Let's get recorded. And thank you everyone for coming in. Today will be our seminar for the spring 2020 to SMP seminar series. And our speaker will be Dr. John Dewey. And so Dr. Dean received his master and a PhD from U-Dub back to 990s. And after that has been working at noaa GFDL lab for years. And currently he is the biogeochemistry, atmospheric chemistry, and ecosystem dimension leader. So case Team, Lisa, effort of the earth system model development or adapting GFDL model for coupling carbon climate studies. Participating in the fifth carbon model into comparison projects, CMU's 5, as well as the recent climate immodest and 4, and couple of carbon chemistry climate, ESM for mottos parties have in the current seem it's six project. And John has more than 30 years of experience developing instruments, collecting field observations, and performing analysis and modeling studies. And also key actually joined our teaching and mentoring at AUD and gay men toward one of our, our undergraduate last summer. And I really enjoyed working with him. And today we're going to hear the most recent updates on the high resolution CSM model. Jump the streets. Yours. Welcome. Great. Thanks very much. So I'm going to talk about climate change impacts on marine ecosystems and biogeochemistry in earth system models. And when I, when I agreed to give the talk and expected to focus on the most recent seem at six efforts. What I what I came to realize was that there was a lot of earlier work that I felt provides a fundamental context to interpret these, these, these most recent studies. So I'm actually going to spend a lot of time giving, giving this background on the evolution of the field over the last 15 years. And hopefully, hopefully it will be entertaining. And I welcome questions at anytime. So so I'm going to start with some, some work in the early days of, of our efforts to look at the mechanisms involved in climate, ecosystem interactions and detection and attribution of change. This the same five, the coupled model in comparison project. Phase 5 of where coupled carbon climate models were placed in, in the, the MIT for the first time post steam at five investigations into reversibility, ecosystem representation, large ensembles, high resolution prototypes. And then our recent contributions to seem at six, including improvements from the previous generation. And then the, the, the application of those, those models for multi-model comparison under Centennial scale change. So one of the so why I was I was hired by GFDL to, to bring ocean carbon into the climate model. Because it was becoming ever more clear that the coupled carbon climate interactions were critical to the long-term evolution of the climate system. But to do so, we had to bring in ocean biogeochemistry and an ecosystems. And because noaa has a mandate for stewardship of living marine resources, this became a big focus for us. So, you know, back in 2006, I might bear unfilled and coauthors looked at the tenure at that time record of global chlorophyll and from the satellite. And they, they, they solve some, you know, some trends emerging. So here's, here's the plot. On the left is the seamless map. And we had at the beginning of this record, the amino of the century with a low chlorophyll in 97, followed by a La Nina in 1998. And then after that, there is what they consider to be a long-term decline in chlorophyll. And the question at the time was, was the weather, whether this change was a signal of, of global global climate warming. Or whether it was part of the variability. And so so what we at GFDL did was up over the years we started with these hierarchy of biogeochemical tools, including the tracers of ocean phytoplankton, allometric zooplankton, Topaz scheme that I developed in the early days. And then we've gone through and, and, and adapted and refined and increased in resolution and done a whole bunch of different experiments sense. But the first application was with every analysis for station where we took the n separate analysis from 148 to at that time it was 20076, I believe. And you have Stephanie Hansen was a postdoc with Jorge, working with Jorge, somebody into and myself. And what she she did was was just trying to figure out whether the model could realistically capture the North Atlantic bloom. So on the upper, upper left you have the sunlight timing of the bloom. And in the lower left, the, the model timing which was, which was pretty good but even better. One on the right side, you can see that the model is able to capture the variability in the timing along that, that chord or between meetups around 45 north. So this gave us some confidence that the model was, was representing the combination of physical and ecological factors associated with, with chlorophyll variability. And so we reached out to iPS cell and an end car and ran these, these, this prototype earth system models with at that time the ESR as a2. Future projections. And looked at what scope of variability was at versus climate change signals were at work. So all of these little panels just show you areas of the world and the chlorophyll variability within the models and the corresponding amount of variability and see what's showing that the, the models are capturing about the same scale of, of interannual variability as being seen and seamless SSO. Then in the upper left, shows a few locations of future projections and these models showing long-term chlorophyll declines. And then the lower right. An assessment of the, the time of emergence of the chlorophyll signal in These, these different regions. And what you see, these numbers are very small, but the essence is that we found that the signal of climate change would not be anticipated to emerge from background variability until for different regions ranging between say, 20222090. So it was a much bigger problem to detect chlorophyll response to climate change than was, was anticipated from the initial observational analysis. So we also are in those early days, we just tried to figure out what kind of mechanisms where it work in controlling ecosystem variability. Looking at changes in sea surface temperature, salinity stratification when light at mixed layer depth and nitrogen and, and, and micronutrient phosphate. And just saw that for different regions, there is a complex interplay of, of, of these different factors involved and and, you know, getting getting them all right At at the all locations was going to be a really big challenge. So for one detailed application we did in their early days was, was looking at this, this model for the California current system response. You know, at the time, there is a theory that because land's would warm faster than, than ocean, that this would result in an increase in summertime. A wins along, along the coast. And, and that that there'd be a change in the Ekman, ekman pumping. Well, what we found. Was it instead that the dominant factor was this poleward shift in the in the winds and the upwelling, which both changed where the upwelling happened. But also the, the source and propagation pathway of the nutrients that got supplied to the willingly in this model leading to a stronger productivity. Once that, that higher nutrient concentration water got to the coast. We also, as part of this poleward expansion of the of the subtropical area, had did a study that projected expansion of the subtropical biome and contraction of the subpolar biome with where in the upper left you see the areas of the temperate, subtropical, and equatorial up-welling by ions as defined by this, this threshold of 1, 35 grams carbon per meter squared. And then within this little, the little black box on the left, you see the surface nitrate part and Productivity, total phytoplankton density, and proportion of large phytoplankton all decreasing over time under this, this future scenario. And then most curiously or are interestingly, the lower right-hand box, you see a new biome that that Jeff Bolivian identified of, you know, 30 degree water and in the SST and that that 30 degree water increasing in time from something that wasn't, didn't exist until this century and was anticipated to grow substantially over the course of the century under warming. So one of the, one of the big questions in, in, in the ecosystem responses to climate change has to do with changing hypoxia and the, the observational estimates of strong declines in coastal pelagic hypoxia. And so what you, one of these set well, we've done several studies looking into how this works. And in that and while the solubility effect and the stratification effects are are well-constrained. I'd say, hi, You see the volume of, of this red line, this volume of high threshold hypoxia increases substantially and robustly. Under climate warming. The volume of, of low extreme hypoxia in the blue line is, is much more difficult to constrain. And we'll see that's a general theme that goes on through the various generations of models. And on the right side, there's this panel shows oxygen change. And you see this, this really red area off the coast of Chile, where in this particular model there's a, a wintertime salinity based convection system that begins as the evaporation leads to two, this will side of convection under, under this warming scenario. And this the, that which has been observed in it, but not to the degree that the model represented. So it just, this is sort of a foreshadowing of how difficult it is to get all of the terms in, in the, in the mechanism, the right level and the right locations to be confident about some of these changes. Alright, so moving on to seem at five. So in C15, the, the reference frame for projections changed from the these ESRD as scenarios, special report on mission scenarios that were that were used in the early days of Climate Assessment 2, these Representative Concentration Pathways which defined specific radiative forcing at 2100. So RCP 8.5 corresponding to an imbalance of, of, of 8.5 watts per meter squared, 2100. As a sort of the business as usual type of scenario versus a strong mitigation scenario of RCP 2.6, which peaks at 23 watts per meter squared and before 2100 and then declines to 2.6 by the end of the century. So, so this GFDL was one of GFDL had two contributions to C15, ESM to G and U. Um, and these were described in this BOP at all paper, where the sea surface temperature change and the pH change are a very robustly characterized in, in, in, in unidirectional where the stippling indicates agreements among the models and the sign. So under the strongest warming scenario, we see global sea level, sea sea surface temperature warming of approximately 30 degrees. And then global sea surface pH changes of about 0.1 to present day and then up to minus 0.4 by the end of the century. So moving on to productivity and oxygen changes, you see that the, the situation gets a lot less confident between the models about where, where productivity is declining, where it might be increasing, and the oxygen change. Showing that the, the, these, these areas where you can see the solubility is driving change in the mid-latitudes, is, is, is unidirectional lay down. But within the tropics where the oxygen is, is lowest, that the models being much, much less confident. Where the, this is shown in the lower right panel. A high degree of of uncertainty between the change in volume of waters with oxygen levels less than 50 millimoles per meter, meter cubed. So one of the, the, the real fascinating outcomes of seeing that five was our ESM 2M model was identified as having a, a a very well where for good characterization of sub-tropical mode waters. And, and lower Svante noted that in this model with characterization of mode waters, that the pH change was much higher in the subsurface then and then in the surface. And that was through the through the combination of changes in circulation and changes in the biological the natural acidification of remineralization of CO2 in the subsurface. And Marion Galen further suggested that while the, the, the surface changes in the subsurface mode water changes were the strongest. The most disrupt, disruptive might be the ocean acidification impact. In deeper waters, where the presence of Indians and seamounts, which are strongly biologically diverse, are being exposed to the most recently formed waters in the North Atlantic with high pH signal. Where this fraction of deep hotspots in sea mounts and canyons, that's, that's impacted with a pH change of minus 0.2 increases from only a couple of percent in RCP 2.6 up to five to 30 percent under the RCP 8.5. So post seem it five. We conducted a, a set of ramp up and ramp down reversibility experiments to look to see what the future ocean would look like after having had a warming and cooling signals. So what we found was that there were a legacy effects of, of warming even, even after the CO2 went back to normal and went back to present day. So The, the, the, while, you see that these, this red line shows a decline in Surface line diatoms with, with increasing CO2 and then the blue line and the, the, the ink that increase in diatoms returning to a lower CO2 value. What we, what we found was larger. Diatoms recover larger non diatoms. The population that doesn't require silica was. Over shooting strongly to, to recover a much higher population than they had available. Before the war, before the CO2 had increased and the warming had occurred. So along that thread, we also look to expand our ecosystem representation to include an explicit zooplankton community of small, medium, and large zooplankton. And this was done in the context of GFDL, Carbon ocean biogeochemistry, and lower trophic Cr, cobalt model that took advantage of the meso zooplankton observations and CocoaPod database along with many other observation of ecosystem parameters to constrain upper trophic levels, along with the biogeochemistry. And what we found in stock et al 2014 was when we, when we switch to this representation of, of, of the upper, upper levels of the base of the food web. That we saw, a strong amplification of the ocean productivity changes compared to the primary productivity changes alone. Further, when these models are what are used for as inputs into fisheries fit fish and fisheries models. We found a further amplification of, of change in, in fish biomass through the, through the ecosystems with a, with a large degree. Here you see the different EIP, the iPS cell and the GFDL models, the upper, upper and lower panels respectively. And they're being forced by a range of fisheries models that give changes ranging from say, 10 percent reductions down to 230 percent reductions in and in the fish biomass. So we've also expanded in, in conducting a large ensemble projections with the GFDL ESM 2M. Where the idea here is to expand on this, this question of, of the time of emergence of biogeochemical parameters. Where you look at the Venn diagram of a in the upper left here of the the the ability to detect and observations among the various measurement errors, natural variability and gap-filling versus in the simulation. The ER for system models as a function of its natural variability. The modal response strength and the scenario forcing, and what Sarah sugar found was a real range in the ease of detecting a signals as they emerge. So the lower panel shows fraction of ocean area that emerges as as detectable. For pH and PCO2. That is, it's very detectable. That first line it goes up to 100% by present day. And then followed by saturation state and the normalized alkalinity. And then follows a projection, a progression of more and more difficult metrics to achieve. So that this, the clerk chlorophyll is in the cyan where even by the end of the century, only 75 percent of the ocean area is detectable. And you know, some, some of these physical variables are, are, are also similarly difficult to detect. Including the mixed layer depth falling at the very, at the very end of the, of the sequence. So here's just a, a regional characterization of, of, of these times of emergency with the color scheme showing the when, when the signals are emerging. And demonstrating that these, these patterns vary quite a bit between, between basins and between variables. So this way we follow this on with a multi-model comparison between the GFDL MPI. Csm and Canadian models, which it illustrated that in the higher climate sensitivity models, CSM and can ESM signals emerged. You'll quite a bit earlier for some of these variables. Whereas with the lower sensitivity MPI and GFDL models, the, the, the signals were emerged later on in the records. And also this. We're able to look at the, by using, I'm making use of a multi-model ensemble. We're able to divide the sources of uncertainty to between the scenario, the, the model itself, and the internal variability in these models. Illustrating that initially internal variability is the biggest source of uncertainty. And then the, than the model uncertainty comes in. And then finally, over long timescales, the, the, the scenario becomes the largest source of uncertainty. And so here and here is just a, a, a map of, of what those times look like. So on the upper left SST, you have early emergence of, of the climate-change signal in the tropics, later emergence at the polls for ARC CO2 flux. We have these, that's very late emergence of the signal at the middle latitudes. And for export, late emergence in the polar. Polar areas. With the sea surface salinity signal emerging much, much later than the sea surface temperature signal. And then, then this is the, this is instead of looking at the time, this is analysis of the strength of the signal at the, at the time of emergence where it shows for sea surface temperature a strong a strong signal, strong positive signal. Whereas the, the, the ERC CO2 flux has some positive and negative signals, with the export showing strong negative signals and in a few, in several key areas. So another, another source of, of interest is that we've, that we've been engaged in has been the concept of perfect predictability experiments to try and figure out how, how much predictability there is in the biogeochemical system following on the questions of seasonal to decadal scale prediction. And what we found is that there are sources of predictability in the biogeochemistry that far exceed that in the, in the sea surface. Temperature alone. As an end as we. So this is a plot on the left of, of several different biomes, which I'll show is kind of a similar pattern of predictability declining over years. The where each winter the productivity peaks and then summertime, the predictability declines but then re-emerges the following winters over, over time. So we've, we've taken advantage of these forms of predictability for, for a set of simulations of actual climate prediction. With the cobalt biogeochemistry, an ecosystem where we see some areas of a strong color for a prediction skills such as the tropical, tropical Pacific and subtropical mid-latitude Atlantic. Where again, you see this is the, these these panels showing different, these different regions in the boxes on the left show forecast skill over forecasts, lead time versus month on the abscissa. So you can see that we have strong winter skill. Then these, these diagonal lines that R0. In some, in some areas. And as well as skill in predicting the long-term very variability of oxygen in the upper right hand panel. And as well as the total fish catch. And and a few key areas where we have a sense that the, that the fish catches is, is a representation of the total ecosystem productivity in a mature fishery. So moving to ocean resolution, here's a comparison of three different resolution models. On the upper, upper left was our cm2 0.1, which was the basis for ESM 2.1 and ESM 2AM and to G, one degree ocean. And then on the lower left, a seem to 0.5, which is a, a quarter degree ocean, and seem to 0.6, which was a tenth degree ocean. And this this comparison illustrated that there were bottom water temperature changes that agreed with observations. And being very much stronger than these coarse resolution models we're able to represent when the, when the model was run as a under, under idealized doubling of CO2 transient simulation. So we've also been exploring the representation of, of ocean ecosystems at this high resolution. Looking at comparing the, the one degree has some 2M with a tenth of a degree as some 2.6. Here's a comparison. Let's see what's in the central panels. And in the lower panels. A, a comparison of temperature, oxygen, and pH variability along the California Current it and this is at a 40 North. Just to illustrate that, that the, the, the verb, the variance associated with N. So variability in this, in this simulation is constrained to less than a degree from, from shore, which highlights the need for, for high resolution to represent what's going on in the coastal systems. And the and the right-hand panel is a comparison of observations and black, the one-degree model and the 10th degree model for if our sea surface temperature and log chlorophyll as a function of distance of shore. Where do you see that? That only the content degree model is able to capture the cold, relatively cold water upwelling near the shore and the chlorophyll response. So what we've, what we've done with that model is then is reconcile the chain that the variability in in, in fisheries catch on this large marine ecosystem scale between observations and, and models through the meso zooplankton metric. Then you see that there's, there's not a great correlation, but definitely def, definitely a significant one over, over these large marine ecosystems. So moving on to CMAP, six improvements from the, from the previous generation. So as, as, as as contexts. The IPCC put out its fourth assessment. The third, sorry, third assessment in 2007, where GFDL contributed with it cm2 model. In 2013. The fourth assessment, GFDL contributed CM, cm three, which is a coupled chemistry climate model and ESM to which we're couple of carbon climate models. And since that time, like I said, we really, there was a proliferation of different science focused modeling efforts that lead to high resolution and different emphasis on chemistry and biogeochemistry. And so for the fourth generation models, we tried to consolidate these efforts into see him for which had a quarter degree ocean, a high resolution. Ocean and ESM for which had coupled chemistry and coupled carbon to contribute to the the IPCC. The same F6 effort and IPCC report that came out just last year. So CMAP six was a vast vastly more comprehensive effort of comparing models than than previous efforts at GFDL contributed to 13 of the different mips, including ones on radiative forcing detection attribution, cloud feedbacks, monsoons. The the the standard scenario. Mips for future predictions and carbon dioxide removal, land use, and aerosols and chemistry and coupled climate and carbon. So lots of different things going on. And to do so, we had this, this model of coupled chemistry climate that represented emissions of, of dust and greenhouse gases. And they're, they're alterations within the climate system. And with a integrated representation of land atmosphere, ocean carbon, and also land atmosphere, ocean dust and iron interactions. That this, like I said, that the CME for effort focused but with sorry, I should say both a atmospheres where one degree, about a 100 kilometers. The CME for a low top, ESM for a high top to represent atmospheric at stratospheric chemistry. And the mom ocean having a quarter degree versus a half a degree in S and 4.1. And in CME for we, we do not use a mesoscale parameterization. So there's, there's only explicit eddies versus an ESM 4.1 because it's coarser resolution, we included a mesoscale, Eddie parameterization. Ocean biogeochemistry with limited to a very simple representation of the carbon and steady-state ecosystem with six tracers called Bling versus the 33 tracers and cobalt with representation full ecosystem. And as I said, the, the dust and CO2 or interact interactive in any awesome for one. Bling includes these DIC, alkalinity, oxygen, phosphate, iron, and dissolved organic phosphorus. Whereas the cobalts includes a much more comprehensive representation of it, the ecosystems. With the dust. We were inspired by by work that Paul's you knew collaborators at GFDL had had done to allow the dust emissions to reflect the variable land conditions such such that drought and fire and other ways of, of diminishing the land vegetation leads to enhanced emissions as a function of, of climate variability. So here's a reflection of the improvement in sea surface temperature through these different generations. Upper-left ESM 2AM, they RMS of 1.2. Down to lower right are ESM for RMS of 0.68. So almost, almost a doubling in our in our fidelity for sea surface temperature. Similarly, for sea surface salinity, upper-left RMS of, of 0.8 down to an RMS of 0.47 parts per thousand salinity. And you can see that the, the patterns are still the same. Where we have low salinity in the South Atlantic and high salinity along in the subtropical Pacific. But, but general improvements. But one of the, one of the things I was most excited about is reflected here and the zonal eastward wind stress in this models. The Southern Ocean is critical for ocean carbon uptake and the maintenance of, of, of nutrient supply into the tropics. And so we really wanted to make sure that, that we were getting. The Southern Ocean, right? So you see the red line is ESM four compared to Mara and ECMWF at represent, representing the both the strength and the latitude of the zonal average eastern wind stress compared to those that the gray shading was the seam at five generation models which were had a tendency to, to be to equatorward. And the tendency to have low wind stress peaks, which very much affects the the up with the upwelling and downwelling and mode water formation into the, the tropical ocean. So here's just a snapshot of the representation, a boundary currents and eddies in the CME for you, awesome. For models, you can see there's on the, on the left-hand panel is the visa, sea surface height and the middle C and four and right panels ESM for showing the the good representation of of a standard deviation of sea surface height and lower, lower panels and the and the quarter to remodel without this harmonization of mixing. And then the representation of, of a mixed layer depth and sea ice in these models. So you can see that's the line that the color is maximum monthly mixed layer depth. And then the black line is comparison of the sea ice extent in observations and models showing that whereas in the DSM to M in the upper left we had sea ice, particularly in the Ross. And what I'll see, I'm not extending far enough north and deep mixed layer depths that that bias is, is removed. And ESM for what? Which leads to some interesting Paulina interactions that I don't have time to go into. And, and then overall representation of of ocean CO2 uptake. That, that agrees with observations and strong improvement in our representation of the atmospheric CO2 and its variability going from really read reducing by a factor of the error in ESM to G versus ESM for more than a factor of two in each case. So we're doing much better for those. And then overall, this is analysis of, of, of improvements in models in general from five to see what six, you can see that in terms of ocean mixed layer, gas flux, chlorophyll, oxygen, nitrate, and silica. That models mostly fall in the green area, meaning performance improved. And only a few in the model degraded. With the models having done very well with nitrate in the past. Which in general, which meant that most of them didn't improve very much. Alright, and then final, the multi-model comparison under Centennial scale change. I'm going to have to go over this pretty quick. So this is, I'm going to focus on this paper from Leicester, quite Karski, which updated all of these results from C15 into that with our CPS, into this shared socioeconomic pathways representation of, of the future emissions. Where again you have these blues strong mitigation and to read the business as usual, warming with strong pH change, a larger windows with the shading being the spread among models for oxygen, nitrate and, and productivity at the global scale. So again, we see that even with these, these better models, we still have some, some strong disagreements on where the, the, the nitrate change will, will be. Among. Among them as well as where the productivity changes will be. And again, so for and then the case for oxygen and the central panel, you see that the change in oxygen due to saturation is, is, is highly stippled, meaning the models all agree. Versus the change. That is, the biological component through ALU has, has areas of, of, of increase as well as decrease in oxygen and much less agreement between the models. And so this is just showing that the models tend to agree on where the stratification changes are going to be most severe. And that there are multiple stressors and involved in and dominating where the ecological changes occur. Whether they're throught the temperature nitrate or, or, and, and where the oxygen changes are severe. And then one of the really powerful ways of thinking about these changes is is through the response to radiative forcing instead of considering each individual scenario. And that is something that is unique to think about the, the radiative forcing ants and sea surface temperature response. And how the various by biogeochemical properties relate to, to warming where you see as a very strong relationship between a change in pH change and somewhat less so for oxygen change and then moving less so for nitrate and even to very little correlation with primary productivity. All right, I'm going to move on since we're running out of time and I want to make sure that we have time for questions. So I'm going to just refer you to the quiet Karski paper for looking at the rest of these. And there's a great deal of analysis in this paper on including the seasonal lot amplitude of SST is it increases over time. And then finally, I wanted to point out that we have also been looking at the representation of of river nitrogen, changes under land use and the coastal nitrogen in inventories as they've changed with, with increasing eutrophication. And in this quarter degree model, the the increase in global coastal primary productivity under, under land use change. And then I'll conclude with just some thoughts that some ecosystem responses to climate variability and change are robust across models. Others are difficult to detect and may only emerge over many decades even to the end of the century. And that many different mechanisms are at work in these responses depending on, on the region that we expect. Strong trophic amplification of biogeochemical responses through both meso zooplankton interactions and higher up or higher trophic level interactions to fish. And that modals are slowly improving and resolution, comprehensive listening fidelity. And with that, I will conclude. Thank you very much for your attention. And thank you. Joan will take questions for now and stood as a question first. Go ahead. Unmute yourself. Sure. Go ahead. Oh, okay. Okay. So earlier in the presentation you're talking about productivity models in more historical models. I think it was like Slide 12 or 13. And still a lot of holes all over East Coast and kind of the mid Atlantic. And then later models you are able to correct for that. So I was wondering what was kind of I mean, I understand obviously the tropics are hard. But what was so hard specifically about the East Coast of the United States. Sorry, I'm being yeah, so like through here you kind of have these that I understand the tropics are hard to model. That makes sense to me. I understand the poles are hard to model that makes sense to me. But these kind of mid Atlantic Zara's, why are they so hard to model? So the, so one of the factors is that's difficult in, in the middle latitudes is the representation of nitrogen versus phosphorus limitation and the role of changes and nitrification over night nitrogen fixation over over this period. So, so you, you have to, it's not enough to get major nutrient limitation, right? You have to get both nitrogen and phosphorus limitation, right? And, and you understand how the the dynamics of nitrogen fixation work, whether the nitrogen fixers are are facultative, meaning they can, they can take up nitrogen if they want or, or, or not. So the, so the, the distributions of nitrogen fixation between the different models are, are, are, are, are extremely disparate. And that's sort of the first answer to why. Why the, why the North Atlantic have a very different response between, between models. Okay. I, I also have a question. So for your input, like that, patterns of diatoms versus not diatoms that input to the model. With that based on empirical observation, genes, like the relative abundances and distribution if those two types of algae or was it just based on theoretical assumptions of which of those algae would predominate under certain conditions. Yeah, so both Topaz and cobalt take the the size structure from an empirical relationships on the ability to, to add larger and larger phytoplankton under eutrophic conditions. So the large phytoplankton have, have a size based restriction on their ability to take up nutrients. But they have a higher ability to, to make chlorophyll because they're larger size. So those allometric constraints are used to determine what fraction of the population is large versus small. So we do not allow the, the, the large phytoplankton to be constrained by silica supply. So instead what we do is we diagnose how much, how much of that large phytoplankton community is composed of diatoms by, by the amount of silica supply. So that's it. It's a diagnostic diatom rather than a prognostic diatom. And that's it. So that's that's why you have separation between how the how the diatoms that are inferred to have responded versus the large phytoplankton community as a whole. And yet, I had a question about the improvements on the Antarctic. Pauline. Yes. Could you just kinda go over that a little bit? So yeah. So what used to happen in the models was because of their C. So the models used to have this very strong sea surface temperature bias. And so this is a summertime bias due to mostly the absence of cloud cover in the Southern Ocean. And so as that and as well as the the inability to represent the ocean boundary layer and, and the, the, and its response to the winds. So you have a, you have a combination of, of too much, too little cloud cover, two week wins, and then two little mixing. And so, so that leads to a warm bias that, that has been improved considerably in this, in this latest generation of models and has allowed the sea ice to extend farther North than in the wintertime. Due to the lack of having to cool off all that water before the ice can form. So as, as a response, the the, the, the waters which used to convect and, and in the What else see, every winner are now kept. And the problem with that is, is that we still, we still have an accumulation of heat in these regions associated with North Atlantic deep water forming too salty and to warm. And 2 and, and to shallow. That gets propagated into the to the Southern Ocean and accumulates until there is a thermal instability that creates these pulling it and these are, these plein air. Are we call them super clean yet, but, you know, they're, they're really a recovery of this, this large release of heat that has accumulated the Southern Ocean and that's, that's, uh, that's one of the main topics of, of continued emphasis in our development is to try and remove this North Atlantic bias that leads us to a stronger than than observed Southern Ocean bias. And thank you John. I recognize that we're two minutes over time and I'm not sure, John, do you have a little bit extra time for me to ask a quick question? Sure. Yeah. I'm fine. Isn't probably everyone wants to know. You'll have shown this great improvement of the performance throughout increasing resolution. So my question is lesser than one ptsd, great resolution and model results are publicly available or steel upon request. So the, the, the 1 tenth simulations are are just too big for us to manage. So we were able to put the quarter degree simulations out with the latest CMAP six effort. And we would look to increasing the resolution in in in future efforts. But yeah, it's actually if this is this has been something that has been a frustration for many years. The, the, the tensegrity simulations were initially conducted in support of this, the so calm project, Southern Ocean carbon climate observation modeling project. And the we had an effort with the University of Arizona to to format a data server to supply that data publically. But that, you know, it's it's, you know, even even after, I think three months I think it was $3 million were used at that it still fits just difficult. It's just a lot of data. Okay. Thank you. I believe in this department, we have a lot of interests are using the earth as a model to look at the change on the shelf and shelf opposing interaction. And thank you for sharing that information. All right. Thank you everyone for coming to the seminar today. And let's again go to room, a path for Java. Great talk. And thank you very much. Thank so much. Okay. Yeah, Let me know if you next time.

SMSP Spring 2022 Colloquia Speaker Series - Dr. John P. Dunne

From Yun Li March 11, 2022

12 plays 0 comments You unliked the media.

Zoom Recording ID: 95225027531 UUID: EF9qq3I7Sa6w9KbMJ7G1Tw== Meeting Time: 2022-03-11 04:17:54pm

Tags: climate modelinggfdlmarine ecosystems

Department Name: SMSP
Date Established: March 11, 2022
Appears In: SMSP Seminars

Comments

Add a comment