FOSS4G 2021 - Integrating Remote Sensed and Modeling data for Local Flood Prediction and .....

good excellent greetings everyone thank you for joining us so uh today for um but i do believe uh is the the first collaboration between the group on earth observations and fos4g and in this particular track which is uh um open earth observations or open eo it's a pleasure to be your moderator today and without further ado since we're running a couple of minutes behind schedule i want to just jump right into our really exciting talks that we have today and introduce our first speaker dr albert kettner will be presenting on integrating remote sense and modeling data for local flood prediction and risk assessment dr ketner is an associate research professor at the university of colorado he is assistant director of the institut at the institute of arctic and alpine research dr kettner received a phd in civil engineering and geosciences in 2007 at delft university of technology in the netherlands where he studied local and global fluvial supply dynamics to the coastal zone he has been involved in dfo or the dartmouth flood observatory since 2011 currently is housed at the university of colorado and is since 2018 the dfo director dfo observes cert observes surface water changes by utilizing satellite data for dfo he designs and implements gis frameworks to automate free available dfo product distribution for global inundation maps and remotely sensed water gauging stations give me one moment and let's bring on dr kettner greetings albert please uh please take it away thank you david thank you for the for the introduction and thanks everybody for attending so let me start sharing my screen [Music] okay and share so if all works well you should be able to see my screen so i'm gonna put on the slide so there we go all right um so my name is albert gattner and i'm gonna present you a study that we have worked on for the last several years and that's obvious not work that i've done all by myself many others are involved and i'm going to mention a few of them so gee shuman and he will be a speaker later on in this session he's from rhs hydro was was involved in this study as well as bob breckenridge who is the founder of the dfo flood observatory michael shuffle fritz polisselli then slayback they're from nasa goddard bob edler from university of maryland and and many others so and what what i hope to uh convey to you all is uh so that's kind of the take take home message um is that you know that there there are many sources um regarding uh flood information for the flood data sources but so far there hasn't been global efforts to to integrate um remote sensing data model simulations and and provide that um in a global manner to the to first responders or to to countries that often have to deal with uh flooding so um i'm going to show you a platform that we have worked on to make this available um and it's not only so this platform is provided through um through website but it's also on a mobile app and i'm going to show you more about the mobile app but it makes use of ogc standards and so those feeds those data feeds can actually be brought in local gis systems so all right um before i start though and and showing some of that uh i want to make you all aware of an another conference that's happening in you know starting in a week that it's spread out over four wednesdays wednesdays uh two hour long sessions each session has four speakers and it's october 6 13 28 and 27 and the link is below here you can just google gfp globalflow partnership as well as the jrc who is who's hosting it so gfp and jrc should bring you to the site or just email email me and i will shoot you the registration links um okay so back on floating so why are we looking into floating so much um well here you see uh the distribution of natural disasters for a 20-year period and as you can see that over 40 percent of the natural disasters is actually caused by flooding and every year more than 20 million people are actually affected by flooding so over 40 percent there are roughly over that period there are seven over 7 000 natural disasters so 3 000 of them are due to flooding and then the gdp the cross domestic product is on a yearly basis affected by 100 billion dollars so you know being able to mitigate flooding um would not only save lives but also would have financial impact and um in the in the near nearby future most likely uh the flooding is going to increase so everybody is probably aware of let me see if i get my mouse here of this diagram of how temperature change global temperature is is changing right um it's getting warmer and warmer and that means that um the the the air can hold um more moisture so it's a little bit harder to see a trend in uh in precipitation but nowhere came up with this this trend you see here um for the continent the the usa continent where they see a increase in in precipitation for the last 100 years so and the world resources institute estimated that by 2030 so so less than 10 years out um over 50 million people are yearly impacted and what i said earlier about 100 billion dollars a year is uh is impacted by the flooding is it they're estimating that to be 400 billion and to look at just europe by 2050 there are actually studies indicate that um there's a five-fold increase in flooding so the flooding we see today might uh there's there's strong evidence that that's flooding increases so having observations forecasts having good platforms uh available only makes sense now it's not only um that climate change will increase this right and i think actually humans are are causing more flooding uh we we're artificially managing our watersheds our river systems and there are many cases i'm just picking out one of the flood fans here in in kerala in india 2018 flooding that caused close to 500 people's life um that flooding was due to uh was during the the monsoon heavy rainfall but it happens every year right but they were not anticipating that so much rainfall would uh would fall and that made that uh many of the dams and reservoirs they they had to open their uh gates basically otherwise it would damage those dams and and that caused most of the flooding there um if you look at uh you know where where are the most uh flood defense occurring then the top 15 uh countries that are listed here on the lower left side they're they're good for eighty percent uh of the population exposed to to river flooding uh so most of the flooding is actually happening uh in asia that that's the top three here right in india bangladesh china and vietnam so although those that region might be good for for eighty percent of the people that are affected flooding is really a worldwide phenomena and the movie that you're seeing here is uh an archive that we're putting together since uh 1985 um and it's it's basically it is so this is based on new sources like uh you know the new york times or the uh any any big news papers that are out there and we're we're scraping those newspapers on a weekly basis and this was started by bob breckenridge back in 1985 and we have over uh 5 000 of such event now in our in our archive and we're listing you know when does it start when does the flood ending how many people are affected or displaced and what what area is is affected by the flooding so so as you can see from this map it's not just acr flooding is happening everywhere so some of the flop products that is currently available so and this is just a flood product that that we make in in collaboration with nasa and with uh much of nasa's support actually uh our flop maps and and many others are actually generating those maps so that the map you see here on the on the left side that's a a map of uh river in india and the light blue or dark bluish that you see here that's a historical flooding and then on top of that is the more uh current flooding in the in this case it was a flood event i think in in 2019 and so you can see that of the red area is the the the more recent flooding that happened there it's not something that um didn't occur uh earlier uh you know it's we met that in a in the last 20 years that the area has been flooded and have has been more significantly flooded in the past so um so that kind of information you can get from earth observation data and then based on that you can actually also get a discharge signal and i will go back into that later so we'll get back to that but you can see here that we can measure using satellites we can measure uh these churches all right some of the other flood products or or refer to products that are available is here the usgs for america the discharge stations and they actually have a product i think it's at least daily updated if no if not more frequently where they show the status of their discharge stations and um and as you can see here the the the dark blue and and the black areas are where where the discharge is very high so most likely a flood of event is current there or close to a flood event there's also um model simulation products out there so i'm showing here two examples uh one on the left uh is from the global flood monitoring system uh that that's made by uh the university of maryland uh led by uh bob adler and one who um and they have a i think it's a one kilometer product where they simulate uh near real time as well as a forecast of the flood extent as well as flood depth uh similar initiatives are going on in europe and that's what you see on the right here there's a they've put something together called glowfest the global flood awareness system also that has a near real time as well as a long-term flop forecast that goes out several weeks this is an operational service it started three years ago or something to be become operational um it's funded through copernicus but uh the joint research center as well as ecmwf the weather forecast center are operating this and this surface and this this data is openly available for for people for emergency relief agencies as well as for the public um and then there are simulated products that are available within a country right so not those the the systems that i showed earlier are kind of global systems or or regional focus systems and the the the the simulations in countries you can think of the the work of within the usa of the work of fema where they produced the 100 year or one in a 500 year return period maps that you see here on the left which they did for many of the urban areas um you can kind of map out the areas that are most likely be affected by a one and a half year old 500 year defense and noah also has and that's the image on the right hydrocrafts with indicating river discharge stations where they the current discharge as well as a a week out forecast of what they think the river discharge will become there so there are inc in country um [Music] how do you say that the observational data sets as well as global data sets and there are in-country simulations as well as global simulation kind of products already available but all on different platforms so and then on top of that there are commercial entities uh that are working towards um making flood products available right um so there is uh the image on the right there's flood tax there they're very futuristic in using uh social media to detect where you know they analyze the twitters of the twitter stream with the global twitter stream to identify flood events where people are twittering a lot about floods they are analyzing that data and and putting that in a map and you can they have an api you can subscribe to that and then there are commercial satellite entities like the matsar maxar sorry the digital globe named previously or planet and they they typically go for optical data and there's eyesight and capella space right there and then there are more the the radar data sets that you can get so there's an open source kind of data stream as well as a more commercial data stream available all have their own platforms and and data availability but what wanna what one disaster relief um agencies uh what do they need right because they they hear okay there is um a flood event going on in in a country how do they need to respond when when do they need to respond so they rely on data of what's made available within that country but also from other sources so typically and this is a slide i borrowed from the world through program and they they say so so there's a timeline here on the xx and the information kind of information uh necessity on the on the y-axis um and they say when you know when it's when there's not much happening there's um there's they monitor uh various countries you know are there big storm events coming on what's uh what is the historical condition of the was there a lot of precipitation in the last several weeks or not so and that data did that comes in on a very much on the same level as what they needed so so they're good there but as soon as a as a flood event hits and that's that's this this orange star here then then the need for information uh becomes really huge they have to decide basically in three days or 72 hours do we need to respond if we have to respond how do we need to respond do we need to send in um you know medical supplies food supplies water supplies can we do we need to use airplanes do we can we come in by boats or can we use the road system still or so they need lots of information um what is the amount of people that are affected where are they affected in the country right well is it a flood event that happens every year or is it a once-in-a-lifetime flood event a one in a hundred year one in a 500-year flood defense is the flood just a kind of a flash flood or is it gonna be there for several days or weeks even right so all that kind of information they need they need answers to know how to respond so we created this um or we start to create this one-stop shop where flood products um are kind of integrated in a one portal people have access to to simulated data as well as uh observational data and so for simulated data you can think about you know forecasts or or more like uh flood frequencies for for certain areas you can get information from about that on the portal but also um you know observational data like discharge stations that are available on the ground or what a discharge that's actually measured from from a satellite and i want to dive in a little bit more about that satellite water discharge so here is on this graph the yx is um is are the years from uh so this is basically a century it's kind of hard to read maybe but it's from 1900 till 2010 or something the number of discharge stations and then and this is global but uh the jr dc has in their database on the left in the kind of blueish color is is monthly data availability and on the right is daily data availability and you can see basically that around the 1980s there is data available for many many stations but then there is a decline which means data is less shared by by countries and it's this is troubling because this is very useful information to see if a flood event is occurring or not right and this is kind of a same graph but then for the for the usa that that's our the the the filled in columns here and the in the in the open uh bars here that's uh gauging stations for canada and this is actually the number of gating stations that were taken out of um operation on an annual basis from 1987 on the left all the way up to uh 27th century less and less gauging stations data made available with hospitals in canada and this is a global phenomena and one of the reasons is because it's expensive to operate to keep those stations all operational the maintenance aspect to it is quite expensive so although i think everybody recognizes that you know discharge information is important there are less in the stations and data is less shared so we developed uh a system to measure what discharged from space and i won't go into it too much here but it's um there are several advantages and you can for example if a river starts to flood it it could wash out your gaging station if it's ground based right so but if you have a satellite you can keep monitoring your discharge and they're disadvantaged as well but i won't dive in that now so the way it works so on the left side is is a ground-based gauging station right and if the water level rises they're monitoring that in this this kind of tower and then they if you did measurements before about you know your your river profile basically and a little bit about stream velocity you can get up to a to a discharge well similar can be done with with satellites where we're not looking at stage height but more looking at changes in width so and the way to do that is with satellites like the am sorry that measures temperature and you can imagine water is colder than the surrounding land right so when water rises you actually your temperature will drop right and there's always cloud obstruction etc so that the signal doesn't look this nice but it's more like this wiggle kind of so you have to correct for that for the ground but by also taking a dry pixel into consideration to to to to filter out kind of the moisture what's in the air and get a real ground temperature and then you can kind of reconstruct that signal into a temperature change and therefore into a water change so this is kind of the same right um i talked about it briefly with depth and velocity makes up for your discharge well you can also if you got kind of your um if you got changes in width and you got model simulations actually that that simulate discharge for a river you can kind of calibrate based on that width you can calibrate your satellite signal to a mold simulation so here you see modeled in in red and here you see the temperature signal of the remote sensing and you can see it's matching up relatively well so you can constrain kind of a rating curve and based on that rating curve you can get from temperature to discharge and we do that for many stages on earth um for many of the large rivers and we have they them color coded there's a daily return visit to those discharge stations so then you can get to you know daily flow and you can say something those daily flows are reusing satellite data from 1998 all the way up to now so you can also say something about you know this river is now in a major flooding or not um as you can see here this is kind of daily data for 20 year-ish or something and you can see the discharge go up and down hi there albert just uh a note if we can uh yeah maybe draw to draw to some conclusions um yes thanks very good thank you okay give me two more minutes so you can combine this discharge signal with a float extent map right so that you can also monitor with observational data and then you can correlate the ones and you can get to return periods for those flat extents and let me skip this so then you have for example that flop frequency that the world food program is after you can say that when when you observe a float extent you can correlate it with the discharge and say something about that flood frequency so that that gives you that information so we put all that data in a mobile app that you can download for android unfortunately not yet but for the apple phone you can just go to your store and type in dfo flots and you you can get to the app we hope to to get something similar like this where it's really in a timely matter but right now there are daily kind of static maps that you you can't go back in in history this is how it looks and it looks like it's all flooded but don't worry it's not but red is where we measure surface water and when you will zoom in you will see actually tiny pixels not this this this totally flooded area you see over here and you can also include other data like i show you here on the right won't go into that i won't go into the time mode um so the challenge is to overcome global coverage we have that but to to integrate you know a spatially skills um so data not only at you know 100 or 200 meter resolution but also on a 10 meter resolution how to integrate that data is still challenging but we're working on that and the amount of various data sources and how to connect them is challenging as well and we hope also to say something about the uncertainty of the data sets and make that available through our web platform as well as through the mobile app and with that thank you very much this is my contact information if you have any questions or if you want to get this link on how to register for next week globalflow partnership conference let me know and thank you very much thank you so much albert really appreciate that um unfortunately we don't have any additional time there are one or two comments in the in the chat on the the venulis site so maybe if you uh you want to if you want to go there and you could address the one or two comments that are there in the chat um but we must move on thank you again albert our next speaker uh in this session is dr g schumann who is an internationally recognized senior scientist

Channel: FOSS4G

Views: 69

Keywords:

Id: _gzndGQD7ZU

Length: 30min 3sec (1803 seconds)

Published: Thu Nov 11 2021