 And our next talk will be also about MTO, but it's slightly different perspective. And it's from Dr. Charlotte DeMott. Charlotte is a senior scientist in the Department of Atmospheric Science at Colorado State University. Her research interests include ocean atmosphere, coupled processes, and their effects on intensity, frequency, and organization of tropical convection. She's in particular interested how scale interactions affect the Madden-Julian oscillation. And she also co-directs at CSU, the Department NSF Reacher Research Experience for Undergraduates Program. Charlotte, I'm looking very forward to your talk. All right, thank you, Judith and Anish, for inviting me. Let's see. So I believe you should all be able to see my presentation now. Great, I'll start my timer. So this is great. So Shudi and I sort of shared each other's slides beforehand. And I think we agreed that the topic of ocean atmosphere, a couple of feedbacks over the tropical oceans, with a particular emphasis on the MGO, is probably a topic that probably requires two types. So some of the material I will be presenting will be familiar to you, based on talks that you've heard this morning and the other day. And some of it may be new. So just to give you an overview of what I'll be discussing, I'm going to give just a simple sort of visual tour of how ocean atmosphere coupled processes work. I'll give my own take on the MGO overview. It's scale, structure, and surface forcing. And then for the bulk of the talk, we will talk about the ocean response to MGO forcing and then the ocean feedback to MGO convection. And I will close with just really a couple of questions on how these coupled processes fit into MGO theory. And I'll try to give you a teaser diagram for the next talk on ocean coupling in S2S prediction. Okay, so I'm going to start with this slide, which is just sort of a picture to help frame the problem of ocean atmosphere coupled processes. So what you're seeing in the background is just an image that shows tropical convection in the atmosphere, clouds, you have sunshine coming down, and the lower part of the figure, you're looking at the upper ocean. And then if I can get my pointer, you can see we have the air-sea interface here. And so there's this statement from Chidang's review paper, his 2005 review paper, that is, I think, really a good way to, or a good thing to keep in mind. He says that the atmosphere does not see SST. It only senses it through surface fluxes. So you may have the ocean heating up here, but the atmosphere does not see the SST directly. It sees its effects, or I'm sorry, doesn't see the ocean SST directly. It sees it through its effect on surface fluxes. So we can think about this a little bit more schematically. Well, first we'll talk about the basics of the surface fluxes. What does this mean? And so what I'm showing here are the bulk flux formula for surface latent and sensible heat fluxes. So you can see both of them involve the air density, some specific heats, a transfer coefficient, and this is a parameterized quantity, but really the two main drivers of the surface fluxes are the wind speed and then a quantity either delta Q for the latent heat flux or delta T for the sensible heat fluxes. And these are defined as follows. For the delta Q, this is just the near surface vertical specific humidity gradient where you take the saturation specific humidity set by the SST minus the near surface specific humidity of the overlying air. So typically at two meters. And then similarly for delta T, it's just the sea surface temperature itself minus the temperature of the overlying air. And so in the tropics and really over much of the globe, really surface fluxes are primarily driven by the winds. So the SST affects the flux through its impact on delta Q and delta T. And so it's important to understand that the surface fluxes depend on both oceanic and atmospheric properties. So what I wanna do now is sort of walk through this cycle of ocean atmosphere coupled processes. So we're going to start by looking at processes in the atmosphere. We'll talk later about processes in the ocean and how these two spheres communicate across the air sea interface shown in gray. So in the atmosphere, we have boundary layer processes that give rise to atmospheric convection that then force large scale circulations. Together, these processes apply fluxes of momentum, freshwater through rainfall and solar heating to the ocean surface. So these are fluxes into the ocean. Once you have these fluxes into the ocean, these introduce imbalances. And so the ocean will be constantly adjusting to these fluxes of momentum, rainfall and heating. And so these processes take the form of upper ocean stratification stabilization or mixing. This can give rise to barrier layers, as Shudi mentioned, which can affect surface currents and these currents can then affect larger scale or longer time scale phenomenon such as upwelling down and downwelling. So the accumulated effect of all of these adjustments is what gives rise to changes in sea surface temperature. So this is where we come back to the point that the atmosphere cannot see the SST itself and only sees it through surface fluxes. So remember, all of these processes in the atmosphere are also generating low level winds and also boundary layer moisture given as Q and temperature T perturbations. And it's really the combination of these things that set up the flux. So these blue lines here are what connect the surface temperature to the air and give rise to the long wave radiative feedback to the boundary layer. But it's the connection between on this side of the diagram that sets up the delta Q and delta T. These delta Qs and delta T's are acted on by the low level winds. And this is what gives you the flux out of the ocean to the atmosphere in the form of latent sensible heat fluxes and then by this pathway, you get the infrared fluxes. So these fluxes in turn affect the thermodynamic properties of the atmospheric boundary layer and this is what completes the cycle. These thermodynamic perturbations will be reflected in boundary layer processes which gets communicated to atmospheric convection, the large scale circulations and then you sort of finish the life cycle in this manner. So this is just sort of a highly simplified picture of these processes. So you can see the surface fluxes between the atmospheric mixed layer, sometimes referred to as the lean boundary layer and the ocean mixed layer, shown with these arrows here. And the point I want to make is that these fluxes are themselves regulated by processes both at the surface and at the upper and lower boundary of these mixed layers that affect the thermodynamic properties. So for the remainder of the talk, I'm going to try to break this feedback cycle down into three different phases. First, we will look at the atmospheric forcing of the ocean. Number two, the ocean response to that forcing and then three of the ocean feedback to the atmosphere. Okay, so why the MJO? First, it's a nice laboratory for studying ERC interactions, the MJO convection effects and response to oceanic processes on multiple scales. And of interest to this group, of course, is that the MJO is a source of predictability on S2S time scales. So just to give you a brief overview of the MJO, I think this slide needs no introduction, but what I want to show here is overlaying on this map is a satellite image of the convection associated with the given MJO event. And the point I want to make is that this is a huge, huge heating anomaly. So you can see that the cloudiness associated with the MJO active phase spans almost the entire Indian Ocean basin. And so hopefully you can see these blue lines. What I'm showing, these are the sort of the sea level pressure perturbations that are driven by this large scale heating. So you can see the Kelvin wave component stretches far, far into the Pacific Ocean. And then you have the Rossby gyres which are located in the Indian Ocean. So to the east of MJO convection, you have low level easterly winds. And to the west, you have stronger low level westerly winds, which are associated with these westerly wind bursts that have a strong impact on the upper ocean. So the MJO forcing of the ocean, this is sort of the first part of the feedback cycle that I showed you. Let's just walk through these briefly. In the suppressed phase, you have suppressed cloudiness, calm winds, large surface solar heating to the upper ocean. Whereas in the cloudy phase, you have the opposite. Enhanced cloudiness, reduced solar radiation. This will result in surface warming in the suppressed phase and surface cooling over the ocean. And so you'll see SST signature like this. Roughly in quadrature with the MJO convection, you'll see warm SSTs to the east, cooler anomalies to the west. So just to summarize this, the suppressed phase is associated with clear conditions, calm winds, warm warming SSTs. The active phase is associated with cloudy and rainy conditions. Stronger winds and cooling or cold SSTs. Okay. So number two, the ocean response to MJO forcing will focus on stratification. And thanks to Xu Yi for giving you a preview of this talk. So now we're looking at the ocean response to MJO forcing. So the ocean has received flexes of momentum, rainfall, solar heating. We want to understand in this slide, the stratification. So if you consider the depth of the isothermal layer on the left side of this diagram, we have a very deep isothermal layer. Precipitation can then reduce, it can reduce the density of the upper layer of the ocean. So this is a form of buoyancy flux. So your surface salinity will decrease and this is the fresh lens that gives rise to a more stably stratified upper ocean. This resists mixing and it can help trap future fluxes of either momentum or heat into a thinner layer of the upper ocean. And so the stratification effectively reduces the depth of the ocean mix layer. And so throughout this talk, I'll have in blue some questions that I'll just let you read. We won't go back and answer these. There won't be a quiz. These are just things for you to think about. So if we move forward, how do these processes of stratification affect the SST tendency? So the SST tendency equation can be roughly broken down in just two terms. One is the SST tendency due to net surface heating across the air sea interface. And the other is processes that are inherent to the ocean. So essentially the temperature divergence driven by ocean currents here. So this is a surface heating term. This is an ocean processes term. Both of these control the SST tendency. The point I wanna make here is that any of these processes that show the ocean mix layer or make it thinner, it reduces H. And by doing that, reducing H means that you will get a stronger SST response to a given surface heating. If your mix layer is shallow versus if it's deep. So this is just repeating that in words. And then the barrier layer again is shown here. Barrier layers as you mentioned are interesting because they limit vertical mixing of surface properties. So they limit mixing both at the top and the bottom of their boundaries. So a barrier layer is a really great description of what these layers do. Okay, so now we're going to talk about some of the different time scales of the ocean response to MJO forcing. So I just repeated our little schematic illustration here. And just to remind you that MJO is large relative to ocean basins. So MJO forcing and ocean responses differ according to where you are in the MJO phase. And I'm going to sort of list a series of different time scales in this gray box here. But I want you to remember that these feedbacks do not exist in isolation. Cross scale communication happens all the time. Okay, so the first thing we're going to talk about are ocean responses on diurnal time scales. As Shunyi mentioned, it's not okay to think of the ocean response to MJO forcing as being much slower. The ocean responds on very quick time scales as well. So what we're showing here is the SSD time series over about 15 days. And what you can see is that every day you see a strong diurnal peak of SST. The boxes and all of these diagrams I'm going to be showing are sort of where the forcing occurs in the MJO spatial structure. Okay, so what you see here are two things. You see the strong diurnal cycle of SST. You also see that the SST increases gradually with time. So the calm conditions in this region plus the enhanced solar heating lead to upper ocean stabilization. And this is why you get these strong SST tendencies. Now, the reason you get this gradual increase of time is that the daytime mix layer showing is greater than the mix layer deepening that happens at night, that is driven by long wave cooling at the surface. And so what we understand is that the diurnal SST cycle rectifies onto the intracesional SST tendency, which you see here. Okay, another form of ocean adjustment can happen on synoptic time scales. So this is a figure from Jim Mohm's 2014 paper from the Indian Ocean. And what you're seeing is upper ocean zonal current as a function of time and depth. And so the red shading is an eastward current or a westerly current. And so this is showing a few pulses of strong westerly winds that last one or two days each. But the point is on the equator, these surface jets can persist for many more days beyond their surface forcing. So that's a synoptic time scale response in the ocean to a momentum flux by the MJO. Now at intracesional time scales, what I'm showing here is a figure from one of Kyla Drushka's paper. This is anomalous salinity as a function of time on the x-axis and depth on the y-axis from minus 15 to 15 days about maximum rainfall. So what you can see in the blue shading is that as MJO rainfall increases as the disturbance moves from west to east, you get more and more of a fresh pool that forms on the upper ocean. As the westerly wind burst starts to form, this gets mixed downward. And this forcing, of course, happens in the part of the MJO where it's raining. All right, seasonal time scales. This gets a little bit to the question of what are some of the Indian Ocean dynamics and how does it respond to MJO forcing? So this is a figure from a paper by Adam Ridbeck and Tommy Jensen in 2017. And what you're looking at are sea surface height anomalies roughly 75 days after a westerly wind burst occurred roughly right here. And so what happens here? So the forcing, again, is the westerly winds to the west of MJO convection. These westerly winds force an eastward propagating downwelling oceanic Kelvin wave that would propagate roughly along the equator. When this wave reaches Sumatra, it is then reflected as a westward propagating pair of equatorial raspy waves in the ocean. These are oceanic waves. These waves take roughly 70 to 90 days to transit the Indian Ocean. So this is an example of how the ocean responds to MJO forcing, but on a seasonal time scale. And then finally, as Shiggy mentioned, there are inter-annual feedback. So this is a schematic illustration of NGO driven westerly wind burst acting on a stratified upper ocean in the West Pacific. This is a figure by Hoshuan Wei that just came out. So essentially your westerly winds can affect the warm surface waters eastward in the presence of a thick strong barrier layer. This will promote, this will reduce SST cooling at the base of the mixed layer and help sustain this warm water advection to the east. Okay, so just to put all of these responses together, we can see what I want to really emphasize is that some of these processes result in ocean stabilization, and some of them destabilize the ocean. So strong surface heating and rainfall are processes that stabilize the upper ocean. These mostly occur in the vicinity of NGO precipitation into its east. Destabilizing processes are largely driven by wind forcing and destabilized. And you can see where the convection is strongest, you have competing effects, right? Rainfall stabilizes the upper ocean, wind speeds want to destabilize it. Okay, so now let's move to some of the ocean feedbacks to the NGO. We will repeat this, but then trying to understand how all these processes that are shown in this panel here feedback to the atmosphere. And so I'm just, the blue little rectangles here are just sort of your markers for which process we're discussing here. So diurnal feedbacks. This figure is from a paper by James Rupert and Dick Johnson in 2015 based on dynamo observations. Here time increases from right to left. And so when you have these strong diurnal SST signatures this makes the ocean surface behave almost in a land-like way. So you have strong diurnal surface temperatures, it can force a more land-like convection diurnal cycle. Typically over the oceans, you find maximum convection and precipitation shortly after midnight. And so you can shift the phase of the diurnal cycle to more like late afternoon when you have these conditions. Synoptic time scales. It's a little bit hard to find a good figure that illustrates this, but the main point is if you have the sustained strong surface jet in the ocean near the equator, this will induce shear-driven mixing at the base of the mix layer. And when this mixing is a sink for mixed layer heat content. And so the effect is you will actually reduce the ability of subsequent surface warming or surface heating to warm the upper ocean. Jim Moehm has a 2017 paper that suggests that this feedback can actually lead to a weaker convective signal for the next MJO cycle. So he argues that the synoptic jets affect the upper ocean heat content felt by the next MJO cycle and can regulate the amplitude of the MJO. Okay, intracesional. You'll notice that I've switched from surface salinity and it's mixing with depth to intracesional sea surface temperatures. This has been really a source of puzzlement. How does this feedback really affect the MJO? I'll talk about that a little bit more later. But what I'm showing here are a lag regression diagrams of MJO convection in the eastern Indian Ocean across the warm pool at lags from minus 25 to plus 25 days. And these are model experiments. So in the left, we have an uncoupled experiment. And you can see that there is eastward propagation but it's rather jumpy, which, and it does not look entirely like what we expect from observations. When you run the same model and allow the ocean coupled feedbacks to occur, you see much stronger eastward propagation. And this effect of the intracesional SST or the intracesional ocean feedbacks to the MJO has been demonstrated in many, many different models, understanding exactly why possibly still remains a question. Okay, going a little more quickly here, let's talk about the equatorial Rossby waves in the ocean, how they can affect the MJO. This is a figure from a paper by Weber et al. As these equatorial Rossby waves slowly propagate westward across the Indian Ocean, they are associated with the warm SST anomaly as large as 0.25 degrees SST, which can actually be enough to destabilize the boundary layer. It also sharpens SST gradients in the western Indian Ocean, which is also favorable to initiating convection. So these equatorial Rossby waves have been associated with the initiation of what's known as a primary MJO event. That would be an MJO event that does not have any well-defined atmospheric circulation precursor associated with it that might cause the MJO to form just by purely large-scale dynamical processes. Okay, and then finally, how do ENSO cycles affect the MJO? This is from a paper by Nick Klingman and myself in 2020. This is just simply the the lag regression of rainfall with itself, or I'm sorry, NOAALR with itself, for El Niño versus La Niña conditions. And the point I want to make is really here in the western Pacific Ocean, MJO convection tends to propagate much farther east during El Niño conditions where you have the eastward expanded warm pool. And the idea is that this can simply help sustain deep convection farther into the Pacific. Although these patterns may also affect MJO propagation speed, as shown by a recent study by Ben Wong and co-authors, and presumably this happens through frictional wave-sys type feedbacks. Okay, so I'm doing okay on time here. So I just want to go back to this and say one of the challenges in understanding how ocean-coupled feedbacks affect the MJO is that when you do model experiments and you make a change that improves the MJO, it's very, you improve the entirety of the MJO. You improve its representation of convection. You improve the large-scale circulation features. You are very likely to improve all of these processes that we've shown here, plus others that are not shown here. So it's really very hard to understand what is the actual feedback path of the ocean to the MJO. And so I recently led a study in this 2019 paper where we examined four global models and we ran these as fully-coupled models, ocean and atmosphere communicate with each other. And then we took the monthly mean SSTs from the coupled simulation and prescribed them to an atmosphere-only simulation. And it took us really a long time to realize that the mean state is the place to look to understand this. So what I'm showing here is the difference. This is the ensemble mean of all four models of column water vapor for coupled minus uncoupled. And what you can see is that in the coupled models, you had slightly more moist conditions or higher column water vapor near the equator and drier conditions off the equator. So the net effect is you had a sharper, more peaked column water vapor on the equator than off. And this is really important for MJO propagation. And to illustrate that, if you look at this panel, what I'm showing is mean state SST in shading and then total column water vapor in contours. And so you can see that the highest water vapor is found in the Western Pacific, slightly reduced in the Indian Ocean. But also you can see that there is a zonally oriented band of maximum column water vapor along the equator. When you add MJO heating anomaly shown with this green dot, you induce low level circulation anomaly such as this. You have cyclonic circulations to the West, but you also induce larger scale anti-cyclonic circulations to the East. These circulation anomalies act on the mean state moisture to a beck moisture from high water vapor content water vapor content regions to low as shown here. This advection happens both through zonal wind anomalies acting on the zonal mean gradient and meridional wind anomalies acting on the mean meridional moisture gradient. And as it turns out, this is really the process that many studies have shown is a key to MJO propagation. It's the meridional advection of mean state moisture by MJO wind anomalies. And so this study by showing that the ocean coupled feedbacks yield a sharper mean state meridional moisture gradient means that ocean coupling is actually consistent with these weak temperature gradient concepts or weak temperature gradient moisture mode ideas. So essentially by sharpening the mean state moisture gradient the same wind forcing will result in stronger moistening to the East of MJO convection. And this is important because advection is the dominant process for MSC tendencies on MJO time scales. The latent heat fluxes play a far secondary role in this process. And even for MJO maintenance when averaged over the entire MJO life cycle surface flexes actually tend to offset or take away from column MSC. Okay, so just quickly, where do the ocean feedbacks fit into these MJO theories? As I just summarized here, this study suggests that the mean state changes with coupling are consistent with weak temperature gradient moisture mode theory by arguments consistent with the meridional advection of mean state moisture. But it's still very hard to rule out these feedbacks via frictional wave sys. If you have a warm SST anomaly shown here, this is right where you would expect the warm SSTs to enhance low level frictional convergence as associated with the Kelvin wave part of the MJO. But in this study, we found that this feedback was only enhanced in two of the four models and it was not the dominant change that led to improved MJO propagation. It was still the meridional advection that increased more with coupled models. And we saw that in all four of the models analyzed. So I think this study is interesting because it gives a new way to think about the role of ocean coupling on MJO. But I'm personally not ready to say this has closed the book on the subject. I think a larger study with more models would be need to really improve our understanding. Okay, so just in summary, the atmosphere feels the ocean through, feels the ocean SST through its effect on surface flexes. The MJO forces the ocean through flexes of momentum, fresh water and heat. The forcing varies as a function of MJO life cycle or phase. The suppressed phase tends to lead to ocean stabilization, the active phase to ocean mixing and destabilization. These processes occur across scales and recent studies suggest that the coupled feedbacks to the mean state moisture may be one critical pathway or critical feedback to the MJO. So my very final slide, this is a picture of mean state SST drift for a collection of coupled forecast models that are part of the international S2S prediction project. So over 30 days, you can see some models exhibit SST cooling, some SST warming. And so trying to diagnose how the ocean feedbacks might affect the MJO in these forecast models and its prediction can be complicated. And so an interesting question might be, how can we diagnose the ocean feedbacks to MJO prediction in these models? So with that, I will stop and take your questions. Thank you. Thanks very much, Charlotte. That was a very comprehensive talk. I learned many new things, many new aspects.