 So, we talked about research frontiers around groundwater recharge and flow and ways we could collect data, especially where in-situ data were sparse or inaccessible. So this actually caused us a little bit of angst. We talked about grace at the large scale as a way of looking at water storage changes, but if there were multiple rivers sort of at a swat scale within that larger scale that we were looking at that we might be able to use swat as a way to distribute that change as a way of downscaling grace. Ed might have to help me here, but we were also talking about some non-event driven river discharge using that as a way of estimating base flow from swat or ice set to potentially. But then we sort of struggled with other options for systems where, especially where things were inaccessible, so we started to move into like vaguely accessible and then we just went like full accessible, but we said if permission was available that we could use some drone-based measurements potentially to look at things like temperature or NDVI to look at vegetation. And all we could get from this is maybe wetter areas. We weren't really sure that this was tied to groundwater recharge, but we were really just trying to find a thing. If the system was actually accessible we said there were maybe a couple of things that we could do. And I don't know if these are research frontiers because I think both of these things already exist, but you know nuclear magnetic resonance is a way of imaging moisture content. So people have been using this for a long time. More and more lately as that technology, as the signal to noise has gotten better and borehole instruments have changed a lot of these measurements, but looking at changes in moisture content is a way of estimating recharge. And if that site was also accessible you could look at changes in some sort of process, either precipitation or actually inducing tracers with some geophysical instruments like electrical resistivity came up over some smaller scales. And that is all we did. We kind of tried to wrap our heads first of all around the scale of the problem and we came to the conclusion that it depends that the frontiers depend upon the size of the problem and so likely we would need multiple technology advancements. We talked about airborne as well as spaceborne, the airborne AEM, the geophysics talk that we heard that was very impressive. The only downside of that is that it's not glowable but it's doable if the political will is in a country if you wanted to fly there. The good thing about it is it gets at the 3D and gets at subsurface property distributions. We also talked about drones as well, large drones. We do a good job of measuring precipitation from space but we need to do a better job of runoff temperature and ET and we thought that large drone-based measurements could potentially help us there. We also thought about the grace mission and you know the challenge of course the grace mission is great for closing the water balance but the scale, spatial scale is an issue so we thought that you know if we had two pairs of grace mission, two grace missions that would help us to knock down the scale as well as the uncertainty. We talked touched a little bit on GPS which is something we hadn't talked too much today. It's a little bit tricky. It needs a ground-based receiver but it can measure surface water changes and so you can track surface water with GPS and then track how that surface water interacts with the interfaces with the aquifer. For some research that's out there, far out there, we talked about isotopes and any sort of water quality remote sensing that's on our wish list here. We talked about the benefit of INSAR for CalVal and then in terms of a research frontier we talked about the fact that all of these measurements are not measuring the same thing so we don't have a good way of integrating measurements. Could this be a framework for example do we need that? Is there a new methodology or we need a new methodology to combine its situ and remote sensing data together? We talked about machine learning and data analysis techniques as a key component to combining these measurements although I think that there was some concern that machine learning can be a little bit tricky. We also talked about the importance of technology that could kind of help to ensure that all people that all citizens are participating in that problem. So we talked about combining geophysics methods and remote sensing for linking surface water and groundwater movements. We talked about also combining hydrologic modeling with remote sensing. I mean a lot of people are doing that already but definitely downskilling. There are downskilling issues such as grace using grace data to its advantage and to bring it down to more useful resolutions. We talked about also combining climate projections for strategic planning for future recharge methods. How we can prepare for future management of groundwater recharge where there might be changes in precipitation and other variables. One important thing that we discussed was there's a lot of new advances in computational methods. It could be in geostatistics, it could be in Craiging AI machine learning that could these methods could be adopted in a smart way so that we can interpret or interpolate the existing data sets better and not just for getting better results but also for guiding where we can where we should focus for getting in situ measurements for calibration validation and also just for getting new data sets and also for reducing uncertainties. I mean reducing uncertainties with the use of these new computational methods. We also talked about just like the previous group we talked about how we can engage citizen science and also the availability of cheaper and newer sensors so that we can get more and more in situ measurements. I think I'll stop there. Then we go to the next question. The question was about identifying when and where aquifers are being recharged in the source of water. So part of this was answered in question one but we talked about some other possibilities for looking at when and where aquifer was being recharged. We have like a bunch of asterisks on this one but like if your aquifer was elastic and you had injections then maybe you could use INSAR to look at locations for recharge. Obviously the Trident U method for looking at groundwater recharge are boreholes but that's pretty limited in terms of where you can do that but the data are continuous in time. We talked about using sort of old-school you know earth science inference when you look at a system and you look at its topography and its geology and maybe its vegetation if the shallow if the water cable shallow you might be able to take a look at a system and figure out where groundwater should potentially be recharged. Grace obviously at the large scale again in terms of sourcing the water I don't think there's like a you know a magic wand we can wave around and figure out the source of the water but we could probably infer source of water depending on what the recharge data looked like so does it look like precipitation or is this happening in a location where there's managed aquifer recharge etc. Otherwise if we really want to get at water source we probably need isotopes or some sort of genetics of microbial communities in the water to parse that. And the last thing we talked about in terms of when and where aquifers are being recharged is just the soft data or interviewing people in systems where data are sparse or otherwise there is an instrumentation. I've talked about the importance of both in situ as well as model data. In situ data are absolutely essential and models in many ways are key to integrate validate data to put constraints on different components. And recharge is very difficult to observe and it's best estimated by models. And I think there was a comment in the group that recharge is actually one of the most difficult things to measure in terms of groundwater. We talked about scale we talked about the temporal scale that's needed. We need to have continuous monitoring which of course comes from in situ measurements continuous in time but discreet in space. It helps us they help us to have the best chance of understanding seasonal changes in recharge. There is also we talked about event based recharge and then using paleo data to look at recharge in past climates to just get a better handle on recharge. Talked about spatial scale in situ data doesn't provide us with high spatial coverage. It isn't spatially consistent and so the question came up what spatial scale do we need to focus on to do monitoring and we didn't have an answer for that. We also talked about the importance of having good precipitation and ET data a better understanding of groundwater and surface water interactions and the need for runoff data for the models. We also kind of agreed on the need for both in situ data and models for these identifying recharge but we kind of wrote down the specifics that included topography, geology, the layers, the land surface information, hydrology and also near surface geophysics. We discussed that the human processes are very important how the land cover has been changing, how the land use change can impact recharge, those information are needed as well. In terms of identifying the sources of the water and the types of water isotope studies and biogeochemical markers those could be useful but we also had a good discussion on the types of recharge and the scales of recharge. Do we mean just the inflow? Do we want to find out the net recharge or can we separate it out from the lateral flow that the basins are having like the Bengal Basin that Holly presented this morning. There's a lot of lateral flow in those regions. How do we really separate that out from the recharge that is happening and also the scales of this recharge depending on the basin or the type of climate or land surface that we are talking about and in combination with the surface water flow of that basin. The third question was about NGA resources that could help make meaningful progress in terms of our understanding of recharge. We actually weren't entirely sure we didn't have the benefit of Tony being in the room to know exactly what those resources were so we ended up asking some questions. So one of the things we asked about was whether NGA has access to satellites or geostationary information that the public doesn't. We assumed that was probably the case. If so it would be helpful you know if some of those data were open to sharing like we could maybe even image where water is and then isn't from some high resolution images to get at recharge. If those data aren't available then maybe there was a way for NGA to think about deriving groundwater characteristics have people that actually do that and then share that with the public. But that was about as far as we got with that because we weren't entirely sure what was out there. We actually had the same question about I think it's the last bullet we need to know what kinds of resources might be available. But we talked about you know the possibly supercomputers and cloud computing resources that universities and other research organizations could access and utilize. We talked about opportunities for partnership or funding to work together toward a common goal. For example the eight countries share groundwater data can we expand that list. We asked if it would be possible for NGA to be a steward to get international partnerships to do the work and help to pay the science community to conduct the research. We talked about utilizing NURRI grants and what we tell lastly we talked about techniques for soil complexity or or tech to simplify soil complexity as a maybe something that the NGA could help with. We actually had the NGA guy in our group so we did not have those questions. So we came up with these three kind of ideas that we want to focus on data fusion from multiple types of sensors and also multiple types of data to come up with better measurements of recharge. Definitely there should be a research focus on transboundary aquifers which is definitely a key knowledge gap out there in terms of managing groundwater and that applies to both the water reach systems as well as fossil groundwater systems. And looking at also how the surface water and groundwater is interacting in these water reach systems that is also a key. I think there's there's a lot to learn in that area. I think that's where we start. We this fourth question was basically the same as the fourth question on the last breakout group which was examples of successful collaboration. So we actually felt like we more or less covered that on the last go-round. So we didn't have a lot of creative new input. There's a couple things up there but other than that I'm going to punt on that question. We did as well. Sorry. Instead of looking at just the successful examples we actually wanted to look at the four transboundary aquifer treaties that are known to be out there and we want to see what works and what doesn't work in those treaties and we thought that would be informative for us. We also want to look at interstate compacts of water sharing and usage in the U.S. and I thought maybe in India we could look at because India has a lot of interstate water issues as well. And also Mike I don't think he's here. Mike talked about the upper San Pedro basin in Arizona. There's a lot of groundwater movement between Mexico and the U.S. where we could learn from that basin as well. So we wanted it's not really successful examples but we thought these are the areas we could focus on to learn more about collaboration opportunities and research on research.