 Hi everyone, welcome to today's seminar. Our speaker today is Professor Ines Alcavedo from she is a professor with energy resources engineering. She is going to talk about marginal emissions and damage factors today. And just a quick reminder, our next seminar is in two weeks. Shabai is going to talk about EV charging behavior. Okay, yeah. So it's great to be again in this seminar series. I think I spoke into it in one of the previous quarters, it's not even a class of quarter. And so I'm an associate professor in the department of energy resources engineering, working mostly on sustainable energy systems and looking at transitions both for the power sector for electrified vehicles. And we do a lot of the work focusing on United States related transitions and now more recently also starting to work on India related issues for decarbonization. And I'm honored to also be a senior fellow for the Hoods Institute for the Environment, as well as for the Precort Energy Institute, now co-leading the Beats and Watts Initiative with Ramana Rappupol. And prior to that in my previous life, I was a professor at Carnegie Valley University in the Department of Engineering and Public Policy. And today I'll be talking, this is going to be a little bit of a walk on memory lane because this is what that started around 2010 on marginal emissions and damage factors. And so putting this in terms of the question and the formulation is as follows. So as we think about any sorts of intervention, think about adding solar or wind to the grid or an electric vehicle, the question that we asked at the time is really what are the emissions reductions or emissions added for both greenhouse gases as well as criteria pollutants as you add or remove those sorts of new loads or generation in the grid. And so in order to understand that we'll need to have an understanding of how the rest of the grid changes as we proceed those interventions. So there will be some power plants that will need to run up and others that will need to provide additional power as we plug in an electric vehicle to be charged. As you add solar and wind, maybe some of those other power plants will scale back and run down and not be needed to provide power. How can we track those sorts of changes so that we can provide a more robust estimate of the emissions consequences associated with those implications? In terms of the big picture, energy services are the larger contributor to greenhouse gases, in particular from transportation and electricity. And the sorts of questions this framework has developed to answer a whole suite of questions. And the questions include things like should we add more solar in California or in Pennsylvania? And we'll get back to this one. Are we helping the environment if we choose a battery electric car? Or is it more beneficial to actually select an iris depending on the location where you are? What are the largest environmental climate change and health benefits from increasing the energy in different locations? And are we really helping the environment by adding storage to our electricity grid? And the questions that we explored, so all of those are examples of papers that then came out using this framework. Think about video streaming strategies on Netflix, Sulu and so on. Could we use that as a climate mitigation strategy by shifting the way the loads of these data centers are operating across regions so that we take advantage of regions where you have a lower carbon emission footprint to be routing those services at different points in time? So the overall umbrella is really how do these different interventions affect the emissions and damages? And I'll get back to what I mean by damages from US electric grid. And this is important because if we just pick one number on a CO2 nationwide average emissions factor, we're really masking some key differences that we have regionally. This one is already dated, but we have developed also a tool that's not the one I'll be talking about today that computes the quarterly emissions factor for every single state and nationwide as well as the electricity mix fact close as much as real time as possible even the outputs of the environmental protection agency data. But what you see is that talking just about an average number for the grid across the US will potentially to decisions of electrifying of the avoided emissions thanks to renewables and so on that would mask really the displacement of emissions that would have in different locations. And this is for CO2 but we care also about other types of pollutants. So for power plant operations from fossil fuels in addition to CO2 emissions power plants such as coal still are associated with large amounts of SO2 NOx emissions as well as your PM 2.5 emissions. Those types of pollutants have very strong implications in terms of premature mortality leading to premature deaths due to health problems. Indeed the emissions and resulting exposure to go direct PM 2.5 as well as secondary formation of PM 2.5 that arises from the emissions of SO2 NOx and other pollutants is the major bubble and US health risk in terms of cause of mortality per year. And here the dynamics are very different from that of greenhouse gases. So in Greece greenhouse gases we have a ton of emissions that will be very well mixed across the globe will persist for decades to several centuries depending on the gas and the fraction that will still remain in the atmosphere versus in one of the sinks. And the damages from climate change are difficult to attribute with precision to one location so they will be global and over the long term. But when we think about this type of other pollutants like the criteria pollutants like SO2 it's a very different story. So what this is showing in this plot and this is from a model output from a colleague Nick Mueller just for an illustration is that the damages to health in terms of causing premature mortality and then monetizing that value transforming into dollars will go over that too may range from a thousand dollars per ton of SO2 all the way to fifteen thousand dollars for the same ton of pollutants. Why the difference? Well it will depend on where the ton of pollutant is generated in terms of the stack high geographical location, the dispersion of that pollutant and the reactions in the atmosphere forming secondary PM 2.5. And then it will depend on how the additional ton of emissions contributes to an increase in concentration of PM 2.5 as well as to exposure to think about it like this the coal power plant in the middle of nowhere with the wind blowing in regions that also don't have many people there's not going to be any major effects in terms of premature mortality. The coal power plant just up to the major population center again same ton of emissions but the consequences for exposure will be much larger and that drives the sorts of differences that we're seeing over here in the back. So over here is just a ton of SO2 out of the stack we'll see the change in terms of concentration of average PM so getting more at the issue not of the mass media but actually the concentration in just a little bit. So tying this together to the issue of the time of those emissions the emissions of CO2 and these other criteria pollutants are going to depend on which power plant is operating at the margin and how we dispatch the different power plants. So this figure is mostly for illustration and not representing one real system but on the vertical axis we have the marginal costs of producing electricity by these different plants which is closely associated to their fuel costs. And on the horizontal axis we have the amount of demand for electricity that will have in a certain hour. And so in a simplified way and ignoring transmission constraints and so on the way the plants will be dispatched is from low operation costs to high marginal operation costs and so you'll see that plants such as hydro power and so on will come here towards the left of the plots followed by things like nuclear and then coal and natural gas with the ordering of coal and natural gas depending on what are the coal prices and what are the natural gas prices and actually in their cheap natural gas prices as we've seen the order kind of reverses and as even like to the requirements of several coal power plants. Now that's the way the plants are dispatched but now their implications in terms of emissions across the plants are going to be quite different so you have the here the vertical axis is the emission intensity of CO2 per megawatt hour and for either that's zero and then you may have some coal power plants that have high emissions factors and natural gas power plants that have kind of half the value of those of coal power plants. And so what happens if you're doing a demand side intervention or the charging of an electric vehicle or adding solar or wind to the creek that is producing a certain amount or displacing a certain amount of electricity in this particular point in time where the demand was close to 12 gigawatts the plant emissions that you are effectively displacing or inducing is a natural gas power plant if instead your intervention is occurring at the time where the demand was around gigawatt hours for that hour maybe what you're doing is inducing or removing the emissions of a coal power plant. So how can we track those things? Well it turns out that we're very fortunate that in the United States there are publicly available data sets that allow us to do quite a little bit in that regard. So the modern strategy and this continued over time and being refined and updated was to one model to US Creek so identifying the plants that are operating at the margin and their characteristics. The second step is that we and policymakers probably won't care as much about the emissions consequences. We care about the consequences of the emissions. That entails going from emissions of a pollutant to the increase in concentration to exposure and to understanding what are the the damages in terms of premature mortality or a dollar value associated with those losses so that we can understand which public policies to pursue that are cost effective at reducing those damages. And finally we use this strategy to model how different types of energy interventions change the baseline health environmental and climate change damages. So in light of this and just opening up for a conversation with all of you too let's think about the case of should we add more solar in California or in Pennsylvania to assume that you are a central decision maker and that can allocate subsidies for one region versus another. What would be the reasons to allocate that funding to California or to Pennsylvania? That comes to mind yes. On one hand California is long years so you might be able to have a more generation capacity with like less per solar panels but you might just place more all in Pennsylvania. So right fine the solar resource is likely we'll look at this and probably we've seen this in other seminars but better in California but Pennsylvania has still a substantial share of coal and electricity generation mix so maybe the emissions that you're avoiding this fact the fact that the resource is not so good are better in Pennsylvania we'll see where you're having the same point already. Yeah so one that also depends on the transition mix so for instance you could add more solar in Pennsylvania but coal is very good enough low to following supply so in some ways if you add solar that it's an emergency you can't run up and down the coast so it might actually work. Yeah that's right on and so in the next slide hopefully we'll address your question by exactly tracking how plants are behaving in terms of going up and down from one hour to the next under different levels of demand and actually we have now a project that is starting with also a Jack Schellender who was here as a speaker on understanding how large amounts of renewables may change that dynamics exactly because only some plants are able to go run up and down to accommodate that change so that's a great question. So please keep those coming so that it's not just me talking. So how do we estimate this environmental health and climate change benefits so for every single county in the United States we have information on the damages per ton of pollutants and by stack height for each different pollutants so for SO2 NOx and PM2.5 this comes from an air quality model I can recall if I have a figure here but it ties back to this sort of variation in damages per ton of pollutant emission. With this sort of air quality models those are reduced for air quality models that have across them we actually use three of them nowadays and compare the results for our research given that they have different formulations. AP2 which actually is now in the AP3 version what it does is that it looks at the Gaussian plume and has some very simplified chemistry to understand what the concentration of one ton across the entire continental United States would be so the one of the things that I've shown previously in this seminar series is that this sort of emissions really carry over from the stack of power plants to other states right they don't remain between state boundaries so you may have a large amount of premature mortality for example in the state of New York is actually attributable to emissions from other places Ohio, Pennsylvania and so on. So mapping out that information is kind of the first step. The AP2 and these other reduced form models both include a dispersion of the pollutants, the resulting concentration across all U.S. counties for every single source of additional ton of pollution from a stack of every single power plant and then they couple that with those response functions that represent the physical impacts in terms of premature mortality as well as some environmental impacts but I'll note that the health damages are by far the largest amount of damages when we contrast those in terms of dollar though that comes also with several set of assumptions and then for health impacts what is done is to multiply the premature mortality by a value of a statistical life which at the time of the first analysis that I'm referring to here was 6 million those numbers are being updated. So what's the value of a statistical life? That reflects the way in which all of us face different types of risks. This is estimates that have been done mostly by economists in the labor literature that associate the types of risks with risks in the profession with wages and salaries and how much risk is being tolerated so that yet again comes with a bunch of assumptions because not everyone has the same flexibility of changes in jobs even if the risk level is large and they try to control for that. So from the AP2 model in this case and similar models for more recent analysis we would have the damages per ton of source pollution. Then for CO2 as I mentioned it is well mixed pollutant that will be long lasting and so at the time of these analysis again way back when we used $20 per ton of CO2 to reflect the social cost of carbon. This was at the number at the time suggested by US Interagency Working Group. So what this reflects is the damages to health to ecosystems to everything that may arise due to climate change and folks get that these numbers by running also integrated assessment models that look at these damages over the course of several generations on time horizons. So for 1400 plants we would have the location, fuel type, stack height and hourly emissions from CO2, SO2, NOx and PM2.5 and so this is where the publicly available data comes into play and I do hope that at some point this is available for many other regions across the globe. We choose something that the Environmental Protection Agency here in the US puts together called the SEMS dataset. The SEMS dataset provides hourly emissions and hourly generation for every single power plant in the United States that is fossil fuel based and larger than 25 megawatts and this goes back for many years right so when we started looking at it we're using data from 2009 and data up until 2021 is already available for download. So one of the things that SEMS does not include is the direct emissions of PM2.5. So for that we have to rely on another dataset, the National Emissions Conventory, that puts those estimates together every four or five years and again we now have updated all of that to the most recent one, the most recent NEI but for sake of understanding how we got there I'm showing what we did at the time. Now how can we get at the change in emissions and changes in damages? Well the first thing that we do is that we parse out the data into regions that operate reasonably independently of each other so that we don't so that we can exclude any sort of major imports or exports that will be used to meet demand in other neighboring regions and so we use the e-grade sub regions that are shown in the map and for every one of these sub regions and for each pollutant we look at the hourly damages in dollars per hour as the damages per plant times the amount of emissions that each power plant was providing at that hour. And next we do a regression so we relate the changing damages with the changing generation so to your point this allows for us to understand how the damages and emissions so we use for both will be changing as demand fluctuate in this case generation fluctuates up and down. So let me zoom in this is for ERCOT so for Texas and just the damages associated with SO2 emissions which will lead to an increased concentration of secondary PM2.5 and so again vertical access is the changing in damages from one hour to the next your example access to changing generation and what we see here is that at hours of low demand the emissions of SO2 are increasing far higher than in hours of low demand so what's going on here in areas of low demand there could be that there were some outside coal and even some oil plants that were operating at the margin whereas in the hours of high demands natural gas was at the margin and so there was no damaging use during those portions of the time. In the model you were assuming that solar power was prioritized over coal at that moment? So we don't assume we observe so all of this all the dogs are actually the behavior that was observed on damages on emissions and we have plant type so when I say that this was this profile over here is due to coal and some oil plants is because we're able to backtrack actually which type of plants were going up and down in that specific hour so this is something that you see from the data that rather than making any assumptions so we're not doing any simulations we're doing regressions on the historical data that was observed at the time and that's a good question. I guess because I understand we're wrong because I imagine that most grids use coal as a base on the supply and so the damages would increase if you're adding natural gases from peak versus sort of from meaning lowers it seems like the drain on that time. Yeah I would say that things aren't changing in that regard right so you do have also coal serving as fast ramping in some instances so meaning all of that in addition to all of these layers is is pretty quite price-driven to some extent but and your right so meaning we have here a mix of also single cycle natural gas peakers that actually were built a little bit earlier on in the early 2000s and that have higher emissions intensity for NOx for example and then some singles some combined cycle natural gas for plants more efficient also in something so all of that is kind of embedded on what we see coming out in terms of emissions and damages from from this data. Now so what what is going on well basically we're tracking that in an hour for where the demand was 12 gigawatts the all the demand is leading to this sort of integral or the total carbon emissions that are producing this hour and as we pursue some sort of intervention we're trying to capture this difference so as we introduce a bit of solar wind we're trying to see okay how has the system behaved previously and that's what we're referring to the regressions when demand went down from one hour to the next by this amount and that's what we use as a factor in terms of the change in emissions. So finally I'll zoom in again on what we found at the time for for Texas and for SO2 for illustration what we were able to map and this was what was produced in a paper in PNAS was the marginal damages in dollars per megawatt hour that are associated with total fossil fuel generation in Urkot for different profiles so being able now to match this with total generation or demand to understand how the profiles were different depending on the level at which the system was operating so to make it clear for what the plot that is showing on the left and the one that is on the right and the linkage between the two let's look at SO2 when we see that in hours of low demand the damage per gigawatt hour associated with electricity generation was this same $90 whereas in the hours of low demand we see that the damages were very small and we did the same thing for every single region and for every type of pollutant. Now comes the issue of how can we evaluate what happens under different interventions and so I'll show two cases for movement solar to start with. So we had similarity data from Emerald for a bunch of different wind and solar sites across the United States and this is what it looks like for every single site. So we had capacity factors on an hourly basis that were simulated for a telemetrological year and then also for other specific years this one was for the telemetrological year that would show us okay for a site in this location it is likely that the output will be something like this. Now we can use that information to now match the generation of winds in that particular hour and observe how the system would behave in terms of reducing or increasing generation in that same hour and what would be the damages associated with that so we're basically for this specific hour over here we see that we had suddenly a drop in wind generation and we could then see okay what would be the damages associated with the plants that could increase a little bit their generation to meet a makeup for that change even how the system had behaved in the past when different types of changes occurred and finally we were able to bring all of that together in terms of the damages avoidance per megawatt hour of wind generation in different locations. So let's jump we did that strategy both for solar and then for wind we had similarly outputs on the solar irradiance and did exactly the same type of exercise and so if we revert back to the sort of question that we're asking on should we add solar in Pennsylvania or in California let's see what this looks like so we know that the energy performance is much better in the southwest so in California, Arizona, New Mexico where the capacity factors meaning the same kilowatt of panel will produce much more electricity in those regions than in other regions in fact a solar PV panel in Arizona would produce 45 percent more electricity than in Maine over the first of the year but now let's look at the avoided CO2 that is associated with installing that same solar panel across the country and we're showing that in both kilograms of CO2 avoided as well as dollars in terms of the avoided damages using the social cost of carbon number of $20 and if we look at it in this perspective the largest gains to get climate change from adding solar would come in Kansas, Nebraska and the Dakotas where the solar resources are moderate they're not they are not the best ones across the country but you are displacing all intensive coal power plants and finally if we look at this in the perspective of avoiding premature mortality from air pollution yet a different picture emerges because now this is going to be influenced not only by the emissions of criteria pollutants but also by exposure so the locations of highly densely populated areas will matter and so as a decision maker if you want to mitigate the effects associated with premature mortality from air pollution we'll look at other regions and for example a solar PV in Ohio would be providing 17 times the health and environmental benefits of a solar PV panel in Arizona even as the solar panel produced 30 percent less electricity and this because coal was at the margin in those areas and upwind of major population centers. Let's look at the same sort of figures that for wind here in the case of wind we know that the largest resources are in the Great Plains and through West Texas and when we look at the CO2 emissions that are avoided we see very different features from the solar case here the energy performance and the avoided CO2 will line quite well where in the Midwest both having the very excellent wind resources and having coal power generators being displaced leads to things that are very similar whether we look at energy or CO2 but once again for premature mortality associated with air pollution when you see that other locations would be preferable if your goal as a decision maker is to mitigate damages from air pollution find that the wind turbine in West Virginia this was seven times the health effects of wind and in Oklahoma and 27 times when compared to California. Now we can go one step further and try to understand for wind distillations that had already been built what are the social benefits that accrue from things to having those those plans and so we did that and calculated that the annual benefits from both mitigating CO2 emissions as well as criteria air pollutants were $2.6 billion so we went and actually looked at the actual output from those wind plants and backtracked the damages that would be avoided thanks to that and we can compare that to the subsidies that were provided at the time in terms of a production tax credit federally for wind projects which would amount to 1.6 billion and so here of course some good news right the annual benefits that society as a whole was having thanks to the addition of wind were larger than the cost of the subsidy that was provided but now if we look at things recently things also become interesting because everyone gets the same PTC level per unit of electricity generated so $22 per megawatt hour at the time but the benefits that arise in different states would be tremendously different so in California the subsidy is larger than the damages that it avoids in Pennsylvania the benefits that you get from installing wind for health and for climate change are much higher than the level of the PTC subsidy so my question to you is is this fair or should we retain the design of subsidies and why yes or why not and I don't have an answer I mean again this is really the question as as decision makers you could argue in different ways about the decision the goal of adding wind is not only to address climate change but also in terms of adding and building the infrastructure potentially jobs which we ignored completely analysis and trade-offs between shops creators and jobs that disappear and the reason why the PTC subsidies higher than the damages avoided in California is also California has already spent a lot of effort in terms of the carbonizing and reducing the emissions from its green so should it be weighted less or more so all those questions more for regulation and policy but they are there right and folks could could decide one way or another and then we can set up arguments yeah yeah I was going to ask if that number has been contrasted to the costs associated with sort of transitioning with carbon right because your closing down plans these jobs this repurposing so we're currently doing quite a little bit of of that not as much on the quantification of the jobs in our group but the changes in terms of costs of infrastructure so meaning I can point you to another seminar from two days ago looking at the costs of building new infrastructure across the United States when we want to meet specific the carbonization goal and how that changes both in terms of these infrastructure locations and in that benefits when you explicitly include or exclude the health damages associated with air pollution and in addition to that we've been doing a lot of work still reflecting some of these methods on environmental justice consequences across different types of regions as we adopt different types of energy technologies which is I think a good segue into this this was another paper but building on this and looking at the air pollution damages that would be avoided by a solar PV but over time so we're showing how over time the benefits actually got reduced because the intensity of the grid was lower and so what was being displaced was decreasing at the same time this is yet a similar set of figures just showing for different pollutants how the damages avoided by one to another solar change from 2006 to 2014 again relying on the sun's data not only that I don't think I have a figure for this but we also looked at how these benefits were distributed by county by demographics in terms of how it provides just another example thinking now about okay is this storage green or is storage not a green strategy for us to pursue and so what what do you think yes yeah why is that because then we can get more non-renewable energy from the environment rather than using the natural gas at full okay so if we pair up storage specifically with renewables right to provide firm power it's kind of a non-renewable that that would be emissions reducing more likely um any reasons why this may not happen it's expensive so there is the issue of deployment that those um so yeah part of mining for I don't know okay so there are all those environmental impacts that we're ignoring the associated with uh with mining that's that's definitely true but let's look just at the operation side and and and the grid and once again relying on the sorts of marginal emissions and damage factors so let's think about about this the storage act in 2013 proposed changes to the internal revenue codes so that energy storage would also have an investment credit similar to other renewable and low carbon technologies and in california the senate passed a 2514 directing the california public utility commissions to determine the appropriate amount of pre-energy storage that would be added and indeed by 2013 the california public utility commission mandated the three major investor-owned utilities that they would need to have 1.3 gigawatts of storage by 2020 which I think has been met now will storage potentially increase rather than decrease emissions and the answer is yes depending on how it is operated um storage is used for different types of services from frequency regulation to energy arbitrage and the way energy arbitrage works is that storage will charge when prices are low usually at night and coal will be at the margin in several regions across the country and then these charges and these used during the peak afternoon or even periods when natural gas is at the margin right all these dynamics are also changing as coal is being retired and as we have more natural gas and and renewables so the actual net contribution of storage will depend on the greed where you're at but if you use average emissions factor you simply assume that your storage is associated with zero emissions when you're charging at least trash but the storage technologies in addition to the time of charging also experience losses as they store and recover energy so a little bit more of electricity is needed for that and therefore an additional penalty and emissions so in this work we used again the marginal emissions factors we didn't go all the way through the damages and developed the revenue maximizing linear programming optimization model to simulate um uh players in the model that could use energy arbitrage so they will buy and charge the device when the prices or elements are low and then sell it when the price are high and we need that under both perfect and perfect information or forecast of what the prices would be and again i'll highlight that this is changing because i will read this rapidly changing in terms of the baseline emissions this was using the greed back in 2015 and we saw that in some instances there could be some annual revenue from the storage devices but that virtually everywhere using storage for energy arbitrage would increase the emissions not decrease it would increase NOx emissions with the exception of some regions where it would lead to a decrease and same thing for SO2 emissions to be at least zero or higher so as policy makers are deciding on specific scales of storage to be added in the greed equally important is the definition of how art storage services is used in a way that doesn't produce unintended consequences for the emissions purpose in particular if they are aimed at reducing the emissions so one of the things that my group has done over the years this whole work started back still at Carnegie Mellon and PhD student Kyle Seller Evans and then continued on with the participation of several other PhD students and we put together an online tool that all of you can use and that we tried to keep updating where you can select what sort of factor you want to look at whether it's average or marginal emissions factor whether you want the output to be the results of the regressions of the sorts that I described slightly different than they evolved over time as explained in the documentation or as the basis of we will also this develop a simulation model we dispatch off plans and contrasts what emissions factors we get out of those simulations rather than put them at historical and you can do by hour of day and then by state or agreed region or balancing area and you can look at this you don't just damages or emissions and using three different types of air pollutants models and you can download all of this as csv files for your own analysis and assessment of communications of different types of interventions the goal is really to have this being used understanding the limitations of the different types of tools that that are out there still but so that one can inform better decision making as decisions and policies are designed with all of that said I highlighted some situations where we found counterintuitive results this by no means means that we should stop the transition to low carbon and sustainable energy systems we need a massive infrastructure transition to renewables to storage and to having other types of portfolios of energies that help us get there in the reliable way we also need there's some work that we did earlier on showing an increase in emissions of electrified vehicles in some regions across the country and we're definitely seeing that electrification in countries like India is actually increasing emissions even the very large coal base on the generation so those things need to be happening at the same time for mitigation goals to to be met what we do want to highlight with this source of tools is that we now have data and mechanisms to start doing much more careful decisions and planning to make sure that these transitions are indeed achieving the emissions reductions that that we need the second point and I'll get to your question I won't forget about these steps sometimes we'll have different answers if we look just at climate policy and ignore the air pollution size of things or other externalities and damages that may also change the societal outcome and so focusing on at least those two major ones climate change and air pollution together is important because it may be changing the change your decision making we found that location temporal patterns and behavior are all important in determining the health environmental and climate change outcomes and interventions I do also want to highlight that as we put together these sort of marginal emissions factors MEFs or marginal damage factors there are plenty of limitations on issues too and if you do decide to use these sorts of data sets for your analysis I would love to talk more about that too but in nutshell they are being designed to model small interventions in the grid not large ones where the system quickly changes overnight so for that other type of modeling tools are needed that account for changes in the capacity and in the fleet over the next decade so that's one one important limitation the other one is that this becomes harder and harder to model in terms of changes as we have more renewables given that the generation the hourly generation and emissions from sense are only for fossil fuel based but the environmental protection agency and you we are now also making available hourly generation or even some hourly for other for renewables for example so we can start using that and hopefully Jack showed a little bit on that in the last seminar so with that I'll open on questions for a little bit I think this was the last slide yeah so when you go to that slide we're talking about the energy storage for arbitrage being more carbon intense would it be similar and I kind of mentioned it with like electric vehicles but is it would it be similar along the lines of like electric vehicles versus gas I know gas wasn't included in that because there's not really like regular gasoline sort of similar approach but the numbers would differ given the carbon emissions intensity of gasoline and diesel and the fact that they're they will be emitted at the ground level so that model that separately so this person at ground level will be smaller when compared with the with the stack so we did find that hybrid or electric vehicles were generally better than a conventional gasoline vehicle across the country but the mild hybrids just with regenerative braking and so on outperforms naturally electric vehicles in some locations so just to confirm that that's because the electric vehicles are charging at night which is when you have that natural gas and coal that's right so it was under but though we we changed that assumption in the sensitivity analysis but that's right so we're assuming in the base case convenience charging and that means starting charging the vehicle after the last trip of the day and so once again in terms of data the the the US has the NHTS survey data that has information for different counties and locations on when the last trip of the day for conventional vehicles is and you can use that as a proxy on how people behave for the charging of electric vehicles yeah so for the price of the batteries like i'm wondering if they the garbage is trapped because of the life cycle of the battery as such or is it the emissions from the sources that is delivering the battery so here is just really operation emissions so we are not accounting for the production of the battery so that would still add add on to what we're seeing here this is just the net emissions associated with charging when coal is on and selling when natural gas or renewables are on and that's differential yeah so out of interest wouldn't the wouldn't the reverse actually why because i mean we have a whole bunch of renewables at like morning and then noon that's when you have that one negative marginal pricing scenario and it's best to charge them but at night is when you actually want to have a full battery to supply so that's a really good question that i think that's why i kept on on highlighting things are changing so by the time we were running this we didn't see that i think then we did it more recently around 2017 and still the results were still holding very much so but as you're saying meaning under the vast amount of renewables in particular if you have a lot of wind things could change with solar it's a little bit more tricky because solar also coincides with the peak right on so i think it will depend it will depend on the low profile it will depend on the amount of wind versus solar that you have in the generation and how that influences prices yeah sorry one more question have you guys factored time and size into this analysis so for instance if you say add you know one gigawatt solar onto a cold base system those cold plants are still probably going to run and you'll be like cold is running on the side but it's so we looked into the grid for instance if that's done in like two like the year for instance um so is that in fact it has not been if i understand the question i don't it hasn't been factored then the this analysis since we're just tracking things as they change we demand we are now looking at how are our weak capacity factors um have been changing for fossil fuel generation across the country as we have more renewables in the grid but that's ongoing we we started recently jack and i thinking about this problem there's one question from an ongoing discipline so i'll read it i'll read it for me i'll put it away it's shown in the screen uh i so we to read it for myself too so she's carbon our pollution can see he um become a market factor when isos select to dispatch order versus today um so only we we actually are about to submit a paper on that and uh explaining how the dispatch order and the prices would change when we incorporate those two aspects actually um and in the same um in one of the the recent dispatch papers and i'm happy to point you to that we actually looked at the change in merit order and dispatch when carbon price was in place but not the air pollution um but that meaning i i think that's the question this would be something that would indeed provide the right signal um the issue that arises is that all of these values in terms of the pollution intensity if we want to get all the way to the dollar value come with a lot of assumptions and judgment values so for the pollution if you want to reflect that in some sort of uh dollar per ton of damages or per megawatts embedded in that is the the value of statistical life and other things and for the social cost of carbon the heated debate of what that should reflect so to give you a little bit of history the social cost of carbon uh was developed and the one and under one of the prior administrations like this was under obama um as a series of studies that estimated the social cost of carbon using different interagency um sorry with different integrated assessment models and they come came up with a range of values of years 20 at the time then they increased to a range and to a new value that was around 36 dollars per ton of c2 i think in 2017 dollars or something i may be wrong here and the value was including global damages from climate change and the low discount rates to bring the dollars from future years to present day conditions uh then under the trunk administration things very quickly changed by executive order to instead account for only domestic damages from climate change so only damages that could occur in between us boundaries and a higher discount rate and suddenly the the um and the agencies were not mandated anymore to use the same value for the social cost of carbon across all of them which they were previously and so instead of using 36 dollars or whatnot for the social cost of carbon the values changed to something around three dollars per ton of c2 so that would be much less um policies that would be justified under that case and then um under the biden administration biden in the first start i think this was first day of uh in the office uh changed these things back again with an executive order and as to include global damages and the low discount rate that the value increased again so meaning it's these things can still change quite a little bit both on climate change and the the um air pollution consequences depending on some of the key decisions that are made on how to value some of these consequences so this was a very long answer to your question oh yes um can you go back to the slide where you have like the link for the online tool sure i mean actually i don't know if this one is working right now my suggestion is if you go to my page and scan for them and their tools there will be a link to to the shiny apps tool okay and feel free to shoot me an email if it's not working or if it's buggy any more questions okay yeah thank you very much very