0:05
Good morning, or good afternoon, or good evening, wherever you are in your time zone.
0:10
This is the second day of the 2022 online workshop.
0:17
I'm going to wait just a minute we have a lot of people that are logging on to the goto Webinar platform, so just wait a minute here.
0:27
More people join.
0:43
Numbers are coming up.
0:49
So, today is the second day, And, we're delighted that you're here.
0:59
To begin with, as I've mentioned, yesterday, we have a what workshop web page that has all the information that we'd like you to know are, and also handouts, and links to the videos, instructions about installation and things like that.
1:21
So, we do update that fairly frequently.
1:25
So, please, you know, keep that book Barker, go back to that if you have a question, chances are you might find the answer there.
1:36
Um, actually, here's a screenshot of part of that. You can see that.
1:42
We have put the, the handouts, the introductions, um, and, for the first day, it actually, I think, in-between the time I made this screenshot.
1:54
And, now, I think that it's been updated to include the handout for today, which actually, are these, this presentation, this introduction, that I'm giving.
2:05
There may be more handouts that we post, um, throughout the day today, as well.
2:13
Note that the video is take awhile to be processed after the session is over, so, you can see, generally, we have a bookmark for the video, and then maybe 2 or 3, 4, or more hours after that session is over.
2:29
The video will be will be posted in that it takes even longer.
2:32
Often when the transcript is ready.
2:35
And again, these are the these unfinished transcripts, they're just machine generated, that we know there's lots of errors in there. We'll try to fix those up.
2:43
But we will post them in case started up fluff.
2:47
Um, so here's the overall workshop we're in this second day, or you know, the first day zero was installation day.
2:58
Yesterday was the first real day of the of the workshop and we talked about meteorological data and trajectory, calculations, and today, we're going to be talking about air concentration calculations and dispersion calculations and go continue on.
3:15
With that in mind.
3:17
More sophisticated ways on the third day and then, and the fourth day with different applications.
3:25
Um.
3:29
As we've mentioned yesterday, there's two ways that you can ask questions.
3:34
one would be in the goto Webinar, Um, platform, you know, you can try to ask questions. There are. We can't always give great answers, too detailed.
3:46
Questions, there's a lot of limitations in that question and answer box: We preferred, if you can, if you asked questions in the high split forum, Um, Because there could be a lot more details. You can, you could provide screenshots.
4:03
You can, um, do special kinds of formatting whatever you know, you want to do to ask your question in, There could be a much more of a back and forth then.
4:13
More importantly, the question is there, so that if other people have the same question, they can look at this and say, oh, yeah, I have that question. I'd like to see the answer to it.
4:23
So, if you're able, and if you're willing, if you can ask the question in the forum, that would be great. You could look at the forum without registering.
4:32
But if you want to ask a question, you have to register.
4:35
Um, it just takes a second and, well, little more than a second, but it's fairly easy.
4:42
And then, actually, you can ask questions in the future. And you can, you know, this is a place for abusers.
4:51
Ask questions, and we try to answer them, even, you know, after the workshop as well.
4:59
So then a quick recap, then up logistics.
5:02
I've talked about the questions, but the handouts.
5:06
So, for example, these slides are available both on the website.
5:11
And as a handout in the goto Webinar interface recordings are available, you know, soon after.
5:20
one thing we mentioned yesterday, that, no, this is a hands-on workshop.
5:25
We, It's, we're hoping that you're kind of being able to do the modeling along with us, but we recognize, it's really quite difficult sometimes to keep up with everything, and it's even hard for us.
5:39
Sometimes.
5:41
And so, know, if you didn't get the model installed yet, or if you get behind in a certain section, something's not working perfectly, Um, it's totally fine to just sit back a little bit and just watch what's happening as a demonstration and then maybe catch up with the next section or catch up later.
5:59
Remember this, the tutorial that we're going through, it's something that's available to you anytime it's. You know, you can do this after.
6:09
Um.
6:10
No in between sessions. You could do this after the workshop is over, you can do this a month from now.
6:17
The tutorial is designed actually to sort of be done on your own in a self paced environment.
6:22
What we're doing in this workshop is kind of going through it, and we're trying to provide additional explanation, you know, as we can.
6:30
But in its essence, for basically just going through the tutorial, which which can be done at anytime.
6:37
So, um, to get into sort of what we're going to be talking about today, um, you can think of two different kinds of high split simulations.
6:50
And I'm going to turn off my camera to give more of a screen screen screen space for everybody. There's a trajectory, some simulation. That's what we were doing yesterday. and you could think of a trajectory as kind of the center line of a plume.
7:08
Not exactly the central hind, but, um, it's basically the central line of a plume.
7:13
And, and this can be really useful if you're trying to figure out, you know, the general direction where something might be going, or if you're doing it backwards, plume, figuring out where's the air came from Medicaid. So you can go forward or backward.
7:29
And we spent, you know, a lot of time, there should be going through these trajectory calculations at one point and what the mission is.
7:34
That trajectories don't factor in any kind of deposition or chemistry or any kind of transformation that might that might be occurring.
7:42
Um, and you know, that, you know, if you have sort of a source of pollutant, and you have a trajectory going, say, downwind, just because the trajectory doesn't hit like a city or a population center.
7:57
It doesn't mean that the plume won't hit that because, as we know, the plume is going.
8:02
Disbursing, as X goes down when.
8:04
So, again, it's just sort of the center of the, of the air transport.
8:10
So today, we're going to be getting into the idea of the concentration that dispersion runs, and, um, you can think of it now as sort of the full three-dimensional transport and dispersion of the plume.
8:23
Um, it includes the transport by the wind, which is sort of like direct trajectory calculation, but it also includes dispersion around around that central line.
8:36
Um, and, most importantly, it gives you these air concentrations and that can be very important if you want to compare against public health threshold answer, or if you're looking at the deposition, two ecosystems, and what have you.
8:56
So, and with, with this more sophisticated, um, type of simulation, you can include chemical transformations.
9:07
You can include wet and dry depositions.
9:14
Um, see.
9:18
So, you know, at its core, what high split does is it just transporting the so-called particles as a blown along by the wind, and we talked about, Sorry, oh my god! We're supposed to get other slides! We're only seeing the first page of the handout. Oh, my gosh!
9:40
Oh, my gosh.
9:42
I am so, sorry.
9:45
Um, huh, my apologies.
9:52
My, my big apologies here.
9:56
I had forgot that two, tell the program to, Um, thank you, Sonny.
10:04
OK, so this was the schedule that I was talking about.
10:08
Apologies.
10:10
This is asking you the questions that I've mentioned.
10:14
You can download these slides.
10:18
A quick recap of logistics.
10:23
Then these two different kinds of hyper simulations, right?
10:27
Didn't make much sense without seeing some of these pictures, but, again, that was talking about the trajectory being the central line, and then the concentration regulations being the plume around, around the central line.
10:44
Uh, see, this seems to be.
10:48
All right.
10:49
Let's back up here.
10:51
At its core, then, the, I split modulus transporting these particles down downwind. And this is what we did.
11:01
Did it yesterday, and with the, um, the concentration calculations.
11:06
What we'll be doing is we'll be releasing a lot of particles, and each one of them is going to have a turbulent component added to it, and we're going to generate a cloud of these particles.
11:18
And each each one is going to be basically A, uh, trajectory, but with an added, um, dispersion component.
11:33
We talked yesterday about the met data for, for high split and just reminding people that you need the metallurgical data to run high, split. It's required.
11:42
Um, it has to be supplied in this high split format, we call it the REL pack format.
11:49
And, um, we have data that you can download from our website, and also, if you're generating your own map data, you can convert it.
11:58
We have conversion programs.
12:01
We have a lot of different datasets that are available, both for the continental United States, but also globally, as well, um, and that for any metadata that you're using, it's going to be graded.
12:15
It's going to be output of a model, high split will interpolate spatially and temporally to estimate the wind speed and direction and other variables that it needs at any given point in the simulation domain.
12:28
And so you can imagine that if the coarser data that you have, the more interpellation that has to happened and the more the sub grid phenomenon could could be important.
12:41
And even if the model was perfect at the grid points, it's not going to be able to give you the perfect answer in this, the in the sub grid in the areas between the grid grid port. So, I split will do, the best it can. But at its core, it's just moving these particles along with the wind.
13:01
And maybe adding some kind of turbulent dispersion, which we'll be talking about today.
13:06
We didn't talk about much about this yesterday, but you can have multiple make a logical grades.
13:10
You can have a finer grid, and then, of course, or grid, and we use both of the grids, will try to use the best graded can then when it goes off that, find our grid, and in the course a grid is available, then it'll start using that.
13:27
That's frequently what you can do is you can have a finer grid near your source, and then, of course, the grid to let the transport continue on past that grid.
13:38
And, of course, there's difficulties with complex terrain.
13:40
Currently, if you're in mountains or coastal areas and things, then you have to be ready, cautious about, um, whether or not you're metallurgical data are really capturing all of the phenomena that are recurrent.
13:57
Um, I think I've gone through a lot of these points.
14:03
The data are in binary form, and actually, when you're, when you download that, you have to make sure you download it in a binary format.
14:12
It can be a little bit complicated, because the first parts of these files actually look like they're ASCII.
14:17
And if you have your FTP browser or download setting to be sort of automatic, sometimes it can be fooled. And it can look at the first few lines of the file that says, oh, this looks like a text file.
14:31
And it'll start downloading that as a text file, but then it'll be corrupted because most of the file is binary format.
14:38
So, when you're downloading data, make sure you put binary format.
14:46
And, again, the metadata has to be available when you're doing the run.
14:53
It has to be in, you know, covering the area that you want to do the run, because once the particle or the electric goes off the grid, it's terminated.
15:02
So, um, no, that's one of the most common errors when you're running high split is that you either haven't provided, you know, the right location of the right file name, and that data and, or the metadata isn't in the right area.
15:19
Or, or, it's in such a small area that, when you do your run, quickly, goes off the back mic rather than when stops.
15:29
Um.
15:30
So, here's what we're doing today.
15:31
We'll be starting out with the air content, an introduction to the air concentration calculations and then we'll be getting into more sophisticated treatment of that.
15:43
We're going to be trying to simulate this cap dex tracer experiment in a more detailed way.
15:53
Then we'll be looking at sensitivity of E simulations to different parameters, that you might set in, in the model.
16:02
Well, then get into different ways to display the results.
16:06
Then, if time permits, we'll start getting into the Pollutant Transformations and Deposition section at the end of today.
16:13
If we don't get to this today, we'll start this tomorrow.
16:18
OK, so without further ado, I'm gonna now switch over to this first section and let me change the, which screen I'm sharing.
16:35
Here we go, and apologies again for sharing those slides at the beginning.
16:44
Um, In any event, so here we go.
16:48
We're starting in section seven of the of the tutorial and if you're just joining us, we hope you've downloaded the tutorial, and you click on the index dot HTML, um, file and the tutorial directory to upwell pop the tutorial in the in a web browser.
17:12
Um, so we're going to start this concentration calculation. So let's begin.
17:23
By opening up, yesterday, we were working all in the trajectory section.
17:29
Today we're starting to work in the concentration grid.
17:32
So let's first we'll push Reset to kind of clear the GUI out the Graphical User Interface out to make sure there's nothing in there.
17:42
That might be confusing or leftover from things you might have been doing yesterday.
17:46
Envelope it up the concentration grid.
17:49
And just like with trajectories, we're going to do this setup run, and when you click Setup Run, up comes. This looks pretty familiar.
17:58
Um, it looks kind of like what we're doing with the trajectories, and this initial part is very similar to the trajectory calculation, but you're going up. We're going to find that there's a lot more things that you can do with these other other venues.
18:14
So, the first thing we're going to do, is for going to, let's see that medium, pull this over.
18:23
Um, one's a little harder.
18:27
Organize the screen ear, right?
18:36
I guess, you guys are probably also in trouble hitting your screen, organized, it's lot of things on, it happened in one place, OK.
18:52
The first thing we're going to do is.
18:53
We're going to enter in the starting time.
18:58
Simulation in 19 83, September.
19:06
September 25, UTC, 17, and just like the trajectory is all the times in high split, are UTC.
19:14
We're gonna set up this starting location.
19:19
This was the release of the tracer during this experiment, Dayton, Ohio 39.9, oh, point 22, 10, 10 meter aye, above, above ground level.
19:38
We're going to set that.
19:39
We're gonna keep the runtime, but at all of ours, we're going to be going forward, top of the model model is find to be 10000 meters above ground level. The vertical motion is its default, which we normally use, which is using the the input and theological data.
19:57
Then we're going to add the meteorological data that we need for this and to go clear clear out this this other file that was in there and we'll click the Add mythological Files and you navigate too.
20:13
You're a tutorial.
20:15
Uh, Directory.
20:18
And you go to the cap taks, full sub sub directory in there, and we're picking this cap text to work pretty 70 UW dot pen ballots, We were using a lot, yesterday.
20:32
We used that a lot, today, unless we're trying to investigate the impact of using different different metadata.
20:39
And so, then that should show up here and, um.
20:46
Now, we're going to save, and as you remember: yesterday, you have just, you know, once you've set the model up, you have to save it and that writes a so-called control file.
21:01
Um, and, that's what Haislip needs to run.
21:06
And if you don't save this step, that file is not written, You won't be getting what you think you're getting, So you, you save. Save this.
21:15
And now, just like you were doing yesterday, with trajectory, you go, concentration, one model.
21:23
And it goes a little bit slower, and I'll explain why, but it's not quite as fast as as it was yesterday, and, um.
21:35
We go, by the way.
21:36
If, um, when you get that up popped up prompt that said, setup dot CFE nameless file found, you would want to delete that file Because it is leftover from yesterday.
21:52
That would include somebody advanced configuration things, which we'll be going into, but we haven't done.
21:59
Yeah.
22:00
So you exit this.
22:01
And now, just as with trajectory, you go to concentration display.
22:08
And now you pick concentration.
22:11
And we're going to mainly be doing these contours.
22:14
There are other display things with which we'll look at later.
22:19
But this is going to be the main display that we do.
22:23
Um, you click on that.
22:26
And just like yesterday we're going to more or less accept the defaults on these.
22:30
Although, um, you can see this is a much more involved menu.
22:37
There's a lot more things that you can pick when you're displaying concentrations. We will go into these later.
22:43
But for now, just just accept the defaults and execute display.
22:50
And you should get something like that.
22:55
And so, I'll make this a little bit bigger, and you can see a Now that looks like a plume. It's not just, it's not just a trajectory.
23:07
And, uh, what we were doing yesterday, we were running these trajectories, and this is just the first 12 hours, and before we were running, I think for 68 hours mostly the full duration of the tracer experiment. But this is kind of going in the same direction that the trajectory had gone.
23:26
You can see we have these contours of, of concentration as you might expect.
23:34
It's higher concentrations near near the the beginning and and then slightly less concentrations and then it gets even even less and, you know, it's fairly dramatic drop off if you think of these as being factors of 10 and You can change the scaling on this. You know, actually back over this menu.
23:54
You can change that difference, the way that scaled in, you can do all sorts of things in, in this menu, too, to display things in, in different ways.
24:06
So, I'm going to quit out of this, and, um, Um, Let's see.
24:15
I think that that is what we were looking at here.
24:22
one of the things they were talking about here is, of why, how you got that concentration, and if you can think about it.
24:33
We actually released 2500 articles to do this simulation.
24:42
Didn't even show you that yet, but that's sort of the default number of particles that you released, and we released 2500 of these little particles, and actually, let me explain for a second that.
24:55
When we talk about particles in the high split model, um, they're not actual particles that you would think of as atmospheric particulate.
25:05
So these aren't, know, the PM 2.5 of the PM 10, or the, or the particles like that you would, that you would breathe of.
25:11
The particle is a, is a computational point.
25:16
And it could be a gas, it could be a particle, could be a water droplet, It could be a collection of things.
25:25
It's a, it's a computational point.
25:28
And, in fact, in almost all cases, it does represent, you know, sort of a vast array of things that have at that location in the atmosphere. So, it's a bit maybe unfortunate.
25:43
We called them particles, everybody seems to do that.
25:46
You know these this is called the crunchy and particle dispersion model.
25:51
The particle name in that term, just refers to a computational point.
25:58
So, in this simulation, we released 2500 of these computational points, and they were blown down when each one of them was sort of a trajectory.
26:10
But, there was also an added dispersion, a turbulent dispersion component added onto each one, which we'll be discussing in just a second.
26:19
But you can think of the concentrations.
26:21
Then, as we have a grid concentration grid, and when a particle computational point gets into one of these grid cells, then you simply take the mass of that particle And divide by the concentration, or the area box, gives me the volume of that grid cell.
26:43
And, um, here, we have a little bit of a calculation showing that, you know, if we released one unit of mass divided into 2500, different computational points, that each zero point would have 120, 2600th of a, of a unit of mass.
27:05
And if that, if the grid cells suppose we had a five, this great actually was five kilometers by five kilometers and actually it's 100 meters, not one kilometer, Um.
27:19
That.
27:19
the volume of a grid is, um, approximately, um, this size.
27:31
And actually, it shouldn't be 10 to the third.
27:36
It should be about 10 to the ninth, um, 10 to the eighth.
27:42
Sorry about that, Typo there.
27:44
So, if you have 125, 200, then you're basically going to get units of 10 to the 14th, I sorry.
27:52
Yeah, one.
27:54
But, 10 to the . The 10 to the minus 13th, or 14th?
27:58
Because, um, if you're talking about a 12 hour average particle would have might've been in the grid cell for part of the time, but not for the whole 12 hours.
28:12
Um, you can see you can see, this is sort of how the calculation is done and I'm going to go into more detail about this.
28:18
In fact, we'll go into a little bit more detail.
28:24
In a second, but let me just go first, too a little bit about how these calculations are done.
28:31
Um, the, they are concentration calculations are, um, done based on the turbulence in the atmosphere.
28:46
And yesterday, we were talking about that particle, or a trajectory position at, at a T plus delta T as, um, as being just the particle position at the, um, beginning plus the velocity Times the time step, and when we're, when we're adding the turbulent component, two, these, these transport equations, what we do, is we we calculate the mean transport. So this would be the trajectory type of transport.
29:28
So this is just the mean wind times the time step, then we add this turbulent component to the turbulent velocity times the time, step two to the position, and, um, you can think about this in different ways.
29:48
You can think of the velocity the actual instantaneous velocity in the atmosphere is being A An average.
29:56
A U bar is the mean wind, plus interpolate component.
30:02
If you re-arrange this equation, you can get the turbidity component equals the instantaneous measurement minus the average.
30:10
And that's actually, if you have an anemometer, and you're measuring wind, what you're measuring is this, um, instantaneous component, And if you, if you subtract out the average wind, then you actually get to calculate what the turbulent component of the wind would be just from those measurements. And you can create statistics from that.
30:32
That's why people talk about the standard deviation of the wind speed.
30:38
You can also look at the standard deviation of the wind direction.
30:40
And, those measurements that you make are actually giving you some measurements of the turbine component of the wind in the atmosphere, and we use those turbulent components to estimate the dispersion of, of these computational points around the, around the central line.
31:06
And, um, there's a lot of different ways we do this. It's, it's fairly complicated.
31:12
And frankly, this is the in a way this is the the essence of of a dispersion model is how do how do you estimate these, this turbulent component? Because this is getting giving you that spread.
31:23
And you can imagine that it depends on a lot of things.
31:27
It depends on all sorts of different aspects about the of the atmosphere, in terms of the stability in terms of the wind generation.
31:38
The wind phenomena all sorts of temperature effects and and touraine effects, and so it's quite complicated the way Haie split or any model tries to estimate this turbulence.
31:55
And this is just, sort of a brief outline and we have, um, a more detailed description of the equations that are used in, in the document section of yore of your installation.
32:08
And in that RL 224, that Chris Locker mentioned yesterday, there's more detailed descriptions of how we get that.
32:16
one of the things that's used is A an auto correlation coefficient, which is this R, which is a function of the time-stamp.
32:26
and this basically this autocorrelation is A a concept that says if you've got this turbulence going on.
32:34
And you've got say a little updraft of, of, of an air parcel, that chances are, if you're looking No, a second later, Then parcel is still probably going to be going up.
32:51
So, it's sort of auto correlated with, with the way that it was moving recently.
32:57
Longer youth, go forward in time.
33:01
The less that the air personal emotion is auto correlated and, you know, after your weight, after a certain amount of time, maybe NaN, or a minute or two minutes, then, maybe that person will be going down.
33:13
And so, um, that's why this auto correlation, No, it's when it's close to one, then you're pretty much getting the same velocity and, um, when it's close to zero, then you're getting this extra.
33:32
Know, then when the water close to zero, then that this is, first term goes away, and now.
33:38
You're getting this extra turbulent component, which is then calculated by, um, um, the model, in, different ways, based on the turbulence in the atmosphere.
33:49
And this article correlation coefficient.
33:54
It's generally expressed as, A, an exponential of the time step divided by this Lagrangian time scale.
34:03
And the Lagrangian Timescale is sort of the timescale on which these air parcels are the characteristic time that they are correlated with, with one another. And actually this, Lagrangian Timescale varies depending on the nature of the turbulence in the atmosphere.
34:23
Um, so one of the things that we do in the model is, you know, where we estimate this random component, based on a distribution of, of the turbulent components of the wind, But we pick a, a random number, the model picks a random number, and it just picks a, a value from the distribution to add to any given particle.
34:54
So, when we were doing 2500 particles, each, one of them, was basically a trajectory.
35:01
And, then, it was adding this, yes, random component, to wait, which is why you ended up getting a Cloud and not all of them following the same, um, uh, the same path, um, now, there's two different ways that we do dispersion in, in high split.
35:24
In the old days, actually, before maybe till maybe 10 or 15 years ago, we actually used what are called puffs and high splits still has a puff calculations.
35:37
Um, and you can use pops in some.
35:40
Some models also use pops the model called ... or some other called ....
35:46
There's other a puff models and but the puff puff is basically a collection of these particles, and it's computationally more efficient, um, because you're, instead of simulating, say, 2500 of these, you know, little computational points.
36:07
You're simulating a cloud of, of particles, and, and in fact, as you might imagine, the random dispersion of those computational points shouldn't be related to the growth of this tough.
36:22
And this is sort of expressive conceptually here, but there is a relationship.
36:28
And in fact, if you do run, I split in the puff mode and you run it in the our three-d.
36:36
dimensional particle mode with sort of the same turbulent parameter, dramatizations, with the same met data, you should get approximately the same, the same answer.
36:47
Um, um, and then, um, whether you're doing a, a three-dimensional computational point, no collection of zero point simulation or a puff simulation, the air concentrations are, are calculated by figuring out how much mass is in each concentration grid, square L square.
37:15
Grid volume, and then dividing, simply dividing by the volume.
37:20
So it's actually fairly easy calculation. It's really almost leakage the bookkeeping calculation.
37:25
I know I don't want to oversimplify what Heisler does, but it's basically moving, these particles are puffs around, through a concentration, grid and just adding up the mass in each, in each bit, squelch square for each, each time step.
37:44
Now when you have, um, a puff and actually we have two different kinds of puffs.
37:52
There's A, a top hat puff.
37:56
It looks like a hat, which is basically, meaning that the, the concentration is zero outside of the path.
38:04
And then it's constant cross sectional area of the path.
38:11
So you can think of this as a cylinder, like a coke Coke can or something, um.
38:17
And then we also have what's called a Gaussian off where the concentration is distributed in a Gaussian way and this of course is probably more realistic of what the concentration distribution would be actually be in a half, meaning that there'll be more, probably near the center line and then as you get closer and closer to the edge, you'd have less and less, less and less concentration.
38:44
You when you run high slit width in the puff mode, you can either pick the Gaussian puff or a top hat puff. As you might imagine, the top hat pops are easier to calculate and easier to.
38:57
So that runs go faster. When you use a top hat puff, and it goes a bit slower when you use a Gaussian.
39:04
Tough. Partly because when you're using a Gaussian puffy, you have to calculate, well, where are we in the distribution, and then how is that intersecting with the concentration grid? And it's just an extra set of calculations that are required.
39:17
Um, um.
39:23
Let's see.
39:25
Let's go on now, and let's start easy doing some examples of how this works in reality.
39:32
And so, the first thing we're going to do is consider A a single particle, moving through a concentration grid, and, um, instead of just a 2500 number, going to see what this looks like.
39:53
So, let me open up this data entry menu.
39:58
Let me try to organize my screen a tiny bit, differently.
40:04
She's not, need a slightly bigger screen.
40:08
And maybe you roll thinking that two, Um, So, what we're going to do then is we're going to open up the concentration, um, Set up run menu.
40:23
And, um, that, we're going to leave the, starting time the same.
40:31
We're going to leave the set up locations, the starting location, and height the same runtime the same.
40:41
And now we're going to open up one of these new menus.
40:46
And this is where the concentration simulation starts to get much different, then the trajectory simulation.
40:52
So, when we click on this.
40:55
But here the pollution deposition and grid setup, up comes this new, kind of, menu, and all this down here, a little bit.
41:06
And the first thing we're going to do, um, is we're going to click on the Pollutant, and it when you you have to just hover over the area on this menu.
41:18
It doesn't light up or anything so you just hover over the species' one and click on it.
41:25
And up comes this pluton menu.
41:30
and when you're defining a pollutant is a four character, um, identification name that you can give it, we're just going to call this test for now.
41:40
And the default, we hadn't discussed, but the default is to emit one unit per hour for one hour, um, And, that's what that last calculation was, was, was based on.
41:53
And, you'll notice that there's these zeros here in the release start time, and we actually didn't talk about this yesterday, much, but whenever you have zeros in high slit for a time, what it does is it kind of figures out, well, what would you probably want to happen?
42:11
And in the case of zeros for emissions, what does it say?
42:17
Well, you probably want the admissions to start at the beginning of the run and that's what, um, that's what it does.
42:25
If you had zeros or the start of a of an actual simulation, it would start at the beginning of the net file that you picked.
42:37
If you had zeros at the start of a back trajectory, it would start at the end of the map file that you picked.
42:46
So you can use zeros sometimes to your advantage.
42:50
No, do not have to kind of over specify things.
42:56
Generally, when people do simulations of emissions, they want the emissions to start at the beginning of their run, but, if you didn't want them to start at the beginning, then you could put a different, a different time in here.
43:07
So, what we're going to do, is, we're actually gonna change this around.
43:10
And, we're going to emit 100 notes per hour, are only good in it, or 100th of an hour.
43:22
This will then emit the same amount of mass, it will still emit one unit of mass, but it will emit it very fast.
43:31
And since, um, the smallest time step in high split is one minute and this hundreds of an hour is less than a minute.
43:41
What we're basically telling haislip to do is emit all of the mass in the very first step because we just want to see one particle emitted in that first time step and then we're going to follow it alon. Alright, so we save.
43:57
Save this and, um, now we're going to open up the grid menu.
44:05
And this, I started talking about this a little bit.
44:08
Before it enters this will start to make more essentially you sort of hover over this grid menu and you click on grid one and up comes this, this next menu here.
44:21
I can put it, because I'll put it here for now.
44:24
Um, so um, just as with these other ideas about the zeros when you have this grid. You can define the center, the spacing, a whole bunch of aspects about this grid.
44:38
When you say the center is 0 0, then high split says well. What would you really want here? It probably you probably want the center of the grid to be at the initial point, and that's actually what it does.
44:48
So when you put 0 0 here, its center is the concentration grid at the emission location.
44:55
You can set the grid spacing.
44:57
Um, the porthole spacing in latitude and longitude degrees.
45:02
So this is zero point zero five degrees, which is about five kilometers in each each direction.
45:10
You can set the span the overall, um, width of the grid in longitude, and latitude, or actually love latitude and longitude.
45:22
You can tell where you want the output to go, and as we talked about yesterday, this dot slash is actually just stuck working directory.
45:29
Normally, that's where everything goes when you're working in high split, um, and that's sort of the default value.
45:37
You can give the name of the concentration grid.
45:40
And one thing I'll mention about the concentration grids in high split is that these grids are binary. When we were doing trajectories yesterday.
45:50
Um, the output files, the T dump, the trajectory dump files for an ASCII text If we could look at them, um, the concentration files are binary and partly because there are a lot bigger. You can imagine if you have a three-dimensional grid with all these grid squares or grid volumes.
46:12
The files get very big very fast.
46:16
And so these are actually packed.
46:18
So for example, they won't necessarily unless you do something special, um, include all the zeros.
46:27
So there's different ways that these are unpacked to make them smaller, um, that we do have utilities to, you know, to read these files and you can process them in post-processing, and you can export an ASCII version of this.
46:45
So, you know, all hope is not lost, but, um, you can't just directly read these output files with a text editor, like could with teed up files.
46:58
You could set the number of vertical levels in the grid, and, um, and that the height to the levels.
47:05
Now, if you, um, if you say, two, um, and you did something like this, what the first level would be zero to the ground level, up to 100 meters and the second level would be 100, up to up to 500. And when you see the output, it will say 500.
47:27
But it really means 100 to 500 certainly stringy.
47:30
Sort of the top of the, of the concentration level.
47:35
We're going to leave this at one, and we're actually going to change this, too, 5000.
47:41
Because, what we want to do is, we're going to be following this particle that's going down wind, and we want to be able to see it in the concentration grid, even if it goes higher, then, say, 100 meters.
47:53
So, remember, when you're setting these concentration grids, this is sort of what you're telling high split to do.
48:01
How you want the information to come out.
48:03
If you set this, you know, to be too low, and your particles go above it, and you don't have any concentration levels above it, then you'll just get zero coming out of yore in your concentration grid.
48:18
It doesn't mean that there's no, that the simulation stopped or that the simulation failed.
48:22
It just means that you didn't center concentration grid to capture all of the mass that you were trying to simulate.
48:30
Um, this is something to sort of keep in mind as we go forward that, you know, you do have to be careful what levels you're setting had and and things like that.
48:42
Then just as with the zeros that we talked about before.
48:45
But the sampling start time zeros mean, let's just start sampling at the beginning of the run and assembling stop time. That means, for zeros, let's start stop sampling at the end of the run.
48:57
And again, you could specify this if you wanted to, to be something different, but the defaults are zeros.
49:02
Sort of the natural, what you'd want them to be.
49:06
And then you could say something about how you want the data to be tabulated.
49:12
Um, the first, um, number here is whether you want to have, now, which is basically, and, excuse me, an average concentration calculation, which is what we have here.
49:29
If you put one here, um, sorry, one here.
49:36
It would be getting a snapshot of, of the concentration at a given time.
49:42
If you put two here, you would be getting the maximum concentration over particular averaging period.
49:50
Then the second two numbers are the hours, or minutes of the averaging periods.
49:56
So what we have here is, we're getting 12 hour averages and no minutes, but, you know, if you wanted to 30 minute averages, you would put 0, zero, and then 30 minutes, and if you want an average is, you'd want to make this to be first day to adhere to be P zero.
50:15
So, um, this tells you the kind of average.
50:19
And again, this whole thing is, essentially, your bookkeeping scheme.
50:24
You're telling it where the great is, how big it is, what the vertical layers are, and then how you want the averaging.
50:32
Or more instantaneous or maximum value is calculated.
50:38
Alright, so we're going to keep this at, at average, we're going to keep it 12 hours, and zero, in it, and we're going to save, save this.
50:51
Um.
50:54
The next thing, we're going to do, so we're going to save, You have to sort of save out of all these menus now, so we've, we've done the pollutant, and we've done the grid.
51:02
We're not gonna do anything with deposition right now, Um, the next thing we're going to do now is we're going to save this main menu. Just like we do the trajectories in this, now create creates the control file.
51:18
And let's see if I can.
51:24
Open this up. Actually, default cons.
51:27
Coping with notepad.
51:32
See, OK.
51:34
Now, this control file is different now than what we had yesterday with the trajectories. The trajectories we basically had.
51:41
Most I think of this kind of this first part, you know, that the starting time, the starting location, the length of the run, vertical motion, and something about the net data, and now you start seeing a lot more things in this control file.
51:59
This is that number of pollutants. The name of the pollutant, the emission rate.
52:03
The map, the time of admission, that when it started, and then we get into the number of grids. one grid, and this is, you might recognize, this is some of the data about the grid that we did. And these last part here, it's about the deposition.
52:18
And transformation parameters, which we, which we haven't put anything in yet.
52:23
This is cryptic, you know, you don't, when you start using I spent a lot.
52:26
You'll start to be able to sort of look at these and start to see what, what they look like, but anyways, it's more the pollutant dispersion control file is more complicated than the control file for the trajectory.
52:41
Now, we're going to do one other thing here, and we're going to click on the Advanced menu.
52:48
And as we talked about yesterday, if you don't, click on the advanced menu and doing said anything, and then, I will just pick defaults for all of these advanced features, But if you, if you do, then, you could set some things different, if you want.
53:03
Um, and we're going to go to the configuration setup, and yesterday, we were going the trajectory, art, and today, we're going to be in the concentration.
53:13
So, you're going to click Advanced configuration, setup, concentration, and there's more menus that are available here.
53:23
And, um, let's see.
53:25
Let me, ear, we're going to pick menu number four, which is the particle poff release numbers and limits whoa, and what we're going to do is this first simulation.
53:40
I'm going to do this sort of pedagogical Exercise, it just releasing one. at a time.
53:45
We're going to set the particles release per cycle to be one. And, you saw actually they were in there in 2500, That was the default. So, that's why, when we did that simulation at the very beginning of this section.
53:59
Oh, we, I was telling you that was 2500 particles, because that was the zip file.
54:06
And let's save this.
54:10
And I think, now I will save this.
54:14
And I think we can.
54:18
We can now run the model.
54:19
So we're going to go, concentration, run model.
54:25
And now, since we have created that setup dot c.f.t. file, which is created when we did the advanced configuration menu, we do want to run with that Because we actually put something in there.
54:38
And we do want to use that.
54:41
We run it. It goes pretty fast. It actually goes faster because we're only looking at one particle as hundred particles.
54:49
And as we exit, and then we go to look at the results concentration, display, concentration, contours.
55:02
Um, they'll click on that, and they'll just accept the, um, all the defaults for now.
55:10
And what we get is sort of a weird looking.
55:17
It almost looks like a trajectory, and that's because what we're seeing is one computational point being moved along by the wind.
55:29
We are adding a turbulent component to it.
55:32
So, if you looked really carefully, you probably see that sort of jiggling about a little bit more than a trajectory would be, but it basically is going along the lines.
55:41
I'll put a trajectory, would be, and the the concentration is basically, more or less the same because it's just this one particle, it's going and it's encountering the different grid squares along the way.
55:57
It's inside of a grid square for a short, short amount of time, and it contributes then to the 12 hour average of the concentration in, in that grid square.
56:09
And we we pick the the level between 0 and 5000, so these grid volume is hard or actually like five kilometers on a side and exactly five kilometers, high, sort of both, interesting, sort of a cube grid great volume.
56:35
So this clearly is not, you know, a very realistic simulation, and what we are going to do, know, very, very shortly, is add more particles in.
56:45
And, um, we talk here about what would happen if you, if you did this with a concentration layer of only five centimeters, you don't have to do this.
56:58
I'll do this quickly here, go back to the liquid out of this.
57:04
And let me open up the concentration setup run and open up that pollutant deposition and grid setup, you don't have to do this with me.
57:14
But, um, clicking on grid number one, which is we just have one grid.
57:20
By the way, since that is one here, number of grids that are defined, you can only click on that.
57:26
If we hit, If you want to do With any of these are the grids, you'd have to change the number grids in this, in this box here.
57:32
If we change this to 500 meters and then save, save, save, concentration on model, we are going to run using that setup.
57:45
Also goes very fast concentration, display, concentration, contours, clicking the defaults.
57:55
But, you see, is that you only see the beginning part, and, um, what happened was that, after this beginning part, it got higher than 500 meters, And so, even though it kept going, and even though the simulation kept going, um, we don't see it in our concentration grid because we only had one level and it was between 0 and 500, 500 meters.
58:20
So, that is, um, while you have to be, you know, be thoughtful about what kind of levels you want in your, in your concentration grid.
58:30
OK, um, let's go on to the next, uh, section: Um, we're going to do a thing here where we're going to animate, animate this, and I mentioned yesterday that image magic was kind of optional that we're only going to use it a little tiny bit in the in the workshop.
58:56
This is sort of the one of the, one of the only sections that we're actually going to use the functionality of an image magic to create an animation. So Don't worry if if you weren't able to install that image magic for any reason.
59:10
Um, It's OK.
59:12
You're gonna you can just watch this this one section and then um, And Then we'll carry on with the rest. You'll be totally fine.
59:21
Image Magic, It's actually a really an amazing program, it's been around for quite a long time, and um, you can do all sorts of things to, you know, to convert different Hmm.
59:35
image formats j.p.e.g.s to geoff's to p.n.g.s, and you can change the resolution and do all sorts of things to images Image Magic, and it's, it is free.
59:47
All right, so let's, um, do this quickly.
59:51
I'm going this over, second year, Um, what we're going to do is, we're going to go, back to the.
1:00:04
Concentration set up run and we're going to leave everything more or less the same, except, um, we're going to change a few things with, with the grid.
1:00:16
So I'm going to click on this Putin Deposition in grid setup.
1:00:19
Um, See me.
1:00:25
We're going to click on grid number one, which is the only grid that we have.
1:00:29
Um, the main I'm going to put this back to 5000, so we do capture all of the of the of the plume.
1:00:38
The one thing we're gonna do is we're gonna change the way that we're accounting for the particles. We're going to do a snapshot.
1:00:45
So, we're going to change that to one that's that now snapshot, and we're going to create just hourly snapshots, we want to see where this plume is going every hour.
1:00:57
So, then, we will Save, and we will.
1:01:04
Save.
1:01:06
And let's see if there's anything else that we were going to do, um.
1:01:13
Now, so now we can save.
1:01:17
Again, you always have to save the setup menu.
1:01:20
Russ, I split won't know what control file to use or use an old old control file. By the way. I spent, Well, if there's a control file in the working directory and you run high split, it will use it.
1:01:31
If it happens to be there from a previous run that you did and you save the new one, you will get a run.
1:01:38
It will be.
1:01:40
What you told high school to do, it'll be based on that control file. Same thing would be for that setup, that CFD file.
1:01:45
If there's one in there, it will use it, and it will run. So, you do have to be careful that you don't have things left over from previous.
1:01:53
Brian is that you have haven't overwritten the graphical user interface, you know, overwrites the control file each time.
1:02:00
But if you forget to save, um, after you've done the setup, then it gives me.
1:02:06
It won't have overwritten it.
1:02:09
So, now will concentration fun model and we do want to use that setup file so, we can click that one using set of file.
1:02:18
It again goes very fast.
1:02:20
We can exit and now we're going to go let's go ahead and do concentration display, concentration, contours, and we can execute except all the defaults execute.
1:02:34
And in this case, now, if you're working on a Mac, I think that you basically just get the images for each frame.
1:02:45
Kind of, sequentially as you go down. But if you're in a Windows machine, and perhaps on Linux, I'm not sure.
1:02:53
You basically have this navigation button here so we can see the this particle, or you can think of it as a plume with one particle moving, you know, downwind, each in each case.
1:03:10
And now we're going to show just quickly how you would animate this.
1:03:15
Um.
1:03:19
We.
1:03:22
We go to C, ones didn't.
1:03:29
Let me, quick out of this display, let me go to concentration.
1:03:38
Utilities.
1:03:41
C is utility, yes, utilities, SVG to image, and, um, by the way, this year.
1:03:51
This new version of high split is using SVG graphics and for people that have used high split.
1:03:56
In the past, we had been using postscript graphics.
1:04:01
And, um, this is a new, you can still create postscript graphics if you want.
1:04:06
If you're running a command line script scripts, if you still want to use pushka graphics, you absolutely can.
1:04:15
That's how you're set up to run, but we changed the SVG because it's much easier for people to use.
1:04:21
And we've actually noticed or we hope that we've noticed that it's been much easier for people to install the model this year because they didn't have to install all these extra utilities Ghost Script Ghost view, which becoming harder and harder to install, um, to deal with the the postscript graphics. We found we're getting more questions about that than almost anything else, and that's what prompted us to change these images, too.
1:04:50
To SVG graphics, which can be opened up basically in any web browser that, OK. So we'll go to SVG to image.
1:04:57
And, um, what we're going to do is, it populated this from the the, um, file that we just created.
1:05:06
So, the Graphical User Interface is trying to help you.
1:05:10
Um, sometimes it doesn't quite get the right answers, and also if you want to do a different file you.
1:05:17
But this was the last file it created, and so it populates it with that number.
1:05:23
Hopefully, it's what you want.
1:05:24
Um, we want to do animate, and we want to do crop crop, will remove the western to the whitespace around, around the image, and we don't want frames, frames would be, would be.
1:05:39
Separate, No animation in each Graeme.
1:05:45
But we want all in 1, 1 image.
1:05:47
Let's go ahead and change the let's say I think we're going to change the.
1:05:52
It changes the resolution a little bit, I think.
1:05:56
No, no, no. You don't make it 150. You don't have to do this.
1:06:00
And let me execute.
1:06:01
The inversion takes a little bit of time to do this on a processing, And you can see that in, um, in the working directory.
1:06:13
Now we have this conch plot dot Js file.
1:06:18
And let me, I click on that.
1:06:20
I think you can open it up in a web browser. Might open up, and I just double click on it.
1:06:26
Yeah, it's on Windows machine.
1:06:28
It opens up and sort of picture on video display, and you can see that we now have a, an animation.
1:06:37
And you know that there were these defaults set up in the amount of time it spends in each frame and if you're using Image Magic or from the command line, you can, you tell it to go slower or different things, So, it's a very powerful program.
1:06:55
OK, um, let me stop that and, um, let's go ahead and save the, the control file and so we're gonna have a concentration setup run, Save As, and this will let us retrieve this later.
1:07:15
It's called this.
1:07:17
Cons, Yes.
1:07:21
Control, dot text, and it'll be saving in the working directory.
1:07:28
Actually, if I, you see it showing up in the working directory there, and I'll save out, and we'll go to the Advanced menu, configuration setup, concentration, and let's also save that, setup, the CFE file, Save as.
1:07:49
Clients?
1:07:52
Yes.
1:07:53
Setup.
1:07:58
Text, save, Save.
1:08:03
You can see here, you see? Oops.
1:08:10
Um, see, that did not seem to Save, like I thought. Let's try that again.
1:08:18
Oh, there. it is, there. It is. All right.
1:08:20
When I X, when you exit out of the Graphical User interface, it also writes these other files, which were the same.
1:08:29
This will let us retrieve it later.
1:08:32
All right OK, Let's go to the next section.
1:08:37
Um, so, we're going to, um, do some more, um, now more realistic calculations, and, um, by releasing more, more computational points, and we're going to start to see how that impacts the, oops, Let me organize myself slightly differently here, right.
1:09:05
They're not making you C sik ear, OK.
1:09:11
Obviously, you know, releasing one particle, it can't capture the whole dynamics of a plume.
1:09:19
So, um, if we were starting from scratch.
1:09:23
And hadn't just done this last section, we would load these files, you could go in and set up run, and you could retrieve one of these files. But since we just did that, we don't have to do to do that.
1:09:39
Um, what we're going to do first is, let's see, is we're gonna go back and change a little bit about not the grid.
1:09:49
So we'll click on the pollutant deposition in grid set up, and we're going to click on the grid number one.
1:09:59
And the one thing we're going to change is we're going to do a three hour averaging time.
1:10:09
And we'll leave it at a snapshot, for now.
1:10:16
And we'll say, save this, will save.
1:10:20
And we'll save, and we're going to go into the advanced configuration setup.
1:10:28
Concentration menu, and we're going to change the menu number for the particle release number limits.
1:10:41
And, if so, instead of releasing just one particle, now, we're going to release 100 hundred particles is still not very many. Not like the 2500 that we did. But actually, it was 100.
1:10:53
You're gonna start to see, um, some interesting new things.
1:10:58
Um, we're going to save, save. And now we'll fun.
1:11:05
The model.
1:11:07
Concentration, one model.
1:11:11
We are going to use that setup file so I can click this green Rohn using setup file.
1:11:18
All right, goes pretty fast.
1:11:19
A little, tiny bit slower you might have noticed but pretty fast because only 100 hundred particles exit, and then we go now to concentration, display, concentration, contours, and accept all the defaults.
1:11:37
And here, what we see, it is in the first time period, and, by the way, you could see, it's telling us that it's averaging between zero and 505,000 meters, and you can see That this is a snapshot.
1:12:02
After three hours.
1:12:04
So, the run started at 7500 and since we're talking about snapshots, it's just showing it at three hours later, so it's the snapshot.
1:12:14
That's why the times here, 20, um, hundred and 2000, D, three hours later are the same.
1:12:21
If we had averaged it, it would be like 7500 to $2000. We are just doing a snapshot.
1:12:28
And so, this is what it looks like after the first.
1:12:31
And you can see. now, we're getting more complicated.
1:12:35
Because actually, all these, these hundred particles started disbursing.
1:12:38
And we're starting to get, you know, a bit of a spread around.
1:12:43
But since we only emitted stuff for the very first minute, it's like this little sort of puffer.
1:12:49
No a bolus of pollutant that was that was emitted into the air, and now sort of this, this mass is now being effected down downwind, and as we go further down wind, um, you start seeing the.
1:13:07
It's things spreading out, and what you see by the end is, this is not a very satisfactory picture, is it, because really, what this was was some kind of cloud, we didn't have enough particles, we only had 100 particles.
1:13:24
And so, what you find is that we didn't have enough particles to fill out the concentration grid, so we had a concentration group that was five kilometers on a side, and, um, but we only had 100 particles. And so, sometimes one of those particles would get into the volume of a grid square, and sometimes it wouldn't.
1:13:47
And, um, when you see a simulation and you see kind of this sort of popcorn version of the output, chances are what you you just didn't have enough particles to fill up your concentration grid and this there's two things you can do. You can either make the concentration grid, coarser.
1:14:08
And, you know, spread things out.
1:14:10
to make the, you know, the grid squares bigger Or you can add more particles.
1:14:17
And, if you know, if you have the computational resources, and at this point, we are not doing grow many particles at all, so it's pretty easy to do. It's there are times when do you sort of can have to make compromises between.
1:14:32
How fine upgrade do you want and, um, and how many how much computational power you have to simulate articles.
1:14:42
So, now, um, let me get out of that, and now we're going to try running with one thousand particles, and we'll see what, what happens. So, let me.
1:14:58
exit out of this display.
1:15:01
We're going to go to concentration set upfront. Actually, we're not gonna make any changes to this.
1:15:07
I'll just, I'll just save this.
1:15:10
Um, we're gonna. do is we're gonna make a change in the advanced media and configuration setup.
1:15:17
Concentration.
1:15:18
We're gonna go to this menu number four.
1:15:21
So, instead of 100 particles, it's going to be one thousand articles.
1:15:29
Save.
1:15:30
Save, and let's run the model again.
1:15:36
seven, concentration my model, we are going to use that at setup file that we, we made.
1:15:44
And now it's going a little bit slower.
1:15:47
Again, not too bad.
1:15:50
But, you could sort of notice that now we're doing one thousand particles, not just one.
1:15:55
So, you know, if you remember yesterday, we're just doing trajectories, just doing one computational point and it was not even adding that, just that, that trivially dispersion to each time step. So, now, we're doing a thousand of those. And, we're adding the ... calculating dispersion as well.
1:16:15
And we're also calculating the concentration in each of the grid grid squares that we B hat.
1:16:24
Let's exit out.
1:16:25
Let's look at concentration display, concentration, contours, and again, we will accept the defaults.
1:16:36
It's a little bit longer, let's see, if I didn't click that right here, we go ahead and click the right, OK, um, here's the output.
1:16:46
And it, you might remember last time wasn't quite this convoluted, but you see some little things sort of pushing up the side a little bit more complicated.
1:16:57
Um, in the last one, we had seen kind of, like starting to see the individual particles.
1:17:03
Now, you can see it's like, we're filling up the grid better, with a thousand, um, computational points, not perfectly.
1:17:13
You start, you're starting to see some of these, you know, individual particles in which squares, you know, towards the end.
1:17:20
Um, and, again, it's it doesn't look as bad as the last time we're starting to see kind of the, basically a plume. But you see a lot of these for the individual things.
1:17:31
So, we're not quite there yet. And and, by the way, this is something that you kind of think about when you're when you're running high split.
1:17:39
Nobody can tell you exactly how many particles to run, or how big your grid square should be.
1:17:49
You kind of have to think about the problem that you're solving, and you can do these kind of numerical experiments, and whenever I'm doing a project, I often spend quite a bit of time at the beginning doing, you know, trying one thousand particles, trying 10000 particles, trying 100 particles.
1:18:06
No different grids, grid resolutions.
1:18:09
Try to see, you kind of want to find the sweet spot in your configuration space, where you're getting the information that you need, but you're not extending too many too many computational resources.
1:18:22
But, you don't need, um, um, at some point, you can add more and more particles, and it won't change the results of your computation.
1:18:33
And at that point, if you sort of have overkill and you probably don't need to.
1:18:39
I've done that.
1:18:40
So, um, it's one nice thing about models, is that you can keep doing them over and over again, and with different configurations.
1:18:49
And you can see what, what the answer is.
1:18:54
Let's see, I think that we more or less got. Yeah, that's that same answer that we got. Let's try 1 to 10000.
1:19:00
Um.
1:19:03
Particles.
1:19:04
Sorry, let me Bring this back.
1:19:08
Sorry, this back.
1:19:13
I'll just quickly try to 10000 particles. I'm going to quit.
1:19:15
I'm gonna go, concentration, and, again, we're not actually going to change.
1:19:22
Anything. So, I actually didn't have to open up this menu, open up the advanced configuration set up concentration.
1:19:29
Um, number limits, 10000, save, save, concentration, run model We are going to use that. I'll now, this is going to go.
1:19:44
Noticeably slower, because now we're doing 10000.
1:19:47
Hmm.
1:19:50
Not too bad, but you definitely can notice now.
1:19:56
Slower, slower, run.
1:20:02
OK, let me exit out.
1:20:04
Concentration Display, Concentration, Contours, Execute Display, and now we start seeing a more.
1:20:19
Uh, even more fully formed.
1:20:25
Boom, it's still starting to see a few.
1:20:27
So, you know, know, if you were doing this simulation, you might say to yourself, maybe even 10000, might not be quite enough, and you might go to 20,000 or 50,000.
1:20:36
Although, it kinda depends a little bit about what you're, what you're after.
1:20:41
Um, know, these concentrations out here at the edge are orders of magnitude less, then the concentrations, you know, elsewhere and so this is my fee, um, this might be, might be sufficient, all right?
1:20:59
Let's see what else we want to talk about here.
1:21:09
I think I'm not going to do this.
1:21:12
But if you set the grid to a much finer grid, instead of zero point zero five, which would be five kilometers, you get zero point zero five, which would be 500 meters. Yeah, 500, 500 meters.
1:21:28
Um, then, if you get back into the situation where you don't have enough particles in any anymore, and it's a little bit hard to see this, but this actually is is similar to that sort of popcorn version where you're just seeing a whole bunch of individual particles, they're so small.
1:21:47
The grid squares are so small that you can't really see it.
1:21:51
But, um, this kind of talks about how the interplay between the grid size and the number of particles that you, that you have.
1:22:00
If you, if you want to have A Feiner grid, you certainly can do that, but you're going to need more particles in order to, to represent the dispersion on that on that grid.
1:22:16
Um, and also, when you make a finer grid, dissimulation goes a lot slower because of the bookkeeping function in the model.
1:22:31
Um, remember, it's trying to calculate for every every computational point, um, what D concentration is in each grid square and if you have a lot more of these grid volumes, then it's got a lot more of that bookkeeping to do.
1:22:51
So totally fine. You can have whatever resolution you want and whoever the number of verticals you want, But just know that there's tradeoffs as you as you do that.
1:23:02
Um, so, let's go ahead and save.
1:23:10
Save this, concentration, set up run.
1:23:17
Save as, we're going to call this C part, um.
1:23:25
Key part.
1:23:26
Control Text, and it will also save the advanced, um.
1:23:37
File as well, Save As.
1:23:42
See, Part setup dot text.
1:23:52
Save.
1:23:53
Safe. All right.
1:23:55
Um, OK.
1:24:01
And by the waiver is mentioning some of the things I've already said, but, you know, we were just releasing one set of particles at the beginning of the run.
1:24:11
In, in many cases, what's your, you have a continuous release, and you, um, are releasing particles throughout the simulation, and maybe you're releasing 2500 particles.
1:24:24
No, every minute or 25 particles.
1:24:27
Please, under particles every hour or so.
1:24:29
No, you can quickly get a lot of computational points in your calculation, and so, um, things can, you know, get slower and slower depending on a faster computer is.
1:24:43
And by the way, there is a multiprocessor version of High Split, that can be run on Linux machines, um, where you can then use more than one processor, too.
1:24:59
To do the calculations, and it can, can go much, much faster.
1:25:03
All right, let's, um, go to the next section, and here, we're going to be talking about a, um, an interesting way to view the results and, um, what we've been doing so far is displaying the results of our calculation in this concentration grid, but there actually is a way to look at the.
1:25:34
The actual computational points themselves, and where they are, and this can be quite instructive, and I'll show you in a second with that.
1:25:43
It's gonna look like.
1:25:47
So it's telling us to, we're gonna go into the concentration this.
1:25:53
Yes, birney go into the concentration setup, and we haven't really done anything, So I'm just gonna go ahead and save, save this.
1:26:06
It didn't really do anything different here, and, um, we're starting from scratch.
1:26:13
We could have loaded that setup file as well.
1:26:17
By the way, if you weren't here yesterday, these files are, or at least versions of these files are available in the tutorial Files Directory.
1:26:32
Let's see, see part, let's see this.
1:26:38
Siebert?
1:26:42
Yeah, See, Part control, cheaper setup.
1:26:45
You couldn't, you could load those files in. The one thing that you'd have to watch out for would be whether or not.
1:26:55
The net data is in the saint, right right place.
1:26:57
Even if you load that in, you might have to specify where the metadata is, because, um, those are created with a certain computer on a certain configuration, where the data was in a certain directory, and it might be something different on your computer.
1:27:15
But it can get you most of the way there if you need to.
1:27:20
OK, so what we're going to do is, we're going to go to this advanced.
1:27:24
Then, you can figuration set up, um, concentration, and we're going to open up menu number nine, which is input and output of vertical files.
1:27:40
And, um, we're going to change these parameters here.
1:27:48
Um, we're going to output particles every after the first output, after three hours, and then every three hours after that.
1:27:57
And we're going to call it this name part.
1:28:02
You could change the name if you wanted to. And by the way, this was created actually for almost another purpose.
1:28:08
It was created so that you could save, run, and then restart the run the next day if you wanted to or you know the next time you wanted to start the run.
1:28:19
So this actually saves all of the information that the run has the simulation has.
1:28:28
It saves all the computational points or puffs and their locations and also other information about how much mass they have.
1:28:39
It actually saves all the information that you might want or run and you could start.
1:28:47
This run from the end of the last run, if you wanted to, that's what this initialization part of the of the menu comes from.
1:28:58
Alright, so it will save, save this.
1:29:03
We'll save, um, we want to set the grid spacing back to zero point five, but I think since we didn't open this up, set up run pluton deposition, which we didn't change that to zero point zero five, so we don't have to worry about about that. Save.
1:29:24
Save.
1:29:25
Let's see if there's anything else you want to do.
1:29:31
Let's go ahead and run the model.
1:29:33
So I'll save, Save this, and Concentration, Run Model.
1:29:42
Go ahead and run using this setup file.
1:29:49
This is going to go even a little bit slower.
1:29:53
We have 10000 particles, and also no writing this extra file each time.
1:29:58
So we're starting to get some input output issues, again, not not too bad.
1:30:03
But each time you do something else, you add a little bit of computational time.
1:30:09
And now we're going to go and look at the results.
1:30:15
But instead of display concentration contours, we're going to display article.
1:30:25
And, um, when we do that, upcoming is sort of a new kind of menu, and let's see, if we want to change any of these display options here.
1:30:44
It looks like we're not changing anything. We're sitting across the cross-section.
1:30:49
We're going to do every particle.
1:30:50
Sometimes, you have solid particles, it's just so many that you can't really display them all, or, it takes too long to display them all equally, 10000, Which is not too bad.
1:31:01
Let's zoom a little bit in.
1:31:02
So, we we it was much like white space, and then we execute the display.
1:31:12
And, um, it's interesting, this is after the first three hours.
1:31:17
Remember, when we're looking at the concentration output, we saw that sort of blob of stuff.
1:31:25
Excuse me.
1:31:27
Pardon me.
1:31:29
Here we're seeing, looking down on the plume, we see that same sort of blob, But, individual particles, um, know, there's so many. There's 10000 that you really can't see.
1:31:41
Maybe they're a edge, but we're also now looking at the cross-section, kind of looking at it from the side, the vertical distribution of those particles, and, I remember from yesterday that the Boundary layer hight was around 2020, 500 meters, something like that, at Dayton, Ohio, when this tracer release was released, and you can see in this first.
1:32:07
three hours, The pollutants are the particles are actually pretty well mixed in that boundary layer they're pretty well mixed Actually, after after three hours.
1:32:20
If we go on to the next fraim, um, you can see we're starting to get material are higher up.
1:32:31
You can see, that, the material, at the top, it's going a little bit faster, at these, these light, blue colors, the ones at the top, are going faster than the ones lower down.
1:32:44
That's, as you probably know, as you go up in the atmosphere, the wind speed can get faster and faster, wind speed near the ground is generally the slowest, but as you go up, she's been hiring the atmosphere. Let me drink water.
1:33:03
It goes faster and faster as we go on into the next.
1:33:08
You start seeing more, You start seeing the, the winds at the top, kinda start to curving around, you know, We're seeing, you know, different winds up above the boundary layer.
1:33:21
And then, finally, the last last frame.
1:33:24
And this sort of looks like that the plume that we saw.
1:33:31
We looked at the concentration grid.
1:33:32
But, this is sort of another way too, um, to look at this, OK, let me stop that, and, we're going to show you one more thing.
1:33:48
Um, um, there's one more aspect of split, that we'll talk about briefly here.
1:33:58
Um, you can output a trajectory based on these particles, and instead of A, a simple trajectory like we were doing yesterday, what you can do is do the full dispersion calculation with all of the turbulent dispersion. and all of these particles moving around and being dispersed.
1:34:22
Then, at the end, calculate the center of mass of all those particles, um, and create a trajectory based on that center of mass.
1:34:32
It's, it's very similar to the trajectory that you would create from a mature trajectory, but really, it's a little bit different because it's taking into account all of the different dispersion, and the fact that you might have dispersed up into different parts of the atmosphere with different wind speeds and wind directions.
1:34:53
So the trajectory that you'll get from this is a little bit different than than than a simple trajectory that you would calculate, And it can be useful in sometimes to look at this.
1:35:05
So what we'll do here is, we're going to see, you go down.
1:35:14
We're going to go to the Advanced menu Configuration setup.
1:35:18
Concentration, um, and we're going to go into Menu 12, which is this Output center of mass trajectory, clicking on this, and you have to give it a name, we'll call it C M syndrome mass bash.
1:35:39
Next, And this is going to create like a t-dap file with the latitude and longitudes and in.
1:35:45
Right.
1:35:48
And we will save.
1:35:50
Um, save this.
1:35:53
Number, go concentration.
1:35:56
On model, we are going to use that.
1:35:59
Um, set up dot CPP file.
1:36:03
Now it's even going to go a little bit slower because now we're doing this extra calculation.
1:36:12
Are doing everything we were doing before, or now we're doing this extra extra calculation.
1:36:20
And we won't get anything different, you know, in this Particle display or the concentration display, but now we can display the trajectory.
1:36:29
And so now we're gonna go back to the trajectory menu, display directory, and you'll see this weird thing that come up.
1:36:37
It says great identifying X gets set up first.
1:36:41
Don't worry, you can.
1:36:42
You can quit out of this, what the GUI is telling you is that it doesn't, you haven't been running a trajectory, and it doesn't quite know how to set things up.
1:36:52
So what you have to do is you've got a trajectory, you set up run, and you just say save.
1:36:58
It is what it does. It's tiny the GUI. Yes.
1:37:02
Graphical user interface. Yes, fruit are running trajectories now. I'm gonna give you some information. It doesn't really matter at this point.
1:37:09
What information is here?
1:37:10
You just have to go ahead and save it, to tell the graphical user interface that that not to worry, We're going to be able to do something here, So now you go trajectory, display, trajectory, and now, instead of teed up what we pick is that C M fresh, that text, that was that end point file that we created.
1:37:40
And let's go ahead and, Um, Execute, um, display, and This now is the 12 hour, um, center of mass trajectory, and, you can see, it's actually a little bit different than the other one.
1:38:04
I'm not going to go back.
1:38:05
Let's see, maybe we Yes.
1:38:09
Similar.
1:38:11
Um, well, I won't go back when we did the trajectory yesterday. It didn't go as high as fast.
1:38:19
And what we're seeing is, it's actually mixing up, and that center of Mass is actually higher earlier.
1:38:26
Then when we did the pure trajectory, that's because of that mixing throughout the boundary layer is, is faster.
1:38:34
And the center of Mass then is going up a little faster.
1:38:38
And I'm not going to do this, but if you do the Python version of, um, this center of mass trajectory, I'll just show you what the picture looks like here, It actually creates a really nice version of this, where it actually gives you the distribution, sort of the standard deviation of the, the distribution of the particles around that center, line trajectory.
1:39:11
So, and that's actually only available right now On the Python version of the of the graphics, and like you said, depth of the Python version of these graphics actually are really, really quite nice and quite sophisticated.
1:39:25
And if you've installed the Python and you wanted to try some of these Python versions, you absolutely can but we're not going to be doing a lot of that in this workshop and I'm not going to do it.
1:39:38
Ear in this section, OK.
1:39:45
Getting towards the end of this section here.
1:39:51
Um, I'm going to show you now a little bit about, I'm using Puffs And, um, Let's go. Let's go back here. Let me click out of here.
1:40:06
Um.
1:40:09
Let's go to set up, run.
1:40:13
I'm going to go ahead and retrieve E C part.
1:40:20
Um.
1:40:25
Setup, Uh, Setup dot text, OK.
1:40:37
Oops, let's see.
1:40:42
Things like my Dewey, a graphical interface, maybe get it confused.
1:40:49
They need to.
1:40:53
Interesting me.
1:40:56
Ha ha. All right.
1:40:59
Wish application. That's the TCL TK.
1:41:03
Let's go ahead and close the program.
1:41:05
This could happen, um, hopefully, not too, major.
1:41:10
Let me, Let's see here. I'm not even showing my.
1:41:19
Love my mom.
1:41:24
one second here. I'm not showing all of my?
1:41:29
Shortcuts. See, I split.
1:41:32
He is here.
1:41:35
Thanks.
1:41:36
Let's see if that's.
1:41:41
Let's see if that's the right one there.
1:41:45
Yeah, OK, Um Let me let me reset although probably we set to do that.
1:41:50
So now let's go to concentration, set up Run.
1:41:56
Um, let's retrieve again that.
1:42:03
C Part setup text OK.
1:42:12
This is not.
1:42:16
Good.
1:42:18
Huh, This is, uh, little bit frustrating here.
1:42:27
Um, fortunately, we were able to get through most of this, of this section before my computer is turning to issues.
1:42:40
Um.
1:42:42
Exactly, sure, why is actually hasn't happened to me, maybe more than once in the last maybe 15 years when something like this has happened. But we are having a problem.
1:42:56
So why don't I, we are coming up on the time for the break. We may be able to come back. I'll show you some of these puffs.
1:43:06
Let's stop this here. So, if something is happening on my computer, I'll have to reboot.
1:43:10
But, I think what I'll do now is Why don't we go ahead and go to the 2, two to the break, I'll try to fix my computer. And.
1:43:26
We'll come back in in about 15 minutes, um.
1:43:33
And carry on with with the next part of the workshop, Alright.
1:43:37
So, see, you all in about, about 15 minutes.
1:43:42
OK?
1:43:49
OK, we are going to be coming back after our break.
1:44:00
And, uh, thank you for the folks that pointed out what happened at the end there.
1:44:08
I was trying to retrieve a setup file into the, the advanced file, into the into the control file settings, the basic menu, and that's what was causing that error. And I did it twice in a row.
1:44:26
So, I think we never thought anybody would ever try to do that.
1:44:31
And so, we didn't have sort of an error catching mechanism.
1:44:34
And, I'm gonna talk to the, our team are code developers and things, and we may want to, at the end of the program, it says, no, you're not actually importing a control file into this control file menu.
1:44:49
Um, try it again.
1:44:52
So anyways, that's what happened. And what we were going to be going through was information about puffs and we'll try to get to that perhaps later.
1:45:01
For the most part, most of you most of us use the three-dimensional particles to do the, um, the dispersion simulations anyway. The puffs are a little bit of an older, more legacy version.
1:45:15
For me, I used puffs for like 25 years, but now even I'm using the prediction of particles.
1:45:22
So, OK. Our next section is going to be getting into more.
1:45:29
More detailed and more accurate representation of the simulation that is kept ex experiment.
1:45:36
And I think Chris Lochner is going to be presenting that, and Chris, I'm gonna make you the presenter.
1:45:45
See.
1:45:47
Sure.
1:45:50
Yes.
1:45:53
Then it turn off webcam, alright?
1:45:58
And we see your screen.
1:46:00
Great, Chris, so take it away.
1:46:02
OK, so, I mentioned yesterday that a bunch of this workshop, we're going to use the second tracer lease of the cap techs experiment to kind of walk you through how you run high split.
1:46:17
There are some other, uh, simulations to what we will perform later in the workshop, but a lot of it is based on the ... Tracer Experiment.
1:46:27
So now I'm going to go over a general setup for this, for this tracery lease. Experiment with high Split and go over.
1:46:40
Just basically the basic way to set it up using some of the default options. How to run that run a dispersion simulation. Look at the results, display the results, and compare them with observations.
1:46:55
And later on in the workshop, you'll explore more of the kind of inner workings of high split, looking at all of the different model parameter ..., how you can change the, for example, the turbulences ..., the option of how you estimate mixed layer height, or PBL height and things like that.
1:47:17
So for now, this is just an initial setup of the cap tax experiment experiment. So let's get started now.
1:47:27
OK, so what I'd like to do before I start anything new and I mentioned it yesterday, is to clean up my working directory. So I have my work, my folder, just pointing to the high split working directory.
1:47:40
I'm going to clean this up, clean up, working directory, continue.
1:47:45
It kicks me out of high split and notice my working directory was changed to working Don't working dot 471, OK, so then I go into my new working directory. I've got don't have much in there, just like yesterday.
1:48:00
Open up, High Split, and my Working Directory is populated with some of the default, uh, scripts and files in there, OK?
1:48:12
And, I also want to mention that, Ah, it Throughout this tutorial.
1:48:19
For example, you already did it earlier today, and yesterday, you, for when you're in setup, you can save setup files, and you can save that, or the control files if you save the setup, that config files.
1:48:35
Now, if later in this tutorial, it tells you load this file that you already saved, and to clean up your working directory, and notice you have a bunch of working directors, they previously cleaned up, you're like, Oh, I don't know where that files. I don't know where where it is, which which directory. It isn't. Don't worry. All of those files, we also have saved in the tutorial.
1:49:03
Directory, if you go to tutorial, Files, has a bunch of these.
1:49:11
Input files, the control file set of files, and in other files already for you, you can just go to that folder to get them OK, So, let's get started. So, That this is an animation of a high split simulation for the second Tracer release of Cap T X Experiment, showing some of the observations downwind, and the general flow, from Ohio into the north-east.
1:49:40
OK, so, the second Tracer release it emitted PM CH, that's per floral model, Methyl Cyclo hexane, or C 7 F 14 from Ohio. And we're going to simulate that. So, let's just go ahead and get started here.
1:50:02
So, if we had concentration set up, Rohn.
1:50:08
And when you use my cheat sheet over over here and look at how, How to set it up to show the Tracer Release, started at 1983, September 25th, 17, UTC.
1:50:24
And, the starting location was at 39.9 degrees latitude oops, , at a height of 10 meters near the surface, and we're going to make this simulation for 68 hours in the forward direction.
1:50:50
And, we're going to use, so, if I hit Clear for the Meteorology, it gets rid of the default Meteorology file that was in there.
1:50:58
And I'm gonna upload that Wharf 27 UW, that we were looking at.
1:51:05
That we've been working with, which is located in the sea, tutorial, cap text directory and find the 127 UW.
1:51:19
OK, and now I'm going to open up the pluton grid fire. So if I hit the pollutant deposition grid setup.
1:51:29
For pollutants, hit that species one.
1:51:32
And the, the pollutant that was emitted PMC H and the emissions rate was 67 thousand grams per hour for a three hour dirt duration.
1:51:48
And it is up to you to keep track of the units.
1:51:55
So I know that the units are in grams per hour. So you have to keep track of the mass unit grams per hour and that means the output will then be N grams per meter cubed.
1:52:08
So if you're meeting, if you're emissions rate, that you have is kilograms would be kilograms per hour and the output will be kilograms per meter cubed, OK?
1:52:19
So it's up to you to keep track of, of the units, and if I leave the release start to really start time, we've all at zero, that means it is going to be released when the simulation started.
1:52:35
OK, so I'm gonna hit save.
1:52:37
And now I'm going to up edit the concentration grid that we have. If we leave the concentration center of the latitude and longitude to zero that then the center of the concentration grid will be located.
1:52:51
Where are the emissions, where the release took place? But we know that the image, the pollutant was transported into the north-east.
1:53:02
So north-east of the sites, so we're gonna, we're gonna move the center of the concentration grid over that area.
1:53:11
So I'll put that 42, -78, and use spacing of zero point twenty five degrees by zero point twenty five degrees.
1:53:25
And the concentration domain will span 15 degrees.
1:53:31
North-south in 25, east west degrees.
1:53:37
And I'm going to name the output file, high split to top 10, The output grid directory, I just left, it, dot slash what's going to go in the working directory.
1:53:49
The number of vertical levels, you can have multiple vertical levels. Here, I'm going to leave it at one.
1:53:54
And the height of levels is 100 meters, so what this means is that, so the concentration grid is going to be from the surface to 100 meters, and note that we had the pollutant being emitted at 10 meters, however, it that the measurement sites are well down when, so we expect the pollutant to be well mixed in the boundary layer Seville.
1:54:18
So, this should be equivalent to it.
1:54:24
This should represent the measurements that also take place at the surface. The concentration at the surface, I would expect to be very similar, to an average concentration from the surface to 100 meters.
1:54:37
And, as Mark mentioned earlier, if we do have a very small concentration.
1:54:45
Vertical resolution, say, we're only averaging from 0 to 10 meters, we're going to need many more particles because that's a very small area.
1:54:55
If we don't have that many particles, we're going to under sample the particles.
1:55:00
So you have that, that kind of thing to think about, is that the higher the resolution, whether it's in the horizontal or vertical resolution, you're going to need more particles. And if you have a coarser resolution, you won't need as many particles. And I know that. The PBL is well mixed. The measurement sites are well downwind of tracer, where tracer was released, that we can increase this height to 200 meters and feel comfortable about it.
1:55:28
Now if the measurement site was, say across the street from the power plant, that plume will not be well mixed in the Boundary Layer.
1:55:36
But we know these sites are well downwind, so this this should be OK but got the sampling start time.
1:55:44
We're going to set up What.
1:55:47
Said, Have a sampling started 1983 September 25th hour, 18, 0 minutes.
1:55:57
samplings stop time in 19 83, September 28th, 15, UTC.
1:56:07
right, and now we have to say whether we want average or instantaneous concentrations and how we want it to average them.
1:56:17
If, so, if I put 0 one here, that's, this is going to give me, uh, 0 one, is, for the first, is, tells you whether you want instantaneous values or average concentrations, we're gonna, we're gonna average concentrations, and then the next is how you want to average them, or how frequent you want those instantaneous concentrations, and I know the measurements that we have are either three hour concentrations or a three hour average concentrations are six hour average concentrations at the surface for the surface observations.
1:56:57
So since we have averaged measurements, I want average concentrations for my high split output.
1:57:05
And three hour average concentrations, we can use those for both the three hour average measurements or six hour average measurements.
1:57:14
So, this is, this is how we're going to set this up now.
1:57:21
OK, so we're gonna save this.
1:57:27
And now moved to the concentration Advanced Setup Mania.
1:57:35
And change the number of particles We're gonna increase this 50,000, and increase the maximum number of particles to 100,004.
1:57:49
This is 100 maximum amount of particles, 100,000, that's more than enough, Since we're only releasing 50,000, that's mainly for allocating the domain. I'd like to have, make sure there's more than enough park, more than enough allocations for that. For the arrays in the model.
1:58:09
Now, uh, this is very important for it.
1:58:14
If you're, if you're running with puffs, downwind, those puffs, can split and if the maximum number of puffs, uh, it reaches this, this amount, but you haven't finished emitting, puffs from the Tracer from the release location They'll actually stop releasing more puffs.
1:58:37
So for so something to keep in consideration, if you are running puffs, because puffs, as they grow, they can split into smaller puffs and then they grow some more than.
1:58:46
Once They get to the critical size that keeps they split, and they can keep splitting down downwind, but particles don't split. So so we don't have to worry about that now.
1:58:59
OK, now, we've got a format.
1:59:01
The, the Particle output file, the particle file output options, we're going to output the first output, will be in the first hour, and then we're going to repeat the particles. The part dump file will keep writing every three hours, where the particles are every three hours into an output file called Part dumped.
1:59:26
OK, so I'm going to save this setup as cap Taks.
1:59:35
Served at T X T.
1:59:40
And I am going to save this as cap taks control dot TXT.
1:59:51
And now I'm going to run the model.
1:59:58
OK, so I take this run is going to take some time.
2:00:05
So while this is running, we can look in the working directory.
2:00:11
See?
2:00:17
So I created this cap tech setup, kept X control file. That's what I saved. And as it's running, it's re writing out the par dump file.
2:00:27
It writes where the particles are, starting at the first hour than every three hours, it's going to write where the particles are in latitude, longitude, and height is also has a message file so that let us look at that.
2:00:42
So, it's continually reading to this message file as it runs, and it gives you some warning messages. And actually, there's a warning file to let me open that up.
2:00:55
So, this warning message is the same thing that's, that's showing up at the top of the message file, So, in, in high Split, it's it's meant.
2:01:04
Kinda built to run, even if it's not configured properly for the met the met file that that's that it's using.
2:01:14
So, for this case, it's using all of the defaults that high split house. And one of the defaults is to run with a wharf Vertical Interpolation Scheme called this will be W ... equals true.
2:01:30
Even if it it's not running with a war file. But then it reads in the meteorological file Notices If it's not a or file it's going to say, OK.
2:01:39
We've got we have to switch this to false in order to have high split to run.
2:01:43
And also to run with this war for vertical interpolation scheme you need some very specific meteorological variables in the file such as these variables here called ....
2:01:56
And if they aren't in there, some other files do if they aren't, then again it's not going to use the Wharf Vertical interpolation scheme but instead use another interpolation scheme which is very, very similar to the wharf scheme. So it won't.
2:02:11
It shouldn't impact the results or results very much.
2:02:15
So also in the message file it shows what meteorological meteorology meteorology in there.
2:02:22
The resolution 27 kilometers how many the grids are in there.
2:02:28
And also all of the name list variables in the setup, that config.
2:02:36
And if you have not set them, it's it's going to set the default values that high split uses.
2:02:45
And for all these fought these, the settings, um, No, even this tutorial, we are not going to go over, not that any of them. And we'll go over, so that the main ones that we think you should be really, really, really concerned of, and for all the other one for. Basically, all of them. You can also reference the users.
2:03:08
Manual.
2:03:09
So, if I open up in, and I'm going to show you exactly where, in the user's document, you can do that, open up the document, file, dot sub sub directory, and open up User guide.
2:03:21
And oh, that popped up on another screen. Let me move it over.
2:03:31
I already have it set to that Location, but let me go to the very top.
2:03:35
I'm going to move down to kind of the table of contents. Did I skip it already?
2:03:44
Let me scroll down to the table of contents.
2:03:51
OK?
2:03:56
And you see the summary of all name list, variable, Variable Defaults, and Yeah, so let's click on that.
2:04:08
So this is a description of all of the options that you can put in that set up that config file. And what's bould. That's actually what the default is, but you can change it. You can click on any of these links.
2:04:21
It will take you to a kind of in this document, description of more about what this is, Borrow, for example, the ... That they it.
2:04:31
You have five different options for Boundary Layer Turbulance ..., and the the default is two, which is the counter-claim and you can click on that link and read more about the vertical turbulence options, OK, OK, so what's still moving slowly, but eventually, you know, it'll get there, we hope, OK, so, let's look at some more the message file.
2:05:05
Also in here there are a bunch of flags this has to do with what is in the meatier Meteorology file.
2:05:13
So, it's telling you that it's, it's set to be It has two meter wind temperature or 10 meter wind.
2:05:22
Velocity is two meter wind speeds and you can look true. You has F, false.
2:05:27
It doesn't have it in the file, and this is important based on some of the model options that you have.
2:05:34
For example, we have yeah, turbulence parameter, ... That is dependent on TK.
2:05:42
If you select that option in the Name List, file, the setup that can take out that set but you don't have TK here. It says three-d. three-d. TK field.
2:05:52
We do have it, but if we didn't have it and we wanted the set it, and we set that in the as we set up high split, you would get a warning message that says.
2:06:05
You don't have TK in the meteorology file, so we're going to switch to this other.
2:06:14
Turbulence parameter, ..., OK, and as as you go farther down, we can, you can see, ah, let me move down here. So we can see this shows how many how much mass is currently in the simulation at this particular time.
2:06:33
So there's about 200 grams, 200 kilograms right here of of the of this tracer within the simulation and there's a percentage of where it is in it in the model and the vertical.
2:06:53
OK.
2:06:58
And I'm going to close this now, and I want to see if it's far enough along where we can see that number change a little bit. So let me go back to the high split working directory.
2:07:25
Click on this message file again.
2:07:33
OK, so it still has about 200 kilograms of mass.
2:07:48
And over here, it starting degrees decrease less than 200.
2:07:53
So so as, as we move on, we're losing mass, but not really I mean, within the meteorological domain, we're losing mass.
2:08:04
This is because the pollutants are now moving off the meteorological grid. As the pollutants move off the meteorological grid, they disappear from the simulation.
2:08:14
And as I mentioned, briefly yesterday, ah, you can create these extracts of meteorological grids for only an area that you are interested in.
2:08:31
If we have a global model Global Meteorological file. So I'd say the GFS model, it's a global model, but we run this simulation with the GFS Model.
2:08:46
And there is no deposition.
2:08:50
This, the Mass, the total mass, here is 200 kilograms.
2:08:56
That will change. That will not change once all of it is emitted. That number will not change for the rest of the simulation.
2:09:04
Because it's a global model. It's never going to leave the domain.
2:09:07
Now, this meteorological demand is just over the northeastern part of the US.
2:09:13
So these particles are capable of leaving the domain.
2:09:23
And also, it increases the simulations. So once it's off the domain, we don't need to keep track of the particle.
2:09:29
We don't need to, uh, transport this particle has to go through the turbulent scheme for that particle, wherever it is.
2:09:37
So it so there's no point really of having a much larger meteorological domain, then you are interested in.
2:09:47
For one thing to be considered, to consider is, the wind can change directions, obviously, so you don't want to have the domain to be exactly only where you're very tight domain where your, where your measurements are, because we know the the winds can go from west to east and then and then change directions to from east to West.
2:10:09
So, you can have particles that transport transport through your area of interest, and then out of the area interested, then loop back. So, so some, some things are another thing to consider.
2:10:27
We're getting close.
2:10:44
Oh, another thing I wanted to show you.
2:10:50
Is in the tutorial, um.
2:10:55
There's also a directory called results, OK?
2:11:01
So some of the lot, very long simulations like the one I'm doing right now There are actually results already pre computed some of them may not.
2:11:14
I've been run up to four this particular version of the tutorial that may be out to date. For example, I know. I mean, I did not replace this file when, when, when we for this latest version of the workshop, so, I'm not sure if it would be identical results or not. So, It could be a little a little bit different, but it should be similar. So, what you could do?
2:11:41
OK, What you could do if you want, is you can copy this into your Working Directory, and then also, then compute dust, No, display the results. Compute statistics off of them, and look look at the output without actually running the simulation, OK? But the simulation just ended. So let's continue.
2:12:11
If I look in my working directory again, I'm going to open up this message file one more time.
2:12:18
And we see as I mentioned, at the very end, now we have 151 kilograms of mass within the demand. So we see that about 50 kilograms left are meteorological demand.
2:12:35
And the output files, the high split to dot bin file.
2:12:41
OK, so let's move on here, So now we're going to display the output.
2:12:49
So if I go to Concentration, display, concentration, kantor's.
2:12:58
And where it's mentioned several times, that automatically kinda sets high split to dot bin because I just ran it, it's that files already.
2:13:08
Kinda saved in the GUI, but you can change this change that, just type in whatever binary I split output file you have, and you can look at other files as well.
2:13:22
OK, let's see what I need to change here.
2:13:29
OK, the concentration multiplier.
2:13:31
So I know the measurements that I have are in units of PICO grams per meter cubed in a PICO gram is 10 to the 12th, PICO grams equals one gram, so I'm going to put in this concentration multiplier of 10 to the 12th.
2:13:52
Change the way. Have a label of PICO gram so that so the out. So there'd be a label now. PICO grams per meter cubed when when I make this figure.
2:14:04
OK, so let's execute the display.
2:14:11
And a new window pops up showing the results.
2:14:17
And if I can hit this and it will move on the Next button, it moves down in time.
2:14:22
Notice the, the, the the color bar changes the contour levels change as, as I move on in time, as it gets no lower concentration.
2:14:37
And the plumes get more dispersed.
2:14:45
OK, so, those are, That's, that's the figure from, from this and now, we can create an animation, I think you all, we also just did that with Mark.
2:14:59
So, let's, let's do that, do that again. So, to remind you, we go to Utilities.
2:15:05
SVG to Image already populated the Khans plot, the HTML and ... GEF.
2:15:15
and we're going to animate it, crop off that white space and hit execute.
2:15:30
OK, It did it once that window goes away it finished doing it OK, so I have now on my working directory, uh This ... plot is the ...
2:15:42
plot dot HTML, that S SVG graphic.
2:15:47
The top one, this is the ...
2:15:48
dot gif, Click on that, and This movie pops up.
2:15:55
So, we just created.
2:16:02
OK, and now we're gonna look at the individual particles, and display them.
2:16:09
So we're going to display particle.
2:16:13
And if you remember, in our setup that config, we had the Particle Output file kind of the default, which is par Dump, we're going to create an output file called part plot.
2:16:25
And Sue, I'll increase the resolution to 80.
2:16:36
And instead of displaying every particle, you can skip particle's anomalous Display.
2:16:44
I'll do every 10th particle, OK, so not every part of every 10th particle.
2:17:01
And there it is.
2:17:04
Or it's coming?
2:17:13
So then, again, increase the size. I can see.
2:17:21
The cross-section as well.
2:17:24
So let me move on in time.
2:17:28
And remember, what Mark mentioned is that you have different colors based on the height.
2:17:38
And typically the winds aloft her faster.
2:17:43
So we have the particles that are higher up in the atmosphere.
2:17:48
They are farther downwind.
2:17:52
OK, next.
2:17:59
Oh, one other thing.
2:18:02
All of these kind of GUI programs that we do, that we're using, that's kind of programmed into the GUI, you can also do them in the command line.
2:18:15
And to work with command line, you'd have to look at the, I would recommend looking at the user's document, and it had, what I went to yesterday, I clicked on all the utilities, and then there's a brief description of how to run each one. Now, every every little, these GUI commands I do is a separate program.
2:18:34
And here it says arrival time. PGM equals ISO.
2:18:38
Chron means that program that you can also do in the command line is called ISO Cron.
2:18:44
And in the high split model, high split directory.
2:18:49
All of these Code utility, all the programs R are located in the exec directory. So ISO kron right here.
2:19:01
So the GUI can adjust runs this program when when you click when when we click this button, but you can also just execute command, execute this from the command line as well.
2:19:14
So let me go back to the working directory. And now. We're going to create a time of arrival plot.
2:19:20
So this this is, it tells you when, in this case, the tracer reached a particular location.
2:19:30
So, let's go to concentration, display arrival.
2:19:37
We're going to call the output file T O a.
2:19:42
A threshold value, we're going to use zero.
2:19:44
So if it's there, it's going to say the time of arrival is at that particular time whenever it, whenever, it's greater than zero, If you can have an increased a threshold value, if you want to know when, a particular plumed with a concentration of, over, whatever, whatever you want, say, 10 grams per meter cube, you can put that, OK? We're going to say if there's a positive value of their time difference, it's a uses the concentration averaging periods.
2:20:16
So if I left this at , it's kinda have intervals every three hours, because in, in the control file, I had the concentration averaging time period of three hours, We're going to change this to six hours, and let's execute this display and see what we get.
2:20:36
And this is what we have. So this tells us what time, at the earliest time period to period between in that 0 to 200 meter layer.
2:20:48
A tracer was in depth particular grid cell. So, you know, right at the very start, at the beginning near that emission source, and then as you go farther down, when it's going to take longer to get there.
2:21:03
OK.
2:21:07
And now we're going to run a program called regret plot, which is going to tell you that concentration of each in each grille it I'm sorry. In each each grid, it's not going to like contour and smooth out like the ... plot. Program does like the concentration contours. That kind of smooths. It smooths it out. We're going to do.
2:21:35
The actual grid values.
2:21:38
So running on this output file contour interval, we're going to change that, change it to the units, are PICO grams, and the concentration out multiplier 10 to the 12th.
2:21:59
Let's execute that display.
2:22:04
OK, so this is what we have sought.
2:22:09
And as you move it out in time, you see that plume move in time as well.
2:22:15
So, this, so, you can see, how big the concentration grid is to every box as one grid cell, OK.
2:22:24
So, that's that. That's, that's what this this program does.
2:22:34
OK, so let's move on to the next section.
2:22:46
OK, so we're going to look at, inside this output file, see what's in there.
2:22:52
OK, so good.
2:22:56
Concentration, utilities, simple listing.
2:23:03
And we're going to get that.
2:23:07
Hi, split file.
2:23:17
Where I split two dot, ben, so as the output file from, when we ran High split.
2:23:28
OK, so, let's look at this file a little bit.
2:23:30
So this shows you it tells you some basic information about what's in the file.
2:23:36
It tells you what meteorology was used.
2:23:39
It was a war file, the met file began.
2:23:44
983 September 25th, 2018, UTC.
2:23:48
The simulation started in 19 83 September 25th at 17 UTC.
2:23:55
The origin point or the the when, where the tracer was released is located at 39.9 -84 to 10 meters.
2:24:05
And it shows us the concentration grid.
2:24:09
How many grids are there in the north-south, east, west direction?
2:24:14
the resolution zero point twenty five degrees by zero point twenty five degrees and the center of that concentration grid Layar Heights are from the surface to 100 meters.
2:24:29
Uh, it, the levels.
2:24:33
There's one level it in the level is 100, so little confusing, that high split for their levels, It has to that the top what we put for the level is actually whatever the level is below to that level.
2:24:47
So there's one level and the height is 100, but what this really is is the the the one Level is an average from the surface to 100 meters, OK? If we at two levels, for example, whether it be a two here, and let's say that second level, what was 500 here? So this would be, our first level would be surface 100 meters or second level would be 100 to 500 meter that the average.
2:25:15
Another thing that I didn't show you that we can also have a level that's just at zero and that would be a deposition value. And actually let me, before I get into that I want to, I want to, I want to show you that.
2:25:31
So if we were in this grid menu, and if I want to look at the deposition and only want deposition alpa, I don't even care about what's the, what's in the air.
2:25:44
This is what output was one level.
2:25:47
The height of the level zero that that would give you the deposition information. If I want to levels what I mentioned.
2:25:56
Let's say surface 100 meters.
2:25:59
And then a second one for an average concentration of 100 to 500 meters.
2:26:04
That's what I want, but if I also want to add a deposition level, I'll do that, I guess.
2:26:10
So, I have the deposition output, uh, average concentration from the surface to 100 meters, and then another level of 100 meters to a 500 meters average concentration.
2:26:26
Got a little off track there, sorry, but let me move on now.
2:26:33
Then also, this concentration in this, yeah. Hi, Split output, concentration file.
2:26:40
That tells you with a simple listing program, you can see the the maximum and where the maximum and minimum concentration is and what it is at each time step, OK?
2:26:53
So the max at is located dislocation in the minutes at dislocation at this in that for this three hour averaging time period and then so on for the rest of rest of the output, OK?
2:27:08
Yeah, say.
2:27:10
Next: OK, so High Spec creates this binary file.
2:27:18
It's a lot easier to look at if it was an ASCII file. So let's let's convert it to an ASCII file.
2:27:25
Then you can use that and want to whatever other utility program that you have important to excel if you want.
2:27:39
OK, So I'm like at this highest point, two, dot, then I want to write out a single file.
2:27:50
The concentration multiplier.
2:27:53
Uh, assays and precision options, How many places add extra digits and concentration Include minutes, have the output, and comma delimited or not deposition multiply. We don't have deposition so we don't have to worry about that.
2:28:15
So let's just execute the display here and see what we get if we look at in our working directory.
2:28:24
Now we have a file called high split, two dot bin as text file, so it's a high split two dot bin dot TXT and I can open that up and this is what I have.
2:28:34
So it year, month day, our latitude longitude, and the concentration of this PM CH value with these numbers. Kinda appended to it which tells you kinda the level information. So this is 0 to 100 meter layer.
2:28:54
And this is only outputting, positive values. It's not going to output zeros.
2:29:04
So but in there, the menu we could have clicked on that include zero, So if you have a program that you can read this in and grid it, nicely button, you need, those zero values also to make it your program easier to work with, or it needs those your values to work with. two, I have latitude longitude every grid cell.
2:29:26
You can include that as well, OK?
2:29:36
OK, now we're going to run a program called Con to STM in a program called Time Plot, within this this other GUI menu.
2:29:50
Utilities convert to station.
2:29:55
So what we're going to do is, at one particular location, let's get a time series of the PM.
2:30:03
CH concentration, OK.
2:30:07
So we're using the high split output file.
2:30:12
What's one thing that's nice about the GUI is that, that's it.
2:30:16
Sometimes you don't like it, but it saves a bunch of stuff automatically in the GUI and moves it into certain menu.
2:30:22
So I don't have to keep putting in this 10 times 10 to the 12th. This concentration multiplier, it's already in there for me.
2:30:30
OK, so at a particular station called C 5, 10.
2:30:38
It's a latitude of 42.25 longitude of , zero point eight zero.
2:30:48
Yeah, and if I hit extract data, it creates this text file.
2:30:56
It just output it.
2:30:59
The actual concentrations at these, this latitude and longitude location at a station. It's called C at 5 10 and it has a time series.
2:31:13
four the whole simulation So now you can we can do other things that we can create a time series plot. We can compare it with actual measurements.
2:31:25
The measurement station is that, that's called see 5, 10 in this, in this particular case, OK.
2:31:32
So, if I look in my working directory, we have that same output file con to SDN.
2:31:40
This is the same same file.
2:31:43
We can plot that data.
2:31:48
Make a plot to hit plot there, and this is this is what we have a time series plot of the MTH concentrations at the C 5, 10 location.
2:32:02
OK, and we can also add measurement data right on top of that, if I can find this measurement file, which is located in the tutorial cap taks, cap text to M 5, 10.
2:32:26
And, actually, before I do that, let's look at that file.
2:32:31
So, you can understand, see how that file is formatted cap tax, two and 5, 10.
2:32:41
So, here's the file, as the year, month, day, the starting hour, and the duration of the measurement, which is three hour average.
2:32:52
The latitude longitude of the location, the concentration PMC H then the station number is 5 10.
2:33:02
So, this is kinda our measurement what we call the Measurement Datum File, uh, Format.
2:33:08
So if we have measurements, we can put them in this format and easily compare them with high split with the utilities that we have.
2:33:16
So, I'm gonna hit plot data again, and it has the time series of high split overlaid with the time series of the measurement.
2:33:31
OK, so, let's now move on to the next section.
2:33:41
And now we're going to perform some statistics based on the measurements and the high split simulation.
2:33:50
OK, so We have more than just measurements at that one location.
2:33:55
So if I open it up cap text to message dot T X T, which is also in the tutorial cap decks.
2:34:04
We have, it's formatted the same exact way as that file. I just opened the M 5 10, but it's for all of the stations.
2:34:14
OK, So as stations, and for the whole time period of the cap Text to Tracer Release experiment, so we're going to use this measurement file to compare with our high split simulations.
2:34:30
And you can read more about the data archive that we have.
2:34:33
We have observations from multiple tracery lease experiments.
2:34:42
Throughout the U S, and for different spatial and temporal resolutions from this regional scale simulation that we're using for the CapEx experiment, and kind of suburban urban spatial scales like this, very high resolution, spatial scales as well.
2:35:05
And you can use this data for your own purposes, your own numerical experiments, and evaluating your your model or dramatizations that you may may want to use.
2:35:20
So, let's use all this data.
2:35:23
So, I'm gonna go to concentrations utilities and convert to Datum.
2:35:28
So we're going to convert our high split output file to a data format so we can easily compare the high split output with all of the measurement data, all the surface measurement data that we have.
2:35:46
So we're using our high split output, conversion is 10 to the 12th.
2:35:53
The Measurement Data File, we need to go find that to then tutorial cap texts maj dot TXT.
2:36:02
Then hit this Create Datum File.
2:36:05
OK, so let's see what it created, the output I had saved as high split dot T X T and here it is.
2:36:13
Click on that, and this is what I have, so year, month, day, starting our duration, latitude, longitude, PM, CH concentration, and station number. So it looks it's formatted, identical to the measurements, how the measurements are are.
2:36:31
So we have all these zeros because it's the beginning, but move on, we have much higher concentrations here.
2:36:41
OK, and now we're going to hit this Compute Statistics button, and it is going to pop up this window with lots of statistics for us to look at.
2:37:04
OK, so from this file, we can see that there are 395 kind of data points that the hour that were processed, our simulation has a correlation coefficient of zero point eight one.
2:37:21
The average measured concentration was 458, PICO grams per meter cubed. Well, the simulation was 264 PICO grams per meter cube.
2:37:30
So we do have kind of a low bias.
2:37:36
And in our high split simulation. And there are many other kind of, statistical measures in this file There, is.
2:37:47
Figure of merit in space.
2:37:50
This kinda describes the amount of overlap of the positive measurements with the calculated positive values, So we see that there's about 69% calculated positive values that also show up in the measurements.
2:38:07
Let's see.
2:38:11
And if we look at the, the 95th percentile of, this is very, very high concentrations of the measurements, uh.
2:38:25
The concentration was 14 50, about 450, PICO grams per meter cubed.
2:38:32
Whereas the 95th percentile in the simulation, it was, it was a little bit lower, 13, 27, but then if we look at the 90th 75th, and 50th percentiles of the measured and calculated, they are very close to each other. They're there and very, very good agreement.
2:38:47
So the low bias is really based on the this very high, the 95th percentile, which you would think would be very close to the trace release location, so, close.
2:38:59
So the closest measurement sites to to the Tracer Release where farther downwind, we're getting very, very, very good agreement.
2:39:09
There's also the, the, the column on Grove ...
2:39:12
parameter.
2:39:14
This describes how similar the results are where, if you have a value of zero, this would mean R Model agrees perfectly with the measurements.
2:39:22
And then we have something called this Final rank, and this is equally weighted, kind of statistical measure, this equally weighted between correlation, fractional bias, Figure of Merit and Space, and the Kolmogorov Smirnov Parameter, where, if R phi in this final rank, ranges between 0 and 4, where a value of four would be perfect agreement between model and observations.
2:39:53
OK, and now, if I hit scatterplot, we can see the scatterplot of our simulation pops up, and we see the highest concentrations up here in the, with a 300 numbers.
2:40:06
We said that, yes, there is a low bias in the simulation where the 1 to 1 line is this dotted line, but below that, the measurements and models are in fairly well. Wellesley distributed along this 1 to 1 line.
2:40:28
OK, so, the next thing, I am not, yeah, I'm not going to go over this, this next simulation, but I'm going to walk you through this because it will take longer to run what it involves running with multiple. Actually, no, I'll set it up for you, but I'm not going to run it, because I don't want to take the time to run it. So what does kinda little experiment?
2:40:52
Does it say what if we run with a higher resolution meteorological model simulation?
2:41:01
Along with this course, course are 27 model simulations.
2:41:06
So, what we would do is hit Add Meteor, I'll add Meteorology Files, find this cap Text to work three dot ben, File OK, then just hit Save and then run.
2:41:21
So, what this would do is when the particles are within this high resolution modeling demand, which is very well confined mainly around the tracer release.
2:41:33
The particles are going to be run with this high resolution meteorology what is outside of this high resolution meteorological domain.
2:41:43
High split will use, the coarser domain byte by default. I split always uses the higher resolution meteorological simulation.
2:41:52
And then if it if we ran this This is these are the, this is the scatter plot. We would we would get if if after we convert the file to data, compute, compute the statistics.
2:42:06
And what we see is that these higher values closer to the Tracer release, are right around that 1 to 1 line, and these other concentrations are us stay about the same stay around that 1 to 1 line.
2:42:21
So to higher meteorological model resolution model, near the Tracer Release improved the results for this for this particular case.
2:42:34
OK, so let's move on to the next section.
2:42:40
OK, so now, what I'm going to do is kinda test, high split, to test the model inputs, to see if how we're configuring it, if we're doing a good job of, or whether it's configured properly, or whether we should do something else.
2:42:58
OK, so I am going to start by loading this cap Texts Control kept at Setup, I don't remember if I messed up the CapEx Control and setup file, so I'm gonna I'm gonna load those files. So I'm gonna go to Set up Run.
2:43:14
And hit retrieve.
2:43:17
And I know they are at, in the tutorial Files Directory, cap texts control, OK, go to advanced configuration, setup, concentration, and retrieve.
2:43:40
Again, in the tutorial files, sub directory cap tech setup.
2:43:47
And I'm also going to decrease the amount of particles.
2:43:52
So let's see what happens if we significantly less.
2:43:59
Particles released.
2:44:05
I go to concentrations, special runs.
2:44:09
And test inputs.
2:44:10
So this isn't running the actual dispersion. It's running through specific routines to test to see how we have the model set up and if I run it.
2:44:25
Then if you run it, I don't know if you heard, you may not have heard it, but there will be a beep or a noise when it ends.
2:44:33
So finished and then I can look at some of these other files, this warning message control set up and run it.
2:44:40
In working Directory.
2:44:43
It created them too and they all end an underscore mod OK so I can look at them here or look at them by just pressing this button.
2:44:51
So I'm gonna I'm gonna click on the warning And this is what pops up.
2:44:59
So similar warning message that we saw before it has to do with the war vertical interpolation scheme by default It's changing it is changing it anyway. So we can choose to change it in the name setup file or or it will just do it automatically for us.
2:45:17
Concentration, Gritz sampling, is saying, we can have the time end at 13 UTC.
2:45:26
And instead, previously we had it, I think, at 15 UTC.
2:45:34
It's saying this non par value, we just change it to a very low value of 2500.
2:45:39
It's saying that's too low for the simulation. We should have it at least 7000 release, at least 7000 particles.
2:45:46
Now, as we, as you Mark mentioned before, if we have don't have enough particles, our concentration grid is going to under sample.
2:45:56
There aren't going to be enough particles that fill up these concentration.
2:46:01
Grid cells do give you a nice, smooth patter of the concentration, and you get very noisy results.
2:46:09
And, if you don't have enough particles, say, for example, if I ran width, a thousand particles in 2000 particles, if I compare those?
2:46:21
Simulations, I could have very different results, value.
2:46:24
It can be very noisy, this, comparing it with measurements, statistically, The measure that their evaluation will be statistically different, but, if you have enough particles or too many particles, the results won't change, statistically the results won't change. The look at the display, it will be very similar. Similar images.
2:46:50
So, yes, OK. So, basically, this is telling us we can change this to 13 and change this to 7000 particles.
2:46:57
And if I look at the message, file, I think there's, I don't think we need to look at that, Hit the control file to menu's popup, OK.
2:47:10
one is my original control file control file contents, and the other is kind of the suggestion of what I should change, control underscore mod.
2:47:20
And if I compare these.
2:47:25
We can see that.
2:47:26
it suggest, changing the stop time this from 15 UTC to 13 UTC.
2:47:32
And the Reason is, is because our runtime, at 68 hours, it actually ends at 13 UTC It doesn't go to 15 28 the 28th at 15 UTC. So that's why it says change that.
2:47:44
So you can change that if you want.
2:47:48
It won't have much, it won't really impact your results, but you can if you want.
2:47:56
And now this change will impact results because it told us we did not have enough particles trying to increase.
2:48:08
OK, so my original setup, that can Figg?
2:48:12
I can find it.
2:48:15
None par of 2500 and my model.
2:48:19
The modified, it recommends 7000, OK.
2:48:25
So if you want to just use these recommendations, you can hit the copy button, You will just copy the, these mohd files to to the input, the input files, that you can use.
2:48:39
OK, so right now, I'm not going to do that.
2:48:43
Then.
2:48:47
I'm gonna.
2:48:50
Do another test, what am I going to test.
2:48:54
Oh, that's in the next chapter.
2:48:58
OK.
2:49:03
OK, I'm going to do another test, but this time I want to change, increased the resolution of the concentration grid, before running this test, to see what changes. So I'm going to go to setup run.
2:49:18
Pollutant Deposition, Grid setup menu, and hit Grid one.
2:49:24
Change the Concentration grid from zero point twenty five, zero point twenty five to zero point zero five by zero point zero five degrees, OK?
2:49:37
Hit Save a few Times.
2:49:41
And then Concentration, special runs.
2:49:46
Test inputs, hit Analyze.
2:49:56
Finished.
2:49:58
It's a warning message.
2:50:01
And similar warning message, except it says, now I have to set the non par, 54,000.
2:50:08
This is because I'm decreasing the concentration grid to a much smaller, much higher resolution concentration domain. So, I need significantly more particles in order to get an adequate amount of particles.
2:50:23
In these high resolution grid cells, too, create and evenly smooth out, not noisy, noisy results.
2:50:35
And this, it will show up in the setup that config to it would have the non par here, set to 54,000 and.
2:50:48
Now, I'm going to do this, run this one more time, but I'm going to change, reduce the runtime, from 68 hours, 25 hours, and see what happens.
2:51:02
OK, so, I go to special runs.
2:51:05
Test inputs, Analyze.
2:51:11
Hit the warning message, and now, even though I have this very high resolution concentration demand, ... is still at seven, only 7000, the recommendation, from 2500 to 700, that, that 54,000.
2:51:28
This is, this is because, as, as time goes by, these particles are disperse.
2:51:35
So, close in tromped close near the tracer release at the ... location, there are lots and lots of particles. So, near those grid cells were getting enough particles with seven.
2:51:48
You would if you run a 7000, you would get enough particles within those grid cells to create a nice distribution of particles.
2:52:01
But after 25 hours, these particles start moving farther downwind, they're disbursing more.
2:52:08
And now the particles are, there aren't enough particles to go into all of the different grid cells, farther down wind.
2:52:20
So if we run a 7000 particles, get farther down wind, that is where we're going to have very noisy, noisy results.
2:52:30
OK, so now I am going to run a simulation.
2:52:37
With two grids changed the non part to 10000.
2:52:43
And what I rank, What we recommend is that that 7000 value, that is a minimum, we'd like to use more than enough particles, OK, in order to not run into, not run into trouble.
2:53:01
So, I'm going to change the non par value from 2502, 10000.
2:53:11
It's safe.
2:53:14
And I'm going to run with two grids. We haven't done this yet.
2:53:22
OK, so I hit Pluton Deposition grid setup disbursed Grid.
2:53:27
I'm gonna go back to my courser two man, OK, that zero point twenty five by zero point twenty five Degree demand.
2:53:37
And now, if I want to concentration grids, you can have as many concentration grids as you want, and they can be located anywhere you want.
2:53:45
You can have a concentration grid in New York, another one in California, or you could have them nested, which is what we're doing where a very high concentration, high REZ, while the second concentration grid.
2:53:58
I'm gonna have a high resolution, a higher resolution concentration grid, or near the source.
2:54:04
And then, of course, your concentration grid that covers a much larger area.
2:54:10
So for this second concentration grad, I'm going to change the center latitude longitude to right over the release location.
2:54:24
And it's going to have a resolution of zero point zero five by zero point zero five.
2:54:31
And only cover span five degrees latitude and five degrees longitude.
2:54:37
So it's only going to be really close to do the emission source, the tracer release.
2:54:46
Change the output Good name.
2:54:50
So, they don't conflict. Give it it.
2:54:51
Give it a new name, and I'm only going to have this concentration grid exist for 12 hours.
2:55:04
OK, after that time, the particles will be morally well dispersed or off the meatier lot off this concentration grid. So, I'll only simulate them in, in the course.
2:55:16
Ah, demean them.
2:55:20
OK, so hit save a few times, and then run the model, and Hopefully I set everything up correctly.
2:56:21
OK, it's getting close.
2:56:24
Yeah. So thinking about how many particles to use.
2:56:28
That's that's tricky at the trickiest. So that's why this, this new option, this test inputs file, which is somewhat new, is it, is great for us.
2:56:39
Uh.
2:56:44
Kinda helps guide you to decide what you should use and make sure, have that under sampling that that can occur if you don't have enough particles.
2:56:58
OK, so now let's look at the results and look at them for the different concentration gritz that we have.
2:57:07
So I'm going to go to display concentration contours first. So we have, we have two buttons up here now.
2:57:16
The course, castration grid, which is the first concentration grid and the second one is the high resolution .... So let's look at the higher resolution one first.
2:57:25
I am going to change the output file name, two cars plot F G and I'm gonna make some rings around the release location at 50 kilometer distance and set the center.
2:57:50
Of the of the Figure to over the release location, though it's already set automatically for me where the tracer releases.
2:58:01
I think everything else, I want to keep you up as the concentration multiplier units of PICO grams.
2:58:11
I'm not going to I don't need the measurements. I'm not going to show the measurements. Now. Let's just look at what the simulation showed.
2:58:18
Hit Execute, Display.
2:58:26
And this is what pops up, So what these rings and the rings set. four rings, 50 kilometers.
2:58:34
That's kinda what we showed showed, you know, what, I'm going to increase this is, uh, X out here.
2:58:41
I'm going to zoom in.
2:58:46
OK, that's better.
2:58:49
So we have these ranked at 50 kilometers.
2:58:51
This is our high resolution simulation, move downwind, a little bit.
2:58:57
This is what we got.
2:58:59
And go back to this time period and now I'm going to, let's look at the same figure before the courser simulation and I think everything else is already set up properly.
2:59:16
Hit Execute display.
2:59:20
And this is what I have.
2:59:23
So this is the course Rohn.
2:59:27
This is the Higher Resolution Rohn.
2:59:29
So, you can see that you get more resolution, closer in.
2:59:35
That, then, then the course run, OK, I think that might be all I wanted to say, OK. Yeah.
2:59:45
So the next up, are the exercises. You guys can do these on your own.
2:59:51
And again, the hints are very good, hence, I saw kinda that this, click on that magnifying glass to help walk you through it.
3:00:04
These problems are kept.
3:00:06
So you can do this on your own. If you have any questions, again, that forum is there for you to ask questions.
3:00:16
And now I'll turn it back over to Mark.
3:00:19
All right. Chris, thank you very much.
3:00:23
And we will go into a longer break now, and we'll go ahead and keep the schedule and come back in, um, at 1 30 PM Eastern Time. So, about an hour and 13 minutes.
3:00:43
I'm going to take that controls back, and we'll see you back then.
3:00:49
And 130 are an hour and 13 minutes for the next section.
3:00:56
Bye, Bye.
3:01:15
OK, we are back from our Break, and we are about to go to the next section of the workshop of the tutorial, and this section is called Air Concentration Parameter Sensitivity.
3:01:33
And we're going to be looking at what happens when you change some of the configuration parameters and how that might change the results of the simulation.
3:01:44
And the person that will be presenting and teaching that section is doctor ...
3:01:51
and fan team.
3:01:53
I'm going to try to change the presentation over to you.
3:02:08
Let me change the screen.
3:02:11
Oops!
3:02:15
Can you see my screen with the menu?
3:02:20
yes, yes, perfect, alright, thank you, ..., take it away, OK, so, hi everyone, in this section, we will look more closely at the concentration calculation in terms of the various mato settings, and how they may influence the results.
3:02:41
Um, instead of running the full duration of CapEx number two release, and comparing the result with the ground measurements, like what you have seen this morning, presented by Chris, we will have configured a motto to cover a limited.
3:03:00
Limited Time period to corresponding with the aircraft measurements, show that.
3:03:08
So the simulations will run quickly.
3:03:12
The aircraft sampling was up, cut off hundred kilometers downwind, and cover a lot of spatial details, Just a heads up. I won't be doing order runs in the in this section. I'll try to explain the rationale behind them.
3:03:30
You'll see, like many of the simulations have, have put his similar results. There are a lot of options.
3:03:38
in the advance menu, we will be looking at, The point here is not to run all these simulations and get the same results as the, as showing you the tutorial. But, to know, how like, why we, we have different configurations And what they are useful did. This will help you to better understand the ensemble section then and talking about the concentration calculation uncertainties later in this workshop.
3:04:13
So now we will load a cap tech setting and dance. Fine tune that simulation to look at the aircraft missiles.
3:04:24
So let's hit the reset button.
3:04:27
Oh first, I want to show you the aircraft data sets. So let's go into the tutorial directory cap tax. And then there's a file call, Fight Cyril 9, 14 dot T X T.
3:04:46
Um this is the aircraft data we are going to use.
3:04:50
But for this exercise, that was X chapter two, and not a file called, data underly, TACE dot T X T.
3:05:01
Which contains data during Civil seal three UTC on September 26. So, this is just a subset of the, the original data, the freight, 190914 dot TXT.
3:05:18
So you can see that the aircraft aid her at six minutes interval.
3:05:26
So, I will leave this here, because we are going to use this in this exercise.
3:05:31
Now, let's go to the menu category, and then click Reset, Whoops.
3:05:41
And, then, bring up the concentration setup manual.
3:05:49
Less louder control file.
3:05:58
It's in, Skirted, the tutorial, kept text directory, and then look for CapEx and the lie.
3:06:07
Control, doc T X T, here we go.
3:06:14
OK, And then let's change a couple of things, set the duration to 13, we we don't need.
3:06:29
Because the aircraft data starts at 0, 3, UTC of the next day, self 13 hours is enough to take us through two lights, us, six UTC.
3:06:41
So, and then the met, they just want to make sure the Directorate has a correct.
3:06:48
So, let us go to the Tutorial Cat Tax here, and then select kept text to an ally was 27 years up to Adapt van File.
3:07:02
And then go into the grid, set up.
3:07:11
So, dissembling starting time, let's change it to September taught us September 26 0 3, 0 0.
3:07:24
And I can say, Save, and go to the advance menu, then, again, a lot.
3:07:35
The file?
3:07:39
Cap tax.
3:07:42
Set up the T X T file.
3:07:46
Click Save, and then Random Model.
3:07:55
Yes, to take a while, to run it, because, Yeah, this is, we're assimilation.
3:08:04
Because we are running for a shorter period of time, so actually is not necessary to run the simulation with 50,000 particles, especially at this concentration grit resolution. So, actually, we are going to change that later to reduce the computational time.
3:08:29
So, yes, ongoing assimilation.
3:08:33
For display, less, look at the chis ****. Yeah.
3:08:39
We're going to Yes, setup couple things for the display, once the simulation is done.
3:08:54
And then, we will be looking at a few more changes in order to make the assimilation better, or more, or less take, cut to compare with the aircraft data, and then, in the next subsection and then, that will be our base configuration.
3:09:20
And then from that point, we will be looking at bunch of, lot of settings, especially in the events concentration, menu number seven.
3:09:31
Like a lot of personalization that we can use and choose in the high speed.
3:09:38
And then to to do the simulation almost done here: 92%.
3:09:49
Thank.
3:09:50
Open the Display Manual concentration contours.
3:09:55
Set the center the map center too.
3:10:00
What are your 1? -82, I have five.
3:10:06
And then the the unit conversion, so, is tend to 12.
3:10:17
And then put the label pictogram, and as set the color scale contour scale to this less. We can copy it.
3:10:42
And then correct execute.
3:10:46
So, you can see that the peak.
3:10:51
That is because the concentration matter is about 5500 PICO gram per cubic meter. Here, the center, like the maximum is located here indicated by this red box.
3:11:10
And then less at the aircraft data, basically this file opened here.
3:11:18
So, you can put ah the aircraft dayna file here.
3:11:25
I select data under like haystack T X T, and then execute.
3:11:33
So you can see that this is cusp, size, and also all these numbers. They are aircraft measurements. So, you can see that the width of the ... is wider than than the calculate at one. So, the aircraft data showed the peak.
3:11:54
Battery is about almost three 30,000 PICO gram per cubic meter.
3:12:02
So, basically, our simulation, the simulation we just did is it's under protecting the peak. So, that's why I mentioned earlier we will optimize the configuration to make the simulation closer to the aircraft measurements.
3:12:20
So, let's move on.
3:12:23
Um.
3:12:26
So, this say ah there are a couple things, we want to consider.
3:12:32
Why is that the F craft sampling is at mean sea level about 900?
3:12:40
You can see that is about 900 meters.
3:12:44
Mean sea level L we will move on to small things up by going to a Gaussian and then we want to reduce the motto output time interval from three hour average, the reality average to a shorter interval.
3:13:00
So, let's go back to the Concentrations menu.
3:13:08
The first thing we want to do is to reduce the grid resolution as open-ended.
3:13:18
So, the the the cap test case like we use 25 kilometer definitely S 2 2 cos.
3:13:26
So, we want to change it to our fina with will use ... degree ...
3:13:34
by suicide and then safe, safe, safe, and then go to the events menu number three.
3:13:51
So, um, sir less less, choose Gaussian, Salt, we will slow things up by by going to use, ah calcium, and then, save it, and they also go to manual, number thought, two, to reduce the particle number year, the way this particular number of mean, we have 50,000 here, and then less change, too, 5000 party cop.
3:14:29
And then Save, Save, and then run the model.
3:14:38
So you will see that the simulation runs.
3:14:44
Faster.
3:14:45
Just because we have feel a particles set up and then put it linear in terms of the wall clock time versus Dipoto number.
3:14:58
Ah.
3:15:01
And I ruined after it's done.
3:15:06
Look at the result.
3:15:16
By the way, when I was doing this earlier, actually last night, and then it ran much faster puppy. I need to restart my computer herself.
3:15:37
Um.
3:15:40
Almost done.
3:15:55
OK, Less to Display See what happen.
3:16:06
This is interesting.
3:16:08
This is not what I am expecting, but the **** here is slight too, right?
3:16:17
While wheat when we go two different setting, like I tried to use the Gaussian, best look at the Yeah, cheese **** here.
3:16:29
So basically I try to use 500, 5000 particles, and then going to calcium here, and also that high resolution.
3:16:45
Let me try to do that again, to make sure I have things set up correctly.
3:16:56
So I try to decrease the grip my solution.
3:17:07
And then go to the Concentration events menil number, number three, menu number three, and then pick the Gaussian Horizontal Particle Vertical option, and then save, then go to the menu number fall.
3:17:33
Changed.
3:17:36
In part, the part petticoats varies per cycle to 5000.
3:17:41
In order to reduce the computational time and then save it, save it.
3:17:46
And then this should be the the the result from that setting and then you can see that.
3:17:57
To pick where to actually goes up to seven is 67,700 PICO Bamford per kilometer.
3:18:11
But the motto premise still light too narrow. You can see that and the content concentration is, is at the south of the absurd pick.
3:18:22
So, to make the simulation better for comparing to aircraft data, we want to change that averaging time and concentration layer.
3:18:33
So let's try to do that.
3:18:37
Highfalutin? Yes.
3:18:39
The resolution, the concentration resolution? Does it needs to be, like, a point O five, instead of zero point five?
3:18:47
Oh, yeah.
3:18:49
Aye.
3:18:50
I typed something else. Attacking something as public. Somebody pointed that out.
3:18:56
Yeah, right.
3:18:57
Yeah, I was yeah. I have my notes here. Is that puzzle five, but actually typing point five?
3:19:06
Thank you for pointing that out. Thank you, Sonny. And thank you for those who pay attention. On the chat, OK, thank you.
3:19:18
Safe.
3:19:19
So, I have a kid moving forward.
3:19:24
I mentioned that we want to change also the averaging time and concentration layers?
3:19:35
Let's do that.
3:19:39
So, OK, let's double check.
3:19:41
Now, there, the great concentration group, my solution is on Civil five base points, they were five, and then change the average averaging time too, from the reality to hourly.
3:20:02
And, also, um, that uh, concentration layers, 2, 2, and then we would do a tangent and one thousand.
3:20:17
So, basically, we will have two layers first, ...
3:20:19
on the surface up to 800 meters, and then the second layer is 100 meter to one thousand meters. So, that the second layer will be will be, compared with the against a will become pair with the aircraft data.
3:20:40
So, click save, save save.
3:20:44
OK, one more thing, we have already reduced duration, 2, 13, 13 hours, but actually because now we are doing the averaging averaging TAM one hour averaging time, that means we can run the simulation for two hour less.
3:21:07
So, we can set the duration too 11 hours.
3:21:14
And one more thing is go to the events menu, number two.
3:21:21
So, here we have the option of Les using height above graph. Because we're going to compare the result with the aircraft data which actually is immensely levels so, we would set it, set, set the motto to, of additive to the minister level, click Save.
3:21:45
And one more thing and go back to the, ah, set a file sap menu.
3:21:55
So you can see that now if you click the starting location since we switched the unit now the manual shell means the level that her away deterring high in this area is about 200 to 300 meters. Means the level has its internal co-ordinate system is sigma and is relative to two to the terrain. So.
3:22:19
So the starting high intent meet her means, starting at the ground level, I mean no matter cannot stop dissemination below the terrain so that's why this is at the ground level.
3:22:35
OK, Ladd, You'd random, the model.
3:22:55
As I mentioned before, when we are going to display the results, we will be selecting the second layer, concentration grid. That is one thousand meter level.
3:23:29
Kay, Scott, who does pay Cantaloupes So now, because we set the concentration, great to two layers, where kone so.
3:23:41
Now, you see, you have option to choose, like, 100 meters, or one thousand meter, so one thousand meters level.
3:23:52
And then, now, we're getting the peak concentration right near where the upset pick where it was.
3:24:02
And, and a model is predicting the maximum about at 21,000 PICO gram per cubic meter. And then the ... S is a little bit wider than the pure assimilation any small compatible to the aircraft data. So this this is our base configuration.
3:24:29
OK, Great Less Now we have our base configuration, then.
3:24:35
Let's go to the next section from now on, we will look we'll be looking at Advance Menu seven.
3:24:45
Let me bring it up first.
3:24:49
Um.
3:24:57
And you can see that there are quite a lot of things here.
3:25:04
So this manual is, it's about configuring high speed, how to compute, Tupper and velocities, ... timescales that is better than to the turbulence, PBL boundary layer stability, mixing layer death and dispersion raise.
3:25:27
Um, let's go back to actually, Section 7.2. Web one to utilize that equation mentioned by Mark earlier.
3:25:45
So, in high speed, to computer transport and mixing of particles, there are two steps, the olive actions and dispersion components.
3:25:57
The advection calculation use the mean wind feel from the input metallurgical file that is essentially the tragic to a calculation component.
3:26:08
And then, the Second Step is to add the top or N component to the particles mean trajectory. So, the dispersion calculation we cry it, the standard deviation of the terminal velocity that is to add the random component component to the advection motion depending on the stability. And turbid and parameters that are available in the mythological Data file user can choose different options.
3:26:42
It has be to estimate a turbulent velocity.
3:26:45
So this is lot we have, in the events menu's number seven.
3:26:52
Ah, let us look at the first part of vertical turbulence setting.
3:26:59
Most of ... datasets do not have a turban fielding them. So that's why we need to estimate the amount of turbulence from the mean wind fields that are part of the ... data file.
3:27:12
And I use for the treachery calculation the mean wind field, temperature and moisture, and when as a function of high, I used to compute a step at a T and turbulence values.
3:27:27
Let us go back to the section 9.3 here, The ... approach is so-called count a caisson.
3:27:36
Um, this is the set of equations used to derive the vertical and vertical and to harvest Santow turbine components.
3:27:49
Um, they add the Turbulence Statistical Radio?
3:27:52
So, if you, if you take the square root of that, that is the standard deviation.
3:28:00
So, for vertical component is W Pi, it also cause sigma W The equation here shows that the turbine velocity as a function of.
3:28:16
Fiction velocity, U stars eight, and the Boundary Layer Boundary Layer Debt.
3:28:23
So, if the ...
3:28:25
that does not have these variables, as it will be diagnosed, that for the ...
3:28:32
data file, we use in the, in, in doing this tutorial, these fields are available. So high speed will just use them directly to compile, to tab around the ....
3:28:48
As you can see in this menu, then nothing actually select a thought vertical turbulence, because when when hudspeth starts up, it actually look at the mechanical data file to determine what option to use.
3:29:06
For instance, if, if the ...
3:29:10
file contains turbulent kinetic energy T K E, the motto is select death for you automatically.
3:29:19
Otherwise, it would it will just go back to the default setting using ....
3:29:26
If you select an option in this in the menu, it will force it to use it to use that option. Unless has been cannot use it for whatever reason. You can you can select TK Option, but if it's not available in your ... file, that has, we cannot use it.
3:29:45
That has to switch back to the default option, which is count a caisson.
3:29:53
That, to add up approaches, you can see here the bow John Huss Stack and the Hannah M F that, they are based on different sets of equations.
3:30:08
If you look at them for sample, here is the equations for the hana Methought, they look quite similar to the kind of Kaizen equations, right.
3:30:23
So the common practices that, like Peephole go to the field to take measurements of turbulence and main variables, and then they tried to atomized them and come up with us cemil, semi empirical relationships.
3:30:40
So desecrations came out of that according to 2, 2 different measurement campaigns.
3:30:51
In many of your can also see Ann, an option car, Mattia variance that refers to when the Metallurgical dataset has U Prime V Pi and W Apply.
3:31:05
So, basically, these three variables, these are turbulence, but ... then husband can just use them directly to compare the dispersion component.
3:31:16
That is a where K is. Surely we don't have to ... in the mechanical file.
3:31:23
You can see that light three hour, five Methought die now. Step 3 of 5 mythos here ... Based upon mean variables and verbose better than that.
3:31:39
boundary layer stability depending on the stability regime and the attitude.
3:31:46
The motto has different sets of equation today announced a mixing of particles. Sometimes they arise due to the process of we diagnose thing where we grab a pulse. May be carried to the dispersion simulation.
3:32:01
But the advantage of this mythos is that you don't need extra elbows.
3:32:07
You just need to provide mean that mean variables. Then you can calculate audi's mixing as Deputy MCH variables for you.
3:32:20
Is if TK he is provided by the ...
3:32:25
Core Data file, as we can use the, that is, can use it to computer, turbulent velocity, by partitioning, the total ... to the vertical and horizontal components.
3:32:41
So this option light really depends on the quantity of the TK protection from dermatological Model and how to celebrate Nimble partitioning ratio between the vertical and horizontal components of TK.
3:32:57
So let's do some run soothsaying, different vertical turbulence options. So I mentioned that the default setting is kinda case, and so this is the base configuration we did in the previous subsections.
3:33:13
So that means this is using kinda Kaisa.
3:33:18
Now let's pick another Vertical turbulence option.
3:33:23
Let's pick our Hot Hannah and then Save Save and Random Model.
3:33:45
I will bring up the advanced menu number seven.
3:33:49
Again, because we will be talking it light throughout this section.
3:34:11
Again, let's display the con until the salt.
3:34:20
OK, you can see that this is the base simulation using ... case. And then this is the one using hana.
3:34:31
So you can see that there they look very similar, um.
3:34:39
Using hana option, the peak value is now tiny 4000 PICO Gram per cubic meter, so, compare to the base case, 21,000, so it's a little bit higher.
3:34:56
Let's close that.
3:34:58
And then, um, another important parameter here is the Lagrangian type Scale.
3:35:06
Let's go back to the question in Section 7.2.
3:35:14
When we come pealed the autocorrelation coefficient ah here it's related to the time step divided by the La ... of the autocorrelation timescale.
3:35:30
That refers to a measure of the timescale or length scale of the turbulence. You can consider it as the time it takes for the turbulence to be kept to become uncorrelated.
3:35:44
So if you have very small eddies, they don't last very long and the things become like random much more quickly. So the autocorrelation chops more quickly.
3:35:58
For larger concha timescale, the eddies have a larger spatial scale and then tang, decoherence, for a longer time period of long for a longer period of time. Which means that autocorrelation stay for longer time intervals.
3:36:17
So, in high speed configuration, the Default Folder ...
3:36:22
timescale is NaN for stable condition and NaN for unstable condition for political mixing, and 12.5 for horizontal mixing.
3:36:38
Sometimes, you will see ... here in the menu.
3:36:42
That means undefined, which will go back to the default setting NaN. If you want to actually see what is used, for the ... time scale, you need to look at the message A file.
3:36:56
So, let us go to the message file like from how peers run?
3:37:05
That is in your working directory. And then you can use.
3:37:14
You can use notepad to open it.
3:37:17
Jeff.
3:37:20
So, here you can see the rescale be scale as we scale, you saw these ah the values for the tan ... timescale.
3:37:38
So, normally, we use ... for the Lagrangian Timescale. We can't force it to be time dependent.
3:37:46
When when we set the staple rather 2 -1 and 0, the model compares dipanjan timescale using the hana question, which is shown.
3:38:03
Which is shown here.
3:38:08
Yeah.
3:38:09
Basically, it is related to the maximum energy wavelength, which is a function of stability and and the existing turbine's values.
3:38:24
So, let's try this, Now we are going to do a run using the hana, but a call turbulence options, and then did time dependent upon gen timescale, so set the stable value here, . Click, save, Save, No Random Model!
3:39:00
OK.
3:39:08
Same thing for the display.
3:39:15
So, as I mentioned, some of the, the other approaches were less sensitive, but this particular configuration is quite sensitive.
3:39:25
You can see that now, the term shape is different from the pyrus simulation.
3:39:31
This is the base base configuration.
3:39:34
And now, here, couldn't shake is different, is narrower and the peak baddiel, is alpha predicted.
3:39:46
You can see that here is 73,000 Pictogram, prepare them pivot cubic meter, higher by more than a factor of two a competitor base case.
3:40:03
So, let's go back to the events menu number seven again.
3:40:16
So we talk about the vertical turbulence options, and for horizontal turbulence the default approach is to compute a value in proportion.
3:40:30
To that vertical.
3:40:32
Not an option is to use the diff deformation for horizontal, what happens?
3:40:40
The idea behind this is, if, if you have large differences in the horizontal velocity, between the neighboring quiz sow's that the field is being shear, shear and deformed. Dan, you may have lots of turbulence on the sub grid scale.
3:41:00
So this is what to account for that.
3:41:06
This is more for the for the longer range applications, because for shut, range, simulation, I mean, here, I mean, smaller than hundred, kilometers the deformation characterization use in.
3:41:22
In conjunction with large-scale meteorological feel, well, will not refract ... in horizontal tolerance.
3:41:30
So, lately, we prefer to use the default setting, which is in proportion to the vertical component.
3:41:44
And for all these settings, Hideaway, normally we, we, you can do, you can lift the lift them as default. In fact, we don't necessary know which one is better. Depending on the weather condition, geographic location, order metallurgical dataset you are using one approach May, may work better than another one.
3:42:09
And different settings can be used to construct ensembles simulations that will be discussed in section 12 type thing OK.
3:42:23
Let us go to the next section.
3:42:30
OK dissection, it's about, How, do we compare stability in the model?
3:42:41
Dad, turbulence equations requires some mesh of step at a tilt later manion, path, length and riches some number.
3:42:53
So, here, we can, we can compare Step additive by using the heat flux and momentum flux, which is the default in the, it has fit, and it is using this set of equation.
3:43:11
You can see that U star is the fixture will last T T star is fiction temperature which is the function of the sensible heat flux H.
3:43:26
Then using this variable, we can compute as deputy parameter Z Alpha L basically, it is a measure of how much buoyancy is being generated by the windshield and how much buoyancy is being suppressed by the thermal stability.
3:43:44
If enough facts this are available in the in ... Call data, file, we can use, you can use.
3:43:53
When and temperature profile to estimate, which are some number following day question here, and Lancome Pill, against the over L S deputy.
3:44:10
Yeah.
3:44:11
Another thing here is, the vertical turbines profile usually varies with height, zero at a graph, and then increasing to a peak value in, in the middle boundary layer, then decreasing again to the top of the Boundary Layer.
3:44:31
This is the T four, but we can't force the vertical turbines profile to be replaced by a by a Boundary Layer Average Fattier.
3:44:43
So let us do a simulation. You are saying, OK?
3:44:50
The temperature and wind profile to to calculate a step at a T, and then also changed, happened in ... profile.
3:45:02
Uh.
3:45:04
Replace it with, by beta P B, our average, let's do a simulation like this.
3:45:30
OK?
3:45:39
Right here, I did not do one thing.
3:45:43
I did not change.
3:45:50
Athena changed the setting back to the base configuration. That's why you are looking at the result is pretty similar to what I did before. So less.
3:46:04
one thing important to, to remember is to do the reset here for the manual seven and then ... dies out base case.
3:46:17
Then now we can try to look at the computer that you're saying the option off the when and temperature profile for stability.
3:46:31
And then the PBL average tourbillon ... file.
3:46:40
Like say say then ran the motto again.
3:47:20
As you can see, you lay the advance manual number seven.
3:47:25
There are so many options and settings, you can, when you up even all doing some kind of light sensitivity tests.
3:47:34
So, it's very important that you'll remember to reset it back to the default and then change one option at a time.
3:47:56
OK, here, you can see, that the, the result is slight, for this case, we are doing here.
3:48:02
The result is not sensitive to change after step, but it's a option and the vertical mixing profile option, again, full.
3:48:12
The reason for all these configurations, anything up to many, OK.
3:48:25
Ardis options, The reason for them to two hat took to be available here is that we have this capability to create an ensemble using different physics and turbine smet thoughts. Again, that will be discuss in a later section.
3:48:51
See?
3:48:58
Um, as I mentioned earlier, normally, we don't touch the default setting, unless you have a good reason to do it.
3:49:08
What an example I want to give here is, when using a cause, we catch a cold data file late six hourly temporal resolution. A lot of global metric amato provide heat and momentum flux feel every six hours, and then the temperature and we feel in an instantaneous sense.
3:49:33
When, when there are huge differences between the day and night, you may be under sampling.
3:49:41
So your six hour average flux barrios, where does not represent anything practical or valid at the instantaneous time, at reach out of variables output.
3:49:58
So in this case, days disconnect between the facts, turbin, Fierce, and instantaneous.
3:50:07
But when temperature fills, so, that's why you may want to we computed the computer stability and mixing parameters from the wind and temperature profile. But if you have ...
3:50:24
coming from hourly output later, ... hub we use here, Dan is totally fine.
3:50:37
Next section.
3:50:39
We're going to look at the mix, the calculation for the mixed layer death dafa option to, to, to use **** hearing.
3:50:52
The PBL bounder layer depth, one is to use the, um, the value form directly from the ...
3:51:03
model file, me tactical data file, and not away is two computed from the temperature profile.
3:51:12
These two option actually gives you put a similar with salts.
3:51:21
And then, the Nahda two options you can choose here to compare the next lay at that.
3:51:29
one is fun TK profile, the turbulent kinetic energy profile, or from a, um, richest number approach.
3:51:41
Let's look at the turnout of some form, an ... core data.
3:51:48
Let's create, quit, and then go to dimity article.
3:51:55
Man, you know, me to Object menu. Display Data.
3:52:00
Tax profile.
3:52:03
And then select the loft dataset, we use the cap text too.
3:52:09
And the lag was 27 ... file.
3:52:15
And then that Time Offset 12 and then the TAM Increment set 2, 6 hours.
3:52:31
And then let's look at the the profile at the starting location 39.9 see it 39.9 Land lead to N -4 84 **** 22 philander too and then Rempel file.
3:52:59
Then you can see this.
3:53:05
So. Let us go to.
3:53:12
Here, the 18 UTC of September 25th, that is the about the starting time of the simulations.
3:53:23
So, if, OK, let's put it here, And let's bring up the Advanced menu again.
3:53:40
So if you use the arm k.v.l. Death from ...
3:53:46
file, that means you are going to use, find this disparity with the PPL H patio.
3:53:58
And use it directly for the calculation.
3:54:02
If you choose to use the temperature profile to compute a mixed layer that's it's going to find a height.
3:54:12
Reached a potential temperature is two degrees warmer than the surface.
3:54:17
So they sister potential temperature. At the Surface, S is 284. So to decrease warmers. That will be 286, 96, Sorry, 296, so it's about here.
3:54:38
So, that is your PBR.
3:54:40
That if you use TK profile, to compute a mixed layer depth, eat will look at the.
3:54:53
It will looks at the TK, TK, jobst, twice K here, this is TK jobs to a certain percentage, so puppy that it will the model would determine the PBL.
3:55:07
That is around this height at this height.
3:55:12
so far, if you choose richa number, and let's go back.
3:55:21
So if you choose switches, the Number eight use Access if temperature for convective cases that is computed from virtual potential, temperature and horizontal and vertical friction velocity to estimate the PBL tight.
3:55:38
Again, in this particular case, we are we will, we are getting put as similar results with all fuss settings.
3:55:49
So if you run the simulation using this, this difference, settings for fall, the PBL height, you will see pretty similar results, um, we can also force d'amato to use to use a constant next Layer heit, The settings here, the reason you may want to do this, is the same reason you may, you may want to use a cause to win in, caused some when input to compare myself to an analytic solution.
3:56:31
So, let's try to do a simulation by setting the PPO, the mixed layer layer that, with a constant bajau.
3:56:47
Let's do, 1250.
3:56:53
12, 50, and then less safe, save, and run the model.
3:57:27
In the tutorial you, I'll show you, saying, the, The fix next layer, death, 5000 meters, then that result is pretty similar.
3:57:41
Tutor base simulations.
3:58:02
Yeah, something is puppy, not pray away in the setting. So let me look at it again.
3:58:09
Events Manual.
3:58:13
Probably I felt GAD two.
3:58:16
Reset it.
3:58:21
Yes, I was choosing, like, not the base configuration. You can see it here.
3:58:26
That's why I'm getting this message out, not what pain I am expecting. So, less reset.
3:58:37
Go back to the kinda Kaizen, and then sat there.
3:58:44
Next layer, death, to accustom radio, 12 50, 50, and then save then let's run the simulations, OK?
3:59:24
OK.
3:59:36
Interesting.
3:59:43
Let me go back to the manual.
3:59:56
Yeah, puffy I did not crack this button, so, it was not using the constant mix lay at death.
4:00:06
I thought it was. He was saying that.
4:00:09
But, anyway, so, we can look at the tutorial graphic in the tutorial. So, this is the result.
4:00:20
You're saying that 5000 effects', Max like that, and then, this is the result using 12, 50, 1250 constant mix layer. that's so, it can see that light.
4:00:40
Using the 5000 meters, mixed layer that night, the maximum concentration is still kind of like Sim, I mean, the result is similar to the face base case.
4:00:53
Um, basically the same maximum value.
4:00:58
But when we change the mixed layer DAF two, tau 50 so, you can see that, now we are getting, higher, maximum concentration, bateau, 61 thousand PICO gram per cubic meter, which is much higher than da base.
4:01:24
Cause simulation and also higher than the observe aircraft aircraft upset max maximum value. So that is how they are these settings will do too.
4:01:40
um, the results.
4:01:47
So the next one we're going to look at the TK option.
4:01:56
So the TK is defined as does some of the the component variables component variance. So is U prime Square plus B Price get square and then plus W Prime Square Are these components at them together and then divided by two, then you get the total TK. Some I mentioned earlier that some ... dataset provide a T field later wolf output.
4:02:30
What we have been using.
4:02:35
So this this the total turbulence, turbulent, valueless.
4:02:41
So for our purposes, we, we need to partition it between the horizontal and vertical.
4:02:48
That is so-called, and as such, I shall, I shall be racial.
4:02:54
So the settings here, you can see that in the events man who was seven.
4:02:59
The default value is **** 18 for the ratio of vertical prime square and the horizontal your Price square plus V pi square. That is, your correct.
4:03:13
The false button. That is the default approach in the menu.
4:03:17
You can also select none, which set the value to zero, then has to be computed based on the counter casing question.
4:03:29
So here, let's do assimilation.
4:03:34
First, that's me.
4:03:36
Click the reset button ah and then pick a ...
4:03:44
option for vertical turbulence and then.
4:03:54
and then click None here.
4:04:01
And then click Save Sade.
4:04:06
And then random model.
4:04:20
Let's see if I can get her, get the results I am expecting.
4:04:34
OK, so then let's do another run.
4:04:40
Go back to the Events menu.
4:04:55
So we're still using TTE Martha forever called happens And then all these other settings I in the fall then now less falls it faster the Use the force button and then set the better to the default **** 18 Save say.
4:05:28
And Then random model.
4:05:40
OK, let's display it.
4:05:49
OK, you can see that this, I'm close this, this is, This is our face.
4:05:57
Without using ... saw the maximum batteries, Tandy one thousand.
4:06:04
And then this is the result using TK methought and the ratio between horizontal and vertical component of T This one is using Syrah.
4:06:20
Um, so the maximum value is about 39,000.
4:06:27
Then when we use the TK methought with Du Pont 18 ratio, the peak value is higher, now is 56,000 T PICO gram per cubic meter.
4:06:49
So, again, you can see that in this advance, number seven menu, you have many options, um, to choose.
4:07:01
Um, so what are this about?
4:07:06
Yes, all these options.
4:07:10
This is giving us the background information in terms of what happens later in the in the physics and sambo section, which is going to be looking at that concentration uncertainties. The primary importance of this is there are approaches within the high speed to give you different answers. Oh, wait, we don't know which one is the right answer. But by doing multiple simulations with different configuration.
4:07:42
and an average of them are so-called ensemble mean they give you a better answer and in many cases, they where are we doing the concentration simulation.
4:07:55
Right.
4:07:57
In addition to the grid settings, how many particles you want to use, all these things. The physics options here.
4:08:06
We you may need to go through some of them in order to try to figure out which one is better, that's why many situation we need to do multiple simulations.
4:08:20
And again the later section, talking about the ensemble S, is one of the the good good way to two pi U of the rule I what can off physics are settings that you may want to look at or do some noise simulations in order to figure out what is the best configuration for your particular application and study.
4:08:53
Um hmm.
4:08:57
The last part of this section is about this new, kind of new.
4:09:06
Because starting with the release has been released version five. Another, I go to them for dispersion calculation is implemented to high speed. So this is cost deal.
4:09:22
Features still is another variation of high speed out there.
4:09:29
So as I mentioned, starting with this version five is this algorithm is available in high speed.
4:09:40
In addition to that, some features coming fun still, and now I saw that a ball in high speed.
4:09:47
We have look at some of them, like, oh, atty, in this Advanced Manual number seven, the hana vertical turbines option, ah, yeah.
4:10:03
The modify richest the number to compute mixed layer death and, uh, Tom, Dependent, the ...
4:10:12
timescale, all this are features coming from as still and now, they are available in high speed.
4:10:24
I think in the very last last section of this workshop, we will have time to look at more about high speed versus still algorithms. So, bye for now.
4:10:35
Let us go to the advanced manual, number three.
4:10:47
So, you can see that we have two options to choose foot dispersion, occultism the device high speed dyes, what say High Speed has, has been using. And still use the new algorithm.
4:11:03
Now, you can choose.
4:11:06
You can choose that to do the dispersion calculation.
4:11:12
Let's choose that, Save, and let us go back to the Events menu seven and then hit the reset.
4:11:19
I just want to make sure order settings, ah, go back to default, Save.
4:11:29
And OK.
4:11:33
Can now we add using steel, agata, fodder, dispersion calculation save.
4:11:43
Save, and then let's run The model.
4:11:51
So, when choosing the still algorithm is sess up several addition, No computation they are listed here.
4:12:01
Um, it does follow backward calculation to look for consistency in the transport and, and define a finer layer near the top of the Boundary layer to make sure that particle dong ... acid, definitely, as a tantalum mix.
4:12:22
So, the top of the bounder layer, and it also has some finite temporal time steps.
4:12:32
So, your hand, you can see, it takes a lot longer to run this.
4:12:40
Yeah.
4:12:41
But when, I mean, we're going to look at the result, ah, in a second and then you will find that actually it takes longer to run.
4:12:50
But at least for this kept tax application, we don't see the resell, play has some has difference like compare to the base. Or default has fit dispersion algorithms.
4:13:10
one thing we may want to look at later, is they opened a message file.
4:13:15
OK, Les Tried to display, though, itself first.
4:13:25
OK, so this is, this, Michelle you saying, skill, I called him, and this is you're saying the high speed algorithm.
4:13:34
So you can see similar the shape of the problem and also the maximum value of the of the simulation, Maximum concentration is about the same.
4:13:44
So, let's go to look at the message file.
4:13:54
And then you can look for the parameter call.
4:14:00
I D S P.
4:14:01
Equal to, that means, using that Still, I called them, instead of the high speed default eckerson.
4:14:13
So, um, the next thing.
4:14:18
Actually it will be the last thing I want to talk about is that the off Vertical interpolation scheme which is also coming from Steel.
4:14:37
So it is like a bit more complicated by taking into account the great staggering in the loft data.
4:14:47
This option, we cry it, certain time averaged variables that are not in the Kevin Wolf data file, we may use. Let me bring up the sub menu.
4:15:01
So, we are using the Wolfe 27 UW.
4:15:06
But, the time average fibers are not there.
4:15:09
So, we need to to pay kanada whilst data file.
4:15:20
So, let us do this.
4:15:22
So, this 1 whilst 27 M C data file the M C first stands for mass Couples.
4:15:32
So, that is this file has the time average variable that is needed for using.
4:15:42
This option this vertical off vertical interpolation scheme.
4:15:47
So, let's save it and then go to the Events Manual None three.
4:15:57
So, make sure we pick Steel and then Manion number 13.
4:16:06
So pick the wulff scheme.
4:16:12
then let's go to the number seven again the menu.
4:16:17
Make sure we have Arm the default setting.
4:16:24
Save Sade and then run model.
4:16:44
So again, if you want to double check, if you are using what you want to use, they, as I said, set them in the in the manual. You need to go to the message file then. Look for parameter I D S P. Equal to, that means using still algorithm and then the wolf vertical integration scheme.
4:17:08
The parameter is W R T equal true.
4:17:13
That means it is using the ... interpolation Scheme. accion now that you set default.
4:17:20
So whenever you do it whenever you are using Oh.
4:17:27
Input ...
4:17:28
data file in which the time average value of this variable and you all you are also using still then you will be just using this wolf vertical interpolation scheme.
4:17:57
And more about Dell configuration will be discussed in the, in the very last section of this was sharp, I think is in Section 17.
4:18:11
So, you will learn more about distill features later.
4:18:23
OK, so, they sister when using the time image variable, mythological data and still agoa some and then also turn on or off vertical inter pleasant interpolation scapes.
4:18:39
So, compared to this one basic default setting, again, very similar result.
4:18:50
Also, if since actually the same maximum concentration, Valeo.
4:18:58
Let's see.
4:19:04
Yeah, I think that's, that's it.
4:19:08
The exercise for this, the, for this section, yeah, you can just follow them and then go back and try to do then, that solutions are so ... here on this page. As such.
4:19:23
It's a good exercise to do in order to learn more about that these different settings and options fall running hudspeth.
4:19:34
Um, then I will conclude this section, and turn it this back to Mark.
4:19:43
OK, thank you very much, Pantene, um, thank you very much, and we'll now go into A Let me change the presenter back to me, Yeah, and hmm, show, Application will now go into a break, and we'll come back. We'll get back on schedule.
4:20:04
We'll come back in about 25 minutes at 15 minutes past the hour would be 315 Eastern time, or 17, 15 UTC. So, we'll see you in about 25 minutes, for the last of sections of today's.
4:20:29
OK, we are We are back now, and we're about to go into the next section of the of the workshop, which is going to be looking at different ways to display, um, some of the results set that you're getting and Sunny Zit, doctor ..., will be presenting this section.
4:20:58
So Sunny, I am going to change the presenter.
4:21:04
over to you.
4:21:07
Yeah.
4:21:11
Turn off my webcam.
4:21:18
Great and Sunny.
4:21:19
We can see the screen you want us to see, I think. So I think you're all set. Take it away.
4:21:24
Thank you, Mark. Hi everyone.
4:21:27
Here, during this section, we will be talking about how to produce graphics using, um, different options.
4:21:42
OK, so let me quickly, so you have Section 10.1 open.
4:21:52
OK, First, Concentration Batch.
4:21:56
Let me open up Folder Or did it go navigate to Working Directory.
4:22:06
Let's open up the console plot dot bat.
4:22:10
If you are running Linux or Mac Taj, it's going to be named as ... plot dot SH shell script.
4:22:17
Say if we open it up using a notepad, what it does is it shows the yeah when the last time when you run the kind of plot concentration plot, it shows all the options that were used to produce the plot.
4:22:37
So, um, if you have your terminal window, Windows terminal opened, and you are in the Working High Speed working directory, you can execute it from the terminal and reproduced the plot.
4:23:00
So, I suppose this was what to you did some time ago, like yesterday.
4:23:07
So, what we are, what I'm trying to say is Dart.
4:23:11
After you run the concentration and contour just did a set up first.
4:23:21
Suppose I did that.
4:23:23
Um, after you create a plot using X, could display, you can custom tail.
4:23:33
I mean, you can enhance your thought by changing one of these options or more options.
4:23:42
So.
4:23:44
If you just execute the Khans plot without any argument, it will show you all the options that I support.
4:23:55
So there are about, you know, roughly 30 or 40 options.
4:24:00
However, only half of them are supported by the UI, so using terminal, or a script, or from command line, you have a final control of what you're going to plot, or how you apply what look like. So for example.
4:24:19
And another important thing is that, at the very end, you are seeing dash colon.
4:24:25
A dash column means that's the end of our command line argument, whatever appears after it, there are ignored by ....
4:24:33
So, OK, I know that the, um, the dash K option is for coloring.
4:24:40
So if you see here, zero is black and white, one is a color, two is no lines of color, for three is no lights, red and white.
4:24:50
So let's change it to zero, save, and rerun the batch file.
4:25:01
And let's refresh the web browser.
4:25:05
Now you see the color, is it gone as in black and white or grayscale, plot? So, that's how we'll use, um.
4:25:13
OK, there's, um, cons plot, duck, batch file.
4:25:20
If you used Python Graphics, it would drop P Y hook Python hook that bat or PWA hook dot SH depending on your operating system. We could take a look at that to various modes.
4:25:38
OK, So Hi left. OK, here we go.
4:25:46
Um, the Python version takes all the arguments that T M ... plot supports.
4:25:53
On top of that, it has more options. For example here, dash dash interactive.
4:26:00
Meaning, you can, you know, can move around the map or plot or zoom in, zoom out interactively and also support different types of street map.
4:26:13
Well, we'll cover this later today, so I won't go into detail, but the same story here Like the ...
4:26:22
you can modify options given here and produce a different I mean uh, Have a finer control of how your plot generate your plot, OK, OK, let's click that away and let's move on to Next Section 10.2, This is where first less As the Yeah.
4:26:48
Tutorial says, let's retrieve the Previous.
4:26:58
Control file, that's a tutorial.
4:27:02
Tutorial I believe it's under Files.
4:27:07
Conch continent case control.
4:27:13
OK, save it, and oh, so, Setup retrieved.
4:27:25
Uh.
4:27:30
Conk case setup, OK, and Save.
4:27:36
and before running the model mix, making sure.
4:27:43
11 hours, yes.
4:27:46
And sampling start this under of this and pollutant definition.
4:27:53
Grit.
4:27:55
three Oh, 9 2 6, OK, that's fine.
4:27:58
And grid resolution is point oh five, OK, we just verify that.
4:28:06
All right.
4:28:08
Good, let's start running Run Using Setup file, because we just save that.
4:28:15
All right, this will take a few minutes.
4:28:20
Clean up.
4:28:26
Just a little more time.
4:28:37
OK, why don't I open up another?
4:28:45
Explorer, Find the buyer that they're going to use, which is a shape file. That the shape file, shape files dot TXT.
4:28:58
I don't see it here.
4:29:02
Mimic kept tax.
4:29:06
OK, Let me try to find where I am.
4:29:15
OK, that's done. Display Contours.
4:29:22
Producer plot?
4:29:25
OK, so this is what I'm supposed to look at now.
4:29:29
Now, let's go back to the tutorial. We'll need a shape file.
4:29:37
OK, why don't we create one here.
4:29:43
Name a file is a shape file.
4:29:48
And I know that yeah, we have these files.
4:30:01
High split, let's verify that.
4:30:05
Let's go up one directory component.
4:30:10
Oh, no, I don't have map there.
4:30:13
I suppose his graphics right.
4:30:15
Yes, Sorry.
4:30:18
Graphics.
4:30:23
Graphics.
4:30:28
Correct X, Yeah.
4:30:31
The shape file is a text file. Each line represents a boundary that will be drawn on your ... plot.
4:30:40
The first is a name, name of the shape file, followed by a number, five numbers.
4:30:46
The first number is how do you select a zero means.
4:30:51
A solid line, a non-zero number is going to be a dashed line and the value here specify the the gap between you know dashes or dots, then line width and followed by three numbers, red RGB components running from 0 to 1.
4:31:11
So, if you have 1, 0 0, that means our only you're specifying are non-zero components.
4:31:20
So the line where the drone in rat completing red no blue, or green component, OK, so we have that ready, and let's go back, remember to remember.
4:31:34
So this is what the plot look like without Specifying, uh.
4:31:46
The shape file background can less change that, too.
4:31:52
Sze file, shape file dot, uh, T X T, and rewrite it.
4:32:04
Oh!
4:32:10
I think this because.
4:32:18
Shape file definition, isn't that? Correct.
4:32:27
Correct. Thanks.
4:32:34
Oh, sorry, I'm.
4:32:39
That wasn't correct.
4:32:44
S: Give it a try.
4:32:49
No, no. That's correct.
4:33:03
Wow.
4:33:09
OK, I'm having some trouble here.
4:33:13
Can't tell what I am.
4:33:15
Not doing quite right.
4:33:18
Hmm.
4:33:25
I send the tutorial maps.
4:33:29
Tutorial maps.
4:33:36
OK, let's give them a try.
4:33:41
Marking.
4:33:50
OK.
4:33:55
Shape Files, Execute.
4:34:03
Ah, now we get it OK.
4:34:07
So, as we expected, say the state boundaries are drawn thicker line.
4:34:18
And then, you see in below county lines, and you see also rows as well.
4:34:27
So that's how you can change plot background, using Shape Files, OK, good.
4:34:36
So next would be I already explained the format of the shape file.
4:34:43
and now.
4:34:47
Up a contour replace background with the datum.
4:34:51
OK, let's Add Data Case.
4:34:55
So, that means, under the concentration display window, at the very bottom right, above the zoom option, there's a data file.
4:35:08
So, you can specify the data case. Where did it go?
4:35:16
Demo Tutorial.
4:35:19
I believe it's under cap tax data cave's, Data Case Data Case, to match that option, we know that.
4:35:30
The previous example, we know we want to plot this and pick a gram.
4:35:36
So, Multiply.
4:35:38
Because the emission was in N gram unit, to display the PICO gram, We're going to put 1.
4:35:47
1 to the 12th, multiplier, and also.
4:35:57
The contour will choose users set and bold use 50,000, 20,000, 10000, all the way down to 500 PICO grams.
4:36:08
And based on that, and let's execute it again.
4:36:17
Appeared, the other.
4:36:22
Yeah, OK.
4:36:24
So, this is what you expect to see, the datum plot file, the measurements file from the measurement data, they are drawn in black crosses.
4:36:37
And Contour is now in pickup ram. OK, so, good.
4:36:42
We have covered Section 10.2.
4:36:49
And now, let's move to Next.
4:36:54
Section, and which is Enhanced.
4:37:03
Graphic label.
4:37:04
So, um.
4:37:09
High split.
4:37:10
There are different ways of configuring or customizing text appearing on your plot, The, uh, um, plot Title.
4:37:23
And, also, you can use Map Box to put text down below here.
4:37:34
Below the plot, so, let us do that.
4:37:37
In order to do that, let's go to go to Advanced file, edit.
4:37:43
And panel labels, I suppose.
4:37:47
And here, you can use enter. Various information, for example, you know, this is a Dayton. Ohio.
4:37:59
Let's bring up the juice yet and Times 700 metrology walk, 27 kilometer, pollutant emission rate one kilogram over three hours.
4:38:21
This was 10 meter lilies.
4:38:26
And none start time September 83.
4:38:35
So by saving this what happens is that you are generating map text CF tiff file in the working directory.
4:38:48
And now go back to that concentration plot menu.
4:38:53
And without you don't need to specify where the map text file is located, it will be read if it is there in the working directory.
4:39:02
It will be read automatically via ..., so let's execute, display it again.
4:39:10
Now, if you scroll down, aye, text box appears on the plot now. What custom text that we entered?
4:39:24
And, um, beside Map Text, CFHA, ...
4:39:31
plot uses labels, thus the F J, throw that you can change the plot title, the units, et cetera.
4:39:43
We won't be doing that here, though.
4:39:46
OK, let's move on. Next section is creating KML, or K M Z for Google Earth.
4:39:54
Um?
4:39:57
Well, they use the same window with a concentration display.
4:40:02
The only change we're going to make is clicking the one button for the JS output under the Mapping 1 to 1.33 role.
4:40:13
So we'll choose KML and K, M, Z and Execute.
4:40:23
So, then, it says, The Gloria said KML, a KML file created in the working directory, and his name is high split conch dot K, M Z OK, that's good to know, And let's try to find that, where that file is, OK. Here we go. There it is.
4:40:43
A KML file is nothing but a zip file, if you change the file extension to CIP on Windows, and you can actually look into what it has inside, say, it has 1, 2, 3, 4 graphics file. The first one appears there, G label this ASTM legend.
4:41:10
Uh, this is too small to say, basically the M, the legend, the control levels.
4:41:21
The KML file, this is A Google earth specific File within an XML format. You can open it with a text editor.
4:41:34
It specifies what the contour look like. I mean, the last line, co-ordinates of each contour, and the value, etcetera.
4:41:46
So let's get go back to Working Directory, and restore the extension name to K M Z.
4:41:55
And on my laptop, I have Google Earth installed.
4:41:59
So if I double click it, it will start Google Earth.
4:42:06
And we'll open the file, and we'll show the content.
4:42:12
OK, Loading, zooming into Ohio area and here is, Um, this Red Cross is the release point or Source location.
4:42:26
And this is a contour just like them.
4:42:32
Yeah, OK.
4:42:35
Good.
4:42:38
Next would be Let's close this up.
4:42:46
Will be producing additional output JS output, as as really generates Format, instead of the Google Earth format.
4:42:59
Let's select that and that option as regenerate an option and execute display.
4:43:09
OK, um, what happens to us then? It will.
4:43:20
Create two files: GIS on those scores 00800 HTML, 0 1 ATT and dot T X T The both of these are text files, listing the points of H Contour contour and that ... file as an attribute attribute file.
4:43:44
Hmm?
4:43:45
Specify the color level and time and name concentration.
4:43:51
So let's click them away.
4:43:54
Um, in order to these two files are not readily usable by GIS, for example, arc GIS, you will need to convert this into shape file.
4:44:10
So to do that, let's go to Concentrate Menu, and Utilities, and there's a GIS to shape file.
4:44:21
Let's bring that up.
4:44:23
And here is expecting to enter the name of Generate File, the GIS file, a text file, OK?
4:44:33
Let's click that, and we need to go back to High split Working Directory, where that file is located OK, This is one.
4:44:47
Let's open it.
4:44:49
Um, we know some polygons and is from a concentration plus that we know we collect concentration and the output shape file based name couldn't change it, change this to well, something else, but let's stick with the default, OK?
4:45:12
Let's click Process.
4:45:16
Sze file created, OK?
4:45:21
Um, OK, let's refresh.
4:45:26
There are four files produced by the process.
4:45:30
Converting GIS to shape file, that's dot ... dot ..., J, the SHP, and that's H X, so less.
4:45:48
Let me open a new window.
4:45:52
Um, I have the arc GIS program on my laptop and I place, usually, the data files under a certain directory.
4:46:03
So let's copy this or files that were generated generated by the GIS two shape File Conversion menu and open.
4:46:21
Our map, and let's click cons plot dot SH, SH, PHA file, and drag and drop.
4:46:30
aren't it?
4:46:33
All right, mmm hmm, OK, this is taking time.
4:46:38
Let's zoom to layer.
4:46:46
OK?
4:46:49
So.
4:46:55
Right, yeah, yeah, So, OK, that's good.
4:46:59
This way, you can import a shape file produced by high split into your favorite GIS system, and look into that way, OK, so this.
4:47:13
good, or let me just delete.
4:47:22
It's, yeah.
4:47:27
All right. Next, what pay?
4:47:31
Section 10.6, talks about shape file overlays?
4:47:37
Um, because you can, since you can display whatever geometry, or drawing, and shape file as a background for ... plot.
4:47:51
Um, if you have trajectory plot, and if you have shape file flutter trajectory, then you can simply impose that trajectory as a background on your concentration plot, or the other way around, too.
4:48:09
So, let's first.
4:48:14
Like that tutorial, suggest this set up a trajectory run.
4:48:20
Wow, forward, trajectory.
4:48:26
Philes, what's the name of it?
4:48:30
Trajectory.
4:48:33
Board control, OK.
4:48:39
Save, and we don't need, there's no need to have setup.
4:48:49
So, we will delete and run that trajectory run.
4:48:57
Oh, that's pretty long.
4:48:59
What did we, ah, OK. We weren't supposed to, according to the tutorial, we're supposed to run on the first 11 hours or less, revise it, and rerun OK.
4:49:12
And rerun the model, OK, that was quick.
4:49:16
Now, go to display trajectory.
4:49:24
We'll need to change the background to shape files.
4:49:29
That's under our Working directory.
4:49:33
Where did it go?
4:49:35
OK, shape files.
4:49:39
And, um.
4:49:45
Let's plot it.
4:49:50
OK.
4:49:54
So, we have that, now. next, we'll modify.
4:50:01
Uh, our shape file is the background file.
4:50:07
Let's make a copy.
4:50:11
And member, we have.
4:50:17
Um, we have mmm hmm.
4:50:26
OK, let's try to change Des cons plot the shape, Um.
4:50:37
Let's rerun.
4:50:42
Uh.
4:50:46
Let's rerun execute.
4:50:50
OK, so if I bring the plot back here.
4:51:02
Um, what I'm showing, what I'm seeing, is that there's a trajectory plot, where the map background now shows that State boundaries, and it was a county boundaries, and the contour plot, which is kind of hard to see here, So less increased line widths, and use something.
4:51:28
What color Let's use black for the moment.
4:51:33
So, 0 0 0 does black.
4:51:37
Let's rerun.
4:51:40
No.
4:51:41
OK, so now, as expected the first line, the conch plus ...
4:51:48
file, is now trone as map background. So this way.
4:51:57
The shape file from a ...
4:52:00
plot is drawn as a background, or a trajectory plot, OK.
4:52:06
All right.
4:52:08
Next is, um, Or, um, if we go back to concentration plot.
4:52:23
Um, I'm sorry.
4:52:26
Probably we need to do.
4:52:32
GIS lines Proteus.
4:52:40
Sze files right, those two.
4:52:45
And let's convert GIS to shape file this time.
4:52:54
What we're trying to do is the other way around. We want to plot the trajectory as a map background on a concentration plot.
4:53:03
So, it's the opposite, well, of what we've just done.
4:53:06
Go to HYSPLIT.
4:53:09
Mayor, working, and this is, yeah, file to be converted.
4:53:19
Conversion renaud trajectories are the lines that we select lines, and it is a trajectory shaped by our last name, treasure plot based name. That should be fine.
4:53:32
Click it, OK, it's done.
4:53:35
so, you see 1, 2, 3, 4 files are generated through the conversion, going back to trajectory and concentration display, concentration, contours, and, yeah, we'll need to modify the shape file again, only this time treasure plot instead of cause plot.
4:54:07
Uh, OK.
4:54:11
Let's hit to exec.
4:54:24
Is it because we are two?
4:54:30
OK, so we see the plume and trajectory has a map background.
4:54:40
I'm not sure why we're not seeing the.
4:54:46
Uh, Colony and State, nest.
4:54:55
Try. Once again.
4:55:00
Um.
4:55:08
Yeah.
4:55:12
Maybe replace trajectory, shape File.
4:55:17
Last.
4:55:20
Yeah, OK.
4:55:21
It's probably because if we play some trajectory shaped by first desktop determinist a plot, or affects the plotting plot extent extent? So we were, too, close to it, and now you say it is.
4:55:41
Uh, the plot is as expected with the county boundaries and state boundaries and also the trajectory.
4:55:49
OK, all right, now, the last section, 10.7, Python displays.
4:56:00
Ah, one thing to check is advanced configuration setup and set directories.
4:56:10
And if you setup Python according to the installation, um, OK.
4:56:21
Instructions, oh, by the way, you're a high split directory.
4:56:28
There is M subject subdirectory called Python. If it's really down, there are a number of files there.
4:56:36
And for Windows Users, uh, if you want to install Python, you can open up this install underscore when 10 ticks J, S a text file.
4:56:48
And follow the instructions there to install Anaconda three on your system and also a high spread Python graphics after that is done you wanted to check the F the last, the Python Anaconda three environment a path is correctly pointing to your high spirit environment F test set, then, it will work.
4:57:18
So, let's go back to concentration plot.
4:57:27
Contours.
4:57:31
Here, instead of choosing Fortran now is Python.
4:57:42
Straight map I'd like to use terrain.
4:57:47
And let's choose an debug messages checkbox.
4:57:53
When you run generating Python graphics, the first time the first plot takes longer than the subsequent plots.
4:58:06
I believe it has to do with them loading all the required member dynamic libraries into the computer memory.
4:58:15
And once it's loaded, second time would go a lot faster.
4:58:21
So it is working.
4:58:27
And here, OK. Good.
4:58:31
So.
4:58:34
They, Python graphics produces almost.
4:58:40
Well, the format of the plot is the same.
4:58:43
The only different thing is that The map area, the background, is now, it's based on a street map.
4:58:55
one thing nice thing about this is, you can click the magnifying glass and you can zoom in into certain area.
4:59:10
OK, it's just like this.
4:59:13
Or it could choose the pen menu and move it around the map.
4:59:27
Alright, so that's good.
4:59:29
So less.
4:59:33
You can click the X button on top to closest graphics or you can type choo to close the interactive display.
4:59:45
So, let's exit that.
4:59:48
All right.
4:59:51
By doing so.
4:59:55
If we take a look at the working directory.
5:00:01
Pyre hook that batch, a window for McIntosh and Linux users.
5:00:07
This file is named as P Y H O O K dot SH.
5:00:13
Less open that using a text editor.
5:00:22
So there's more information here.
5:00:27
Um.
5:00:30
So just like the ...
5:00:32
bat that I talked about in the beginning of this session, it does likewise.
5:00:38
So at first, activate the High Split and icon the environment and runs Python.
5:00:47
And the script name is ... dot P Y, then using the options given here, the follows, OK.
5:00:57
And if you want to take a look at the additional options that the Python version ProVis, you want it to open up a Python and our conduct enabled Terminal.
5:01:17
So I am opening up that, Anakin, the 3 6 12 Bit.
5:01:23
And a conduct prompt, OK?
5:01:28
So that stare and first is, but I freshly open AM, and I called the prompt.
5:01:38
Uh, It says, I'm base, right? That means I'm not using the high split environment environment, So I need to activate, activate high split environment by running condor.
5:01:51
Activate high split.
5:01:54
All right.
5:01:55
Now, let's drill down to the working directory, and we know that.
5:02:05
Uh, The script S, cars, plot dot P Y.
5:02:11
So, just like Cons plot, the Executable, the Python script, if you do not use any options, it will print all the command line options that is supports.
5:02:25
So, starting from there all the way, 2 dash nine These are all the same options that the conch plus supports.
5:02:37
In addition to them, there are 1, 2, 3, 4, 5, 6, 7 additional options, the Python version ProVis.
5:02:46
They're all use a double dashes to distinguish them from the Cons plot.
5:02:54
The executable cause plot options.
5:02:57
So debugging, whether to print debug messages, or, no, to use different types of street map. You can use them, too.
5:03:06
OK.
5:03:07
And another nice thing about the Python version is that it can output not only in one format, but also multiple formats.
5:03:16
For example, if you specify dash dash and more dash, Form S is equal to J P J, comma, P D, F, comma PNG, without any space between them.
5:03:28
At one go, the same plot will be produced in three different formats, OK?
5:03:38
I believe that, close this session, and their exercise for this section to, I know what you need to do with them. Now, I will turn this over to Mark. Hi, Mark.
5:03:56
OK, hi, Sunny, thank you very much.
5:03:59
Um, and I will take back over control, um, of the present presentation.
5:04:10
Yes.
5:04:14
Second, OK, thank you.
5:04:18
See.
5:04:23
Give me one second here.
5:04:27
two.
5:04:30
Yep, my screens are organized.
5:04:34
one more second.
5:04:37
Sure.
5:04:40
two, OK, let's see if that is working, OK.
5:04:46
So we have, come to the end of today, we, we were going to consider starting.
5:04:54
The pollutant deposition and Transformation section, I'm having some computer problems instance, I was going to be doing, that section, I think, would be better to, need to resolve those this evening, and then be able to start fresh tomorrow, and do that better for, for me and better for you, I think.
5:05:14
So, I think we'll just go ahead and, and give a short wrap up today, then, and, and start fresh again tomorrow.
5:05:24
So, just give a quick summary then, of what we were doing today.
5:05:32
As we've said, many times, we're posting things on the workshop webpage and we're updating those each day with new files and links to videos, things like that.
5:05:48
Here's just a picture of that, and it actually is it.
5:05:51
So, we've updated it since this was generated, but you can see the links to the, um, introduction files, and the videos, and things like that, and also these wrap up files, things, things like that.
5:06:05
Um, so today, we have done a lot of stuff with the air concentrations, which is, you know, a little bit different than that trajectories.
5:06:15
And tomorrow, we'll get into some more details about the pollutant transformations deposition, and concentration uncertainty, source attribution, and getting into some applications with smoking, industrial arms.
5:06:30
Um, As we have been asking you, and you can ask, you know, sort of general questions in the GoToWebinar question box, and we're hoping you'll X, ask most of the more detailed questions in the High split forum.
5:06:48
As we've mentioned, if you, you could look at the answers, questions, and answers in the forum without registering.
5:06:53
But if you want to ask a question, You need to register, that's for early, easy. It's free, and it's fairly quick, and that lets you do this, going forward, in the future, as well.
5:07:06
So here's just a quick recap of logistics. Let's, I think I've said all these these things already.
5:07:14
If you get, you know, too far behind, it's good to watch this demonstrations, that happened to me today, actually, I was trying to follow along, and got a little behind, and it's hard to catch up.
5:07:23
I know, been doing this a long time, and it was hard for me to catch up at some point during this afternoon's section.
5:07:32
Um, let's see. So, just to summarize a little bit of some of the concepts that we're doing.
5:07:39
We started the first day of the trajectories, and you can think of that as the central line of a plume.
5:07:47
Um, and, the first day, we did all sorts of different ways to do trajectories in the second day, here we've been doing.
5:07:56
These concentration dispersion are plumed calculations, and, um, when we think of the computational points, they don't follow necessarily the central line, up the trajectory, but they they move around with sort of a random component due to the turbulence and, um, what we do To simulate a plume is released many particles at a time.
5:08:29
Each one then follows its own turbulent path, and so you end up with sort of a cloud of particles, and here I'm just showing an example.
5:08:37
It's just six particles.
5:08:38
But, as you know, we were releasing hundreds thousands, even as many as 50,000 particles, you know, degenerate.
5:08:46
A realistic plume.
5:08:47
But anyways, it, here, I'm just showing you that as things sort of spreading downwind with these particles, that's hombres regenerating, this disperse.
5:08:57
Um, Um, see, we haven't really talked about this very much, but I just wanted to just mention that these computational points, we keep calling them particles. But really they're computational points.
5:09:12
And, like I said, I think at the beginning of the day, it could be, you know, gas phase molecules could be atmospheric particulate.
5:09:20
And the actual amount of the pollutant associated with any point is actually determined by the amount that you remitting, divide it by the number of particles that you're releasing, and you set the.
5:09:34
So, um, just to give an example, suppose we are looking at nitrogen dioxide emissions from power plant.
5:09:41
And suppose we know the power plants eating one thousand pounds per hour and suppose P two assimilation simulation, it releases 500 of these computational points per hour, You can calculate, I'll show it on the next page, how you do the calculation, but there's no 10 to the 25th nitrogen dioxide molecules per heisler computational point particle.
5:10:05
So, no, we're not simulating.
5:10:09
Um, every single molecule in, or every single particle, we're, we're aggregating these things into these chunks of these air parcels.
5:10:18
So, if you released 5000 particles, then it would still be, you know, a huge number molecules per for particle, and I have shown on the next page, how you would do that calculation. I think you can look at this later, if you want.
5:10:34
I've even tried to include the units so you can see how things work out.
5:10:39
And here's another example with particulate emissions, and, again, I'm, I'm showing that, you know, suppose you have a wildfire, a new meeting, thousand pounds of particulate matter per hour, and we've released 500 computational points per hour. Again, you're simulating.
5:11:01
You're representing a very large number of smoke particles per heisler computational point particle.
5:11:09
And here's the, the compilation details for, or that, if you'd like to look at later, we'll post this file on our workshop webpage very shortly after that session ends today.
5:11:21
Um, and so, um, what we talked about today, then, is that you need to have a, when you do your simulations, you need to have a concentration grid defined, and these concentration grids.
5:11:37
They don't actually affect the calculation, but they do affect what you get out of the calculation.
5:11:44
So, depending on what you specify for your grid, if you don't specify it in the right place, you might not get anything, you specify it, too low or too high. You might, you know, get something unexpected.
5:11:56
If it's too fine or to course, you might get things that you don't don't totally expect, um, but the the concentration in each grid cell um is going to be equals to the number of these computational point particles divided by times the mass of the pollutant on each particle divided by the volume of the grid.
5:12:19
So, we talked a lot about how you, if you change the resolution of this concentration grid, you can get different answers.
5:12:28
And if you change the number of of particles you're doing, you can get different answers, but we also talked about how when you have enough particles, then you will stop getting different answers when you increase the number. So at some point, you will level off, and it won't change anymore.
5:12:45
And that's a way, too, to know that you've got enough, is when you keep increasing them, in essence, and the results stop changing. You know, you've got enough, and you might be able to back off a little bit to get down to the point where it was.
5:12:59
No, just stopped changing. You don't have to have overkill necessarily.
5:13:03
And again, the concentration grids are completely independent of the net data.
5:13:08
The data grids that data grids are driving the whole simulation.
5:13:13
And, you know, you have to have those, you could do a simulation without a concentration. Great.
5:13:18
You just wouldn't get any output output, but the simulation would go along just just the same, same way as it would have with the concentration grid.
5:13:28
And you can have, I think, Chris talked about, you can have more than one concentration grid, and often you do this, you can have fine grids.
5:13:35
That might be near the source and, of course, grids further away. Here, I put a grid, it looks like it's actually kind of in the wrong place.
5:13:43
And so there wouldn't be necessarily any concentration on this grid.
5:13:48
It might be all all zeros, although later on, if the wind changed direction and went over to here, and then maybe we could get something.
5:13:58
Um, I think that's, that's it for today then.
5:14:02
Um, and so tomorrow, like I said, we'll be, we'll do the pollutant transformation, a deposition.
5:14:09
Hopefully my computer will be working better, and then we'll get into some discussion of uncertainty and source attribution and getting into the, some of the applications with certain kinds of sources.
5:14:22
And let's start with the Wildfire Smoke and then de storms.
5:14:26
I think that is it.
5:14:28
I put some extra slides in here, that's some additional information but I think I'll stop there for now.
5:14:36
And so with that, I'll say good evening or good afternoon or maybe good morning if you're unlucky enough to be in the morning where you are and we'll see you tomorrow.
5:14:49
Were the the third day of the workshop.
5:14:52
So bye for now.