 Well, first of all, thank you very much, James, for inviting me, and this is my first trip to the annual meeting, and I hope it's not my last. I'm really impressed by it, thank you. The topic of the meeting is such that I have to wonder why, what part of it I fulfill. I think I'm pretty much the uncertainty, actually, as I look at it. But I did notice that James was suggesting that one of the things that we're continually doing is looking at new models, changes in models, and that's what I'd like to present today. Some of my co-authors are here, and they will answer any of the particularly curly questions, and one of the ones who isn't here, I will show some of his work, Sam Roy, who with his wife is expecting a child momentarily. So if we have questions, perhaps I'll cover those. First of all, the motivation, the roadmap, and what I'll be talking about. The Firm, which is the failure earth response model, is a model that we put together in response to concerns that many of the existing process models were not capturing some of what I was particularly interested in, and also it was not producing the kind of coupling between the geodynamics and the theory of geodynamics and the theory of the surface processes. So the two main principles of this are first one, which I promise not to read the slides in general, but these I think are things which will be centered to what we're doing. That is that the material detachment, no matter what the stresses applied are, still occurs as a function of the response to the stresses. So plastic failure or a discontinuous process occurs as the material is stressed and that is independent of whether the stresses happen to be applied by geomorphic processes or by geodynamic or tectonic processes. Secondly that the large displacements affect the strain field, the stress field, that is the strain affects stress and it also affects the strength and the strength field is going to be hugely important and this gives us a predictability in much of what we're going to talk about and a relationship in the topographic memory of the strain field and it's a very important component of what we'll be doing and I'll probably start out with that. So I'm really particularly interested in using the high frequency information that we have, that is the low frequency information, something I'll briefly go through, but we've been dealing with it for quite a few years now and it's really time to move on to the tougher problems I think and this is from the geodynamics, I think the rest of you are dealing with high frequency stuff all the time. So I will start out by looking first at the strength and strain relationship that is mostly number two of this, the failure earth response model firm and then we'll take a look at how that couples to geodynamics and how it's predictable and then from there we'll take a look at the introduction of an alternative formulation for surface dynamics that I hope we will develop in parallel with the existing one. Well first of all, let's clarify the kinds of frequencies that we're talking about, the low frequency behavior in geodynamics is something that we've been thrashing about for a long time and it's the relationships really between long term climates and converging or extensional or strake slip tectonics and they generally describe something like the, oh yeah, so that's not, if that pointer is offensive to anybody let me know, I won't pay any attention to you but you can let me know. This is a cross section across the Himalaya from India up in Tibet, it's a standard one, it's one of the early kinds of cross sections that we modeled and we modeled often and there's very few geodynamics modelers who haven't brought this up but all of this is, I think we're self-agreed to learn with it but it's also things that we can do relatively simply. What I'm more concerned with here and in fact in most of my research and most of what we're doing is looking at groups and back so maybe this isn't such a good idea with a pointer. It's mostly looking at the very high frequency spatial ridge valley relationships and the high frequency spatial relationships are usually associated with high frequency temporal problems too and I'm interested predominantly in this talk dealing not just with a gross Himalaya and the Tibetan plateau but also the small ridges and valleys and the rapid fluctuation of those ridges and valleys from time frames which exist at let's say in the hundreds of years down to the milliseconds of seismic displacements. So I guess one of the reasons is why bother and in fact as I look at this talk it's mostly about why bothering and what we're trying to produce. First of all the modern solutions these are solutions the kinds that I would prefer to work done by programs like child and so on do a fantastic job of representing the long term and sometimes short term fluxes and climate variations to the sediment that's being transferred but one of the obvious problems and one that is I find personally difficult to get over is that we look at the earth in terms of different physics depending upon whether we're looking down on the earth with surface dynamics or up at the earth with geodynamics and I'm trying to bring those two together. In addition the properties that we use in much of the surface dynamics are often difficult or impossible to measure things like erodibility and so on are important parameters intuitively they make sense but they're very difficult to measure but there are standard measurements of material properties that come from the engineering world and those are ones that I would like to move into. And third the kind of advances that we're seeing and many of them I think will be presented here allow us to begin to look at the full 3d stress field and rapidly fluctuating and slowly fluctuating and so the concept that we deal only with principal stresses and so on is something that we can now move away from and move on to something new. Another reason is because there are a number of parameters of a number of processes that occur in natural dynamics which are important processes I'll do it a couple times because I quite like it which are important processes we recognize their important processes such as seismic generated landslides and in addition there's a there fluctuations in poor pressure which we fluctuations in poor pressure fluctuations of poor pressure which also cause sliding and failure and they're very difficult to handle within the present scheme of our surface dynamics problems that doesn't mean that they can't be handled but the it becomes increasingly ad hoc sort of like a British automobile of the 1960s with more and more things hung on to a rather attractive initial model but by the time one gets it well developed it really no longer is the I feel is quite the correct way of handling many of these problems. So the I apologize I really have given a power point before in my life. Okay so let's go back first of all we're going to whip through the low frequency variations and in this case I'm taking a model of the Himalaya so it's about a 2000 kilometer model by a thousand kilometer model by about 100 vertically maybe a bit more than that and these are the ones that I suggest are relatively standard in this case of the in this particular region here we're indenting India or running India into Tibet from the left to the right and these are the representative strain fields above and I don't want to spend too much time on it except to say that much of what we're dealing with in geodynamics is of course is a very large field and we'll be looking at embedded melt models for the kinds of scales that we're interested for the surface dynamics and then feedback to the large models. In terms of the just advection link which is an important link then it's quite a straightforward picture that is where there's advection there's vertical displacement then we get topography constructed so if we start to one way in which we can couple the geodynamics with the with the surface models is by coupling it in with rheological changes as a function of rapid removal of material and one of the examples of this is the tectonic aneurysm the tectonic aneurysm model is developed really as if in response to these rapidly uplifted blocks which as they rapidly uplift they carry their temperature signal signature but also their strength signature they weaken the region there's a nonlinear feedback between the weakening and the erosional scheme can produce very large mountains and three examples which we've discussed are things like negaparbat where the erosional regime is a focused river power namtrabawar at the other side of the Himalayan chain which is also focused river power and another one that we've looked at more recently the St. Elias where they're focusing us through a vicious leisial erosion and that's a useful coupling and it's something that we've talked about for a while but one of the ones i'd like to look at now is the higher frequency one in which strength is coupled to strain that is to displacements and this is not something that is new certainly to those of us who deal with faulting and we recognize that a faulted region produces a gouge zone and weak zone shatter zone associated with it and the question that we're addressing is what does this how does this impact in the three-dimensional way on the evolution of the topography so we'll take an example which is from the channel river channel fault in New Zealand and the picture of the fault is looks like a this is this is actual data here by the way most of what i have won't be so let's take a quick look at it and the picture is a core zone with fault gouge very low that's a very low cohesion on the order of kilopascals and then going into the megapascals schist on either side with a zone a fault zone which i think everybody is familiar with but it's going to be important as we go on which is a decreasing intensity of fractures as one gets further and further from the zone if erotability is defined in most of the models in terms of an inverse function of cohesion it's actually of the one over the square of cohesion then a change in cohesion makes a big difference makes a huge difference in terms of what we're going to see and it also means that the orientation of these planes is hugely important so here's a generic model that's been published off and on for some time now and represents pretty clearly the way we see many of these faults although specific faults differ again a weak core and strong edges so the cohesion structure varies vastly in these systems from something which and they represent a more cool home space is down on the kilopascale level to something which is associated with the intact rock on the order of several orders of magnitude different and if we were to take a look at this we'd see that rivers in general have this kind of strength and the buttresses of negoparabhat in other places have a strength that exceeds that so the difference is huge particularly when we put it as a one over c squared term well what does it do to the to their standard model and this is the same standard model that we had before except in this case we're allowing the material the earth surface material to change its cohesion as a function of strain so this little video that we just played and that I'll play again is a is the result of a deformation field a three-dimensional deformation field which is imposed allowed to grow as a function of the strain localization so it's a strain softening those you could think of as fault zones those fault zones have the cohesion which is which is on the order of orders of magnitude weaker than the intact material and the erosional scheme is dominated the river scheme is dominated by that so the net effect of this is that the the strain field dominates the strength field and it the orientation 3d is something that's represented in the topographic field so at higher orders when these are fully coupled together with a flak model excuse me a deformation model and together with child we get something that in which we can begin to see the relative importance of the various components as it evolves and I'll play it again because it's quite nice one and in this case what we're saying is to sort of pick this apart a little bit more that is so the strain and strike field are coupled consequently this will be something in the memory of the topography which is associated with effectively mantle kinematics the in the first case there's the the elevation field that gets superimposed by developing cohesion field and in this cohesion field we see the strike strong bits the high peaks and the weak bits of the valleys and that then further if we put them both together we begin to see that elevation and cohesion aren't always the same but there is a relationship between the two and it also controls the velocity field so there's a complete feedback between the strain weakening which weakens both the erosional field and the material parameters the finite element solutions for the for the deformation field and clearly it is it's necessary as far as we're concerned to be able to represent this in a three-dimensional way so these patterns are very persistent that is the very strong regions can exist for a very long time the weak ones get eroded very very quickly which gives multiple multiple characteristic times to an origin and those characteristic times give characteristic rates which are very very different and an example of this is given by Sam Roy's earlier work in which he looks specifically at an uplifted block and the propagation of a in this case of a nick point like behavior as a function of weakening and the weakening occurs on these two fault zones one which is dipping at 30 degrees one which is dipping at 90 degrees the weakening has the same kind of shape as the fault zones that we were looking at before with a weak core and a more solid edge and if we run this it's not too surprising the the material outside of this by the way has no weakening it's on the order of intact rock and Albert again what Sam showed pretty clearly is that the rate of propagation is a function of certainly the weakening but also the orientation of the weakening so the fault zone that weakens very quickly and goes straight across as a strike slip fault equivalent whereas the other one is a dipping reverse fault equivalent in terms of the orientation of the strain plan this is important in that the the propagation of the rates is therefore coupled to the the strain the nature of the strain so the nature of the strain produce or the signal of the nature of the strain represents in the rates as well as the orientation Sam has since taken this with us and particularly helped from Greg has taken this I think in somewhat further and began to look at what happens when the dipping fault moves across into a system by which we have a the whole landscape develops as a function of the deforming region and what is quite clear is that the asymmetry associated with the dipping fault and erosion back along the fault plan produces a signal which again is characteristic that is it sits it will sit in the topographic memory and allow us to unpack these unpack from the topography something about the memory in the future so the intermediate conclusions of the strength strain relationships are that first about values a week and that the heterogeneity starts from the very beginning it is as soon as the origin begins to form heterogeneities there and from a modeling perspective almost all of probably all of us but certainly all of the things that I've done in my in my modeling life to assume homogeneity being with and it's a bad assumption how much of what we do it's we we will still continue to do it because it gets us started but it almost always gives us incorrect intuition from the beginning and I think that shows up particularly well the strength the anisotropy it's basic control in all origins and it reduces the complexity of the surface that is the increased complexity of the processes reduces the complexity of the remaining surface it's one of the nice things about systems which could be fractal otherwise and the 3d history is is a permanent memory memory in that topography but there's still a number of problems and opportunities that I'd like to go on about and one is the specifically the physics of the of the mechanisms which actually allow the displacement and bringing together the two sets of theories for surface behavior and geodynamic behavior so in this particular case what we will what I will try to do is look at the 3d strength field and this 3d strength stress field particularly and as examples take a look at the New Zealand Southern Alps this is a photo looking from the west to the east across the alpine fall and we'll talk about this region to be a little bit to begin with one of the problems of talking about new theory is that by necessity we criticize or implicitly criticize old theory which is not really my intention here so I want to make it clear that I'm responsible for some of this old theory so I can at least criticize myself on some of this and that is one of the points that James made he said that some of our results look like they look actually like some of the observations and therefore we think that they're right and in fact this is something which has got me a lot of free beer over the years because it does look very it is reminiscent of an origin in which we produce ridges and valleys and it's produced predominantly by a diffusional mechanism and diffusional advection coupled together it's been very useful it will continue to be useful but I think that we have to move beyond that or I have to move beyond that certainly case that I'm going to take a look at is the plate boundary of New Zealand and the area is in the Mount Cook region we take a look at two things one which is a glaciated valley and one which is a badly slumped valley right on the fault there so the point about this is that one of the one of the standard geomorphic surfaces that's examined are the iconic nick points and nick point behavior and nick points are generally described often described in terms of the relationship between the stresses which are generated from fluvial system and the response of the earth that is there's a threshold if it's overstepped and material can be removed in this particular case we're looking up a valley from the that I the first one I pointed out Franz Josef at Glacier and we're looking at the at a nick point which sits right about at the edge of the of the valley fluvial stresses are represented by that little piece of principal stress space that sits down in the same orientation as the river and this would be a fairly standard way of representing whether the stresses are going to be overcoming on in addition to that though there are also glacial stresses not too far away and the glacial stresses also operate on the same piece of rock but they operate with a different set of of conditions roge age shear stresses and so on in addition at the same place there are also the slope stresses those slope stresses not only act in the direction of the glacier in the direction of the valley but of course they also operate very strongly in the direction of the ridges which is right angles to the direction of the fluvial channel here and those stresses can be very large and they can be large enough to contribute to the tectonic behavior too so there's already a coupling between topography and tectonics through the stresses of the topography but in fact it gets a little more complex than that too so if we take a look at this we can see that that the the stresses at the bottom of the slope are very crudely represented by a reverse shear sense at the top of the slope they're represented by an extensional shear and there's also tensile failure possible further up as well in other words one slope has multiple different stress states it can't easily be represented in terms of a transport or failure in terms of a single value and then we put on tectonic stresses so again that same poor piece of rock is also affected by the tectonic stresses and then dynamic stresses those of earthquakes and that's really where i'm hoping to go today so if we then look at this through time that piece of rock is affected by all of those stresses plus the time dependent or climate dependent at glacial load on top all of these happening at a particular place so in order to construct firm to deal with the fact that these various stresses are responsible for the the failure therefore whether a threshold is overcome or not we reformulate the failure all of those stresses into a single tensor referred to as the total stress tensor and it has all of those components and it's this is mathematically not a difficult thing to do by the way it just means that we put into a single reference frame which isn't a river-based or glacier-based or slope-based but rather a geocentric based we then look at the earth failure in terms of the effective stress formulation allowing fluid pressure solutions this is something we normally do in geodynamics and engineering applications so it's not hard and we apply we allow a time-dependent strength material behavior and then we solve in 3D and in this case we use a constructed meshless method and that meshless method is something which has i think great hope for the future i'd like to talk to other groups about producing something here but this one is not a sophisticated one particularly and doesn't have any any transport in it so let's take a look first of all at the other system which is that great big landslide landslide system and federa has a poster on this that you can ask her questions about it later if we take a look at a model from this then that excuse me the model here is one which is a topographic model which then the stresses are solved for in this particular case the cohesion fields come from measured cohesions in the region and so we're looking up the valley across the alpine falls doesn't really matter too hugely at a three-dimensional model and the representation here is red is relatively strong blue is relatively weak and the strength stress relationship is such that we're going to look at the evolution of that strength stress relationship james is eyeing me so this is a strength stress relationship and its evolution as a function of fluid pressure if one takes a look at it you can see that there's a the weak region that is less than one migrates across the system as a function of changing the partial pressure fluid pressure we all know that that landslides are a function of partial pressure the local fluid pressure local fluid propagation we can solve for that now and in addition this other one is another example which we put on we take the same system this happens to be dried and put the wet one here but we look at the applied tectonic strains at the at the boundary and we look at the evolution of the failure system as a function of the application of those if you take a remember the first animation and this one you'll see that the blue marches at a very different way as that blue marches i don't know if you can see it in this slightly scruffy impact but the topography is changing significantly and it changes differently as a function of the kind of input that we that we would put on well there are ways in which we can i want to skip these because there's something else i want to get to particularly like we can because we now have the surface and the geodynamics in the same same reference frame we can solve continuously for both and therefore we no longer have an ad hoc replacement of one on the other they're actually fully integrated system but what i'm really interested in because i'm getting old and i'm interested in actually moving this thing forward fairly quickly and i'd like to talk to you about is the dynamics and that is the thing that i find most exciting about this is that we have the opportunity to look at some of the older paleo seismology in terms of a rather new technique and use all of the information available in the topography and that information is associated not just with a single scrap that may decay but rather with every piece of that of the topography that we can measure says something about its present history in terms of its stress state and its previous history in terms of acceleration and stress state so one of the things that i'm particularly keen on is and in fact is not particularly difficult is looking at the response spectra of a single of any particular part of the earth this gives two regions one in which is the response spectra of parts of new zealand which are calculated of course they're calculated for california and many other places and that response spectra is a function of the it's a function of the period and not only function of the period but also the nature of the earthquake so strike slips faults at different distances produce a different acceleration spectra than other faults and of course they have a different spatial relationship and the spatial relationship is also a good indicator of the spatial relationship of acceleration and the acceleration response spectra has a fantastic amount of information about the nature of tectonic geomorphic coupling at very high frequencies so there we go and we'll quickly look at one other example that we're doing i'll go through very quickly which is an ice sheet model applied to alaska in which we couple the presence and absence of the of an lgm load with a very high prism data set and we look in this particular case at the position of the high velocity ice streams and we couple it to our three-dimensional geodynamic model and these are some of the implications and i think that it's time to finish for questions thank you this could be quite disconcerting to me what you just said because if we have a block of homogeneous material and then you apply a a stress field to it propagate strain weakening let it interact with the surface processes then the steady state solution can has a lot of information in it that can get us back to what the stress field was and so on but in real earth materials with an inherited history hundreds of millions of years pre-existing faults and strains and so on it seems to me what you're saying is we're screwed because you comment on it. Rudy's asked me to comment on whether we're screwed or not with these and in fact my feeling is that we're less screwed with these than we are otherwise that is that these the weakness zones and so on are controlling variables and i think that we can demonstrate that in a number of ways so by not bringing them in we are we are making a very large mistake but beyond that i think that the the coupling through the firm approach allows us finally to look at the effect of the position of weak zones versus strong zones for instance in a river migration in valleys and ridges if we suggest that the the strength structure is only related to the one in the valley then and we play the same physics and the same extrapolation to the ridges we know the only thing we really know is that we have to have it wrong and so in this case this allows us to solve by a way which is in which we can go out and measure things like cohesion tensile strength and so on it allows us to solve and compare it directly to the to the observations so i appreciate the comment about the complexity which i think is real but i think this is a solution to it rather than the problem thanks you uh you noted that this approach can give us some solutions to things like poor fluid pressure landsliding that have traditionally given us trouble are there things in the surface processes that go the other way that you know you're modeling less than optimally by forcing into this other framework thank you yeah i think that there are a number of things i one of the real strengths of the of the current process directed regimes is that it allows us to look at integrated fluxes and basically it is an integrated flux if you use q over a certain amount of time and this is not particularly good at that but i don't as i look increasingly at this scheme i think that this does not lose any of those of the capabilities that is if we can define the flow field then we can calculate the stress field in addition the more we see about the the really exciting hydrodynamics where they produce a three-dimensional stress field the more we see that we can immediately put it on here so i don't except for the fact that we're early on on this i don't see it that there's any particular drawback other than my own clumsiness at bringing them together