0:07
Good morning, and welcome to day four of the 2022.
0:17
Annual HYSPLIT workshop, and, um, thank you all for joining us, and thank you all for being here for these four very interesting days.
0:27
This will be the last day, I'm going to give a short introduction and then we'll get on to the content of the workshop for today.
0:37
Think I'll turn off my webcam to give more screen space.
0:42
Yeah, and.
0:45
Um, What I would like to mention here at the beginning, again, is that we do have the workshop webpage and we're putting more and more things on there.
0:57
Um, on the web page section for yesterday, we tried to put a lot of the slides, extra slides that people had shown, and of course, the video, as well, that usually takes a couple of hours, at least, to be, to be processed.
1:18
Today, we will be going through a section on radioactive pollutants and dose and how you model that with HYSPLIT.
1:28
And then another section on, it says volcanic eruptions, and it will conclude that.
1:33
But it will also include information about EMITIMES files, and some of the other special features of of HYSPLIT with with regard to emissions and also particle sizes will get into some, um, some special applications, some custom applications that will be interesting just to help most of you.
1:55
And, then, we'll have a section on, uh, questions, And you can, you'll be able to ask questions, And, we'll have a panel of all the instructors, We'll do our best to answer, answer them. You'll be answering, asking the questions in the question box, and we'll try to get to every single one of them.
2:12
If we can, if we can't, we'll try to follow up offline with you.
2:19
So, today, as I said, first, the radioactive then, the volcanic eruptions plus emit times and things custom and then the question and answer Period.
2:33
Um, as we've said, we're asking you to ask most of the technical questions, if you can, in the, in the High Split Forum. Which, which you can, you can look at the forum without registering and see the answers, but if you want to ask a question. You do have to register just takes a minute.
2:50
Um, people have been asking questions in the forum, and we've been giving answers, and, um, it's really excellent questions, and we really appreciate all of that.
3:00
And, the nice thing about this forum, is that, you know, let me see, other people can ask, can see, see the answers, and see the questions.
3:08
Also, it's a much better platform for asking things.
3:11
You can include attachments, you can include screenshots, you can do things like like that. It could be a better interaction with us.
3:21
We can see what you're doing, and we can give you better advice about what might be going wrong.
3:27
Um, here's a quick recap of logistics.
3:31
I think I've mentioned all of these things already, this morning.
3:37
I wanted to mention, just to say a few words about, about Scripting.
3:43
Um, so we're not really covering scripting in the workshop, but we recognize that.
3:50
It's a very important topic and it's, some of you are already probably doing scripting with high split, and some of you are probably thinking about doing scripting with high split.
3:59
Um.
4:01
the graphical user interface that we're using.
4:04
Um, it's very powerful, and it's a very good teaching tool.
4:10
It shows you how to use I split, and, and it helps you figure out the kinds of things that you can do, but you can see that, you know, even even for us, you know, if you forget to enter, you know, something on one menu.
4:24
Or if you forget to load at something, then you might not get exactly much what you want.
4:31
And, um, I will just note that, you know, all of the programs in high split can be run from the command line. Or, you know, that means that can be run from from scripts. Because scripts are basically just a bunch of commands that you're issuing, that you've written down for the computer, as if you were typing them from from the command line.
4:52
And, in fact, there's a lot more features that are generally available, um, front the high split programs, then that we, then we programmed into the graphical user interface.
5:06
Partly because, you know, that can only be so many things, aren't on a given menu page. Before, it just becomes just overwhelming.
5:16
And some of, the, some of the, options are a little bit less commonly used.
5:22
Um, so you know, the advantage of using the graphical user interface like we're doing here is you can clearly see the context of what entries you're making.
5:31
But, as I said, no, even for us. You can see it's possible to forget to do something.
5:37
And the advantage of using a script is that, it, it, it might take awhile, you know, to get it working.
5:44
Especially if you're just beginning to learn how to do those scripts.
5:47
But once you do get it working, it works. No, every time.
5:52
And you don't have to keep remembering, you know, to do each thing in the script.
5:57
Some of the scripts sit you'll, that we've written, the net, you're right, You know, have hundreds of commands in them, and you it would be difficult to remember to do every single thing, every single time.
6:08
You know, the other thing about scripting is that, once you've got something to work, and you want to make a little change, you can just go in and make that little change. You know, if you want to change the latitude and longitude or change the runtime or something like that, you simply go in and edit the script.
6:22
And I'll say, and this is not, I'm not an excuse at all.
6:28
The people that have been teaching the course, myself included, we don't usually use the graphical user interface in our day-to-day work.
6:36
We use scripting a lot, and so part of what, yeah, you're seeing a little bit is that, um, we're a little bit out of practice ourselves. And using the graphical user interface.
6:46
However, however, I will say that when when we try to do something new, like doing a clustering, analysis of a frequency analysis, or some kind of source attribution analysis, we do go to the GUI, and we see, OK, here's this is how it's done.
7:01
These are the kind of results that we would see.
7:04
And then we then try to recreate that and then and then customize it in our own own script. So again, not not, not an excuse.
7:15
That's part of what you're seeing in this workshop.
7:21
Um, So as I said, a lot of you are probably familiar with scripts and use them in your own work, although maybe not with high split yet.
7:28
The graphical user interface is actually brilliant big script. It's written in this TCL TK language.
7:34
and it has this additional point and click functionality, where you're filling in forms.
7:38
and then the forms that you're filling in, and then writing the files, and doing that varies operations that you intend to do. So, there have been many other scripting languages.
7:53
It's the old versions of the dos batch scripts that work on Windows machines.
7:58
It's still work on Windows machines.
8:00
There's Linux shell scripts, a number of different kinds of Python, which is a scripting language.
8:06
But also you can do a lot of calculations as well, or which is sort of similar to Python in that way that you can do scripting as well as calculations in it.
8:15
Any others?
8:17
I think you've heard us mentioned, there are scripting examples that we give on each page of the tutorial.
8:24
The little gear.
8:26
Uh.
8:28
But an icon will show you an example of a dos batch script. And the little Penguin icon will give you an example of authentic shell script. And, again, you may be scripting in other languages, and that's fine. You'll, you'll just see the examples, and you'll see the kind of, syntax and the kinds of things that you could do.
8:48
Um, so, not to go too much more into this, but basically, the high script, high splits scripts, generally follow this type of structure, generally define the parameters for the simulation, things like this starting time, and then that data.
9:07
The grid's size isn't everything else that you would want to define, sort of at the beginning, then based on those parameters, you'll, you're right, the control file, and the setup dot C F G five, which is these are the two files.
9:21
I split definitely needs to control file. And if it has a control file, it will just run with all the defaults.
9:26
And if you want to set up anything more advanced, you also need to write that setup slash c.f.t. file.
9:32
So you write those files in the script, then you run high split and that's actually usually just one line.
9:39
You know, once you've set all these things up, you know, you run high split.
9:44
And then after it's done, you run any kind of post-processing programs that you might do, things like that display programs that we've been running.
9:52
You know, throughout the workshop.
9:54
This, is it, a trash plot is, is the graphical the plots. The trajectories are ...
10:01
plot the the program the plots the plumes but there's other programs as well and some of the some of the workshops workshop sections have, you know, run these other kinds of programs as well. So whatever you want to do after high split is done.
10:17
Then those are the remaining lines in your, in your script end.
10:22
Just to show you an example, I think this is an example of a Linux shell script.
10:27
This is just the first few lines, you know? It's setting some of the directories.
10:31
And it's, it's writing this ASCII data dot C F G, which tells us something about the boundary files and then it's setting some of the parameters for the run.
10:42
Then you can see it's starting then too, right?
10:46
The control file, and it continues on and finishes running the control file, it writes the setup file, runs high split and does additional additional things.
10:56
Um, you have given sort of just a workflow that you might do if you're doing a trajectory simulation, um, in the graphical user interface.
11:06
You see we, we did the trajectory setup run, we might do the advanced configuration setup trajectory, we got trajectory run model, um, we might view that messages if we want to. We then do the trajectory display.
11:22
We might do some additional post-processing with these special ones, like frequency things, things like that.
11:29
And um, this is sort of what you would do in a script along the lines of what I just said.
11:35
You know, you're writing the control file set of file you'll run, high split, which is just one line.
11:43
You might then run, you know, some of the other programs, and then other, either call additional scripts or run additional things in the, in your script.
11:55
Um, here's some of the items that are important for these different aspects of, of the run.
12:04
And here's the same thing for a concentration.
12:07
Simulation is really very similar, you know, you you have to write the control file, the setup file. You run high split. So it's actually a different program name that ICS That's the sea as concentration.
12:19
The other one was T S for trajectory.
12:24
Um.
12:25
You run ...
12:26
plot usually at the end and then the additional programs and you might do things like con to ASCII, which is one of the things that converts to the concentration.
12:39
Binary output file to an ASCII file.
12:42
You might run Conda station which is something that converts the concentration to a particular location, a monitoring location.
12:50
These are things that you can do in the GUI, but you can also just do from from the command line as as well.
12:58
Um, so, I mean, we may have shown this at one point or another. This is actually an older view of this, is, I think, even more files now.
13:07
But there's a lot of programs in heisler, and a lot of these are called by the graphical user interface with you pushing different buttons and different different menus.
13:17
Not all of them are, know, some of the ones that, you know, of course, we're using in the workshopping of Piteous Standard .... That's the trajectory program, ICS standard exceeded the concentration program are using ... plot trash plot.
13:33
And here's the other couple of I showed on the last screen, but there's, you know, more than 100 other other programs.
13:40
And each one of these can be run from, from the command line.
13:45
And a few people showed you yesterday that in the help documentation, there's a, at the very end, there's a, um, a section that goes through each of these programs and tells you kinda what they do.
14:01
But, the other thing that you can do with any, with almost all of these programs, is if you're in the command line, and you just type the name of one of these programs in without anything else, then it'll either give you a sort of a help screen of how you would run the program, or it will start, aye.
14:23
An interactive process of asking you what's the net data, or, or what time do you want to start, or things like that.
14:30
And in fact, here's an example of sort of the help screen that comes out. It's a little cryptic.
14:35
But you'll get to know these things pretty well once you start using that. So here all I did was I typed in.
14:42
I'm on the command line, the name of one of the programs, trapp trash plot.
14:47
And without any arguments, then it knows that you would like to figure out what the arguments are.
14:54
And you can see, there's all these different um, arguments that you can use.
15:00
And most of these are available in the graphical user interface.
15:03
Through through the menus, but not not all of them are necessarily available. This is actually an older.
15:10
But I think there's more things available, and I'm almost sure there's more things available now since the time I I produced this slide, but when you run something from the command line.
15:20
Um, it's fairly simple.
15:22
You can use any of these flags, so, the dash, I is like the input file names. that would be the T.
15:30
Dump file, and you, you just put that dash I there, and it in the program then knows.
15:37
The next text, whatever it is, is going to be the input file name.
15:43
And then the next one here in this example, is the output file name, and I'm telling it when it creates this this map up the trajectory.
15:52
I wanted to call it fire and it'll be fired um, postscript although, you can see, this was before we created, made the graphics into the SVG graphics.
16:04
So now I think you probably specify whether you want a, it's probably a flag here, to say whether you want a postscript output or SVG output.
16:14
So, and, you know, this minus a three is the GIS output, and I'm telling it I actually want to create a KML to Google Earth File, and with with that, I've actually added another option here that dash Capital A I'm saying three both.
16:36
No extra overlays.
16:38
Um, and no endpoints, so I'm basically just writing the line, not not the end points of the trajectory.
16:48
Anyways, this is the basic structure that you'll use with all the different programs in High Split, and without, there's one more thing I want to mention before we get started, this morning, there's a, there's a conference coming up in, about a month, and I'm actually in, excuse me, a month and a half or so that's at the George Mason University.
17:11
It's a conference on atmospheric transport and dispersion modeling.
17:15
I think some of you may have actually attended this in the past.
17:19
And I think it's of interest to anybody that uses high split. Our users models like high split or twice to solved.
17:26
Problems like Pi Split can solve.
17:29
This is a really great conference.
17:30
It brings together people that are doing these kinds of of calculations.
17:35
And usually there are a number of talks where people have used high split and in fact, this year we're actually co-sponsoring that the conference And we're hoping that if there's enough high split related, actually abstract Submitted, hopefully, from some of you, are some of your colleagues that we can have a special section on just on high split. But the deadline for submitting abstracts is July 16. So there's about a month, and I'd encourage you to check this out.
18:04
And maybe, considering attending, there'll be virtual options, as well as in person options.
18:12
So finally, then we'll get going for today and the first section is gonna be on radioactive pollutants and dose.
18:21
And doctor ... chai is going to be presenting this section.
18:26
10 thing is R, our resident expert guru.
18:31
he's in charge of all of our radioactive pollutant activities of which we have quite a number.
18:38
And so, 10 thing, I'm going to turn the presentation mode over to you, C, O.
18:50
Going to turn that off, OK, great, we can see the screen that you want us to see, and let's see if we can hear you, however, we can hear you great. Alright, OK.
19:04
Good morning, and take it away, OK, good morning, Good afternoon, good union to you.
19:12
Well, before I dive into the tutorial, like, OK, well, turn off the camera, the introduction of their has failed nuclear applications.
19:27
As I mentioned, in the Sten 2015 paper, well, he sets up all the history of a has to be a model and has a general description about the model. Actually, the model, initially, it was inspired.
19:42
Bye.
19:43
There's nuclear application, as it says in the paper, the scientific foundation and the inspiration for has been the trajectory capability can be traced back to 19 49 when the hour saw predecessor of the US.
19:59
Weather Bureau, charged with trying to find the source of radioactive debris originating from the first Soviet atomic test detailed reconnaissance aircraft.
20:14
So after that, there are so many applications related to their stance about nuclear application and steal one of the most important thing, although it doesn't happen a lot, but it has a huge impact.
20:30
So, on the lower left, I just gave you a live demo, all the Fukushima simulation You publish that in 2012 by traxler undraw. Ralph?
20:45
It's going to actually, in the tutorial, we're going to show you how to do this simulation, but, that's going to be ... the simulation.
20:57
In fact, the Rhine All and no obvious how.
21:02
two operations your scene has faded.
21:06
To do a right to our table pollutants, simulation and forecast.
21:12
On the left shows our back tracking support to comprehend 70 auclair Test Ban Treaty Organization.
21:21
And for this application, whenever City Beat Hill, they detect some hybrid actel measurements in the air.
21:32
They are going to send a request tools, several centers, including ours at Noah, and they are going to create a powerful day backward. That is partial model results, to back to them. And they are going to do analysis after that, trying to detect the source of the radioactive material.
21:54
Another thing it's called the Ares MCA Emergency Response.
22:01
RSM C stands for Regional Specialized Meteorological Centers, and this activity is related to the nuclear emergency response.
22:11
In order to prepare us for a possible nuclear power plant accident, it's our way to mostly exercise.
22:20
And in those exercise, the way, specify some location, most of the time, the the power plant, location, nuclear, power plant.
22:34
And they provide the trajectory, dispersion, exposure, and also some time alpha rival for the greater ... Taylor pollutants.
22:48
So let's go back to other tutorial. And.
22:55
So the first one is a very simple, simple one, waste, still going to your stock cap text tool.
23:03
Just assume there is going to be extend from there.
23:09
Though, that's a Euro, you need to press the reset button.
23:16
Before I do the exercise, I think tool, most important feature for these kind of application is your how to consider that take K, all for the radioactive material and is can be treated inside the code to all be post processed is inside the holidays, very similar to the other transformation.
23:44
And you will have the half-life or the parameter they find in the control file so that you can calculate the loss of the mess inside the code.
23:56
OK, and first, I'll follow the two at Harvard step two.
24:02
So we can see we are going to retrieve the control file.
24:13
Going to browse in A Tutorial Directory.
24:22
And can you you can find that enough.
24:27
Files, directory, and cap tags are control open OK save.
24:38
Similar array for the setup file true.
24:46
Told her rafael's set up OK sale.
24:57
So the in this exercise we assumed that a result in the megawatt reactor had been running for one year and it created a volatile.
25:08
Tend to 17 bakker. All of cesium 137 is a long living. Radioactive material has half-life of about 10 years.
25:18
Will assume like 10% off the ... is already stay an accidental over a one over a period.
25:25
So let us go to us set up here.
25:30
Yeah, we're going to radials runtime to a kind of five all verse and the pollutants manual.
25:42
Go into that and change. Share the name too.
25:51
Cesium 137.
25:54
And the release rate is going to be one plug. There.
26:03
10 to 16.
26:06
It's going to be just one hour release.
26:12
So, nice check, OK, same, save.
26:24
So And then go to the concentration great menu, And change the name of the output to seize on line $37 Bill.
26:48
This is very important here.
26:50
You'll need to change the number of vertical eyeballs to the dementia tutorial that you need to do this.
26:58
Because you'll have to add the deputation layer, that's the surface there, and then you will have the 100 meter for the air concentration.
27:11
So, you can see, is highlighted here by the didn't highlight this one.
27:18
You need to, to how this changed to two.
27:25
There are other things that you don't need to change.
27:40
The last part is deposition.
27:44
And here you will help configure that for season 137, just click on the Radio button, according to how the default for see them once erythema here, You can see the half-life of ... is more than 10000 days. So it's bulky.
28:07
Is about in like 30 years, I think.
28:15
So, I see after you click on the Radio button, I racing is going to be automatically changed, because you pre configured here.
28:27
And, uh, we are good to go, I believe.
28:32
So.
28:34
Yeah.
28:42
And then, we can run the model.
28:50
....
29:04
OK, Molly, here it says says the model is going to dry. It's going to take up our ceremonies.
29:19
So, as from the setup, you can see a lion parlance is, you'll have to, they find that a half-life of days.
29:27
And another standards are because, for these kinds of applications, it's important to get deputation results. So you'll have to set two levels. one for the deposition and the other for the lower level.
29:48
Soleil, calculate those later on, you'll know you are going to how is she lacking that those coming from the invitation and those coming from the air concentration.
30:02
And uh, those coming from the deputations, a lot of times, is called a ground shine inside because pick all the radial active material, deposited to the ground. And from the air concentration is call it?
30:21
Claus shrine because, they say in the air.
30:26
And a lot of times you can gather more detailed a dose calculation, including from the inhalation and some other organs by the theory, just, uh, trying to gather those after we get the results back onto your application rates all with an air concentration to calculate the dose.
30:56
So before the race is over, if he can help Little cat, you'll kind of get more information about those conversion factor from the some lab sat linked here.
31:13
So here's some. You can get some of the information.
31:19
And, uh.
31:26
Here's a US.
31:27
NRC Labs said the Nuclear Regulatory Committee, it has say because his, It's Men of God, who is dealing with the nuclear accident. So, if you have more information about that.
31:47
Yeah.
32:04
And, uh, and for this application, the Arab unemployed well assumed that those conversion factor for the, for cesium 130 salmon, is going to be three upon the full 10 to -11 rim power.
32:24
That's the dose rate for every backhaul killing me tareq air concentration.
32:34
And for that application you'll have 1 on 1 times 10 to 12 gram per hour bould Yeah.
32:44
Average becquerel per square meter deposited material.
32:50
Because, later on, we are going to calculate, ah, that dose over 24 hours. So, we need to multiply those numbers by 24.
33:05
For the deputation and multiply the air concentration by sree.
33:18
Because there are concentration units over three hours.
33:24
It's almost 94%.
33:29
So, that's why we look at the results later.
33:33
Those are the conversion factor of AI, how to use to convert the concentration and deputations into dose.
33:55
92%, almost don't.
34:06
As I mentioned because I see them once every summer has a really long half-life so the policy, the material are going to stay there.
34:15
So after we do the dose calculation for one day, then we are going to calculate equivalent of this fall over year.
34:29
That's great.
34:34
So they are calling into your concentration display and the concentration contours.
34:45
So, they have this already generated and, uh, as shaq.
34:56
So, because we're how to levels that, we need tool shows at both levels, we need to change this And, uh, Adaptation Multiplier, Malraux dimension, Hartwell, Gatos.
35:20
Can they need to show the total off the deposition and the concentration?
35:30
Yeah?
35:35
Yeah.
35:38
Care redefined.
35:41
We'll use the extra potential scale.
35:45
and fixed.
35:51
Here.
35:54
Inc.
35:57
May have everything changed according to the tutorial S X Cubed Display.
36:12
So you can see here is, uh, concentration.
36:17
In the very beginning, Yeah, it goes, It's the plume.
36:22
You can see the takes 10 day, then transport it, because there's only one over release.
36:28
Cell is calling into more wildly from the source.
36:39
The deposition rates, all states only, integrated altogether, so you only have one.
36:45
And that's OK.
36:50
So actuary here.
36:51
what we show here is, uh, uh, for the devastation, and you can see the maximum is 9.8, 10 to -7 per square meter.
37:07
Well, the unit should be the ring.
37:10
Although.
37:15
Probably they need to change the graphic for label, no order to show the correct units for the fault.
37:26
And before that, I'm going to check dose map.
37:34
OK, so bring you to the mic, that change for the labeled file, so we can go to the file I did, So Advanced File, Edit, and Powder Label.
38:00
So, the title, how cesium one sorry, seven, and laptop, we change that too.
38:19
Dos a prevalent.
38:25
They don't international and because they all write A convert, the mass concentration to those and the neonatal doses ram.
38:39
So save.
38:48
And execute display again.
38:52
So you can see here, back at the unit of change of terrorist, re modify the powder label.
39:03
Right?
39:05
So that's it.
39:07
Almost always first exercise and because of ghost mode.
39:18
Sir.
39:19
So again, exactly similar results as in the tutorial.
39:25
And then they can also do the tech one year dos calculation by changing, uh, that conversion rate tool.
39:45
While you look at this number by multiply the original number of aids, a solid earned 706, hey, because as the horse, you know, a year.
40:19
So still, you'll have the concentration every three hours, and you can see the dose from the air concentration is pretty low.
40:31
You know, for the, for the last one, it's, uh, air concentration ***** the law, but because of the deposition, that's the much longer, longer.
40:46
How safe act because they deposit their and foresee them once or salmon.
40:53
Half-life is so long, so it doesn't change during the year.
40:59
Yeah, excuse me.
41:02
So the mouse is going to be a surf on safe, tend to my nurse told rim, and trays still.
41:10
not high.
41:13
And because that's like either If I set a cold case and we're going to work on other malware, at least one in the next section.
41:28
It's called ... before we call that.
41:31
Just Reveal was mentioned here about mailroom a year is made for the house a fact.
41:41
So, what we get is still lower than that.
41:45
But a lot of times, uh, the law of ideal you'll get is that because you don't have very high resolution meteorology data.
41:55
If you use the High Resolution means larger data and they also calculate the your air concentration and deputations. High rates ocean are great.
42:09
So yeah, according to how song, close to the source.
42:13
Region base very high concentration that they saw you, how high those over there. So that's just trying to emphasize.
42:23
Yeah, the simulation results give you a long range transport.
42:27
A lot of times, you'll made May some high concentration close to the source.
42:36
So, let's work on the next one.
42:40
Right.
42:43
This one is going to calculate the long range transport, awful ... from Fukushima nuclear power plant for the accident.
43:00
Are done, once everyone is ready to add radionuclides, which has a much shorter half-life, it's only about eight days.
43:12
And during the accident, the madmen location at Harbor, Alaska, it's actually the matter very high value of that.
43:22
So welcome to today's simulation and first, still which Retrouve?
43:33
Japan control, the control file is called Japan. Control alt Text T.
43:41
It's going to be in the tutorial, derived trays up, their Drive Tree is going to reach a pen.
43:49
We'll have the control file here.
43:55
True.
43:59
So.
44:02
For the configuration you do the same configuration set up.
44:07
Conficker rituals.
44:11
From a trip and directory.
44:13
And, uh.
44:17
Setup fell.
44:22
So.
44:26
Go to the Setup menu.
44:35
So open the Setup menu and said, the models start time till 11014 taro taro is already from the scene, the control file, so you don't have to make changes over there.
44:51
That's just revealed what's there.
44:56
Release location, latitude 37.4 and longitudes 141, and release height is 100.
45:09
Molly, how to mention that here is a 100 meter is compromised because the actual explosion poverties, they should be much higher than this could be 100 meter about.
45:27
We don't have ASAP information for the *** or for the exercise, which has 100 meter for this, but for the long range transport.
45:37
This doesn't change the result a lot, although it's going to be mandatory for the local transport.
45:53
And, uh, they have allowed the runtime.
45:58
Yes, 196, All worse, for eight days.
46:05
And we have the Global re analysis data with.
46:14
With a 2.5 degree resolution, so the ride is going to be faster.
46:25
And nobody can check the concentration great here to see.
46:34
The spacing concentration, and resolution is the one degree.
46:41
And you can see the spaniards 181 for the latitude.
46:46
The reason 180 why is because it includes both the North Pole and South Pole.
46:52
So you'll have one actual point and longitudes 400 saved here, so we have the whole globe covered or the concentration grade.
47:07
OK?
47:11
And, say.
47:16
Before proceeding.
47:20
If, anyway, so you will see in the lab portion you how to download that RJ data and uh, make other changes to apply to the location of the file.
47:32
So open another thing in the Chat tool to see the ..., we have the ...
47:41
one and the animation rate is said to be ..., 10 to 15, yeah.
47:51
And the duration is 24 hours.
47:55
Really Star Time, March 14th to 12 UTC.
48:03
Already saved?
48:08
Hopefully, the concentration, great, lot more time, we have the level as 500, it should be, OK, Because we are going to calculate the long run transport. So the deposition and also important, here we only consider the air concentration. They only housing 1 level 500 meters.
48:33
So, they are raising a fund.
48:38
And now we are going to calculate the 24 hour average air concentration.
48:45
Save.
48:50
So, next, uh, that's revealed that application manual.
48:57
And here.
49:04
And you can see the pre configure, the hired on line 31.
49:07
It has two options: L Y and the forecast they are there for a particular edge, and we need to check the radio bartella awful, one circa, one caches.
49:20
And it here you can see then everything is going to be changed to the half-life all for Airtime eight days.
49:33
So you can also see the ..., et cetera, It's constant, zero point zero one.
49:44
Yeah, Say.
49:49
Before starting, We also need to change the setup.
50:02
Kotler Setup table Sub menu will fall.
50:08
We're going to change the particle numbers.
50:11
It's already a fyodor retrieval from the Geoff Hinton setup.
50:16
It's already changed, that's 24,000 per cycle.
50:21
And the maxima is 25,000.
50:28
So, it's what they want. And another thing is to reveal the menu or supplemental one. The time-stamp is going to be sad as a constant 30 minutes.
50:46
Yeah.
50:48
Sylvia, however, is the inside out.
50:50
All right, so We are going to run the model.
51:09
So.
51:18
The one that runs stars you can see the first 10% is really a falsehood because it doesn't have it released during that.
51:28
We can see that does release Star side, totally OTC on the day, but B, how the models start had several tail, so.
51:51
Again, they are going to, after the rally is down, they'll go into showed.
51:59
That, those creation on, about four days, one and A First, we're going to.
52:08
To check the boundaries also, based on the measurement.
52:12
So, we're going to convert a station after the rhinestone.
52:23
Well, once they are in poverty for this kind of oils calculation, one thing you have to be careful is for the airtime.
52:33
Because the half-life is really short.
52:36
The way you refer to the emission rate, your, how to, know, what's a reference time for the animation, because it's going to decay off from that time to allow the USA, is called a decay corrected to a certain time.
53:01
And then most of the time, the The release of the material is happening.
53:10
After the shutdown saw the inventory, emission your mentoring. Maybe they find that a different time.
53:18
So you'll have to be careful about the decay correction.
53:31
Or less.
53:37
Turn to see what we are going to next to go to the states For the sampler text Dot Tech CFL is just they find the location and all for the ... Dark Hopper in Alaska.
53:55
So, Sir, Important Night, latitude, North, 166, longitude West.
54:13
After the raised only need to remember to use a multiplier. one solitary, because a unit cell for the measurement, is malay becquerel per cubic meter.
54:26
And we had the ...
54:28
set up ads, up becquerel Paul R. So, the results should be converted using a multi-player when salt. And.
54:59
seven digit, 2%, by the ratio, will be just about three minutes to complete.
55:07
So, because kind of normal time for this.
55:16
Well, if you have time, it's poverty worse folio to try, you're seeing higher resolution.
55:24
Global data mythology data.
55:26
You can use half degree, G das data, all budget.
55:34
Here, your usual held, that ...
55:37
data, already downloaded to your tutorial directory, so that you can do the exercise, but the simulation is going to take a much longer time. So we are not going to do it here.
55:53
And the world is like a you need to know is because when you use a higher resolution with RJ. And the results, again, will yield higher resolution concentration, as well.
56:05
So, close to the, uh.
56:12
Instead of the 2.5 degree, very solid, you are going to how are, like, a different results for the peak.
56:21
So we're going to just show you the results of the, the run we are doing here.
56:30
Yeah.
56:33
OK, It's down.
56:37
That's it.
56:40
So we'll go into deals.
56:43
Concentration, utilities, come word, tool, station.
56:54
And here, we just, uh, Import the longitude latitude directly to the to the table here.
57:09
Think that's, Yeah, that's right. Yeah, OK.
57:15
That's the name of the station and latitude is 53.9.
57:23
Longitude is -1 6 6 on the file.
57:38
So due to how the multiplier one thousand.
57:45
Yeah.
57:51
OK, we have already seen to fund correctly so the eighth track data.
57:59
So they see the peak is the forte full.
58:03
Oh, mainly becquerel per cubic meter.
58:14
On March 8, 18 is sauce for days after three days.
58:22
And they can plot the data.
58:28
In order to compare, is the measurement of a how to define the supplemental data here.
58:39
See, it's going to be a Dutch harbor dot TXT, and the content of that file inside here is a Dutch harbor.
58:50
So you can see it has merit, the Mayor Battle, Insider, Files, you can call it comparison, so that's exactly as given them totoro, so here.
59:10
That riot is from.
59:13
From the runway, just finished, the black is from the measurement, and so you can see the estimation for the concentration.
59:28
And Here, which are the finish of that discussion?
59:35
I bought her on another run using a higher resolution global data.
59:41
So, we're not going to do the run by that, Show you the results here.
59:47
So you can see the high merriment value is around 100. But if you have a higher resolution ...
59:55
at zero point five degree, you're going to overestimate the peak.
1:00:01
Is not going to, I'm not going to say the which raised certain bad her, but her probably prefer to have the higher resolution results.
1:00:22
So what? Uh.
1:00:25
Well, the care, the last paragraph, is just trying to emphasize the importance of all the heartache data.
1:00:33
The results are going to be very sensitive to those.
1:00:41
And next, we're going to, to a more drastic test of those calculation, close to or to the source.
1:00:54
Create all the tables here.
1:01:07
And first, let us retrieve those control dot TXT.
1:01:13
And then, ... reveals some other materials.
1:01:22
Browse You know the same Japan directory You have that?
1:01:31
Thing you might want to reset before you retrieve those in possible. Yeah, OK.
1:01:39
Maybe not, but it might help OK, Power is sad and rich true that file again.
1:01:51
Those Control Coltan, Coke sale and a sellout fell Betrayals phone Those set up, OK?
1:02:14
OK, reveal the required changes here.
1:02:20
Go to the Setup menu.
1:02:28
Silly is going to be the start time is going to be March 11, 2 LT.
1:02:36
And the Total runtime is going to be $180.
1:02:44
Star location is still the same because it's the same Fukushima accident.
1:02:55
And, uh, this time they have the ... data.
1:03:00
Because we're going to, to a shorter range transport is not going to go to the global one.
1:03:14
So Harris here and open the Floor Turns menu.
1:03:24
So you can see, we have two pollutants, Kara, one, actually, they are like, uh, They should be like a surrogate for two different radionuclides.
1:03:39
And here are, the first one, is called, the R N, you will see, you can see the animation rate.
1:03:47
here, the one, and the number of ... 24 and release the starting time. Everything went on, I don't need to change.
1:03:58
Kate, for the specious too.
1:04:02
Um, everything, I'll say the same except for the, the name. Here your, how is a gas.
1:04:13
Which stands for noble gas smell, kay?
1:04:22
Then open the deposition manual for particular species one.
1:04:31
Yeah.
1:04:36
And, uh.
1:04:39
See, the default value of System 137 is almost same.
1:04:45
If I click on this, the only change is that there's a half-life Actuated Cyril's.
1:04:55
Originally, it's a role, is consider not going to decay, but for the system, one sorry, silent, 30 years almost isn't is identical asimo decay for the next five days, simulation or even shorter time we have here.
1:05:14
Yeah?
1:05:17
And, it doesn't matter. So here how, that all changed to zero?
1:05:25
Thank.
1:05:27
Specious, too.
1:05:34
You can see that, for the noble gas, everything is set to zero.
1:05:45
So, there is null decay for the, for the two species that they find here.
1:06:05
OK, we are going to check the concentration grid again, see, by the way, how would that application defined?
1:06:14
See, a number of our vertical levels. We have to ...
1:06:18
the deputations zero, and how 500 meter, OK, fine for the air concentration, ...?
1:06:27
hello, Much higher spatial resolution, and zero point zero five for this like a regional or local simulation.
1:06:44
So, average income is defined as 36 alvares for this the assimilation.
1:06:52
Starting time?
1:06:56
March 14, after ..., and Nikolay, until the end of the simulation.
1:07:07
OK, Yeah, then I called Till the Configuration, or the Events configuration, set up a concentration.
1:07:18
In the manual 10.
1:07:22
You will notice that, here, we are how the single particle mass dimension as tool.
1:07:29
This is meant to how one particle represented two different species, in order to save for the computational time.
1:07:47
OK, so they revealed, And then they can go to the manual for, to check the particle Numbers definition here.
1:07:59
So, where how staff salta inch particle is released per cycle.
1:08:07
That means 500 particles. Valerie period is each animation haller.
1:08:17
And then they can run the model.
1:08:27
Again, the first 70% of the road a FAFSA because Yeah.
1:08:33
Because the Simulation has Stars results released at first.
1:08:52
So here, I think that the most important part for this test cases to see hall, the alcohol and to yourself to surrogates to represent the French radionuclides.
1:09:09
For the power plant, nuclear power plant.
1:09:15
Although they have a single field, but they are going to be many different profiles are going to be generated by the reaction.
1:09:28
And, again, this, you.
1:09:35
Our normal practice is, you'll see several, a surrogate for this case, so they have to surrogates to represent different radionuclides, such as the noble gas, and the other is dissimilar to the system, once or examine, is consider as a particle.
1:09:55
So that weekend.
1:09:59
After we got, we got the data, which in February for those two, we can calculate the total concentration by yoshino TCM approach, but even if you don't have to scan for this, why is only a single run?
1:10:16
You can still apply daycare, op terror, the model rise completed.
1:10:22
So all pi, the the dilution factor and decay and take a ... rate for different radionuclides.
1:10:38
And while they're waiting, we go to see the activity, test EFL, they'll go into your ... tool.
1:10:50
Calculate the total dose.
1:10:53
Yeah.
1:10:56
So here are the 10 most important radio Neil Cries.
1:11:04
And the first column should, at the beginning you chose the name of the book, the radionuclides are the element name.
1:11:14
And then it's the half-life of the stock, the last, the next four columns.
1:11:21
Actually, you can see, they are actually the same, because they say is to represent a single accident for the Fukushima. So our original format is going to show your faulty friend or after the foot ... graduates and the high energy and T Recreates and ...
1:11:43
and the for the Power plant. On how the Uranium 2 solid 235 as Fuel Source. They can also have an approved ... to 39.
1:12:00
So they are going to be, like four different nuclear power plant profile.
1:12:10
But for this one, everything the same as the Fukushima one.
1:12:15
And then, as mentioned earlier, we need to calculate the dose.
1:12:19
Solvay, how to calculate the was the conversion factor ...
1:12:26
from the air concentration hodo can get that those strayed from the concentration and how to gather, Quantum Shine, that all straight from the deputation.
1:12:42
The last column is about insulation a favorite we are not going to use here.
1:12:55
Yeah.
1:12:56
So, the rainiest, 91%.
1:13:01
And then, we can go back again to look at how to use the activity, alt text T this profile.
1:13:08
Because we, no calculation, we have to surrogates one for ..., such as ... care, you can see that.
1:13:19
And now the line is for the noble gas, so actually, for the first, the 10 most important radionuclides, only one of them is treated as Noble cause poverty, You can see.
1:13:37
It's uh!
1:13:42
Is seen on one start A three. You can see the quantum shy is conversion rate is zero because it, as long as it doesn't deposit solid, it has such zeroes. So, you can, you will say to differentiate from noble gas from the other partakes, alright.
1:14:05
It's 95%.
1:14:17
OK?
1:14:28
Well, another thing about this activity, the test is the importance of different new radionuclides are going to change from the beginning to to later.
1:14:43
Because of the different decade, hi, half lifetime.
1:14:50
Because the short-lived radionuclides won't have much effect hurtful to follow date or time.
1:15:03
Almost all 98%.
1:15:09
Yeah.
1:15:12
So for this results, we're going to change the dose tab to Total, and we're going to use another yo, Yo ... .
1:15:24
The output of well-being in total sievers rather than sailors proverbs.
1:15:42
OK, OK, great.
1:15:48
So they call it to kill Taillight Harris, and convert to dos.
1:16:01
So, let's try, Whether we have the activity ..., in the working directory.
1:16:07
If you don't, you can create a new.
1:16:12
Yeah, already created has, oh, check is, should be the same as the ... tutorial.
1:16:23
So, it's the same with C, now, zero conversion of Health Care folk range, quantum shine.
1:16:34
OK, right.
1:16:40
Well, another thing like, if you will, How are the applications probably for today's activity? adult has to you? how to gather data from some other sources, because here is the only some represented tail full day?
1:16:58
Reactor or detonation.
1:17:01
And it may not be the exact what you'll want. But if you don't have any information, you can start from there.
1:17:10
Oh, already have that, and.
1:17:15
OK.
1:17:19
So If a created, not why it's not going to change the result.
1:17:29
Toilers handles the according to change, our total output have a claim to your dose and clung to some species.
1:17:43
And those tab, they are gone tool.
1:17:45
Changes to total price, the Draco bought in off the doorstep total.
1:17:52
And we're going into yours, sievert as a neonates apply take K.
1:18:00
yes, because this, the half-life of information is included in today's activity dot TXT file so they can do that after regenerate the activity and ...
1:18:12
OK.
1:18:15
So, we can create the ... file.
1:18:28
What it does is applying this equation, although, the share TCM is only from a single run, but they still need to apply the decay factor for each species.
1:18:42
And also, the emission rate is, uh, here, this column is the emission rate for different radionuclides.
1:18:55
OK, Still running, Yeah.
1:19:07
Actually, you can check out the firewall really generated when I did that. It didn't show the message over to, come back to Jeff Dahn, but it was already generated, OK.
1:19:26
Mmm hmm.
1:19:42
Yeah, I think it's already there already.
1:19:47
OK.
1:19:54
So, yeah.
1:19:57
It doesn't give you the complete, uh, message, I think is a hunch, will mesh in poverty down, because the sez's looks all right, So they come cloud, OK.
1:20:18
So here, the first plot is, uh, A combination dose, in the malay Sievert Unit, Average between zero meter to 500 meter.
1:20:32
So we can see a very narrow region.
1:20:36
And, if you look at the air, the results, it's about the same.
1:20:45
And the ... two Millisieverts.
1:20:55
And we can go to see the deposition, dose, accumulated dose at ground level, so you can see the much larger region, and you can actually have the, take the maximum value. Much larger than the air dose.
1:21:17
You have a 15 malay Saylor close til they're close to the Fukushima really sad.
1:21:28
While you may note, is where here is to Kendall Square Care.
1:21:33
That's because of the resolution here, is very close to, uh.
1:21:40
To the table upon five degree, off of that, she had asked.
1:21:47
If you have even higher resolution, me hard data, property, you're not going to how this kind of artificial effect.
1:22:01
... CA.
1:22:13
You can check out the, because for that accidental, a lot of data is still not finalized. There are still going to be some updates. If you want to do the simulation of that one, you can go back to the law.
1:22:28
Uh, to the Japan Euclea Recreation, ..., to check, then you'll see information.
1:22:44
So now let's see here is, uh, that Apigee does show some ....
1:22:50
And the year, a premade hadar for the general public is 10 ...
1:23:00
above that house level.
1:23:06
And the last paragraph, Gabriel.
1:23:11
Some information about the ..., how very simplified version of the exercise. So you can use this example as a starting point, but layout to the real calculation and how to consider a lot of other six.
1:23:26
Just be careful about it. And another thing is, like, hey, you'll probably need a lot of post-processing program to generate your results and the ...
1:23:40
not not going to gateway, You'll all have a better taste you'll want.
1:23:47
As Mark mentioned, alert air, take the scripting. Poverty Cailliau mall, pick apart the two.
1:23:56
To do the simulation and for the real applications.
1:24:01
Have some time poverty and hunger, too, tracks the exercise. The first exercise is about to check the top 10 degree to new cries to see which, one, it's more important for quantum shy. And the cost sharing ... tried to emphasize here is the top 10 list can be very different one day after they are released, as it became, you know, maybe it's one radionuclides is important, but later on is a dwell changed to another one, because self the decay of some of them there faster.
1:24:45
Second problem is, if you do, you are going to notice the effect of all different resolution of the ...
1:24:54
RJ and different resolution for the, for your concentration, for your simulation.
1:25:05
And the last day in some poll shows OK.
1:25:11
Some other leg amputation may last day.
1:25:14
A huge impact on the results.
1:25:18
If you have the deposition velocity, I said other differently, you are going to how a Deputation mega, very different deputation rates all set.
1:25:27
For the nuclear applications, the deposition is going to be very important because it's the gateway or the most important, like a dose for the most of the cases, especially close to the really sad.
1:25:45
OK, then maybe I'll just call sir, oh, here's a summary.
1:25:53
For the Radio Alto pollutants and the dose, we need to calculate the decade. So that's the poverty of the most important thing to consider for those kind of applications, and for the dose calculation, you have the quantum shy.
1:26:11
From the air concentration, you also have crawshaw from the air concentration.
1:26:17
You have the quantum shy from that application, and, you know, all there to make the Grandma efficient, You can house surrogates for multiple retinue cries and, uh, to the final, I'll pause for sizing to ... class. And to do that, you are going to need a ...
1:26:42
file that they find this. They impart that and create a new crisis, including the half-life, all those.
1:26:50
And the emission rate.
1:26:53
Uh, also, they kuang shy Claus shrank conversion factor from the devastation, and the air concentration.
1:27:05
And for the real application, you need to be careful about the animation decay correct here. At certain reference time, everything has to be consistent.
1:27:16
And the ... sometimes like Gabriel, that's information, but you'll have to be careful.
1:27:25
And last thing I want to mention here is for the real application, property, you how to how well ... approach.
1:27:34
Stand for the Transfer Coefficient Matrix is basically a Jessica will yield a dilution factor bahrainian that a friend.
1:27:45
It runs for that ... SAML approach.
1:27:51
It will also update the forecast to be a quickly produced ...
1:27:55
estimates made available, especially for at least nuclear applications.
1:28:01
The release rate is not going to be accurate at the very beginning, They are going to be updated at a continuous rate. So, that TCM approaching them hardened. It. So every time when the animation is going to be updated you don't have to run the model. You just do the post-processing.
1:28:27
As the last example shows how to do that.
1:28:31
And for the TCM approaches you divided do you are ready time in tool, or smaller time sacraments.
1:28:39
So, each segment is independent, calculation your random separately.
1:28:46
And then you can think, multiply combine them together. And in fact, the husband model provides Samuel ...
1:28:54
to do that including, uh, OK, mega Kong Merge and say to our array to a January, it's also array.
1:29:05
Well, they have some fun, I'll show you another scan on Powder Fukushima.
1:29:14
a source term estimation of a deed using classroom model and it's using the TCM, as I mentioned and for that Fukushima.
1:29:27
the sauce reduce are not going to be finalized.
1:29:33
Poverty, technic perfect results is not never you can only estimate the reduced rate from the measurements.
1:29:42
And the the results are going to be updated, continuously.
1:29:48
Weighted estimation using the global observations.
1:29:52
And, uh, we ran the model from different times segments, using those measurements.
1:30:01
We can estimate the emission rate, and we found that over a mesh has made it very close tool to the others ratios are much more information and the local information.
1:30:22
I think poverty. That's what I'm going to cover.
1:30:29
Mark?
1:30:30
Yes.
1:30:32
Did it, do undertake over, is that OK?
1:30:36
OK, very good. Thank you very much, Tim, thing.
1:30:40
Really excellent, and you can see why.
1:30:43
10 thing is our expert on nuclear nuclear simulations, and in the unfortunate event of a nuclear accident. I'm going to change the presenter back to myself.
1:30:59
And let's see here.
1:31:02
Yes?
1:31:05
Show.
1:31:09
Screen.
1:31:12
Yeah.
1:31:13
So we're gonna go into a, um, a 15 minute break now, and we'll come back at actually 15 minutes ahead of time.
1:31:23
So we'll come back at a quarter to the hour and start the next section on volcanic eruptions. So we'll see you in 15 minutes.
1:31:35
Bye for now.
1:31:38
Welcome back.
1:31:40
We're back from our break and over, going to be starting the next section, which doctor Alice Crawford will be presenting, and Alice will be demonstrating a lot of the some of the more advanced but really important features with the emit times files and emitting different particle sizes and and restarting runs and things like that from from the particle dump dump files.
1:32:11
And she'll be doing that in the context of we do volcanic simulations of volcanic emissions.
1:32:18
So, Alice, I'm going to change the presenter over to you.
1:32:25
OK, can you hear me?
1:32:26
Yes, we can hear you, and I'm going to turn off my mic.
1:32:33
Aye.
1:32:36
There we go. And we can see the screen that you want us to see. So.
1:32:40
Perfect. And take it away.
1:32:43
All right. We've made it to Section 16, right? So we're we're in a way. He renamed this section, the share from volcanic emerson's too.
1:32:51
Complex emissions has been found that a lot of people who who don't model volcanos actually use the material in this section and they were kind of having a hard time finding it.
1:33:03
Um, a lot of it is very relevant to volcanic eruptions as well. So we're going to be using that as a kind of a simple example to show you some of these.
1:33:15
These capabilities, so we're going to look through, so we're going to start with doing multiple particle sizes as well as with gravitational settling.
1:33:25
I will talk a little bit about the gravitational settling schemes available in high split, also restarting the model from what we call a particle dump file.
1:33:35
This has been mentioned a little bit before, but we'll talk about it in more detail today and, um, usually we consider kind of the main outputs of highest, but the concentration files or the trajectory files, and the part I'm files are actually can be very useful, uncertain situations.
1:33:55
There are another binary type file: The rotating and shifting the plume will probably just touch on that.
1:34:03
And then, the particle size distributions, as well.
1:34:06
Yeah.
1:34:07
That's what we'll go over that and then hopefully we will, the aim will be to kind of get through those things and be able to spend a bit of time on these emit times files which you've seen in use for the smoke.
1:34:19
You've seen them for the wildfires, but we haven't really talked about how to, given a lot of detail into how to construct and for various emissions scenarios.
1:34:30
And, you know, this is really the tool that people, people use highest, but for all kinds of things. I, one thing I miss about having these workshops in person, is that you use to talk to the people that came to the workshops, and you'd find out all the things that they were doing with high split. And it's really amazing to see, you know, the type of emissions scenarios that people are trying to deal with, and so forth, So.
1:34:57
Um, a lot of times, there's not like a single button on the GUI to deal with your emissions scenario that you have that you have.
1:35:06
So you have to be able to, um, construct it yourself and the tool that you can use SSM at times file.
1:35:14
Alright, so we'll go right into the particle sizes with gravitational settling. and this is of course important for volcanoes because the volcanic ash has a particle size distribution, and the gravitational settling is important because you get separation of the different sizes over time, especially when you have fairly long range transport that takes place over more than a day. So, we're going to, as usual, are going to set up our run.
1:35:44
And we're going to start by retrieving pre written control file, ah.
1:35:56
And silky in.
1:36:01
I forgot to see where it then, let me see.
1:36:06
Controlled bulk, right?
1:36:30
Save that, and then we're gonna get a setup time file as well.
1:36:45
OK, and We're going, alright, we're gonna start running it, and then we'll talk about, um, What's in what we just set up?
1:37:00
So run?
1:37:02
I'm using the setup file, OK, so what did we just do, if we open?
1:37:11
Clinton you can see the starting time is 414-2010 um.
1:37:24
And there's two starting locations.
1:37:28
So this is the location of the ... volcano.
1:37:36
And then you'll notice that the Latin line of the two locations is exactly the same, and the only thing that's different is the start of the heights.
1:37:48
So, when I split sees this situation, where you have two starting locations, the same in two different heights, it assumes that what you want is the mass to be distributed evenly between the bottom height and the top plate height.
1:38:03
So this will create like a uniform vertical line source at this point, and so that's kind of a very simple representation of a volcanic plume where you just have the mass distributed and a uniform line from the height from the event to whatever the plume top height is.
1:38:26
You might say, yes.
1:38:27
That is, You know, you're realizing that as a simplification that plants usually have not a uniform mass distribution in the vertical and there are ways to take that into account.
1:38:39
On the other hand a lot of the operational setups at the volcanic ash anti israeli centers, still utilize a uniform line source Friday or initial, at least for their forecasts.
1:38:49
Because a lot of times we just don't have information about the exact mass distribution and the, and the line. And there was at least one study that was kind of comparing, you know, can you get a better forecast, if you?
1:39:05
If you tried to do a more realistic distribution, and the answer was kinda like, Well, Unless you have some information informing what that distribution is, a lot of times a line source kind of covers all your basis because you're kind of, you're putting mass everywhere and not assuming.
1:39:21
You know that there's none, low down or that it's at a particular location.
1:39:26
OK, we also have our total runtime as 12 hours here, You'll notice the top of the model is probably larger than we did in some of the other.
1:39:34
Um, France and that's because the volcanic eruptions are a bit different, and you know that usually you're dealing with dispersion much higher in the atmosphere than, say, like a chemical release or a refinery fire or something that's released closer or even a wildfire.
1:39:54
All right. What else do we have here. Now, it's done already, all sorts kept looking.
1:39:59
So, we have, one thing that we have is that we have four particle sizes defined, and they each have their own emission rate.
1:40:10
But, they each have the same hours of Emerson, So, it's, it's set up so that the smallest particle size is only like 1% of the of the total mass emitted. The two micron is like 7%.
1:40:27
six micron is 25 and 20 is 67%.
1:40:31
Um, and this is from, ah, you can get different estimates from different studies about what's the proper particle size distribution for a volcanic eruption.
1:40:48
Clearly, you know, using just for particle sizes are somewhat, um, course estimation. Sometimes people use more particle sizes.
1:40:59
Uh, sometimes that distribution is a little bit different.
1:41:03
Um, I think USGS, they did a huge study trying to create a particle size distribution that would be appropriate for using for ashfall.
1:41:14
So, for instance, this particle size distribution would be geared toward only modeling, airborne Asch, that would stay aloft for a significant period of time and possibly be a danger to aviation.
1:41:27
If you want to look at the ash fall, like the deposition around the volcano, you would include a particle size distribution that had a lot more larger particle sizes. Because those fall out very quickly.
1:41:41
And they are Hazzard toward, you know, nearby area of the volcano.
1:41:47
And so there are particle size distributions that have been kind of constructed to give you better deposition forecasts.
1:42:02
You can in the US. The USGS generally does deposition forecasts. New Zealand has actually using high split further deposition ashfall forecasts as well as their airborne ash forecast.
1:42:14
So it's just, it's a little bit different of the setup for doing one versus the other.
1:42:23
All right, we're gonna save that. So also, so here's the second species. Identification number, there's nothing magic about that number.
1:42:32
You have to, it has to be less than or equal to four characters and otherwise you can create it.
1:42:38
However, what helps you, remember what that particle is corresponding to then SPC is three and species four.
1:42:51
And if you go to the deposition menu, you'll see that these have a particle diameter.
1:42:59
They have a density. And they have a shape factor, which we usually just set to one.
1:43:06
Even.
1:43:06
Now volcanic ash particles are probably not spherical.
1:43:12
We don't really know what their shape factor is.
1:43:16
And all of these numbers are a little bit uncertain, right?
1:43:20
So, like, all of these are just used to calculate settling velocity, So if you have your shape factor a little bit wrong, um, the error in your settling velocity is not going to be going to be that large.
1:43:37
And it's sort of like you'd want to create these so that your settling velocity is cover the range that you would expect.
1:43:46
Um, there's also wet and dry deposition.
1:43:52
Fred Volcanic Clouds, which, for the ash, and, and this can be quite important, although it's a little bit tricky, because sometimes the precipitation is not always forecast and the correct place and the numerical weather prediction model. So there's some uncertainty associated with that.
1:44:10
For our grid, we have a pretty simple grid for this, mean, this is not really realistic of what you'd actually want to output. You'll see that there's one vertical lever level. Usually, you know, you'd want a whole bunch of different ones if you're trying to get the concentrations and the average time is six hours that you'd probably want at one hour.
1:44:29
We're just doing this here because we have already really in the past sections talked a lot about the concentration grad, so we're not really going to focus on it and the section.
1:44:39
Um, even though it is important if you're modeling volcanic ash.
1:44:43
Uh, the particle we already went through this. Yeah. Oh, and if you look, there's one of these for each of the species, right?
1:44:56
OK.
1:44:58
and the other thing, um, oops.
1:45:06
There we go.
1:45:10
That we So, since we're not looking at the concentration output files, well, we are going to spend more time on is the ...
1:45:20
file, so, that's these, And this one. Again, we didn't do a first out. We're just looking, like, we're going to start looking at the 12 hour 1.
1:45:32
So, we're going to look at the concentration display, and instead of looking at concentration, we're going to look at that particle file.
1:45:47
So, Oh, And I do have us looking at the, ah!
1:45:59
OK, well, it does have us looking at that first, sorry, Let's go ahead and do that.
1:46:10
So, you'll see that.
1:46:13
And the current the Plotting program, you do have the option of looking at the concentration that's an output of all the pollutants that we looked at, or only one of them. So, here, we'll look at all.
1:46:28
Then, there's really only one level that we define, so that doesn't make much of a difference.
1:46:34
Um.
1:46:38
And where can I create these?
1:46:53
OK, I think I didn't do the sam. They can say that.
1:46:56
That's, um, that's the plume.
1:47:03
OK.
1:47:05
Then, let's go ahead and look at the.
1:47:13
I got the particle file, as well, ah.
1:47:22
Yeah, so, that says the output of the the particle, the power dump file, you can, this is the vertical projection onto the vertical, and you can see that this is the van try it.
1:47:38
So, the particles are being released, and the kind of a uniform fashion from 100 meters up to, uh, for outlet it was already, um, 6000 meters six kilometers, and then they're moving forward, and you can see the effect of the wind shear here, right?
1:48:01
Like, particles that are higher up, happen faster, there's clearly a faster windspeed at the higher altitudes that are carrying the particles farther out, and then, that lower particles tend to stay closer.
1:48:19
All right.
1:48:22
So now, what we're going to do now is actually, we're going to do another run, but we're going to restart it from the last run.
1:48:34
So we're going to use the part on file as input.
1:48:42
So the last simulation only, we tickets to 12 hours. And now we're going to restart the simulation at Chelsea.
1:48:50
Ah.
1:48:52
So we're gonna cut a 12 here.
1:48:57
All right.
1:49:00
And then we're going to change.
1:49:08
First, 2, 12 kilometers.
1:49:18
And the other thing we have to do is go to our working Directory.
1:49:31
And we're gonna take that part on fire.
1:49:34
We're gonna rename it and we're gonna call it Har in it. you can actually name it whatever you want, but that's what we usually call it when we're using it as an input file.
1:49:46
Um.
1:49:50
OK, oh, they told us file, OK, they wanted us to us, I'll make a copy of that.
1:50:02
So we're going to carve a copy paradigm 14, 12.
1:50:12
OK.
1:50:15
And then we're gonna go, huh.
1:50:20
Advanced Configuration setup, We're going to go back to nine and now we're going to say Um, Yeah, we're going to keep so there's a couple of options if you want to use if you want high split to read a particle file and initialized from there.
1:50:45
And you can either say don't do it.
1:50:48
Like, don't even look for the file, Uh, you can say read the file once, like, look for it, if it has particles at the starting time, add them, but don't add any other particles that might be in there.
1:51:02
Or, you can say, OK, read the file, and, know, look at it each hour, and add particles.
1:51:10
And then, there's also like, read the file a tower, and basically replace the particles.
1:51:15
So, kind of get rid of the old particles, and just keep that, the new particles that have been added from the particle file, maybe, this is a good time, actually, to see what's actually in this part of the file.
1:51:30
Actually, we'll start the run first, and then I'll show you what's in the particle.
1:51:34
So, we're gonna save that, We're gonna keep that reading the file once at startup.
1:51:40
And we're going to keep this, because we're going to create another ... file from this friend.
1:51:46
Save.
1:51:48
And.
1:51:55
Right.
1:51:58
OK, so one thing is that this capability it depends what you're doing.
1:52:06
You're either like, I don't see how this could possibly be useful or you're like, wow, that's like what I've needed my whole life um.
1:52:15
Or maybe you're somewhere in-between. and you're like, maybe I could use this, but I'm not sure how.
1:52:19
So one way it can be used, and that's kind of what this example is getting at as if you're in like a really a forecasting scenario where you're in the middle of a volcanic eruption and you've been running model runs.
1:52:32
And then you have new emissions coming in and you need to model those.
1:52:37
But you want to keep modeling the emissions that you were modeling before, and you don't want to have to start your simulation back from the very beginning.
1:52:44
So basically this is a way of like incorporating the model runs that you've already done and then propagating the old emissions forward in time as well as any new emissions that are coming in.
1:52:57
So, it can be very useful in like a forecasting situation.
1:53:01
It can also be useful.
1:53:04
I think Mark was doing some of these, if you have a really, really long run.
1:53:09
And the model for some reason fails, like, I don't know, like your, your computer decided it had to reboot or something, right?
1:53:19
And if you are outputting your particle them files at a certain interval.
1:53:25
Then, instead of starting back from the very beginning of your really, really, really long run, you can just start from the last time the particle dump file was written.
1:53:36
Um, all right, let's look at what that did.
1:53:47
And let's look at the particle file again.
1:53:55
OK, so now you see, it's like, this is what is kind of the old emissions here, right, that went up to the six kilometers.
1:54:03
And then you can see, right here, the new admissions are coming in at 12 kilometers. So at the simulation dead was it carried forward. The older machines that kept them going.
1:54:15
Just like if you had done like a 24 hour run for those and it brought in the new emissions. and that started at the time, the last simulation ended.
1:54:25
And you get this very nice little structure. The volcanic emissions are fun because the plumes conform these really complex shapes high in the atmosphere.
1:54:36
And part of that is because the, the turbulent mixing up this high is not, as it's not as large as it is, like close in the boundary layer, right. Like, close to the surface, you don't get as much mixing.
1:54:50
And so that means that the plume is really getting kind of stretched and folded by this complex flow, and if firms really interesting structures.
1:55:01
Um, all right.
1:55:04
So, now, we're actually connecting you that, again, I think we're going to quit out of that.
1:55:16
And we're going to.
1:55:20
And change the start time.
1:55:31
OK, right, so we're going 12 hours later, Sarah, Sarah.
1:55:49
We also have to go back to our working directory.
1:56:00
And we have to.
1:56:04
Fine.
1:56:05
So, now, we're going to change, are going to copy that.
1:56:14
So, now, that part M phi, I think, is, can it become 15, rename it.
1:56:28
Uh.
1:56:32
1900.
1:56:46
And then we're going to make another copy.
1:56:52
And again, we're going to call it Par. And at this time.
1:57:06
Now, I just wanted to delete my other pertinent file, I guess.
1:57:11
Ah!
1:57:25
All right, so, now we can go ahead and run that again.
1:57:33
Um, so, let's look at our message file.
1:57:37
So, if you're wondering, it's actually opening your message file and reading it, me or emit times file, You can go into your message file, and you should be able to see something that says particle file opened for initial conditions, then the name of the particle file.
1:58:02
Now, though, it still doesn't really tell you. It will tell you this.
1:58:07
Even if it opened it, and decided, like, oh, there's no times that match, Like a very common error is that you use that, No, you tell it to look for a, for any particles that have the same as your simulations start time.
1:58:24
And then, you don't realize that your admit, time, start time is like the particles, then there have a slightly sorry, system at times as this particle dump.
1:58:36
Sometimes if the particles in your file don't match, your simulations, start time, then it'll just be like, well, there's no particles there that match the time that I need, then it won't actually read any N.
1:58:48
The same thing happens with the M times file, so that's why I was getting confused, so that's so it tells you that I opened it. So, that's the first step, right?
1:58:56
Um, but then you can see.
1:59:03
OK.
1:59:06
So at Nest.
1:59:11
Let's close that now and let's see what it did.
1:59:21
Alright, so now we see that we had emerson's up its effects kilometers, they went up to 12 Kilometers, they came back down to six kilometers.
1:59:28
I got this really nice looking plam shape, Um.
1:59:40
All right, so.
1:59:47
Excellent.
1:59:49
I'm gonna go through this, there is.
1:59:54
There is a utility and if you want to see what it's called, um, it's called power shift and it basically acts on the part on file.
2:00:07
So you can only use aes utility if you've created a par dump file, it doesn't work on the concentration files.
2:00:14
Um, so basically what it does is it reads the par dump file and then creates the ...
2:00:21
file, which is the same, it's the same file format.
2:00:24
right, so it creates another part on file with the locations of the particles shifted, or rotated.
2:00:33
And and the GUI, you can access this utility from the GUI.
2:00:41
You can also access it from the command line.
2:00:44
And these are kind of the things you can do with it.
2:00:46
You can rotate your particles, you can shift them, um, and so forth, and then after that's done, right, then you can take that part on file with the particle positions, adjusted and read it back into High split.
2:01:11
Alice can explain why you might want to do that, OK.
2:01:16
Well, I mean, this Yeah.
2:01:20
In this particular case, you know, they showed that and it's kind of hard to see here, I think that the plume was actually a little bit off of where you might it was observed.
2:01:33
And so, you know, the idea is, if you just wanted to do something pretty quickly, you can just rotate the plame around to where it was observed, and then keep running it forward, so that, it's, like, a little bit of that.
2:01:47
Um, you know, a very, very simple data assimilation, like, like, a, kind of.
2:01:52
So this could be useful a lot of, times, you know, if you are, unlike a situation, where you're doing, like, a forecasting situation and you're like the plume is a little bit off. And this happens a lot, because.
2:02:06
Winter actions are often, off and numerical weather prediction models.
2:02:10
I know it's not uncommon for them to be off by 10% or 15%, which, for a lot of applications, it's not a big deal, but for forecasting, you know, the transport of a plume.
2:02:27
And over time, that translates into quite a large distance.
2:02:31
So, the idea is that, OK, well, if you're going to, if you're going to run this forecast forward, you know, rather than these, if there's a little bit of an error in the plume and the wind direction, rather than that kind of like, keep propagating that forward, let's try to correct the position, and then run forward.
2:02:48
Again, that kind of what you had in mind, Mark, or did you have, Do you want to add anything?
2:02:53
No, no, that's perfect, thanks.
2:02:57
Um, Alright, so now let's, we're gonna go over a particle size distributions, a bit.
2:03:06
Uh, so, And then, we're going to get into the ... file, which we'll spend a bit of time on.
2:03:13
The particle size distribution, so this has kind of a nifty little capability and high split.
2:03:19
We're going to, remember, we were talking about, how we only had, like, four particle sizes and our particle size distribution.
2:03:26
Kinda, it's actually a very effective, more effective than you might think. But initially, it looks a little bit crude to just have, you know, for a particle sizes representing all the particle sizes that would be found in an ash cloud.
2:03:39
Or even, you know, just this also could apply to something like dust, or a smoke, where, clearly, it's not, it's not just one particle size.
2:03:51
But I will say that I think Mark touched on this earlier.
2:03:55
Sometimes, when you, you try to make things more realistic, and you think you have to do a little bit of a test to see how much improvement you're getting.
2:04:06
And so this is one thing you could try, could be like, well, as making a more realistic particle size distribution with a lot of particle sizes, how much does that gaining me, aye.
2:04:20
And an improvement in the forecast?
2:04:22
And a lot of times, it's not as much as you might think.
2:04:28
Alright, so we're gonna open the concentration grid And we're going to, oh yeah, we're gonna do that.
2:04:37
So what we're gonna do is add a deposition layer so that we can look at the deposition because that says, OK, so what I just said is more or less true.
2:04:48
But before I was talking about, you know, that trying to forecast the deposition or the Ashfall. And actually, when you're trying to forecast at this deposition or ash fall from a volcano, that is one case.
2:05:01
We're actually having a good particle size distribution, or maybe a more complicated one, can really make a difference.
2:05:12
All right, so we're going to do, Yeah.
2:05:19
Ah!
2:05:29
OK, and then are supposed to get rid of our parent files.
2:05:40
OK?
2:05:59
OK, so we're gonna run with, like, just the regular particle size distribution first.
2:06:08
Um.
2:06:20
It's gonna run pretty fast, OK?
2:06:27
Then, we're going to look, I think, at the, uh, we're going to look at the deposition.
2:06:37
So, look at all, um.
2:06:45
And I look at the ...
2:06:49
layer, I think, as we want, And, then we're gonna do the user sat, and.
2:07:03
Um, think this was an old, need to replace that graphic, I guess, with the newer, the New York Gallery one that I think, yeah, so they want it.
2:07:19
all the way down to it.
2:07:28
Mmm hmm.
2:07:32
Let's see, what did they use?
2:07:35
All right.
2:07:43
Ryan?
2:07:48
5000.
2:07:53
Else, you need to 100,000, not 10000, and the third entry.
2:07:57
Oh, thank you.
2:08:00
Oh, and then do I, Yeah, that's 50,000, right, yeah, I'm Sam ...
2:08:09
Now 10000.
2:08:19
All right.
2:08:24
Mmm hmm.
2:08:34
I'm Anna.
2:08:39
You just need to go to the next frame, say Think Oh, Right that always gets me when I was more than one frame, OK?
2:08:53
OK, So I saw probably need to do the Zoom.
2:09:11
All right, there we go.
2:09:14
Does this a deposition?
2:09:17
Yeah.
2:09:20
Alright, So, the idea was, now, what we're gonna do as.
2:09:28
Um.
2:09:32
We're going to modify our setup file.
2:09:36
We're going to find it in our working directory here To open it up, And I'm gonna put this, um.
2:09:50
And the P type equals five parameter in there.
2:09:56
So, ah.
2:10:07
OK, and then we're going to run the model again.
2:10:12
I freakouts who are methods file, let's see what that did.
2:10:19
So at the beginning of a message file, yeah, let's see.
2:10:24
Pat now instead of having just um for each pollutant.
2:10:31
So, we had four pollutants defined, 1, 2, 3, and four, and now each pollutant is associated with five different bins.
2:10:41
And so, um, do you remember our first pollutant?
2:10:47
Let's look at them, actually.
2:10:51
It had a, actually, that's on the other one.
2:10:57
So, it had a diameter of zero point six, and now, five fab, the fans, have diameters that range from zero point three to zero point nine.
2:11:13
Um, and then the same.
2:11:17
For the other one, I think the diameter was two and now, there is, I have different ones that range from one point to 2 2.5.
2:11:27
So, basically, it takes your kind of coarse resolution, particle size distribution, and creates a finer resolution, particle size distribution from that.
2:11:43
I mean, you could just, I don't think you can. I can't remember in the GUI and the GUI, there's a limit to how many particles sizes you can define.
2:11:52
But if you're, if you're just editing the Control file as a text file.
2:11:59
Um, Know, here.
2:12:03
So, in the control file here is one of the particles.
2:12:09
So, this tells you the four particles, and then there is 1, 2, 3, 4.
2:12:15
And then.
2:12:18
there, that other menu that tells you the particle size and the wet and dry deposition are here. So if you're creating the text file, you can put more than seven particle sizes in there.
2:12:29
You can put, as many as you want, I think I've seen at least one with at least 50 particle sizes in there.
2:12:35
Um, but, still, you know, that's, so, that said, there's, this one is a little bit of a shortcut for doing that by hand, Or, you wouldn't do it by hand, You're dealt with a script, right? All right. So, let's see what that did to our deposition.
2:13:00
And latest data, same one.
2:13:06
So, um, I should have saved the other one.
2:13:11
I think the Parent point, as you will see differences.
2:13:17
In this case, I don't think the differences are really large. It does depend a bit on the meteorological conditions and and so forth. And actually, if you're really doing deposition, we would probably include larger particle sizes in there as well beyond the 20 microns.
2:13:33
Which also brings me to, it's actually isn't in this section, but maybe I should add it.
2:13:41
two.
2:13:44
So, haislip does have, um, two different ways of calculating, gravitational settling, and this is detailed in.
2:14:03
So, I put this in here.
2:14:08
Yeah.
2:14:09
Um, So, this is detailed in the user guide.
2:14:14
And there's actually, if you want to read in more detail about it, I would really recommend this report by Richard Darr from Australian Weather and Climate Research.
2:14:27
They asked they use HubSpot as well.
2:14:29
And they developed this new settling algorithm for high split using what's called the gans our formulation.
2:14:38
Smokestack, it's it's more accurate for larger particle sizes. So if you're using particle sizes that are greater than like 10 or 20 microns, you probably want to use that one.
2:14:49
And the way you do that is you actually use a negative shape factor.
2:14:57
So, if you look at the High Split, let's go to the ... User Guide real quick.
2:15:02
So.
2:15:05
Um, So it's actually in the.
2:15:15
Sure.
2:15:27
Yeah.
2:15:27
So, if the shape is set to negative value, then, again, the calculation will be used.
2:15:33
And so, we can go ahead and do that.
2:15:37
The differences are, see if I have.
2:15:42
So this was a plot at once.
2:15:47
Yeah.
2:15:51
There's two that the Stokes formulation is the one that we just use by default. And then gans there is a new one.
2:16:00
There's also another one called Wilson, which is used by some other models, and so I hadn't thought Inherence file, but you can see that that says, the settling velocity versus particle size, assuming, um, no, it's a little bit different depending on density of the air.
2:16:19
Atmospheric pressure as well as these are for different density is of the particle, So 2500 kilograms per meter cube just generally the default of what we use for ash. That's another **** you can kind of tweak.
2:16:37
So you can see that no, there's really not that much different until you get no more like 20, at least, 20 micron particles. So if you're using, if you're modeling like PM 2.5, RPM 10, then you really don't have to worry about this too much.
2:16:53
Um.
2:16:58
But you can always go and you know, turn your shape factor out a negative and see if it makes any difference for you.
2:17:03
All right, so we're going to quit out of that and So wrong.
2:17:18
All right?
2:17:23
OK, so what next we're gonna go into the image timescale that there was actually one more thing.
2:17:29
Start a promise to you, and then brands for too fast today, Content have time to get to it. But I wanted to show you, what was in the particle is part, is protocol temp files.
2:17:42
And so, this is detailed in the highest Split User Guide.
2:17:46
There's this file formats thing, and then you can look, and these are, these, are you all to file formats that we use, right? So there's even one for the N times, such as what we're going to be talking about.
2:17:58
Um, and there's one for the ... file.
2:18:03
So, basically, the part on file contains all the information, Um, that high spike keeps track of for the computational particles.
2:18:14
So, there is like a header record that tells me like how many computational particles there are, and also, the number of pollutants per particle.
2:18:23
I think we got a little bit into that.
2:18:25
In the earlier section, Um, usually, it's just one.
2:18:29
But under, you know, if you're doing some simple chemical transformations, you could have more than one species on a computational particle, Um, and the time of the that doesn't fit.
2:18:43
This file is good for, even though the particles can have different times for them.
2:18:52
And then for each part a call, it keeps track of latitude, longitude, the height of the particle.
2:19:02
Sigma H and sigma W and sigma V are four paths.
2:19:07
Um, so paths have a width to them, so highest, but we need to now hereto backend like what is the width of the path at that point.
2:19:16
It also keeps track of the particle age, how long as the particle been in the simulation.
2:19:21
Um, it's distribution again as whether it's a particle or whether it's a path, and that includes whether it's a Gaussian puff, a top hat puff puff, and the vertical, and a particle in the p.u.f.
2:19:33
and the horizontal and a particle and the protocol And the pollutant name.
2:19:38
So that would be like the P 0 6 0 that pollutant ID, or maybe that's the particle, but that might be a number um so that you know it tells you if you have more than one pollutant in there, which one it is, um five.
2:19:56
The meteorological credit was on sale.
2:19:59
Remember how we talked about earlier, that high split can utilize more than one, um, meteorological grant?
2:20:08
So sometimes you might run with like, uh, a fine grid like, you know, like maybe a three kilometer waterfront close under the source, and then the simulation would be handed off to like a coarser meteorological model. I mean you can even do like three right.
2:20:26
Like you get up a really fine resolution handed off to something that covers like the CONUS or whatever model and then hand it off to like a global model like the GFS. And so that tells it that and then there's like a sort index which is kind of some bookkeeping.
2:20:44
So, that's basically what it is, and that's why these files, ..., can actually read them in, and you start the particles from, um, where it left off.
2:20:58
Now, we're going to go to him at times files.
2:21:02
All right, so, Um, Before we, we did kind of this emission scenario with the height of the emission changing over time.
2:21:14
I like stopping high split and restarting it, which would be kind of realistic if he were, add, you know, if it was like a forecasting situation, but if you're doing, like, a retrospective, fran, or like a study, no, that's probably not how you would do it.
2:21:31
In some cases, I would say, it might be, like I was doing a very long series of runs once and I wanted to break it up into more manageable chunks, and I did start doing that, where IOT is part of files.
2:21:43
And then start the next one with that, with the new Emerson's and the ... files. And.
2:21:50
But, you know that, for a short one like that, right, you just want to put all the emissions on at once.
2:21:58
So then you could use the image type of style.
2:22:02
So we're going to do this.
2:22:05
And you can do this a little bit through the GUI, but the cooling is really not setup to create complex and at times files.
2:22:11
So, if you want to use these, you really need to, you know, setup some kind of script that will write them.
2:22:22
So, let's see.
2:22:26
All right, we're gonna.
2:22:31
And then we're going to retrieve the control file, so we can start fresh.
2:22:42
Um.
2:22:46
I think we're asked. I can agree, trail the setup file.
2:22:57
All right, and now, the other thing we're going to do, so we're going to set all the emissions to zero.
2:23:06
And the reason we're going to do this because, if for some reason, we set up our times, we were going to set it up. So we want all the emissions to come from the image file.
2:23:18
And if we set it up wrong, we want to know about it.
2:23:23
Like, Wait, what could happen, Right? If you set it up to read from the ...
2:23:28
file, um, and some, it doesn't read it correctly and doesn't actually get the message from them at times file but I just, It's, it's still emitting stuff from, um, from those.
2:23:44
But if it doesn't do that, it will look here, and I'll try to get the emissions from here. And they'll run with those Emerson's, then it might be a little bit. You might not be completely obvious that it's not getting the emissions from the N times files, but if you set this to zero.
2:23:59
If it doesn't get emissions from the met Times file, that I just won't have any emissions, and that will be pretty obvious.
2:24:05
Um.
2:24:08
All right, so now we're going to.
2:24:15
And the advanced configuration setup, Um, Actually, I thank her Advanced File Edits Emissions file, and.
2:24:37
I'm just going to press Configure locations.
2:24:54
OK, so.
2:24:58
We're gonna keep all of that.
2:25:02
Save that.
2:25:07
Um.
2:25:15
OK?
2:25:23
Parent: Safe to Fail, OK. So, now this file is created. And let's look at it.
2:25:32
So, what we see is that it created.
2:25:40
Um.
2:25:43
This one, this tells you P.
2:25:50
It says, OK, so this is a mission cycle header. So that tells you what's in here.
2:25:56
So, this is the year, month, day, hour, duration.
2:26:03
So, for this, we only have one, let me call this an admissions cycle.
2:26:07
There's just one admission cycle in this file, and it's just reset the duration of it, and that's a really large number, because it doesn't really matter.
2:26:16
It just has to be larger, then, um, then the emission time, so if you start doing multiple Amazon cycles, and we'll get into that, then this number will said, it says something besides just a really large number.
2:26:32
Then eight is the number of records, so we can see 1, 2, 3, 4, 5, 6, 7, 8. There's eight records here.
2:26:39
Um, so you have to have that number here, because the way high split is working is that's going to read this line, and then it's going to allocate an array.
2:26:49
It has eight, as you know, it has a length of eight, At this number, isn't right.
2:26:56
Then it won't allocate the array length correctly and you'll get, like a segmentation fault or something, Nice like that.
2:27:04
Um, then each line, I'm going to actually bring up my PowerPoint here.
2:27:13
So, um, because this emissions file that we're creating, it has to things in it.
2:27:25
one is that, it has what's called stacked emissions.
2:27:29
So, um, the emissions are, it's at the line source, where are you going from?
2:27:39
You aren't in Mississippi, from 100 to 6, 100 meters to six kilometers, and the other thing that's going on is that you have four different particle sizes.
2:27:51
So, if you only had one particle size here, you'd actually just have two lines.
2:27:57
You would have, one, you would just have, and this line, and then this line, um, but since you have four particle sizes, you need aligned for each particle size.
2:28:11
And, again, that has to do with array allocation and highest split.
2:28:16
High split actually reads the Control File first.
2:28:20
So, let's open the Control File.
2:28:28
It will read the control file first.
2:28:31
And it'll note that there's two locations, And it will there's an array that's allocated for that. A number of starting locations.
2:28:39
Then it reads S for particle sizes. There's also an array that's allocated that corresponds to that.
2:28:45
So it allocates those arrays after it's the control file.
2:28:49
Then it goes and reads them at times file, and it expects the information in the Met Times file to have to correspond in, like, kind of size two, What it got from the Control File.
2:29:03
So it's going to expect this number, right here, the number of lines in your message cycle, should be equal to the number of locations in your control file, multiplied by the number of species you have to find.
2:29:19
So, that is why this is eight here, because you have well, it's eight here, because you want it to be eight, but in the control file, then you have to have two here, and four there. So, 2 times 4 is 8.
2:29:32
Um, if you don't have it set up that way, then it will read the current control file, it'll allocate those arrays, Then when it reads seem at times file, um, it will, you'll usually get some kind of failure because it won't be able to fit the data into the array that's allocated.
2:29:51
So that's one kind of, once you understand how it works, it's not too bad, but, if you just get into it, it can be a little frustrating because you're like, Why?
2:29:59
No? Why is my, why am I getting a segmentation fault?
2:30:04
So, this particular one.
2:30:07
It's gonna, this is emissions for the first cycle, or for the first 12 hours.
2:30:14
Right, so, so each line here, this header tells you what's in the line, and that's, again, the year that the emissions start.
2:30:22
The month that the Emerson starts today, the hour a minute even so it goes down to the minute, then the duration of the emission and the duration is hour with two and minute, So this one is emitting for 24 hours 0 minutes.
2:30:42
So, with the M at times file, you can specify emissions all the way down to like a resolution of a minute.
2:30:48
So, if I just wanted to less emissions, the last five minutes, it would look like that.
2:30:53
I wanted to last 10 minutes.
2:30:55
It would look like that an hour and 10 minutes would be like that.
2:31:01
Ah.
2:31:05
So, the latitude and longitude of the amison points, the height of the missing point and meters, this is the rate per hour. So all of these are emitting just one unit. Mass.
2:31:18
It says it's an area that you can specify if you want to. And I'll put it over an area.
2:31:24
Um, and it says that the heat which we don't use her volcanic eruptions, but you saw that we did use it for fire in order to use utilize one of the plume rise algorithms that would calculate plume rise from that heat input.
2:31:43
All right. So, we're going to edit this.
2:31:48
Yeah, so they want us to create, say, I don't, I think I did something wrong, because these are all supposed to be 12.
2:32:04
And we're going to change, we're going to create new admissions cycle, and each emission cycle is going to be 12 hours long, that's gonna look like that.
2:32:13
We're going to copy these.
2:32:22
This 1 starts 12 hours later.
2:32:26
And we wanted to go up to 12 kilometers.
2:32:40
Then, this one started out.
2:32:45
15, 4, 15, at zero, still last 12 hours.
2:32:51
We also have to change.
2:33:19
Alright, good, So, that, I can save that.
2:33:28
And.
2:33:35
Then I can test.
2:33:42
Oh. Well, the other thing I have to tell you, I think.
2:33:46
Ah!
2:33:50
I think I have to tell it, Let's see.
2:34:00
Right.
2:34:03
So.
2:34:04
You and the fact that CFTC file, you have to tell it to look for an admit times file.
2:34:10
Just like you had to tell it, to try to read a par dump file and if it doesn't see something, the setup telling it to look for the M times file and what the ... file is named at one to it.
2:34:23
Um.
2:34:29
I know I said that from the GUI Venue six: Thank you, Mark, Amazon cycling.
2:34:37
Ah!
2:34:49
Yes, I think.
2:34:53
I Think it's just called. Me just cut it to him at times.
2:34:58
Um.
2:35:12
OK?
2:35:22
And, then, there Oh, yeah. So, now, you can see that this E phi L equals M times, such said, Look for it.
2:35:31
Oh, nice.
2:35:32
The Queue cycle might have had set that as well for, Um, I don't think so, for this. I think that's only, I may be wrong, but I think that's only when you have a regular revisions, and you want to keep repeating those over and over again. I think you're right. All right, well, let's have fun.
2:35:56
Let's save that and we'll go ahead and run the model.
2:36:05
OK, so, a few points, Um.
2:36:12
A few are just examples to build on that, so this is an example.
2:36:18
Um, I have, OK, Excellent.
2:36:23
I'm just gonna go take this example first.
2:36:26
This is an example of kind of a more complex emission scenario. This was where.
2:36:33
Emissions were determined for a small eruption using an inversion algorithm, like TN Painless, describing an earlier section.
2:36:42
And then those emissions that were determined from the inversion algorithm.
2:36:48
They were determined as a function of time and height with a spatial resolution of a kilometer and a temporal resolution of an hour and so there was we put them back into him and him at times. I ought to run forward. And so you can see, for this one, the event was that, like, it was the event itself as 2.8 kilometers high.
2:37:08
And then the eruption reached, um, no, mostly reached around 9 or 10 kilometers.
2:37:14
So this was just a ran with one particle size.
2:37:19
So that's, it's a little easier to see when it's just one particle size of its multiple ones.
2:37:24
As you saw before, you just, each of these lines has to have, like, however many particle sizes there are, you have to repeat it.
2:37:32
Um, so one thing is each emissions cycle must have the same number of lines.
2:37:39
So you'll see, I have some lines here defined with zero emissions, So, you know, there were no emissions from like that 2.8 to like almost seven kilometers.
2:37:52
So no.
2:37:55
But I couldn't just I couldn't just make, uh, three lines, and this is from cycling and put 10 lines on this mission cycle.
2:38:06
So to keep the number of emission lines the fame, like I added, some lines with Sarah Omissions.
2:38:13
So in this particular one, that sounds like the first admissions cycle, just last one hour. And at the next slide, I'll say, you have some flexibility with this.
2:38:24
So to make this pretty readable and just, I liked how to organize this way, I made each emission cycle an hour, and each emission cycle contains the emissions, like the vertical emissions from the event to the that's the top height.
2:38:39
So I was looking at emissions from 2.8 to 3 13.8 kilometers, so just looking at this little portion right here.
2:38:47
No, this MSN is from 6.8 to 7.8 kilometers 1.547.
2:38:54
When it's doing the line Suarez that actually will use the Amazons from the bottom of the line source.
2:38:58
So, 1.5, 4 from 6.8 to 7.8, but then 0 from 7 zero point eight to 8.8, then three, E 9 from 8.8 to 9.8, and 1.2, a 9 from 9.8 to 10.8 and so forth, So that's how you can build.
2:39:16
No.
2:39:17
A vertical mass profile that's more realistic, and just a vertical line sorest.
2:39:23
I also here used an area for there volcanic emission, which is a lot of times a plume as can be pretty wide.
2:39:33
This is actually a pretty small area, because it was. It wasn't a very big eruption.
2:39:38
And just to emphasize, this could be the same one, but I just wanted to say, you could set this up differently, for instance.
2:39:50
If I wanted, I could make to our emission cycles, and in that case, no, this would be changed to a two right here.
2:39:56
And then, I would just get rid of this header. So, these guys would be, kind of together in the same emissions cycle.
2:40:02
Um, in that case, the control file would need 24 location lines, OK, because that's, that's a change to two, This would tends to 24.
2:40:10
You'd have to change our control file to have 24 location lines, and so forth.
2:40:16
Ah, so, that none part that that you set, when you, that's the number of particles released per emission cycle.
2:40:24
So, in this case, if you have none parse that to like 10000, it's going to release 10000 particles for this one, and then 10000 particles again here, and each emission cycle, it will really set number of particles.
2:40:38
If you then combine these two right, and then if you kept, and I'm part of the same, then you would have 20,000 particles covering all of these emerson's. So, there's, there's, like, a little details like that.
2:40:51
Um, Once you kind of, if you do this a lot, you get used to it, but when you first start out, they can kind of trip you up.
2:41:00
Um, yeah.
2:41:02
And I already went over there, so, that test, senescent cycle, the number of lines, and your Amazon cycle has to be equal to the number of location lines in your control file, multiplied by the number of species in your control file.
2:41:17
All right, let's see what are, OK, now look at the particle plot again.
2:41:26
and, OK, doesn't work.
2:41:31
Oh, I know, because I don't think I output the particles today.
2:41:39
I'm not sure you ran the model. I did, I think I did, but you know, what I did, was I, the particle files only output every 12 hours. And the first 12 hours, of course, it's gonna look like that, So, let's do this.
2:41:52
Um, actually, it took me repeating that I not, I really now run the model.
2:42:07
All right, let's see.
2:42:09
Let's look at the message file.
2:42:17
Hmm, So it said, Say.
2:42:24
It doesn't.
2:42:31
I've met times. That's good.
2:42:39
OK, it open the Amazon Fire masons updated.
2:42:44
So these are what you want to look for of updating the emissions.
2:42:48
All right.
2:42:49
Let's see if it works.
2:42:56
Oh, that's, it's way. More than one.
2:43:01
Hmm. Hmm. Hmm.
2:43:03
OK, I think, I need to hand configuration setup concentration.
2:43:11
Particle files, ah.
2:43:17
I have said, not simply, there should be more than one output.
2:43:28
May look at my setup fire.
2:43:33
And AMP and Cycle 12 and 12 so I said be saying more than if it's something with the plotting program.
2:43:47
You might check the set up to make sure that you're running for 36 hours.
2:43:51
Oh, that's what it is, Mark. Thank you.
2:43:56
I only have a total orion's him of 12 hours.
2:44:05
All right, so I was only getting the first 12 hours of adolescence.
2:44:20
All right.
2:44:26
OK, the other thing I'll mention is that maybe we can do this.
2:44:32
Ah.
2:44:34
There's now When you were.
2:44:41
Um.
2:44:45
When this area number is specified, ah.
2:44:50
It used to be that was always like a square area, but we added an option to make it like a I guess it would be mostly if you're doing this line, sorry.
2:44:59
I said, Be, like a cylindrical area, Um, in highest, but version 5.1.
2:45:07
Wow, um, What you can do is Gowdy sat at that CMT and This isn't, this isn't available from the GUI yet again, but you can do area equals one and then, it will give you Well, it will create a cylindrical area.
2:45:34
Ah!
2:45:36
So, maybe, we'll do that. I think we have time to do that.
2:45:41
Oh, Wow.
2:45:43
Still running?
2:45:56
Um.
2:45:59
So these are met Thomas files.
2:46:00
I mean we're focusing on the volcanic eruptions here but clearly, you know, they were used they're used for smoke emerson's, um, they can be used for dust emissions.
2:46:13
They could be, is for destinations, like if you are going to use your own test to miss an algorithm, um, then you could run your destination algorithm.
2:46:22
Get the emissions from that, and then put those emissions and one of the summit times files, Basically, for any application where you have your own emission algorithm to calculate the emissions, um, if you want a couple of that emission algorithm with high split, this one, most be a very, like, likely way to do that.
2:46:48
All right.
2:47:00
All right, So now, there should be, Yeah.
2:47:02
So that's the first 12 hours, The second 12 hours, the third 12 hours.
2:47:07
So, basically did the same thing we did at the very beginning.
2:47:11
We just did it with him at times file instead.
2:47:17
And, I guess we can go ahead and we'll save that with the area one.
2:47:27
Then, well, that's Open iron at times, file again.
2:47:33
Um.
2:47:41
Let's try to run till 2.5, E nine, and it's what we're going to put in there.
2:47:46
So, the area is here.
2:47:54
OK.
2:47:56
Think, I'll just do this for the first 12 hours.
2:48:21
OK, I think, got us to a 12 hour run.
2:48:28
Yeah.
2:48:44
Oh, someone's asking if we have a character limit. And the image times duration. Yes, actually, so you do.
2:48:53
Yeah, If you're looking at this, the field, it has to be, oh, you mean, well, for this first duration, Yes, I believe these are all they have to be exactly for me like this.
2:49:08
So, if you see, did I, um, you see here, even though it's happens here, there's erawan, um, then this duration has to be our hour minute minute.
2:49:20
Uh, so, I guess someone's checking the actual code, but I'm pretty sure that you do need to format it with, you can't, I don't think you can do over 20, I guess. So that limit would then be, like, 99 hours.
2:49:40
I think. You're right, Alice. And you know, if you need to do longer, you can just create more records.
2:49:46
Yeah, I mean, it's not really a limitation, right? Because, if you want, like, if you want the emission to last longer, you just do more emission cycles.
2:49:55
Um.
2:49:57
So if you had an admission that lasted 200 hours, you could do, like.
2:50:06
Instead of doing Y and emissions, like all with a mission of 200 hours, you could do like 10 emission cycles with the emissions at 20 hours eat.
2:50:17
Um.
2:50:24
Alright, so we're gonna see what that does to Iran.
2:50:35
OK, and I mean, it's not hugely obvious, but if you look at the point of it thoroughly, he said, as I can.
2:50:43
There, it was clearly, Yeah, more of an output.
2:50:48
Actually, if you want, Huh, OK, all right. I realize I have like 10 minutes, right, Mark?
2:50:56
Yes, actually, we actually aren't scheduled to end till 12 30, but you can end.
2:51:03
We started 15 minutes early, so I don't want to, like, you know, drag it on too long, but I'll take the 10 minutes to kind of play around with a few things.
2:51:13
So, one thing, let's see, that. So, let's look at the particle output, like every hour.
2:51:56
Alright, so that's just, like stepping through.
2:51:59
Senate says, the first hour average, or a section at an average, right? The particle positions are always like the instantaneous particle positions.
2:52:08
So, this is where they are, like, one hour after the run started, then out two hours, sorry.
2:52:16
So, you can kind of step through and see where they're going.
2:52:23
That was the end.
2:52:25
Um.
2:52:29
OK.
2:52:33
Wow.
2:52:34
Were there any forum questions that came up, that had that, were not addressed?
2:52:43
That are worth discussing.
2:52:53
As this is Fans, Hear Me?
2:52:55
Yeah, I can. Yeah, we mean, suddenly we were looking at the code to check the limit for the duration in the midterm file.
2:53:07
So I think we, we see, we see that you, Nicole, is not limited to four teachers.
2:53:16
Even when you read it, like, so, when you do the read of the file, because the reading stamen actually saw as the columns separated by character.
2:53:29
So, yes, it's not reading of formatted right.
2:53:34
Asked stamens, Yeah, OK, well, let's, we have, Yeah, so, and then, um, after reading in that duration column, then actually, it will calculate the, the hour by divided by dividing that Barrier hundred, and so that would be ours.
2:53:56
Yeah, OK, so you're saying that I could do.
2:54:02
Yeah, if you put the in light, 168 ...
2:54:07
sulfide ditches, I think that the code will read it and then calculate duration hour, 168.
2:54:19
Oh, for this one?
2:54:21
Like I do that, Yes, OK.
2:54:27
Let's try that.
2:54:34
So I guess I need to actually change that one, too.
2:54:41
Ah.
2:54:46
Alright, so, what we just did, is, we just, hopefully, uh.
2:54:54
We just made the first, the only one particle size.
2:54:58
I guess, one particle size of can emit 468, um, hours and zero minutes.
2:55:07
We're not, the runs aren't going to go on that long.
2:55:09
So, actually won't admit that long, but, but, will admit, as long as the Rent is going on, let's see what happens with that. So, we're going to close that.
2:55:21
Then we are going to go to Setup, and we're going to make our total run time.
2:55:28
Let's make it like, Let's make it 48 hours.
2:55:42
OK, Well.
2:56:04
Look at our message file.
2:56:24
OK, so let's, Um, OK.
2:56:37
I have too many.
2:56:41
Hold on a second.
2:56:43
Need to close out some of my window.
2:56:46
Ah.
2:56:49
OK, so, you'll see that the reason there's less particles at the beginning writers, because that was all in one admissions cycle, but we didn't actually increase our particle number. So, highest, but it's like trying to stretch the particles that we specified over the whole emission cycle, which is now much longer than it used to be.
2:57:09
That means it's going to Like, you know, have less particles at the beginning, because it's kind of saving them, so it'll have enough for the end, as well.
2:57:19
Um.
2:57:27
So, yeah, I mean, you can see that there are particles still being emitted.
2:57:32
So, it's clearly, yeah, it's working, OK, good, Good to know.
2:57:44
All right.
2:57:54
I guess I'm going to hand it back to Mark, and and if you have any further questions, you can put them in the forum, and we'll try to answer them.
2:58:05
OK, thank you very much, Alice. And also we have that question and answer period at the end of today's session for an hour, and be the way to answer to ask questions as well.
2:58:16
Alice, I have never heard a clearer explanation of the times files. I feel like I understand them.
2:58:25
Really appreciate the loneliest spent, you know, like 10 years, learning how to create them, so, thank you very much.
2:58:36
So, we're gonna go now into A A 60 minute break, which for us is sort of a lunch break, and we hope you can have lunch or dinner or some other break.
2:58:50
And so we'll come back at 115 Eastern time and are 15 minutes after the hour, an hour from now what time zone you're in.
2:59:01
See you then.
2:59:04
We are back from our break and we're going to go into the next And last section of the workshopped proper.
2:59:17
Before we go into a question and answer period extended question and answer period.
2:59:23
And the next section is on custom simulations and Chris doctor Chris Lochner will be taking us, through that, looking at a number of sort of specialized and very powerful aspects of high split that we haven't dealt with yet in the workshop, Or haven't dealt with yet, very much in the workshop.
2:59:48
So without further ado, let me turn the presentation duties over to Chris.
2:59:57
Yeah.
3:00:00
Get rid of them. Why?
3:00:03
webcam?
3:00:09
There we go, Chris, it looks good. I think we're seeing the screen you'd like us to see, OK, great, and we can hear you perfect, All right, take it away.
3:00:17
OK, so, this, this chapter is called Custom Simulations, but what it really is, is just off a whole bunch of other highest split options or routines that you can do, that don't really fit into any other of the chapter. So, it's just a whole bunch of different things put into this last chapter.
3:00:40
This first chapter called counting particle trajectories is fairly similar to why when we are working with trajectories and calculating the frequency plots based on a whole bunch of different frequency based on a whole bunch of different trajectories.
3:00:59
However, instead of the frequency plots based on different trajectories which are based on the mean wind, these are more or less frequency plots based on the turbulent motion of the atmosphere. So, we'll go over a little demonstration on how to set this up and do a simulation for this.
3:01:20
So first, they opened up High Split.
3:01:24
I'm starting something new. So I'm going to clean up the Working Directory.
3:01:30
Open it back up.
3:01:36
OK, so what we're What we're going to do is just count the particles, and then based on then later on, we'll create a figure to show. like, for example, you could see what percentage of the particles released at one that dislocation is later in time at the at another location.
3:02:00
OK, so to set this up, we're going to start with S cap T X control, and Cap Tec setups that we've been using.
3:02:08
So I'm going to go to setup and Retrieve, which we know is located in the tutorial.
3:02:20
Files.
3:02:23
And here it is cap tax control.
3:02:28
And the cap tech set up retrieve it, the same directory tutorial files called cap tech setup.
3:02:41
OK, and now, we're going to set up or run for only 19 hours.
3:02:55
And under the advanced concentration menu, menu number four, we're only going to emit 10000 particles, set the maximum, max par value to 20,000.
3:03:14
The advanced menu number eight, uh, we're going to, for the output units, typically what we've been doing, for the output units, it's it's mass per meter cube.
3:03:27
So if we we emit unit, if we are emitting something that it's in grams, the emissions rate is a grams per hour. The output has been in grams per meter cubed.
3:03:40
For this particular case, we are not going to add a divide by the volume of the grid cells, so it's just going to be, if you're emitting grahams, it's going to be the grams of Total Grams in that grid, So we're not, we're not dividing by the volume.
3:04:01
Or, in this particular case, where we're not even got it, have really units, the unit is just going to be particle.
3:04:08
So, we're going to count how many particles are in a particular concentration grid cell.
3:04:15
OK, so we'll hit, Save That.
3:04:22
Then go to the back to set up run.
3:04:28
The pollutant many unlooked the species.
3:04:31
And we're changing this too, one third of 10000.
3:04:38
So 3333.3, we are emitting for three hours.
3:04:43
So this emissions rate times three, it's 10000, which is the same amount of particles that we set in the advanced menu number for non power of 10000.
3:04:58
So, every Emitting 10000 particles, over three hours of time, that's what's shown here, OK? So, we're going to save all these menus and now run the model.
3:05:16
And while that's running, we're going to set the map label for, when we create a figure, that the labels will make sense, file, Edit, and it's Border Labels.
3:05:34
And the Outputs, we're not looking at air concentration, or look, just counting the particle, so we're gonna put particle number.
3:05:45
And, the.
3:05:48
The units are just particles, not a mass per volume, so, put particles there.
3:05:57
That's all.
3:05:59
OK, and, now, we're going to display the results.
3:06:08
And for the label, you're going to put particles.
3:06:15
And define our own user set.
3:06:16
So, remember to click that user set button, and then we're gonna from one thousand down to one.
3:06:41
And hit display. And let's see what pops up.
3:06:45
OK, so this figure is telling us how many particles are in each grid cell at the average over, at the average over this particular time period, where we emitted 10000 particles over three hours. So we know the lowest concentration where it would be one, or not concentration for particle count would be one at the very edge. So there's only one particle along the edge of of this kinda plume diagram here, and the maximum, very close to the Tracer Release location.
3:07:20
There are 150 particles, and then we can move on in time and see how.
3:07:28
Basically count the particles over time and where they were, the distribution of them, how they, how they were distributed.
3:07:37
Uh, and since we emitted 10000 particles, if we wanted to, we could change the concentration multiplier in our concentration display.
3:07:49
And two, instead of just showing particles, we could also change the units too, say percent of emissions.
3:07:59
So if we divided, for example, divided that divided, the results by, the amount of particles that we emitted, which are 10000, and then multiplied by 100 to make this into a percent, so that then this red kinda colored bar, instead of being one thousand particles.
3:08:19
But represent 10% of the particles that were emitted over three hours would be would be considered this kinda this red bar area.
3:08:32
The other thing I want to note is that this red box, you see in the middle of the plot, that's not really the red contour here in all of these plots that you've been seeing.
3:08:45
You've been seeing this red box that denotes the location of the maximum concentration at that particular time.
3:08:53
So, where this red box is, that's where we're seeing 24 particles in that, and the in the size of the box represents the size of the concentration grid cell.
3:09:05
Now, by default in it, we, this red box shows up in the GUI, There's no way to get rid of this breadbox but if you are able to create plots without this little red box, if you don't want it by using the command line.
3:09:23
So if not, going through the gateway.
3:09:24
There are many, many other options for plots and other things, if you, if you run all of these high split utilities' Through the command line.
3:09:35
OK, and yeah, that's, that was a quick section.
3:09:42
Uh, that's all I really wanted to show there.
3:09:47
So let's go on to dynamic Lagrangian sampling.
3:09:53
So, throughout this, we, we made these concentration domains in, averaged already provided, instantaneous. Values are over this, kinda this concentration demand.
3:10:07
That's kinda fixed in time, and over a certain resolution, but we also can get output over kind of moving.
3:10:18
Moving sampler, for example, Case, where, where, if you have aircraft that flies through a plume, we can kinda follow that aircraft and come up with estimates of the concentrations along that aircraft, with highest point along that moving path.
3:10:41
In addition, in not just an aircraft or moving car or whatever you want to do, you can also have this moving sample based on a passive sampler, or if you think of a Balloon, that is just transported horizontally with the wind.
3:11:00
You can do that with us, this similar technique that that I'm going to describe here.
3:11:08
OK, so, oh first I want to show.
3:11:17
So here's some data for that aircraft's flight, that we, that you've been working with a few times at that 914 meter meters above sea level. and this flight kind of made a few different paths through the plume.
3:11:37
Downwind of the cap to X tracer release.
3:11:40
And what we're going to do now is, we're, during this one, kinda set, the one pass, through, through the plume and, and, and measured concentrations shown here.
3:11:52
And we're going to kinda simulate that one part of the flight path.
3:12:02
OK.
3:12:05
So, I'm doing something new.
3:12:08
So, whenever I do something new, I like to just clean up my working directory.
3:12:17
And open up py split again.
3:12:21
So just to start off, I'm going to load these cons case control on ... case setup.
3:12:28
Hit the retrieve button, browse, move to tutorial.
3:12:38
Files and cons, case, control.
3:12:47
And then in the advance, configuration, set up concentration, retrieve, browse again in the tutorial files and the cast case setup dot TXT.
3:13:02
And now we're gonna check can maybe alter it.
3:13:06
Some of these parameters told a runtime we're going to run for 11 hours, OK, We're running with the war of 27, UW.
3:13:22
Not that.
3:13:23
The Pluton Deposition grids menu open up the Grid menu and how the Sampling Start Time?
3:13:33
September 26, 0 UTC.
3:13:36
And this first, kinda set this first zeros.
3:13:41
That tells you, if it's a zero there, that's going to give you average output.
3:13:46
here.
3:13:47
I'm going to put, want to give you out the output instantaneous output into into this file.
3:13:56
This, the next value, the one, we're not too concerned with this, because we're actually going to look at different we're not looking at going to look at the concentration up a file. We're going to look at a different text file that, that the data that we're going to configure, to have output.
3:14:18
And to kinda tell high split, too, kind of sample along a particular, whether it's a blue path, path, or car, or airplane or whatever.
3:14:37
To configure this, we're going to open up this new menu that we haven't done yet, if you go to Advanced.
3:14:44
And then file edit and dynamic sampling, This is going to create a new input file in the working directory.
3:14:54
And now we can set the number of samplers.
3:14:57
So we can have many different moving samplers during one simulation.
3:15:03
So, for example, if I open up this flight path this data again open it new tab.
3:15:12
So you can see that it made a pass up here another flight pass through the plume here and another one here.
3:15:22
So, so, we could kinda have three samplers simulate three samplers over different time periods going through this plume.
3:15:34
But right now we're only just just just doing one.
3:15:38
We're just going to do this one in there, and another reason we can't really well, I'll show you show you shortly.
3:15:45
But we can't have just one sampler for all three of these kinda transacts because we have to set the direction and the velocity of this moving sampler. We can't have the sampler change direction.
3:16:02
So if this, the airplane is moving straight at one altitude, we can we can do this dynamic sampler like I'm going to show you now, OK. So we're just configuring for one sampler.
3:16:19
It can take your sampler, and Sampler hashtag one, OK?
3:16:25
And the release location is, we're saying 48 or 40.86. This is kind of telling us the initial location of the sampler when when when we're starting, OK.
3:16:43
And we get this value right here from the flight data, that's the same as right here, 40.86 degrees latitude degrees longitude, OK. And it's 914 meters.
3:17:01
Mean sea level above mean sea level OK, so we're going to put that information here.
3:17:16
The next line shows the velocity vector so that the direction the of the aircraft and the velocity and meter per second or speed if per meter meters per second.
3:17:31
And you can calculate that based on the latitude longitude, or if the aircraft has that measurement.
3:17:41
That information directly is just use that you can calculate based on the latitude, longitude and the data.
3:17:46
The Sampler Release Time is basically the time that where we're starting, the moving moving platform is starting at this particular location, which is at that, which us, at this first time, 926 to 48, OK.
3:18:07
So we're gonna put that and here.
3:18:10
So in 19 83, September 26th, Power two, minute 48, and then now the sampling start Time.
3:18:23
This could be later than the sampling release time. If the measurements started after the, what we're saying is the release location. However at this point that for this particular example, we're starting this release location at the same time as the first sample.
3:18:45
So we can just, we're just having the sampling start time be the same as the sampler release time, and for the averaging time, and the right output interval for reporting for six minutes, OK.
3:19:01
And it is going to write this the concentrations along this flight path to an output file called lag out dot TXT.
3:19:14
We're going to save this Return to Previous menu, Hit Save File, and now if I look at my Working Directory.
3:19:25
Highest split working, I have this file called Lag Set and this is what it looks like. That's basically what we just configured, OK.
3:19:35
There's one Dynamic Sampler: Starting Location, Lat-lon Heit remember high splits, typically, the height is above ground level, but in that advanced configuration, that what the, what, we loaded before we had in there, is that switch to switch from above ground level to mean sea level, the velocity based on direction, wind speed, the sampling start time, the measurement start time, and then the averaging and output time interval of six minutes and the output file.
3:20:16
Now if we wanted to sample not along.
3:20:23
A set path, if we set this, is this line, this 290, 50 to 0 and 0? What this would do is sample along.
3:20:38
Along Kinda based on the wind direction and the mean wind direction. So if you release a balloon, wherever that balloon travels, it will kind of make a measurement along that path.
3:20:57
But in terms of vertical motion, both with the balloon and this particular case with it with the flight path, it's, it's based on ISO barack.
3:21:09
Vertical motion, OK, so, it's X out of there, and now we're going to run.
3:21:53
one other thing I wanted wanted to show you is that, notice, this, this flight is at 914 meter, so we have a concentration grid.
3:22:06
There's 1 from 0 to surface to 800 meters and another one from 800 to one thousand meters, so, since it's going to get information from that in-between the 800 to one thousand meters, another thing to keep in mind, you notice these are our datum information from flights shown here.
3:22:33
It has the latitude and longitude, but it doesn't have information on the height.
3:22:40
We just keep track of that based in the filename or in the header here.
3:22:45
So when you're comparing data, actual data with high split output, you need to be careful with making sure the data that you are using is the same height level as what you're using from in your high split output.
3:23:06
Make sure your, your concentration levels are equivalent to an represents, your measurement heights.
3:23:17
OK, so let me get out of here, the run is finished, and if we look in our working directory, and now we have this output output called Lag Out, and this is what we have for our results. Get date, latitude, longitude, these heights, the output heights are above ground level. So just to keep you on your toes, we had the inputs and above mean sea level, but these outputs are above ground level, and here are the concentrations.
3:23:50
Keep in mind, These are our grams per meter cubed, whereas the aircraft over here, PICO grams per meter cubed. But now we can compare these with with with the aircraft flight path.
3:24:04
Oh, another thing is the time is a little bit different here.
3:24:09
This time, the highest concentration is an hour, 3, 6 minutes, Uh, this time it is in-between this time period, our 3 and 6 minutes, two for the next six hour or so, 6 to 12.
3:24:32
However, in this output, the time is considered that kinda the after, ah, that, that the measurement took place.
3:24:42
So, this measurement at you said, you think it's 36, is the equivalent it, too, but our 3 and 12, which is also the peak measurement for this, for this particular simulation.
3:24:57
So, the peaks match, they're at the same the same time, but it's a little tricky here.
3:25:06
No, just average concentration from 3 6 to 3 12 is here for, here. It shows up as, that is 3 12, So, little, that another little something to keep you on your toes a little confusing.
3:25:23
OK, on to the next section.
3:25:31
OK, so, next, this talks a little bit about unit conversion and air quality.
3:25:41
Gasses are typically measured in Value Misc mixing ratio. So, PBS, PPB, PPT.
3:25:50
Aerosols are in what we've been typically been using grams per meter cubed or micrograms per meter cube. PICO grams per meter meter cubed.
3:26:01
But somebody's measurements are, format a different differently, and volume mixing ratios, so we need to do some unit conversion to kinda compare with, with those types of measurements. And kinda, if you're unfamiliar with this unit conversion process, you can read up on this page. I'm not gonna go over too in-depth. But in general, High Split has an option, too.
3:26:35
Have the output units BN at its max mass make mixing ratio, based on this Unit Conversions.
3:26:43
So it'd be a, for example, grams per PM CH per grams per there, right? So then then you can multiply by the molecular weight of air over the molecular weight of the pluton you're interested, and to get the volume mixing ratio.
3:27:08
OK, so let's go over a little example of how to do that. OK, so I'm doing something new again.
3:27:15
So I'm going to clean up my Working Directory.
3:27:20
Open up, I split.
3:27:26
And set up, run, retrieve khans case control.
3:27:34
Oh, one thing also thing I want to know is that see this red text here?
3:27:41
Very important, OK.
3:27:44
We just created this lag CET dot config, which is telling high split 2, two, you know, you come up with these concentrations along this flight path.
3:27:58
Now, we're not doing that anymore, so we want to make sure for sure that this is not in the working directory, otherwise.
3:28:07
It's it's not going to do what we're going to expect. It's going to calculate the concentrations along that same flight path, which is not what we're doing now, OK.
3:28:15
So, this is an optional input file that we have for the dynamic sampling, but if it's not in the working directory, it won't do this dynamics sampling so we're not doing this dynamic sampling. So, we do not want in the working directory.
3:28:32
You can either delete it, or if you clean up your working directory, that would, that will take care of it for you as well, OK?
3:28:41
So, now, let's find Count Case control, which is back in Tutorials, our Tutorial Files.
3:28:53
Cons Case control, that text.
3:28:57
Then Advanced Configuration Setup Concentration can retrieve.
3:29:05
Cost case setup.
3:29:11
OK, and then the only thing I need to do to prevent it from out pudding and mass per meter cubed is: Tell High Split, too.
3:29:28
Divide up, I'll put mass by air density to sum the mixing ratio, OK.
3:29:34
So, B, Now the outputs will be Mass per mass.
3:29:45
OK, and then just run.
3:29:52
And while it's running, I am going to configure the labels for our plots that we're going to make.
3:30:06
Simulating P P M C H concentrations.
3:30:13
And the units.
3:30:16
Fluid liters per.
3:30:18
We're.
3:30:40
And now that it's done, we're going to display it Israel results under display concentration Contours.
3:30:49
And let's look at that 800, The one thousand meter level, which is the same level as that aircraft flight.
3:30:58
And concept, we're going to put in a concentration multiplier.
3:31:03
What this is, is that molecular weight ratio, the molecular layer, the molecular weight of air, ...
3:31:15
molecular weight of PM CH, uh, times the conversion for to get.
3:31:24
Leaders over leaders.
3:31:31
And switches D point three E to the 10th.
3:31:39
And we're going to use a user set.
3:31:46
Value is here.
3:31:57
OK.
3:32:01
And this is what we get.
3:32:04
So now, these these same kind of simulation is now in volume mixing ratio instead of mass per meter cubed.
3:32:17
OK, and then if you want, we can add those flight observations, which are also in these same units.
3:32:27
So these are located in tutorial cap tax.
3:32:35
Data rich.
3:32:40
Let's display them, and here they are.
3:32:43
So it, it's hard to say you have to really look at the numbers, but you can see that it's hard to see those numbers.
3:32:51
They're pretty small, but the peak is measurement is pretty much exactly kind of right in line with where the peak is in the simulated plume.
3:33:05
Uh, off to the side, where there's just these plus signs, that's a measurement location, where there weren't any measured value, so it is zero concentrations were measured there.
3:33:18
So, I split did, it looks like did a very, very good job for this particular case, for that, that flight, that one flight pass.
3:33:31
So, let's get out of here, and move on.
3:33:40
OK, so we've been looking at it throughout this week, a lot of about this cap Texts to experiment, which was a Regional Tracery Lease Experiment now. We're going to look at something very different, very high resolution localized, tracer experiment.
3:34:02
Very big differences where for the CAHPS X the measurements where F O, about 200, 300 kilometers downwind of their tracery Elise location, here the most distant measurement was, uh, 1.6 kilometers downwind. So a very, very different different type of case.
3:34:26
And if we look at, here's uh, highest what simulated plume in Google Earth?
3:34:36
And it's a little hard to see.
3:34:37
But you see these little these kinda circles around where the Tracer was released.
3:34:46
This, this is a location where there have been Tracer Release experiments for it, for decades and dislocation, is outfitted with measurement sites around these these kinda circular arcs. And some of these arcs also have towers to make tower based measurements the outermost.
3:35:04
Arc over here Circle is 1.6 kilometers downwind, away from from tresor at least location, in the center here.
3:35:19
Now to do this, I kind of simulate very, very high resolution simulations in it, there are limitations and some things to really, really think about.
3:35:33
So in high splits, the minimum time step is one minute, OK.
3:35:41
And the time step in high split is calculated based on the maximum wind velocity in your meteorological model.
3:35:53
And the resolution of either through your concentration grid domain or your meteorological grid demand, whichever one of those is a higher resolution, OK.
3:36:07
And then it uses the maximum wind speed and the highest resolution, whatever that resolution is for either mere, meteorology, or concentration, grit.
3:36:17
Too, calculate how far a particle would travel, or how long it would take a particle to travel three quarters of a grid cell.
3:36:28
And the reason, this is, is because high split does not want a particle to kinda skip over a grid cell in one time step.
3:36:38
However, for very, very high resolution simulations, it is possible that that can occur.
3:36:44
So, what High Split does to kinda remedy this, which isn't perfect, but the, it, if that one minute.
3:36:54
If, if that time-stamp that textbook calculates is less than one minute, it sets that time step still the one minute.
3:37:01
But it, but when high school, it gets to that kinda one minute time period.
3:37:06
Which is when high split takes all the particles and sums them and grids them into the concentration grid to say for this concentration output file, it will linearly kind of interpolate.
3:37:23
Where are those particles, which grid cells, they were, maybe in those increments, incremental time periods from time, say, time T equals zero to T, equals that one minute, minimum time step.
3:37:40
OK, see if there's something else I want to say about that.
3:37:45
Oh, yeah, So currently, I want to say, this minimum time step is one minute.
3:37:53
Our high split team. It is kind of on our to-do list to decrease this time step within the model to one second, but that's still, not there yet, but it's on our to-do list.
3:38:04
So, should be, should be coming soon, OK. So now let's configure for this very high resolution Tracer experiment that that occurred.
3:38:16
So again, doing something now, somebody clean up my working Directory.
3:38:23
Open up High Split.
3:38:29
Concentration setup.
3:38:33
I'm gonna cheat a little bit. Since I know.
3:38:38
Oh, where the, the control file is, it's in the sea tutorial and then this one is in that sub directory called Sage.
3:38:55
It's Sage five control dot text.
3:38:59
I'm going to check to make sure everything looks good, but I do OK. there. We go.
3:39:04
OK, so Tracer released at 2013 October 18th, 19 UTC we're going to run for three hours we're using a warf simulation without one kilometer horizontal resolution.
3:39:22
Starting location looks good 10 meters above ground level.
3:39:36
And we see the center of the concentration domain.
3:39:40
Shown here 43.59 -112, very high resolution concentration, to man zero point zero zero one, by zero point zero zero one, and a span of over zero point two, zero point two.
3:39:58
Vertical levels were going one, well from the surface to 25 meters.
3:40:04
Remember the Tracer release set at 10 meters, the surface concentrations I think were, I don't remember if they're two meters or 10 meters, but we, this.
3:40:20
This, this height of 25 meters should be representative, we hope should be representative of the surface, observations, large enough to where, where it won't under sample, based on the number of particles that we're going to choose.
3:40:41
But, also, should be representative of where surface observations are made. Sampling, start time is shown here, sampling stop time, and we are outputting average concentrations shown by the 0 0 here.
3:40:57
Every 10 minutes, OK, So, zero hours and 10 minutes.
3:41:05
OK, now let's look at the pollutants, pluton releasing SF six at emissions Rate of 3708 grams for me, grams per hour for 2.5 hours, but they're really stuck. Start time, shown here.
3:41:29
Um, I want to show you one more thing.
3:41:34
Because I saw this, I saw this, something configured like this, yesterday, but I don't think anyone explain this to you yet.
3:41:48
When I saw in one of, uh, one of the control files is configured like, something like, We know, you should know by now, that the sampling start time, where by default, it's all zeros.
3:42:04
The sampling start, it's the same time is the starting time of the simulation.
3:42:10
Now, it can also do something like, just put a one there.
3:42:15
What this means is now the sampling start time, because we were setting the year and months to zero, or just a month to zero, this is telling us that this sampling start time is one hour after the starting time, OK.
3:42:36
So, that's Canada.
3:42:41
Something to think about, you can either say, start the sample, and at certain time, after or after the start time, or just put in the actual, actual value.
3:42:54
So.
3:42:57
Another trick that maybe you will find useful.
3:43:03
Maybe not.
3:43:07
OK, so now let's configure this, the go to Advanced Configuration Setup, concentration.
3:43:16
And we're going to release 50,000 particles.
3:43:25
Now, you might be thinking 50,000 particles, But this was actually, this was enough for the cap tax experiment, which had a much larger resolution concentration demand.
3:43:38
But this, this simulation is a much higher resolution, both on the horizontal and the vertical, However, since we were only simulating for three hours, this, this should be sufficient for this very short period of time. Remember, the ... experiment was a much regional location.
3:43:56
So if we had, if we simulated with this type of high resolution concentration demand for cap tax, over a much larger area with only $50,000, If you move a few kilometers downwind, then we're going to under sampled. The grid cells where the measurement locations are which are 200, 300 kilometers away from from the sampler. But here we are only looking at measurement locations, within 1.6 kilometers from from the sampler, or from the release location. So this should be sufficient.
3:44:37
OK, now we're going to Advanced menu Number 10, and we're going to divide output mass by air density, OK, so we're going to get these two, the volume mixing ratios, again, instead of instead of the mass concentrations.
3:44:56
And now, we run, OK.
3:45:11
OK, while it's running, I'm going to set up my labels for my figure later. So I got a file edit.
3:45:20
Border labels.
3:45:24
Sagebrush experiment hashtag five.
3:45:28
Similarly, simulating air concentrations between because it's going to say like one level between different layers.
3:45:39
You'll see in the figure, units, parts per trillion.
3:45:49
So, since we're doing volume mixing ratio, remember, we're going to have to have a conversion factor.
3:45:55
We're making these figures, which are going to be molecular weight of air over a molecular weight of SF six, Then, also, 10 to the ninth, times 10 to the ninth for the PPT.
3:46:14
Thank.
3:46:22
Here is a description here of, uh, this conversion factor, OK, Molecular weight of air, 29, molecular weight of SF 646 times that of the nice stuff: convert to PPT.
3:46:45
OK, so now, let's convert to data.
3:46:57
So it's not this old conversion factor that we use the other day, it's 1.986.
3:47:06
Eat it.
3:47:11
The measurement file.
3:47:13
It's located in tutorial Sage, Sage five, mezze dot TXT.
3:47:21
And let's call this sage fat dot T X T.
3:47:28
And create data.
3:47:31
OK, now let's compute our statistics.
3:47:37
It's trying to make it bigger.
3:47:40
There we go.
3:47:45
OK, and what we see is the average measured concentration.
3:47:50
Yes.
3:47:52
PBT, so 1800 PBT calculate it a little lower, 1200 about correlation coefficient of zero point eight.
3:48:06
Uh, like the 95th percentile of the measurements, 11,000, whereas the calculate is 7000, so we have a low bias for the high values.
3:48:22
But when we look at the 50th percentile, the measured has 64 PPT simulated as 120 PPT.
3:48:30
So we got a high bias for kinda this 50th percentile, but the rank is very good, it's 3.26, remember the rank kind of goes from 0 to 4, but for us, is the best, OK.
3:48:47
Now, we can look at scatterplot.
3:48:56
And, we see that kinda what, what we're showing in the statistics, We got to little, low bias at this upper level.
3:49:08
Little high bias, down here, OK.
3:49:17
And I should have showed you what I clicked on.
3:49:22
So, when you create data, it also creates this, this file called Data.
3:49:30
A, OK, and this data, A, has the number of measurements us, number of site, basically the latitude longitude, month day, our measurement location, and calculated calculated value.
3:49:48
So you also have this file if you want to, uh, do your other, I don't know, other.
3:49:57
Plots are calculations or evaluation with your own. Utility, is your own programs import it into something else. You have this file here with direct direct comparisons with the measurements and calculated values. So, that could be handy.
3:50:19
OK, and now let's look at, uh, make it pretty picture.
3:50:25
So go to Concentration, Display, Concentration, Contours.
3:50:32
Let's make some rings, like five of them. The distance of one kilometer.
3:50:39
And remember that concentration multiplier for converting from mass over mass to Volume Mixing Ratio and PPT, 1.986, E eight.
3:50:55
Label's PPT and I user set and come up with my own values here or Copy what's done here starting at 10000.
3:51:31
OK, and also in here, Is that the data, the measurements.
3:51:38
So, let's let me zoom all the way in and now display, and let's see what we get.
3:51:54
OK, this is what we got. It's hard to see the measurements The I don't know.
3:52:00
Maybe I should take that out.
3:52:04
They're there on top, on top of these figures. so you can see go move on in time.
3:52:11
And you can see where the plume is.
3:52:15
Actually, I'm. I am going to take this out. You can't.
3:52:19
I'll see it better.
3:52:35
OK, so this this is our plume that you can see here.
3:52:47
OK, so let's move on.
3:52:57
Now we can also have a concentration grid and polar co-ordinates, so let let, let me go over how to do that.
3:53:09
So I'm going to start by loading Saige five control ah.
3:53:20
Stage five setup files.
3:53:28
Then go to setup run, pollutant deposition Grids.
3:53:33
Click on grid and where it's spacing.
3:53:39
Instead of putting in the degrees spacing we're gonna put in an angle and spacing and kilometers.
3:53:50
So the spacing is going to be one angle or one degree by one kilometer zero point one kilometer.
3:54:03
And the span is going to be, by definition, it has to be 360 degrees.
3:54:12
And then five kilometers out from the center and let us name the output file P dump.
3:54:23
OK, so where it says Spacing is, a little confusing, because it says, Put it in degrees for the Spacing and Span. But for polar Grid's, it's not, it's an angle.
3:54:35
And then kilometers and always, 360 degrees, and then the kilometers for the span here.
3:54:48
But then we also need to update the setup that config, which is the Configuration, Setup Concentration, menu number eight.
3:54:57
And click on Polar Grid. So this kinda tells it has arc in the distance.
3:55:02
So that's why we put the angle, and then distance in kilometers, and you have to press the update button because there could be multiple grids, multiple concentration grids. So, where are you?
3:55:16
So, if we had three concentration grids, for example, and want two of the first to be irregular Box Lat-lon box, it would be we we would change and only want the third concentration demand we would change that, the three, But right now we're only using one concentration grid and one, which is a polar grant, OK? So, you have to hit Update.
3:55:46
So, it changed the concentration grid for want to have a CPAC of three, which is the mess of which is the polar grid.
3:55:55
Hit Save, and now run.
3:57:26
So for polar grids or polar grid outputs, not all of our utilities are compatible with, uh, kind of post-processing with with polar grids, but some of them are, for example, The contour is the con to ASCII's.
3:57:47
I'm not sure all of them, I mean, which ones are, and which ones aren't, but here, we're going to here's the contour.
3:57:57
Oh, with polar grits.
3:58:01
OK, and let's change this step, the zero point one, the kind of resolution, and that's all we need to do, Display.
3:58:33
Notice, OK.
3:58:34
So here, here is our figure here, and we can move on in time to see that plume change evolve over time.
3:58:49
And as I mentioned, previously, like all of these utility is if you don't like these colors, and want more control over them, you will have to become more familiar running high split from the command line, where you where you will have more options to kind of have a little more control with what, What year?
3:59:10
Oh, what's being displayed what you're producing.
3:59:16
But the GUI is a great place to start.
3:59:19
Great place to learn how to do this, OK. Now, in the last section.
3:59:25
This, I'm just going to talk about, there's really not much to show.
3:59:28
A lot of it give, That's kind of here, you've just talked a little bit about.
3:59:38
So, this section is called Configuring stilt Options and High Split.
3:59:42
So, I don't know how many years ago quite awhile ago. I guess it says this paper with 2003. So since since then.
3:59:54
This group got an old version of high splits.
3:59:58
And kind of put in a few other special model options for the sole purpose of estimating greenhouse gas emissions, using a combination of high split, and then version tool, version modeling tool.
4:00:21
And, and that, the model is kinda diverse for a little bit. So this, this old version of high split with these other options. We're now going to be called stilt model, and.
4:00:35
The high split that was just constantly keeping developed and maintained at NOAA RL is still at highest point, But it didn't It didn't have these other options that were instill. But then Stilt model.
4:00:50
It was there but it wasn't being updated.
4:00:54
And all the bug fixes and updates and improvements that us at IRL as have performed in high split, it wasn't transferred into the stilt model.
4:01:06
So that the, kind of the foundation of stilt was becoming out out of date.
4:01:13
OK, so it was, Yeah.
4:01:16
So what we did a few years ago, as we took these other options that were in Stilt and not in High Split, and put them in High split. So now, just kind of a more robust problem.
4:01:29
Robust program for people to want to use the same configuration for stilt, in instead of using this old, outdated, still model that's based on this old version of ..., but now they can just use the update I split, which is continually updated and maintained and then run their conversion program based on high split output.
4:01:55
Now, the big, the main thing.
4:02:01
Does this still option in this code, is it out we can output a footprint? And what a footprint is?
4:02:10
Is it, it tells you kind of the sensitivity of a particular measurement is, two, the surface or a possible emission source, OK, This may sound a little confusing, but To calculate a footprint in high split, let's say we want to make a backwards simulation from a measurement location at the ground.
4:02:37
Make it backward, dispersion simulation.
4:02:41
And when that part of a particle is close to the surface, typically, it's set to, well, if it's in the lower half of the planetary boundary layer, we say then that particle is sensitive to the surface, insensitive to a possible emission source at that location.
4:03:04
So, therefore, if emissions takes place at that location, a portion of that, those emissions, or then kind of part of the measurement location, where the measurement was, what was taking place.
4:03:19
So we calculate the footprint, which is a large area downwind of, of measurement location, and usually it's multiple measurements, and then users can combine those footprints.
4:03:34
And then within their inversion modeling system, they can estimate kind of greenhouse gas emissions based on this.
4:03:44
This approach based on the emissions and based on the based on the measurements, should be upwind.
4:03:50
Not done yet.
4:03:52
Yes, sorry. Sorry.
4:03:55
OK.
4:03:56
So in order to kinda tell high split to calculate this footprint, you need to set, I can equals to eight here, so that's located in Advanced Configuration, Concentration, Conversion Module, and then there's stilt Load option.
4:04:16
OK, so it does a couple of things. It says that Layer two, like half of PBL height.
4:04:24
Um, so it basically telling you that we're sensitive to the lower, lower half of the PBL and that Mixing Ratio option to this, this Divide output, mass spec air density that we, that we've been doing here.
4:04:39
OK, and if, if you don't have a set up that config, if you are, if you don't have these options sets in the setup, that config, these are the options that are the default only when I cam equals to equal equal to eight.
4:05:00
You can run, hi split with ICANN. equal to eight to different values of these.
4:05:06
Just just the default we have set set to these these values.
4:05:12
No, real particular reason why they are better options for when I can equal eight, but it's more historically that stilt users ran with these options.
4:05:27
So we just, when we have ICAP equals eight.
4:05:31
We have these be the default but you can run these values with I can equals or something not to it.
4:05:38
You can run with the set to something else with when I can equals to it, OK. And then most of these were mentioned previously.
4:05:50
What, what this says.
4:05:51
The IT SP equals two. That's distilled the version dispersion scheme.
4:05:59
Which was mentioned briefly, I think yesterday.
4:06:02
And that's set in the you think it's huge.
4:06:08
Aye.
4:06:09
Don't know if I remember where exactly it's set.
4:06:16
Um.
4:06:27
I'll just hit this button, this will tell me right?
4:06:35
Here, Menu three, there we go.
4:06:39
Particle Dispersion Algorithm.
4:06:41
So, we typically, in here, we've been running with high support, except in this particular test case that you did yesterday, you ran with the still option.
4:06:51
The main differences difference between the high split and the stealth dispersion scheme is the stilt Scheme has a little stronger lid at the top of the Boundary Layer.
4:07:02
So, there's less, uh, venting of pollutants up between the free troposphere and ...
4:07:10
with a stilt option than that and that high split option.
4:07:14
OK, Kavi LTE option, that was mentioned yesterday to a different turbulence scheme, modified Richardson number for estimating and another match way to match. Estimate the PBL height, the mixed layer height.
4:07:30
Also mentioned yesterday, as well as this, the, the time and space faring Lagrangian Timescale which can be set with the scales equals . And then this grail convective, mixing.
4:07:44
This skill only be used.
4:07:48
If you run with Warf: With the grill convective scheme and output convective flux this and then, it will use the same kind of convective, mixing algorithm that is used in the wharf model with the grail scheme in high split.
4:08:03
As as done in wharf.
4:08:07
Uh, think that's mainly all they want.
4:08:10
It wanted to say, it's just, high split is a tool that you can get this footprint.
4:08:17
And also, output is a file called particle dot dat.
4:08:24
Here ...
4:08:24
dot that and in particle stilt Dat Dat which are basically the same kind of the same Information button, is it a different way to format the file?
4:08:37
And then this is used into different people's inversions software systems that they have, which is not part of the highest modeling systems. So we're not going to go over that here.
4:08:52
So I think I'll I'll leave it there.
4:08:56
Mark.
4:08:57
OK, Chris. Thank you very much.
4:09:03
So, now what we're going to do, I'll take back the presentation.
4:09:11
Yeah.
4:09:13
Let's see. Let me not show that.
4:09:18
Give me one second.
4:09:21
Me just, for a second here, just not show anything.
4:09:26
What we're going to do is go into a, I guess, a almost a 20 minute break and we'll come back at at 2 45 Eastern or quarter quarter to the hour in your time zone.
4:09:40
And we're going to have a one hour question and answer session.
4:09:45
And you'll ask your questions in the in the goto Webinar question box and we will try to get to every single one if we can. And if we can't get to every question, then we'll try to answer you off offline.
4:10:00
So, you can go ahead and start putting questions questions in there if you want, Or take a little break, and come back and put your questions in. And then when we do come back, the ... will be here, and we will.
4:10:14
Like I said, we'll do our best to answer the questions as we can.
4:10:17
All right, So we'll see you in, I guess, 18 or 19 minutes. So bye for now.
4:10:35
OK, we are back from our break, and we're about to start the question and answer period, and I see that there are quite a few really excellent questions in the question box.
4:10:53
And we'll start with the first one, and I'm going going to invite my colleagues who are on the line as well.
4:11:00
two, just jump in if they would like to try to answer a question there.
4:11:06
The first question has to do with computational resources.
4:11:10
And it's it says, I noted that during longer runs I split doesn't seem to utilize a lot of my PC's resources.
4:11:18
What is usually the bottleneck capping the calculation speed?
4:11:22
And what can I do to improve it?
4:11:28
Would somebody like to jump in and maybe talk about the multiprocessor version of pie split?
4:11:40
Um.
4:11:44
If not, or let's see, an assembly Oh, yes, yes. So, so if your computer has NPI, you can run a high split with multiple processors in the exact directory. The standard high split utility you'd have to, I think you'd have to run it from the command line, but is, uh.
4:12:08
It is H by C, S underscore, STD is kinda standard high split.
4:12:15
What's one processor? To run with multiple processors?
4:12:19
It's H Y C M, underscore STD so that changed the S from single pot processor to M for MPI And there you go.
4:12:30
Yeah, and I think that usually, um, the multi node MPI Multiprocessor interface, I think it's called, is available on Linux boxes, but may not be typically available on windows, Windows machine.
4:12:48
Um, the second question, um, says, in the emission rate and the machining time option, we've set hours of emissions too, zero point zero one, hours are 0 0 1 hours.
4:13:03
And if I split runs on a minute basis, at least how is it allocating the emission rate corresponding to this very short time which is shorter then a minute.
4:13:15
And the second question or part of the question is when would this second based?
4:13:21
And that minute based haislip be released?
4:13:25
Um, let's see if you want to take a crack at that one.
4:13:31
So Mark, I can answer the first part?
4:13:33
Sure.
4:13:34
So the first part of, yeah, you really can't have emissions that will last less than a minute.
4:13:39
I mean, is it to the model?
4:13:44
The minute is the smallest time increment.
4:13:47
so anything that's less than a minute it's indistinguishable from a minute, so what will happen is that will cycle through, it'll release the particles, and they'll get transported.
4:14:01
Whether you put, you know, zero point zero five, or five.
4:14:07
Um, so that, I don't know that. I think that's the answer to that question.
4:14:12
The other thing I realized, is that you do have to be a little careful when you're doing very short and miss ends of a minute or less, and you might need to set your model time step in your setup dot c.f.t..
4:14:27
Because if you do the automatic time step, I don't believe this may have changed, We may have updated it, but I don't know that we have, um, When it's setting the automatic time step, it actually doesn't check your admissions file.
4:14:43
Right? Chris was talking about how it decides, what the time steps and be.
4:14:46
So it's possible that if you have very short emissions and you're using the automatic time-stamp that it will, it will actually skipped some admissions.
4:14:55
So if you had, like a 10 minute time step and there was supposedly emissions happening in the middle of that time at 10 minute time step, then highest, but might not use them.
4:15:07
Say you're if you're doing very, very short emissions, you might need to set your model time step to coincide with the shortest of your admission times.
4:15:20
Does that make sense, Mark?
4:15:21
Yes, And I just wanted just to emphasize something that you said earlier on, even though the time step is the minimum of one minute.
4:15:34
When we did those earlier calculations, where we said emissions rate of 100 per, um, then hours of admission is zero point zero one, even though that point zero one is less than a minute, it's still an emission of 100 times zero point zero one.
4:15:52
So, what happens is that, instead of mitigate in whatever NaN or NaN, it's just going to admit that one unit over the over the first minute.
4:16:01
So, we do get the right emissions that when people to applied by these two, Same thing would happen, if you put a thousand and zero point.
4:16:11
So, it's still emit one unit, and it all happened in, in the first minute.
4:16:17
Then, the second question about, when will that, one second version of February be released?
4:16:25
Maybe next year.
4:16:27
We're hoping number of people have been asking for that, and, it is definitely on our, on our list of things to do.
4:16:38
The next question is, in which typical cases would you recommend a trajectory calculation instead of a dispersion calculation?
4:16:48
Do you want to jump in?
4:16:58
I'll start. Oh. No, go ahead, Mark.
4:17:03
Um, calculations is, as we talked about, are really sort of like D Center Line of of aspersion calculation.
4:17:13
And so, if you're trying to get sort of a quick answer of where where a plume might be going or where an SMS that arrived at your site might have come from, trajectories can be done. You know, very quickly.
4:17:27
And it gives you a, you know, a basic idea of where things are coming from or where things are going to trajectories don't take into account things like deposition and any kind of transformations or anything, um, and don't really capture the entire plume. Because, because, remember, we talked about that center of mass trajectory.
4:17:49
Which could be a little bit different than, then the simple trajectory that, that you would run.
4:17:54
So it all depends on, I think what you're, but you're a problem, is that you're trying to solve a lot of times when people take measurements, and then they want to see what, what, what's contributing to their measurements, they'll do back trajectories, and that seems like a very good application of trajectories.
4:18:15
And it's really when you want to sort of do forward dispersion of from wildfires, or from chemical releases, or any other kind of release where you're trying to get the concentration downwind.
4:18:31
You really want to do a dispersion calculation.
4:18:38
The next question: I see the Haislip can be run in, our versus are split, I think it says how Comprehensive Is that package doesn't have all the features of high split via the GUI or command line whatever pros and cons of split are, has anybody on our team run that?
4:19:05
Now, I have, in front of us answer that I know it's pretty popular and there are several. There's a that's like an R wrapper for high split.
4:19:14
There's also several Python wrappers.
4:19:17
I'm including some code that we have here, but there's also ... which was created outside of Arielle.
4:19:27
I don't know if I can say pros and cons.
4:19:29
I guess a Pro is, if you like, are, in your, into our scripting, that it's, uh, you can use. You can utilize that to help you write our scripts to run high split, and look at the results.
4:19:41
Um.
4:19:44
And just to jump in, this isn't just another model.
4:19:51
This is just a wrapper similar to, uh, scripting that you can, that you can do in Seashell or bash that we have some scripts here.
4:20:02
And there's also the stilt wrapper which is also based in our as well.
4:20:10
Just want to throw that out Yeah the rappers are really useful, right, because we kept talking about a conscript this, you can script that's, then you can script it in different languages, are Python, and so forth.
4:20:23
Like Chris said, it's not another model, it's a wrapper.
4:20:27
And if you do scripting in our, you can write your own scripts and are, and like Alice or Chris said, you can, you could look at the, split are, and seek examples of how they've written their scripts in R and you might be able to use some of those scripts, or you might decide to write your own scripts. Same thing would be put with Python. So, we write.
4:20:50
I split scripts in Python and R at AOL too, um, just because we use those kinds of scripting languages as well, so, but these are these are people that have made a package and it put together a bunch of scripts, which can be very helpful.
4:21:04
And we do appreciate people, people doing that.
4:21:10
Next question: for unknown source locations: we have seen the receptor matrix approaches a step: let's find sources, but my understanding is that it does not run an ensemble of trajectories.
4:21:22
How do we account for the uncertainties associated with turbulence when using the source receptor make matrix method to find source, location and strength?
4:21:33
Let's see, 10 thing. Or if you're on the line, this might be your ally here, yes.
4:21:39
Uh, I think, yeah, the ... can be used to this kind of problem, and the ... based on the dispersion model, not the Trajectory model, so that you use random distribution model. And actually, you can specify the uncertainty using different ways.
4:22:01
But that says, Like a part of the research field poverty right now, don't halla.
4:22:10
There's rarely a standard way to do that.
4:22:14
The one way I think, you can do ensemble runs, and so you can get different sorceress after matrix and you can compare them to get the uncertainty.
4:22:26
Yeah. So I think that's, that's where it our open question.
4:22:32
Yeah, no, 10 thing was mentioning that in the GUI we had sort of these ways to combine ensembles and things but but really you don't have to use those particular tools.
4:22:47
You can just do separate runs with it with different settings, different configurations, different parameter ...
4:22:56
and see the kinds of differences that you see And that's one way of getting at the, at the uncertainty. You don't have to use the different ensemble.
4:23:06
Combination and synthesizing tools that are in the graphical user interface. You can create your own outputs and look at them. You look at them yourself.
4:23:15
Um, next question, regards arm.
4:23:19
If for any recommendations on, on literature, um, that would help us.
4:23:26
People get a deeper understanding of some of these underlying concepts of dispersion and turbulence, and I think we gave it an initial answer of looking at the review paper by starting it off from 20 15 and had a lot of references.
4:23:41
Um, there is also this, arel 224. Let's see if I can quickly find that.
4:23:51
Let's see, give me one second split document.
4:23:59
Arielle, it's in the document, rectory, and if I open that up, Oops, opened up on the wrong screen, of course.
4:24:12
Ah, can't win when one second, let me move it.
4:24:17
Over, um, this is a document that started out, you know, more than 20 years ago, but it's been updated, you know, um.
4:24:30
Very frequently and you'll see basically, a lot of the equations and there'll be references to where these equations came from.
4:24:40
Um, this would be one place to start, at least you'll see the references for the equations in, in high split.
4:24:50
Does anybody else on the team have any suggestions about, particular, good textbooks are articles that they could recommend? But somebody's underlying concepts?
4:25:09
I know there was a book on the crunchy and modeling, and it was an AGU monograph or something like that potentially be be helpful, You know, when you're looking at um, literature high split as a Lagrangian particle Dispersion Model, or L P D M, and grungy and meaning that's following the plume. As opposed to ..., which it'd be a good model.
4:25:36
So you're looking for, um, articles about the grungy, and modeling, if you tried to understand more about about islip.
4:25:46
Let's see.
4:25:47
Do you have any advice for running high split dissimulate, insect dispersal?
4:25:52
I notice there are many references in the literature that have done this.
4:25:55
But wonder if anybody on the team has any broad tips, tricks, suggestions, and things about that.
4:26:03
Sonny, do you wanna answer that?
4:26:08
You're going on the line.
4:26:11
If not, I can I'll take a crack at it.
4:26:14
Well, we we were contacted a couple of years ago by the Food and Agriculture Organization of UN, because there was a big, locust desert, locust Upsurge.
4:26:24
and we created a special version of high split that simulates locust dispersal.
4:26:34
And, the basic point of it is that, um, locusts, and some many insects just follow the, when they're kind of lazy flyers, or they don't know, they can't fly against the wind very well.
4:26:52
If so, that basically a trajectory model actually is almost ideal for, for simulating insect dispersal. Because, basically, especially focus this deadline and a swarm staying together.
4:27:09
It's almost like they are trying to mimic what the, what the trajectory model is doing.
4:27:15
And, we have a, um, a website that you can launch your swarms off at different times and it will ...
4:27:26
the one thing about insects is that for locust, at least, they stop at night and sleep or eat.
4:27:32
And then, they get up the next morning. That was so, that was the one big difference we had to do.
4:27:36
Um, for locusts, as we had to stop the trajectories each day, and to start them up the next day.
4:27:43
But, we've seen papers, do all sorts of things live with insects, and, no, basically, um, I think you can do it and you can run trajectories, or you can run dispersion, um, and it should work, you know, reasonably.
4:28:02
Um.
4:28:05
Let's see.
4:28:05
Next question is: can you talk about the differences?
4:28:09
Influx part and high split, and what would be more applicable?
4:28:13
two High Resolution, North American grigg: My understanding that Highsmith pushes the plume up quickly, for long range transport transport. The flux part does not.
4:28:22
And so, it would be a better Would be better for short range dispersion.
4:28:26
However, in today's workshop, we discussed Stilt option that puts a lid on upward motion. So, that might be useful for modeling, sharpens dispersion.
4:28:35
Thoughts?
4:28:36
Anybody want to take a crack at that flex part versus high split?
4:28:40
Or I split stilt versus high split or short range dispersion.
4:28:45
I am not going to say use flex part or use high split.
4:28:50
I think you need to do your own evaluation to make that decision or see what other people have done. As if someone else has evaluated Flex Barton High split with the same same tracer studies. In RL we have a ton of transfer studies that can be used for evaluating our our model.
4:29:12
Um, in terms of the Stilt dispersion scheme versus the high split dispersion scheme that we now have an high split yet are I have found that the Stilt Dispersion scheme for the cap tex experiment gives you a little bit better results.
4:29:40
due to that Qaeda that what I mentioned and what you mentioned, your question.
4:29:44
That there's uh Yeah, there's less PBS venting, flesh transport pollutants between the PBL and free troposphere with the ... scheme than than the high split scheme.
4:30:02
Oh, go ahead. Go ahead.
4:30:04
But there's a follow-up question that, maybe while you're answering this, you can think about, Is still the better option to go with an urban areas with high density of sources and mapping, needing to be done at short distances? I don't know if you want to factor that into your answer as well.
4:30:20
Uh, aye.
4:30:24
Not for those reasons. Not for those reasons.
4:30:32
The reason would be is if you have, if you're running highest split and you're finding that it's, it's venting too much, and too many pollutants are being transported from the PBL to the free troposphere with the high split scheme, then moved to move to the stealth game, and there'll be a bit sharper lid.
4:30:55
That's all I have to know.
4:30:58
And I think that, um, no, Flex party, high split, I think, are?
4:31:04
Are very similar, I think. It.
4:31:07
And fundamentally, and I know there must, there's differences. Yes, go ahead. So, I'll just jump in a little bit.
4:31:14
As Mark was saying, I mean, they're both pretty similar and type of model that they are in their capabilities, and they're both being actively developed, so it's a little bit hard. Like if you ask the Flex part people, they might not be completely, no, you know, everything that's in high split.
4:31:32
And the same for us when you're asking us we, We don't know, necessarily the most recent Flex Park developments, as far as you know, like, I think both of them probably have different defaults for their mixing schemes and so forth.
4:31:49
But we do have taken part in some inter comparison exercises and in which both models are used and it's really, you know, like sometimes one does better sometimes. The other does better. It depends on what meteorological model they're being used to, to drive.
4:32:08
Being, you know, driven with.
4:32:09
So, you know, if you use flex part and high split both driven by like GFS, they're probably going to be more similar output. Then if you use Flex part driven with like an ECM deaf ECM WF Madalyn High Split driven with TFS.
4:32:23
Um, so yeah, I guess that's kind of the answer to the question.
4:32:31
I mean, there are capabilities and then to some of the times, one has like a capability, the other one doesn't have I think flex part does have that kernel density estimator and that you can use.
4:32:44
Um.
4:32:48
That's OK, thank you, thank you asked, and I'm showing on the screen here actually, we have this datum website.
4:32:57
It has all the tracer experiments that we that we've done.
4:33:02
And so you can download that data and you can do so do your own tests.
4:33:06
We even have meteorological data that can be downloaded for each of the phaser experiments in high split format so that you can run high split on those.
4:33:18
And then you, if you wanted to, you could run flex part on some of those as well, and see how well the different models did.
4:33:28
The next question also deals with sort of comparing models.
4:33:32
It says, a key advantage of high split over Cow Path, and air Mod is the ability to conduct batch trajectories.
4:33:39
It looks like a URL is using high split, the short range dispersion for forward modeling as well.
4:33:45
I can see the value of doing Aramark at very short distances since it model sources to a great detail.
4:33:50
What would be application scenarios to use cow path?
4:33:55
Um, oh, or, or high split.
4:34:00
Let's see, I'm not quite sure that the wording of the question I guess they're asking, you know, what are the relative merits of using arum odd Cow path, high split, especially maybe in short distances.
4:34:15
I'll start: XA ...
4:34:17
is a Gaussian plume model, and it's a very sophisticated Gaussian plumed model with really interesting features regarding, like building down wash.
4:34:27
And, um, so in certain ways, Aramark is very, very much sophisticated and more sophisticated than then high split in terms of a very short range.
4:34:40
Um, modeling what ... does, though, is it uses no setup.
4:34:48
Sort of a constant meteorological winfield anything, you can change it.
4:34:53
Um, but it's just sort of assuming.
4:34:57
I think that the wind is going a certain direction, and then if the wind blows a different direction in another, the next 10 minutes for the next hour you will have a new set of plume calculations. That may be perfectly fine.
4:35:11
That may be actually what, what, what you want with, with high split.
4:35:14
You can use, know, this continually changing meteorology, but for sure, um, with the one minute time step, you know, the typical wind speeds is going 100, 200 to 300 meters in a minute.
4:35:30
Um, you're not getting near the kind of the short range resolution that you would get from a model, like, like Armand.
4:35:38
And I thought That familiar with cal pop.
4:35:40
I think that's I think that's a path model And that's I think it would be similar to using high puffs. And I'm not sure, what could a short range?
4:35:49
Capabilities that has, anybody want to add anything?
4:35:53
Yeah.
4:35:53
I'll just add about the Gaussian trim models.
4:35:56
I mean, one of the reasons that we haven't yet, it's taken us awhile to think about no decreasing split time step and so forth, as what Mark kind of touched on as that.
4:36:09
If your meteorological data no only has a temporal resolution of even an hour and you know spatial resolution, you could, you can decrease the time second highest split, and, and make your concentration grid really fine, and you can do the calculation.
4:36:24
But what you're gonna end up with is something that's very similar, that you would get from a Gaussian plume Model in the short range, but what you've just done you've done, you know, the Gaussian Plumed calculation but at a much higher computational expense.
4:36:38
Um.
4:36:39
So, that's one reason that, uh, there's that kind of dichotomy between the Gaussian plume model is being used, you know, at short range at short timescales because they provide that, basically. An almost like an analytical solution, right? That you're going to get this Gaussian plume.
4:36:57
Whereas, with the Lagrangian Particle Models, if you want to reproduce that, which you can, you have to use a lot of particles are very fine concentration grid. And you're not taking advantage like your meteorological model. Data is not changing at that time in spatial scale, right?
4:37:13
So, that's right. That's right. And you know ultimately you should get roughly the same answer and in fact if you use a Gaussian p.u.f.
4:37:22
coming out of high split, it'll probably even get closer to that.
4:37:26
To that answer, um, but absolutely I think if you're looking at really short range dispersion like within a kilometer and you're really concerned about things like you know building down Washington street canyons and things like that then um you're probably better off using a model, like like like air moth.
4:37:44
Um the next question has to do with scripting, and I'll kinda paraphrase the question.
4:37:50
Um, having to do with it, some students, you use Windows machines and some students use Mac OS, and, um, they were frustrated that they had to have sort of two sets of scripts.
4:38:08
And so, therefore, um, seems like R or Python would be better Because they're more platform agnostic on that to be run on both.
4:38:17
On both platforms.
4:38:20
I will say, I've heard recently that you can run Linux Shell scripts on Windows now, and you certainly can run them on a Mac.
4:38:29
So I'm not sure if that's quite as true as it used to be.
4:38:33
But absolutely, writing scripts in Python or R would certainly be a good way to go.
4:38:40
And you don't necessarily have to use Pi split or Split are.
4:38:46
You can, But you also could just write your own scripts in Python, or, So if those software packages don't have everything that you want to do, they may not then, you know, you should feel free to write your own.
4:39:04
When you just want to add anything to that.
4:39:06
Yeah, I'll just add it unlike if you are a Python user, Um, We have some code available from GitHub.
4:39:14
They're not really like false scripts that will run that if you're not a Python. Either. This might not make sense to you, but if you are, we have like some Classes that will read, like the sea dump Files into like an X Ray format.
4:39:30
And, uh, so we have some classes that will read the binary, high split, files, and pack, and read them into a quick we use the X-ray package module, and Python.
4:39:47
OK, thank you.
4:39:49
Let's see, Ah.
4:39:53
one second here. Next question. A little bit more about the Locus forecast capabilities. Just when you go to that website, the locust, if you search for locust, Noah, RL, or something, you'll come up with this. This website. We have some documentation.
4:40:09
Um, we have these different modes. You can do a single swarm, you can do a batch. You can also do a matrix, if you like.
4:40:18
If you know where the source location that's sort of a big area is, you can do, um, a matrix. We can do forward trajectories, or you can do backward trajectories if you know where the swarm, why isn't. You wondering where it might have come?
4:40:34
Come from.
4:40:35
We have a lot of output in GIS shape files.
4:40:41
And, I'm not going to go to a lot of detail on on this, but it's, it was used a lot in the last couple of years.
4:40:48
But there was a big, a big upsurge in locust, something like 10 to 50 times, higher numbers of locusts, in, in Northern Africa. And in the Middle East, due to some climate change impacts, it rained in some areas that don't usually get rain. And there was more breeding than then a lot more breeding than normal.
4:41:11
Um, so, I think I'll, I'll stop there, but there is a lot more information on our website about that.
4:41:21
Let's see.
4:41:22
Um, you're going to find, OK, ah, sorry.
4:41:36
Sorry, one second here, Do you have a recommended resource, or documentation on how to create the various types of input files we've use? Such as the comparison measurements?
4:41:48
for the Flight data?
4:41:51
So, how do we, how do we create all those, I guess, the data files that we're using, come to, to plot Somebody want to take a crack at that one?
4:42:03
Hi.
4:42:05
So maybe you can go back to the webpage, you show that?
4:42:11
Yes, Yes, Yes, So there is a document PDF file there actually showing how that then, format.
4:42:23
Is created, and by the way, other a text file, or a, especially, two measurements, like the cat tax measurements, aircraft measurements, stacking.
4:42:38
So caught at that, and all that stuff.
4:42:42
Oh.
4:42:46
Let's see.
4:42:51
Can you go back?
4:42:55
Just go up the top of the page.
4:42:59
So the first link that Adam Yes, one OK thing OK.
4:43:07
Yeah, this this?
4:43:11
I think at the end of this because this document talk about couple of Chase experiments originally included in the program, but you also talk about the format, the text file format we use in high speed.
4:43:30
So, following days, you can create basically karlan textile, um, then it can be used, and it has fit and can be read by many programs, utilities, programs, in High speed.
4:43:50
OK, thank you, ..., and thanks for directing me to approximately the right right place.
4:43:57
Um, OK, let's see.
4:44:03
If a smoke event and a dust re suspension event are produced at the same time, how can we model the total concentration considering both candidates?
4:44:15
Yes.
4:44:17
Oh, go ahead of the event. And to test the event, that data about a particular matter about the characteristics of particle sizes are quite different for a smoke case. And this case, though, for the fire, it is more likely fine.
4:44:40
More of a fine Patrick curves, and therefore, that does, it has more, of course, oh, I'm not sure about what the Internet can be done for the Islam.
4:44:52
But I really don't let command to do that.
4:44:58
PICO today are totally different, OK? They seek to tilt their product effects.
4:45:04
And one thing you could do is simply do the run separately, and then add them together at the end.
4:45:12
If you use the same grids, might be possible to do that or you can simply, you know, just have these two different runs and add them together.
4:45:24
Uh, perhaps, using, I think there is a cost.
4:45:29
Yeah. Yeah.
4:45:31
Can, we can add them? For the total mass.
4:45:37
Right, right, yeah.
4:45:39
Um, OK.
4:45:43
Let's see.
4:45:45
Could someone explain more details about the footprints and how they are used to estimate concentration?
4:45:50
Any missions.
4:45:52
Chris? Yeah. Yeah. Maybe talk about a bit about the inversion.
4:45:57
I think you're probably wondering what's that next step, OK, Yeah, so, to calculate the footprints in the high split, really, all you have to do is set ICANN equals to eight, and then, it will calculate a footprint file that's that particle dot dat.
4:46:13
What the footprint is, when you also, when you run, you'd run with a unit emissions.
4:46:19
And from your measurement location, if you're running backward in time, from your measurement location.
4:46:24
So, you have a measurement concentration, and then you run backward in time and you calculate this footprint, which is basically the sensitivity of that measurement to the surface.
4:46:35
And if there is a missions at a location, whether there's a footprint near the where, the backward dispersion run, uh, was over an emission source.
4:46:47
Then you can use that footprint, two, which is more or less, hey, atmospheric dilution coefficient and kind of calculate the modeled concentration based on what I think T and thank covered that chat that section. And in Section 13.1, you can calculate an estimated model concentration based on the footprint.
4:47:14
Then, you have this model, concentration, and this measured concentration, and there's some error there, OK, and then you would need to put that the error, that error information as both those model and measured concentrations into an inversion system. So, whether it's a Bayesian inversion system, or some other inversion system, too.
4:47:39
It as well as this.
4:47:42
The another emissions, a priori which would be an emissions dataset, which could use to calculate, have the estimated emissions from your simulation.
4:47:56
Put it all into an inversion system to estimate, uh, try to improve the emissions from bottom-up inventory.
4:48:08
And, Chris, I think there is some software and some examples on the cell website, and I think you could look there to see, um, some examples of how people thought people do that.
4:48:21
Yeah, You can look at the ... website and there's also the carbon tracker, Lagrangian that our colleagues at in Boulder. NOAA grew up in Boulder. There.
4:48:34
That's their inversion system Bayesian inversion modeling system that they use.
4:48:39
Where they would, yeah, take high split output, put it into, there are Carbon Tracker Underground system to estimate emissions.
4:48:48
OK, and the next time, Oh, go ahead, yes.
4:48:52
OK, I'll add a little bit to what Chris said, I think footprint is the basic they, just the sensitivity matrix say you're going to gather, escaping the relationship between the source terms and that your measurement.
4:49:08
And so, you can gather footprint by running the model backward from the Betterment locations. And a photo you are sharing yourself footprint, is not different from a huge round for a model to gather saucer receptor matrix.
4:49:26
So, what it does to the inversion, is, basically, you launch your, adjust, your animation, trying to make the your model results. Matching the observations so that you're going to improve the luck a priori emission estimates, you'll get a better estimate of the mission.
4:49:53
Thank you.
4:49:54
Thank you, Tim thing.
4:49:55
Um, and the next question that keeps related, where can I find more detail about doing this Still put, put calculation. I think we've sort of been answering that.
4:50:04
I think you could look at the stock Web site to see how, once you calculate the stilt footprints with high split, how they can be used.
4:50:14
And I think they give you examples of how um, They do it.
4:50:20
The nice thing I think that Chris talked about is how still it was originally based on high split.
4:50:26
And then the codes diverged a lot now that we brought them back together and then sort of put those stilt um, feature is back at the high split now.
4:50:36
I think even, there's still people using still, no further inversions are now using that new version of, of Haislip, but they're still in it.
4:50:45
So, um, let's see the next question. For simple photochemical modeling.
4:50:51
Could we use a Box model to mimic an equivalent linear rate to be put in high split?
4:50:58
Asked?
4:50:58
He's asking, because box models are clicks to run and can be combined with high split, instead of setting up an entire new, Um, C T M C has probably a chemical transport model that has a steep learning curve and requires long time setup.
4:51:12
Um, this is a good, good question, I'll take a crack at it at first.
4:51:18
If you can make an estimate of the typical sort of linear type rates based on a box model, um, simulation, then, yes, you absolutely can put that rate.
4:51:32
That estimate, that approximate rate in into high split and that would be, you know, um, a fine way to do it.
4:51:40
If you feel pretty confident that the rate that you calculated, it seems reasonable.
4:51:45
But there's one important caveat that I'd like to make, and that is, that, so put in most non-linear chemical situations like, for example, photo chemistry, what you have is the emissions from one source are interacting with emissions from other sources, and mixing together in the atmosphere and then, creating, doing all the photochemical reactions.
4:52:08
And then the ozone creation and the reactions with ... and everything else.
4:52:13
It is, everything is being all mixed together, and you need to take all of the emissions into account in order to do the chemistry.
4:52:21
So that's not what Haislip would be doing.
4:52:25
So, um, or, for real photochemical modeling, where you really need to keep track of all the sources and all the emissions that ended up in a certain grid square to do the chemistry.
4:52:38
Then you're really going to have to do probably use and Olivia Model Alexi Mac, or one of the other culpable models.
4:52:47
Ariel Stein that the Director of RL for his PHD thesis he did a version of high split wear brought in the CD four, which was, at that time the good the photochemical mechanism. And the way he did it was, he did the dispersion with particles, and then brought them all together into a grid square.
4:53:08
Then, mixed all the mass together and all the chemicals together did that see before the big, big chemistry mechanism, and then reapportioned the mass, back onto the, back onto the particles.
4:53:22
And so you did that transport and dispersion with particles, and they did the chemistry in kind of a ...
4:53:29
way to kinda go back and forth. Between the two. you see essentially turned.
4:53:34
I split into an ...
4:53:36
bottle.
4:53:36
So, bottom line is this, if the emissions from one source impact the emissions from another source, then you probably should use in a eulerian model.
4:53:48
If it's just sort of a non-linear chemistry where, um, the emissions in your plume are just sort of reacting in certain ways that are just non-linear.
4:53:57
And even if you can estimate that reaction rate in a linear way, then absolutely. It will work.
4:54:03
Sorry, long winded answer, but, but I had to deal with a lot of that with, with the modeling of mercury, so I have a strong feelings about some of that stuff.
4:54:16
Let's see.
4:54:18
Next question for ICANN equals eight. Is there any recommendation as to the length of the hours to be run?
4:54:26
Uh, let's see, Chris, or let's see, I'm not sure who might be the best person to answer that CC.
4:54:37
I can, I'm sorry, I missed that, was the question mark for ICANN equals eight.
4:54:45
Is there any recommendation?
4:54:47
That's the length of the hours to be run.
4:54:51
Oh, Yeah, I can be. Equals to eight is. Oh, no, that's sorry. It's not that still, I'm sorry.
4:54:59
Yeah.
4:55:01
No, there's no. You can run it for us as long as you want.
4:55:04
There's now set time.
4:55:09
Yeah, I mean, people all have, yeah.
4:55:12
two people in you have been using?
4:55:18
Part of God, that dot, that footprints, they still type footprint files for a variety, from spatial scale. So from short scale, looking at urban areas to regional, too.
4:55:30
Continental scale type of experiments, so no, there's no time period or resolution requirement.
4:55:43
Basically depends, I think, on this scale of phenomenon. That you're trying to model. Whether it be local, regional, continental, etcetera. So, yes, yeah. Yeah, I can add more to that.
4:55:55
I think for their problem was whole time they have Becker area of interest. And one day to day, even worse, still run together for the print. If the footprint covers the area of interest, that's the time period that you need.
4:56:18
Thank you. Thank you, can. Actually, you might be the person to ask the next question, Sue, so keep your mic on. The next question is: Is it possible to model the transport of mother daughter nuclei chain in high split?
4:56:33
where one ready nuclei decays into another with different transportation properties?
4:56:38
For example, radionuclides that decays into a noble noble gas.
4:56:45
Well, that.
4:56:46
Yes, I think of a way to have that capability.
4:56:49
But That's something that I have not tried, but I know we have that type of analytics.
4:56:57
I remember there is something in the tutorial over 50 theories that equity.
4:57:06
Yeah, we Oh, go ahead.
4:57:08
Oh, no, no, no. Yeah.
4:57:10
Yeah, in the, um, On the ...
4:57:13
webpage, when you go to the Tutorials, we have so-called advanced tutorial, which is a little bit out of date and not necessarily been updated with every single change in every single menu.
4:57:26
On the other hand, it has a lot of these more advanced concepts, especially with nuclear simulations, and.
4:57:35
It would be, even though it's not quite.
4:57:40
100% up to date, it actually, I think, has enough really good information that it could be quite, quite, quite useful.
4:57:47
And that's one of the things that Ken thank showed, is that often when you do simulations of nuclear, pretty good guides, you run three, or four, or five different surrogate species, you know, you run a noble gas, you run particles, you might run particles of different sizes, And then you can combine, you run a unit, the emissions of those different surrogate species, and then depending on what's happening in, in the environment, and in the emissions, you can combine those in different in different ways.
4:58:24
We don't have A capability in high split right now.
4:58:28
Where you can convert something from a gas to a particle, or a part of a gas on the fly.
4:58:35
Um, actually, we are thinking of adding that in shortly. We had done that, for some special applications.
4:58:43
Need to need to import that type of code into the, into the mainstream model.
4:58:50
Let's see, I think we are, we've reached the end of the questions, and, um.
4:58:59
The question is really excellent, and I really appreciate all of them.
4:59:03
And I think what I'd like to do now is do a quick wrap up, and we can, uh, get on to the rest of our.
4:59:15
Let's see, let me find one second here, is this the right one?
4:59:22
Now, give me one second.
4:59:25
Let me open up the right.
4:59:34
I'll.
4:59:43
All right.
4:59:45
So, I'll just do a quick, quick wrap up here. Let's get the right view.
4:59:51
Yeah, you, then.
4:59:57
You, I do the full screen. There we go.
5:00:03
Oh, Kay. There we go. OK.
5:00:06
So, um, you know, first of all, thank you all very, very much for attending, this has been a really great workshop for us, and I hope for you as well.
5:00:18
We've gone through trajectories, air concentrations, and then some sort of specific applications of, of, of the of the model.
5:00:31
And we hope we've shown you sort of a little bit about the underlying features of the model and underlying theory of the model.
5:00:40
And also, sort of, how you, how you apply it, and why you might make different choices.
5:00:45
Um, as we shown you, the model can be run in these puff modes, that split, or these particle modes, um, you can go forward and backward.
5:00:58
You can do trajectories of dispersion, no different kinds of deposition and transformations.
5:01:04
In this workshop, we've basically download of the model and we run using the graphical user interface, but you can run it online with the ready system.
5:01:14
Um, when you do that, there's a lot less features that you can do, but that may be sufficient for your use. And the advantage there is you don't have to download any metadata.
5:01:24
When you run that aren't on your own computer, you do have.
5:01:26
To download the data, which we do have for you, but it's an extra extra step, And then, of course, if you've downloaded the model, you can also run with scripts, and it runs on different platforms.
5:01:40
It's used a lot at Noah and at other agencies, and also the scientific community.
5:01:46
Um, just to give you an example, just for the record, we should review paper 5 or 6 years ago, by Ariel Stein, the Director of the Air Resources Laboratory.
5:01:57
That's been cited over 3000 times.
5:02:01
And they're ready papers decided, you know, over 800 times.
5:02:06
So that's a lot of citations, a lot of people are using high split to help interpret their measurement data, and to do other kinds of other kinds of analyzes.
5:02:17
Really quick, a little bit about the resolution, people always ask the question: What is the resolution of Pi split?
5:02:23
There's different answers that you can give, you know.
5:02:26
On the one hand, it's driven by Gritted meteorological data and so you could say, well, its resolution is sort of the resolution of that logical data. It's a quarter degree or 10th degree or 27 kilometers.
5:02:39
But that certainly affects the accuracy.
5:02:43
But the particles in, the puffs are being transported in between these grid cells.
5:02:48
And I split is inter-related in time and space to get the wind speed, wind direction, and other things to move those hops and particles around.
5:02:59
So, I think what we're saying here is that the accuracy of the transport and dispersion will be affected by, oh, finer how of course the gridded data is.
5:03:11
But, but that's not really the resolution of the model.
5:03:13
It just means that you're you're driving it with either a coarser or finer set of data.
5:03:19
Um, then there's just this concept that the concentration grid.
5:03:23
So that perhaps is, is it a little bit more related to the resolution that if you have a very coarse grid, like you know, one degree or something like that, then clearly you're going to get averages for over that whole area.
5:03:39
And so, so what you've done is you've told the model only two.
5:03:44
create averages on a very coarse grid, but you could have picked a finer grid.
5:03:49
So no, it's partly dependent on what you pick for you, for your output and you can't go crazy and pick grids that are like one meter or something like that.
5:04:01
We've talked about you know maybe one kilometer might be a grid that it could be used as a very fine grid, but it might be need to be careful by going much smaller than that, partly because of the time-stamp of haislip being being one minute.
5:04:16
And that's sort of a little bit about what I'm talking about here.
5:04:19
Um, um, no.
5:04:24
The wind speed of three meters per second, which it might be Typically, you get 180 meters in a minute.
5:04:29
So you know that the time steps are moving stuff along to 80 meters, or more, little bit less.
5:04:36
Now there is a interpellation platoon, high split that tries to, to account for that leapfrogging, but it's not using any kind of metadata finer than one minute, And in fact, you rarely would have that data anyway.
5:04:51
So, no, you can make a really fine grid, but are you really improving the calculation, any, if there's different answers that people can give give to the resolution question?
5:05:04
Here's just a little thing about the concentration grid's you need to make them coarse or fine.
5:05:11
Um, we talked a little bit about the cap tax experiment and we did a lot of tactics simulations in this workshop and we did one, um, of the project sagebrush with certainly short range transport. We have we've done a lot of tracer experiments over the years.
5:05:31
Here's a map up showing sort of the locations and spatial scales, and we have that data on our website in the data system. And you can download it. You can download metadata. You can also do some runs, right, on the system itself.
5:05:47
So, um, this is a good tool that you can use if you want to do some of your own model evaluation, and we, we use this to try to figure out how well the model is doing and how, you know, how we could maybe improve it.
5:06:02
And even though the Tracer experiments were done awhile ago, in most cases, you can still test out new versions of parameter ..., new versions of mythological data, simulations, new versions of dispersion scheme's things like that.
5:06:16
So, they're still quite quite useful.
5:06:20
Um, I made a quick summary of some of the recent updates we've had in the model.
5:06:25
We've talked about a few of these, not too many of them.
5:06:31
And then some of the recent features that we but we've added, you know, I've mentioned the SVG graphics for the whole, big set of Python graphics that we didn't talk about much, that are really quite quite quite powerful.
5:06:44
Alice talked a little bit up a little bit about the density estimation via the Gaussian mixture models.
5:06:50
There's that new program that we showed you about the test program that can help you diagnose some configuration errors or configuration problem.
5:06:59
So, this just gives you a little bit of an idea of what we've been working on in the last several years.
5:07:07
Um, almost done here.
5:07:09
We just had a, recently, we had a five year science review of the air resources laboratory, and I'm mentioning it, I have a link here because there were, oops, sorry, just go to the next, is that the group that was doing all the teaching during this workshop, they each gave short presentations of, I think, 10 to 15 minutes about certain topics that, but close to their research interests. And on this website, you can see their presentations.
5:07:43
You can download them, Also, there's a movie of them giving the presentation that we'll hear there.
5:07:49
There are voices and things.
5:07:51
Also, there's an overview presentation that I had, and I put together. So, if you wanted to dig a little bit deeper into the kind of things that we're doing on a day-to-day basis.
5:08:02
Ah, you could look here.
5:08:06
Um, Here's a little page.
5:08:08
I've put together about some haislip tips, and I guess everybody probably has their own version of this, and and.
5:08:15
But this was my version, then couple things, you know.
5:08:19
The control file is the main file for high split, and you know, if if you're having a problem, sometimes you can look at that and see some obvious errors like misspellings of things or things that are obviously wrong. Things that should be numbers, that are, that are letters, things like that.
5:08:36
When you're using the graphical user interface, all your files are in their working directory.
5:08:41
When you're doing scripting, you often create a new directory, because, you know, it's only in the graphical user interface, where everything's in the working directory. But when you're doing your scripting, you probably want to make a new directory.
5:08:53
Um, and you can have too many directories for different projects and things.
5:09:00
When you're picking your net files, you have to have the correct directory name, and it has to encompass the time and spatial domain of your of your simulation.
5:09:07
If this is a common area where you don't have the right file, or it's not big enough, or it doesn't have the right time, or just long enough in time.
5:09:17
So, and you can have more than one that file, and, you know. You could have a lot of the met files.
5:09:22
We have are only one day long, so you can put, you know, seven of them.
5:09:27
You can specify 70 files to do a week long run, if you want things like that.
5:09:32
Um, some of the output files, or ASCII text, some are binary, so you are going to be a little bit careful about that.
5:09:40
I think our, um, I'll let you know. Look at some of these things.
5:09:46
If you want, you know, the idea that numerical experiments, that you can just try things, we don't always know what the perfect answer is.
5:09:53
But you can try things and see how they work.
5:09:57
And generally, if things are changing a lot, based on the settings you're doing, you probably need to keep looking into work, finding where things might stabilize.
5:10:09
Especially in terms of number of particles in your simulations.
5:10:15
OK, almost done here, here's a sheet, There's a bunch of different high split learning, um, resources, including that, that, that, the document I showed earlier with all the equations and some other other things. This is the paper from 20 15, that's beside the lot and also the ready paper.
5:10:35
All right, there's the workshop webpage, will keep updating. This will be adding a few things like some of the slides from today.
5:10:42
Um, just another image of that worksheet webpage is just part of it.
5:10:50
For course, certificates, We will make them available to all of you will probably give you an e-mail in the next week or two, asking you if you'd like the certificate and you can let us know what name, you know, the name that you would like to have on the certificate, um, and also obesity, or survey.
5:11:11
We really hope that you will give us feedback.
5:11:14
When we we would like to be able to improve this workshop and we'd like to find out what you think worked well and what didn't work well.
5:11:23
Or are some ideas that that we could use to to improve it. So we really hope you'll take a few minutes and fill out this survey and it can be anonymous or you can provide your name.
5:11:34
It's up to you.
5:11:36
Um, then finally, one more, one more plug for the ISP Forum, you know, now that you're, you know, more versed in high split, we hope that you will keep using it, and we hope you'll continue to ask questions in the forum.
5:11:50
And you can see, we were, we were asking questions and sort of the workshop version, but there's a lot of other sections, and you can see there's a lot of questions and answers, that have been, um, given, you know, about different topics, trajectory, dispersion, bugs, And these are the Peter bugs, not the insects.
5:12:09
But we have, you know, other sections for you know, platform specific sections, questions about meteorological data, um, then different research topics, you know, ash, smoke, dust, chemicals into different things.
5:12:25
So then then the high split workshop on which we were working in this this week, and we had, you know, these different sections for, or for different, different workshops.
5:12:36
So please register and use this form.
5:12:41
It's simply wait a great way for the community and for us too, to make the model better.
5:12:48
Um, final plug then for this conference coming up, in about a month, if we get enough high split related papers, we could have a special I split section, no, No pressure.
5:13:02
But, if you're using haislip for your work, and we'd like to present to a group of like minded people that are, that are doing this kind of work, from around the world, this would be a great conference to to attend.
5:13:14
And that is, yeah. Thanks very much to everybody for participating, Thanks to the support team, and the REL IT team, as well.
5:13:23
And thanks for your extra. I split, and we bid you, adieu.
5:13:27
Thanks very much, and bye bye.