing on to Laura. Yeah, thanks. Hi, everyone. I'm Laura Gregory, and I'm at the University of Leeds. I was lucky to work with Patience for my postdoc when I was working with Gerald Roberts and on active faults in the Apennines. And, you know, one of the things that I just enjoyed the most about collaborating with Patience was the way that she attacked a problem and kind of never let go of any aspect of it and always sort of delved as deep as we possibly could into kind of questioning and trying to figure out a problem and her capacity to listen. I really have to echo that, but, you know, she treated everyone as equal and really listened to what you had to say, considered your ideas, even if they were completely bonkers. And that was really greatly appreciated. Yeah, so just a bit of housekeeping on the way the session is going to run. We have six speakers each speaker will has about 15 minutes, which includes approximately three minutes for question time. During this time please feel free to contribute with a question or comment or even just a memory related to that talk in those three minutes using the Q&A function in Zoom. We'll do our best to bring those questions up in the three minutes and then we'll also kind of save them to the end. And I've just copied some texts that explains that again, if you if you missed anything. But yeah, so we'll then have 20 minutes at the end of the session for general discussion, and we particularly like to think about how Patience's research has influenced all of us and will have long lasting influence in the future. So I think forward to future research and maybe come up with some questions or comments on that and of course you're welcome to share any memories that you have of Patience that you'd like us to bring up during that discussion session. And we've been told that we can keep going a little bit over if we need to. And we also have a separate Zoom link if we want to have a drink afterwards with with everyone that we'll share later. Okay, so I'm going to now move on to our first talk. So I'm very delighted to introduce our first speaker. Our first invited speaker, Professor Chris Schultz from Columbia University. Chris was fortunate to mentor Patience through her PhD at Lamont. And the result of this collaboration was insightful research on faulting mechanics and scaling that continues to have a huge and lasting impact on our field today. And so I'll hand it over to Chris to hear more about that. And he's going to talk to us about Patience Cowey and the inception of modern fault mechanics, a recollection. So Chris, if you want to share your screen and kick us off. Hi, I'm Chris Schultz. I've been asked to share about patients as early work with me on fault mechanics. Patience, Patience PhD thesis resulted in three papers published in 1992. They're pretty well known. And I'd like to give some historical context for her work so you'll understand sort of the breadth of her contribution to this area. Patience approached me first about being your thesis advisor in fall of 89. At that time, my own research was an impasse. I just finished the first edition of my book. And in so writing I discovered the major gaps in our field. The most glaring of this was that we knew nothing about the mechanics of folding. We had no idea how falls formed or how they propagated and grew. All we had at that time were Anderson's theory, which of course didn't say anything about those questions. And most brittle fracture theory up to that time was restricted to tensile fracture problems. If you tried to apply fracture mechanics to shear fractures, you immediately ran into a dead end. Because all that fracture mechanics told you was that if you try to drive a shear crack, that you just get these arrays of tension cracks along the edges as shown. And if you do experiments, you get the same result. So, but what we really want to know is how the shear crack grows in its own plane, which is of course, what must happen to explain faults. So we were absolutely stuck. So I suggested to patients that she should try to see if she could apply the Dugdale model to faulty. Now, I have to go back and explain some background on that. And fracture mechanics was developed in the 50s. And at that time, there were a lot of issues about it. Which were later sort of brushed under the rug. The most important one was a stress singularity was implicit in elastic crack theory, and it's non physical, because a real material would yield by some inelastic process in the high stress region, the crack tip. And also this issue dug down later bear black development are called cohesion zone models to sort of look into this issue. So this up here on the upper left shows a sketch of the dug down model and what it's trying to do. It assumes that there's this yield stress applied around the crack tip, such that that yield stress has to be exceeded in order for the crack to open. And you're going to open along some breakdown zone like s. So as a real result down here, the bottom you see that gets rid of the stress concentration, which is now spread out over the breakdown zone. Now the way patients knew this in terms of faulty was goodness. It was shown up in the upper right to assume that beyond the crack tip, the fall tip, the rock was all broken fractured to the high stress there. And that was observed experimentally already. And that went further in these cracks started to coalesce into forming a fault, sort of a proto fault, which eventually formed a full fault. And so on the bottom here you see the stress. As patients observed, there's an applied stress A. And at the tip of the crack that reaches the yield stress sigma zero, which gradually breaks down until it reaches the final friction stress material fall. And a particular figure from her paper is the most important figure in her paper, and I'm going to spend a few minutes on it. Because it has to do with this way that things must scale. So in the upper right you show the idea as the fault grows. What happens to this breakdown zone s in this first assumption, it's assumed to be constant scale independent. This breakdown zone underneath gives the fracture energy. So that means, over the left here, that fracture energy G is a constant. And in that case, D max the maximum slip on the fall scales as length of fault from square root. In that case, the driving stress to drive the fault will decrease with fault length. As shown here. So it decreases the fault length. That means the fault gets weaker as it gets longer. And that implicitly contains an instability. And of course, that's the group with instability. So this is the group of model. So the idea that there's this instability doesn't seem very pro very likely in terms of our observation of faulty. So she looked at this lower lower right version in which s scales linearly with fault length. In that case, G scales with L and D max also scales linear with L. So in that case, the fault grows self similarly, and it has constant strength. So it's a stable fault model faults are stable at all times, and they just grow so similarly. And that's much more physical. But at that time, there was some data that indicated that maybe D max increases as L and exponent greater than one. And so what's the consequence of that. In that case, s must grow with faster than linearly with with crack length. In that case, the strength of the fault increases with length. And eventually, the applied stress will reach the yield stress, in which case the fault can't grow any longer. The rock just crushes it doesn't make a fault. So that means there's an upper limit of fault growth. In this case, and G increases with L. So that's kind of that upper limit of phone growth sounds a little unphysical, compared to what you think about faults. So we better look carefully at that data that indicates this nonlinear relationship between D and L that showed on this slide, where we show results of D max going as L to the end power over on the right. And then goes as one and a half. And I want to look at this data turns out the data. It's pretty questionable in these plots. First, this Menard plot, which is all the long, all the longest of these data points is shows the offset on the rise at fracture zones plotted against the total length of factor that's a pre transform fault hypothesis paper. And of course to transform fault hypothesis falsified that whole idea. So those data have to be thrown out. Then this big group of data from Macmillan turns out to be an unpublished bachelor's thesis from Carlton University. And if you bother to actually track down that thesis, you find almost all these data points are from plate boundaries, like the San Andreas fault and plate boundaries don't have well defined ends. That is, they don't have ends with a displacement goes to zero. So those data also have to be thrown out. So over here in this plot, we have another pernicious problem. Notice these, these is diamond shapes. Those are the original Canadian thrust faults of Elliott, and Elliott himself in his paper showed that those scale linearly and we can see that you sort of go off at an angle to this block. There was trying to use data that has a linear scaling to sort of prove that it's quadratics really is making a procrastion argument. So what the main thing is that patients recognize that D max on L must be a critical strain for faulty. So that must be a stress drop over sheer modules, both of which are dependent on mythology and on pressure and therefore depth. So that means it's the whole concept of plotting data from faults and all different mythologies and all different depths on one spot and expecting to get a uniform scaling of that is the whole ideas fallacious. That's what she could do with this data this crummy data that time where she cleaned it up a little bit. And after cleaning it up a little bit. She saw, okay, it's sort of linear down here from small faults and sort of linear up here, but then there's a street strong climate here. What's that. Well, that's at the range of one kilometer to 10 kilometers. And that's the range in which faults, of course, going from that length range are digging deeper and deeper into the crust. And so they're, so they're encountering stronger and stronger rocks, and they're higher and higher pressure. So it actually be L ratio has to climb there. And most of the spits were in that region. So all she could do with this data she said, she could fit these, these small faults and soft rock and shallow depths sort of to a linear scaling of rock, scaling along sort of these long faults created the 10 kilometers and hard rock to another scaling rock. But she said, this is really crummy data. So what you really need to do is if I have to find a place where there's faults in all the same rock type over large scale range of to actually verify whatever the scale is. We immediately went out and look for a place like that. And we found the volcanic table lands in California, where we have these normal faults in in welded tough. And here you can see the data scales beautifully linearly scaling between D and L over three orders of magnitude. So we thought the whole issue was settled at that point 1993. But no. People kept repeating like a mantra for the next 30 years. Oh D goes as L to the end power, where it is between one and two. The muggles just didn't get patients is argued. So here's the most recent results over here. So in this case, you see, oh, these authors sort of make a big fuss about this variations down here. The final result is linear scaling over five orders of magnitude, where strikes the faults and granted. So maybe that'll finish that issue. Who knows but anyway the last leg of patients is thesis was to compare that linear scale law for faulty with the non linear scaling law for earthquakes, which shows as shown here D. So it shows as linearly with L for earthquakes. And if you compare those two results turned out they're now compatible with one another to show that you could take the linear scale and all faults from the accumulation of earthquakes. So what so prior to patients were only know about false was that there were these discontinuities, which produce these displacements in in rock formations. And according to read read first show the relationship between earthquakes and faults, but that took a long time for that to be. What was that how a folk be treated as a sure crack with friction going by the collapse collapse of cracks and damage on a tip. Both were now seen as growing in three dimensions, following well defined scaling laws as slip accumulated by earthquakes or creepest case maybe this new viewpoint has guided the work. Mechanics of faulty for the past 30 years, much progress has been made, which greatly expanded and elaborated the basic template proposed by Cali. So the inception is modern era fault mechanics stems from the foundational work of patients Cali. Thank you. Yes, that was really interesting to kind of hear your perspective on on that work. We don't have any questions yet they'll give people a minute to put some in otherwise we can move on to our next talk. Maybe you could comment on, you know, did. Did you ever speak to patients about kind of adding things in like creep and and all of the other weird and wonderful ways that that folks can slip besides an earthquakes into your thoughts on on displacement length and mind to I think that was obvious there wasn't a particular point in this but however you get displacement. It's got a, it's got a scale of length because if you increase it is, if you're at the equilibrium D to L ratio. And you increase the displacement, then you increase the stress concentration chip chip must like. Whatever man, you do that. Yeah, I guess you don't have the kind of, I always think of an earthquake as having a damage zone around it because it is such a, you know, strong force that's exerted by the slip by the rupture itself. You don't have a little bit damage zone, probably getting damage zone that goes and hits the damage bigger damage zone and it goes plunk and increases it. But if you, but if you slip default by creep, it's also going to increase the stress concentration chip and increase the damage zone there that way too. Do any of our panelists have a question. If not, we can move on. And one thing. Yeah, lately there have been a lot of people who've been talking about faults that grow by lengthening without increasing the slip. And that of course can occur by coalescence chip and coalescence will, of course, increase the length, but then the faults out of equilibrium, the length is too long for the displacement. The displacement can keep increasing on the fault without increasing the length until it gets back to the equilibrium. The L ratio, and then it starts increasing again. So there really isn't any conflict between those two ideas. I agree. Great. I think given that we are pressed for time. Let's move on to to our next speaker. But if you have any questions or comments and kind of your thoughts on patients is very early work with Chris. Please add them to the chat and we can bring them up later. So now I'm very happy to introduce our next invited speaker, Professor Zoe shipped in from the University of Strathclyde. So it was patients is first PhD student during her time at the University of Edinburgh. So it was patients is enthusiasm for applying range of techniques like fieldwork and modeling to understanding problems was really imparted on on Zoe, who has since conducted ground baking research into fluid flow and fault systems 3D fault structure and fracture network interactions and geological models. So I'll hand it over to you Zoe to talk to us about a retrospective work of patients cowie, the interactions of faults and space and time and their influences on subsurface fluid flow surface processes and earthquakes. Thanks very much. Can everybody hear me okay. Yeah. I've put a slightly different title up here musings in memory of patients carry out I'm not giving a science talk here some of our later speakers will be telling you about the really cutting edge science this is a bit more rambling. And as Joanna said I was, I was very lucky enough to be patients is first PhD student. I first met her when I was working for Rob nights rock defamation research group, and went to help out on an undergraduate field trip and there was this real buzz amongst the lecturers on the field trip about this person who was coming to visit patients is coming patients is coming. And when I finally met her. She had the most brilliant day out in the field it was it was pouring with rain. And I got to show her and the field area that we're looking at which was like this really classic section of the moin thrust. And the two of us for the last women standing in the field everyone else went home to escape the rain. And we just had a great time and I was really excited when she asked if I would come and do my PhD with her. And so I moved to Edinburgh University and straight away we went out to do field work here which is along Bob Krantz's for mapping out field mapping area from his PhD along the chimney rock fault array. My brief really was to look for that tip zone that Chris talked about based on the dogdale barren black model, you know, can we see evidence in the rock record of that yield zone at the fault tip. I was in the field by John familiar who is one of patients is good friends from her time at the month and working in the field with patients was whoops was incredibly good. The bottom photograph here and thanks very much to Nancy Dawes for sending this to me there's this John patients and I in a fit of giggles and in the Canyonlands Grubbin district. Patience was a was was very tough in the field she had us up at the crack of door making coffee. Very loudly if you slept in she the coffee like making got louder and louder until you woke up. And then during the day in the field we were taking surveys of using a total station to measure the offset. It was a good fun, immensely hard work and incredibly challenging. And then every evening we would, we would sum up with discussions over a fire. And then I'm quite sad to say, I didn't get the chance to cycle into work into my office because of COVID I have a napkin that we wrote ideas on and raised haven in Green River Utah I still have it registered into one of my PhD notebooks but I didn't have a chance to go and get it to scan in for this talk. So, the, the, what I want to do in the next part this talk was to really reflect on the kind of philosophy that patients had towards her work. So, and present some of the technical aspects but really to think about the approach that she had, which was different from, or maybe not different but really is what she passed on to me. And this is an adaptation of the figure that Chris has already shown us that fault length displacement scaling. And so the first thing that I learned from patients was that stress really matters, you know, you need to be thinking about mechanics. You can't make observations of paleo stress in the field you are making observations of the finite geometries that are left behind by the actions of those paleo stresses, but thinking purely on the basis of the geometry and kinematics of faults doesn't get you answers that can delaying the observations that you make and we've already seen this graph from Chris but I want to point out these blue dots here was from a paper by Ralph Slisher and co-authors in 96. And it was actually the product of an undergraduate dissertation where some an undergraduate went into some really tiny faults that really blew the idea that the scaling relationship was at a was an equals 1.5 or two out of the water once and for all by going right down to these these low levels. In the carrying Schultz 92 paper. I love this quote the unambiguous resolution of this question will require some significant improvements in the existing database. So, this is an example of where an undergraduate student brought those improvements to the community. The second aspect of the patients was really that that space and time matter. And, you know, when you're teaching first year geology or as I do nowadays geology to engineers, they're often looking at very simplified block diagrams, where they're little cubes and you have a normal fault cutting through a cube or a thrust fault cutting through a cube it doesn't have any ends. It's basically a very small slice through a really three dimensional structure. And in fact what patients got to thinking about was that they're not just three dimensional structures they're four dimensional structures and they grow in time link and coalesce. And a lot of this work she did during her postdoc time in France she she moved to France after her PhD to work with Sornette and Vanessa and produced a series of numerical simulations models based on numerical simulations that showed faults growing and interlinking and coalescing through time. And that allows you to start thinking about how faults influence things like landscape through time and that really seeded the later work that patients did on landscape evolution on the influence of faults on on river patterns and sediment deposition patterns and so on, which are important for understanding modern day landscapes and in in converse using modern day landscapes to understand fault growth and earthquake hazards, but also were really important for the oil industry and thinking about how faults, how growth faults may have influenced the deeper centres of where or reservoirs would be in the subsurface and enable people to make predictions of what might be happening at length scales that were much below those are image of along on 3D seismic data. So, a quick segue into into my own work with patients we, we went to the site in Utah, and on one of our first days we were at this site this is the, the blueberry fault, which is to the south of the chimney rock fault array and there's clear bond in the University of Aberdeen for scale there. And quite close to the to where I took this photograph we found a pole with a tin can nailed to it. And inside the tin can was a very raggedy old piece of paper which we pulled out and it turned out that this was a claim steak for a mineral rights claim. And later, many years later I got chatting to one of the locals who said that that his father was looking for silver. I don't think he ever found any silver I don't know there is any silver in this district but the name of the claim everyone every claim was given a name and this one was called patients claim, which was kind of spooky but there we go. In my PhD, we, I can say we were looking at surveying the stratigraphy to try and look for these, these tip zones that Chris was talking about, and we didn't find them. What we found was the displacement gradient to the fault tip. So here is an image from my paper with patients and 98 the distance from the fault tip the displacement grew pretty much linearly. And again, that that didn't phase patients. What we then turned around was to think well that's the observations so the models must be wrong, and all models are always wrong so how can we explain this. And it must be through the repeated accumulation of slip, each one of which might behave in that that way with with a process. The additive nature of a false slip smooths that out. And the symmetrical nature of the loading and each of these rupture events moves out into a roughly symmetrical triangular shape displacement profile. And then the final sort of philosophical point that I learned from patients was exactly how much, you know, data matter, and the models are always have to be underpinned by detailed data and for geologists that almost always means fieldwork, not always. But this is an example of some really beautiful, careful data that were collected in Mexico by Estelle Mortimer, looking at how the sedimentary architecture of a series of deltaic sediments could give you information about the temporal evolution of this fault, and the evolution in space how it was growing in time. In a sense what Joanna and Chris were talking about in the last question session about how these things are not uniform in time that they accelerate and decelerate and accelerate and decelerate. So, stress matters, space and time matters, and fieldwork matters. The thing that mattered immensely to patients, the two things were her family and top right here we've got her dancing at Kaylee at the University with her husband, Leo. And her mentoring of younger career academics and at the bottom right it's the same Kaylee, an image of her with her then research group at the time that Leo very kindly sent me recently. She was a music mentor, she was a great listener, she was extremely challenging, she didn't take fools gladly, you know, if you said something that was daft and non thought through it she would prod you on it until you unpickled yourself from your own mistakes. But she was also quite happy to admit when she was wrong. I'd say one of the big things that she passed down to me as well as these sort of philosophical approaches and the rigor was the sheer infectious joy that she had in the field. And when I heard the news that patients had died it was absolutely gutted and went and had a look through my photo collection and on the top left here is one of my favorite photos. Patience is grinning with absolute joy at the sheer beauty of the vault that she's looking at there. And that that joy is something that I really enjoyed being in the presence of and I really enjoyed taking one passing on down to my own mentees in my career. So, thank you very much. Thank you, Zoe. Do we, anyone from the audience you're welcome to submit a question through the Q&A. And in the meantime, maybe I'll ask a really quick question and then we'll move on. But so do you think that, you know, we can take information from those really small scales and extrapolate up to the big scale. What are your thoughts on that? Yeah, when we're at the large scale and we have, you know, really complex earthquakes happening, do you think it averages out enough that that we can compare all different fault systems and fracture systems and gain an understanding from that? I think you can think about what happens at different breaks and scale. I mean the scale that we can make observations on is inherently limited by our own scale, right? We can scramble around field areas a certain distance over a given day, depending on what the field area is like. We can look up a cliff to a certain distance. And more recently, you know, within things like drone photography and satellite photography, that's changed that length scale that we can make observations on immensely and in really exciting ways. So you have to think about the scale of processes and Chris has shown that, you know, natural break in the data sets. And so I think you have to take, you're using the small scales to try and gain a bit of an understanding of the mechanics and to kind of ingest, digest and reflect on the understanding of the mechanics and growing cracks in analog models, you know, breaking food, doing it in the kitchen. You can do a lot fraction mechanics in the kitchen. But cheese is not the same thing as rock. So you're always having to question, question, question, what assumptions are we making at this scale or in this material that are the same as or different from the the situation I'm trying to make predictions in, whether that's earthquakes moving through the cross store or whatever. So I think it's that questioning your assumptions is really key. Yeah, that was another thing that I think patients always did very well as to question our assumptions constantly and never let go of reminding us of what our assumptions were. So yeah, we don't have any other questions right now but if anyone else would like to comment, you can if not then I'll hand it over to Michael to introduce our next speaker. Yes, questioning or assumption is something that, yeah, it's always there. And, and I'm sure we're going to hear more about that. And it's my pleasure to introduce our next speaker, Joanna for Walker, who is an assistant professor at UCM. Yeah, she's been in, she's been working on the upper nine faults with with patients and with Gerald Roberts, and she's going to talk about some work with Francesco yes he as well, looking about fault evolution skating relationship and has ours. Thank you very much. And that was lovely to hear sort of how patients his work has developed and I can certainly second to what Zoe was saying about always being questioned by patients and being aggressively sort of made to be challenged as much as I could. I'm today I'm going to talk about fault evolution that scaling relationships and hazard, particularly looking at the role of fault geometry and how we can use that to understand the evolution of faults. So this work, I started actually back in probably 2009 and the first paper I woke on it worked on this with patients was a co-author so she has very much influenced all of this work over the last 12 years or so. So I really want to think about how faults grow and this actually builds on the last point that Chris was talking about. I'm going to consider the case where we have two fault segments which are offset across strike from each other, and how they connect and grow and look at the different stages of the fault evolution. So here we look and we see we have some fault tips growing below the surface gradually growing and increasing their throw. And eventually we're going to have some linkage at depth there might be soft linkage initially, and then the faults will actually connect via on this long connections they might then form a relay ramp breach faults and eventually a fault bend. And in order to connect if they're offset from each other, we're going to have to have a connection that's a bleak, so the strike will form a bend, and we'll also have a higher dip. And I'm going to give away my conclusions at the start and then I'm going to give you a lot of examples to show you how we've used our evidence and our data along the way and as Zoe said, patients are certainly someone who says we must have evidence we must have data and we've been collecting more and more as we go along. So here we have an idea of let's just take these two faults segments which initially have no linkage, and we have those classic fault throw displacement profiles along the fault we get an increase in throw at the center and decrease at the end. But as we connect these faults somehow these parts need to catch up. So what we propose is that we actually get a higher flow rate in these sections which are connecting so that eventually the total throw profile will indeed form what we expect which is that total displacement for a profile. But how does it know to do this how do we actually have a system that allows this to occur because the fault isn't just told all you need to connect it needs to actually have a mechanism for doing this. So let's think of these different stages long fault growth and as I say I will show you some examples along the way so we have our sort of fault and soft linkage initially we have our two separate displacement throw throw profiles. We then have perhaps a breach fault scenario where we actually have an oblique breach fault for me across it. At this point, our total throw profile will still have this what we call a double displacement maxima we sort of have the hump with a smaller for in the middle. But what we're proposing is that we need this higher throw rate so that eventually we can get to a stage where we have this total displacement profile and eventually the breach fault will develop into a fault and what happens across this. So let's explain this through the idea of looking at strain across the fault. So if we imagine our fault bend coming across the and the strain across this spend. If we presume that we have a strain rate that's constant across this part of the fault now as a fault as a whole we have maximum strain rates in the middle they gradually decrease towards the end but on a smaller scale. So if it is either constant or at least that it's only changing smoothly. Well, if we introduce a fault bend, we actually need a higher throw to maintain a same strain rate. And the reason for this is essentially just taking the equation of principle strain rate and I'm not going to go through all the details. But if we simplify this to the case of just considering one fault segment here, rather than all the fault segments that we might have in the area. If we do then presume we have this constant strain rate across the fault which we do presume if the fault is linked because at that point it's just deforming in the response to external forces. Then we can see at the bottom here that our throw is proportional to one and this is this over the sign of the difference in the angle between your strike and your slip vector. For a pure dip slip fault, we would expect that value to be one sign of 90 is one. And as we get more oblique that number goes down which means the flow rate we expect to go up. So here's the theory behind it and now I can show you lots of nice examples from Iceland and Italy where we've been studying this. I'm going to start off in Iceland on the thing veiler rift services in the southwest of Iceland here and in particular I'm going to look at a particular fault here, which is a north or the thing the Latin lake apologies for any of my talk. In this part of Iceland, we have lava flows which have flowed over and they'll cover any of the pre existing topography, they form a some nice flat surface. And here in particular the latest flow we have has been dated from sort of between 10,200 and 8200 years ago, using carbon 14 dating and tepra chronology. So once we have this lava flow coming down any fault activity that breaks the surface since then will offset this and hence we can get our displacement sorry as this figure has gone wrong. But essentially here we can see the offset from both sides of the fault and this displacement we can then divide by the time period to get our through rate. And if we just focus on this little section here where we have soft linkage, we can get a maximum in our throw rate here showing this soft linkage at depth showing we have this highest grain rate here, so that the fault is already connecting through soft linkage. We're now going to sort of jump to the southern Epinines in Italy, and we're going to go to the second scenario the breach fault scenario. So here we have we're in central Italy, and to give you an idea lack was here and we're going to focus on a very small fault here, and the parisano fault, which you can see the mountain fault here it's got lovely fault Scott, and it has a breach fault at its center. So this is a breach fault not just because of the angle and the geometry but because the total offset since the upper Cretaceous and eosine is actually much lower in the central section. But actually when we then look at our rates since 15,000 years, we get a higher throw rate across the breach fault. Now we can do this in the central Epinines, a bit like in the same way we did in Iceland, but instead of looking at the offset of lava froze, we're now looking at the offset of the last glacial maximum surface from around 15,000 years ago. So here we have the last glacial maximum surface and the same idea that this has been offset in this time and we can form SCAR profiles across the fault and use this to estimate our throw rate over the last 15,000 years. And this again is showing that we have this higher rate across this breach fault, even though the total profile profile which is shown here we have that double displacement maxima like I showed earlier in the cartoon. So we have this breach fault but a higher throw rate across the center in order to keep that strain rate looking as if we would expect if it was just behaving as a normal fault or I say normal I shouldn't use the word normal when talking about if it was just behaving as a fault, you know with a maximum in the center and to decrease into zero towards its tips. So, again, we're showing here we've gone from a soft linkage we're now looking at a breach fault and we're now going to look at the case of fault bends, which is where we're saying this hard linkage has occurred and here we're looking at another fault in the last 15,000 years. So, same mechanism for measuring offsets. Here we've actually got some terrestrial LiDAR. And this is, we've got lots and lots and lots of throw profiles across the fault. And where we see a strike where we see this change in strike where we have a fault bend, we again see this higher throw. But despite having the higher throw, we actually have a very smooth strain rate profile. So, again, this is showing us that fault geometry is playing this role, giving us clues about how these faults are connecting. Starting to have as one fault, then being hard linked, and then this is continuing throughout its life cycle. If we go back to Iceland again, back to Finvala, we have examples here. Here we're actually looking at the offset of subglacial erosional surfaces. And again, if we look at these faults and look at the offset across them, we are seeing maximums across the fault bends. And likewise, we've seen the same thing in Hengil volcanic province. And here we've got a slightly different age. Here we're looking at about 7,000 to 13,000 years. Because we've had dating with exposure dates from cosmogenic helium isotopes in olivine and this gives us an idea of these offsets here in the time periods over them. And once again, we're seeing this maximum across the bend. So lots of examples from different stages of fault linkage and from different tectonic settings. Both are normal fault settings. We've got extension. One is sort of intraplate and the other is in an interplate setting across a mid-Atlantic ridge. The role of rivers and geomorphology has been really important in understanding all these examples. I've sort of said, oh, this is an area of soft linkage. This has been linked at different stages. And we use lots of clues to help us really understand where we are in these cycles and to understand how do we know that these really are two paleotips that have connected. I mentioned in the Central Appanise example where we had the total offsets and then we could see there was a minimum. So that gave us a clue. In Iceland, we've used examples where we've actually looked at the relationship between the monocline. So here we can see a monocline up against the fault and with the monocline to determine tips. And we've also used river offsets. So here's an example where we've actually had a river migrate north in response to an increase in throw rate across the fault. And we can see that this is actually quite recent because we have a waterfall here. But you notice there's very little incision, very little incision as a result of the waterfall. So we know this is a relatively sort of new example. So how does this occur? How does this faster throw rate occur? These settings are both examples where most of the deformation or all the deformation is through earthquakes. And we're actually showing that we can get this relationship where we're seeing that higher throw across a bend in individual earthquakes. So we actually think that in individual earthquakes, we are seeing this greater throw across the bends. And that's how this long term profile is growing over time. And here we have examples from Mexico, Greece, the USA and Italy. And what we actually did here was we use the measurements across the cosize Mcthorough measurements. These ones we actually did ourselves. These were from other authors. And these lines here are green and red lines here are where we've actually modeled what we expected the throw to be given the model. So the red lines are showing you sort of out of fault, if you like, or where we have the continuous strike. And then these green bars are what we would calculate the throw to be across the bend in our model and then the actual lines here and the actual dots are all the actual data. So we're getting a really good fit using our sort of predictive results compared to the real data. So we think we've got a really good strong relationship here. And this is really important for things like scatter relationships. And when we look at looking at the D max versus surface rupture length, I've got two examples here from wells and copper Smith and Managetti. And we see that we get quite a lot of scatter when we're trying to relate D max to length. And we think we can explain this through fault bends. So if we take these examples here where we look at the examples are examples here in blue. The orange line shows where these orange dot show where they would plot on the wells and copper Smith plots. And you notice that all our examples we had that maximum displacement across a fault bend. And all of those plots above the line and we think that's because they're in a location of fault bends. So they have a higher flow. And that can actually explain, you know, perhaps not all the scatter that there's other influence as well but quite a lot of the scatter that we're seeing these scaling relationships. Another important example is, sorry, my computer's going crazy is from looking at fault bends in terms of seismic hazard. And I just want to make a point here we're looking at that parisano fault again where, if we look at all our data, or if we use simplified data so a simple profile across a fault. Then we don't get the same results we get very different rates for recurrence intervals. And likewise we get quite different results for our seismic shaking individual sites so here I've the blue line here is showing the case with all our data. So we have ground motion prediction equations, spectrum acceleration at individual sites. And if we start using simplified results we actually come out of the uncertainty. So we need to understand fault growth and the relationship of the fault geometry and throw in order to really understand, you know how we use faults in seismic hazard assessment. And so with that in mind, we've produced the fault to show our database which uses a lot of detailed data, including all the fault geometry on the faults and detailed slippery data. And this sort of fault to show initiative is actually something that patients was initially sort of one of the people involved who really wants to get it going so I'm really pleased that we are sort of keeping all this going and hope this sort of subject to this talk has continued on some of the subjects that she would like to see to be continued on. And here we're showing you the importance of using the detail in our models. So I'm saying thank you, it's a double thank you it's a thank you to everyone who's listening but it's also a thank you to patients, because that has sort of basically inspired the last 10 years of my research. And I also leave you with some information about an upcoming special issue in a journal. Thank you very much Joanna. That was really, really nice. So I think do we have any questions. I just want to encourage people to ask question we've got the Q&A so at the bottom of your screen you shouldn't have Q&A button so you click on that and you can put questions here at the moment we don't have many things like they are some questions coming from the chair. So maybe I'm going to start with that. Zoe is asking a similar flavored question to the one Laura asked me, the work you present is all on surface traces of active faults and needs to be, and needs to be to get that spatial resolution. How well do you think this translates to the linkage of faults and depth in the crust. What assumptions if any that you have to test in order to say that your surface interpretation are valid and dead. So, yeah, it's a it's a really important question. And so I think there's sort of two different parts of this. One is that we have looked at the total strain rates we calculate across the whole of the Appanines and compared them to the topography and actually compared that to a quartz flow law. This is work that she did with patients and Chris and Gerald Roberts. And so, this actually can show us that we have a mechanism for showing what we think the driving mechanism is it's all to do sheer zones of depth, and how these then relate to the topography of the surface. So I think the fact that we do have an explainable mechanism helps us think that you know these results are not just randomly they are important. We also have compared the results we get at the surface to things like the GPS and historical seismicity. We get good matches when we look at the total strain rate across the Appanines. But it's different spatial resolution there's a different geography of hazard when we look at shorter time periods. So there is some sensibility checks we've done. In terms of evidence for how this sort of propagates to depth. We have looked at some of the micro seismicity following the 2016 lack of earthquake. And that's sort of, you know, within resolution there's agreement. It's not precise it's not going to give us the same details as we get to the surface but again it's giving us a broad agreement. We also looked at examples where we've located the focal mechanism at depth for particular earthquakes so for example the Colferito earthquake, and we've compared the depth and location of that earthquake and compared it to the fault of the surface and showed that the propagation of the fault of the surface does meet the detailed examples of where we expect to see that fault. So there are you know checks we can do in terms of you know really knowing does all the detail propagate and exactly how does it, it's still very difficult to do. If anyone could do that and get the real detail at depth then I'd be very, very happy and would love to talk to them about it. But I think that is one of our really big challenges overall is trying to relate what we see at the surface to depth. But I think in geology it's one of those things we have so little information we have such a short time span that we see the detail for and we have such a small fraction of the fault that we get to see these details for we have to use what we can. Great. Thank you very much so we have some question. So, so I'm going to start. I'm going to start with the most question. We've got another question. Have you ever observed any noticeable change in the displacement and fault length relationship with mythology. Sorry, I didn't quite catch the second half of that. Oh, sorry. How the that relationship, which I think Chris mentioned at the beginning as well as how the displacement the relationship between the displacement the fault length, fault length changes with mythology. And I haven't actually looked at that myself. And so I'm not sure we've, I've got a good answer for that to be honest. That's fine. Is it is anywhere. Is anyone aware of data that show that doing you said. I can't remember off the top of my head what they found, but I would make point that as soon as you start getting faults that are large and by large I mean large with respect to this logical unit boundaries. They're not sampling a lithology anymore. And when you're looking at faults to scale at Joanna's looking at those faults are sampling multiple mechanical units. So, those faults are in a lithology they're in lots of lithologies that are spatially variable but also potentially temporarily variable if you're talking about long time periods then they may be changing their mechanical properties due to this or conversely changing their mechanical properties due to fracturing, which can have big effects on Young's modulus of a material so any studies that have been done. I'm sorry, I don't know any off the top of my head to point you towards but I would say you need to take anything other than very very small faults with a bit of a pinch as old. I think that. I think you can only see that for the reason that Zoe just mentioned this for small faults, tiny faults. So for example if you look at the faults in the British coal measures which are really soft rocks. They have much smaller DL than those faults, those tiny faults of switching at all with your court site. So you really see it from small faults. And just as you respect the stronger for stronger faults have much higher DL ratio. But of course that's all smeared out for long falses. Thank you very much. There is just one question from Ted. I'm working in earthquake direction and how it influences the atmosphere, how we could contact if it's possible so if some people have some ideas about about that you can answer to Ted directly in the chat. In the meantime I think we need to move on so I'm going to introduce our next speaker. Thank you very much Joanna. And our next speaker is Alex Whitaker, who is a senior lecturer at Imperial College London. And we did this PhD with patients at the time I was doing my postdoc in the upper night and is going to talk to us about 14 landscapes and I hope you got it. Thanks for it. Thanks very much, Miguel. I'll just share my screen. There we go. Right, hopefully everyone can can see this now. So it's, it's, you know, a pleasure in a way to be able to give this this talk and remember patients so I really got started working with patients in 2002. You know, I'd come come up from from Cambridge fresh faced and wanting to work with patients really on thinking not just about folding but how that affected the landscape evolution. And then I'm going to give some examples of where I think the ideas and work really influence some of our ideas on folding and seven from. I'm really going to start with the background to some of this stuff. And that really is all about how fluvial landscapes record active folding. So, you know, when, when people started to think about all the landscapes panel landscapes can rivers record folding a lot of that initial work was all about numerical modeling and I would, I would kind of point point you to one of my own papers by kind of metal and Greg Tucker, which really attempted to implement detached limited modeling bedrock river modeling stream power modeling of rivers responding to a base level and what that that initial work kind of showed was that in response to a to a model base level change. Detachment limited bedrock rivers, basically developed a kind of nick point that migrated upstream in response to this, this change in base levels these were numerical examples, but you know, in 2002, and this was published nearly 20 years ago, you know, it was possible to use a figure like this where you would kind of show a long profile elevation against downstream distance with a kind of kind of nick point. And the question was, well, actually, wouldn't it be great if there was some good field examples of some good field case studies. And this is really whether I think the brilliance of patients came in because she had been talking to Greg. Greg Tucker, and she was basically saying, well, I've been working out in the Italian Appanines, and we think that we know what the faults have done and we've got good examples of faults that have changed their rates over time. So I basically joined this, this project in, in, you know, 2002 and I've been working on rivers with Niels Hovis and I kind of found myself working on a, on a super exciting project with an amazing mental. You know, as, as John others, I've talked about the Appanines, the normal fault array in the Appanines has got quite a complex history, but essentially it initiated, at least in this area around three million years ago. And some of these faults have increased their rates through time as the fault array has interacted. That means that some of these faults have increased that rate. And that basically meant they were good examples of where we could then go and say, okay, how is the landscape responded to that. And, you know, patients was always really keen on making sure that we have like some really well constrained field examples to test numerical models. Because, you know, she often said to me, your models can make predictions when we need to see what, see what, you know, the real world actually shows. So I was really, you know, I dug out some photos and I was really fortunate to be able to do lots of field trips to Italy and elsewhere with, with patients. You know, I mean, it was an amazing experience because, you know, patients was, as Zoe was saying, it was like, I remember working breakfast, close to seven, you know, you had to be on, on the wall. But it was, it was a, it was a fantastic experience just in terms of the, you know, kind of a field training and, you know, when we started patients, you know, it was funny because she was like, oh, Alex, you, you already know more about fluid genealogy than I do. And of course, this wasn't really true at all. Because, you know, I rapidly realized within like a month of starting a patient's authority read half of the papers that I would be thinking of. And, you know, it was, I, I, it really taught me how the level I needed to work at to really try and tackle, tackle a problem. So we spent a lot of time surveying rivers and really thinking about some of the assumptions that people had made in, in, for example, detachment, limited erosion models, we spent a long time, for example, to thinking about channel width. So we had, we collected a lot of data on, on that kind of stuff. And one of the kind of key outcomes from, from this work was to, to show, I guess, for the first time that we could actually detect the signature of transient response to fault slip rate increases in the, in the landscape. And one of the, the poster child for this was the, the Rio Torta River cutting across the Fiamino fault weather. The River Long Profile has got the kind of poster child, Nick Pointe's steeper reach upstream of fault that increased its rate by perhaps a factor of three, around three quarters of a million years ago. And, you know, one of the things that really then got us thinking this inspired so much additional work was actually the fact that while, you know, upstream of this area wasn't just the landscape was responding. And this was the point where all of the erosion was concentrated. This was where all of the material was coming from. So we, you know, we could partition the landscaping this football between the area that had high incision. And then upstream of this point in the area that had low incision. And although, you know, you can look at this as a geomorphic problem. This was also a problem for thinking about selling supply from from catchments as well. So, you know, the sort of the classic things that I and others took, took away from this was that it was in general the case that all of the rivers that were crossing faults that we thought it increased their rate in the last million years. All of the rivers have shown classic signatures of transient landscape response to tectonics, whereas a lot of the rivers crossing these blue faults that have not really changed their rates or the inactive faults, but not, you know, they really were largely concave up rivers they weren't really showing a transient response that we could particularly detect. So, you know, broadly these rivers in Italy were consistent with a Bedrock Detachment Limited model of erosion. But one of the key things from this was that the landscape response to tectonics could be bracketed between one and three million years. And I think that was a really important revelation because it meant that, you know, landscapes, the tectonic tape recorder, if you like, did last over over million year time scales. And that of course then comes back into one of the things that we talked about a lot, which was sediment supply and I remember being in like some gorge with patients and looking at a whole bunch of landsliding. And you know, we really got thinking about where where is that material coming from what characteristics. And one of the first things that the patients did, and this is still work that inspires me is a is a paper that I and others including Mikhail who was involved in this project and Greg took on a whole bunch of us did was really to what patients really was trying to couple her fault growth and interaction model with a surface process model, and we use cascade for this but where we basically looked at how the landscape evolved when, when these faults grew and then subsequently interacted. And this was kind of important because one, these individual catchments they grew and then shrunk as the drainage areas competed with each other. The example catchment for was bigger, you know, this time scale in the model, but by this time step catchment five and captured the lot, but also of course that the center released from the catchment varied and increased obviously as the people started to interact and link together so this really demonstrated that, you know, faulting. You know, we're having a big role on sedimentation so I guess there are three parts that I would like to kind of stress as things that motivated me and my, my career my research students but it also motivated a lot of work from other students and they're still working on that, you know, 10 years on one was the realization that that sediment supply signal itself, then can modify fluvial incision dynamics. So one thing that, you know, we initially originally did was to try and compare, for example rivers crossing faults in Italy with rivers crossing faults. In this case in Greece, near the Gulf of Levia, which have got similar catchment sizes and faults with similar rates and slip histories. And you know what was immediately apparent in this and I should thank them for some of these slides, is that, you know, the rivers in Greece, they'll, you know, although they're crossing the faults, you know, they've got, you know, relatively long profiles that don't show a huge nick point compared to, to the rivers that we had just been talking about in, in Italy. And of course one of the key differences is that the upper parts of these catchments are dominated by the sands and gravel units that are supplying abundant sediment to the channel even though the channel down sister is cutting through bedrock. Whereas in comparison, the Italian examples is relatively little sediment being supplied. And we were able to show that they made a substantial difference to how effectively the rivers managed to cut across faults. One of the things that we saw, for example, this is a fluvial efficiency, if you like, on the Y axis, which is related to the sediment supply, and this is the sediment supply related to the transport capacity. And we're able to, not only in Italy, show example that we do have a tools effect in other words, if there's not much sediment putting more in increases the erosion efficiency. But we will also to argue in the Greek examples that if you've got lots of sediment in there, and you put in more, you basically cover up the bed and dampen the erosion efficiency. And that some sort of parabolic model best explained our observations and these are relatively consistent actually with the theoretical predictions of sclarin D trick and others. And this kind of view of how rivers behave, you know, lots of people have taken this forward in the last, in the last 10 years. The second thing I want to focus on is that, you know, in the last 10 years or so I've really ended up thinking a lot about normal faulting and particularly grain size. And that's one of the things that I think, you know, really came from some of the work with with patients. So, you know, we, you know, we, we looked a lot at, you know, a bunch of these rivers crossing, you know, for example, Fortuno fault. You know, they've got classic signatures of landscape transients upstream of the faults. But, you know, they are, you know, the rivers are actually dominated by, you know, upstream in these incisors and dominated by landslides. So we actually look at the grain size on the bed down system towards the fault where these black dots are the 84 percentile D 50 and these white dots, the medium grain size. You can see that the grain, the grain size really increases in that zone upstream of the fault. So it's not just that you're eroding more sediment but it also has different characteristics. And I think that matters a lot. And, you know, those kinds of questions were also taken forward with patients. And so I work with Stal Mortimer, for example, you know, this is the recent compilation of mine I'm still kind of putting together where these are channel grain size against uplift rate for catchments that are broadly the same size, whereas the ones in red are all of the catchments that are known to be undergoing a transient response to a tectonic flativation. And even where fault slip rates are similar, it's actually the catchments that are responding to faulting are the ones that are actually chucking out all of these coarse grain sizes. So that, you know, that matters, that matters a lot. And the final part of this is actually really to do with basin implications. And that's often, you know, this is, you know, my initial conversations with patients have kind of inspired where I've gone with this and also where some of my research students and others have gone. So, you know, if you change the amount of sediment and its characteristics coming out of the catchments, you also change what you expect to see in stratigraphy. So we started this problem just by doing some relatively simple numerical models. This is some stuff that John Armitage, Rob Duller and I were involved in where we built simple models of catchments and alluvial fans, that are bounded by, for example, a normal fault, but we could scale our grain size release to some of the apennine examples and include appropriate catchment tectonics. And then we basically balance the stratigraphy in the hanging wall and then find the grain size deposits. And what I'm going to show you is actually not the catchment, but just the base and stratigraphy that we see. In this example, we basically have run the model to steady state and at this point here, we have increased the fault slip rate by a factor of five and we've also, it also increases the grain size that is being released and that's scaled to the Italian examples. And I actually produced some really thought provoking insights. Sure, when you increase the fault slip rate, you create a combination immediately, but it takes time for the landscape to respond to that. So you don't have enough sediment to fill up the space, but the sediment that you do have is coarser. And that leads to the counterintuitive result that close to the fault if we look through this borehole here at A, you get an apparent coarsening in stratigraphy, whereas in B, on the same timeline, further away from the fault, you get an apparent finding in stratigraphy because you've run out of sediment by that stage. So some of these things really kind of challenge some conventional notions of what you expect to find in deep, you know, stratigraphy and that actually tied in a lot with some of our field examples from the, from the base. And I think that the key takeaway here is that, you know, transient landscape responses are producing transient stratigraphy over a million year timescales. And I guess I will just finish with saying, well, we kind of know that that's also true in the Italian Appanines where we first stopped, you know, started working. I remember talking with patients about this because the, you know, we were like, oh, well, we're eroding all of this core sediment. But you know, the Fuccino basin was a lake until we drained it, or until it was drained 100 years ago or so. So the story is actually that core sediment just doesn't make it into the center of the lake. It basically produces these thin fringing alluvial fans so you just extract all of that. So, you know, fits very much with the central Italian App examples. So I guess my three messages are that, you know, a lot of patients work in terms of landscape and faulting was, you know, really seminal but in terms of some of my particular focus with supply, we've really demonstrated with some of this field work that it controls the original dynamics of rivers. The grain size really matters. And particularly that, that, you know, then feeds into understanding basis stratigraphy, and that if we believe in transient landscape response to, for example, active faulting, we also have to believe in transient stratigraphic responses to active faulting. I think that's the really important message. So I'm going to stop there. Thank you very much Alex. And yeah, that was a long talk. So we're just going to move on to the next speaker directly. I'm afraid. If you have any questions, you can put them in the chat and Alex can respond directly Alex or the speaker so feel free to contribute to the questions. You can also share your experience with patients as well. This is what this is about. And I'm going to let Annaline take over for interesting on next speaker. Thank you very much. Yeah, thanks Alex. So our next speaker is Philippe Steyer. So he's an assistant professor at Ren University. He did his PhD in Montpellier and Paris, and he came to Bergen to work with us for five months. But then he met patients and so they work together on their own on a nature paper, nature paper from 2013 on viscous roots or faults. And I think he's an example of, yeah, who has a career that's really inspired by patients work of the interaction between surface processes and tectonics. So, Philippe, the floor is yours. Yes. Thank you very much, Annaline. So I will share my screen. Is it okay? Yeah. Okay, so I don't see it. You can't see it. Here it is. Okay, good. Thanks a lot, Annaline, for inviting me to give this talk in memory of patients. So I will talk about trying to go to have a better understanding of the impact of erosion and topography on deformation. And the subtitle of this talk could be why we need to look for debates and provocative research rights. This is a bit of history, how I met patients. So I was in 2011 a young postdoc working with Ritzke-Ritzmann on the numerical modeling of extensional settings. And we rapidly hooked up with the patients when she arrived in Bergen a few months after me. We had very intense and very insightful discussions about the links between surface processes, topography and solider deformation in extensional settings. So at this time I was, for instance, running this kind of coupled model between surface processes and tectonic deformation. And we quite often ended with this kind of hook-shaped fault interaction zone that were favored by surface processes. Patient was not directly supervising my work, but she played a role of a mentor and she also was a kind of role model for me. And I think I will always remember this advice she gave me that it is good news when a paper is provocative and fosters debate. And at this time I think it was quite a shock for me because in the scientific community we generally don't like when our works are commented or debated. But actually for her it was a very good sign that at least our work was useful and was making the community making progress on the field on which we are working. So because of that I will therefore present to you today not only one, but several provocative papers. The first paper I will present is a seminar paper by patients with also Chris and Joana about the viscous fruit of active seismogenic faults revealed by geological slip rate variation. So patients was looking at the apennines and she was in particular looking at the links between the elevation of topography and Olo Sen extensional strain rate that were measured along faults caps that you can see here on this image. So what patients basically did was looking at compilation of this data along the apennine so you can see on this map the shaded topography and you can also see this green and orange line showing the direction and intensity of strain rate, that were measured based on these faults caps and patients she simply actually use some kind of sliding windows along the apennine and for each of these sliding windows she measured the maximum elevation and the accumulated strain rate during the Olo Sen. And she ended up with a very simple relationship between strain rate and elevation where she observed a correlation and that the strain rate was proportional to elevation at a power around three to three dot three. So for those of you who are familiar with flow low viscous flow low and dislocation creep this exponent of three or three dot three must ring a bell on the bell of dislocation creep. And so to make this a link between this empirical relationship and viscous flow low patients needed to show that elevation could be related to differential stresses. And she noticed that when you are at field in a crust. You need to have actually a proportionality between the differential stress and the maximum stress that is in the vertical direction, and that is proportional to elevation in a context where you are where forces are driven by buoyancy. And so she ended ended up with a proportionality between elevation and the differential stresses. And if you combine this empirical observation between strain rate and elevation with this proportionality between elevation and differential stress, you obtain a proportionality between strain rate and differential stress at a power of about three. And this law directly suggests viscous flow low by dislocation creep. So in this approach, she developed a simple model of crustal deformation where if you have a driving stress you will induce both elastic and viscous flow deformation in the lower crust and frictional sleep on the in the upper crust along the fault. Because you need to remember that the strain rate data were obtained along frictional faults. So in a way, if you have some surface uplift of your topography, you will increase the elevation you will increase the differential stress that must be relaxed because the crust is at field. And this will be relaxed by permanent deformation with frictional sleep in the upper crust that lead to this strain rate that we observe here and viscous deformation below. And these two deformation will play in parallel and will be in a way proportional to each other. And so the main message of this very nice and very similar paper by Komi in 2013 is that we with this kind of relationship we can constrain with geological data, this exponent of the dislocation creep flow low, which is I think quite unique. And the results that patients obtain also suggest that viscous deformation was actually localized just below the frictional faults in some kind of fault roots that were deforming in the viscous way. We had, of course, a very good coupling between full sleep and this fault root deformation. And the model of patients assume that at least during the whole cell deformation in the upper nines were governed by buoyancy forces imposed by topography. And I think at least for me as a young postdoc at this time it was this paper by patients. And the very important conclusion is that topography and topographic changes could govern deformation, which I also feel is a provocative idea. And later on I try to extend this idea and to apply it to other type of settings to frictional faults in the mountain ranges where erosion rates could also maybe influence the stress loading rate of stress faults. So I will now present you another paper that I developed based on this first paper by patients, looking at how surface erosion can lead to fault loading and seismicity. So the concept of the idea behind this paper is that if you remove mass at the surface of the earth, you will change the stress distribution at depth and you will change the stress on potential stress faults. And what you will do is that you will include the fault and increase the normal stress, and you will also increase the shear stress. And if you combine these two increase, you will also change the coolant stress on the fault that is equal to the effective friction times the increment of normal stress, plus the increment in shear stress. So in the following, I will compute stresses along this kind of faults using the Boussinesq point load model which applies to an elastic space model. And this model allows to sum up individual point loads to mimic some kind of surface distribution of erosion. So I will compute stresses on the faults by projecting the resulting stress tensor on the fault plane. So two very simple results, even very obvious results, is that the coolant stress increment is proportional to the amount of erosion you have in surface. Coolant stress increment also decreases with a distance at a power 2. So in other words, further away you are from the place where you make some erosion, further the lower the stress increment will be. And I apply this approach to erosion rates that were measured in Taiwan by Datsuneta in 2003 based on suspended flexes that were measured before cheat sheets. And these erosion rates are representative of some kind of inter-cystic phase. So you see here this map of erosion rate and I computed how this erosion pattern impacted the stress loading of the stress faults on the western foothill of Taiwan. So what we obtained is that the impact of erosion on the stress loading of these faults was quite moderate and we obtained some value of about a few times 10 to the power minus 3 bar per year on this fault. In order to integrate this value of stress rate over the length of a seismic cycle, we could obtain up to maybe 0.2 megapascal over a seismic cycle, which could in a modest way maybe influence seismicity. So this concept, this idea of the impact of erosion on faults was later applied to other type of settings, and in particular to settings where erosion is not from natural sources, but due to anthropic activities. In this paper by Qan et al. in 2019, they looked at the Tianjiang earthquake that occurred in Eastern Sichuan Basin in China, that was, that has a particularity of occurring at a very shallow depths of one kilometer. This earthquake of magnitude about four was occurred just below an automotive testing site that was built just two years before. And to build this automotive testing site, they removed about 10 meter of bedrock over a surface area of about one square kilometer. In these studies they computed the change in the coolant stress due to this anthropic surface erosion, and they ended up with an increment of about 0.1 megapascal, which is quite significant. And they concluded in these studies that possibly this anthropic erosion could have triggered this Tianjiang earthquake. In other words, I think very interesting cases where we could also involve this link between anthropic erosion and seismicity is the recent earthquakes that occurred in sources in France, the Thai earthquakes that occurred in 2019 at a depth once again of about one to two meter. And this earthquake occurred just below a very large cement quarry here that has a size of about one square kilometer. And with a total erosion of about 18 meters over a few decades. And in this study by the novellis et al. they computed that this erosion due to the quarry could have increased the stress by about 0.2 megapascal, which could have also resulted in advancing the earthquake by by about 20 So in conclusion, I think patients was one of the first to demonstrate using feed data, the role of topography and topographic changes on current deformation. In addition of a better understanding of this course deformation in fault routes. I think the seminal work has led to several daughter studies. And these studies these daughter studies have demonstrated that natural or orthopedic erosion can increase stresses and potentially trigger earthquakes on trust foot. And I think this, these results have strong implication for earthquake hazards, in particular in regions of low tectonic activity, where the probability of earthquake occurrence is therefore quite likely underestimating. I would like to thank very much patients, and also to thank all the co-authors of the paper I have presented. Thanks a lot. Thanks so much for this great talk. So, if there are any questions. Yeah, welcome you to put them in the chat or no in the question. And so, I was just wondering so, because you were showing the paper of that quarry. So probably that there must be many examples then are you trying to compile. Yeah, all those sites where you have major quarries and. Yeah, I'm not trying directly to do that but I think actually people have rediscovered a big this link between quarries activity and and earthquakes, I mean triggering of earthquakes quite recently, especially by this paper in 2019 by and more recently by this Lotte earthquake in source and friends. So I think there will be in the following years some additional paper on that. I'm not directly working on trying to compile data about this link between quarries and potentially triggered earthquakes. But probably some other will will try to do that, I guess. Yeah, I see this one. Okay, maybe we can only maybe you want to ask it yourself, or. Yeah, we can do that. Yeah. So, so my question that's just fascinating and I mean considering how much stuff humans are moving at the surface of the earth at the moment, I mean, we're moving more dirt than natural processes now and like is it not surprising that there are no more of these earthquakes happening because I imagine you can trigger earthquakes but you can also potentially like suppress them if you have sediment deposition or things like that. What is your thinking here. Yeah, you're completely correct. You can also have the reverse impact if you do for instance surface erosion at the top of normal foot, you don't expect to have exactly that. But I think also it's not. You don't expect to have all always is directly between, for instance, removal of rock mass at the surface and earthquake triggering I think these cases are quite spectacular but in the kind of probabilistic way will not maybe be so easy each time to have this kind of cosal links. So yeah, so it's a good question. Thank you. Thanks. Okay, thanks again, Philip. We have to move on to our next speaker. So it's my pleasure to introduce Sophia. So she is from Greece, but she's now in Bergen. So she did a PhD at the University of Tesla Niki. And she came to Bergen in 2014. So she has worked with patients for many years. And she's mainly on Greece. So first on spatial space and now mainly on current. And she's using a combined approach of numerical modeling. And now she's also involved in the Iodp. Stratigraphic analysis. So that's the one of the recent course that was drilled in current. So Sophia, the floor is yours. Thank you very much and Elaine. Thank you very much for the invitation. So it's a great honor for me to be here today in this symposium dedicated to the outstanding work of patients. I was extremely lucky to have patients as my postdoc supervisor at the University of Bergen where we worked very close over the last six or seven years. But all these years I've learned so many things from her, not only related to science, but I really got lifeless from her and I feel very grateful for that. So as it has already been said by the previous speakers, patients used a lot of numerical models in their work and she was always trying to validate this somehow with field data. So here you're looking at photo from the current risk in briefs where a wealth of field observations exist both onshore and offshore, but makes this place as an ideal natural example for model calibration. Of course, this area attracted patients attention a long time ago, and she was actually one of the persons involved in the initial Iodp proposal for the current goals that took place a few years ago. So today I will show you the modeling work I did with patients aiming to understand key controls on sediments production within the current risk. And as I'm also a member of the Iodp expedition, I will show you some of the recent results that we got from this expedition regarding serious transgraph development. So the current is in central Greece and it's one of the most actively extending areas in the world with extensive rates reaching up 15 millimeters at the western tip. It is a still controlled trees that it's now at sea level and it is connected to the open sea to the west through the real straight where water depths are less than 60 meters. Therefore, during sea level falls this connection was lost and the Gulf was turning into an isolated lake. It is generally considered by many geologists and geomorphologists as an ideal laboratory for understanding structural and surface processes interactions. Because this is a young group with a relative single economic history and without overpins, we think began further south and has migrated north into the Gulf at around two million years ago. And subsidence is now controlled mainly from this north deepened normal falls along the southern active drift market. Another reason that makes this area as an ideal place to study structural and surface processes interactions is the unique combination of high strain and sedimentation rates with the fact that this is a self contained small risk with a cloud drainage system as you can see here from a Google Earth image, but actually allows us to characterize an economic and sedimentary risk system. Finally, as I mentioned, there is an excellent field data cover up with the series strategically the last two million years well preserved offshore and other phases are uplisted and preserved on shore, mainly along the northern and southern active risk margin. So, I will move on to the surface process modeling. Our aim was to evaluate the key control from sediment flags into the Gulf over the last 130,000 years. We use this time window because this is simply where we had the best constraint for model calibration. The main input that we use is by badlands, and here you're looking to present the topography of the research that serves as the initial surface into the model. This model simply wrote sediment on shore based on a stream power law and deposit them into the marine domain using a marine diffusion. The main inputs that we use was the first of all the cumulative vertical displacement map that you are looking to the slide that was imposed uniformly over the 130,000 years. This map red shows subsidence wind shows uplift. This map was generated using a linear elastic dislocation model based on the geometry of the map offshore fault that you see here with red lines. And it was validated based on the quaternary uplift rates of this fault. These are the slip rates that we use to validate this map and notes that the highest up to states are concentrated mailing along the central part of the reef. For this time period, please. We also use the ability maps that you're looking here based on the map on shortly followed these we define three main ability groups with low intermediate and higher abilities that you can see with different colors. We also use a global sea level curve that we corrected based on the lake current flow stands to include fluctuations of the base level, the flat areas here are corresponded to the time interval when calling towards late. And finally, the precipitation map that's based on the present day climatic data. And as you can see, there is a great variability in precipitation rates between the western and eastern part of the risk, and with high precipitation rates along the western part using the existing relief of where the need is for the prospect in this area. So, these are our results. Here you're looking the cut fence that are draining into the gold for that color coded based on the total volume that has been eroded over the 130,000 years. You are also looking the model offshore sediment thickness with this green is the blue is colors to validate our model. We basically compare the volume that we got from the model with the volume that we got from the interpretation of offshore seismic data, based on the work of mix on about 2016. As you can see from the table we've got a very good match between model output and sizing data that provided a lot of confidence in our results. So now let's look at a little bit where this sediment comes from. If we consider the risk margins that the hardware margin and the north margin that is not deciding. Our model implies that actually both margins contribute almost half of the total offshore sediment volume. The cut fence at the western end, due to the large area contributes a lot of sediment and we've got a really good match between sediment volume and cut cement area as you can see here from the plot. However, a key result of our model concerns cuts and other its erosion rates and time other its erosion rates. And this to plot you're looking at the variation of erosion rates for the solder and the northern region. For the solder margin we built a very good correlation between erosion rates and the combination of maximum release and up is great, but we didn't see for the northern margin. However, erosion rates by a fairly similar between both risk margins. And this is something that we didn't really expect. The total erosion rates along the northern margin are similar to that of the tonic daily active solder margin, even though this is the one that is subsiding. And in other words, we would expect that the north margin would contribute significantly less than 50% to the total offshore sediment volume. But this is not the case. So, to explore this a little bit more with it fairly simple experiment. We ran our model without imposing any electronic forcing, and we compare that to the original model model run. Here you're looking at the result of this comparison, and you're looking at the catchment draining into the gold again, but now our color coded based on the difference in erosion rates between these two model runs. So the erosion rate that we got when running the model normally in the phone mind the erosion rates that we got when run the model with the phone switch off. So what we see here is that the the catchment that are lying along the southern risk margin. So a negative difference in erosion rates. You see that most of them are called with this fluid colors. Which implies that we are having higher erosion rates in the run when the tonic is actually switched off exceptions are these areas close to the fault that you see with red and both of some areas upstream but are related to high purchasing relief. So in contrast, the catchment that are lying along the north margin show a positive difference in erosion rates you see that all of them are colored with red, which implies that we're having higher erosion rates in the run when the tonic is switched on. So the explanation behind this counter intuitive I would say results is actually the flexural building of the week. The catchment that are lying along the southern market margin are being backfielded at the footholds of active folks. As you can see here from the sketch that leads to decreasing erosion rates at the upper parts of this catchment at the headwaters and also leads to drain of diversals that we know that have a cure to especially in that area here. In contrast, building forward the wrist along the north margin leads to the increase increasing erosion rates because this flexural building increases risk directed channel slopes. The flexural building of course plays a key role in controlling serious sediments like instant actually results in assist in sediment source area such as the north margin would become progressively more important compared to the southern active wrist margin. This result has implications for sediment starvation in this that actually might be too big to increase in erosion rates and degrees in erosion rates and not necessarily imply needs and increase of the current crops that has been merged. So what I've shown you so far where the modeling results of the code that were actually average over this time window found with the 30,000 years. However, patient vision for that area was to use the high resolution record for model calibration. And this record become possible through the IDP expedition 381 that provide us with metronome records of 1900 meters of section code. These are the three drilling sites of this tradition as I will talk only about the central site here. So, for this site, I will go to many such a graphic units, a lower one that corresponds to another rich face and another one that corresponds to the most recent rich face, and that will consist of this alternating marine and isolated environment. Now it's also low sedimentation rate for the marine environment and highest imitation rates for isolated lake environment. However, a key result that came out from high resolution photograph analysis of this unit one is that most of the beds that consist this unit one are gravity flow that's actually forward more than 60% of the total sedimentation. The rest is covered by the background sediment. The gravity flow show a great variety in terms of bed sicknesses as you can see here from the plot below. And what we actually trying to do now is that based on this record, trying to investigate what controls observed thickness variability of this gravity flow. And since there is actually an external signal at the tonics or climatic preserved in the strata. These questions sort of brings us back to 2003 and the work that patients did that we say Hugh Sinclair about the scaling of their diet. I would just want to close this patient's idea about that area that was by combining this also drilling data with surface process modeling. We might have the opportunity to understand the relative roles of the tonics and climates on sediment production and stratigraphy development within with patients in general. And I would like to close with this photo from patients. It's one of my favorite sense is from our field within this is not smiling but I can tell you had an amazing time back then so thank you very much. Thanks so much for your great talk. Thank you. So, I don't see any questions yet. Maybe. Yeah, so I will just, because there's also another question to believe that still left. But I was wondering so you about your tipping the balance work. So, the starvation of sediment that's just because the effects on the southern margin is bigger is larger than on the northern margin right so that the back tilting effect is. Yeah, you have back building along the southern margin but actually reduce erosion rate. But the forward building along the north margin actually increases erosion rates and yes, we can say that relatively speaking, makes the north margin, like, quite important, compared to that the southern margin. But we actually showed. Yeah, so this is the tipping to the balance effect. Yeah. Okay. I'm not sure if this question to Sophia here it says. Oh, this feeling with question I think maybe Philip you want to ask it yourself or. No, I think that's, I think that's really responding to the question that I had. Which I could be the answer because we've got the way we're now going on at the same time sorry and we've got questions here. Yeah. Okay, good. I had a question for first to fear I this work is so this work is so cool. I want one thing that I remember talking with with patients about was the extent in current that whether the long term sediment supply is also governed by drainage network evolution so the loss of for example catchment areas on on for example in the in the south for example so do you want to could you comment a bit on on that because that's always something not just in current but in other places that you know I've struggled with I guess. Yeah, I mean for the last well for the time interval that we were investigating. We know that drainage reversal and well drainage device migration was minimal. So, actually that was another reason that we use this time we didn't really want to mess with so much complexity complexity in this model but yeah. But this building effect actually leads because it reduces the slope of the reverse upstream. It leads to drainage vessels and I think this is also something Mikael has shown on one of well also huge in one of your papers. So, yeah, how this back building can lead us to drainage reversal over time, and therefore do that reducing sediments like. Thanks so much. Yeah, we have to move on. So thanks to all the speakers. So we have yeah just six minutes left. So the plan. Oh yeah so maybe Mikael do you surely want to ask the question to Philip or. Okay yeah I can do that quickly because I tried to copy and paste the thing and you can't copy and paste from the zoom chat so it's a bit of a disaster so. So there was a Julian Julian Rossi had a question for Philip, which was congrats congratulations on your nice presentation in the relationship between the differential stress and elevation. What about if we consider also the poor pressure and saturated media. So, do you want to just respond to that. Maybe Chris also can answer if you prefer because it's probably more especially that I am. But I think it will not change a lot of things we will just need to to account for hydrostatic pressure. Instead of like little static pressure in the model. But I think it will not change the, the fact that the crust is at yield and when you have surface to please you will lead to deformation. The same type of relationship between strain rate and elevation, it will just just change maybe a bit the relationship between strain rate and differential stress, I guess. Thank you. Okay, so five minutes left I was just so I will just show some pictures of patients. Just to finish off with. So, yeah, I wanted to ask Chris was on final words about. Yeah, you wanted to say something about patients character and to transformation as a student. Do you want to do that Chris now or. Yeah, okay. You know, it's a, you know, when you're teaching PhD students you're always looking for a point where they change from being a student and just accepting knowledge until they realize they're creating new knowledge and sometimes you see a big change at that time I never saw a greater change in that with patients. Patients that actually been at Lamont for about two years before she started working with me and she'd actually been working in marine geophysics and had a blow at a start pulling away with her advisor. So she stopped, and she was looking around for her for some new field and some new top again. So she came in some some me and at the time she was sort of sort of lost in the press perhaps and anyway I put her on this project and within a month she just changed completely just enormous I think enormous change she just, I think she just got the deeper implication of what she was working on. And it just affected her enormously. So she just adopted, she just took off. And she just, of course, her thesis was a tour de force. And then I was just in law, watching her develop her career afterwards. It was a most amazing thing. That happens to rare people that actually suddenly get that inner inspiration of what it's all about to do science. And she is a great example that everyone's talked about. Thanks. Thanks, Chris. Okay, so I think I will just end with thanking everyone because we have only two minutes left. So the webinar will stay open for a bit longer. Otherwise it will be a very abrupt ending of this session celebrating patients. So I welcome everyone to stay a bit longer. So, yeah, again, I want to thank all the speakers and you for supporting this session. And I think it's, yeah, this session illustrates very well her great contribution to science but also, yeah, the way she inspired so many of us and especially younger scientists, scientists. So thanks everyone. Thanks for joining me today.