NASA ARSET Overview and Analysis of NASA Terrestrial Water Storage Data from GRACE_GRACE-FO Pt. 1

Speaker 1: Hello everyone, Welcome to this applied remote sensing training on monitoring groundwater changes from Water Resources Management. My name is Amita Meta from our SET, and today we will be talking about overview and analysis of NASA terrestrial water storage data from Gravity Recovery and Climate Experiment or GRACE and GRACE follow on missions. We have a guest speicker today, doctor Matthew Roddell from NASAGA Space Flag Center, and I will be introducing him shortly. We'll start with a brief introduction to our SET program.

Speaker 1: Our SET is part of NASA's Earth Action Capacity Building program and it provides accessible and relevant, cost free trainings on remote sensing satellites, sensors, methods, and tools. Trainings include a variety of applications of satellite data on these thematic areas including agriculture, disasters, ecological conservation, health and air quality, water resources, and wildline fires. Trainings are tailored to audiences with a variety of experienced levels. Our SET trainings are online and there are in person trainings offered as well.

Speaker 1: There live and instructor led like this one, or there are asynchronus and self based trainings available from our SET website. As I mentioned, they're cost free. Most of our training material is translated into Spanish and many trainings are bilingual and there are multilingual options available. Our set only uses open source software and data and accommodates differing levels of expertise. Again, please visit our set website for more information. We will start with this webinar series on monitoring groundwater changes for water resources management.

Speaker 1: Now. Groundwater is water from precipitation that percolates into the soil and moves downward. It fills up cracks and openings in the rocks and sand below the surface. Depending on the porousity of the soil, it percolates down and it can extend from just below the surface all the way down to several thousand feet. You can see from this diagram that from total global water, most of it is saline and it's in ocean or it's brickish. Just two point five percent of total water is fresh water, and out of that thirty percent is in groundwater.

Speaker 1: Groundwater ages can range from months to millions of years, depending on the source of the water and how deep it is. The total groundwater volume in the upper two kilometer of continental crust is approximately twenty two point six million cubic kilometer and out of which small amount abouzero point one to five million cubic kilometers is less than fifty years old. It is the most extracted resource worldwide and its renewable resource depending on regional and environmental conditions. Traditionally, water wells are used to monitor groundwater levels.

Speaker 1: Here you can see from National Groundwater Monitoring Network blue dots show wells in the United States. So water level in these wells are regularly monitored to estimate what the groundwater levels would be at that location, and there are no direct measurements of groundwater from remote sensing observations. Measurements from GRACE and Grace follow On. They are used to estimate or infer total terrestrial water depth, and these total water depths are then used, along with additional hydrological information to derive global large scale ground water distribution at their solution of about one hundred and fifty thousand kilometer square.

Speaker 1: Additionally, there are two more sources of groundwater data. One is Global Land Data Assimilation System or GILDAS that assimilates GRACE and Grace follow On data and data as global groundwater data, and the second one is an opera surface displacement product which is derived from synthetic aperture radar measurements from Sentinel one satellite and so these two are also groundwater products available from NASA.

Speaker 1: So the overall training learning objectives are that by the end of this training you will be able to identify characteristics of groundwater quantity data sets from GRACE missions, GILDAS and opera displacement product. Access analyze and visualize GRACE data to monitor seasonal to inter annual changes in large scale total terrestrial water storage. Access analyze and usualized GILDAS data to monitor interannual to interdictal changes in groundwater at regional scale, map surface subsidence related to groundwater extractions with OPERAS data, and recognize applications of these groundwater data sets available at multiple spatial and temporal scales for monitoring drought and flood conditions and for planning groundwater resources for a variety of uses such as for drinking water or for irrigation, etc.

Speaker 1: These are the prerequisites fundamentals of remote sensing. This is a self paced course available from our set website. It provides background about satellites, their orbits, different sensors, their special and temporal resolutions and other characteristics of remote sensing data. Then there is an introductory training are set provided in twenty twenty on groundwater monitoring using obervations from Grace Missions that also has some useful information. This training will have three parts. Today, as I mentioned, will be focusing on overview and analysis of terrestrial water storage data from GRACE and Grace Fallen.

Speaker 1: Next week on April twenty eighth, Part two will be on overview and applications of gildask groundwater data products at regional scale, and on thirtieth of April, Part three will be on overview and applications of opera disc to monitor groundwater changes. There will be one homework posted on last day of the training that is thirty eeth of April, and the homework will be due on fifteenth of May. The homework will be posted on the training web page and a certificate of completion will be awarded to those who attend all live sessions and complete the homework assignments before the given t day.

Speaker 1: Start with today's session and specific objectives for today are that by the end of part one you will be able to identify characteristics of terrestrial water storage data sets from Grace and Grace Swallow on monitor seasonal and interannual changes in terrestrial water storage using GRACE analysis tool. So the outline for today is that our speaker, Doctor Matthew Rodell, will provide an overview of GRACE missions, Race data products and applications. Then we will have a demonstration of race data analysis tool that allows analysis and visualization of terrestrial water storage data.

Speaker 1: There will also be a hands on exercise that you will be conducting using the same analysis tool and some of your homework questions will be based on that. At the end, we will have a question and answer session and just a note about asking questions. Please put your questions in the questions box and we will let us them at the end of the webinar. Feel free to enter your questions as we go and we will try to get to all the questions during the question and answer session. After the webinar, the remainder of the questions will be answered in the Q and a document which will be posted on the training website about a week after the training.

Speaker 1: With that, I want to introduce our speaker for today, doctor Matthew Rodel. Doctor Matthew Rodel is the Deputy Director of Earth Sciences for Hytrosphere, Biosphere and Geophysics, or hPG at NASA Corder Spaceplight Center. HBG comprises more than three hundred and fifty scientists and engineers within five laboratories involved in remote sensing and numerical modeling of land and ocean processes and related applications. Doctor rodel has also served as Associate Deputy Director for HBG and as Chief of the Hydrological Sciences Lab.

Speaker 1: He is a member of the science teams for NASA's Gray's follow On Mission and Future Grace Continuity or Grace Sea Mission. He leads the Global Land Data Assimilation System and projects focused on monitoring groundwater storage changes, mapping and forecasting drought wetness, and detecting climate related variations in the watorcycle. Doctor Roddell is a past associate editor for the Journal of Hydrology and a current editor for the Journal of Hydrometology. He has also chaired the Hydrology program for the American Geophysical Union or AGU Fall meeting from twenty nine to twenty eleven and led various national and international scientific working groups.

Speaker 1: He received a Presidential Early Career Award for Scientists and Engineers in two thousand and six, a NASAJSFC Earth Science Achievement Award in two thousand and seven, a Robert H. Goddard Award for Exceptional Achievement in Science in twenty eleven, and an Arthur S. Fleming Award for Outstanding Federal Service in the area of Basic Science in twenty fifteen. He was elected to the rank of AGU Fellow in twenty twenty two. Doctor Rodel has more than one hundred and sixty peer reviewed publications and appears on clariveet Web of Sciences twenty eighteen to twenty three lists of highly cited researchers.

Speaker 1: He holds a BS in Environmental science from the College of William and Mary and a PhD in Geological Sciences from the University of Texas and Austin. With that Midnight, doctor Rodell Matt take it.

Speaker 2: Away, Hey, thank you, Amita.

Speaker 3: I might to provide an overview of the Grace and Grace follow On missions. So, Grace and Grace follow On are two separate satellite missions. One launched in two thousand and two and ended in twenty seventeen, and Grace follow On picked up in twenty eighteen and is still active. They were both jointly developed by NASA and the Germans, and both of them are twin satellite systems meeting their two satellites operating at once. And the two satellites are critical because actually the key measurement is the distance between those satellites and how it changes in time.

Speaker 3: So one is following the other, or about two and twenty kilometers apart, about four hundred kilometers above the land surface or Earth surface, and as they orbit the Earth in a near polar orbit, they basically observe all different regions of the world pretty well over the course of about a thirty day period. So these satellites again provide global coverage. You see the dates of the two missions, and right, the key measurement is the distance between the two satellites, and that distance is measured by a micro rate K band ranging instrument and actually on grace follow on, there's also an experimental laser instrument.

Speaker 3: And so imagine at two hundred and twenty kilometer distance and you're able to measure that distance every five seconds with the precision down to the size of a red blood cell.

Speaker 2: That's how accurate these measurement measurements are.

Speaker 3: In addition, there each satellite has an accelerometer on board. When an accelerometer does is it measures non gravitational forces on the satellite, in particular things like atmospheric drag. So you don't want the atmospheric drag to leak into your estimates of the changes in the distance but then and then cause errors in your observations of the gravity field. There are also GPS receivers on board to provide the precise locations of the satellites. So what happens is, you know, as these two satellites are orbiting the Earth, imagine what we call a mass anomaly, like a mountain range.

Speaker 3: That's where there's a bit or actually a lot more mass at the surface than there is on average. So imagine these satellites floating along in space and they come closer and closer to a to a mountain range, and that mountain range has extra mass, which means there's extra gravitational potential. And what it does it pulls the first satellite forward, and that satellite is basically the separation between the two satellites increases because that first satellite is sort of speeding.

Speaker 2: Up as they pass over the mountain range.

Speaker 3: The first satellite sort of held back by that extra gravity, and the second satellite speeds up in the distance between the two of them become smaller, and then things sort of even out again once they pass the Mountain range.

Speaker 2: So that's sort of an extreme example.

Speaker 3: Actually, you know, Mountain range is actually a huge amount of mass, but these satellites in the measurements are so sensitive that they can actually detect changes caused by changes in mass over time over the same location from month to month. So for example, if there's a big snowstorm and there's a lot more mass on the ground because of all that snow, that's actually enough mass and enough gravitational change to affect how the satellites orbit in a way that we can detect and then infer the amount of snow on the ground.

Speaker 3: So again we use sort of fundamental physics, you know, or understanding of how mass and gravity are related to translate the grace follow on measurements these distance between the satellites into gravi and then from there into mass concentrations, meaning like where there's more or less mass on the surface of the Earth. So, as I mentioned before, there there are subtle shifts and it's gravity field that.

Speaker 2: Are caused by movements of mass around the Earth.

Speaker 3: If you think about what moves or how do you have large movements of mass on the Earth. The biggest one is the is the ocean, So you think about ocean tides and ocean sloshing around in the ocean basins. That's a huge amount of mass. There's also atmospheric mass. Think about surface pressure pressure changes of the surface.

Speaker 2: That's really a.

Speaker 3: Measurement of the amount of mass of atmosphere above given location. And then over land changes the mass are primarily associated with redistribution of water. So again you know something like a snowstorm or a drought, or or a flood, or even you know, just smaller changes in water storage over a large region. Those are actually huge quantities of mass that again are are big enough to cause changes in gravity that then affect the orbits of the Grace satellites. So what we do is, after accounting for the atmospheric oceanic changes using models and measurements of the Earth, you sort of strip those off from the Grace observations.

Speaker 3: There are also some solid earth mass redistributions.

Speaker 2: For example, if there's an earthquake, that.

Speaker 3: Could be a large amount of mass moving and we can estimate how much that is, and there are solid earth tides that are pretty well understood. So after you remove all those other components, what's left is the changes in mass associated with changes in interestraal water storage. So I'll get to what trustra water storage is in a moment, but spice to say, it's of the it's all the water on and in the land surface.

Speaker 1: Uh.

Speaker 3: And so the data that we get from GRACE are provided as an equivalent water thickness.

Speaker 2: So imagine if you.

Speaker 3: Took all the water on and in the land, all the groundwater, so is your surface water snow ice and just created, you know, just ponded it on the surface, and then you watched how that how the the depth of that pond changed over time, you know, sort of like the depth of water in your bathtub. And and that is provided those those an anomalies of trust water storage relative to some arbitrary data are provided in centimeters or millimeters equivalent water thickness.

Speaker 2: So the so you know, when we.

Speaker 3: Say level one, two, three for GRACE products, Level one is basically the the measurements of the distance between two satellites. Level two is is the gravity anomaly fields and level three or the trust water storage anomally data that we as hydrologists can use. The effect of spatial resolution of these observations from GRACE and GRACE follow under is very coarse relative to other measurement systems that you're probably familiar with. So for GRACE, the effect of spatial resolutions about one hundred thousand to one.

Speaker 2: Hundred and fifty thousand square kilometers.

Speaker 3: To give you an idea of how big that is, the state of Illinois about one hundred and forty five thousand square kilometers, So we're.

Speaker 2: Talking about very large regions.

Speaker 3: And actually, when GRACE first launched, a lot of hydrologists through up their hands said I can't use this, it's much too much To course, I was actually among the first first hydrologists to be involved in using GRACE data.

Speaker 2: And we figure out a lot of ways to use it.

Speaker 3: And the bottom line is we have no other global measurements of total trust or water storage in a particular ground whatter, which we'll talk about later. So if you're clever, you can find a lot of ways to use these data despite the coarseness. If you look for trust or water storage data, and we'll provide some links later. They're often provided on a one degree or a half degree resolution grid and I want you to it's an important takeometric, important take on point that's really just for convenience.

Speaker 3: In order for those data to be meaningful, you can't just take one grid point and say this is how the water storage is changing over that grid point. You have to average over a sufficiently large region and I'm talking about greater than one hundred thousand square kilometers in order for your your results. You know, the time series of trust to water stores you might get, for example, in order for that to be meaningful, and the errors or the uncertainty is on the order of one to two centimeters quite one height of water change each month for a three hundred thousand square kilometer or a larger region.

Speaker 3: Okay, so what can we do with these these trust water storage data? So again I mentioned before trust water storage is some of the groundwater, some I sure surface water, snow and ice.

Speaker 2: As you can see sort of the chart on the right there.

Speaker 3: So Grace cannot and Grace follow cannot differentiate among these different trust water storage components. They're only showing the bar on the left there this total trust water storage and and in fact, I can only tell us the the changes in trust water storage, not the absolute amount of water. So don't expect that grace is going to tell you how much water is left in the Ogalalla aqua, for example. It's only gonna tell us how that water storage is changing, usually relative to the long term mean, from the GRACE data.

Speaker 3: So one of the ways that we can we can sort of tease out what's happening with the individual components is to combine the GRACE data with a with a landserface.

Speaker 2: Model through data a similation.

Speaker 3: And because that model has other information going into it from other observations and and our understanding of the system, it's able to do vertical, horizontal and temporical temporal disaggregation to make the GRACE data more more useful. So you know, if you have trust your water storage data, you know a simple approach for disaggregating would be to simply do sort of a water balance approach.

Speaker 2: So you see the boldled equation at the.

Speaker 3: Top there, groundwater equals trust your water storage chain or ground road change equals stretcher water storage change minus soul moisture minus snowater equivalent minus surface water. And you know early on in the GRACE mission. That was one of the ways that we that we evaluate the GRACE data is is to do use you know, land based observational time series of these components and and sum them all up. And so do we actually see the trust water storage changes we expect to see based on grace answer is yes, again, you can.

Speaker 2: You can do a you can do the equation.

Speaker 3: You set the equation up so you find groundwater storage changes as a residual if you have estimates of the other components from either observations or models. A more sophisticated approach is data assimilation and and this involves again you have.

Speaker 2: A land service model.

Speaker 3: The LANDSERFCE model has various inputs going into it in addition to the GRACE data. The LANDSERCE model on its own can provide estimates of how the Trust water storage change components are changing in time. But then when you assimilate the GRACE data, you have an additional constraint on the model. You basically get better results. So we're gonna be talking about GLEDS version two point two. And in this version, we have a landsforce model called the cash Man Lanserce model that simulates groundwater storage in addition to the other components, and by using GRACE as a as a constraint on that model, we get a result that's better than either the model alone or the Grace observation alone.

Speaker 3: So we'll talk about some of the GRACE products that are out there. So here's some links.

Speaker 2: When you have a moment, you.

Speaker 3: Can get the slides and follow these links. There are several different Grace and Grace follow on trusher water storage data products available. Protter Spaceflight Center, Just Propulsion Laboratory gf SAID which is in Germany, and the University Texas all of their own products. So, for example, the GPL Grace Data Portal has a few options here. You can download the grids as global mass cons. A mascon is sort of an advanced way of deriving trust or water storage data from the GRACE observations, and all of the centers now use.

Speaker 2: Mass cons as sort of the preferred method.

Speaker 3: There's an interactive Grace data browser and you can download the monthly grids for the land or for the ocean.

Speaker 3: So there are you know, a lot of applications of Grace and Grace follow on data. One of the early ones or not so early Actually, this is publication from twenty eighteen where we looked at the changes in trust water storage over the peer twenty two to twenty sixteen. Basically, at each point on Earth, we determined the rate of change of trust your water storage by fitting a linear trend after removing the seasonal cycle. So if you had you see a blue here, that means trend was upward. It was an increase in trust or water storage on average over time.

Speaker 3: You see the bar at the bottom of the units are centimeters per year, and where it's a yellow, orange or red, that means that the trust water storage was decreasing over time. And so what we did with this this study was was try to understand why trustal water storage was changing. And in some cases it's just natural variability and those are that's indicated by green. And what I mean by that is maybe you had a drought in the beginning of the time period around two thousand and two, and maybe there was you know, just sort of a wet period towards the end, and so it looks like you have this increase in trust your water storage, but we don't.

Speaker 3: We don't expected to continue a long term. It's just sort of part of the natural fluctuations. On the other hand, there are regions where where we might be you know, there might be a human a direct human impact, like groundwater pumping. And one of the first ones that we saw and the gray data was in northern India where a huge amount of water is pumped out of the opera there to use for irrigating crops. When you pump water out of an aqua faster than it can recharge, and you pour it on the land surface derigate crops, and most of the water evaporates or transpires, then over time the operas.

Speaker 2: Can be depleted and and there's a there's a.

Speaker 3: Very fast rate of depletion in northern India that's clearly anthropogenic. And then the other thing is there might be climate change signals. And the most obvious place to look for this is is someplace like Greenland, where the ice sheets are blating or melting away into the ocean, and that's why most of Greenland is red. There there might be there's some debate over whether the blue in the middle is an accumulation of snow or just a just an artifact of the data processing that might be related to something called post glacial rebound.

Speaker 3: Another more operational application is is drought monitoring. And one thing we've done is we've developed these drought indicators that are based on the GRACE data assimilation into the LANDSERFCE models. So the data simulation allows us to do again the vertical, temporal and horizontal downscaling of the GRACE data. Remember, GRACE data are monthly, they're very coarse and and in fact they're only available a few months after a near real time after real time. So to make them really useful for an operational application like drought monitoring, the data.

Speaker 2: Similation is really essential.

Speaker 3: It allows us to run up to neural time, it allows much higher spatial and temporal resolution, and we can also do things like disaggregating vertically. So this is the top two examples are a drought indicator, which basically tells us the relative conditions or the conditions each location relative to the long term going back to nineteen forty eight, which is based on the model results. Obviously, GRACE doesn't go back to nineteen forty eight. And what we see right now is we're get into some pretty dry conditions over much of the US.

Speaker 3: If you look top right there, you can see that around the world. It's a mix of dry and wet conditions around the world, which.

Speaker 2: Is what you'd expect.

Speaker 3: And then there's an example at the bottom there showing drought conditions in Brazil, which matches we see in several news reports. Another thing we can use GRACE data for is looking at flood vulnerability or flood potential. And so we have enough data from GRACE and GRACE follow on going back to two thousand and two, we have a pretty good handle on the range of variability of trust your water storage that we would expect, and when when the trustre water storage is up near the top of that range, we would say a region is vulnerable to floods, meaning you know, additional water that comes down the Lancerfruce's rain or snowmelt is likely to overwhelm the river system and then you start to have you start to have flooding.

Speaker 3: And so it's a useful way of basically preparing for floods, you know, looking for regions like this, this.

Speaker 2: Region is very wet.

Speaker 3: Maybe it's not flooding yet, but you know, look out if there's a storm or something.

Speaker 2: And that's that's some work that J.

Speaker 3: T Rieger worked on a while ago, and we've also used our drought wet syndicator maps for this purpose. Another recent, more scientific result was looking at what is apparently an abrupt decline in trust your water storage around the world, which happened around twenty fifteen. The time series is not here, but basically, if you look at the time series of all the average trust water storage for all of the land excluding the ice sheets, where there's a long term persistent trend, excluding the ice sheets and the glaciers, i should say, and you looked at how the time series evolves from two thousand and two to present, you would see around twenty fifteen the time series suddenly has this decline of about one centimeter equivalent head of water average over the entire Earth.

Speaker 3: And what's really interesting about that is that, you know, twenty fifteen lines up with when we had a series of or in the midst of a series of the warmest years on record, so twenty fifteen through twenty twenty three, or the nine enormost years in the in the data you know, in the temperature record for the Earth's surface, and that of course continued in twenty twenty four and twenty twenty five, and this decline has also persisted, or this this lower level trust your water storage has has persisted.

Speaker 2: The map here is showing it's twenty fifteen.

Speaker 3: A significant portion of the landsurface has has hit its lowest on record trust your Water Storage anomaly, which is another indicator of what's happening. And just for fun, here's an animation of GRACE data starting from twenty two and running through twenty sixteen, and just to see how it sort of evolves.

Speaker 2: You see these.

Speaker 3: Blobs of wet and dry moving around over time, and then it's going to zoom in on the Acabango Delta in Southern Africa and for this region that's outlined here. This is how the trust your Water storage changes over time, and you can see during this particular period there's a very wet period and the delta, which is home to a lot of cool, big animals you'd see in the zoo, was wet. There was a lot of water there and the elephants are happy. And then towards the end of this period, you going towards twenty sixteen, it's sort of coming back down towards normal.

Speaker 3: So that's where I'm going to leave it off, and I'm going to hand the mic over to Amita.

Speaker 1: Thank you so much Mett for your presentation on GRACE Missions data and applications. Next, going to have a short demonstration of GRACE Data Analysis Tool or interactive browser, which allows visualization of terrestrial water storage data from the missions.

Speaker 1: Is a brief outline we'll learn to navigate the GRACE Data Analysis tool. There are several features that are useful to selecting special and temporal domains. Then view terrestrial water storage or TWS maps and time series using the tool. And will use these two regions for case studies Colorado River Basin and Ogola La Aquifer. These two are very important for water resources in the Western US. So I'm going to share my screen with you. This is the data analysis tool and you can ask questions if you have any and submit questions.

Speaker 1: For launching tool to just click on launch tool here and then a welcome window opens. At the same time, you can see graced WS data in the background. So let's just quickly go through some of the features, but we are going to demonstrate all of them. So this symbol here this is about different data layers. There are multiple data sets available. You can use plus sign to add data layers. This symbol here it toggles between different layers. The visibility you can turn it off and on. I next to any data source provides information about that particular data, and this symbol here that we will look at.

Speaker 1: It allows you to draw a box to select a region, or drop a pin to select a point, or even choose river basins as we will see this one. Once you select the region, you can look at map by zooming in, or you can make time series and then you can download data by using this button. This errow here allows you to toggle between different layers, and this animate layer over time allows you to animate TWS data over different time range. So we'll start with looking at the map first. You can zoom in by your computer mouse using your computer mouse, or you can use plus and minus signs here to zoom in and zoom out.

Speaker 1: Note that here is the time window. You can change time. Last data available here is January twenty twenty six. You can change time, month and year, and you can see TWS changing. Here's the color bar given for that units are in centimeter and the range shown here is minus thirty to thirty centimeter approximately. But you can change a range of these values by sliding here. You can also change visibility or transparency of any layer by changing transparency by moving along this line. And there are some more features here.

Speaker 1: This is just a different map projection. This allows you to turn off and on, say timeline or there are some options here, but we'll keep them on. And this one allows you to label the map in the sense that once you turn it on you will see country boundaries, which is useful. And if you just click here on the home button, you will be taking back to where we started. You can again zoom in on any region and you can uh look at the year or month that you're interested in. What we are going to do now is uh look at the analysis options.

Speaker 1: But let's just look at the different data sets available here. Right now we're looking at the maps. Map that you'll see here is from GPL Water Equivalent Thickness over land. If you click here for data search, there are four options, So one way looking at here is land GPL water equivalent thickness, and the similar thickness is available for ocean. Then there are two data data sets CSR water equivalent Thickness. This is from University of Texas, Austin. As was mentioned earlier, and then this is for ocean, so you can be looking at this.

Speaker 1: You can add this data as well if you like, and then it shows both of these two different color tables show up, and then you can turn this on and off. You can select right. Now, let's just stay with GPL water equivalent thickness, and now you can look at this symbol here. This allows you to drop a pin, draw a box by clicking on the map, and you can select a basin. When you click on select a basin, you can see that all the river basins are now demarketed. And when you point to any of this river basin, it highlights with blue color and it shows which basin.

Speaker 1: So this is Mississippi, this is Rio Gran and this is Colorado your basin. So let's start with Colorado. This is Amazon and you can you can these are global rivers. Major rivers are shown here, but let's just focus on Colorado River. I'm going to click here to select the river basin and now you can zoom in to look at map. You can change h values here to edges to see proper colors here, and to look at time series. Now you can keep operation as time series and choose the time domain that you want to look at or time range.

Speaker 1: These are monthly data as we saw, and let's keep the entire range right now and say create chart. When you click on that, you will see time series appearing here. Once you have the time series, you can see monthly data. This is the time access and here you can see terrestrial water storage change is shown in centimeters. The values are fifteen to ten centimeters. And if you see on top here, you have several options. You can de season this data by clicking here. You can have either line or column.

Speaker 1: You can fit a trend type here, you can have either linear or you can fit polynomial. You can see that not only there is annual variability, but that is inter annual variability as well, and there is a steady decrease in terrestrial water storage as you can see over this period. And this is the polynomial fit. And you can get coefficients here as well. You can download this time series either a spn' G your GPEG image or PDF, or you can download this data as CSV or as XCELS file. You can have this or data table.

Speaker 1: So multiple options are there, but this allows you to quickly pick an entire river basin and see how TWS is changing over time. Also not to one thing. Here, you will see some between there's a gap between grace and grace fall on, so you will see missing data here. So when you look at maps, it's best to look at where you have the data. Also, once you have the time series, you can move along this xxis and you can see this vertical line. It shows year, month and TWS data for that month. This is for the entire basin, averaged over the basin and when you are clicking through the time, if you look at the map, it will show you map for that particular day.

Speaker 1: So this is when we find minimum TWS and you can see colors if you see this is mostly in negative TWS. If you go to say January, February or twenty two thousand and five, you can see now this is in blue zone. There is positive TWS animally. So you can look at each individual month and corresponding map here. So average value and map you can see here. So we looked at Colorado basin. We want to now look at Ogula lacquifer and for that we're going to turn the basins off and draw a box. Now approximate location of the Acrefa Gogola Lockwifer is thirty two to forty four north and ninety six two one oh six west.

Speaker 1: What I'm going to do is you can click here and make a rough box. Then the coordinates will appear. And if you know exact coordinates, you can enter here, or you can just click and draw a box wherever you want to if you're interested in some other region here, I'm just going to put exact latitude longitudes inside. Know what the region I'm looking for, and once you do that, you will see a box appears here the multiple states in Ogola lack Pifer. You can see that, and now you can do the same. You can create chart and this is for that box.

Speaker 1: Averaged over that box. You can see clearly when you can see that there is a degrees in terrestrial water storage over time, especially after this period. And you can see map also here and then when you go down here there is a slow change, but you can also see that there are this inter annual variability period you can see in the terrestrial water storage. So you can explore this in the area of your own interest. This is just to show the main features. So one more things to see here is that over this box.

Speaker 1: We want to compare different data sets. We can go back and add let's look at CSR land DWS and you can add the data and now you can see that you can turn this on and off. So this is JPL and this is CSR, so you can compare this. Resolutions are different here. That's why you can see this is three degree. I believe this is half a degree. So you can click back and forth and see this is more smooth. But basic features are the same in both the data sets. But so you can compare two data sets if you like.

Speaker 1: And if you want to make time series, you will have to pick this data set and go through the same procedure and compare series. So this is the basic browser that we wanted to share with you and you can explore other regions. You have an exercise that is based on this browser. So I want to conclude this demonstration. Now, this concludes our demonstration of the increased Data Analysis tool and you will have chance to work with this data tool. There is an exercise available on the training page that you can download and follow the steps to work on a case study.

Speaker 1: Before we start the exercise, let's summarize what we saw today we saw that grazed missions provide unique measurements of variations in mass or gravity changes over entire Earth surface, producing monthly maps of the gravity field. These variations in gravity are primarily related to the movement of terrestrial water and they are interpreted in in terms of change in equivalent water thickness or terrestrial water storage t WS. Using the GRACE TWS along with model based hydrologic components, groundwater can be estimated and this is going to be the topic of our next session.

Speaker 1: Finally, GRACE data are used to obtain t WS and groundwater change information globally and have been useful in monitoring flood and rout conditions and large scale groundwater depletion. Now. Prior to the launch of GRACE, there were no global measurements of groundwater. The only way to estimate groundwater was from well water levels wherever they were available. The GRACE missions have made it possible to estimate TWS and now we have more than twenty years of time series of global water storage data and this is a major advantage of GRACE missions.

Speaker 1: There's some limitations we should keep in mind. GRACE and GRACE swallow based estimates of t ws and groundwater are available globally. However, their special resolution is relatively low. It is three hundred eighty by three hundred eighty kilometers square approximately, and so it cannot resolve small watersheds. T WS is measured in centimeters or meters, which is much smaller compared to the Earth's radius which is approximately six and seventy eight kilometers, and so TWS has an estimated uncertainty of about two to three centimeters.

Speaker 1: In addition, gravity change is caused by mass distribution in the solid earth, such as large earthquakes or glacial adjustments. They must be removed from the measurements before deriving t WS thickness, so that also should be kept in mind. With that, our next ssion is going to be about view of groundwater data from a Global Land Data Assimilation System version two point two.

Speaker 1: As I mentioned earlier, there is going to be one homework assignment that will open on thirtieth of April at the end of the training and it will be available from the training web page. Answers must be submitted via Google forms and the homework will be due by fifteenth of May. A certificate of completion will be awarded to those who attend all three live webinars and complete the homework assignment by the deadline. You will receive a certificate via email approximately two months after completion of the course.

Speaker 1: Once again, we want to thank doctor Matthew Rodell for his excellent presentation about race missions, data and applications, and he will be our speaker next week as well talking about jailed as peace groundwater. Contact information for doctor Rottel and also you can contact our set anytime with any questions. Our set, website and YouTube links are given here. For questions, comments, or to share how you have applied our trainings to your work or studies, Please email at our set at gmail dot com and join our quarterly newsletter to stay up to date on our latest trainings and you can do that by joining our list serve.

Speaker 1: Here are some useful resources for your information and we want to thank you for attending today's session. We have a few minutes for our exercise to start the exercise now and then we will have our question and answer session. Okay, so we'll start with the question and answer session and if you have any questions about the exercise, you can email us or we can talk next in next session on twenty eighth alsome. So we'll start with the questions. Question one is that's the distance between both sensors change sometimes?

Speaker 1: And magic can unmuted and answer the question.

Speaker 3: Yeah, hi, I'm sorry, I just uh, I was looking at something. Can you say the question more? Oh, does the distance be both sensors change sometimes? So yes, it does. And actually that's you know, that's the that's the key measurement, right, the distance between the two satellites, and normally it's about they're about two.

Speaker 2: Hundred kilometers apart, but.

Speaker 3: You know that's sort of the two satellites are free floating, so it may vary from you know, by a few kilometers, so if you tens of kilometers or the course of the mission, I guess, but but yeah, I mean that's that's the key measurement, is the the distance between the satellites. And really how you know they've measured down to you know, how fast is that distance changing? More of a distance range rate we call it, and even like the acceleration of how it's changing.

Speaker 2: They can look at that as well.

Speaker 1: Thank you. The next question is how can multi source satellite data be combined to reduce uncertainty in groundwater estimation and what are the limitations of this approach for decision making.

Speaker 2: Well, so the answer is yes and no.

Speaker 3: Here uh, the the grade data are really the only satellite observation that provided direct you know's not even a direct measurement, but really a measurement that you can use to understand changes in groundwater, you know, fairly directly.

Speaker 2: There are others satellites that can give us sort of sort of a.

Speaker 3: You know, an idea of what might be happening with groundwater, like the recently launched nice Our satellite, which can tell us very precise changes in the elevation of the land surface when an.

Speaker 2: OKFA is is is dewatered.

Speaker 3: When you remove a lot of water from the ocfur, the the the OCFA compacts and so the land surface declines a little bit, and we can see that with the nice Our satellite. But but it's not really a one to one. You know, the Earth isn't perfectly elastic, so when you add more water back to the OXFA, it doesn't necessarily, you know, the landsurface sort of bouncing back to where it was before. And then of course there are other observations things like precipitation and solar radiation we can be incorporated into our landsurface model as as we talked about, and those helped you constrain the overall water balance, which then helps us to you know, understand how much water may be entering and leaving the Aquifa.

Speaker 1: Thank you. The question three is is the Great Satellite Mission suitable for monitoring interest and what stortage changes in archipelagic countries such as Indonesia.

Speaker 3: Unfortunately no, And the problem there is that the islands are really smaller than the effective spatial resolution of GRACE or Grace follow on, and so you have a lot of leakage of the signal from the ocean into the island.

Speaker 2: So it's it's very hard to.

Speaker 3: To sort of isolate the mass changes that are happening in an archipelagalo or the or an island from from the often much larger changes than the ocean that surrounds it.

Speaker 1: And the next question is how can uncertainty in satellite dare groundwater estimates be effectively communicated to policy makers? And are there probabilistics or machine learning approaches such as measure density networks that can better characterize the multi source uncertainty inherent in groundwater estimation.

Speaker 3: Well, so I'll take the first part of that is effectively communicating to policy makers. What we've really found is that you have to put everything into their language. It's not enough to say, uh, here's here's a change in trust your water storage, or here's a trend in trust your water storage. You know, they might just sort of say, yawn, I don't know what, I don't know how that I can apply that. So so we we found that it's important to work with stakeholders and policymakers to develop products that are really sort of tailored to their specific needs.

Speaker 3: And and I think, you know, you're sort of answering your own question here. Of course, you know, we've only just begun to look at how machine learning and AI could help with with improving that communication and developing the sorts of products that would really be more more valuable to these these end users, so that we can ensure that they that they really make good use of the information that's available.

Speaker 1: Thank you. The next question is does vegetation contribute to these animalies or depletion that we see, for example in the himalay and north of India and maybe the mountain glacier retreat.

Speaker 2: So the glacier retreat.

Speaker 3: Absolutely, there are areas in the world where glaciers have been retreating rapidly and that's and there can be a very large mass change signal associated with that. And we've actually, you know, especially areas where there are multiple glaciers, we've we've looked at grace and seen at least the component of the grace trend that we can attribute to to the the mass loss from the glaciers recording vegetation. I did you know a study in that twenty one years ago, and and we found that in terms of vegetation water storage, the largest changes happened actually over agricultural regions, you know, between you know, the time of the times of you know, plant growth to to when they're their their maximum and most you know, most massive to when they're either harvested or they go into sin essence.

Speaker 3: But those changes are still pretty small. They're basically within the uncertainty range of grace. I'd say it's on the order of half a millimeter per year of equivalent height of water in terms of mass change in vegetation water storage. So it's really it's not something we typically think about too much, and it's you know, if you're interested in that component. It's unfortunately not something that's going to be you could easily detect within the grace signal because there are so many other larger components of the grace signal like the soil moisture, changes in groundwater, et cetera.

Speaker 1: Great, Thank you so much. The next question is I would like to know what the observation that of grace data is in the subsurface. In other words, is it useful for evaluating for example, animal is in deep confined a prefer systems.

Speaker 3: Well, there's sort of two answers to this question. The first is that grace is not limited by depth whatsoever, because it's because the grace system is really measuring changes in gravity that are associated with changes in mass at all depths. There's no there's no limit on how deep grace can measure.

Speaker 2: However, if you're talking.

Speaker 3: About confined octfers, there's not so much of a one to one between a change in in headed a confined oct for and a mass change. So confined oct for head is really you know, if you mentioned if you put it well into a confined oct for, you're really measuring changes in the in the water pressure in that confined oc for uh and and that cannot be easily translated into a change in in mass storage. So it's it's difficult to to say, you know with certainty how much of a mass change is is how might be happening in the in the confined oct for But we do know that there are areas, you know, where unconfined and and confined octoras are being over exploited and over you know, average, over that very large area that you know, at the at the spatial scale that Grace would observe, we can see mass changes that almost certainly would be partly attributed to changes in those confined oct for us.

Speaker 1: To learn that. The next question is also is interesting. Has anyone investigated groundwork the storage changes in area experiencing mysterious earthquakes not related to fourth zones, like in the Great Plain or Louisiana.

Speaker 3: That's a good question. I don't I don't know that anyone has done that, and and I'm not sure exactly what the linkage would be there. I guess the idea is that if you've removed or added a lot of water to the to an opfer, that can then either lubricate the faults or or cause other changes that would then initiate.

Speaker 2: Such earthquakes. But I don't.

Speaker 3: I don't know of any specific research that's looked into that.

Speaker 1: The next question is, how can we design a unified probabilistic data fusion framework that integrates multi scale satellite observations and hydrological models to robustly estimate groundwater under uncertainty?

Speaker 3: Wow, that sounds like did AI asked this question? I'm not even sure how to answer that. This sounds like something that you might write a proposal to do. It's not something I can I can just answer off the top of my head. But certainly, you know, data fusion, machine learning, artificial intelligence, those are all things that we're beginning to you know, incorporate into how we analyze satellite observations, including GRACE and and combine them with hydrological and other other models.

Speaker 1: M The next question is from the Philippines. Given the archipelagic nature of our country, how useful would GRACE data be? Is that a way to overcome the pixel limitation on island nations?

Speaker 2: Well, again, unfortunately not at this time.

Speaker 3: It's it's again an issue with you know, islands being surrounded by ocean and they're being very large mass changes in the ocean.

Speaker 2: And because you know, Grace is not.

Speaker 3: Looking downward like a like you know, like a camera or an imager, it doesn't it can't get you know, crisply does define a region. Say this is what's happening in the region. This was happening outside. Remember, the key measurement is the distance between the satellites, and so we're very limited in terms of the spatial resolution we can get. And it's you know, you just cannot distinguish, for example, Philippine Islands from from the ocean around them in terms of in terms of the mass changes. Now if we you know, we are researching.

Speaker 2: You know, next generation satellite based.

Speaker 3: Grab imagery and there are some you know, potential systems that would increase the special resolution, maybe we get to the point where we were able to distinguish an island from the ocean around it.

Speaker 2: But that's that's at least a decade away.

Speaker 1: The next question is about I think it's about the browser. Why are there not river basins in Scandinavia? And we check and let you know, I'm just checking the browser. Major river basins, of course are there? Uh, they are very small river basins, they may not be included, but I will check and it might. Your question the next question is to what extent can groundwater depletion infert from graze based terrestrial water storage animal is act as a predictor or amplifier of compound climate extremes such as heat waves and wildfires.

Speaker 2: Yeah, I would say this is this is an active area of research.

Speaker 3: And uh, you know, when there's you know it say, it's more like the thresher water storage when that is is lower than you're you know, treasure water stort is lower, that means that the land is drier, there's less water available to plant roots, so the vegetation is is not as as moist as it could be, and those conditions make it more likely that you're going to have a wildfire. And similarly, when there's less water available for for evaporation and transporation at the surface, you know, when they when water evaporates or transpires, the air cools and when there's less whata are available, there's less that evaporative cooling and so of course that can that can amplify a heat wave.

Speaker 3: And and again I think this is the type of thing that the people are very interested in and researching right now.

Speaker 1: Great, thank you. The next question is again about the brows. The data analysis tool in the chart section for gris follow on why not replace no data with nuts and we'll give that feedback to the browser developer. Next question is what does D season mean? So, and that is about browser.

Speaker 2: No no, no, I think, I think, I know this is so? This is uh is a great question.

Speaker 3: D season means if you have a time series of observations from Grace. So let's say you have a time series of trust your water storage anomalies from GRACE over a particular region of interest. In that time series, you will see a natural seasonal cycle. You know it'll be there will be more trust your water storage during the wet season and less trust your water storage or in the dry season. And you know, depending on what part of the world you're in that you know that seasonality is different to D season means we use a statistical approach to basically remove that seasonal cycle so that we can then look at the the sort of non seasonal variations in trust your water storage.

Speaker 3: So, for example, well, if it's you know, if the average stress of water storage anomaly in the region is is you know, plus ten in the in the spring, and the averages is uh minus ten in the fall. You remove the seasonal cycles so that you have you know, basically zero in all seasons, and then it's easier to see like, Okay, we're drier than normal this spring, or we're you know, wetter than normal this fall. It's easier to sort of to understand where we are in terms of thrush of water stewards relative to the normal for that time of year.

Speaker 1: Thank you. The next question is how how can you get the actual volume of d w S in water volume rather than water thickness or equivalent thickness.

Speaker 2: That's that's pretty simple actually, just multiply.

Speaker 3: The the the equivalent water thickness by the area.

Speaker 2: So so if you have a.

Speaker 3: River basin that's one hundred thousand square kilometers and you have a trusted water storage anomaly of of of ten millimeters, multiply the ten milimeters by one hundred thousand square kilometers. You'll have to convert the units to you know, to probably be you know, cubic kilometers or cubic meters or whatever. And then and then that's how you estimate the volume.

Speaker 1: The next question again is about but the browser are the data an is tool. I'm interested in the analysis tool and I have a couple of questions. The option dtarend data. How is it detrend in the data set? So it's just fitting either line or a polynomial as you can see two time series and you can see the coefficients that it finds at the bottom. And if you're talking about removing seasonal trend what explained It just statistically removes seasonal variations. And the second part of it is that JPL water equivalent and CSR water equivalent.

Speaker 1: There are two data data sets. I see that CSR has better special resolution. But can you expand on the differences of these data sets? So MET can add to this. But I did share a link earlier and I'll put it here as well. It describes how to choose a data set or how you can average available data sets and then use that. But MET, if you have anything to add to that, please go ahead.

Speaker 2: Yeah, I just want to I.

Speaker 3: Want people to be careful about, you know, what's viewed as better spatial resolution.

Speaker 2: You know, just because CSR provides their.

Speaker 3: Their trust or water storage anomalies on a finer grid, maybe I forget if it's the zero point five degrees versus one degree, the effective spatial resolution revolution of grace remains the same. You still have to average those pixels over a sufficiently large region for your results to be meaningful. So that especially large region is you know, one hundred to one hundred fifty thousands core kilometers at a minimum. So, but in terms of other differences, you know, they do their processing in different ways, and you know, JPL may be better in some ways and CSR in in other ways, so it's hard for me to say which one is better overall.

Speaker 3: You know, sometimes people look at the differences between the JPL and the CSR and the Goddess Space Flight Center and the GFZ products and use that the spread among those different products as a measure of uncertainty. I don't really like doing that very much because I think it only tells you measure of the uncertainty in the way they process it, and not so much other sources of uncertainty, like the the limitations of the satellite observation itself, which would be you know, inheriting all the products.

Speaker 3: And I prefer to just focus on just pick one and focus on the science, is always my advice.

Speaker 1: Thank you. So much. The last part of the question is, I was wondering if these results would be the same similar as downloading the mask ons and plotting the data myself. I think, yes, that is true, it's the same data that's been shown here. Next question is could a grace like mission detect temporal mass changes on Mars and what magnitude of water movement would be required for detection.

Speaker 3: So so if you know, if there were significant mass changes on Mars, then yes, a grace like mission could could detect those. You know, for Earth, we know that you know, the the uncertainty level for a monthly gravity field or trustre water storage derived field is on the order of one to two centimeters equivalent height of water. So you need to have changes in water storage on Mars that were at least you know, one two centimeters equivalent height of water.

Speaker 2: You know.

Speaker 3: I don't know if anyone knows this for sure, but my intuition is that there's you know, pretty much no water cycle on Mars, so you're, you know, very small changes in mass associated with water storage. You know, there may be water know in you know, frozen in the polls on Mars, but if it's not moving, then a grace like mission isn't really going to tell you much information.

Speaker 2: I can't tell you, just.

Speaker 3: Like on Earth, they can't tell you the absolute amount of water there, can only tell you how it changes over time.

Speaker 2: And my guess is on Mars not changing much at all.

Speaker 1: Yeah, I think you next question is can I find the water depths through this in a twenty square kilometer area.

Speaker 3: Not using grace alone. You know, if you use one of the the GLEDS two point two data Grace data simulation based products, those are available on a on a quarter degree grid, so you know, twenty five by twenty five kilometers, which is not going to eat you down to your twenty square kilometers. And you know, as we discussed, there are a bunch of caveats in terms of interpreting the Child's two point two data.

Speaker 2: In particular, you.

Speaker 3: Know, the model doesn't know anything about you know, groundwater pumping and other human water management, so there are some fine scale properties processes not captured by the model, and therefore it's not it's quite imperfect in terms of downscaling the GRACE data.

Speaker 1: Yeah, just to note that next session on twenty eight will be about Jilda's two point two and noctre Rotal will explain this again. Question eighteen high everyone, I'm asking that Grace twenty five kilometer resolution data is available. I think you just answered that we will see that next week and with some caveats, but yes, Gildas has quarter degree data.

Speaker 3: Right, and again that's not purely Grace data. That's yeah, Gray data to you know, assimilate into a landservice model and integrated with other observations.

Speaker 1: Next question is can you please help me regarding how to create or get started to prepare such interactive web browser, So that actually is beyond the scope of this training, but we will get you some reference material for that information.

Speaker 1: Next question is is it possible to upload our own basin in a shape file format or is there another way to download data more systematically and more locally. So next a week we will this browser will not allow you to upload your own shape file, but you can download the data.

Speaker 3: Uh.

Speaker 1: The links and information was given in the presentation slides. So the data available in net CDF and GOOT format, so you can and get them into a GIS QGIS platform. You can also specially subset data so your region can be extracted and then you can in GIS, you can upload your own shape file and do the analysis. So if there is interest, we will have a small information We will provide information on that. Next question is do we need the newest QGS version for the next exercise. I believe we're not using any special features, but the procedure is based on the newer QGIS version, so when you follow the steps you may have to make some adjustments.

Speaker 1: The next question is for identification of a regional signal. Can you recommend practices for combining multiple mask ons for a single product or for multiple products to estimately W S and the uncertainty or of d W S I G E G. If you calculate the meantime series, should some waiting be applied to de emphasize mascons at the edge of the basins or would you say that the full range of all mask on across all products represent the uncertainty for a region.

Speaker 3: You know, the the mask on products they've already done some You know, they put some effort into trying to

Speaker 3: trying to account for for leakage across the basins. So so I understand what you're asking, and and you know it's possible that you could do a study and show that, you know, it's better to to you, you know, wait more heavily the interior mask ons than the ones around the edge. But I don't, I don't know that's that's ever been done. And like I said, you know, the mask on products do attempt to to counteract leakage. So you know, my general advice is just to use all of the equally weight all of the mask ons within your region of interest.

Speaker 2: But I'm you know, I'm not certain that that's the optimal way.

Speaker 1: The next question is what does water thickness mean?

Speaker 3: Yeah, so we use the term water thickness or equivalent height of water two as a sort of a simplified way of explaining a mass change. So so if you have a mass change it's you know, in terms of uh, you know, kilograms of water. We can then say, what what would that equate to, you.

Speaker 2: Know, averaged over a region of interest.

Speaker 3: So if we have a you know, grace region that's hundred thousand square kilometers, you know, you have a change in mass over that region, what would be the equivalent height of water that would cause that change in mass? And and you can think of it like, you know, again grace measures all the different components of trustre water storage. So there's the groundwater, the sawmash or the snow, and the surface waters. If you took all of that water and ponded it on the surface of your region, what GRACE is measuring is like changes in the in the height or depth of that water.

Speaker 3: So think of like a bathtub, right, So maybe you're you might maybe have a bathtub that's you fill a pathway with water and then you drain some of the water out and you see that you're, you know, one centimeter lower than you were before.

Speaker 2: And maybe you don't know.

Speaker 3: The total amount of water in that bathtub, but you know you've lost a centimeter of water.

Speaker 2: That's basically what we do with the GRACE data.

Speaker 1: Thank you, Matt. It's a great explanation. And the next one is two level three products account for regional earthquakes such as the one in twenty twenty three in Turkey, Syria. If studying a region like this for groundwater depletion, should the impact of seismic events be taken into account.

Speaker 3: So typically what we've seen is an earthquake needs to be greater than magnitude eight in order for it to have a larger mass change to be you know, to really affect the time series in a meaningful way. And when there are those large earthquakes, you know, like the Sumatra earthquake, then we there are some some specialists who work on earthquakes and related you know, gravitational changes that have provided data that that the the Grace team has used to sort of try to remove that signal.

Speaker 2: It's not perfect.

Speaker 3: It's actually surprisingly difficult to to provide even a mass change estimate for an earthquake because because you'll see sort of this ringing effect where there's you know, large mass change in the middle and then then the opposite sign mass change as you uh as you go outward from there, and uh, it's it's difficult to to remove that signal.

Speaker 2: But they do make an effort.

Speaker 3: Again for those earthquakes that are larger about magnitude eight, that's already incorporated into the products, but it's not perfect. So if you see, you know, if if you look at a gracetime series around you know, Sumatra area and you see some large changes, you know, some of those might be real, but a lot of it might be you know, part of the earthquake signal that was not completely removed.

Speaker 1: Well, that's interesting. The next question is how good is the GRASE data if I want to check the correlation between levee subsidence and groundwater or is it good when analyzing a large area.

Speaker 3: Well, if you're talking about a levee, I mean that's you know, then you're on the order of, you know, less than a kilometer, right, if I'm understanding the question correctly, and you know, again, the GRACE effective spatial resolution is about one hundred thousand square kilometers. So so even a you know, even a pretty large reservoir, because it's a small area, it can be sometimes hard to see the changes in water stored in that reservoir behind a behind a levee or a dam. For the larger reservoirs in the world, as they're filling, you know, for example, the three Gorgeous Reservoir in China, we can see that signal in the GRACE data.

Speaker 2: But if you're talking about you know.

Speaker 3: A smaller, smaller reservoir or levee, then it's probably going to be too small of a signal for GRACE to detect.

Speaker 1: The next question is the tool indicates that the inflammation is water thickness, does it mean the thickness of the water above and below the ground, is it ignoring the layers of soil in between?

Speaker 3: So yeah, I sort of answered this before, But the answer is like you take all the water on end below the surface and combine it into one. Almost like you took all the water and you ponded it on the surface, and that's what the thickness of water is. You know, changes in that in that ponded water, So it's not is ignoring the layers of soil in between? I guess the answer is sort of yes, because that's the soil layers are not changing, right, You're the mass of the soil is not changing unless you're having huge amounts of erosion.

Speaker 2: But you know, so it's really just looking at the water.

Speaker 1: Thank you. The next two questions are about the exercise. So does this exercise have to be done now? I mean in the live session or there is a deadline to it. We just gave some time so that you can explore the tool and if you have any questions, but it's not due today, you have almost between now and fifteenth of May when the homework is due. Your homework questions will be based on the exercise so you have time to finish that later on. The next is also about exercise. Yeah, it says month here July.

Speaker 1: Okay, there's a question in the exercise that says to move the month here to July twenty twenty one, and then it follows with a question regarding July twenty twenty two, should we used July twenty twenty two? So the question is about It goes from you're checking twenty one, twenty two, twenty three, twenty four, twenty five, all years, and then you are just then looking at one month, so one July twenty twenty two. Next question is could you please repeat what the vertical line on the plot was. So vertical line just it's the it's the place for month and year.

Speaker 1: So when you move that line along X axis, it shows which month and the year you're looking at, and then you can see the actual TWS animally at that that month and year along that line. And also as you move the line, you can see the maps changing in the browser, so you can see corresponding distribution of DWS animalies. Next question is are there any attempts to make to downscape the grace data for smaller watersheds or high resolution impact studies.

Speaker 3: Yes, there are, and and that's one of the things we'll be talking about with GLEDS two point two next week. GLS two point two is when we perform data assimilation, we take the grace data when we use it to constrain a land service model that incorporates a lot of a lot of other higher information, higher resolution information like precipitation and solar radiation and and soil types and other things. So so the answer is yes here, there's also you know, people are starting to look at you know, machine learning approaches and artificial intelligence and that sort of thing.

Speaker 2: But you know, none of these are perfect answers.

Speaker 3: You're never going to get, you know, a truly high resolution grace only product question.

Speaker 1: I think it's a follow up question to the one that you answered about the coastal region. It refers to the question about the identification of a regional signal given multiple mask ons and products and the uncertainty for the signal. In answering that question, please also consider the case of a coastal region. I think you covered that right, you mean.

Speaker 2: Yeah, I'm a little I'm a little lost because I don't know what the original question was. I'm not I'm to answer this one.

Speaker 1: Yeah, I'm not sure either. The question was about River Basini that you know near the edge. Would you just consider if you're looking at a region, how would you wait different mask ons. We can revisit this later on a question that you too. Can you plot multiple data on the same chart for comparison? I believe that you will have to download CSV files from two different data sets and then make time cites, either in Excel or in some other software. Question thirty three when will the present data March April twenty six be available?

Speaker 3: So the Great team has a goal of making the data available within about three months of real time. Often it may be more like four or five months, and it just depends on the degree of processing that's required. And sometimes the satellites are in a better orbit with fewer repeat pass overpasses, and other times there's more uncertainty to deal with. So I don't know exactly the answer to that because I'm not one of the product developers, but it's typically typically I would say it's on the order three to four months after real time, And this is actually one of the things we address with GLS two point two and data assimilations we can and much closer to real time using other information to extrapolate to near real time.

Speaker 1: Great, thank you. So currently I think February twenty six is available, but then March April or not there yet. Question thirty four. I wonder if we have options for downscaling this grazed data so it can fit for analysis in the archipelago, and whether there are some approaches to mitigate the leakage signal problems on the edge of the island. I think tot total is answered this question, but meant, if you want to say anything more, yeah.

Speaker 3: Again, I mean, unfortunately, there's just not a whole lot we can do to get a good result. You know, Grace observation over island where the ocean signals leaking in. I mean, it's just it's almost impossible to mitigate that effect. So it depends on the side of the island. Of course, you know Greenland, you know it is a huge island. Then we can monitor changes in Greenland. But if you're talking about you know, the Philippines, those islands are going to be too small for Grace to resolve.

Speaker 1: Right. Next question is can be interprete water equivalent thickness data to identify each in regional wet and dry periods influenced my flood events.

Speaker 3: Yeah, I mean so so if you have a flood, I mean, you use there's there's more water there than than average, and so you should see you know, a significant positive trust your water storage anomaly. But it all, you know, it does depend on what we're talking about localized flooding. You know, if it's just if it's a small area, then that then then it might be too small for grace to resolve.

Speaker 1: Uh.

Speaker 3: And if it's something like flash flooding where it's really just you know, heavy rain and water ponding on the surface temporarily, that can be less of a signal, you know, less detectable than something like you know, continued heavy rains or what happens after a hurricane where you've just had you know, many many inches of water.

Speaker 2: Rained down.

Speaker 3: Yeah, so it's it's yes with a with a caveat there.

Speaker 1: Thank you. Next question, Please correct me if I'm wrong. But since this is a monthly average data set, it may have limited usefulness for water logging analysis influenced by surface water and groundwater flooding, especially in a country like the UK where rainfall events are frequent and highly variable.

Speaker 2: Yeah, that's true.

Speaker 3: I mean it's it's monthly, so so we don't have any direct information on what happened within a given month.

Speaker 2: It's only the average over the course of that month.

Speaker 3: And I'm going to talk about GLS two point two again. You know, GEOS two x two with the data simulation and the corporation of other higher spatial and tempered resolution data sets, means that we can look at some monthly variability.

Speaker 1: Yeah. The next question is, I'm not quite sure which CEA basin to use for California. I believe this is about the exercise. You're actually not looking at any basin. You're looking at the state of California. Just look at the map to answer the questions. So that's basically you're not looking at time seed is you're looking at the map.

Speaker 1: Question thirty eight. Given the course resolution of GRACE data, what is the best way to apply it to small basins? Is it better to use mask on products or combine GRACE with hydroologic model.

Speaker 2: Well, but you know, it depends on the definitive of small.

Speaker 3: So if small or is smaller than about one hundred thousand square kilometers, then you know, if you just average over the the the grace grids that are available on one of the one degree or cattery products. Then then you can you can come up with a time series, but the errors are gonna be very large because there's gonna be a lot of leakage, and I don't recommend doing that. It's you know, it's not very meaningful. So so the integrated products with the hydrogical models like GLS two point two are going to be better for that sort of application.

Speaker 1: Thank you so much, Matt. We are almost at the end of the hour here and it's the end of our webinar time. There are questions remaining that we will be answering on the Q and a doc and post on the website in a week or so. So we really want to thank you for returning to this session and we hope to see you on twenty eighth of April, where doctor Dorot and will again be talking about held US two point two with quarter degree resolution to look at daily groundwater storage data. And with that we want to thank our set team for help, especially our coordinators Anatasha Johnson and Jerry Morris, our helper and editor Maria Marabito, Jonathan O'Brien, Arrakashell, our coordinator Prop Levins, and our our set team Sean McCartney, Rica Podees, our instructional designer Susan Monty.

Speaker 1: Thank you also very much for your help with this webinar, and we hope to see you on twenty eighth of April at the same time. Thank you. Thanks Matt very much for your presentation answering one of the questions

The Story