 Okay, thank you very much for the invitation to talk. My talk is going to be more data-intensive than model-intensive. I would suppose that the majority of you are not very familiar with glacier hydrology, so I'll use a chunk of the time here to introduce you to the complexity of the problem. That'll be the going on throughout, really, the presentation. I will use examples from the field from our work on the Bench Glacier and the Kennecock Glacier and try to show you that what we're finding out about alpine glacier hydrology is indeed applicable to the outlet glaciers from Greenland. I'll be talking a bit about each of the components. I'll talk of the hydrologic system. I'll use some simple models but I will end by emphasizing the challenges and the challenges lie in several aspects. One is the scaling. The hydrology of the glacier acts on sub-daily time scales and yet as a geomorphologist I'm interested in the evolution of landscapes over many glacial cycles, so hundreds of thousands of years. How are we going to get between those scales? I'm not going to run my long time scale models on sub-daily time scales. I'll also let me very quickly tell you the components of the system. There are pieces of the system, in particular the snowmelt part of the problem that's the source of the water for the glacial hydrologic system. That's essentially a Vados zone problem with all its complexities but we're dealing with a substance that's vanishing as its hydrology changes too. Even the snowmelt sort of source is a complicated part of the problem. Then the water gets into the glacier which by the way is opaque. We can't see it as it travels through and it operates, it moves through it in pipes that are much like karst systems but the pipes can refreeze and collapse by creep in the winter. So the whole hydrologic system has to be reborn each year. In addition at the bed, there's motion of the at least in some glaciers, there's motion of the ice against the bed, that's a fault. So there's fault hydrology as well with all the attendant complexities of that. So that's by way of warning about the complexity of the system. Because we can't see through it, we have to use proxy information about how the hydrologic system is operating. Part of the proxies for that are actually the sliding the glacier which responds to the water pressures at the bed. These are various motivations why one might want to understand glacier hydrology. You might be interested in as a fluvial hydrologist in the water output from the glacier in the time series of that, hydrograph of that, to drive sediment transport down the fluvial system. That's the coupling between the fluvial and the glacial systems. It turns out that the water transport through the glacier acts as essentially a heat source. It can warm up the glacier which can alter significantly the effect of viscosity of the glacier which allows it to speed up. This is one of the hypothesized speed up mechanisms of the outlet glaciers of Greenland. I'm going to focus on sliding because I'm a geomorphologist interested in long time scale evolution of mountain ranges in the face of glacial and interglacial variations. As I've said, we're going to use sliding as a proxy for telling us something about how the hydrologic system is working. One might be interested in the surges of glaciers as well which are definitely associated with blocking of the hydrologic system of the glacier for a while and then release of that water. Anyone interested in askers is worrying about how water moves at the base of the glacier in addition outburst floods which can be right atop the glacier hydrology, the annual hydrology you need to know about how the hydrologic system evolves. This is just a way of advertising what we can do or what we'd like to be able to do is evolve glacial valleys from the dashed line to the present red line as glaciers come and go but the time scale over which this happens is thousands to tens of thousands of years in order to do the sliding right which causes the erosion we have to do, we have to have a model for the hydrologic system. One of my points is going to be that these systems are distinct from many others and that the plumbing system collapses on an annual basis maybe even in fact it has much shorter time scales when the ice is indeed quite thick and as I said before we're finding that there are analogs between what we're finding on alpine glacial systems and Greenland. This is an example of the complexity of a daily hydrograph so snow melt happens during the day and the system goes to sleep at night, reborn during the day during snow melt but this is the, pardon me, this is the surge, hang on let me go back so this is a video taken of a time lapse image taken of the variegated glacier in Alaska and you can see the surge coming around the corner right so the ice is accelerating it right here it's going about a meter per day up here it's doing 30 meters per day and there's all kinds of nifty ice tectonics that goes on in this compressional mountain range that's happening in real time here okay so in order to understand how the surge works we would need to know how the hydrologic system in the glacier works yeah we better go get this camera quick, obviously we did right like we better do it real soon, there we go okay so this is an illustration of the kind of hydrograph one gets this is the water input into the glacier actually it's the water output from the glacier pardon me here water output from the glacier and notice that over the course of the season there's not much in the way of daily oscillation at first and then it gets to be significant daily oscillations in the water discharge from the glacier that has to do with the daily oscillations of the snow melt at the surface early going in the early going there's not there much of the melt that is occurring is going into the snow pack and warming up the snow pack so that it can ultimately release that water as melt into the glacier the components of the glacier system include the snow melt at the surface which is occurring on snow that varies in thickness from significant up here to at least late in the season to no snow down here so you're just melting ice it's going into the end glacial system and ultimately feeding the subglacial system so in general we often see this that the melt rate exceeds the output hydrograph or at least early going that's certainly the case when the water is simply going into the snow pack and warming up the snow pack so that the output hydrograph would be low even if it's not just going into the snow pack but going into the hydrologic system early in the season the hydrologic system is all closed up because it it had 200 days to creep closed during times of no snow melt so that the pipe system has gone away no pipe system you're putting water into the top there's very low efficiency of transport of that water through the glacial system and therefore what water does get in there's a significant difference between the melt inputs and the water outputs from the glacier that means the storage in the glacier is going up that raises water pressures and promote sliding so we see what's called a spring event associated with this mismatch between inputs of water and outputs water components of the surface system include channels on the surface which in themselves are beautiful look at this meandering system on the root glacier in Kennecott one thing that's different about the surface system is that it disappears in holes these are moolins that provide water input to the end glacial and subglacial system at points so the water that is in a distributed way being produced on the surface gets input into the subsurface at points where that's not been taken into account in glacial hydrologic models one of those places where the surface system gets interrupted is in crevasses which ultimately lead to moolins the end glacial system is all I'm going to say about it it's hard to see we can drill holes in the glacier when we count up the numbers of voids that those holes intersect it looks like the porosity of the ice is something like 0.2% very low porosity substance and that will translate into very very reactive oscillation of the water table within the glacier because it's so low porosity and that in turn will translate into variability in the water pressures okay the subglacial system very likely looks something like this and we talk about it as having a slow system and fast system the fast system is a set of conduits which ultimately exit the glacier at one or two outlet rivers the slow system is a set of interconnected cavities which are simply riding back behind bumps as the glacier slides across the bed it opens up voids back behind those bumps and those voids develop some interconnectivity so here's a question from left to right over a complicated bed gives rise to a set of cavities which interconnect through thinner orifices this view of the subglacial system put forth by CAM is nicely illustrated when we look at deglaciated beds and Hari Rajaram here in the Civil Engineering Department has done work on the hydrology of faults that shows the isotropy of the transmissivity of the subglacial system or of a fault system if you fault with sliding left to right you open up voids that tend to connect in a transverse way it's also analogous to and the evolution of that fault system might well be analogous to or fault hydrologic system it might well be analogous to what happens in a karst system where in the case of karst you're dissolving the walls in the case of the subglacial plumbing system and melting the walls the melt rate goes as the discharge of water so as the water discharge increases the melt rate increases developing bigger and bigger pipes which compete for the water ultimately leading to a few raining pipes that are interconnected with these spotty sort of cavities but all of this happens underneath an opaque substance so when we tell you about how we use proxy information from glaciers themselves in particular how fast they go as a means of telling us something about the subglacial system so we've gone to work on the bench glacier shown here in May and here in July and one of the first things that we saw on the bench glacier was that there is a spring speed up event that tends to happen on an annual basis where much of the sliding of the year takes place in short order and this is told the story is told by GPS monuments we contact to the ice surface or traveling with the ice surface we can tell their horizontal position and the vertical position with the three GPS coordinates this is the raw data I've taken out of that raw data the background speed which is associated with internal deformation of the ice any variation in that speed has to be sliding at the bed you're not going to change the viscosity of the ice on short term so that has to be sliding so this bottom plot is a plot of the displacement time series associated with sliding and you can see that there are a couple events one that gets it going here and then a big event here complicated plot I'm not going to go into the detail of that for various GPS monuments on the ice surface each one of it shows an event that starts it off and then a big event and then it goes quiet so here's the speed the derivative of the displacement and this is the vertical coordinate the vertical looks like a significant displacement is displaced vertically by 20 centimeters at GPS monument 3 and then as soon as the sliding stops that the vertical comes back to bed parallel I've taken out the bed parallel motion so this is actually uplifted the ice surface the first event looks like it translates up glacier and then the second event is essentially isochronous at this at the same time at each of the GPS monuments and it's associated with the very actually hot winds that came up valley most winds on glaciers come down valley this really put a lot of water into the glacier and it sped up and then went quiet despite the fact that the air temperature remained high notice that the water discharge took a big leap upward at this point and then stays high what that tells us is that the subglacial drainage system the conduit system has been put in place during that big input of water and at that point you have an efficient bleed of water and hence water pressure out of the system and the glacier has done its sliding notice the time scale for the decay of the uplift of the ice surface is on the order of about 8 days this is in fact about the time scale we would expect for collapse of the cavity of voids at the bed if water had simply been taken out of those voids given the thickness of the ice that we know from radar it would take about 8 days to collapse so here's my poor man's animation of this right here's a GPS monument tacked to the ice surface it's going upward hang on let me go backward here so it's tacked to the ice surface so it's both going horizontally and vertically with the ice as voids get created at the bed but now the water pressure water has been put out of the system because conduits have been put in and now the ice surface is going back to bed parallel so the pieces of physics include the growth and decay of cavities at the bed the growth is associated with sliding times the thickness of the height of the step in the bed there's possibility of melting of the cavity walls due to translation of water through it and then here's the collapse and importantly the collapse goes as the pressure minus the water pressure to the end power and is the rheologic component exponent so it goes as the cube it's a non-linear rheology so the strain rates go as the cube of the stress that's where the end comes and you can see that if I drop the water pressure significantly then the collapse rate is going to be dictated by the ice pressure so rho g h the thickness of the ice cubed so the thicker the ice you double the thickness of the ice the collapse rate ought to go up eight fold right and we can actually use that to model the data and we can do quite well modeling the vertical uplift and decay of that uplift with this both sourcing of a cavity source by sliding and subsequent collapse after water pressure is taken out the conduits have the same physics effectively the rate of change of the of the cross-sectional area of the conduit goes as the melt rate and here it's driven by the water the melt rate of due to water translating through the conduit and then again same physics behind the creep closure of the conduit so one can link all of these elements together Mark Kessler and I did this a few years ago linking an algorithm for cavities which are linked to a conduit which starts at essentially a minimal conduit at the beginning of the season and I'm simply going to show you the animation of that we get this rug flap like speed up which translates up glacier as the conduit gets more and more efficient it bleeds off that water pressure and stops that sliding from occurring just as an advertisement we see the same kind of rug flap like sliding history on some of the outlet glaciers in Greenland this is simply a compilation of sliding histories for many years on an outlet glacier on the west side of Greenland each one of these gray lines is a sliding history let me go on to the other experiment we've worked on which is the kennecott glacier so from the bench glacier 200 meters thick had a decay time for those cavity closures of order 8 days go to the kennecott glacier is twice as thick or so it ought to collapse and be much more reactive to variations in melt we didn't say much in the way of daily oscillations in sliding on the bench glacier what's going to happen on the kennecott one of the cool things about the kennecott is has this blocked drainage here which results in hidden creek lake being dammed up by the glacier itself another cool thing is that it's got Donoho Falls lake right here which acts effectively as a monometer in the system which fills during the outburst flood from hidden creek lake as it travels through a tunnel 15 kilometers between here and the terminus there's a lot of water coming out at mccarthy over a steel bridge we can mount stage gauges to and here's Joe Wilder out in a boat sounding hidden creek lake so that study was done in the early 80s and showed that their target was trying to understand the hydrology of an outburst flood which happens on an annual basis here this is Falls lake right here at the crook between the main tributaries it's dry on this day the next day it looks like this right so 45 meters of water that's clearly subglacial water fills this lake for a day and then it drains again this is hidden creek lake before and just after the flood 100 meters of water 30 million cubic meters of volume is lost in a day through the flood and this is the stage graph from a pressure transducer in the hidden creek lake and here is the deduced hydrograph of water input into the lake sorry into the glacier here's the output hydrograph in blue and the kennecott river and you can see there's a translation took about a day to get from the input to the output and along the way donahoe falls lake in that little crook goes high stays high and goes low it acts as a little pressure gauge okay here's the pressure as it comes by so we went out there a couple years later and said well what's the response of the glacier to this outburst flood that hadn't been studied so we put out GPS monuments one two three four five we put in pressure gauges in donahoe falls lake hidden creek lake eerie lake and measured the output hydrograph same kind of thing in the record of GPS motion here is the displacement associated with the flood you take out these are the raw data you take out the sliding you can see how much is accomplished 4 meters of motion during the flood very complicated plot same kinds of variables that we talked about before I'm going to jump to the next slide which has an animation of the data that I hope you find illuminating I'm a mac I know you're a mac help you're there I just need to point to this okay here we go animation of the data so here's the here it is in real space here this is Tim Bartholomew's animation so here's the hydrograph the stage record in hidden creek lake donahoe falls lake and the terminus in proper order this is a GPS monument we're seeing the x y motions of that monument we put it in on mother's day and then went back to school so for a month here we'll just see the mother's day GPS and you can see there's a significant oscillation on a daily basis even at this time of year in May donahoe falls lake is actually dropped in stage since we put in a stage gauge here we hadn't gotten up to hidden creek lake we'll do that in another couple days when Tim gets back into the field so we've got 10 days of running record of the hydrograph okay he's back in the field got GPS 2 put in which is right adjacent to donahoe falls lake got in the next day and to put in GPS 1 notice even now you see the difference in the background speed it's thinner ice down here thicker ice up here so the speeds now he's got GPS 4 and 5 and he put in the pressure gauge in donahoe falls lake donahoe falls lake is the stage is rising as water is put into it down this river interestingly the daily oscillation of these monuments is big at sometimes and small at others donahoe falls lake as drained is now below the sensor nice big oscillations of daily oscillations of speed now donahoe falls lake is topping up just about to drain here it goes big response of the glacier during that outburst flood there went the flood past donahoe falls lake and out it came and determined us okay so that kind of data can be used this is the kind of thing you publish in JGR you don't publish the animation unfortunately but we could see for example here's just the few days before and then the flood itself you can see the down glacier translation of the speed up as the water moves its way down the glacier okay so we basically find that in both daily and annual and seasonal time scales you see sliding when there's a greater input of water into the system than output of water in the system in a couple more minutes I want to show you an analogy thinking about that daily oscillation we think about that from the point of view of porous material this is from more and Neil Iverson's work out in geology they were shearing a granular medium and they had pressure transducers in this medium and were very carefully monitoring the horizontal displacement and the interesting thing was that when you got displacement events as soon as you got a displacement event the pore water pressure went down that's because you dilate the material as you shear it well that's quite analogous to what's happening underneath the glacier where when you slide you create cavities those are effectively the pores in the glacier system we can quantify that by the same kind of system where we basically have a model for the generation of cavity space back behind these elements in the bed roughness elements in the bed it's a relatively straightforward calculation just analytically to show that the rate of change of sliding speed should go as minus the sliding speed and that's because as you create pore space you fill that water that pore space with water from the glacier which pulls down the water table which decreases the pressure which therefore slows down the sliding and the time scales for this negative feedback are indeed on big glaciers on the order of a day ok so I'm going to stop there and simply say that there's much left to do on all of the components of the system and that formal thinking of those systems is going to be challenging when we want to cross from sub-daily to glacial cycle kinds of time scales ok thank you