 So, what we've done so far is having, we've mixed some serious physics with serious observations and sometimes the mixture is uncomfortable. But the idea is that, of course, we're dealing with natural phenomena and it's good to have some physical guidance with the data. So, what I'm presenting is the physics and an interpretation of the observations but I think the physics are always there and, of course, there's, you must always allow for alternative explanations but the idea is that when you have some systematics as you're in the field is to look for some of the underlying physics, you might have an alternative physics but I think it's the best way to guide you through the observations. So, the models that I've presented and some of the models that I'm going to present should not be considered as true. The physics are true. Whether they do account for what's happening in one particular system of course is open to discussion. So, the last bit is we're going to stretch the time scale a little bit more and talk about how magma reservoirs get in place and how, what happens to them as they get larger and I think that's an important part that you have to deal with that we have the plutonic record. Now, plutons are fully solidified magma reservoirs whether they were once volcanic reservoirs is open to question in some cases, I will mention them briefly you can associate them with volcanic deposits so it's clear that they fed the volcanic and you can demonstrate that the volcanic are indeed compatible with the composition of the fully crystallized body that has been shown in the very large intrusion called the bushveld in South Africa which I will mention briefly. So, in many cases it's quite clear that these plutons did act as magma reservoirs and they're usually found in clearly magmatic and volcanic areas. So, that plutonic record needs to be understood and needs to be accounted for. So, we start again with the same diagram you're sick and tired of it but it's a good diagram. So, on the whole we know that we've been placed in the crust magmas that did differentiate that's what we see in volcanoes and we know that the bulk composition of the continental crust is evolved with respect to primitive magmas which are always basaltic in compositions regardless of where you look at them in or the big mantle plumes or whether you look at them in the seduction zone they're always initially basaltic and yet we end up with something which is between andesite and desite so there's two things that you need to do is to differentiate your mafic magmas but then if you want to get the residue which is fully evolved then you have to get rid of the most mafic part that's always been an issue and so I think on the whole you can see that we know that somehow we've got two stages in the evolution of a magmatic reservoir we have the differentiation and we see that in an active volcano we saw that at Mount Adams for example but then there's a final phase in which we have to lose some of the mafics now whether these are very separate in time or not is an open question so we know that magma scent is buoyancy driven so we know that we're in placing mafics that we'll show you examples that magmas are the rise to some level and many of them get in place at shallow depths maybe due to the edifice also maybe due to the fact that they encounter rocks that are less than them and I will show you one example that's based on a gravity survey of the rum intrusion in Scotland I will show you another photograph later on about the disintrusion that's a mafic intrusion and what is left now is the mafics so what you have is the cross-session that was built on the surface observations and gravity survey the deeper part given the coarseness of the gravity survey is not reliable we'll come back to that later on the shallow part is reliable gravity survey is given the scope and the size of the survey you have enough resolution for the shallow system you have the dips of the... you can see that this magma reservoir got in place close to the interface between the super-crossed hole, the sedimentary cover and you can see that it did deform it and we'll come back to that later on what's left now interestingly there are some volcanics in the area and volcanics are clearly... you can show that co-genetic with this type of liquid that might not have come for that particular system but there's a lot of volcanics in the area you can see what's left now is very dense and that's one of the problems that we'll come back to that what is left is the mafic part, the fully crystallized part and this is much denser than a country rock so there's going to be... the story is going to be emplacement is dictated by the density of the initial magma and as the magma evolves then what's left in the reservoir changes and it changes in density so first phase one is emplacement so if you look at the run you can calculate the initial... that's for a dry magma there were some volatiles but we don't know that so dry magma we calculated that the pressure at which it was emplaced so you do that by using some codes like the melts that comes from the melts which allows you to figure out what's crystallizing out of initial liquid at any pressure you do have to specify the redox state but this is standard redox condition and this is what you have for the incoming magma for the run you can isolate the incoming magma from primitive dykes that are found around the body and that tells you as a function of temperature what happens to this magma so that's the bulk residual liquid plus crystals and this is the crystal cargo and of course there are breaks in the curve depending on the crystals that appear on the liquidus but you can see the basic story is that as you differentiate you start initially with a magma that was not very dense about 2.65% and then it differentiates and then becomes dancer and dancer and of course the crystal cargo is even denser very very large amounts of mafics and here you crystallize a lot of plagio clays which is less dense which is responsible for this but overall you can see the big bulk density change that occurs in a reservoir like this now of course this is erupting so the story is going to be more complicated than this but the basic story is going to be still going to be the same you have a density increase as a function of time now you can see this I've summarized a few intrusions for you that's a small intrusion which I'll talk a little bit about because it's smaller and it's been looked at in the field in Scotland also and there's no information there but it's about the same system as the rum and we have other information later on the rum is there so the host rocks they are well documented in the rum the incoming magma was about probably there you can see that the density was intermediate between the density of the supercross souls and the density of the crystalline basement Settil is a very large intrusion in Quebec very large mafic intrusion beautiful intrusion again we do know the surrounding rocks very well and you can see the incoming magma was probably slightly buoyant with respect to the country rock the bushveld in South Africa is an even larger intrusion and again the same story the incoming magma had this density was intermediate between the country rock and you can see that in all cases the density of the fully crystallized the sample that's left the mafic cumulates are much denser than the host rocks so there's a very large density change between the initial liquid system and the initial fully crystallized mafic cumulate pile so there's going to be a second phase in which they're sagging and foundering and they might not be separate in time but of course for what I will do I will consider them separately so to go back at the rum this is what you have there's a fault that runs through the system you can see the size of it that's important to have the lateral extent and we have good controls on the lateral extent even at depth you can see slightly larger at depth than just on the outcrop there's a fault that runs through the system for those of you who are interested in it it's a beautiful outcrop there's some interesting layered series we'll not go into that here and this is characterized by a big gravity normally you can see the courtness of the gravity is not great and that's an island and of course there's not a lot of gravity data I'd see unfortunately but it would be very nice to have more gravity data on those systems because gravity data of course is going to tell us a lot about what's going on the Adnamaker system that's a Google Earth system you can see this circular outline the erosion level has just cut through the magma body and therefore you can do a lot of mapping the lateral extent is about six kilometers and that's from a nice paper which summarized so that's the present level exposure there's no gravity data so you must definitely think that all this is imaginary but the important thing is that you can see here that it's definitely a thicker body because of the dips so that's what the trial is called Lopolis which is a basin-like structure which is thicker in an actual zone but what's not clearly like Lopolis and I'll come back to that is that the tilting of the magnetic layers the inward tilting increases as you go towards the axis and that's a very important piece of evidence but it's a nice intrusion it's also interesting because it's not a very large volume intrusion so there's an aspect ratio associated with that intrusion which will play a role the thickness is a few kilometers and the width is about a few kilometers so the aspect ratio is about one now how can we replace reservoirs like this if you put them very fast there's a big issue about how you can deform it in the post crystallization phase the whole system has been heated up we know there's a stress that's going to be imposed by the Mayfix we know the amount of rough estimate of the stress magnitude is simply given by Delcero the density contrast with the country rock g times the thickness of the intrusion we know the density contrast that's very well documented the intrusion thickness is also quite well documented and we know the analogy of crux of rocks and I'll show you the examples you may assume several constitutive relationships between stress and strain rate but basically for these stresses these things are characterized by effective viscosity that are in a range such that these strain rates are large so these bodies will deform so just to give you an idea and that depends on temperature so this is an effective viscosity for that even stress level that I was mentioning these are non-linear logical laws and so the diverteric stress magnitude is calculated as I shown previously as you can see as a function of temperature you can have very small effective viscosity so country rock will deform if it's carried to temperatures basically about 400 degrees C depends of course on the exact constitutive relationship between stress and strain rate but the story is that if you heat up the rock they will deform over a small time scale geologically you can relax the stresses in a few tens of thousand years at this rate so in placement first phase and the one thing you see is that there is a relationship between volume and aspect ratio so this is the Adnan-Larkin intrusion small volume 30 cubic kilometers as per ratio of all the one rum intrusion that I've shown previously larger volume 500 cubic kilometers as per ratio of 0.5 Setil, this beautiful intrusion in Quebec it's unfortunately only a bit of it are found on the outcrop and most of it is under the St. Lawrence river but this has been mapped using gravity anomaly almost perfectly circular outline beautifully and large volume very large intrusion aspect ratio of smaller the Bouchvel this is even larger that's in South Africa that's older this is about 60 million years old but I forgot now it's about I forgot about maybe sorry I forgot it has to be someone linked to the St. Lawrence river so maybe 400 million years ago and the Bouchvel is about 2.1 giga years ago so 2 billion years ago it's a much larger system and the outcrop is it can see stretches over about 400 kilometers it has an irregular outline which I think is an important thing so you can see there's a relationship in volume and aspect ratio and that's one thing that I'd like to emphasize so we've done experiments and I will show you some simple theory to account for to find out how you can inject the magne reservoir and what determines the aspect ratio that you see so you inject liquid at the interface between two layers it's a stably stratified system the upper layer is less dense the lower one and we initially inject the liquid that's hot and at its high temperature its density is between these two densities so it will get in place right at the interface but when it cools down it will become denser than the lower layer of material and then it will sink so that's a simple way of representing the big density change that I explained that is due to cooling and crystallization in a simple way but I think it tells you a lot about what happens in terms of physics so you recall this, that's the tank and these are tricky experiments because you have to monitor density to a few parts per thousand so it's a big... but you can do it if you're careful so let's first do the physics so now we're in placing the system let's assume first that it's constant viscosity, constant density and you're in place at the interface between two fluids and we might allow for different viscosities in the upper fluid and lower fluid they have different densities it's stably stratified so density here is less than this one and the intrusion has a viscosity that's intermediate between the two so we're going to write down exactly the same balance equation that we wrote before so I will go fast it's the same type of argument but the same force is the viscosity at the top and bottom of the intrusion now the interesting thing here of course that we have continuity of velocity at the interface so we're driving flow in the surrounding liquids and the important thing is that we're driving flow over a vertical extent which is scaled to the lateral extent of the flow that's simply a result of the fact that in the Navier-Stokes equations that's the initial terms and you have a Laplacian so the Laplacian imposes that what happens in one direction affects in the other direction over the same kind of length scale so that's an important thing so you'll have the strain rate which is going to be linked to the spreading rate over that thickness and that thickness is scaled to the radius of the intrusion, the radial extent so for flow real distance r kinetic boundary extent of this r the Karatek velocity we'll call that U.S. now it's not U the subscript, spreading velocity because I'm going to compare this to another velocity so the strain rate is U.S. over r and the bulk momentum balance we've seen this before, it's exactly g' is the scaled gravity given the stratified system but it's g for all practical purposes and there's some stress at the top and bottom so you have to allow for the two sides so there's two viscosity involved you add the volume conservation it's a constant volume flow and you you get these spreading relationships so the same stuff as before no, no particular the important thing is that you'll see that the radial extent will increase as t to the 1.5 the thickness of the flow will decrease because the constant volume so it's the same thing as the constant volume flow at the surface which is spreading and becoming thinner as it spreads here it's in between two layers but it's the same principle and you can see a very simple set of relationships that of course was assuming there was no cooling so just give you an idea of what happens in experiment that's a side view and a view from the top so we're putting a fluid whose density is intermediate between these two densities when it's hot so it spreads but you can see that it spreads and then it starts deforming and then it starts thundering and it starts thundering because it has cooled and as it cools of course it has to go back down and you can see the view from above initially it does spread you see this radius here is larger than here and then we'll show you later on when it starts to go back down the radius will go back down because of course you have to feed this down occurring flow so this is what you see it has a regular extent of the flow as a function of time these are scaled because you want to compare experiments independently of the volume the viscosities but if you're dealing with one experiment it's just a straight radius in typical units you can see that this thing is spreading and then eventually it starts to sag when it's becoming denser so what I'm interested in is the aspect ratio at the stage where these things is reached at a natural extent so it's a maximum aspect ratio that you can reach before it starts to sag and that's what you find from the experiments the aspect ratio it's the thickness over the radial extent as a function of a dimensionless number which I'll describe it to all so we have data for 12 experiments and you can see there's some ratios here in between the two which I will explain but so 12 experiments it's a lot of time so what can we say? well the end of the spreading phase is such that the cooling proceeds faster than spreading initially if you spread fast then cooling will effect like the flow you remember we had this when we dealt with a lot of homophology but eventually the spreading rate has to go to decrease we have a constant volume so as the flow becomes thinner it has to go slower and so eventually we're going to have cooling because it proceeds faster than spreading if cooling spreads faster than spreading this is the time at which basically you're not going to be able to spread much farther and then you're going to start sinking because the cooling will overtake spreading so the cooling rate for some thickness h is it's simply a diffusion control so kappa is the mode diffusivity so that's a cooling rate the spreading rate you can calculate for what I've shown previously and the time for buoyancy reversal is equal to the spreading rate there's a constant proportionality and that defines a critical time at which basically cooling overtakes the spreading so you know that time you know you have the spreading relationships that specify how the radius and the height increase with time so you substitute this critical time in those values and that tells you the critical aspect ratio which is there and the nice thing is that time disappears of course and the only thing that you've left is a simple relationship which introduces this dimensionless number where gravity is reduced gravity and time velocity and that's a dimensionless number that's basically a measure of the the propensity to spread and the efficiency of cooling you can see that kappa if kappa increases this thing et cetera so a very simple relationship for those of you who like a convection it just looks like a really number and the predicted slope was one minus one third and you can see that you can calculate for the experiment and that's the best fitting relationship the experiments are not as good as you would like them to be but you can see basically the basic argument is okay and that tells you that the aspect ratio is going as this four times to the power one over three so that's important thing AI was increasing with V so that tells you that if the volume increases the aspect ratio decreases and that's basically buoyancy driven flow the thicker the flow the faster it goes and so the larger it can spread so you can rationalize this number again by simply a ratio between two velocity scales spreading rate versus cooling rate it's the same argument you can interpret therefore this number as a ratio between spreading over diffusion cooling so if this number is large spreading is faster than cooling intrusion extends to large distances smaller cooling is faster than spreading intrusion doesn't spread very far so I think that explains the relationship that we saw between the intrusion size, the intrusion volume of the aspect ratio it's a very simplified argument of course because we must feed these volumes into crust and crust is not infinitely thick so this is a very simple physical argument but it does bring up the point I think that's the largest intrusion associated with the faster propensity to spread and they will go to large distances so the bouchevelle is very large and the bouchevelle extends to 400 kilometers the argon americant is a small intrusion it can't spread very far and it chills before we can spread so that's already telling you that the size of its thickness, its lateral extent there is some control on cooling by cooling of course in nature there's going to be complications to that that's a basic story now phase 2 sagging and foundering so now we have this thing has been cooling it's going to become denser than the surrounding rock so if it's denser than the surrounding rock and so I think it's interesting also that there's clearly a trend in the complications in the way the foundering proceeds so a small intrusion regardless of what goes on this is going to be associated with a small negative buoyancy because a small buoyancy is the force is the density difference times the volume times gravity and you can see here there's the classical structure is interpreted which is interpreted as sagging of the floor but the fact that the tipping of the planes increases as you go towards the actual zone is not consistent with the flexural type of deformation it's more consistent with the foundering the fact that you start sinking so the sagging I think is another the correct way of describing this because the inward dips increase towards the actual zone it's not consistent with some flexure of the floor and it suggests an incipient foundering process if you go down to other intrusion that are larger that's a very beautiful intrusion called the Great Dike in Zimbabwe that's very well mapped because it hosts a lot of interesting mineral deposits including platinum deposits and that's the cross structure through the Great Dike in Zimbabwe so it's very well characterized because you have a long strike of the Dike, the Dike extends over a total distance of 500 kilometers so you have depending on where you stand you have observations at different levels and we'll come back to that later on but so basically you have the tilting of the igneous layers and you have the basic structure from the gravity anomaly the observations we'll see them later they stop short of about this level so classically this is interpreted as a feeding structure that's a feeder zone I think it's a wrong interpretation as a feeder zone it's just too wide you have a kilometer wide system here and a kilometer wide fracture that would feed magma it's just way too big it would allow buoyancy driven flow faster than the speed of light so this is not a proper width I think it's a foundering feature and clearly you have a different shape you can see that maybe the adenamercan is an intrusion that's recording some of that feature but stopped before it reached that point and so there's the suggestion of an increasing deformation with increasing volume sagging fellow structure sinking if you sink completely then of course you have no record left the other thing we can see there's complications when the reservoirs are very very large that's the bush vell and there's an interesting story for those of you interested in this about the way the gravity anomalies were handled it took about two decades to sort out the story here but that's the end story it's been checked by gravity anomalies local and large scale it's also been checked by seismics because that's hosting the largest platinum deposits in the world and that's the story now but 20 years ago this was not established because the gravity data had been misinterpreted and for those of you who are interested maybe in a discussion session and that's the structure that you see you can see this thing is extended over very large distances I think that's on account of its large volume these larger insurance can spread very far but it's very regular it has lobes northern lobes, western eastern lobes, southern lobes the southern lobes is below the sedimentary layers in cross-section it has a complicated shape interestingly enough there's a difference between the basal series the most mythic parts there are ultramaffects basically you have theonites and paroxonites for those of you who like these types of rocks and then later on you have periphytes you have more complex mineral assemblages but the central part is missing the ultramaffects very interesting there's a general lobulis shape and at the edges you have these very thick lobes at the end the lobes are thicker at the terminus and that's also very so you can see a complicated shape the central part of the intrusion which is thinner and which has lost some of the ultramaffects and thickening at the edges you can also track the dips of the layers at the edges so there's an inward dipping layers over there and so there's quite a bit of information about the structure so I think the story is a sequence of increasing deformation with increasing volume complex shapes for small aspect ratio intrusions even more complex for large aspect ratio intrusions so we go back to our experience and see the kind of flow patterns that we can develop now we have to deal in 3D and so smaller volumes for large aspect ratios you're in a teardrop regime where you just developed this teardrop if you think in terms of the final shape structure like the gray dye it doesn't stretch the imagination very fast we see that this is consistent with these type of shapes but if you increase the volume and the size these things will spread farther and you get into very complicated patterns so that's side view there's a reflection of the light reflections of light there's a too complicated thing here but then you can see from the top what happens that's what we call the jellyfish regime it's nicely outlined in the view from the top what you have is thickening at the edges and the edges start sinking but there's also this pattern of lobes going down what you might call diapers going down in the center of the pattern and then eventually it goes down you can see the ring starts to become stable to smaller scales and drops forming out of the periphery so a beautiful picture and this is perfectly controlled you can make experiments with slightly variations of volume and come back to that later on it's perfectly reproducible pattern now what you see is therefore you look at the thickness of the function of distance at different times eventually the flow starts to spread but then it starts to become stable and it starts to thicken towards the edges and sinking in the center and then developing into this jellyfish pattern so the jellyfish pattern you can observe it at different volumes so within the limited experience you have only a factor of about 5 between the smallest volume and the largest volume and of course the largest volume the larger the structure and of course the bigger the structure that you form bigger diapers so there's some nice scaling that come out of that there's a simple physical control of everything that goes on in this system there's another regime that develops when you have a slight difference in the viscosity ratio between the intrusion and the surrounding material and instead of using the jellyfish you only thicken at the center and it's a matter of just a wavelength of the most unstable instability that develops and there's only sinking at the edges and you develop these low-bate structures thick at the edges, low-bate structures but you can call that as a variant of the jellyfish regime I'm not going to develop the why there's different regimes the aspect ratio is an important feature you only develop these regimes these complicated regimes at large aspect ratios for very good reasons I don't have time to go into the details but we're going to track these regimes I'm going to find the equivalents in nature so this is what you see so small volume so it's large they would produce a teardrop regime if you have a larger volume you would produce larger smaller aspect ratios and you would go into these regimes here so larger volumes, larger values of AR smaller aspect ratios and the break is about there but definitely there's more theory to be handled so the story is that the larger the volume you would produce the largest intrusions and you would produce the most complex patterns it's not really that surprising and again these patterns are nicely developed I think this is very similar to the low-bate pattern that we had in the annular regime you produce these thicker edges and these low-bate structures and then it's tempting to associate with this type of structure that we had and in the idealized cross-section with staining it's a bit idealized because there was no gravity data at this stage it's clearly second at the edges and it's exactly what we have in these structures the low-bate structures and thicker edges and I'll come back to that later on with slightly better models for these structures so what we've seen is that intrusions are things that might evolve with time they will evolve slowly with respect to the time scale of an eruptive sequence these things are, if we are right about the typical discourses and typical rates that the crossroads can deform you will change the dimensions thickness, size of your reservoir over time scales that are maybe some things like 100,000 years and so this is an important factor and so this is slow compared to the time scales for the volcanic system itself but it's something to be taken into account when you want to interpret the igneous record we have the igneous record with the plutons the plutons are recording the final stages of a volcanic system and so of course it is not exactly what happened during the volcano but it is recording things that did happen as the volcanoes were active now the last thing the experiments have done I've shown you before they show you these very complicated patterns that can develop there's just a typical instabilities that you would expect in stratified food systems no surprise there the pattern however, like this were never described before because it involves the air spreading and the cooling and the thickness that the igneous things like this have never been reported but these instabilities are in food layers so as soon as you've got some buoyancy force the food deforms and everything goes down so the end result of all our experiments was all the fruit ended up at the bottom of the tank which is not what happens in nature because in nature the rocks can deform but the rocks have some strengths so the rocks can sustain strengths and we of course we see plutons the plutons are there they are shared with gravity anomalies so they represent a load on the crust and this load in the case of the bourgeois this load that's being there for about 2 billion euros so it's safe to say that the crust withstands that load and so we have to go to more complicated physical considerations to see what is the preserved shape that you can get so what I'm going to show you is a result of numerical calculations done with a code that handles elastic deformation brittle deformation and plastic deformation ductile deformation so I can't describe the code in any detail it's been checked we checked it against the laboratory experiments that is in the limit of no elastic behavior no virtual behavior and the purely Newtonian relationship between stress and strain rate it does reproduce things it's a 2D code and it takes quite a bit of time to run calculations but this code is going to be able to show us what happens when the reservoir goes unstable but also what is left because the shape of the body that will be left is the shape of the body that can be sustained by the strength of the rocks so what you have is that's the basic setup of these calculations we mostly did calculations over the crust so at the base of the crust we allow for definition of the crust if you deform the crust you have a restoring force because the crust is buoyant with respect to the underlying mantle so this is so-called Winkler restoring forces so if you deform the crust there's a push-up because you're trying to push buoyant fluid into a denser fluid and at the axis by symmetry there's no velocity you can't spread away from nowhere and no shear stress it's a maximum and at the far edges and the calculations we made sure that they were not influenced by the lateral boundary at the top were a free surface the free surface can deform the surface and that's included in the calculation so we start with an intrusion we start with a half width of L now it's another radial intrusion it's 3D of a thickness H we're going to track deformation using several layers so these layers are initially the rock has the same density of rho i so we're not allowing for different densities between the different layers but the color scheme will just allow you to track deformation as this thing goes down and initially this thing is we start at the stage where the intrusion has spread it's now denser than the country rock and we're going to follow what happens again a stratified system and we're going to track deformation now so we're only looking at the foundering stage not the initial emplacement stage so we found three regimes one is the sinking regime if you're dealing with a big density difference and or very weak surrounding rocks and weak surrounding rocks can be either because the geology is intrinsically weak depending on the geology you can be weaker than others but also you can change that by temperature these two can offset one another so you have weak country rock or big buoyancy force you sink and you don't leave much at your initial emplacement so that's the initial intrusion and it starts to sag it starts and then it develops this very thin down-going you might call that dive here and in that particular case we're spreading at the morrow because that's the base of the crust and we can discuss that later on if you want to that's also an interesting phenomenon for coastal differentiation of this talk plan and that's the temperature nothing particular, that's the temperature field and you can see this thing deforms now I would like to stress the fact that you can see the time scale hundreds of thousands of years the geology is just a geology that's a soup of coastal rocks and the load is the load that we know so this deformation unless these calculus are completely far off it's exactly what would be predicted and you can see that indeed this initial state is not stable it has to go and deform into these bodies we haven't imposed deformation at all the other thing I'd like to stress is that in this particular case not much is left at the initial emplacement level and then at final several years there's still a bit of deformation but not much so what you're going to be left with is a slight sliver of stuff over a very thin zone which might be taken as a dike actually the temperature field is you start hot because you're reservoir is hot and of course it sinks and of course there's an advection of heat as you go down and of course that's the intrusion because it's hot heats up the surrounding rock and weakens the surrounding rocks so why does the green layer of the black no because you have the return flow that's the question which one over here there there it's okay everything is entrained by this thing so it's things here and because you've got this temperature difference here there's a bit of uplift there also it's been the intrusion is out sorry? well okay we will discuss that later because I'm not allowing for my boundary gradient stops here so the only way I can go is there so depending on the density of this thing in reality this will continue to go down in this denser fundamental and it might spread if it's that's where we're very deep you're here I've taken a cross of 60 kilometers to stick here that's the cross of 60 kilometers to stick here yeah but you can change the cross of thickness if you're I'm not sure what you want to drive at there's two things here the cross of thickness is assumed to be 60 for the bourgeois now is about 40 kilometers we know it was thicker but we've lost a lot of overburden and so that was to be able to restore the lost overburden and the spreading here is an artifact of this calculation depending on the density difference between the stuff that goes down and the mantle if it's dense without the mantle it will continue to go down but in the bourgeois it didn't go down and that's something that's very interesting we can discuss that if you want to so that was the sinking regime at the initial emplacement level there's not much that's left there's a residual intrusion regime in which you're having rocks are slightly colder the tank control here is the weakness of the surrounding rock they were all intrinsically weak and now that's a consequence of defamation and you produce a final look at the time scale here these things deform very slowly because they're colder so the definition occurs much less rapidly than in previous case and in the end you end up with a stable structure again spreading at the morrow but that's an artifact of the calculation and that's a final structure that is stable at 9 million years there's no more defamation there's a bit of defamation left it's negligible, it's less than 10 to the minus 20 second minus 1 so negligible temperature field nothing surprising but that's the important thing we have a residual intrusion regime because there's a big body that's left at the initial emplacement level however it's been strongly deformed it has, you can see the difference between here and there there's no, there's not a simple relationship between the two the residual body is not the same as the initial one but basically we've lost half of the volume to the morrow and we've left half of the volume at the so we'll come back to that later on you can see the layers how they deform, these are supposed to mimic the individual thickness layers and you can see that they thicken toward the center, you can see that the basal layer got stretched and in fact only lined up the wall to below a certain depth so there's apparently some discordance that you would record in the layer of the universe rocks and I'll come back to that later on so there's some very specific features of this type of instability and the way the defection of the body is reflected into the the layers that are left so the foundering proceeds differently depending on the intrusion aspect ratio these calculations are done exactly the same conditions just to show you so now we have a different larger aspect ratio and what you start developing instead of a centrally central down-sigging what you will produce is a series of instabilities which are akin to what are called relative instabilities but the instabilities always occur first at the edges because this is where the intrusion is coldest because there's a competition between these different and producing these residual intrusion the interesting thing is that you've thinned the central part you have a thicker outer part and with dipping layers and there's no dipping layers there a complicated pattern which reflects the you can see again the time scales because we're in this residual intrusion regime the rocks are not deforming fast so the time scale is large now this is the best we could to do the bushveld so much shallow intrusion and because it's shallow, at shallow depths you've got very cold rocks so it's going to be very hard to deform those rocks so most of the defection will occur at depth and you can see how the pattern evolves and the end result, 2 million years then basically nothing happens after that you're left with a thick outer lobe and a thinner central region and the thinner central region has lost its basal parts very important features that we saw in fact I'll come back to that later on so that's the residual intrusion regime it's the most interesting one how can we match this to with a geological record so before I go into this let me summarize I think what we've done so far is to say okay do we want to look at the emplacement stage because of course as we emplaced magma in the Earth's crust we have a buoyancy force that's acting and it's unlikely that the reservoir will remain static it will deform there's going to be eruptions of course which will withdraw liquid etc but that thing is likely to deform and if it's likely to deform it's likely to spread and so to spread to large distances and that's going to be a function of the volume that's available it's also important to track the aspect ratio and the lateral extent of this body because as it crystallizes and cools then it will generate different types of behaviors when it starts to founder there's no doubt that most basal intrusion should be involved in foundering because we know that the mafic accumulates that they produce are denser than the country wrote and I've shown you examples from the geological record geological record of course you might have a complex situation in which you're still having a layered magma body with dense ultramafix and lighter evil rocks, they are anothozytes at the top and we can discuss anothozytes for those of you who are interested anothozytes have plagioclates accumulate and so there's going to be a density stratification within the reservoir that's of course the next step this characterization was done a year ago there's many different things that we can investigate but the basic story I think is that you're going to start losing your mafics and there's a second stage in your magma body when it's almost fully crystallized that's going to generate a completely different structure and this is the structure that you're going to be able to observe in the field when you go to plutons so does it make sense that's a cross-section of this great dog of Zimbabwe why it's because we've been studied in the last section you can see the lowest exposed stratigraphic level is over there because this is stretched over 550 kilometers the beautiful structure I could have shown you in fact it shows up very nicely on Google because it's so long and so big you can see the width something which is more than 10 kilometers wide a big fat intrusion body and there's always some puzzling features for this type of intrusion is not the only intrusion of its kind there's many other ones there's a nice intrusion called the Muscox which is nice in Canada there's also the Jimbalana in Australia and it's always been a puzzle to petroleum the fact that the layers could be dipping at such large depths in fact the Jimbalana is almost vertical and all the mechanisms that can produce a new slurring are always gravity driven so it was always a puzzle and it disappears by gravity processes against a vertical wall so of course this problem disappears if you assume that this thing was not horizontal when crystallization occurred first thing, second thing there's always these funny discordant structures at the wall you can see the border group which is the earliest crystallization crystallized part and you see that it lines up the wall but it disappears somewhere there's a discordance between you would expect crystallization sequence to proceed from border the night this is a typical sequence that you see in many of these intrusions but the border group has been truncated and the other layers are also truncated so these discordance are always been a puzzle and of course the best way is to find out to assume that this is not the way the intrusion crystallized it's the way the intrusion did deform after it has crystallized and could do this very dense cumulates if you look at this structure here that's the final structure that we saw and you can see we have the same features this lower layer that's the border group it gets truncated and there's of course you disappear some of these disappear as you go up along the wall and also the other thing that we notice is that the layers are thicker at the center so all these features are I think can be simply explained by this type of post crystallization definition same story we could do higher resolution calculations these calculations are not that long but they take now on a fast they take about three days we could have done something a little bit more softly again you have to deal with the definition of a larger distances laterally because as we saw if you deform over 40 kilometers vertical you have to at least have several times that distance horizontally to make sure that you're not influenced by your lateral boundary so your domain is much larger than this blow up if we go back to that structure to the rum structure if you remember it's an interesting structure because it has again I think this route now this route is not well outlined by gravity so that's the best the gravity people could do since then people think the gravity can't extend that deep so I'm not going to say this is proof that there's a funnel structure there but it's proof that there's this big fat thing here which is dense because that's mapped by the gravity field and I think it's very likely that it's due to the foundering we don't have the the interesting thing is I will show you things near the edges that's what you see near the edge of the rum that's the country rock it's layered you're sitting on the top and very nicely there's some interesting structure here so the rock has been obviously suffering but you can clearly see that these the same areas dip towards the intrusion, the intrusion is here you sit on the intrusion here and so that they form and I think that's also consistent with this type when you're higher up in the intrusion you deform the country rock but you don't of course completely you're still left and you just have this amount of deformation in the country rock and I think that's remarkably consistent with what we see here now last but not least the different regime that we saw so for a given reology the depths of emplacement plays a role but a small role so here we're not very sensitive to the surface but I've shown you cases for that's the summary of the different intrusion regimes and different final regimes sorry, as a function of background temperature so as you increase temperature you of course make the country rock weaker and so you're going from sagging to sinking what I would like you to introduce here these caverns were made just with the hot initial intrusion the country rock was cold or at its ambient temperature so there was no thermal halo around the intrusion and you can see how sharp the transition between these regimes are the residual intrusion regime is confined to a temperature interval of about 100 degrees C, not big and so you go from sagging to sinking quite rapidly sinking there's not much left in your geological record sagging there's a lot of left intrusion record and residual intrusion on it but if you allow for thermal halo this is the end stage of your working system you've had a lot of magma inputs in your system you've developed a thermal halo the thermal halo was scaled to for example the dimension that we see in the bush veil because the bush veil we have the metamorphic oriental if you have thermal halo that very sharp transition translate into a much wider transition because simply you've got a wider area of weakened rocks so the residual intrusion regime is in fact occupying a much larger parameter space and that's the end thank you