 Thank you very much for reconvening. So as I had said, we will have, we have four panelists to present two or three slides on regional freshwater budget under major use scenarios. They'll give about a five minute to seven minute presentation. We can hold questions till the end or if there is a very burning question, it can be asked right away. But either way, we have time after the presentations for a Q&A session, which will be followed by lunch. So without any delay, the first panelist to speak is Upman Udall of Columbia University. Okay, thanks. Okay, great, thanks. Thanks very much Venkat. I decided to not follow the script and talk about budgets because I found a water balance discussions interminable and not conclusive. So what I decided to do is to try to talk about something that I thought might bring things into focus from when people do all their detailed analyses. And so I've titled it a decision analytic perspective, but the idea here is that if I want to make decisions under uncertainty, which is basically going to be the case here, one of the things I have to define is after I do all these wonderful water balance calculations, what's the probability that my supply exceeds demand and over what time period and when into the future or when in the past. So that's sort of the way I'm going to pitch this and there are three points I want to make. There are significant uncertainties with any of these analyses that we do that came out in the two talks that we already heard. But what people don't often talk about are significant biases that go into all of these models and estimates that we make. And some of these become apparent as you start actually looking at the situation, but many of these are latent. So I'll talk a bit about a couple of those. The climate community has made an out of this. They bias correct all their models, which you should not get me started about because that's a big disaster. The second thing I want to talk about with regard to groundwater is the issue is not really how much water you have. The issue is, is it accessible and who is able to access it? And that is a key issue that I think this group needs to think about if they want to get somewhere with that. And I'll give you an illustration of shallow versus ground just as a macro story on that. And finally, the point I want to make is that we focus on hydrology, but it's probably non hydrologic information that may be critical in terms of defining that risk and that goes towards the access and the demand side. And there may be remote ways of harvesting that through natural language processing and so forth that we should be thinking about. So here's the first comment that I had. I want to think about this in the context of water security. And water security can be a dynamic or adaptive context if you allow that it is changing over time and you want to be able to do something with it. You can have an early warning or predictive context. You can have a retrospective context or you can have an asymptotic context where you say, well, we just want to know what is the risk in this particular place. And that could be time dependent or not. So these are things that are well-established in the water resources management literature, but not necessarily in the hydrologic literature which is highly fragmented and I never have understood that. So let me talk a bit about biases and I'm going to jump right into the issue of the biggest water use which is irrigation because I'm mad to establish that so I don't have to dwell on that aspect. Typical way we estimate that is from biophysical equations and which have been calibrated in some places and with perhaps some remote sensing data on which crops are being grown where and what state they are in. Let me narrate to you from five years of working in Northern India and Punjab what I learned about that. These equations have no connection whatsoever to actual irrigation practice. The actual irrigation practice is dictated by an agricultural extension service which says every three days for rice you need to apply two inches of water. It doesn't say don't do it if it doesn't rain or anything like that. The farmer extrapolates that and says I need to water every three days. Electricity is not reliable so I'm going to leave my pump on, okay? So the amount of water that's applied on average that we estimated from field surveys was in excess of 2.2 meters per season for rice whereas from the biophysical equations the amount of irrigation requirement I would have had would have been about 0.6 meters. Okay, so that's the bias of a factor of three. Now think about that in the context of the analysis that people present to you on the global basis and I will leave it at that. Okay, the second thing is the urban use and so forth. So here we use very crude numbers without regard to actually who's getting access and not and the big issue here is the social equity or injustice issue because the populations that are disadvantaged typically will lack access and they are the ones likely to migrate or cause problems. So those are two things that I think I wanted to bring out here. Let me skip through this and then the other point I wanted to make here was with groundwater. This is the thesis of this particular meeting but I would argue that you cannot isolate groundwater from surface water sources. And more importantly, even within groundwater the access dynamics for shallow and deep are very different and the way I would like to illustrate that is by going to this graphic that we did for the US. So the upper part is the trend of depletion in shallow groundwater defined here as aquifers which are being pumped with wells that are less than 30 meters deep. The lower one is deep. You can see deep is where the action is. Most modeling that is being done at the larger scale focuses on the shallow. It's not relevant, sorry. The other aspect that I wanted to show is that there is dramatic sensitivity to climate variability, okay? So that people understand. It rains more, there's more recharge but remember I told you it is a deep that matters. So the real issue here is that we talk about conjunctive use in academia but in reality what happens is that when there's a long persistent drought people pump the deep groundwater because that's what's accessible. There's not the shallow. So you get dramatically higher sensitivity of response to climate from the deep groundwater than for the shallow groundwater because of human action and not because of the natural system behavior. Unfortunately, most of our work in hydrology is focused on natural system behavior without mapping in the human element into it. So that's the question that I wanted to highlight there. The other thing that I wanted to quickly say I know I'm out of time here but I'll just make one more quick point on this and that is for example in the Punjab area in India when I was arguing with the government people and the groundwater hydrologists in India that this is something that is solvable and it's important that these people don't pump this much groundwater. Their reaction was your estimates are all wrong because the excess groundwater that they pump is recharging back into the system. So we should look at the net change. Well, you pump from 60 meters you recharge the upper strata which is only five to 10 meters thick. This is a delayed evaporation system leading to selenization of the soil. This is a much more serious issue than what we are talking about. So I think getting into the details of some of these things in the interaction between the human and the natural system is what I wanted to bring up. Thank you.