Episode 12: HEC-RAS Version 6.0- Discussions with Gary Brunner (Part 1)

[Music] hello everybody welcome to full momentum and an hcc razz vodcast i'm your host ben carey and here joining me as always today is chris goodell along with a very special guest who is kind enough to join us to discuss one of the more exciting version releases in the history of hec razz and chris i'll let you introduce our special guest today thanks ben and uh yeah thanks for joining in and watching i'm super excited to have gary brunner here gary is not only a good friend of mine a mentor but he is the father the leader whatever you want to call him of the hec-ras development team and he's been there from the very beginning so he knows all the history but uh certainly all the recent cool stuff that's come out and we're super excited to have gary here to uh tell us all about it hey gary hey you guys thanks chris and ben for having me i'm i'm uh really looking forward to doing this so i've been watching your vodcasts and i think they're great so i love the format by the way because it's i like the relaxed format and kind of like a little banter between the two of you is always good so it's fun it is a lot of fun and we get to talk about hecraz which is something all three of us really enjoy um and so it's great to have you here one thing i want to mention really quick to everybody is today is gary's birthday january 1st happy birthday gary i think maybe what did you finally make it to 40 yet yeah 40 plus i've been 40 like many times i think so we won't talk about the age but yes it is my birthday but uh thank you hey i gotta i got a trivia question for you gary uh oh what famous famous hydraulic engineer that we all know shares the same birthday as you really hmm i'm just gonna say manning no but you're on the right track oh okay so it's very old in other words yeah it is somebody that we all learned about in hydraulic engineering class and you will definitely know you guys will definitely know who this is chessy you you're getting close you're you guys i'll give you a hint he uh he has a very famous diagram that we all use or used at one point not so much for open channel flow more for pipe flow though oh moody yes that's right moody you share a birthday with moody i would have never ever guessed that not a chance it's funny i would not have known that but uh on linkedin today just popped up somebody posted hey it's moody's birthday and wow this hydraulic engineer with the moody diagram that's like oh gary shares that cool now the big question is who has contributed more to the hydraulic engineering world moody or gary burke uh we'll probably have to go with moody i think who's more famous who's more famous certainly in the heck ass community i think it's gotta be yeah but different time frames in life though right i mean well hey one thing i know moody actually i don't know this for sure but i'm willing to bet moody cannot play the guitar like gary can you never know yeah the electric guitar would have been invented back then in his time but uh one thing i think a lot of people don't know about gary he's not only the lead developer pat graz but he can play a mean electric guitar and i know this first hand because when i worked with gary some years ago he recruited me to play in his band and we used to uh we used to hang out in his house and gary had his had a drum set there for me to play and i played the drums and and he jammed on the guitar and one of the results of that was the intro to this vodcast which um if you look closely when we play that intro it gives a little credit to gary bruner because he's the writer of that song and he plays the guitar on it too and uh actually i play the drums on that too as well but uh one thing i wanted to we got to get credit to get it to martin too so martin fisher played the basement martin fisher's on that oh yeah mark martin is a another hydraulic engineer who is in in our band too so uh but uh i have this very special video that gary shared with me that i want to play for everybody so everybody can witness what a talent gary is on the guitar so let me share my screen here real quick so here's gary playing van bruner which is also the intro to the full momentum broadcast [Music] here comes my favorite part right here the walk down gary you got the rock and roll look right yeah old time rock and roll [Music] [Music] [Music] [Music] i don't know we might have to cut some of this oh no man i'm enjoying it i'm enjoying [Music] it [Music] so [Music] [Music] [Music] nice [Applause] are you better are you better at making heck razz software or playing the guitar gary because i can't tell i think i'm much better at hecker as software than playing guitar there's guitar players are a dime a dozen out there so no offense to the guitar players of the world by the way that's true that that is true uh being the uh hecras developer is kind of a unique um a unique skill to have so that that's for sure but um actually i've been fortunate that uh gary ben and gary's been giving me guitar lessons lately and so uh i can't play that yet but you know i'm getting better i'm getting better for sure the first electric guitar was came out um or was created uh before moody's diagram was great all right wow so booty could be a dammer as well it's quite possible it's quite possible so hey that that reminds me gary um you know the the days that we played in the band together so i don't know uh if anyone knows the history that i have with gary but um i was fortunate enough to work for gary for almost three years at the hydrologic engineering center and uh those were three my best years um boy it was just so much fun working with you and the other members of the hec-ras team and working in davis california just a fantastic place and uh it was it was an awesome time for me yeah it's funny because a lot of people after we hired you accused me of hiring you because you knew how to play the drums did that watch the interview no it didn't actually it did not but it came up very quickly afterwards i was like what you can play the drums really want to come over to my house yeah i don't remember how that how we got on that topic but yeah it was like when gary found out i was playing the drums i think it was more like you will come over to my house right you'll be playing drums so yeah we had a great time working and uh i learned a whole bunch under gary and of course that uh that started my my path into the world of hec-ras modeling and ultimately where we are now you know we're uh we've got the razz solution blog and the podcast with ben and um you know i get to chat with gary a lot about heck razz and and do some testing and stuff so um anyway yeah yeah i i do have to say chris was the original sediment transport guy for the hcross team he was our first well he's a hydraulics guy of course that's his underlying foundation but he was the first guy to move the ball forward and set him in transport in raz so and then yeah that was a lot of fun um and there's still a little bit of legacy of my time there in the software i think i have one of the buttons at the the main screen the the sediment transport button uh no no not the sediment transport the hydraulic design button right i can remember it used to just be the letters hd and i told gary once i was like that's so boring all the other buttons up here have really cool graphics and and you know say something about what's behind it why don't we do something different he said well why don't you make it then just if you don't like it make it so i think i went home that night and i got onto microsoft paint or something and i just uh i put something together and it's still there so that's kind of fun there you go thanks for uh leaving that on there gary no problem so anyway uh we're going to talk about heck razz 6.0 today that's that's the main reason we have gary on uh more than just uh a little walk down memory lane but um ben did you want to say something about our sponsors yeah thanks chris um yeah like chris said the main topic for today is i think the topic that's that riveted the entire hydraulic modeling world and that is the release of of hec-ras 6.0 or i should say hec ras 6.0 um and but before we get into that really exciting topic today i just want to take a moment to thank our sponsor um for this episode which is our firm kleinschmidt associates kleinschmidt is known throughout the industry as a firm that provides practical solutions to complex problems affecting energy water and the environment you can learn more at kleinschmidtgroup.com so thank you kleinschmidt for sponsoring this episode we're really really excited to have gary on here to discuss 6.0 um obviously there's a lot of different directions we could go here gary i just want to start off with kind of how we got here you know 6.0 is a pretty significant release in the history of hec-ras but just for folks i guess who are who are interested um how did we get here yeah well it's it's actually been a long road for me to this point so uh the history of raz is pretty long actually i started working at hcc in 1985 and at that time all the computer programs were what we called back then black boxes and what i mean by that is you would develop a text file of the input you would run it through this black box and the black box would write a text file of output and then you would look at the text file filled with numbers and that was the results that was computer modeling in 1985. there was no such thing as a graphical user interface for inputting data we did have some programs to read the output file and then like draw a plot like a cross-section plot or a hydrograph plot but that was it the other thing was at hec at the time we had hec2 which was the main river hydraulics program just did steady flow hydraulics then we had hc6 which did steady flow hydraulics but with sediment transport okay we had unet which was our 1d unsteady flow program and we had a program called wqrrs which did water quality modeling and then at that time the big 2d program was rma2 and we had a special version of rma2 at htc was htc dash rma2 because we changed the input to be similar to htc input and we changed the output to be similar to hcc output and we kind of made it a core of engineers version so here we had five programs the problem was they all required different input different input formats different output and they actually did hydraulics slightly differently so ht2 and hdc6 didn't even do the backwater exactly the same there were slight differences and so i looked at that and i said this is crazy because if you you do a hydraulics model but then if you want to do sediment you got to start completely over and recalibrate your model if you want to do water quality then you got to start completely over so and actually in 1991 i got asked to head up what we call then the next gen software development for river hydraulics we were we were doing next-gen software developer everything at hcc so the director at that time daryl davis and then the other person art paps were kind of the spearhead of we're going to redo our software technology also what had recently come out about that time was windows 3. windows 1 came out in 85 but nobody used it windows 2 came out in 87 it was slightly better but when windows 3 came out they actually had a graphical user interface with multiple windows that you could actually do things in and at least had the appearance of multitasking anyway even though it's a single processor so in 91 i got asked to be the lead to developer ass and at the time they were thinking we're just going to put a windows interface on top of hc2 and i was like ah no i don't want to do that please don't make me do that you know my boss was vern bonner at the time and and i remember him saying well there's all this experience in hc2 and i said yeah but vern we want to be able to do 1d hydraulic steady flow one density flow sediment transport water quality and maybe even 2d dump someday all in one piece of software and to do that we really got to start over with everything so he literally told me okay well i'll give you six months you have if you don't have at least a user interface and backwater computations for cross sections in six months we're changing gears and and just putting a uh an interface on top of hc2 and i said okay so bam there's the challenge and fortunately i had two fantastic people working with me mark jensen and steve piper so mark jensen and steve piper and i were the original three to start ras and those two are if you know them at all they're both phenomenal you know they're just exceptional uh engineers and thinkers and and doers so i was really lucky to end up with those two working for me so it was a little bit intimidating working in that group with uh with you because mark and steve are super smart guys and they they are at a level uh um a lot higher than i do so it was always me trying to trying to keep up with those guys yeah so the other thing is we each had skills that the other did not have and that's what made for a great team i think so yeah so gary what do you think would have happened if maybe you weren't working with those individuals and you weren't able to get something done in six months would we not have hec-ras right now no we we would have something called hec-ras but it would be a pretty face on top of hec2 and then maybe and maybe on then bringing uned in okay and hc6 okay yeah and if it would have never grown to what it is it wouldn't have been possible because the the code back then was horrendous old fortran code with go-to statements and all kinds of crazy crap you know that you just didn't want to deal with it was not modern even for 1991 so what's what's kind of interesting to think about too is back then i mean windows is a part of everyone pretty much everyone's daily lives today right back then it was new and yes had the foresight to link up this new next-gen software into windows was there ever any concern that hey what happens if in two years from now windows is no longer microsoft goes bankrupt and now we're screwed absolutely and and there was um in addition to windows sun on the unix side had their own user interface and there was several groups at hcc that decided hey we don't want to work on the pc we want to work on a unix box okay so they went the sun root and this was before java even existed there was no java so there was a development language i don't remember it was called for the sun machine to develop a graphic use interface but i quickly looked at that and said you know windows was way superior to what sun had going on in the unix box and other things that were going on and i i really looked at microsoft as a company and said okay what do i think is going to really be around 10 years from now you know and you know and it's just obvious to me that microsoft was just taking over in that area um and so to me um that that was really you know an easy decision for me to to be on a windows box the other thing was that i had along with i was one of the first couple people to get a pc with windows on it there was only like two of us at first and but i also got a visual basic compiler and so i started learning visual basic you know and and i just learned everything about it and the first thing i did is actually developed the interface rate you see one a simple one just on my own and then one for htc two to run those programs okay not complete input but to run those kind of a graphical way so i saw that in addition to windows there was a programming language that could easily be used to develop interfaces and if an engineer like me could do it like almost anybody could do it and then once i started working in on raz the first version of raz mark just kind of took off with visual basic and and just was just like became an extremely good user interface programmer and uh and so uh it became obvious that he was going to do more of the interface programming than i was so much better at it than me um but yeah so we started in 91 and the the first beta version didn't come out until uh may of 93. okay and then we had another beta version in 94 and then the first actual release of raz was july of 95 and people asked me often ask me well why did it take so long to come out the first version and the reason is i knew that the first version had to do everything that hec2 did otherwise people wouldn't accept it and so we just kept adding the features that hd2 had in hd2 had been around 20 years so there was a lot of things to add in that first version and that's why it took from 91 to 95 four years to get that first version now so wow yeah that's that is really interesting i i actually gary i learned a lot even having worked there i didn't know a lot of what you would just uh mention there that was really interesting and one thing i i want to clear up too for everybody out there what is the actual real proper way to pronounce your software well everyone says heck razz but i know having worked there that that's not the right way well though if there is a correct way it's hec raz but i even say occasionally say hecarez i don't care hecariz hcc raz whatever you want to call it it sounds good to me so i know what you're talking about yeah i i got in the habit of calling it atc razz when i worked there and uh probably within a year after i left and and was teaching it everyone outside of hcc calls it hec-ras and it's just like it was inevitable that uh that's how how i was gonna call it from now on but i still i still try to when i'm in the company of hcc employees try to say i know some people have a little issue with it but one thing i do know for a fact is you should never call it the hex center right right [Laughter] people do that still i hear them say oh yeah gary he works there at the heck center yep what's funny is we we had a softball team though and our name of our softball team was what the heck let's see okay so you embrace that difference it's all good so gary you know the the first religion the first version gets released in 1995 was it pretty popular immediately did it take some time to grow into kind of close to what it is today it's like everything when something new comes out there are groups that just want to embrace it right and that tends to be younger people let's face it younger people love learning new things and as we get older we get stuck in our ways a little bit not because we're stuck but we we get comfortable let's call it so a lot of the older engineers didn't take to it right away but the young engineers just took off with it you know and even the middle-aged engineers and but to be quite frank it was probably oh gosh over two years before every corps of engineers office even accepted there was one core office i'm not going to say which one refused to use it for over two years would not use it period because they kept saying well i prefer to do the input the hc2 way you know and i was like okay so what i decided even early then is you the way you get people to use things is to make it better and the people who didn't want to use it or the people are actually listening to i would ask well why why not what tell me about it what's the reasoning and they would give me very often some good reasons that they could do something more efficiently this way or that it would run faster or the output they had you know graphical programs to plot this and this and that and and i would make note of those things and say okay well we're going to do that and then once we do that then maybe you'll want to use it and that's the way you get people to use your software so i think that's that's behind that's one of the main reasons raz is so successful today is it's usability um and so that that doesn't surprise me at all that that's the philosophy you have gary because um one thing that i've noticed about hecraz over just about any other hydraulic engineering software that i've used is how user-friendly is how intuitive the the graphical interface is it if you know windows you know how to do heck raz already because things are where you think they should be you know the menu items at the top how do you save something how do you open something how do you navigate around it's all very intuitive which makes it a very user-friendly program yeah when i started doing the design the first thing i did is i went up to microsoft and i got all the user interface design books i could find because i wanted to know what was microsoft's expectation for user interface design what were the standards what should we follow why is that relevant you know it became obvious if you do things by their standards it will make it easier for people to use the software because like you just said if they know how to use windows they'll know how to use your software so yeah so you literally went up to seattle no i didn't i couldn't see how much i went to microsoft figuratively on the internet and got all their all their design their design manuals for user interface design though but i haven't been to seattle but i didn't meet bill gates though i was going to know when you had said that you had kind of looked at everything that was available and kind of forecasted that microsoft was going to be you know a primary player in the space i hoped at that point that you bought some microsoft stock so no i'm not that smart i was thinking about that like gary you saw the light man yeah i've said that to my several times after the fact when i have not bought stuff that i should have although you know to be honest if if you had bought in a decent amount of microsoft stock at that point um you probably wouldn't be working for hcc right that's true i'd be a multi-millionaire and i wouldn't be here right now i guess i guess we should be glad you didn't then gary i was gonna say so thank you gary thank you um very cool so so you know the version gets released it slowly kind of becomes more accepted within you know core offices i'm assuming it begins to catch on a popularity in the private sector at what point did you i mean obviously you had said that even when you guys were initially starting to develop the software that you had kind of forecasted and said hey down the road we'd like to do this we'd like to maybe even add 2d and those types of things well believe it or not the design from the very beginning the reason i called it raz river analysis system was the design from the beginning was to do 1d steady flow 1d on steady flow sediment transport water quality and there was an inkling of 2d but i didn't want to say that too loud because there was other groups in the core that were doing two differential modeling that i didn't want to ruffle their feathers but the design all along was to be a river analysis system you know a complete package in other words this is the thing i need to do river hydraulics and and sediment transport and water quality and graphics and mapping and you know the whole thing so that was really the vision from the start so from the start was it were you guys anticipating doing the multiple version releases that we've seen happen over the last 20 years is that kind of something that from the beginning you said hey every couple years or whatnot we're going to release a new version with new features and we're just going to keep on pushing or was that something that kind of just snowballed that was the idea right from the beginning but in order for that to happen there has to be this thing called money to support it and money kind of goes up and down in these areas so uh you know at that time we were solely funded by research dollars when we started out raz it was a research work unit that we were using to develop the next-gen software but that kind of money kind of goes and comes and goes and there's good years and there's bad years so in the bad years we had to rely on district offices so there were a lot of core district offices ponied up you know thousand hundred thousand for this feature that feature um along the years so there's so many districts that have contributed to the development of raz and not just with money with ideas too so it's been really good okay but yeah even when we'd make a release we were already working on features for the next version that that's always been true for every every release so yeah okay um all right very cool yeah that's that's really really cool information i didn't know a lot about that chris you know you mentioned that you had learned some new stuff there too i'm sure everybody is going to get a lot out of that particular conversation um so that's going to be one maybe one other thing just real quick yeah go ahead so just because people often ask me about the history of razz but so version one we talked about but version two there were there were one point one and one point two and one point three but then version two came out and the big thing about that it was it had inline structures it had the whiz pro bridge routines it had scour computations that had hydraulic design stuff you know stuff like that that came out in april 97 and then version 3.0 came out in 2001 and that's when we first had 1d unsteady flow okay and then also hec georaz came out in version three that with the arc info add-in okay and the spearheads behind gioraz were actually cam ackerman and um he's the was the lead kind of idea of jio raz and cam's kind of been our gis lead ever since then on the razz team including razz mapper stuff that we have now and then rasp 4.0 the big thing about that was sediment transport and water quality so that's raz 4.0 is the first time we released something with the did sediment transport and that was 1d sediment transport in one day water quality and and that's when chris was gone then and stanford gibson was the lead of their sediment stop and stan just done a great job carrying the ball that chris started and and moving sediment transport forward so now you know that was 1d sediment transport and then now we've got 2d sediment transport which we'll see we'll be talking about maybe another time another broadcast um but uh and uh we've got mud and debris flow capabilities and essex xo and et cetera so that stuff's really taken off and then water quality mark in a jensen in addition to be a phenomenal engineer slash user interface programmer is really uh interested in water quality and he spearheaded the development of water quality in wrath so that's mostly mark's effort as well as dr cindy lani that works with mark so and then i i frequently have have said this to gary and others but my my greatest gift to hec was leaving so stan could take over that because stan has been as amazing and and yeah we hope to we hope to get him on this vodcast because he's not only an amazing engineer sentiment transport scientist uh but a great uh a great speaker too a lot of fun to listen to so we're hoping to get him on here yeah definitely definitely and then more recently in 2015 raz50 came out with 2d and that was a major release and there was so there was a big gap between 40 and 5.0 i think five years um and uh so it was a long time waiting but we spent a lot of time working on the 2d stuff and there were other new features besides 2d in there also but that came out in 2015 so now here we are um releasing the 60 beta version and at the end of 2020 so crazy yeah i remember i got my first job out of school in the fall of 2015 and the consulting firm that i got hired to did a lot of hecras modeling and when you know 2d had recently been released and so there was a lot of excitement around that and i just remember thinking like why is this such a big deal and now it's come full circle where i'm hosting a vodcast talking to the developer when 6.0 is released because of how excited we are so circle of engineering life there you go now you know but yeah you didn't you didn't slog your way through uh trying to make 1d models fit into a 2d system for years and years having done that you would definitely understand the excitement behind 2d yeah chris had a great opportunity in one of the earlier vodcast episodes gear i don't know if you saw to to highlight um his 1d and then subsequent 2d modeling effort of the missoula floods oh yeah and so you know that that's a great example of the excitement and the efficiency that was gained when 2o was released sorry when 2d was released so very cool awesome well yeah i think gary you did a great job summarizing that you brought us right up to kind of where we our main topic for today which is the new 6-0 release there's a lot of of really exciting features in here if we talked about all of them we'd be here for probably a whole day um but we're going to we're going to touch on as much as we can in the in the time that we have um and and like chris mentioned we're hoping to have stanford gibson as well as cam ackerman on to the vodcast to talk about other portions of the new 6-0 release because there's so much new exciting stuff and each of those individuals along with gary have been you know very influential in both the development and the release of 6-0 and so we're looking forward to hopefully having each of those folks on eventually but gary today we're going to talk about a number of different things and at this point we're kind of going to kind of kind of hand off the conversation to you and then we'll have an opportunity to you know talk about things ask questions and and what not kind of nerd out on the new on the on the new release but we're going to we're going to briefly touch on um spatial precip infiltration 2d bridges wind forces 1d finite volume pump stations for 2d and improved computational speed those are going to be kind of the topics that we're going to touch on today so with that gary i'll go ahead and turn it over to you all right so obviously we're going to be talking about what's new in raz 6o today and we're going to limit this discussion for today though as you said to basically some of the new major hydraulics features and ben already talked about what we're going to talk about so i don't need to go through that if we do have a little bit of time we might show a little bit of the 3d viewer at the end but if we have time um so first of all let's talk about spatial precipitation and we've if you if you've gotten 60 beta you'll see that in the ncaa boundary conditions there's a couple of new tabs there's a meteorologic data tab and observed data so if you're going to add spatial precipitation you go to this meter logic tab and you simply say i want to enable spatial precipitation evaporation once you do that the window will morph you pick the mode in this case i'm doing gridded but there's also point gauge availability and you can also just do a kind of a constant rainfall too if you just want to do an experiment and once you've decided on gridded then you need a data source and right now the data sources for gridded data are limited we have the ability to store grids in dss okay and there's a there's a computer program from hcc called hccmetview that you can download and it's free and it takes all kinds of data and it can make grids and you can write it to dss so that's how you can get grids in dss but then there's also this national weather service has some raster file formats so if you pick that we support the national weather service net cdf file formats and what's called grib g-r-i-b-b so we support those and that's right here on this screen so there's the grip format in the net cdf format um but as i mentioned you can also do point gauge data and so the point gauges can be nhcds as time series as regular or irregular interval or you can enter point gauge data and just into tabular and if you do enter the point gauge data then what happens it actually interpolates it into a grid which we're going to show here in a second to make a point gauge you basically go into the editor and say okay i want to create point gauge stations and so you hit this create button here and it brings up an editor and you can either hit the detail tab and enter one at a time or you can hit the table and add them you just got to give it a name and then the latitude and longitude to start out and how high the gauge is above the ground in this case 10 meters for these and then you have to pick a an interpolation method there's lots of different interpolation methods because the reality is point gauge data is not really great data spatially we all know that and there's no interpolation method that's perfect in there so we we basically created a bunch of them so you can try different methods okay what is the intention there gary to kind of be able to do almost a sensitivity analysis on that yeah i mean the reality of it is is point gauge data you're just missing what's in between the gauges right you're missing the spatial what happened and so you're just doing some kind of interpolation okay so no matter what it's not going to be as good as gritty data to capturing the spatial variability of the rainfall just it's not possible unless you have just a really dense gauge network which from my 35 plus years of experience doesn't really exist in most places so here here's an example of point data and this is our friend bald eagle creek that comes with the software right here the reality there's only one precip gauge in in bald eagle creek and that's right at the dam sayers dam and then there's some other gauges outside uh bald eagle creek around here so i'm showing a lot of gauges here but probably only these gauges right here are going to affect what's happening at bald eagle creek and so this interpolation right here is actually what's called decent polygon so it's very similar to what you do in a hydrology model okay and where you you connect the gauges and then split them halfway between and here's what it would look like if i did inverse distance squared between the gauges here's inverse distance squared but triangulation limited meaning you triangulate between all the gauges and then you only interpolate on those triangles so only three gauges can affect the magnitude within each triangle and then here's a new method though we call peak preservation method and and a guy by the name brian mccormick came up with this idea of doing gauge precipitation interpolation to preserve the peak and intensity of each gauge so you kind of line up the the centroid of mass of each gauge and then you interpolate from there and then you then you go back and split it back under time to where it was and then you interpolate the time base from that so that's kind of an interesting method would that be more of a conservative estimate yes yeah because one of the problems with interpolating between gauges if you have the peak at one gauge it occurred at noon and the peak of the next gauge occurred at two o'clock and then you use those two gauge interpolate well you just squashed both peaks to half right so you never get the intensity of the rainfall in the event that you interpolated so this method preserves the peak intensities so it's kind of a cool method yeah that's really cool it's just a plot then showing the one point in the um the basin showing the rainfall for this example and then this is the accumulated rainfall from the various different interpolation methods and the highest one giving you the most rainfall is actually the mccormick method or what we call the peak preservation method okay another thing to point out is you can use medview and you can actually develop pmp problem maximum precipitation events in a gridded format in met view and then raz can read those in so here i've got a pnp event over centered over a watershed that i developed the 2d model for and we applied that and here's the accumulated rainfall one of the point here's the rainfall for the pmp at the centroid so you can do those kind of things too yeah well along with the spatial precip we've got spatial infiltration i'm not going to spend a lot of time in this but we we added basically three methods of really simple deficit constant scs curve numbers kind of an intermediate method and then green and amps is a more physics-based method um each of these methods we're applying the infiltration directly to the hydrograph we're not looking at the water on the ground and continuously infiltrating so it's more of a simplistic approach for this first version you subtract the infiltration initial losses from the hydrograph and you have rainfall excess and you just throw that onto the land surface and say it completely runs off and no more infiltration okay so the deficit constant method um there's an initial deficit and a constant loss but you can actually recover the initial deficit based on evapotransp or periods of dry so gary that that makes me wonder because i get this question a lot people will run a model um an event model and you'll get water that goes out of the channel into the flood plain and then it comes back into the channel and and goes back towards base flow but then it leaves behind these pools of water around and i've had people ask me a lot hey is there any way to have that water go away or infiltrate like it would naturally over time and it sounds like this current version wouldn't do that yet but maybe a future version will no the current version will not do that that's correct and but we already talked about you know having options for continuous infiltration but when you go to that route where you can have continuous infiltration you have to have some kind of groundwater model okay because if you allow water to continuously infiltrate okay then you have to account for that as possibly being interflow or groundwater flow that gets back into the channel so we didn't want to go the full route of the con you know the full kind of surface subsurface groundwater model to raj yeah but that is something we are going to work on for future yes cool so right now it's kind of a more simplistic approach when it when it comes to infiltration methods one thing i want to point out with the ses method is it's not the straight ses method um because we we basically encourage people to specify percent impervious separate from their curve numbers which means your curve number would be based only on the pervious area okay you can do it either way though in raz but the other thing is you there's a user-defined initial abstraction it's not a default to 0.2 times the storage which is the ses method you can specify whatever you want and the other thing is you can specify a minimum infiltration the ses method once you have a curve number it computes a storage amount and let's say that ends up being like six inches well if you've got six inches of rain then the infiltration and the scs method is zero then from that point on in raz you can say no i'm going to specify the minimum filtration maybe it's the saturated hydraulic conductivity and so even if you completely satisfied the storage from the curve number method you can still get infiltration and then in addition to that we have recovery options where you can recover not only the initial loss but the the whole infiltration rate from the storage based on a period of dry between rainfall events and last but least we have the green and amp method which is a little bit more physics based if you like that you can you can do that it's got more parameters obviously to calibrate a little bit more difficult to calibrate a model but um it is a method a lot of people like so what i want to do now is switch over and just bring up an example data set that's a spatial precept and infiltration data set so gary as you're opening that up um just a question for you who was the biggest proponent or who was pushing you the hardest to incorporate spatial precipitation and infiltration was it um was it was it yourself was it you know outside consultants who who are the biggest proponents for this well there's lots of people who said they wanted in raz but i i've known since conception that if we want to do a better job of hydraulic modeling even if we're working in a large watershed and the hydraulic model is just a small piece of it yeah we want our hydrology model for the entire watershed to provide boundary conditions to the hydraulics model but if we want a more accurate answer at the hydraulics model level we still need to capture the spatial precipitation infiltration in the hydraulics domain okay so i've always wanted to do that for the longest time um so we're not trying to be a catch-all everything hydrology model but what we're really trying to do is allow people to have more accurate hydraulic modeling in the hydraulic modeling domain okay and you know obviously you might you might want to take a watershed approach with it too but there's a lot of things that razz can't do that general hydrology models like hms can do so we're not we're not taking over what that hms does sure um all right so here we have our friend bald eagle creek and this data set this example comes with the software so this example precipitation infiltration uh data sets in the software ras60 that comes with the beta version if you download the example data sets and so first let's look at the the rainfall so once you here we have a run and we can look at the event conditions for the rainfall and there's two kind of layers under a precipitation layer and then a accumulated precipitation so we turn on the precipitation layer and i'm going to zoom out so you can see more of the precipitation okay and then we'll just go ahead and play that and so here's the gridded rainfall that's happening over time during this event so you can see what's going on you can see how the storm's progressing okay where the intense parts are in time and how that's hitting the watershed et cetera et cetera et cetera okay are the precipitation files um are those pretty significant in size or are they not um they're actually not as significant as you would think because precipitation by default right now is is for the most part stored in two kilometer grids now we're getting it we're getting it in finer grids now and the weather service is moving towards even finer grids so as the future goes on they're going to get bigger and bigger because the resolution is getting smaller and you know that's going to be great for even better rainfall but right now the grid size is pretty large so they don't take as much space as you would think okay so here's another approach to look at this preset this is a cumulative precip plot so here we're just going to look at rainfall accumulating over time so you can see the most intense part of the this event was outside the bald eagle creek watershed it was just kind of on the edge of it in this case that's really cool yeah and so outside there was like six and a half but in the watershed there was like five to five and a half inches of rainfall inside the watershed when you say in the watershed you mean in the 2d area right well yeah that's not true the watershed is actually bigger than this so just just so everyone's clear about this this is this is just capturing the rainfall on the 2d area you still need a watershed model to provide the inputs for the the main river and the tributaries right yeah i mean you could model the whole thing with raz okay if you wanted to but um probably the the more standard approach is to model the watershed with a hydrologic model like hms and then model the hydraulics portion where you really need to do flood inundation mapping and need hydraulics okay and then use hms to provide you boundary conditions at the exterior points so you know there's probably five or six locations where you're going to get flow hydrographs from hms but you no longer need the interior points from hms because you can compute that yourself spatially okay so it's always been kind of a pain in the butt to get hydro gas computer from hms that were computed maybe like here and then here and there and then put that inside the hydraulics model but you couldn't distribute it spatially right so it was not quite as good of an answer now you don't have to do that you just take the external boundary conditions from the hydrology model and compute everything internally directly from the spatial rainfall and infiltration yeah that's really cool and and that gets to another question that that i get a lot too gary but let's say we're talking about maybe a smaller watershed what would be some things people should think about when deciding do i want to do uh just a full 2d reign on grid hec-ras model or should i use hms or maybe some hybrid between the two or a combination of the two what are just briefly some things people should think about to make that decision um some of the things are but what are going to be the applications of that model let's say the model requires uh snowmelt well raz doesn't have snowmelt hms does so if you're in that kind of area right away you need to use hms okay what if you're going to do continuous simulation soil moisture accounting and you want to run 100 years together well raz doesn't have continuous simulation soil moisture accounting hms does okay so there's quite a few hydrologic modeling things that hms can do that we do not have in raz okay yeah and because rasman you know totally focused on hydraulics and with this is our first step into spatial precip and infiltration so it's got some limitations from a hydrologic perspective when you're just talking about let's say strictly the rainfall runoff process uh what are the differences how does hms do it versus how how does raz do it with the the new uh rain on grid feature um well it depends on your approach in hms so in hms you can still have gridded rainfall right but maybe you're doing sub basins when you're doing unit hydrograph runoff but the the current beta version of hms has a diffusion weight 2d diffusion wave from rats so we've added into hms the 2d diffusion wave solver into hms so you can even do 2d in hms now too with diffusion wave modeling so in that if you take that route then they're very similar you could have the same good rainfall we're limited in rats to three infiltration methods hms has more methods hms has continuous simulation um infiltration and kind of um subsurface tank kind of modeling for first ground water flow etc hms has various methods for base flow on and on and on but so from that respect though it depends on your approach in hms and you're approaching raz so there's some overlap of course and but there's quite a few differences also interesting um yeah that's cool and the other difference in ras for the infiltration methods we add these extra capabilities to like for like the curve number method to have the minimum infiltration rate the recovery of the initial and storage capability that's currently not in hms to do that so a little difference there that's that's great information gary what if people were wondering about where they may get into trouble with doing rain on grid are there any limitations that they should be aware about um in heck grass for rain on grid well the the first limitation we already talked about is that raz is subtracting the infiltration from the hydrograph and it's not doing continuous infiltration based on what's on the land surface okay the second one is raz doesn't have a bunch of base flow methods so you kind of have to pre-wet the system and run the model for a little bit to get base flow into the stream and you may even have to have some inflow locations to model base flow as inflow hydrographs and ras so that's some some limitations but other than that i think it's not that limiting so what about what about larger cells gary in an area that has very very steep slopes um that was well like like anything if you if you're just trying to compute the runoff and you're not really concerned with the water surface on those large cells no problem right but if you're trying to actually compute the water surface and the travel time and velocities well like everything if the cells are too large they're too large you need to make smaller cells to get a better answer because inevitably in any 2d model you only get one water surface per cell and one average of velocity per face and that's true of all 2d models so yeah one thing we've noticed is is when you have really large cells in steeper terrains you can overestimate the the time sorry underestimate the time it takes for water to move through so in other words it's it moves through a lot faster than it would with smaller cells because you do have that level pool well part of part of that well there is actually an option in raz when the flow is shallow it doesn't model as a level pull it it actually it actually slopes it as a surface when the flow is shallow in fact you can prove that to yourself just take a perfectly plane on a slope and run raz well if your cells are large it's not modeling it flat pull flat pull flat pull it actually says oh this is really shallow flow i'm going to take that volume of water and i'm going to slope it across the plane and then figure out appropriate depth at the face so we actually spent a lot of time on doing that because we didn't do that it would never be able to waddle water down a plane shallow flow down a plane accurately and it can and you can prove that to yourself by just doing some simple planar surfaces rainfall runoff experiments and i've done these experiments and compared raz to lab data even for uh where we have lab data where people have done kind of like watershed water running over a plane kind of thing and raz can model that okay very cool so another thing that people do probably in raz that they need to kind of start understanding is when the flow is very shallow the relative roughness is higher okay so if you have really shallow flow even going over pavement well the roughness value isn't 0.012 it's probably more like 0.2 and so um what i've been telling people is away from the channels they need to use higher end values for those areas where the rain is going to hit and it's just going to be shallow flow because the relative roughness is so much higher and in future versions of razz we're going to have variable roughness with depth we don't have that currently and once we get that in there though that'll handle this problem where you have these really shallow depths and the rough relative roughness should be high and as the depth increases it gets lower and lower cool so yeah okay now are you going to have a nice table in the manual that tells you what end values to use there absolutely that is so hard to come by there's just a lot of research out there on on sheet flow type of n values but uh right and more of that right and then especially how n value changes with depth so yeah yeah i'll say so continuing on the story let's talk a little bit about infiltration well how do you do infiltration well the first thing you have to do is you have to bring in a land cover data set and the way you do that is you right click on map layers and you say create new raz layer land cover and this is a familiar thing that you've brought had before if we're doing n values in raz and so you add your your land cover data and whatever grids or shape files or combination thereof you have for land use and land cover and you make a a layer so here i made a land cover layer and i used the uh national 2014 data set and i just brought that in okay and but in addition to that though we've added this ability that well a lot of places the national data sets aren't that great you're going to need to override it so if you notice there's this thing called classification polygons okay well you can right click on that and say edit layer and then you can draw your own classification polygons so i did that i'm going to say no to saving that for this main channel i made a classification polygon as my own and i called it main channel and the purpose of that is i wanted to establish a different end value and even percent imperviousness for the channel versus the other land cover types because the national data set does not cover the channel very accurately okay so once you've made that it's that more or less the same thing as an override region but just but it's only overriding the land cover classifications okay so think of this as just i'm making land cover and i can make my own land cover classification so i made a land cover classification is called main channel okay so this would be similar to in the previous version if you were to bring in say two different land classification or land cover data sets and bring them together to make your right n value grid right i could have had a shape file of this right yeah and then a grid of this standard lan cover and now you can do this right you can do this right here in raz mapper that's cool yeah that's really cool so then once you have that then you can do right click on it and say edit land cover data and here's where you get all of your land cover types and then you can enter your manning's end values okay just as before optionally you can enter percent imperviousness now you wouldn't you don't need a percent imperviousness unless you're going to do infiltration modeling okay and even then you don't need it one approach is to put in curve numbers that include the percent impervious another approach is no put the percent impervious in separately and then develop curve numbers only based on the previous area that is my preference it's more physically based so i would encourage anybody to do this keep the percent impervious separate enter it right here in this table okay and then when you develop your curve numbers only base the curve numbers on the previous area which we'll talk about here in a second so here i got manning's n values and percent impervious values for all the land cover types okay well the next step then if i'm going to do infiltration is optionally i can add in a soils group so the way you do soils is the same thing you right click on map layers you go create a new raz layer create a new soils layer you bring in soils data and there's a whole bunch of detail in the manual on that we'll go over it okay and here in this case i brought in some soils database but i decided to only extract from the database the scs kind of soil types the a b c d so if i move the cursor around you can see this is soil type b this is a c this is a c d b d etc here's an a okay and i can look at that in a table okay and just to see what that is so i basically have um eight soil types in this watershed based on the scs soil types then the next step is to create an infiltration layer and the way you do that is you right click on map layer so you say create new raz layer and you can say i want to create an infiltration layer from land cover and soils or if you already have a shapefile out there that maybe is from like hms where you've already defined um your infiltration parameters for let's say subbasins in hms and you export that as a shapefile you can just directly read that in shapefile and now you have to tell raz is it deficit constant is it curved number is it green and amp for that shape file okay so it'll just read those polygons in and then you can specify the parameters for each one of those polygons from the shape file so in this case i did land cover and soils though and when it comes up you specify the infiltration method i didn't i did uh ses curb number and then you can optionally say which land cover you have and you can do land cover and soils or you can do land cover just by itself or you can do soils by itself you don't have to do both so you can base infiltration just on land cover just on soils or both the combination thereof and i did the combination thereof for this example and then here's where you also enter that time between rainfall events to establish the initial loss since i pissed picked the scs cure number method so like maybe you might want to say 24 hours okay after 24 hours of dry reinitiate the complete initial loss and the and put the soil moisture back to zero and start the infiltration over again and then you would create the layer okay in this case i've already created it okay and again once i create the layer then i go to edit infiltration data and it pops up a table and here's where you're going to see the conjoining in this case of the land cover and the soils so here i have mixed forest okay and i have mixed forests soils b soils a b d d c c d and a and then i have deciduous forest then i have open space then i have evergreen open water and i get down to where i also have developed low intensity okay uh wetlands developed media intensity developed high intensity and then different soil types along with those so so for example developed medium intensity okay and soil type d this curve number is based on only the pervious area because i specified a separate impervious amount for developed medium intensity over on my land cover data set okay so i'm not taking into account the impervious area when i when i compute these curve numbers now you got to compute the curve numbers outside of raz and then you enter them in these tables and then you also decide well what initial abstraction do i want ratio and then do optionally you can put in this minimum filtration rate it's not required it's just an option in this case i did i put in the what i consider to be the saturated hydraulic conductivity okay so that's how you specify the infiltration parameters the other cool thing is once you have this stuff in there if i highlight the infiltration layer and i move around not only does it show me the land cover type and the soil type but it shows me the curve number the initial abstraction ratio and the minimum infiltration rate right in the in the that bar there so i can quickly just move anywhere and see what curve number do i have there okay et cetera et cetera same thing if i if i go back to the land cover and i highlight that i can see well what's the name of that land cover type what's the end value and what's the percent impervious okay it shows you that right there so it's showing you can quickly see what your data is for any point just by mousing over top of it so once you have those layers and you have your infiltration layer in your land cover layer then to wait the way to make it work is you go back to the geometry and you right click and say manage geometries and in this editor is where you tell it okay for this geometry i want to use this terrain i want to use this land cover data set to define manning's ends i want to use this infiltration layer for the infiltration and i want to use this layer for the percent impervious now the land cover layer in this case has the manning in and percent impervious in it so i specify the same name twice but i can have several different infiltration layers that i so i could say maybe i want to do ses curve number and one and initial encodes another one or maybe i want to do ses curve number with a certain set of curve numbers and then i want to do another geometry with different curve numbers to maybe represent some future condition and i can do that so i could have multiple infiltration layers and and do that so once you associate those mannings and infiltration impervious layers then it automatically is going to do infiltration if you have rainfall and that's all you have to do so um we specified the rainfall we showed earlier in the unsteady flow data boundary conditions under the meteorologic tab is where we enabled precipitation we picked gridded we picked the data source that's all you had to do for the rainfall and then we just went through what you do for the infiltration parameters of laying out the land cover hydrologic soil group then infiltration and then associating those with the geometries once you've done that then you just make a run like you've always made and it just works that's really cool gary one thing that i'm not sure i'm 100 on is you said that you can use for determining your infiltration um grid that you can use either soil or land cover or both and you're using both how does it figure out what the infiltration is with both is there some kind of averaging technique or no it can it conjoins the polygon so let's say i have a polygon that is let's just say it's deciduous forest but underneath that polygon there's three different soil types well if i just use land cover only i'm going to get one set of infiltration parameters for deciduous forest in that polygon right but if i say i want to use land cover and soils then that polygon gets broken into three polygons deciduous just citrus forest soil group a deciduous forest yeah so it's smaller and smaller polygons the polygons get smaller based on their uniqueness of both land cover and soil okay and then you have to specify infiltration parameters down at that smaller level so that's why they're in my case there was a lot of curve numbers because it's the conjoining of both the land cover polygons and the source polygons to make a much larger set of polygons that are unique okay got it that's really cool yeah so that's that now let's just look a little bit of results so here i i purposely started this model out dry completely no water in it okay and i do have an upstream inflow hydrograph up this is the dam up here so i have an upstream inflow here but if we start running at first there's going to be water coming into the dam but there's not much rainfall downstream because the initial losses are occurring but if you zoom in further we can start to see while there's water going in the channel right away why is there water going in the channel right well because we said it was land cover type channel and it was 100 impervious so there was no infiltration so any rain that's falling in the channel is going to just be right there so it's going to start to fill up the channel with rainfall in this case now in a real model that would already probably bend water in the channel okay and then we're going to get more rainfall occurring after the initial loss is satisfied here in a second but the thing and remember is if you have a steep surface it may not look like there's any rainfall falling but if you zoom in you're going to start to see well yeah there is water in every one of these cells but and the reason it looks the way it looks is remember that razz only computes one water surface per cell so even though it's raining over the whole cell each time step that water gets thrown into the lowest elevation of the cell okay so that's a limitation so and the larger your cell size the more gross approximation that is okay so if you want to see more of kind of like sheet flow type you'd have to have much smaller cells okay right yeah okay so sure enough it it you know it did the infiltration and we can move forward here and look at the runoff okay and here comes the runoff from there if i turn off the uh the grid for a second you can see it more so we'll just kind of go back from dry and then we'll just move through and here it's getting wet and now it's raining everywhere and then it's running off and there's the runoff very cool yeah all right you make it look so easy gary [Laughter] when you've done it you know uh a billion times talking about it is easy right so it's yeah it's like anything else uh you know just for for the folks out there who are maybe hearing about this for the first time just give it a try try it many times and eventually it'll be easy right you know that's that's exactly it because it's it's all foreign at first but once you try you know learning the pieces parts what makes up how you do the rainfall how you do the infiltration and then running it and getting comfortable with the results then it will make sense yeah yeah so i'm really excited about spatial rainfall infiltration i've wanted to do it gosh like 10 years i think it's really going to be a great feature to improve hydraulic modeling so gary how does real quick before we get to bridges um how does that work with 1d reaches the same way so this is really cool we not only does precipitation infiltration work in 2d but it works in 1d reaches and storage areas i forgot to mention that so thank you for asking that question uh yeah so i realized that people are going to have not just 2d models but combined 1d and 2d and if they're doing 1d channel into the overbanks they still want rainfall on the channel right so what it does is the space between each cross sections is a polygon and it and it figures out an amount to rain and a volume of water for each polygon it first subtracts any infiltration if you had infiltration because you might have infiltration in the overbank right yeah okay but not in the main channel so it figures out the infiltration and gets the rainfall excess it has an area for leftover bank channel right over back it gets a volume for each and then it takes that volume as a lateral inflow between every two cross sections very very cool and for storage areas it has the the polygon that represents the storage area so it knows the area of the storage area so it knows the rainfall it has the polygons inside the storage area for infiltration it computes the infiltration for each of those polygons it sums up the excess times their areas to get a volume each time step and it says that's the volume going into the storage area then yep so yeah you could do 1d you can just do 1d only if you wanted and still do rainfall and infiltration on 1d models awesome awesome yeah all right let's hear about bridges man this is the one everyone's excited about big time yeah this is the thing that has um made me feel that razz was inadequate for a long time because this is the thing that most people would ask me about well why can't you do bridge hydraulics in 2d when are you going to have bridge hydraulics in 2d and i kept saying while we're working on it well we're working on it you know and i i sounded very much like a broken record but yeah um so this is something i'm really excited about and and this is just the first phase of bridge modeling in raz okay and we'll talk about what it is and why it's the first phase and we'll talk about maybe what's going to come in the future but so what we've done is we've taken the concepts and methodologies from 1d bridge hydraulics and we're utilizing those bridge modeling techniques right inside of 2d area and the way you're going to be able to do bridge hydraulics you're going to be able to do low flow pressure flow pressure flow and rear flow and low flow and workflow so the full range of bridge hydraulics modeling okay well why is this important well as we were talking about currently in the current version of raz there's no method to handle bridges inside a 2d floor area so what modelers tend to do bridges are often very important piece of a river system from the standpoint that they may be a controlling factor of the water surface right at the bridge and then upstream and if you're going to model any kind of significant events you want to include the effects of those bridges in the modeling so what people are doing right now a lot is 1d modeling for the channel and 2d modeling for the overbanks so that ends up having you know these more complicated 1d to 2d connections it's more time consuming to develop the model these lateral weirs that we use may cause instabilities because we're using a weir equation with submergence okay so there's a lot of care in figuring out how to do that correctly um the transfer of flow between the main channel floodplain is less accurate because you're using your equation and you don't really know the where coefficient and maybe it changes with time in fact it probably does okay you're thinking about it in terms of rear flow okay and it likely changes spatially too along your lateral structure especially if you're using longer lateral structures for multiple connections right because it's a function of depth it's a function of the angle of the flow going across that point there's a lot of factors okay and i would even say it's more computational time to develop the model but it might even be more time to run the model because of those connections and the fact that it has to possibly iterate more back and forth between 1d and 2d okay whereas if it's all 2d it might actually run faster now that's going to depend on how many cells you put in the channel right so if you have a model that has you know tons of cells in the channel and you just created a huge 2d model well that's going to run slower but you can create a model with fewer cells for the channel it's still 2d cells and have the whole thing 2d now you're not going to get maybe with few cells in the channel detailed velocities but you're still going to get the carrying capacity of the channel the right water surface as long as you have the right end values because with raz sub technology capability inside each cell is the full benefit of the terrain and each space is a detailed cross-section of the terrain so fewer cells you're still picking up the details of the train and the volume of the train accurately let me just let me just tell you gary i've done several of these 1d 2d models because just for that reason because of bridges where we've had to make the main channel uh 1d 2d in the floodplain and the complexity of those kinds of models cannot be overstated or understated sorry it's it is super time consuming it's really hard to troubleshoot as well you get some errors and it's really hard to pinpoint and then once you do figure out where they're coming from how to fix them um boy this is going to be a major game changer for these types of models just being able to reduce that amount of complexity and time spent troubleshooting and even the setup time too to put in all those lateral structures and connect them all up uh trim your your cross sections just right trim your 2d areas this is going to be so much better yeah and the other thing the way to think about this is that if you're not looking for super detailed velocities inside the channel there's nothing wrong with raz anyway because of some good technology of having fewer cells just to compute the flow accurately the water surface accurately and conveyance accurately okay you're sacrificing velocity from that standpoint but you can still get a very accurate model and you might think of it as kind of pseudo 1d when like if you use few cells in the channel but it's still 2d and even that's going to be better than doing 1d 2d because at the banks if you align your faces of your 2d model using a refinement region then you're going to have bases along that high ground of the banks and it's going to accurately figure out when the water comes up high enough to actually start going out of the overbank and the other thing is water can go across those spaces at any angle okay and the velocity momentum is carried across those faces which also makes it much more accurate than separating 1d and 2d because you have the full benefit of the momentum of the flow being accounted for crossing those faces so a lot of significant benefits yeah so to carry that kind of discussion on so there's there's going to be less need for combined 1d and 2d modeling you're going to be able to develop much more detailed 2d models because of this because you're going to do bridges now and there much greater modeling flexibility okay to me more stable numerical solutions because you don't have this 1d 2d connection thing and it iterating back and forth between 1d and 2d more accurate flow transfers because of the reasons we just talked about and quite frankly depending upon how you do it faster computational speed okay so i think it's a win-win definitely very cool um so what's the approach well basically approach is we're going to use the 1d bridge geometry basically layout cross sections but it's done in an automated way which you'll see inside the 2d floor area and from those cross sections and the bridge geometry itself depending upon the bridge hydraulics methods you select it's going to develop a family of radian curves the headwater tail water flow curves just like it does in one day now okay the difference is though is we're going to use those rating curves and for a given flow there's a tail water and a headwater so there's a water surface differential well you can equate that water surface differential into a force and you can take that force and apply it into a momentum equation so what we did is we wrote a special momentum equation just for bridges and it only gets applied at the center line of the bridge faces which we'll talk about a second but basically this special equation it takes from the family of curves what would have been computed as friction loss pressure differential and convective acceleration those forces were already used to compute the curves in a steady flow mode so that water surface differential is already accounting for friction loss pressure differential and convective acceleration okay so when we get that water source differential we equate that to a force that's accounting for those terms the only term it doesn't account for is the local acceleration term which is the change in acceleration with respect to time so this special momentum equation gets the force from the curves but computes every iteration the local acceleration term okay now the other thing that happens is that you can have more than one space so the the force from the curves is distributed across multiple faces based on the amount of conveyance crossing each face and that's done iteratively each each time step each iteration okay so it's still completely 2d it's just getting the forces that are occurring due to the bridge from the curves and the curves are pre-computed to the model running so these bridge bases are solved just like any other 2d face it just has a special momentum equation and so there's really no increase in computational time now gary you said the force was distributed based on conveyance area but right is it is it reasonable in some instances where you know your force is going to be more concentrated in the center of your bridge opening as opposed on the outside so for instance like you know a five foot section in the middle of your bridge is going to be experiencing higher force than the five foot section on the bounding side of your bridge absolutely and so conveyance is not a perfect way to distribute um the forces due to the the curves that were computed in 1d but it's a good way but the other thing you've got to remember is um those cells and the water coming into those cells are still the flow and the angle of the flow is still based on the full 2d equations upstream of that so when the water gets to that face it's already kind of been distributed based on the what's happening upstream and downstream of that so the only redistribution slight redistribution is occurring just at those faces at the bridge okay okay that makes sense yeah so it is an approximation but it's not a not a horrible one by any means okay um basically this is the geometry editor and you can lay this on razmapper i'm going to show you how to do it in geometry but you can do the same thing in razzmapper but but the editing of the data you have to do in the geometry editor as far as putting in the deck and roadway computers and abutments and all that but you basically how you do this is you're going to use the sa2d area connection just like you have now to draw from left to right looking downstream a center line then the only difference is when you bring up the essay 2d hydraulic structure editor is there's this structure type and you can pick now pick bridge and once you pick bridge the editor morphs into a bridge editor and you get the deck and roadway button the peer button sloping abutment bridge modeling approach and the eight hydraulic tables parameters just like you have in 1d for the bridge editor okay we'll demo this in a second but once you've entered all the bridge data and you picked your your bridge modeling approach and you entered in the parameters for defining the hydraulic tables that are going to get computed for the family occurs you run the geometric preprocessor and here's the family occurs that gets developed so this is the family occurs that were developed for this bridge and center of this 2d and so this family occurs is what's being applied to these centerline faces so in this example i've got one two three four five six faces so the other thing is if the water is only crossing these two faces then we might be down at a lower flow and the and the differential on the curves is then only applied to those two faces because that's where the water is so it's based on conveyance as the water gets higher and engages more of the faces then and then some of the force gets distributed to some of those faces from these curves before we go to the demo i just wanted to show a little example because i often get asked well how's this going to compare to doing it in 1d and since i knew people wanted to know that i said well let's just do it so here i took a data set that i had a 1d model for already and i made a 2d model and i put in the 2d bridge okay and here is the same model in 1d with cross sections and a 1d bridge and if you notice here's the bridge curves from the 2d run here's the bridge curves from the wendy run they're the same because it's the same bridge geometry and same terrain there might be a slight difference because the cross sections outside this bridge are a little bit farther apart than what was used in the 2d model we'll talk about why that is in a minute but that might be the only difference in this case okay same end value same bridge data same terrain slight different location of the outside cross sections which might influence us a little bit so here we're going to look at the results and we're going to look at simultaneously the water source profile for 1d and 2d 2d is the lighter blue and 1d is the darker blue and even at low flow the 2d has a higher water source right now i use the exact same n values and this is common that 2d if with the exact same values and train ends up at the higher water surface and the reason for that is there's way more cells than there is cross sections okay and so the 2d model seeing way more of the terrain the contractions and expansions the ups and downs whereas in a 1d model between the crossings it assumes a linear transition a smooth transition and that's just not the case so it's very common for the 2d model if you use the same exact envelopes to end up with a slightly higher water surface so that's why if you're going from 1d to 2d you really have to recalibrate your end values if you have a calibrated 1d model let's go ahead and run this animation and so here we're going to go from low flow to high flow and right now the the water surface is like less than a foot difference and that's just n value difference our friction differences because of the different number of cells versus cross sections but as the water is getting higher and higher you can see they're getting closer together and they're only less than half a foot apart as far as up in here upstream of the bridge now one thing to keep in mind is the actual bridge is not inside the 2d faces we're just using the bridge curves to equate the forces on top of those faces so here we're going to see the water looks like it goes through the deck and it is going to go through the deck but the important part is here is it's getting the right force transfer and force loss from downstream to upstream okay now here we're over topping the bridge both from 1d and 2d and as you can see once we get over top this bridge is controlling the hydraulics and now the answer upstream of the bridge is about the same because the bridge is controlling the hydraulics okay there are some slight differences just upstream and downstream that's because of cross-section spacing and so forth versus the number of cells okay and so it's going through the full transition from low flow to pressure flow to pressure and rear flow and then back again um in the 2d just like it does in 1d absolutely so that's super cool yeah so no problem and it's stable which is the other really cool thing very nice does it in a very stable manner so it's it's it's really this approach uh was developed by myself and a guy named ben chacon who works for rma who helps us work on raz2d he and i came up with this idea for this momentum equation to use the 1d bridge curves so yeah i want to ask you about that gary because what i really like about this is that the way you guys have set this up is going to make it very familiar to people who have done bridge modeling in the past in 1d because it's using i mean the editor looks the same it's got the h tab parameters just like we were using before and unsteady flow modeling on bridges and so it's very familiar which is great because people are going to be much more comfortable trying this out learning it applying it was this approach something that was there like an aha moment that you just said ah this is how we're gonna do it or was it kind of a slow evolution well tell us about the process that got you to this well there's two there's many approaches you can take to modeling a bridge in 2d so this is one approach and it's basically incorporating 1d hydraulics and forcing it on to a 2d face and but the key aha moment here was well how do we do that and the first aha moment was well we're going to take this force from the curves and we're going to write a special momentum equation that that captures that force and as far as the momentum but also captures the other terms required in the momentum equation that aren't depicted in the curves okay the second one was distributing that amongst multiple faces okay and so you you're getting different force across different faces being distributed um based on on conveyance and you know et cetera but the cool thing is it can handle going into pressure flow it can handle going pressure over topping flow and then transitioning back in a smooth manner so as you saw once it started to hit the deck the water surface rose very quickly okay if we go back okay if i can go back once we get to to it's going back too far okay once we hit that bridge deck watch how fast the water surface starts to rise even though the actual bridge deck isn't in physically in the face but the effects of the bridge deck are in the curves look at how quickly it starts rising both for the 1d and the 2d and then the tail water comes up because the you know once the it rises first because a lot of water is being trapped behind the bridge and it's filling up a volume and so there's actually less flow downstream and once it fills up that volume behind the bridge then the the flow starts to increase and the tail water starts to rise and then they both start to rise together um yeah so it's really this is i this came out way better than i thought it was going to when i had it as a an idea and then ben and i sat down to kind of work out the math okay now another approach for 2d hydraulics is something we're going to do and that is where you actually try to you don't use curves you actually use the 2d equations and we have a method i'm not going to go into the details because i don't want to give it away to anybody else because it's kind of very unique that we're going to be able to handle pressure flow and over topping flow straight from the faces at cells without doing any 1d curves or 1d cross sections and we kind of already kind of did some proof of concept of this already so that's the next thing for super detailed bridge hydraulics we're going to work on and that'll be a future version of raz excellent cool you'll have both options then nice hey you know one thing i did i i noticed gary that is not in the bridge editor for 2d areas is the multiple opening option and i imagine that's no longer needed because now we can use a 2d area to serve that purpose is that right well that's exactly it you you wouldn't want to do multiple opening 1d curves anyway because that's an approximation a heavy approximation so instead when you have multiple openings you would just have a bridge opening and then you'd have a covert opening here and maybe another bridge opening here and you just model them as separate hydraulic structures in raz yeah with 2d cells and then you're going to get much more accurate does the water really get over to that opening right okay and it's no longer you know based on a single water surface across the upstream of the the roadway uh depicting that you know okay so it's gonna be much more accurate and just like uh you know a lot of a lot of things when you're going from 1d to 2d uh you no longer have to define where these uh points of flow separation occur uh right yeah you don't need any of that because he takes on automatically in the 2d yeah yeah that's absolutely so let's do a little bit of demo on the 2d stuff from from razz so here's our friend bald eagle creek and and the reason we do so much with bald eagle creek is because terrain models are huge and you don't want to send out software with lots of training models it'd just be way too big of a download so that's why bald eagle creek gets used over and over okay for so many things but it's been a really good data set from that perspective so let's zoom in and this model comes with the software also so you have it if you downloaded ras6o beta you have this model this model actually has seven bridges in it but we're just going to look at one so i had a 2d model this and again the way this works is you come up with the essay 2d hydraulic area connection button you click on it and let's say there was a bridge right here and this is upstream and this is downstream so i'd i'd draw the bridge from left to right looking downstream and i give it a name okay i'm not going to do that but and then what you would get is just a center line then the next step is you want to make the mesh along that center line so you're going to click on it and you're going to say edit internal connection cell spacing and you're going to give it some cell spacing you want along the bridge and let's say 50 feet you'd say okay and then once you did that then you'd say enforce and i already did it and then it makes a really nice smooth mesh so notice it made six cells across from start to end so they're approximately 50-foot cells but they start at the beginning of the bridge and they end at the beginning of the bridge okay and once you have that match carries is just a coin is it just a coincidence that the 1d um lines on either side of the bridge are very very closely aligned with those cell faces does that does that have to be done it's going to it it automates those so we'll walk through that so at this point all i would really have is a center line and i don't have any 1d cross sections yet all i would have is this center line this black and white black and red line no gray blob yet okay so once i hit the black and red line and i would say edit then the editor would come up and it would be blank but the first thing you're going to do is you're going to say structure type because the default structure type is weird gates culverts out rating curve the thing you're used to from the sa2d hydraulic connector but then you'd pick bridge okay once you pick bridge this thing morphs you get the deck and roadway piers about men's bridge wiring approach etc and then right from there it's going to have still just one cross section so the first thing you have to do to get any cross sections is you got to go to your deck and roadway editor and you're going to put in this distance from the upstream side of the bridge to the outside cross section which we have in 1d okay and you're going to put in the bridge width and then there's a default where coefficient of 2.6 where you can change and then you can go ahead and put in your stationing high cord low cord for both the upstream side and the downstream side which i just did really really simple for this bridge it's sloping okay once you do that and you say okay then all of a sudden you're going to get four cross sections so there's going to be an upstream outside cross section there's going to be a cross section inside the bridge but at the upstream end so upstream inside bridge there's a there's actually a fifth cross section called center line but we'll talk about that's a special cross section in a minute there's a downstream inside and there's a downstream outside these four cross sections the two outside and the two just inside at the ends of the bridge are what's used to do the 1d modeling and create the family occurs so it's basically going to do 1d steady flow backwater using your bridge method and those four cross sections to create those family occurs which is just what it does currently in 1d it creates a family occurs in the exact same manner now one difference is there's also this center line the center line is going to actually be an interpolated cross section and the reason is this this red center profile is what's going to used it's going to override the terrain of the faces at that center line it's a combination of what you did for the deck and roadway and the interpolation of the ground inside the bridge at the upstream end and inside the bridge at the downstream and the reason is this we optionally allowed you to go in and edit the terrain of these cross sections so you could put in a scour hole by just editing this internal bridge cross section both in this upstream inside and the downstream inside and so that's why this center line then ends up being an interpolation of those two cross sections because you can manipulate the terrain by changing the cross sections to represent something that's not even in the terrain model if you want okay so the center line again is an interpolation of those two interior cross sections okay so if we look at this outside once i give it the deck and rotary information there's four cross sections that gets built that deck and roadway 50 foot distance says okay 50 feet from the upstream side of the deck i'm going to have a cross section it's an exact duplicate in width and it's perpendicular to the bridge so it just says make a cross section 50 feet away same width as the bridge parallel to it okay then there's the cross section inside at the upstream end the one inside of the downstream end with these red dots and then the one outside also gets placed 50 feet away so this distance in case of 2d is defining how far the upstream cross section is from the bridge but also how far the downstream cross section is from the bridge okay and there's no need to put a skew coefficient on the belt because we're working with 2d modeling and so it's going to take that yeah no no skew coefficient required so that's normal and that those uh those outside bounding sections um are only aligned with the cell faces just coincidentally use the same distance as the size of those cells if i would have said 25 feet that upstream cross section would have been in the middle of this cell yeah it's just coincidence that i said 50 feet and the cells happen to be uh large enough that they ended up there okay but they could be inside the cell or outside the next cell downstream whatever all dependent upon what you entered for this distance parameter yeah so the way this is the upstream bounding cross section is this one upstream inside is this one downstream inside is this one oh that's the center downstream inside is this one and then the downstream outside is that one those are the four cross sections that it uses during the geometrical processing and the bridge hydraulics information the deck roadway appears and the bridge modeling approach to develop the curves terry would you say just given what we know about um how hec-ras 2d reports results along cell faces uh and the fact that these you know internal bridge curves are generated along those bounding cross sections do you would you say it's good practice to try to have those cross sections align with cell faces or does it make zero difference at all computationally um you don't want them to be too far away from the cell the two cells that are bounding the cross-section because the reality of it is what it ends up computing for these curves only gets applied as forces at these centerline faces so if i made that distance really large such that the 1d cross section was way up here and the one outside was way down here well i would be over accounting for the amount of friction loss because there's a long distance on both ends i'd be over accounting for the amount of friction loss and maybe even contraction expansion losses that gets built into the curves that is then only being applied at these centerline faces so you do want to keep these cross-sections relatively close to the bridge and i would say at or within the two bounding cells now it doesn't have to be that way so i can make these cells a lot smaller yeah okay and i can still have the cross sections outside those cells okay but it is going to account for friction loss from here to there into those curves so that's that's an approximation yeah in in 1d modeling the the kind of the convention or the the guidelines the guidance for setting up those in this case this is cross sections two and three the the ones just yes outside just outside the bridge it's always been to put those at the toe of the embankment so that right you're as close to the bridge as you can without actually having the bridge structure itself part of those cross sections with the same guidance supply here yeah if you notice the toe of the embankment here's the embankment notice that this is right near the toilet embankment which is why i picked 50 feet okay both on the upstream and downstream side now the other thing you might notice is i've only made the bridge the actual bridge opening why didn't i make the bridge go all the way from here to way out here well i could have if i wanted to however i think it's going to be better just to limit the bridge to the bridge opening because i could do something else out here i can have these faces on top of here and i can just model this flow going over with 2d equations or i could have another structure over here that i use the the standard essay 2d structure for where i just use a weir and the benefit of doing that is i can model it as a weird flow if i want to but i can also breach it so the bridge option doesn't have a breaching option right now but the so either it's uh even though it's an sa 2d area connection that you can breach once you put a bridge into it no the breach no longer works for we compute the curves a priori and we can't change the curve dynamically with a breach got it but if i want to breach this embankment over here over here i just model them as a separate essay to these structures with the standard weir type of approach and then i can have a breach there so i think this is kind of a good approach where i limit and you don't have to do this though i limit the bridge to the opening and part of the abutments okay so there's fully i've got the full abutment in there okay right but the rest of this a roadway approach i'm just modeling with 2d faces or i could have used another sa2d structure here and another one here or i could have done the whole thing as one bridge all the way across and it still would have worked because it wouldn't it wouldn't engage this part of these faces until water was going over the roadway yeah but doing it this way gives you a lot more flexibility for i think it's other things more flexibility and also more accurate because then you're limiting to the influence of the curves to this this domain yeah and this is not influenced by the curves and obviously you still have the capability if for instance you're doing a low flow model maybe your bridge doesn't have any peers you don't really anticipate that the bridge is going to have too much of an impact hydraulically you still have the option of just leaving that out and just having the terrain control the flow through that opening as well absolutely so that's a really good point that if you could if you have the another thing that i think what you were saying and correct me wrong is it let's say you the terrain has the peers in it then or there's no peers yeah or if there's no peers right then you don't have to have them in there but let's say you want a more detailed model and you want to see the flow going around the piers well by doing it this way the piers are just and the pure losses part of the curve so you're not going to see like flow separation around the piers but you could still have a bridge here to get the deck part but don't put the piers in here put the piers in the terrain and then have much smaller cells to represent around the piers and then you've got kind of a great combination of super detailed model for low flow and flow going around piers but when it starts to hit the deck the curves would take over for the pressure flow and overflow part yeah you would just want to know probably in the descript you would just want to note in the description like hey you know peers for this bridge are including the terrain not in right and so if the peers are in the train don't put them in this editor because you're double accounting the loss then yeah okay good yeah yeah so lots of options here lots of things you can do and tremendous amount of flexibility from simple to detailed modeling so really cool stuff yeah that is that's awesome so let's look at some way to tell that yeah let's do it so i said this this data set comes with the raz so you have it if you download the beta version the test data sets here we're at that same bridge is right here but there's a bridge just upstream here i'm plotting velocity okay on water surface and and we can you know we can obviously turn on the velocity tracers and we can see what's going on and so the cool thing is there's a couple things i want to show here first of all this bridge is heavily skewed yeah right but if you see the tracers are going straight through yeah and the things that are influencing the water surface this corner underneath this you don't see it but underneath this gray area is the abutment and so the water has to turn to go around the edge of the abutment and it turns into the flow field then it comes out and you have this you know circulating water here this eddy zone okay because of the corner of this abutment and the velocity but the so it's the velocity is you know being pushed a little bit this way and the higher velocity is in the center and then down here this there's an abutment sticking out down here that's causing this flow to go more towards the center so the flow along this edge is slower again because of this abutment so the full 2d factors are still in play here because it's solving this fully 2d the only difference is that along these faces we have the special momentum equation where the force from that crossing those faces is from the curves plus dynamically accounting the local acceleration term each iteration of each time step so really cool stuff yeah yeah yeah and a bridge like that would always give me fits like what do you use for your skew angle if you're doing it in 1d i mean it maxes out at 45 but that's obviously more than 45 and yeah that's uh and the other problem is the the angle of the flow can change during the event so at low flow it might be turning more but as flow gets higher the momentum might carry it more straight through that'd be interesting to see is this an event model or is this just cause i setting yeah i'd like to see let me turn off water surface here so we'll zoom back in here in this case the channel is pretty much going this way okay yeah but you can still see a little bit of a of a you know of a change so there's a small effect here from the abutment there's no eddy obviously now as we start increasing flow whoops yeah it was it was moving it was just going okay i'm just not far enough yeah yeah okay and now we'll stop there and turn on that so now we've got more flow there's some contraction more contraction occurring in a little eddy forming a little slow down over here let's keep going stop velocity tracers more influence pushing in bigger eddy higher velocity yep very cool yeah so definitely you know and if i wanted even more detail i'd be putting smaller cells upstream of this as as well as through the bridge depending upon what level of detail i wanted to depict and the intricacies of what's happening with that water and velocity changes through that reach okay here i have fairly large cells upstream so it's kind of a simplistic from a velocity perspective up here until it gets to the bridge where i have more cells then oh man there's so many things i want to do to play around with this i want to put a breach in there i want to make it over top everything yeah that's that is so cool yeah so here's another cool thing i just modeled this bridge but remember i said well i line the faces with the high ground here yeah so notice that this water is turning and going through right so all this water that's out here is contracting i don't think this gets high enough to over top that face no it didn't i should have made a higher flow but i could have had a higher flow and then it would start over topping that and then you'd see a whole different kind of flow distribution then yeah you'd lose your mac in that area that's that's the important thing there is to have those faces aligned to that uh up approach way the center line of it so you don't get that leaking through there right all right yeah i purposely added added a brake line here and a brake line here in addition to the bridge yeah so just putting the bridge in won't do that so people need to to put those extra brake lines in there too yeah right absolutely so that's 2d bridges love it oh man that's that's really cool gary well uh before we get into the next topic that gary's going to dive into i just want to let everybody know um we're going to have the rest of this particular discussion in a second episode that's going to be linked in the video you guys are watching right now so if you want to continue listening to the discussion move forward we're going to be talking about wind forces 1d finite volume pump stations for 2d improve improved computational speed as well as what's next for hec-ras feel free to click on the link in in the video here to continue watching the discussion before we conclude this particular video i just want to plug the fact that chris and i will be teaching our winter 2021 1d2d hec razz class which is going to start february 17th and run through the 24th of march if you guys want more information on that class please visit chris linkedin page as well as the kleinschmidt linkedin page there'll be lots of information coming out on that class shortly here we expect that class to be very very very popular we're going to be conducting the class in the new hec-ras 6-0 software so there'll be an opportunity to learn some of the new features of 6.0 as well as just a really in-depth foundational 1d 2d modeling discussions that chris and i get into so again if you're interested in that class please see our linkedin pages there'll be more information on that shortly um and uh but yeah if you're interested in following the discussion please continue to watch on the on the next video here and yeah with that we'll go ahead and close this one out and pick up the next video thanks everybody

Channel: The RAS Solution

Views: 7,262

Keywords: HEC-RAS, International, 2D modeling, 1D Modeling, HECRAS, Engineering

Id: 5zszZC1XcQU

Length: 106min 38sec (6398 seconds)

Published: Thu Jan 07 2021