Transcription
All right, so welcome everybody. I see we already have a handful of people online. Welcome to the GWEX hydroclimatology panel river experiment, also known as RIVEX. This is our um community meeting that we have every 3 months. Today, as usual, we have we have speakers, outstanding speakers might I add, will be talking about observations of of rivers today. We have um uh Simon Michel from the Global Runoff Data Center and MJ Turian from the University of Stuttgart. Um and I'm excited to see the story you have to share with our community today. We'll start with with you, Simon, and if you'd like to start sharing your screen, I'll I'll hand the baton um over to you. Thanks again, everyone, for joining. We'll be recording on YouTube and and recording the meeting and posting it on YouTube. So, if those that are not online in real time, are welcome also.
Welcome, everybody. So, um I will >> [sighs and gasps] >> start my presentation now. Okay, yeah. So, can everybody see my presentation, the slides? Yeah. Okay, perfect. Yeah.
Okay, so thanks you again for having me here. Today, I would like to talk about the role of GRDC, the Global Runoff Data Center, in the framework of RIVEX, delivering the observational foundation. My name is Simon Michel. I'm the head of the Global Runoff Data Center at the GRDC at the Federal Institute of Hydrology in Germany. Um yeah, so I start.
So, why in situ data matters. Long-term meteorological and hydrological observations are crucial in serving the needs of current and future generations for long-term high-quality climate records. In the slide below, you see daily values of river discharge from a station in Germany. It's called It's in the city Dresden and it has more than 200 years of continuously recorded river discharge data and it's still observational and running today, so we can assume that there are 200 more years in the future, but it's important to have >> [clears throat] >> these stations running. So, they are a unique source of past information about atmospheric and hydrological parameters, which serve as reference for climate variability and climate change assessments. So, quality-assured, globally consistent in situ river discharge data is often the foundation for analyses of climate change assessments and variability in terms of hydrology.
So, the Global Runoff Data Center, we have an official mandate, which means we operate under the auspices of the World Meteorological Organization, WMO, at the German Federal Institute of Hydrology. So, this is German's contribution to the WMO. We have extensive experience. The GRDC was established in 1988 to support research on global and climate change and integrated water resources management. Um we have a global database. We hold the most substantive collection of quality-assured river discharge data on a global scale. Marked in bold is quality-assured. In our data database, we have only verified data released from official sources like national hydrological services from member countries to WMO. And this is one of the main outstanding parts of GRDC. Here you see a map of all our stations. >> [clears throat] >> At the moment, we have around 11,400 stations all over the world. We have stations from 160 countries in our database and the colors on the map represent different end years of the time series. Um the bluish greenish colors are the time series which end in the last 5 years and you can see on the left-hand side the legend where you can have a closer look at what the colors basically mean. So, there are a lot of sta- stations which end in 2021-2025, but there are also a lot of dots in the map which are white reddish in color, those are um data delivery basically stopped in the past.
Um some statistics. On the upper left panel, you see the downloads from the GRDC data portal in 2025. In total, it was about 12,100 downloads in 2025 with a mean of 1,100 downloads approximately per month. Um on the both maps, you see the one is a country of users on the right top panel where you can see a diverse um >> [clears throat] >> um picture because science is spreaded all over the world, so users are basically also spreaded all over the world. And the download by countries, this is the lower panel map. It shows which of the data is most often requested, basically. The darker the color, the more this data is requested.
So, um yeah, the availability of discharge data in the GRDC data base is depicted in this picture where you can see since the 1980s or 1990s, a slow but steady decrease of stations, which does not necessarily mean that these stations are not functional or operational anymore, but the data is not transferred to GRDC. Um the last 5 years, this is not a real decline in number of stations. This is only due to the time the countries need for data verification. This is a lengthy process which needs to be taken caution of, so data delivery to GRDC takes between 1 and I would say approximately 5 even more years until the data reaches GRDC, but then we are sure that the data is quality-controlled.
So, the GRDC has a user-friendly data portal where the user can navigate world map. Every one 11,400 stations can be downloaded. Um the stations can be downloaded one by one. You can select different regions. You can select different subregions and so on. Um at the moment, we have some technical limitations allowing unfortunately only 700 stations to be downloaded at once, but this is also in in the makeover, so expect in um yeah, some new development in terms of data portal, but this is in in the next future plans, basically in 3 to 4 years, basically. But at the moment, we have the data portal. The data portal is working. Everybody can use the data portal and can access all the data which we have. It's based of daily and monthly data.
Um the GRDC has a lot of dedicated project data sets. Some to mention because they are also connected with GWEX is the BALTEX data set spanning the Baltic Sea. And there is a GRDC Caravan extension available, so these data sets are out there and can be found navigating to the website of GRDC. Um so, GRDC is a well-known source of river discharge data and the data for from GRDC is widely used with more than 100 publications over annually for the last 8 to 9 years, which can be seen in the chart on the lower panel. So, this is quite an impressive achievement of GRDC, so the data is widely known and is readily used in science, basically. Um marked in red with a red circle, we also have a collaboration with University Stuttgart and MJ and colleagues. This is a space-based discharge, so this is also part of the dedicated project data sets of GRDC.
Mainly what comes out of the collaboration with project is different publications or different data sets. One is the GRDC Caravan data set or as an example I did show here is the BALTEX data set. This is a data set where the GRDC provides original Baltic stations plus or in addition or GRDC stations would comprehend to the original data set and increase and extend the original data set, which is quite impressive because we have a lot of stations draining into the Baltic Sea.
So, um bridging the gap, there are more working groups Yeah, finding that GRDC stations are declining. One of the collaborations with in situ observations and Earth observations is the work from Elmi et al and Sameian et al, where MJ is also part from and here University of Stuttgart used Earth observations to gap fill and extend GRDC stations where data is not flowing anymore and you see the decline on the left panel, which is the same picture basically as I did show earlier and with the help of the Earth observation data they were able to extend the GRDC stations plus remote sensing discharge and the whole data is available um for download and you can see the red line marks the available stations in this data set.
A closer look to the data availability, the most recent year Basically, GRDC works together with WMO for the state of the global water resources report. In this report from 2024, the data from the last year is included. This is about 2,600 stations and GRDC has two different mode of data import basically. So, there are countries they have manual deliveries. The data is delivered to GRDC via email or other means of transportation and these are 635 stations which have data from the last year and the reference period of 1991 until 2020. The catchment size is marked in red on the map on the right panel. The catchment size is 32 million square kilometers. GRDC is able to connect to countries providing the data via API. At the moment, colleagues from GRDC developed a new package. It's called Hydro Downloader connecting to more than 21 APIs worldwide, but in this report and covering the last year is data from 10 countries with 1,600 stations. The catchment size is also 34 million square kilometers marked in blue in the map here. And the remote sensing part is 412 stations and the catchment size is around 85 million square kilometers. So, area of GRDC stations in collaboration with University Stuttgart is covering more catchment area than the in situ data, which is quite impressive achievement. The data set from GRDC covering the last year and the reference period and only the reference period is published at Zenodo. So, 1,800 stations can be downloaded using the Zenodo link covering the reference period and the most recent year.
So, coming to RIVEX, what's in my mind to bring to to to RIVEX. So, I compiled a special data set, which is for me the RIVEX data set. It compromises or it comprises of 5,300 stations from 57 countries. The mean record length is 67 years and 811 stations have data with more than 100 years. All stations cover the core period of 1991 to 2025 and in this map um the colors represent the time series length basically. So, blue are the long time series and the more reddish colors are the short time series. So, this is basically what comes to my mind when building the foundation for the RIVEX project. Um Yeah.
So, this is but not only an achievement by GRDC, but it's a data foundation which we hopefully build together. So, the RIVEX data set is a curated quality assured in situ river discharge data set. All stations have data at least until 2020 or newer. The whole time series is provided. In the WMO peer data set, I only provide data from 1991 until 2020, but in this data set the whole time series is provided. The data provision at the moment is via the data portal. We probably should or can discuss if there is a possibility to make this data set internally for the RIVEX community available as a complete data set where all the time series are included as a whole download. So, basically a one-stop whole data set. So, there are a lot of white spaces. I just go back to the map here. You see a lot of white spaces and this is where hopefully MJ can contribute to and the Earth observation working group can also contribute because there are a lot of white spaces. We know it and this is what comes to my mind would would be beneficial for GRDC is a further collaboration with the Earth observation community and the collaboration with the modeling community would also be great output out of the RIVEX project here. As the last point, I have a wish list of GRDC, probably a credited runoff product as a joint RIVEX output and yeah, now we can start discussing the GRDC data and the RIVEX effort and if the community at this point in time needs other data or even more stations and so on and so forth and what's going on in the GRDC technical stuff and so on and so forth. And this is basically the end of my presentation. I would say thank you to Claudia Centis and Henning Pleßl because without them, the work of GRDC would not be possible and thank you for your attention and questions are always welcome. So, thank you.
Simon, that was that was so great. I I I have to say I had a little bit of a moment when you showed that map and you said GRDC RIVEX, that that's awesome. It is it is really good. Oh, and hold on, you're sharing your screen. I'm seeing myself multiple times now. Um Okay. Yeah. Oh, here we are. Question of clarification and and also if I haven't said it enough, thank you, thank you, thank you, thank you for being willing to to join us on the adventure. I think one of the words in your title was fundamental or essential and I totally agree. Having the ground truth that that your team is able to provide is is essential for for the work that we're trying to do. A quick question of clarification. I love the data set that you're that you're proposing and and so my understanding is is you said there is data daily data available for all of these stations at least between 1991 and 2020 included. And then there is some more. Is that is that accurate? So, so we know we have a is it a 100% full daily discharge data for at least how many stations you have? A thousand and some? 5,000? No, it's it's 5,000. Yeah, so so some some stations have only, let's say, 1 to 10 years. So, they are not covering the 1991 until 2020 period, but these are only them basically. The majority of all these stations cover at least 1991 until 2020, but the whole time series is provided. So, one station is, let's say, 200 years. So, this station is inside and it has a time period. So, this is a massive data set. Yeah. It's all in basically. Yeah. I think this is outstanding and you know, I mean, when we had a study a couple of years ago when we looked at a different time period, but I think it was 1980 to 2009 and we did our best, but we could only find something like maybe a thousand stations that we cleaned up and looked after. Like 5,000 is almost an order of magnitude greater. I mean, it's nuts. It's really great. So, I love it. Any other comments from the from the crowd? Yeah, I wasn't sure if the the floor was open for questions. Can you hear me? Yeah. Yeah, cool. You know, thanks Mark for the the presentation. That's a very informative and yeah, you as you showed like we all use GRDC one way or another. That's like probably the most popular data set among hydrologists these days. Yeah, but I just want to clarify something cuz you know, like very often very very often I see publications and presentations and you also mentioned at some point that uh the No, data people say that oh, you know, gauges are declining globally, but you know, that and and they use often use GRDC as the reference for that. The the graph that you showed. You know, and I I think that becomes some sort of misinformation spread because you know, we don't really know what's you know, what's available out there. You know, what what The one thing is what you have access to or people make it available. The other thing is what you know, the governments make available. So, it's stating that you know, that there's a declining in uh uh you know, in in monitoring it's I think that we don't know that. But I we do know that a lot of countries, for example, Egypt, India, they don't openly provide the data. For example, you showed Uh if you look in in you have some data in Egypt, but it ends in 1990. They didn't stop monitoring the rivers in 1990. I I mean, I'm pretty sure that the the Nile was pretty essential for the livelihoods for them to not have a monitoring system, but even though we don't have that. So, I I just want to be you know, just uh state out there for whoever is you know, attending this meeting to do not state that you know, it's the monitoring is declining because we don't know that. You know, that's and the data availability from GRDC doesn't mean that it's all we have out there. It's a fraction of what is available. It's actually could be available. Yeah. That was that's my comment.
Yes, this is a great comment. I'm and I'm always a bit cautious with stating that stations are declining because there might be different reasons. One reason is stations can break down. Stations but data just cannot be transferred to GRDC. As an example, a lot of Chinese researchers are asking me if I have Chinese data. Normally, I would say yeah, but normally I would come to you and ask you if you have any data for me. So, this is uh political reasons, transboundary basins are often problematic, dams and uh basins Yeah, you know, everything which is covering water extraction and so so, countries are really cautious when they share data basically. So, and there are regions where every data is shared via API. So, there are regions which have really great covering of the coverance of data and some regions are poorly covered. Yeah, but I agree that most of the data it is somehow available, but it's not shared. So, and we don't hold any data back because we got often questions, ah, I need this and that river. Do you have data? And yeah, we you can only go to our data portal. We don't have a secret database. So, all the data which we got from the countries we are sharing on our data set database basically.
So, Augusto, that was a great comment. I agree that that semantics are important here. Most serious publications on the topic talk about declining data availability rather than in the in data or in or in monitoring. But for the story, I I made the exact same mistake that you did. The first time I made it was at a GWEX meeting in front of Ziemon's predecessor and Uli was not particularly pleased with my presentation when I said discharge data is going away. But since then, I've learned to use better words. So, let's all be careful as a community. Anybody else has questions for for Ziemon? Gee. Uh yeah. So, thank you. I mean, great presentation and yes, I'm one of those people that have actually used GRDC also a number of times. Um yeah, not only to look at gauge data, but to look at um statistics as well, but I had one question for you maybe from a very different angle. If you because this is a huge a huge task, you know, keeping this database alive and making sure that you still get the data from those that may have provided to you last year because I know being uh sitting in Luxembourg, you know, where you would think well, the gauge data that we have we have not a lot of gauges in the country. It's very small. But you would think it's very easy to make that available open. And we do have a very strict open data law. But yet, the gauge data on rivers are not easily available. If you go back and request now from 2021, the very disastrous flood event, you would see no data available publicly even in Luxembourg. And then if you request the reason, they would state well, not really calibrated yet. The stage discharge curve is not really great. We can't release that data, right? That's been going on for years now. Um so, I know how difficult it is even in your own country to get data that should be open. So, congrats on keeping, you know, this alive. But and and as we all know, this is really important for hydrology, but I have a a very silly question maybe, but what's the biggest challenge that you have encountered making sure this is kept alive? Is there a single country where you were surprised of what's what's the biggest challenge that you had to face? Is it more like storage? Is it making sure they understand that they should keep feeding the data to you on a yearly basis? What is it? The Is it technical? Is it political? Is it what is the one biggest challenge? Thank you.
The one biggest challenge of my my whole working for GRDC is basically finding the right contacts. So, basically there is a list of contact points, hydrological advisors from WMO. >> Yes. And I can contact everybody from this list, but there is often a response and then it's okay, but when there is no response, you don't know if the email is to the correct person, if the email is to up This is one of the major prob problems to to get the right contact points basically, you know. Yeah. So, yeah.
Thank you. I mean, this this explains also why I can imagine that it's so difficult to keep it alive, right? Because you may maybe the contacts that you had last year are not the same for this year and will change next year. So, yeah, again congratulations on this and and thanks for answering my question. I'm always interested how people that work on efforts like this are actually making it sustainable. You know, what's the biggest challenge? That's my question. But thank you. Thank you. Amazing presentation. You're welcome. Are there questions for Ziemon?
What I can contribute to that is that getting data one time is is is hard sometimes, but getting data continuously for the next year and the next year, this is the major challenge we are also facing at the moment. And one of the main door openers at the moment is the WMO report because everybody wants to be part of the WMO report, but we as GRDC we asking the countries for an extra effort because the data needs to be ready basically now, but verification process is often lengthy. And this is something which puts extra workload on the countries and we are always very thankful to the countries because they are delivering data which is not a given basically. The example was Luxembourg. I had problems with Luxembourg data because the WMO said allow only 10% missing data in the reference period. And the Luxembourg data start in 20 2001. So, one whole year basically is missing. Per definition, this data should be excluded. But there is no missing value whatsoever. So, this data this data set is quite perfect basically. And when I go to Luxembourg because they did a lot of extra work to provide the data to me and I said no I cannot include your data because this is uh not complete um they would stop delivering data to GRDC so complete basically. So this is something you have to be flexible when it comes to data and data availability and percentage of missing values and stuff like that. So this really depends because the countries are doing a good job and they need to be honored and say thank you to the countries because it's most of the time it's the countries work which supports GRDC to be able to provide uh data to the user basically.
Uh let me this is really interesting. Perhaps let me ask a final question before we transition to to MJ. In your um presentation you suggested that in occasions you you share data sets on Zenodo and in others uh the primary location for download would be the GRDC data portal. I I don't have a strong opinion for what we should do for RIVEX um apart from you know having somewhere an official list of the stations and perhaps the expected number of of daily data points that could be downloaded so that you know even if so if we could share the whole thing on Zenodo I would applaud that. That would be wonderful. But if that's not allowed for a variety of uh you know licenses or or anything that stands in the way having a place somewhere that says this is the official list of the 5,000 and some stations and when you download everything from the data portal you should get I don't know 22,601 uh daily data points. Um I think that would be tremendous for our community. Uh and but but I suggest we we have a brainstorm about that with the entire community when we like propose a plan to everyone. Um does that does that sound okay to you?
Yes this is basically what I was thinking. Yeah. Uh wonderful. Well thank you again Simon for your excellent presentation and for and for your willingness to engage with with our community. It's Sure. Of course yeah thank you for having me here. And thank you.
Hey MJ Yeah. >> [clears throat] >> Okay so let me share my screen. Um That's the screen. Let me go in presentation mode and let me activate the Okay. So I I've taken a rather very general title Rivers from Space which is actually connected or motivated by the lack of in situ data which have been talking about and being because of political issue or availability of stations. Um so and I've narrowed down the presentation on two satellite missions specifically altimetry and imaging satellites. And when comes to progress and challenges you can talk about it for couple of hours so that's why you see all my slides are prepared for as as a at a very high level. So don't expect any anything technical here so rather we stick on the upper the surface.
Um Yeah as I said we have been talking about the in situ availability and and discharge but we have not discussed the water surface gauges. I think this is also a very important point. We have we had a lot of discussion with Simon here whether of course for many of the discharge gauges there are water surface elevation data available uh and and are taken out of rating curve rating curve and was provided to GRDC and and the question we had always whether these data are available or not. Uh and of course the response from Simon and and others are that these this is diverse. For some of stations these are available for some not. And this is challenging. And I would say if we stick to if we want if our aim is to monitor rivers uh finding out water surface elevation gauges are much more difficult than the discharge. So um with the help of Payman and a PhD student here we have tried to gather as much as water surface elevation gauges we have globally and that's what we have obtained in the past couple of I would say years. So these are in total 6,000 gauges that are publicly available and some of them are we are not even publicly available. My student had to buy it in from from somewhere. So in principle uh this is actually I would say one of the challenges of monitoring rivers globally. So the water surface elevations are not available.
Um but of course there are alternatives and that's why we have these satellite missions the the standard and kind of very well known satellite mission is satellite altimetry. The concept is simple transmitting the pulse pulse receiving the pulse back uh through travel time you measure the distance and then you know the position of the satellite then you have the height of water surface. Um and of course since 2023 we have the SWOT mission which I wouldn't talk specifically about SWOT because I know that Guy will talk about SWOT in in the next in the next RIVEX presentation. So we have uh altimetry missions many of them are radar or actually predominantly are radar and they are designed in the nadir concept. So that means the cell the pulse is transmitted toward nadir. So that means anything and not in a sort of swath way. So anything kind of profiling method. So anything beneath the satellite will be monitored. So that means that's a matter of chance. So here I've prepared sort of let me check if this works. Um kind of slight animation. So Here you see um I just activate different satellite on the Hydrosat website. So I Topex Poseidon and and then Topex extended Envisat Sentinel A Sentinel B and Cryosat. So you see if the satellite passes through certain rivers or lakes here we zoom in on Lake Constance. So you see we have couple of couple of tracks over the Lake Constance for instance. So that means if you have a satellite passing over the lake or river so here we see a lake but the same if you have a river the satellite passing over so there is a chance that you have water level time series from altimetry. And of course this for that you have different sources starting from C C said Geo said but then effectively those satellites that are that can be used to generate water surface elevation over rivers would be I would say from Topex and then and then Envisat Jason 1 Jason 2 Sentinel 3A and 3B which are SAR based and there are loads of technical uh information behind which are very relevant in terms of capability of the satellites for monitoring rivers but as I said you wouldn't I wouldn't really go deep in the technical information but the bottom line information is yeah there are nadir profiling satellites that could be um that that could generate water level uh time series for rivers.
Um and on the top of that we have the uh satellite imaging. So I mean and this has been the this is probably one of the oldest remote sensing or space based uh system. So basically we have optical satellites and some of them are or many of them are in sun synchronous orbit. You typically take kind of like as you can consider picture or image of the um uh Earth's surface and over rivers essentially you can you can monitor the river extent or surface water extent and of course river widths and and and that could be a very helpful information if your aim is to monitor rivers from space. The same of course we have SAR imaging satellites with a different concept of course radar but you can also use it in a kind of interferometry concept um the benefit of that is yeah you can do some more analysis over rivers than than classical optical. But information is you have we have a lot of satellites monitoring both water surface elevation and extent. So the question would be and of course these are you see starting from 1980 for altimetry from 1992 with Topex So, the question would be how much of how much of the rivers are monitored with all these satellites? Of course, if you want to do that, of course, you have to look at level three data or data products that are producing water level time series from altimetry. And here I've provided a list. So, I mean the many of us who are in the community, we know the website like Hydroweb or the River and Lake initiated by ESA a long time ago, the Dahiti website, our website Hydrosat, the G-REALM, GRATS, Altex. Uh so, there are a handful of level three databases providing water level time series for rivers and lakes, of course. And here you see a kind of roughly the coverage so from different data sets. Um which is published in the SAIM dataset which was mentioned by a Simon and lately. Um so, you see it's more or less giving a good picture, but probably if you zoom in, you don't see the entire river system. In terms of surface water extent, the number of datasets are less compared to altimetry, surprisingly. Um so, um we have in Hydroweb the area time series for lakes in Dahiti as well. And the I would say the only website that provides surface water extent for rivers is the Hydrosat website which which is maintained by us. And that's actually the surface water extent available in the Hydrosat website generated out of either Landsat dataset or MODIS dataset MODIS satellite. So, essentially, when I say when I talk about surface water extent for river, it would mean that you typically consider reach, you calculate the area area divided by length, that would be your effective river width. So, [clears throat] with that, I would say, yeah, you have altimetry satellites, you have imaging satellites. So, that means you can make a claim. The satellite observations are becoming the new in situ. Yeah, so I think we are getting [clears throat] into the age that we have to start we have to really start to think about this new paradigm. Why not? I mean, what's what's so um um different about in situ that the satellite could not provide. So, in that respect, I think this is would be a viable idea. Of course, there is a bot. That's not actually as easy as you think. There are challenges. One of these standard challenges is spatial temporal resolution or coverage. Yeah? So, we have been talking about these altimetry satellites and and and here it's a diagram showing the temporal resolution in day in logarithmic scale and also spatial resolution in degree degree on the equator. So, if you show both in logarithmic scale, then and then put the spatial temporal resolution of all altimetry satellites, they are lined up very well. So, that means for instance, let's look at TOPEX, you have a 10-day temporal resolution. So, that means TOPEX comes to the same point or any of the successor of TOPEX mean Jason 1 2 3 or Sentinel 6 Michelle Freilich or Sentinel 6B, they come to the same point after 10 days and they cover overall so the ground track spacing would be something about something about 120 km or so 100 yeah, something something about that. So, essentially, uh that means anything any hydrological phenomena that is happening in this I mean, if you want to understand the daily variation of the river, this line and many of the cell how to combine information along the river, the idea of densification which we have been also active on that. But still, if you rely on solely on one satellite, then you have to live with the temporal resolution problem. The same, of course, on spatial resolution. We have we have seen the profiling concept. So, that means if a river or lake is missed by one each of these satellites, then it's missed. So, your ground track is typically fixed and then Yeah, so and and this is actually kind of um a kind of teaser for SWOT. I mean, SWOT with its measurement with the InSAR concept can can deliver a very good spatial temporal resolution compared to classical altimetry missions. So, if you put the imaging satellites or I mean, I've picked some sensors here, Landsat, Sentinel 2, MODIS. So, you see the situation is far better, of course. They come with with a better temporal resolution and spatial resolution. So, that means in for many of application, probably would be far better to target the river widths rather than river height. So, I mean, you would be better in terms of temporal resolution if you target satellite images and generating surface water extent from that. And if you put that, of course, and with with some classical bubble like a flood and then rivers, smaller rivers or variation or flooding in small river happening in this area of spatial temporal resolution which are not really covered by these classical satellites. So, it means they are limited with it means if I want to kind of complete my claim that the satellite would become the the new in situ, they come they can become the new in situ with of course with the caveats that they are limited. The other source of limitation is of course for at least for altimetry satellites is their measurement concept. The So, as I said, these pulses sent and then transmitted back. So, and that is if you know a bit of radar, it means that the footprint is rather large. It's not actually you can't you can't send you can't hit a point, you hit an area. And if you hit an area, you you get a sort of return power which we call it waveform rather noisy if you hit as large area. And if you have that, the question would be which of these peak uh peaks would be my my targeted water surface. So, for instance, here when you have in your footprint couple of water bodies, this one, this one, and this one, you probably have three peaks. And the question would be which one you are you want to uh take for measuring the height. And and the algorithm to find this is is the so-called retracking algorithms. And there are loads of different retracking algorithms. So, if you look at this uh figure, you see there are different algorithm empirical, statistical, physically based, adopted for pulse limited SAR or both, adapted for ocean, lake, or river, implemented in level one, level two, or level three, and for different satellite missions, but I would say we are experiencing an age after about 30 years of altimetry, we don't have a retracker dedicated for hydrology. As simple as that. We still use the retracking algorithms that are kind of uh uh defined or developed for ocean, but then kind of used with a certain modification for inland application. So, we have we have tried to kind of deal with that problem. So, with this beast algorithm just to show how how could change the result. So, you see here this is a lake Val Beze in US, very small tiny lake. So, you see if you rely and and this is the in situ water level. So, just to uh highlight for you, if you rely on standard retrackers typically provided in geophysical data records, which I said, these are kind of ocean retracker modified for inland, so you get really large errors. But if you try to be careful with how to retrack the waveforms, you can actually improve the accuracy by a lot. So, Um, the future should be uh should go in that direction to uh improve um to reduce the noise in water level time series.
Um, also in terms of satellite images and extracting the uh um the river widths. Uh so, these are basically the work of Omid. Um, so people typically think that if you take a satellite image, generating river or finding out where water is uh is a simple task. So, you have a histogram and then you typically expect that you have a kind of bimodal histogram. You define a threshold and you say, "This is my water and and this is this is land." It turns out that it's not as easy as you think if you if you aim to monitor the dynamic variation of a river. This is just one example which Omid showed long time ago in in this paper. So, he simply So, green is in situ river discharge and and blue is water area uh obtained from four different thresholding algorithms. So, actu- I mean, if not simple one, rather sophisticated one. But still you see that when it comes to finding out the proper dynamic of the river, extracting the river widths is not so easy. Uh so, of course uh long time ago in his PhD, Omid showed that if you could benefit from special temporal information, things get get can get far better. And and here he means in principle that you kind of make the pixel your your in you make your algorithm informed about if a pixel isn't neighbored by a water pixel and and also neighbored by a water pixel in time. So, in principle, if you make your algorithm clever, then you can generate uh yeah, rather uh proper water river widths time series. So, uh just to mention if you have done that, then uh as mentioned by Simon, then you can generate your rating curve um and and estimate river discharge. So, both from water level you can connect it to river discharge or river widths connected to river discharge generated generated rating curve. This can be simultaneous or non-simultaneous. We have shown in these two papers that also from non-simultaneous time series, you can uh generate these rating curves and estimate river discharge which uh which are also maintained in HydroSat website and also uh shown by Simon.
So, um talking about RIVEX, in principle, whatever we have in our HydroSat website in terms of uh surface water extent, uh water level, river discharge, we would contribute openly to RIVEX and and uh uh discuss with uh uh Cedric about that. So, we have to discuss in which form, but uh these are Yeah. W- One one uh disclaimer, of course, that these are not all belonging to same time period. These are generated out of different satellite missions. So, um yeah. I mean, one can start think of how to harmonize all these data. But still um this could be something uh for future studies. So, where we are now um in terms of what we can estimate and measure over rivers. >> [clears throat] >> So, I would say river discharge can be um estimated through satellite imagery and imaging and there are many studies, many data set by now. And of course, by SWOT. Uh as I said, I don't talk about SWOT which will be discovered by Gil next time. The river surface height can be measured by uh altimetry as I've mentioned and SWOT. And the river surface slope, which is also fundamental for many of the studies, uh was also shown to be estimated by altimetry by Christians uh Christian uh and and colleague. And also um by laser altimetry, the ICESat-2, Daniel showed that uh they came up with this kind of global data set with uh for for surface slope. And of course, that is something that G- SWOT can also do that. What I wasn't uh what I did not mention is the new emerging techniques uh techniques of GNSS-R and GNSS-IR, the GNSS reflectometry and interferometric reflectometry that they come to the game for monitoring both water surface elevation and slope. So, this should be also discussed in probably one of the other session of uh RIVEX. The river flow velocity is something that has Well, uh people are starting to kind of finding ways to estimate. There's a paper by Nick Everard for kind of using videos to uh estimate river velocity. And that's an essential information. One point of time we would have global we would have global river velocity that could be extremely helpful for river analysis. In terms of river depths, of course, um when it comes to inversion methods, uh you can use uh SWOT, you can use satellite altimetry plus uh imagery to uh do the inversion and get the river's uh depth. But um depending >> to interrupt you. We're getting close to the top of the hour. I'm I I'm How many more slides do you have in Are you >> That's the last slide. This is Okay, thank you. Thank you. Yeah. Uh the uh in when it comes to the depth, laser altimetry could be also one option that it can penetrate. So, river widths I've mentioned and groundwater discharge, of course, you can't measure from space. And with that, I would like to thank you and I would like to acknowledge all my uh colleagues who contributed a lot in this presentation. Happy to answer your question.
Okay, that was so great. I don't think that's ever happened to me that I said, "When are you going to be done?" and it's going to be literally 2 seconds later. I apologize for the interruption. Uh thank you so much, MJ, for the uh for the excellent presentation. I I really appreciate a lot of the graphics that you put together. Uh tell a a nice story. I might I might need to recycle some of those. Um do I I have an immediate question which is, you know, how do you propose that we engage your team and the rest of the community to bring something to the river experiment? We we have this approach that, you know, those who do the work have the have, you know, the majority decision power, but we do want to engage as many people as possible. And you you've mentioned alternatives. There are other teams that you mentioned in many of your slides. I'm also hearing of new methods that I think paper was recently published. You call that, I don't know, the neighbor the neighbor observations or the you had a clever word for that. But how do you think >> Densification. Tell me again. Densification, you The densification. Yes, thanks. That stuff is really cool. Um so, how what do you think we should we should do as a community? You know, we don't want to be too distracted by a million data sets. That's that's for sure. But having a handful of representative data sets would be helpful. What's What do you think we should do?
Well, I think the probably the hardest uh task is to harmonize all these data sets. I mean, if I come back to this slide or this one, I mean, you have There are different available data sets and then one should sit and harmonize all of them. So, for many of these, I mean, which we call virtual stations, you have data set from different multiple sources. So, the question would be, do are they deliver the same answer or the same water surface elevation or not? And the other question would be whether we want to use all these data data set together because these are publicly available data set. You can actually do whatever you like with them. So, and then harmonizing in the sense of So, if any comes to water surface elevation, you have to understand whether they have the same reference, the same data, same jewelry, same So, all these kind of complicated uh things should be kind of standardized or harmonized so that you have a full-fledged data set that can be uh can make make can be made available for everyone. So, these would be I think there would be kind of task force a task to discuss and then do the harmonization of different data sets available.
Uh that's great. Can Can Can we look to you, MJ, as as, you know, one of the top experts in in in this field to work with the community that you're that you're showing here and and see who else might be interested in participating? I think we want to be welcoming of everyone who cares to to contribute and and also come up with a recommendation for for everyone. So, I I think our thing thing that we're trying to do is the as you know, one of the one of the topical co-leads for observations here to to have you guide us with what you think is the best thing we should do. And and so, I would encourage anyone on this on on this call here or watching later to contact me or any of the other co-leads of of Rivex or MJ and Zemon to engage with with these topics if that's something that that you'd like to participate in.
Um, all right, we're already 3 minutes past the top of the hour. Do we have any burning questions from from anyone before we close? I I think we're suffering from 3 minutes late. Everybody wants to go to their meeting. Hey, listen. Thank you both MJ and and Zemon for your for your excellent presentations today on behalf of our entire community. And for those of you watching, you click on the thumb, you click on the bell, you subscribe, you do all that business. We'll have the next presentation in 3 months. And in the meantime, we should hear from our co-leads on recommendations. Thanks again, everyone. See you next time. Thank you. Thank you. Bye. Thank you. Bye. Bye. Thank you very much.