 Now we have Sankar Arumugam from North Carolina State, and North Carolina State University. Segway in terms of what I want to say. So I think this is the slide that Mike showed in terms of the trends in national water use. So talking to Antar, he mentioned that why don't you talk about the reason paper on the water use, irrigation water use, which is kind of under review. So I thought he's either an associate editor or a reviewer of the paper. So then I thought that I should do a good job in making the case for why water use is important. And one discussion that we had from yesterday as well as today is the importance of looking at the human side and how water use and lack of data on the water use. So what you are seeing here is how the water use has changed over the years from 50 to 2015. This is from the USGS report. And you can see that thermoelectric is the biggest chunk. But of course, it's not significant consumptive use. Most of it returns back. The next to that is the irrigation. And the irrigation is anywhere around, depending on the year, 45% to 50% age. And so if you take out the thermoelectric use out, I think it comes to around 66% age. Public supply is next, which is around anywhere between 12% to 17% age. And so this is over a period of 1952, 2015. And this is at the national level. And USGS puts out this data at the county level. And so when I mention about USGS data, and sometimes there is a lot of gas in the sense that, oh, the data quality is bad. This is one report in which it really digs deep. So if we just go and download the once in five year survey, then this is what we look at. But this report is pretty detailed and very nice in terms of looking at how each state water science centers, they collect the data, what methods they use. And they also provide guidance. I mean, survey is a voluntary effort from the WSE. And what are the sources of uncertainties? And it's a systematic review. And pretty honest and also has a lot of useful recommendations on how we can reduce the uncertainty in this data. So I'm going to now stick to only irrigation withdrawal. So I think some of the key sources of uncertainties, what they report is the acreage irrigated. And then the next one is the application efficiency. That is, what is the efficiency in application? And also heterogeneity in irrigation methods. So they compare in this report very nicely looking at USGS water science centers reports, as well as reports from USDA, as well as FRIIs, Farm Irrigation Survey. And so these are all collected at different years. And at the country level, the acreage irrigated is anywhere within 4% a year. And this is over a period of 2000 to 2010. And the water withdrawal, on the other hand, which is because if the acreage irrigated is really different, it's going to throw into uncertainty in estimating the actual withdrawal data. So they do a systematic study on that one by using a different method as opposed to what water science centers they use for collecting. And so based on that, like, for example, Utah, Arizona, they are within 2% a year. And you can look at the bar chart in the bottom one. And the main differences are primarily for states like Texas, California, and Florida. And I think they also make a point that one of the reasons is these are large states. That is a significant heterogeneity in irrigation methods. And also the other point what they are mentioning in the report is very relevant in the sense of that these states, they have pretty large agricultural surveys. And they do collect extensive data. And they do make valuable information out of it. So the independent assessment method even could be wrong. I'm not going into the details of the method. So I think, and also there are guidance in terms of rating. If you're going to use the state data, whether it is good or bad, you can use this report nicely to assess. So there is quite a bit valuable information that exists beyond just that five-year water use. So we have to dig a little bit deeper. I also want to make one point that is on the slide. There is no data, to my knowledge, that exists publicly, either with USDA or with USGS, on area irrigated under surface water and ground water separately across the entire country. So you have it's lumped, area irrigated for flood irrigation, sprinkler irrigation, micro irrigation, but you don't have anything available. So this is the paper under review. So you have two plots here. One is the surface water irrigation withdrawal on the top. And the second one is the ground water irrigation withdrawal on the bottom. And you see the numbers on it, which primarily shows whether it is increasing over the five-year. That is all the way from 1985 to 2015. And these are all in 1,000 meter cube per day. So basically divided by 1,000, and you are looking at million meter cube per day in terms of withdrawal. And you can see that the surface water withdrawal in general has been increasing. But on the other hand, ground water irrigation withdrawals has been, surface water withdrawal in the West has been decreasing. But ground water irrigation withdrawals over the country, it is increasing. It's not just whether it is blue, it is decreased. But also you want to have zero on your state, which means the temporal aspect of how it has been changing over the years. And we show that, as I mentioned earlier, the acreage irrigated and also the heterogeneity in the irrigation methods, they are much more important. In fact, in the WRR paper, we show that the climate comes later because at the climate division level, it doesn't really matter because most of the irrigation withdrawals, they heavily depend on other aspects, like surface water storage as well as ground water storage. So given that, I think there is plenty of useful information that one can draw from on the management aspect of the story. Just to switch gears, this is a project that we have started in China. And in the case of China, they have water use data primarily available every year collected. And the North China Plain is one place where the withdrawal is pretty significant. And in fact, most of the river, they don't see any water primarily because of the pumping. I mean, Beijing is one. You can see that this is anywhere around 70 feet deep. And some other provinces around Beijing, like Tianjin and all, is much more than that. And significant extraction of ground water. I mean, I was fortunate to get the data here primarily because I had the faculty from China Institute of Water Resources and Hydropower. So I think there's some other key points that I want to mention here. The bullets are in there just to do these are all collected. The irrigation efficiency has significantly over the time, primarily thanks to drip irrigation and sprinkler irrigation. And what is the space that we want to look at? And that's the question. I mean, if you look at, I mean, in our WRR paper, too, we show that at the county level, there is nothing there we can model. But at the climate division and beyond, we can talk something meaning. So this is a nice article in New York Times on water is broken, data can fix it, and it kind of complains about water use data. But it says glowing things about the energy and the power use data. But if you really go and look into what's how the power data is estimated at the monthly level for each day, it's some survey, and then autoregressive model. So I think when we look into the social and economic aspects or the human side of the aspect, some level of surrogates and some level of approximation is needed so that we capture the key variables in terms of mapping up what the management is. So this is, I mean, we have discussed quite a bit about the lack of data on groundwater, water quality, and the water use. And I also want to mention one more thing. Most of the dams and other things, surface water, this is a workshop on groundwater. But most of the, for example, details on, like, for example, inflow releases and other things from the reservoirs is not available. And if you look at, for example, interbasin transfer, most of them are like anywhere in the 80s. And I think the key point here is these data are going to control how we are going to appropriate on the surface water, how much is left is going to basically determine how much groundwater we are going to withdraw. So I think you cannot look at these two as two separate entities, issues such as water law, water appropriation, they are heavily interlinked and it should be looked together rather than in a separate way. So just to wrap up here my thoughts, I mean, data issues and uncertainties exist in all aspects. I mean, including geophysical data, like, for example, precipitation in early regions, is a big challenge and the same thing exists with model two, process representation and coupling. So what we need to look at is, my last point is, how the data, additional data of what we collect, how it will translate to decision making. And also it's important maybe identifying the right spatiotemporal scale in decision making, the decision making is important. So it starts with maybe conflict resolution or water sharing. Maybe you don't need like a how early model at a fine grain to run the model. Maybe annual is good enough or if good data is available, monthly is good enough, then goes the information on monthly if it is like a actual allocation and the extreme conditions like forecasting, remote sensing is extremely useful and like anywhere between how early to weekly, site specific. So I think the question is, what is the decision, what is the problem and how do we look at it and what is the spatiotemporal scale of the problem and map that one to the data rather than the other way around. I have a model that runs at this time scale and spatial scale and I want to fix it to the problem. It doesn't work that way in a management aspect. A little bit more.