#107 QGIS for groundwater applications

welcome to the australian water school the home of demand driven industry design training for the global water sector hello and welcome to today's australian water school webinar covering qgis for groundwater applications i'm craig and i'm excited to be your host today and to introduce to you the expert presenters that you'll be hearing from but let's start first with you our attendees who are coming to us from around the world looking at this map it's always an awesome sight to see the global reach of any topic but for today's topic in particular there's been a tremendous response to this and it really helps us you know seeing your response to these really helps us realize that we need to be doing more in this space whether it's qgis or groundwater or the nexus of both of those we do want to bring you more content that we'll get into here shortly so as far as today's topic we've got some experts on board here for groundwater modeling which to some of us is a bit of a dark art you know there's sometimes surface water expression that comes out of the groundwater sometimes you can see it when it's sucked out of the ground but while it's doing its thing in the ground it's pretty much invisible to us so being able to visualize it can be very helpful to someone like me who is used to seeing things on surface water so this is a topic that i think is great where we're going to use some free tools in some of our courses to be able to assess the data get it into our software and then be able to visualize it and these are the experts who are going to help us do that today so kurtz hans conrad come on board here um you can read their qualifications here deep extensive qualifications that you see from each of these and i won't go through that in detail but do you want to find out you know where you're coming to us from today and maybe we'll start with curt who i understand last time he did a webinar for us was coming to us from the us now this time you've managed to move overseas in the middle of a pandemic kurt how did how did that go yeah thanks great went quite well yeah i moved from the united states to denmark in january and i've started a new position with a great open source gis company named septima here in denmark awesome hans uh where where are you coming to us from and what time is it uh where you're at it's already morning here in rotterdam in the netherlands so uh i'm close to curt uh during this webinar first time i think that we have a webinar close together so we're looking forward to that yeah we've usually tried to balance the three three corners of the globe when we get the three of us on board but now we have conrad as well just say g'day where are you at hello everyone i'm in perth at the moment in australia it's early afternoon just after lunch it's great to be with you here and i'm happy to help if there are any questions thank you excellent thanks so hans kurt and conrad have been part of our previous webinars as well i do invite you to look some of those up they're numbered on youtube on our on our youtube channel i think there's number 69 was modflow which again has over 10 000 views along with our two qgis webinars as well number 84 and number 93. have a look at those now right from the start i need to put out a disclaimer though in this one hour session we're going to be able to touch on some of these awesome tools that will be available to you but if you want to do this yourself and really dive into it and get into it hands-on sign up for the courses and you see here a bit of a screenshot on some of the things that you'll be able to do with it starting just with google really these courses are available to you on demand right now you can go on online sign up for them we did them live and then we recorded them and all the content is available to you you'll be able to start with google and see that map there and with legends and line types and everything else be able to do the delineation you know look at the 3d terrain and get a catchment map in qgis and get your groundwater model as well so these courses are available for you do take advantage of that if you get a chance and give us again your feedback at the end when i moved to australia been over 10 years now my first exposure to the australian water school was groundwater courses that they were offering and i do get a bit nostalgic when i think of the live real live classroom setting that we had back then lately we've had a bit of a focus on on surface water and flood modeling over the last year or two and and it gets a lot of attention but uh we want to sway back and get into the ground water space here and really explore that some more so give us your feedback what do you want to see more of in this space if we have a look at those poll results before we turn it over i wanted to see where everybody's coming to us from what's which uh industries this is i think the first time ever that commercial consulting hasn't ruled in general um academic is is there somewhere in second or third place so we do have a lot in the academic space esri has come up as uh one of the most popular ones but this is a qgis webinar so as expected it's coming out on top one of the things i wanted to just get from the audience here is you know whether people who are attending have done the actual modeling or are we you know mainly from the gis space and it looks like less than half have actually done any ground water modeling at all so hopefully this will give you some of the tools to get into it some of the tools that are available to you are absolutely free thanks to the taxpayers of the u.s government who fund things like mod flow and hecarims and things like that looking at groundwater versus surface water and water quality versus water quantity we've got all four of those listed here it looks like groundwater quality has actually been getting more attention than groundwater quantity now today again we're talking groundwater quantity if you want to see more on the water quality side of things do put that into the feedback at the end so that's enough for me let's hear uh conrad uh would anything surprise you as far as the programs people have used for the modeling since you come from that background wow probably free software it's looking good i would say not only because it's free but i believe that the quality is has improved definitely over years i remember that for special analysis and mapping i was using different software before and now i move entirely to qgis and i'm very happy also when it comes to groundwater modeling my choices are offer related with something that's affordable because it's there is a larger group of community group that supports it and people support each other and that's that's great to see from my personal point of view excellent and what we want to give you here and give to the industry is resources that you can go to if you've got free software especially when it's government-funded software you might not have somebody you can call and you might need to get links to some of these resources where you can get in touch with the community of other users who have run into the same problem that you're running into all along the way so again from the name of my company you can tell i'm a bit of an outsider here to the groundwater side of things but i am excited to learn from our presenters today and really you know get these things going on interaction we've got groundwater people and surface water people but we want to be uh non-binary if you will here and promote that space in between the groundwater surface water interaction um because there is a lot of interaction between the two and uh the more the surface water modelers can take advantage of groundwater modeling and understand how that works and vice versa the better off the industry will be as a whole so we're going to start with hans if you can go ahead and start sharing your screen hans with no further ado let's kick off with some of the applications that you're going to use in qgis and then kurt will follow on with the visualization over to you hans excited to hear from you thank you craig thanks for the introduction so i'm hans from the quest i work at the iit delft institute for water education and we do everything that has to do with water so completely non-binary surface water groundwater management of water we're located in the netherlands and i'm there a senior lecturer in gis spatial data management so in the next 15 minutes i'm going to talk you through the pre-processing and collecting of groundwater data using qgis but also where to find data and when we talk about a workflow for pre-processing groundwater data we first need to look for the data we need data from boreholes wells that we can use for our models later that we often want to build a database their spatial data non-spatial data we want to join the non-spatial data to the spatial data so it can be also used in a spatial sense that can be for example groundwater quality or level data that is stored in separate tables then we need to sample sometimes data from other sources at the location of those boreholes such as elevation for example often we need to clean up the data and for our models we need to interpolate the points to to rasters or mesh or other formats and then clip to the study area that we want to specifically look at for our case studies and additional field data collection might be needed then we want to do 2d visualization to interpret the landscape that we're looking at and 3d visualization the last two components will be covered in kurtz presentation i will focus on the data pre-processing part so let's start with where to get the data i'll see that from this map that was shown by cray that you come from all places all over the world and if you look at global data then it's interesting to look at spatial data infrastructures that exist these are not just portals they have connections with gis and applications and an important one that you can look at is this one from igreg the global groundwater information system where you can find a lot of data from all over the world but there are also local sdis where you can get data from your country hopefully if they have one otherwise it's time to get started with one because it connects really to the data to gis into modeling and that's what i want to illustrate a little bit in the next slides so this is another sdi from the sodex gmi which has groundwater data for the southern african region and you see that it has a nice web portal it's based on geonote which is open source software and you see here this trans boundary aquifers of the world layer but the nice thing is it's not just a portal we can have a real live link with qgis and you'll see that here demonstrated where we make a connection to geonote that's the standard feature in qgis you can give it a name and then you just put the url of of the sdi test the connection and you see here that it supports wfs which means vector layers and it supports wms which are rendered pictures of the data so the connection works so i can connect to the sdi when the connection is established i see here all the data sets that are available with the type of web service that i can use and i'm looking for this trans boundary data set the aquifers and i choose their wfs and then i simply get all the vectors in qgis that i can further edit it has an attribute table it's just as vector data that you're used to when you work locally now since uh we're back here in a webinar for the australian water school we zoom into to australia this is a national groundwater portal it's the australian groundwater explorer and here you find a lot of oral data for australia and we're going to look at this one and at a specific case study where i'm going to apply this uh this workflow uh it's the namoy river and what we're going to try is to to define a study area for the alluvial aquifer the shallow aquifer around this river and if you go to the australian groundwater explorer you can look for rivers so we do that namoy river and you see that it supports different formats it doesn't have this live link like geonode which i demonstrated before but we can download different formats there's a sv file geodatabase but that's not so easy to use for us and we would like to work with specific data that we set up in our own database so we're going to make a geopackage based on this shapefile that we download when we download that so it will send you an email that you can download it and then it will give you access to these files and it's always good practice to in windows if you're a windows user to check this box file name extensions because then you'll see that this is a csv file although it has an excel icon and it says here it's a micro microsoft excel comma separated file well it has nothing to do specifically with excel it's human readable text which is stored in a csv file and you see that there are shape files which you also understand hopefully that it's not just one file that belongs to one shape file but it's a set of files so we see here some different data another good practice advice is to always check the readme files that come with these kind of downloaded data and here we see that there are water levels in the csv some water quality data which we will not cover today and that the shapefile contains the the boreholes and there's some other data so we're gonna combine all these separate files into a database and we use the geopackage format to have really everything together in one file so we take here the the shape file uh with um the boreholes we're going to export this from qgis to the geopackage and we use the projection that was used with the shape file so that's the projection that we're going to use in our project so after exporting we have our first geopackage the database that we created and we called it tamoy groundwater data and if we expand the bohr dataset then we see all these attributes that we originally had in our shape files imported we see that there is some reference elevation that there are some some id codes there that we're going to use now the other data sets spatial and non-spatial can also be dragged into that jail package that we created so we really have everything together it's a more efficient way of working with that data so if the import was successful then your geopackage will look like this with this icon showing that it's a non-spatial table and then we have some points with boreholes and we have some line vectors with other information now we are interested in this case study of course in the water levels in those pore holes and the csv with the water levels has a common column with the borehole data file the hydro id and the hydro code and we are going to use this hydro id to make a so-called join so we take the non-spatial table and we join it with the point vector based on the hydro id and the water levels are stored in a field that is called result so we can join that we can also do that with other data that is provided other non-spatial data such as the hydro hydrogeologic unit it also has the same common field the only thing is for this case study it didn't contain information so that's also something you will experience that this groundwater data even if you have a whole set of points that some a lot of data is missing that you might need for your studies now when we look at that attribute table of the boreholes we see that the water levels at the boreholes from that data set were were missing but we have that in a separate csv file so that's no problem we just joined that these water levels are the depth from a reference to the surface of the the water in the borehole so there's also a field which gives the elevation in meters with reference to the to the datum so it's an absolute elevation level but we found in the data set that there are values around 200 200 300 but there's also a lot of value zero which is a big difference with the other values and in that way i found out that that must be no data so the zeros have to be considered as no data are not really zero meters but yeah it would be pity to throw out all those features which have no data for the reference so what we're going to do is to replace that reference elevation with the elevation from the dem and i know that we make a little error there because that is not really the the level at which the groundwater level measurement is taken but we don't have any other data we add a little uncertainty there but we need to get our elevation data also from the internet and for australia there's uh there's elvis still alive and it gives us elevation and depth in different formats and we need to just define the search area and in latitude longitude so i use a tool in qgis to get the extend extract layer extend so i give the borehole data set as an input and it will extract the extend and then i write it to the geo package we store everything there and i need it in epsg 4326 which is latitude longitude and then in the layer information i can here get the bounding box of the borehole layer and use that to search for dm data now there's a lot of dm data there but it's in many small tiles and there's a limit to the size that you can use for downloading the data so that's a bit of work an alternative but a bit coarser is to use the srtm downloader plug-in which will work everywhere in the world between 60 north 56 south i think and you see that our area is covered with 18 tiles which we'll download with the plugin and then the next step is to merge that into a virtual raster so that will not create a real new big file out of 80 tiles but it will create a little xml file that links them up together and contains all metadata to to have that mosaic of the 18 tiles and what we want to do with the dem is to sample the elevation at the borehole data we can use the sample raster values tool where we have as an input layer the boreholes and then we have the dm it is good practice to reproject your dm to the same projection as your borehole data but this tool can handle it if they are different it's not recommended but it works but you just as a general rule you are not always sure that the tool can handle different projections and in the back end we'll reproject them to the same to do the analysis so good practice is first to reproject it but in this case it works an alternative is to use the point sampling tool that is especially useful if you want to sample from multiple layers but in this case we use the sample raster values which is a native tool in qgis and that will result in a new field where we see the dm values here and here result that those were the water levels and we see that those that also has a lot of null values a lot of missing data so the next step is to clean up those missing values because we can't do much with that because we're interested in the water level so we can make this select features select by expression where we say if the result which is our water levels the depth from the reference to the water is not null select them and export those selected features to a geopackage so we have a new layer which only contains boreholes at which water level is known then the next cleanup action is that we want to correct those no data values which are in fact the zeros in the ts ref elevation field because here you see it the ts ref elevation it has values in meters of the elevation of the the reference height for measuring the water level but it also has a lot of zeros and those zeros are replaced with this case when function that said like an if then else if the reference elevation equals zero then replace them with the values in the dm field else use the ts reflect value and then we can finally calculate the water elevation which is the level of the water in the boreholes the piezometric level in reference to uh to the data and then the equation is quite simple it's ts ref minus result and that will give us water alef which will be a new field in the attribute table now in this case study we are interested in the alluvial aquifer we can make a selection of course uh based on the the depth of the water level we can say all the shallow ones are probably the alluvial one but if we look at the attributes we can also see here the attribute f-type class which gives different types of use of these boreholes and here we can make an assumption that the alluvial aquifer in this area is mostly used for irrigation so what we are going to do is to narrow down our selection of boreholes by looking at only the irrigation ones again with select by expression and exporting that again to our geopackage to keep everything together in one file after that selection we can do a spatial interpolation we can for example use the idw inverse distance weighing in qgis there you can access these interpolators through the menu but if you use the ones from the toolbox you have a bit more options from the processing toolbox to give the extent otherwise it will just use the convex hole and you can specify the pixel size so here i export it to 100 meter pixels and then this is a result but i'm schwerker it's going to present it in a more neat way than i do here and another method of interpolation also often used in groundwater analysis is using decent polygons which is called nearest neighbor analysis in qgis and has a similar dialogue and that will result in a map like this now we also want to have a real vector boundary of our study area so we could use aquifer information if we have that but we don't have it in this case so what we then assume is that the alluvial deposits are logically close to the river so we can use the quick osm plugin to download the waterway river tags from open street map and from there we can select namoy river and i also selected this river that is below i forgot the name so we retain only that part of the river dataset from openstreetmap and then we want to create a buffer around it and yeah we already start modeling a bit with assumptions so i choose a buffer here of 10 kilometers i use the buffer pro geoprocessing tool to create this buffer around the river and then i want to preserve those pore holes that are within that buffer and discard the ones that are outside the buffer and in that way i can also clip the interpolation to that buffer around the river so this will be our study area now once we have defined our study area we did all this pre-processing we found out maybe that we missed some data we need to go out in the field probably want to to map springs wells boreholes we want to also create attributes if the springs are protected or unprotected maybe you have some additional depth values that were now missing or water quality parameters signature also many of you are interested in water quality in that case it would be useful to create an app for that for surveying in the field i mean a mobile app there's some great tools developed by lutra consulting the input app where you can create in qgis a field form that you synchronize through a cloud service called merging with your mobile phone and then you can use your phone in the field to map wells and take a picture and fill in the fields based on what you need it will take the gps locations of your mobile phone and then you can use merging to synchronize it back with qgis so this is quickly the workflow you design your field form in qgis and all the map layers all the gis layers that you want to have on your mobile phone you synchronize it with the merging plugin and you can also then monitor it in their cloud service and then you have this project on your mobile phone ready for mapping features and once you're back from the field you synchronize back with the cloud service and you'll have all those data points in your qgis project in maybe the same geodata geo package that you created for this project to further explore or to use in modeling so that was all the content i wanted to present if you're interested in for example creating a groundwater mapping app or other topics related to qgis and water then have a look at the gis open courseware website at gs opencourseware.org where you can find free tutorials and of course at ihe we also offer paid courses with support and with certificates and if you're interested there's currently with unesco of course running on programming for geospatial hydrological applications we have the same materials here in opengs courseware but i'll share another link for that where you can get a certificate until the end of april if you successfully complete the course and it's an essential course on learning really from scratch how to use the command line gdal and programming in python for hydrological applications so that was my presentation hope you enjoyed it and i looking forward to see how kurt continues with that later uh in the visualization part back to craig so kurt if you want to start sharing your screen um we do have some q a time here just real quick to uh maybe address a couple of the questions that have been upvoted the most and it looks to me like most of the people uh the ones that got up with the most are people looking for data elsewhere you so hans did a great job showing us how to get it for australia but a lot of people are wondering how do i get it for my own jurisdiction so i'm going to back off here and let kurt dive into it hans is now going to switch over and start answering those with his fingers frantically typing so um yeah keep him keep him busy and keep conrad busy as well so thanks for those over to you kurt thanks craig so i'm going to pick up where hans left off and i'm going to talk about visualizing groundwater data and qgis i'm going to be walking through the steps that hans went through and kind of explain how we can make that data look nice in qgis again just introduce myself i work for a company named septima in copenhagen denmark and um just a quick shameless plug hans and i have a book related to this topic called qgis for hydrological applications with locate press and i've published another book on discover qgis 3x and these are both available from locate press if you're interested in those so launching into it at the very outset when i would start an analysis like this and i've added the boreholes to a map what i'll often do is use this plug-in in qgis called quick map services to add a base map just to make sure that the data is falling in the correct place and so you can see here i've added an open street map base map to this and this plug-in comes with over 150 base maps that you can use to quickly add to qgis once that data is in there and it's it's falling in the correct place i might want to begin to explore it a little bit by styling it so here i'm turning on the layer styling panel which you can use f7 to open up that opens up a panel on the right hand side and i'm going to show the boreholes by type by using this categorize renderer against the f-type column and just choosing random colors and a nice neat thing in cube just you can shuffle those random colors if the first set of random colors is not pleasing to you you can shuffle those around and then quickly see what all the different types of data are in this data set once we start working with the dem it's nice to be able to visualize that in ways other than just a black and white hill ramp and so i'm going to show you how to do a color hill shade and generate contours and first hans was working with this in gcs which worked fine for what he was using it for but when we want to generate an accurate hill shade image the dem really should be in a cartesian coordinate system in other words a projected coordinate system so i can reproject the virtual raster by right clicking on it choosing export and save as and i'm going to put this into the epsg code the projection of the other australian borehole data 3577 in this case australian albers and this basically puts the xyz values of the dem all in the same units so once i have done that i'm going to color this dem and so to do this i'm going to again using the layer styling panel choose single band pseudo color and choose a color ramp and i'm going to create a new color ramp from this cpt city option in qgis and this is a set of color ramps that ship with qgis and they're divided by category so i can go in and find the topography category and find a nice color ramp i can also go into the min max settings and choose how i want the values represented on the map so here i'm going to use the current canvas extent and then finally i'll click classify it's easy to forget that step and have a nice color dem from the min max values on the visible extent and once i've done that i can use another rendering tool and cue just i'm going to duplicate the dem so i'm just making another copy of it here and i'm going to name this copy hillshade i'm going to drag it above the dem in the layer stack and i'm going to return to the layer styling panel and for this copy of the dem i'm going to render it as a hillshade and that's a nice quick render available in qgis and then i'm going to use a multiply blending mode so that i will see the colors blended with the hill shade and you get a really nice full saturated color hill shade effect i've also when using this hillshade renderer there's a section down below in the layer styling panel called resampling and it's often helpful to play with the resampling settings here to get to eliminate a stair step in the result so here i've chosen bilinear for both the zoomed in and zoomed out resampling for this hillshade renderer and again this is just really a nice quick way for you to render the dem without having to actually generate a separate hill shade image you're just rendering it as a hill shade we can do something similar with contours so i'm going to now make a third a second copy of the dem and i'm going to name this one contours and i'm going to render the dem as contours and so this you'll see you pretty quickly will have this dem rendered in three ways a color dem a hill shade and now contours this contour renderer allows you to choose the contour interval here i'm going to stick with the default but i can make the the index contours a little thicker so that those are visible and change the blending mode to um overlay to make them blend into the map a little bit and so you get a really nice visualization of the the contours and topography of the study area that we're working in i should note that there's also a contour or is actually several contour algorithms in the processing toolbox in qgis that would allow you to create an actual contour vector layer which would allow you to then label those contours as well so there's other options for doing this this is the quickest way so i've added the streams and of the area and the boreholes and i'm going to just quickly show you how you can use an svg marker to basically give a nice icon to these boreholes so i'm choosing a kind of a blue svg icon marker here and making it a little bigger so i can see where these bore holes are on the landscape a little bit better so the svg marker is one option for labeling points like this and you can you can find a variety of svg icons for different um borehole and groundwater symbology sets it's also possible to symbolize these the final set of irrigation wells by the depths the water after after the analysis has been done that hans showed you i can use the graduated renderer against that water elevation column that he created and render those so that you see which wells have the biggest depth of water and which have the shallowest steps to water here and so you can see the biggest depth of water showing up in the southwest southeast corner of the study area now that we have those points interpolated we can also go in here and style the interpolated surface so the in this case the idw and so i'm simply changing it to single band pseudocolor like i did with the dem choosing a nice blue color ramp for it and then changing the opacity so that i can see the underlying hill shade a little bit through that so this is what the data set ends up looking like and you can see i've also labeled the the points here for with the depth to water so you can see how that corresponds to the interpolated surface and the coloring of the points now han showed how to also interpolate using a feast and polygon interpolation and so here a similar styling has been applied to that interpolated surface and you see the kind of the polygon look of the raster and you know you there's many ways to interpolate data as hans described so it's kind of up to you to choose which one but in terms of styling them we can use basically the same method for any of the interpolated surfaces that are produced now i've brought in the the study area and what i'm going to show you here how to do is create a mask around that study area to make it pop a little bit so i'm changing i'm actually going to let this reboot a little bit and i'm changing from a single symbol here to an inverted polygon renderer which renders the opposite of the study area so everything around it and then i'm going to use a shape burst fill to basically create a blend from dark gray to white outside the study area i can set some of the shape burst settings here like the blur strength to make it look a little bit nicer and the final touch is to expand a layer rendering section and increase the opacity of this little bit so i see the area beyond the study area but the study area itself really pops off the map and i finish it off by adding one more symbol to the stack and i'll make this extra symbol a simple line which just gives a nice definition to the boundary of the study area so very quickly you can create this effect which is called an inverted polygon shape burst fill to highlight a study area finally it might be really nice to view these interpolated surfaces in 3d so cue just does come with a 3d viewer and i can access that from the view menu choose new 3d map view and the map view will open as a new panel here and i want to click the little wrench tool icon and choose configure to set up the 3d view initially and here what i'm doing is setting it up as a dem raster layer as the surface and choosing the idw surface as the elevation and clicking ok at that point there's a series of controls off to the right which allow you to manipulate this scene in 3d tilt it and then rotate it zoom in and zoom out there are also tools in the 3d viewer for creating a 3d animation with keyframes but you really when you bring this into 3d you really get to see on this idw surface the the cones the points where the the water depth of water is the highest there if i bring the other interpolated surface in you really see that stair-stepped effect uh created by this kind of thesin you know basically what it amounts to threesome polygons represented as a raster around the sampling points so i already showed how to create contours for the dem you can use the same renderer on any raster so here i've used the same contour renderer against a copy of the idw surface and chosen contour interval of 5 and 25 for the index contours and so we get a nice isolines around this surface so you can really um visualize that data a little bit better and to bring the whole thing home we can bring this into a 3d scene and really have a nice view of that data and you could do the same thing for the peace and polygon surface as well so that's a quick show of what you can do for styling this kind of data in qgis and i can take any questions as we enter the q a period here excellent thanks kurt for that what you see again i want to want to stress and highlight for everybody uh on on the call is that uh boy we blew through a lot of information there and obviously if if you're you're doing it yourself what i've found in some of these courses as i have to stop pause follow along and something that somebody presents in 20 minutes might take me an hour or two to do on my own but no threats um if you uh you know if the pace of this and you've got a lot of information uh in front of you now um is too fast um then sign up for the course and we'll make sure that everybody walks through all of the steps um from beginning to end a lot of people have asked the question well what if i have no experience whatsoever in qgis that's fine we give you some background materials those who are experienced sometimes can just skip it and move ahead but we assume that those starting the course come with no background knowledge uh in qgis now what we do want to make sure we've also highlighted though is that we're showing you how to how to visualize data and a pretty picture can sometimes be just that it can be a pretty wrong picture of the data so getting that data and and how you interpolate it and what you what what assumptions you make as far as what happens in between the data points that's going to be up to you we had a few questions come in about the dem data and i i guess one of the things i want to make sure that everybody is aware of is that if you just go take and download srtm data s for shuttle meaning it's being flown from space watch out those are flown and put out into the brasser grids at uh one arc second so 160th of 160th of one degree which ends up being about 30 meters so sometimes you'll have some channel that you know that's 5 meters 10 meters wide taking srtm data to try and find out what your ground elevation is at that point might not be sufficient and what you'll need to do then as you saw when hunts showed you the download there at least in the australian data set maybe you need all 942 of those one meter grids and you have to put them together it's not uncommon i i'll get i'll get data sets sent to me that are you know one file that's 20 gig just for one file these days it's getting bigger and bigger and bigger and it gets you better and better data but you got to make sure your hardware is keeping up with that so those are the few of the questions that i saw on there that um that i've been hitting maybe let's uh start with since we've just heard from hans and kurt you'll uh maybe share some of the ones that you've responded to with the whole group but let's hear from conrad first anything you've seen in the background conrad on the q a line that you wanted to uh to mention to the group yes i think there are very very interesting questions you probably answered very well introduced very well uh craig what we always have to remember and in some of those questions it popped up that everything depends on the quality of data yes so it's like whether this is it's not this just surface data particularly when we are talking about groundwater this is everything it's hidden no one unless we are at this at the spring no one really sees groundwater uh so the uncertainty related with this information is way bigger than the uncertainty of the of the surface so we have to remember and i think that in hans and court presented very well the way how we can basically map and how we can process some of this information but just bear in mind that the quality of of your data set is crucial so there was lots of questions about where we can get data from we unfortunately we don't know much about your country often sometimes for our project we actually prefer to collect our own data we go to the field we take our data and then we we know what we what we have and what's the quality of this data and this is what i would just would like to probably point at this moment thanks yeah sounds good so over to you let maybe let's start with with hans uh go over to hans the questions that you answered during kurt's presentation and then let's flip around um but what we want to focus on i think is those that got upvoted the most because uh that way you know if 12 people upvoted it then that means uh 12 people had that question so we want to make sure we hit those uh first i see at least 50 questions here so we probably won't hit them all so we need to be a little bit selective so over to you hans yes thank you and i think uh the most of the questions are continuation also of of conrad's discussion uh on the data quality so i would like to to handle a big chunk of questions that relate to to that so first of all for uh for elevation data uh what we see also more and more if you want really detailed elevation data is the use of stereo photogrammetry with drones so there's a nice new course on gis opencourseware.org to use web odm which is open source to extract point clouds and then you can really make accurate dems if you also have good gps data and it's always surprising that in these international webinars people ask us how to find data in your country i still find it very surprising so i think there's a lot of effort you need to do on your side with your governments and with your tax money to get uh the data that you need collected yeah and if you are watching and you work for such a government and you are having a mandate to collect that data you see what we can do what we have data we can process it and use it in our models now related to that is also a set of questions about interpolations and assumptions which is important to address so we've showed uh decent polygons that is very commonly used in in groundwater modeling and also in hydrology if you have no other assumptions and you have a sparse point data set so basically it says the location that you don't know has the same value as the location that you know that's nearest neighbor but most people like to look at idw because it nicely smooths out the data and if you have sparse points you will see artifacts in your interpolation you see as you could also see here you see that there is mountains around the points or valleys around the points when they are low because it uses an exponential decay function in the interpolation as a weight with the distance to the point now many people have learned in a university about creating all very cool and sophisticated but if you have uh as less point density as we have here creating doesn't make any sense for creating you need point pairs to construct a semi-variogram and you use that semi-variogram to interpolate and uh it's based on on spatial autocorrelation then there are several questions about the assumptions of 10 kilometers for the buffer this again is up to you as a modeler a model you should know as a modeler that you need to make assumptions so here i just made an assumption by measuring my my points and looking at the points and how far they're from the river and my knowledge from this kind of river so i thought 10 kilometers might be a good way to separate the ones that are further away and that are closer by but you need field studies or you need geological maps or your morphological maps to find out where that floodplain is or use land use as a proxy so with modeling there are all kinds of assumptions and this especially with groundwater because it's so invisible we need to add a lot of assumptions while we are processing the data and interpolating and as conrad said the uncertainties in the whole process of modeling and uh it is much larger than uh than your uncertainty on the dm from srtm for example although i agree with craig that the more precise you get that the better but yeah be aware that along the whole processing chain there are so many uncertainties added that if you have nothing else in srtm just use that okay so yeah kurt um over to you anything that you uh you were frantically answering uh questions there during hanza's part um any that you wanted to highlight uh no i think hans covered them most of the questions i saw around from han's presentation had to do with the interpolation and um i guess one thing that hans didn't talk about was um that was kind of interesting is questions about determining the accuracy of um an interpolation so i don't know hans if you have any thoughts about that well the best thing is to have ground measurements uh of course and you can split your sample of boreholes uh the ones with the data that you use for the interpolation and another set that you use to verify the interpolation i don't know conrad maybe of more experience with uh validation well it's uh it's just pure mathematics so it's uh i i i think that obviously interpolation will be always some kind of interpret mathematical interpretation of of the observations so it's uh i i actually don't have basically what you said hans it's it's it's correct uh about creating it's uh sometimes we do but very rarely because we don't have enough enough data so we have to choose a method that it's mostly it's a it's a bit like a hydrological intuition what's what we feel is the best and if other people agree with this the reviewers of your project their supervisors then it's normally accepted it's uh sometimes it's it's good to see what other people do and apply the same the same methodology but basically the common sense applies here yeah no i think that's true and i think if you well let me just share my screen real quick um and i'll uh i'll just show you some an application here when we talk about the srtm data this is just something i wanted to share and then i'll show you something about the the ts and polygons this one right here that you see this is an application and i'll put a link in the chat line here where i've downloaded dem data one from the srtm and then another one from uh the one meter grid uh data that's actually publicly available from noaa and i've got links to those as well this is the la river and you've seen that little low flow channel where in greece people go up and down that that channel racing you can see that low flow channel in there in the one meter data look at the srtm data so if you had a point up here you know you could be way off 10 20 meters off the next thing that i wanted to share then just based on the this and polygon questions that came up um yes this is for everything for hydrology for hydraulics and and many many applications even outside of water resources for interpolation this is just out of the hec-ras manual and you can see some things about ts and polygons in here and how you might take precip data and move it around and look this looks very similar to what some of the grids are that that hans and kurt were talking about if you also look at i think computational grid we can see that there's many applications like for example in necroz when you draw a grid line all you're doing again taking these grid points and going halfway in between drawing a perpendicular line many many applications for this so i'll stop my share with that but those questions did come up hey what can you use t some polygons for and how accurate is it this is something that is dealt with across a number of uh industries let's see let's go to just yeah any last comments but let's back up and just go in reverse order again from what we just did have conrad then hans then kirk just wrap up with any final comments based on questions you've seen or anything that you want to share before maybe we interact with you again on another webinar or we get back to the attendees again in a course and interact with you live in that format um in a live workshop where you get to do exactly what you saw kurt and hans just do you get to do it yourself so uh conrad any closing remarks for today wow it has been terrific to listen to to hansen and court i also learned lots of things and wonderful that so many people listen to us i think that your interest about groundwater and the application is excellent and it's uh i admire this and thanks very much for your presentation guys but also for your attention all right has been terrific thanks thanks uh hans uh closing remarks yes uh thanks also i appreciate so many people interested in in this session and all i just want to say is yeah use common sense when you use data place yourself into the modeler's perspective and uh yeah don't don't be disappointed if the highest resolution is not available use your assumptions and make the right decisions there sounds good over to you kurt yeah i'll just echo what hans just said you know you may have to actually collect your own data if something doesn't exist so that that's always that's why hans was showing the input and merging tools um i think um i'm just really uh excited about the number of participants we got for this so thanks everyone for tuning in from around the world at whatever time it is for you and we hope we see you in the course coming up excellent well thanks for that uh thanks to all the presenters the panelists let us know what you'd like to hear a lot of times we show you how to do things but maybe we could get some more content on there on whether the data is correct um do we have correct data in there so we're taking these groundwater points and we're saying okay well have a look at that we put them into a map but you know maybe they're taking it different seasons maybe they're taken from perched aquifers you know maybe we need to understand a bit more about the groundwater hydrology and get into some of the concepts like you would have in a university course these are the things that um now in our online world interconnected world we want to make sure we're giving you the most relevant content uh and giving you the information that the industry needs right now um to be able to continue improving even when we can't meet face-to-face uh sometimes at the moment so have a look at the screen here these are some of the courses that are available but the offerings are as uh can be we can get as creative as uh the input from you on your feedback forms so do fill those out let us know subscribe to the youtube channel uh and uh to the australian water school subscription list so that you will hear about future webinars thanks so much for your attendance thanks to hans kurt conrad for attending today and presenting and sharing that information with us we look forward to seeing you again thanks bye-bye thanks for watching subscribe by clicking the link below and click on the notification bell to stay up to date with new releases for the latest in significant innovative and critical advances in water science technology and management subscribe now to build your skills enhance your technical knowledge and learn from leading experts in water visit the australianwaterschool.com.au and discover our online training courses both live and on demand [Music]

Channel: Australian Water School

Views: 2,359

Keywords:

Id: J0WpzLkAvBc

Length: 52min 35sec (3155 seconds)

Published: Thu Apr 01 2021