GeoSci.xyz GWB Lectures, 6 Feb 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2020 02 06 Day4 Part1 3LayerApp

Description

In this video, we use Jupyter notebooks to predict the sounding curve for a 3 layered Earth model. We show how the sounding curve changes as different parameters are adjusted.

A

So anybody questions with question one you just change a be upon and fill in the answers. The.

B

Answer I may be yet I'm go over, it being mean it's not at that move, jump down at all. They bailed out, Peter body.

A

For question two: we're going to look at the sounding curve as we change the value of the middle layer. So.

B

Du/Dr questions are oh la la torre de la mano, yeah I'm, gonna turn boho youngjun, happy we're being wary of Ljubljana, always repeats and ethnic. Oh man, okay,.

A

So, let's just let's just pick one example from the table in questa, so.

B

That's what we ended but to put no matter or big table yeah. So how did ya so.

A

Let's look at Row two is equal to zero point. One.

B

Row, what's your point, one locality me.

A

So I change rome 2 is equal to 0.1.

A

So it's very conductive, the apparent resistivity at a small AP upon 2 is 500, and then it just quickly drops and even the winner ray it still hasn't as it bottom up so.

A

You're asked to do look at what a B upon 2 is equal to 5 meters, and that is going to be right here. So it's basically two three four, that's 500 old meters, which is the background.

A

1, so this is 200 meters.

B

Just say it be it be: wasn't in this I should say so.

A

The this tells you, then, that that's the lowest apparent resistivity. We we're much lower than 500, but we still haven't got to the true resistivity of 0.1.

B

So Judy on 10.1 hotate, oh hey, Carissa, Steve you're charged at nighttime, whatever man so do I want you to the mentality.

A

Does that does that answer your question or do you have something else.

B

A

A

A

B

Means that you happy will be found in the path of your cash. So all done a lot PP, not a thing so.

A

You write your name on your question sheet and you can pass it in.

C

A

Yesterday, yeah for both so put your names on each and then hand the madness.

C

A

Yesterday we were looking at the three layer inversion app from.

B

Bonanza there are three layer, a fashion application, Gianna.

A

How many people.

B

Last-Minute scrambling.

D

A

Yes, it's not going to be that you're going to pass or fail the course, but we need to get information about how well you're doing so. We will mark these but you're not going to.

A

Perhaps while the this is happy, maybe I'll go through question three: the relayer sounding.

B

A

Just watch up in here, so the question was and this three layer model. Let me.

A

A

Okay, so here's our model.

B

A

And we're going to do the winner array, there's sounding or over this.

B

Cheating in it, it was our beds, are oh we're not Reina sounding Obama and.

A

We're interested to know what a be upon to value you get, the smallest, the.

B

Smallest value so baby, while we top it up time, I, don't know, there's no there's this DVD, we got never, never know it's.

A

A work we've got a low, resistivity layer, water, so.

B

This is Stevie need a nanny Leah and watching me, no GMO, and we.

A

Want to discover it and what we're going to investigate. Oh, no. The sounding curve changes when the thickness of the top layer changes so.

B

I posed on Wyatt to bow down wanna name, the resistive develop young man is one who looked at me. Do you shower or you shop with everything? Okay,.

A

So we're going to change the this thickness and make it smaller or greater and then look see. How does that change the sounding curve and at what value of ay B upon to do I actually get the smallest number so.

B

The words are not the two one offers on my a to go down below mop it up a a be tangle of tamati and hazel hazel resistivity. She Manoir shadow name last without Peter.

A

Okay, so let's start will just actually carry through this month, so the first layer thickness was five meters.

B

Does it smell so annoying to God page well, not? Finally,.

C

A

A

We're in a winner, so this is the red curve here when.

B

I was around the young man with tlme and.

A

The lowest value that we see happens when a be upon tube is around 20 meters, so ap.

B

Neatness I done. We done this. A not mine, Faisal NATO company I buy from anybody. So.

A

The the total array length is 40 meters.

B

I read the culo in the ads are leaving us alone and.

A

Remember this was only 5 meters, so the arraignment is very large compared to the depth of burial.

B

Sama-Sama shiranami table and Oahu right. No thank you to nation enter Katara array and I hope you are a machine. What.

A

Is going to happen to this curve if I make the layer thicker the.

B

Dirty boson, novel to a Lutie, lightning.

A

Somebody can answer that what happen to this curve, if I, make this layer 10, o meter or 10 years,.

B

So de bozo ma what the semi Doc's, who know the lightness, a diamond blue balls, are under each other.

A

Does the minimum stay in the same place.

B

Eddy, a halo number god I didn't near machinability na.

B

A

Change and which way: where will the minimum.

B

Number of human development.

B

To row, H 110 look down my man now my job, you don't see that would be a B / I'll just stay at my gosh I, don't know so. Ain't Niaga, hey baby.

A

B

Who says it's gonna do that way, but you gotta go 1 million of these nothing a zone above you in milk, any food grease, there's an IAB year and they are boots. You are my body, nobody.

A

Agrees with you Oh one bar one person.

B

A

How many to go this way.

A

What so you think it's going to go this way, and some.

B

People think it's gonna go this way the rest. It's.

A

B

Way, no something they don't raise their hands and.

A

Some they don't yeah, that's right! So there's some that say this way. Some that say that way and the rest of you.

A

Unacceptable unacceptable, you have to be.

B

A

Do you have to vote okay now we're going to do that again? How many say this way? No.

B

You're gonna bleep announcer.

B

A

How many say this way, so the university says that way. The DRD says this way.

A

So now remember the minimum before was that 20 in.

B

Slo-Mo she did not run oh honey now.

A

The minimum is 20 30 just about 40.

B

Because they'd be in this valley day, lazy.

B

A

What about the apparent resistivity very.

B

Resistant, its job.

A

So go back to five. It's a.

A

Hundred and eighty eight.

A

When they later is shallow, you see the maximum effect of the layer at smaller electrode spaces. Close.

B

On my body say the speech today lean it looked at it Miata bubbles of a bond. You know and.

A

You're also most sensitive to this, because this is 188 compared to.

B

50, when our dining like Joe DiMaggio, such as rock also my body dude Yahoo I, was a bro. I am kind of oversight and.

A

Now, if I make it 10, 10 meters thick.

A

I put this down ten beers thick. This curve shifts to the right, so the size of the array has to get bigger, but the apparent resistivity is also increasing. So I don't see this as.

B

Well so suppose, oh my, oh the two lengths and then now you give the area a be over. Oh man, yeah actually I'm on it. You know I would have a protein allergy. So you know, let's ignite, that would even.

A

Okay, so now we're starting to understand if I make it even deeper, so I make a CH 115. What's going to happen so.

B

It's not h1, go me to say mommy time.

B

Somebody question.

A

Yes, what happens goes this way, and what about this number go up or down a zero I.

B

A very resisting there about the bottom and.

B

A

Going to get bigger, good good job, okay! So what do that so we're going to make.

A

So now the the PICUs is over here ready- and this is just about 300 meters.

A

Now my model its way down at 20 meters, down to.

B

A

I'm well over 300, o meters and the maximum is out here.

D

A

6070 meters.

B

Maybe, with minutes on what does that mean? Am I okay, no valid Tirana, so.

A

This is the power of these interactive notebooks, because you can, you can change things and you can see instantly what happens. Here's my here's my model, here's my data, I changed my model. I changed my data. You start to understand. What's the relationship.

B

A

Now, let's go back to the app that we were doing at the end of yesterday.

B

They located the three.

A

Layer inversion app, so we have made this. We have made a couple of changes to make it easier for you to run and also is going to allow us to many different explorations with different data sets.

A

So I want everybody to follow me exactly. Okay,.

B

A

You're going to sign on to your binder and get your notebooks. Can everybody first do that.

B

A

Who does not have this.

B

App the app coma. She didn't really tell me well.

A

Everybody has three definitely are coming missing.

B

One little passion this Thanks.

A

Just check make sure everybody has got it.

A

So then I want them just to bring up the app they don't need to run it though we just want it, because we're going to restart everything.

B

A

Can there be a local hotspot in the necklace.

B

A

D

A

Take you out to launch binder, but once it's launched, okay,.

C

A

Actually, you're, not that cheap, most yeah most people actually have a shaft.

D

B

Yeah me: no, no, okay, okay,.

A

So that's great, so everybody has got this app. As I told you yesterday, this app is a little bit different. This is really truly a jupiter notebook and it is designed to allow you to write code and to see images and see results. You are not going to be writing code. All of that is going to be hidden, but there are still things that are referred to as cells that you will need to consider so UC regents that look like this and then there's some blue stuff over here.

A

So this is something this is number one. It's just telling you which cells- and this just does something- you don't really need to know what it does, but you need to make sure that it's going to do it. So the.

B

Indo-Pacific or not, the cons are the cultivar enabled a yellow yarn upon it emitted the monitor or cookie Duvall, a blue bow, but some of you had pasta, I know as I roll down a I know, I think delusional journey thinking doesn't wanna. So.

A

The way to activate this is to do a shift enter through act available.

B

Tomorrow shift and they body.

A

So watch what happens when I put my cursor on here my little piglet leg day so that activates the box and now I press shift or shift return or shift enter, enter.

B

A

You see, then it goes down to the next block.

B

And Allah bees on that doo da bomb squad job so.

A

It executed, whatever commands, were supposed to be executed in this box, so it it completed the tasks that were asked in this box. So.

B

They're gonna be box them I thought I imagined so.

A

This is an important concept, there's blocks of texts that are cells and they aren't. When you click them, they carry out some commands so.

B

There's no people other than it. The general be like at me. Do I, like them. Okay,.

A

So then it's it's just like okay there's something needs to be done. I click here like okay, that's not, and then you go to the next task. If.

B

There's an IV to I bounce, my dad, who loves your name, yeah I, can see. Now she did yes yeah. She does in an egg house lollipops with whatever a fast. They wouldn't worry. B minor food bloggers are not cool yeah.

A

Some cells are just text. There's no action. That's really required there just to tell you what is happening. It's about.

B

Steam odds are only it's not a busy day so.

B

A

Notebook: here's the purpose its to take, observed vertical electric sounding data and we're going to invert them to find a layered model at death. I'm.

B

Upping your nah good good. Take about that song. You guys should go Sami gudetama, Lizarraga Sami turned obamaís, ok, Jana, Howser, Oh, yummy, yummy, yum resistivity have model yeah, there's.

A

Two parts: the first part, is manual where you are going to change the layer parameters to fit the data they've.

B

Done it, thank you until later on TV ditto. For me to be body's really part of me touching.

B

A

Second part is done automatically using a computer code and it's going to give a result.

B

Automatic computer I looked at our company in Cairo, we reserve automatic to manual automatic in the museum, so.

A

Then you can compare your best estimate for your results compared to a computer algorithm nicely.

B

Otherwise have been the cuvee, but our divine mom who ranked except you don't need a. How do you see we are below so.

A

We step through ship return and then doubter enter and now we're going to load the observe data. So we'll click on next.

B

A

Here is where we've made the change down.

B

On me, huh yeah Bobby I know yesterday.

A

Automatically we had loaded the DRD data into our app Germany.

B

Got D on da cidade.

A

But what I wanted you to work on was the synthetic data set that was representative of the field data, so.

B

Jenna, looked happy that day job with my open, a name to get you down over there since that takes away the moon. Apples I know you've had it they don't and.

A

Then we also have the field data that we collected yesterday so.

B

Madrid or a moment I hear that D dollars she may an.

A

Adult a minute so there's many data sets including four more from the Union Dr Dee that are from villages around mom state.

B

Won't be any machine, you want to die naughty naughty another sheet, but I really, lady, that a lot of times down to about me so.

A

We're going to use this app to look at different data sets and try to invert them.

B

A

Improvement to this app is given in this box here, because once we once we loaded that you see there's this little arrow, that's that's here.

B

I'm out yeah baby, do you know that's.

A

Going to bring down a drop box that has all of the different data sets in it and we can choose which one we want there.

B

I know you'll need a new list, each other hey mama, droppin data, oh you're, not unless we do happen. Okay,.

A

So I want everybody to.

A

Run the first cell second cell third cell.

A

Actually, maybe B I just want to have a should get them to restart their turtles and just okay.

D

A

Yeah, so let's let's go back up here! Okay, so there's.

B

Something about you buy that both a domain called.

A

B

Out they meant, if.

A

You click that no.

B

A

Good at unable there's a line here that says.

E

A

Start and Twitter upwards so.

B

That the I'm y'all medievally man with.

A

Your nose look with your notebook sometime. If you run something and then you go and try to run it again, it remembers what you had before and it gets confused. So sometimes it's best just to every time you run the notebook just start it from fresh.

B

A

Everybody press Colonel, Conley, Omega bar and then press restart and clear output that.

B

You know how maybe that Reese, technically awkward.

C

A

We got Colonel restart so then you should get up something like that, but.

B

Do you believe in everybody you got docked or not? Yes, yes and click that any game thing mom.

A

Okay, now you're done so everything is everything's, cleaned out. There's nothing about the diction.

B

A

So now we start that's down, so it's activate the first cell, so shift-enter safe and there everybody did that yeah. Actually.

B

They're gonna make.

A

Okay shift enter except in love; it.

C

A

A

It's everybody on this.

C

B

You sharing this one.

B

A

Let's, let's see how far we get because now most people are to this stage, and this is now the important thing that I want to say pop.

B

Tv as a young as I believe we finally Tiziano.

B

A

You bring up dance and you notice that there's it's commodity data locations, so there's there's one of the DRD soundings in.

B

Here so a DJ Magua Molina, two delegations with pretty well no, no.

A

E

B

Sorry did I add a coupon, but.

A

I want you to see this little arrow or they.

B

Know you click.

A

On that, lady body.

D

And then, which one should but yeah.

A

Okay, so this brings up an array of data files, I'm.

B

At the data yet so terrifying to leave me misty, the.

A

Oneness automatically done by default is the DRD dataset.

B

At Emma, the odd things up, momentito location would be, but.

A

There's other datasets in here we have the data at the park from yesterday when.

B

They are located when I really look at it. Dwa and.

A

We have the three layer synthetic so.

B

Not ones are synthetics I looked at New Moon about on the moon upon sanchita. She meant so.

A

Rather than having to type this end, which was difficult yesterday, all we have to do now is click on this well.

B

That's no no I'm, not only next I ain't able just on the Internet kept it. Ok. Click on deleted on the net would be a lot higher up you in a little bit. So click.

A

B

A

Then now click on this button, which says remotely.

B

As semantic they does the else we would never maybe I can load. Sorry blue.

A

So that should now.

A

Okay- and it says underneath here that the jsf has been loaded so I'm.

B

On three, like me, tomography or Xavier, the egg will not be ready, daddy, daddy loved it. So everybody go through. Do you get a.

B

A

Okay, so now we're over the stumbling block yesterday.

B

A

So now we enter.

A

As everybody got.

B

This image dpi.

A

Everybody know.

A

Perhaps for for those people who have got it working, you can also help try to help people that are not successful. Only.

B

Nobody knows in Vietnam.

F

A

D

I'm trying to.

A

Working so let's go through.

B

C

A

Remember yesterday, I talked about fitting the data Germany.

B

God detail both 15 and dummy Holly I know.

A

B

A

Is brave enough to try to explain what we mean by fitting the data.

B

Knowledge about my 15 days answer it we have many of you are. Are you yeah yeah.

B

Did Aria pigmented or quickly equation, alitalia,.

B

Y equal to X equal to Y yesterday, well.

A

Yeah but I'm, so here's our Peters.

G

Our observations.

B

Okay, that's not yet tonic. Okay I did that the day that we selected yes,.

A

So here's the day of that week collected don't.

B

You materiality that not be X and.

A

This is the earth ball, so this is def against recent statistic.

B

Dragons are awesome, yeah the Guv medium order, a little bit of all Tommy, so.

E

B

Like guitarra yeah, a net or not yeah sure resistivity, so.

A

This, what what kind of bird is is this does.

B

A deal ooh more about no mean he was not limited it what's.

A

The resistivity structure of the earth.

B

That resisted these struggles.

C

A

Is this is the debt that this is the resistivity and now that's just a straight line.

B

A

It was just one resistivity and it's 218 ohm meters, but.

B

In the other session, only done now.

A

I can take this resistivity and carry out a winner experiment and get observations.

B

So no we're now trading at D means u19 ago.

A

A

What is this blue line? Doesn't.

G

A

Me well, it is row way, but it's how? How did we get? How did we get this line.

B

Booyah I do the language.

B

A

I do a winner sounding with a B over 2 at 10. Then this is the result that I'm going to get.

B

It was a beo if you hit a sonic come up. 10 meter got a position. My tiny time. I got the resistivity, oh yeah, so.

A

For a uniform earth every time I expand by when RA I get the same number I get the true resistivity of the earth. So.

B

Division of the ions of yours in a bilateral debate, maybe most minutes out of the same anti-gay, they loved I, never know JV, tiny, D versus TV. Our own unique forms in eternity news now that is engineer needs a new style yeah. So.

A

Does a uniform earth? Does it explain the data.

B

Uniform after uniformity or not data.

A

Given or given these data, is it possible that the earth could just be a uniform half space with 218 all meters.

A

So, given these data, is it possible that a uniform earth, which we also call a half space, that a uniform earth with resistivity of 218 ohm meters, does that except ibly reproduce or explain the data.

B

So you do not have space, but not even a half space communion I'll take half Spacely using any stop. You know da McDonough me and immunity need Towe have space with your Katara hippy not eat our God. We have space for resistive, Disney artists, a ship, an inertia destro of young Nikita.

B

Can I can I make a short vacation again just mean I. Don't want to confuse them like those are those data's like describe the half space like uniform, not half space with the resistance, so we.

A

Went out to the park, okay.

B

A

Go to the park, we get these data. Ok,.

B

Bonnie Buratti Diyala here and.

A

Now we want to find the earth model that reproduces those data. The.

B

Donor, our motor, it looks in it, Mariana Jimin is Ian Moir. Well, no did Arianna the same me more. A young girl to sit down.

A

Today, if we have a half space, a uniform earth of 218, all meters, then this is what we would expect so.

B

No, no, not me either said versus Omega are a new G news, not Toby love it and inet1. Lady love with the resistance of a resistivity Shira, so I know the Romulan. We are I'm already developed and I'm already a member. So.

A

My question is: is this blue line a satisfactory in satisfactory agreement with.

B

The dots nearly entire general can I have the unique online DD yeah I got me two rock bands and want analytic be a Saladin. Is wrong? Did Nicole put Shana Katara mr.

E

Bhushan imagine of like a nebula there.

B

Are the more appealing because I am worried about depends on what I need a more a very turbulent Malaga? No.

A

Perfect, but if we look at these blue lines, which are like the predictive data from this model, it gives this that is not in good agreement with the.

B

Observation so general, okay, but Dino. It did not get down a general comment below so climb or anything.

A

So it doesn't fit the data.

B

So these D Model D I did a net fit typically, so.

A

We don't believe that the earth is just a uniform half space of 280 meters.

B

Dinah dinah dinah get an EGD. The other section of below then do the painting chip is alright. You have space for the rope. Ok, ok,.

A

So we need to adjust the earth resistivity.

B

But I'm gonna general D D D done it guys, ok, make resistivity, go don't get spinner on this chain here. How.

A

Are we going to do that and.

B

Our money in a Cheney will do so.

A

What's your thought process, you have these data.

B

The Italian, honey and.

A

There's information on here: how are we going to use that to change our earth model.

B

Does eternity Martita in Moorea also dude? Are you happy done? Do you think I you don't get off, gonna be ready, he'll, never come already at model box, yet model go Munna group gambling.

A

So what is the first thing that we're going to do to change? Don't.

B

B

A

And how are we going to change the role? Why do we want to change it down.

B

Below, if I jump down, then I'm.

E

B

B

The first part is like a five hundred row: yeah.

A

Right so this is good look at the apparent resistivities and this is small- a be upon two. So that means that you're sensitive to something that's near surface, so you're thinking that maybe the near surface here should be equal to 500, ohm meters.

B

The peons are not even without the gravity like that property later Kakuna horizonte' X, booty, like that, has a lot to up my enough. She know me Tommy, daddy, well, I, know Matt I'm, Holly, AP, yeah, love, Medina, so nee nee, nee, nee, nee, nee, nee.

E

Amma Tanganyika.

D

Although Marrero commend.

B

You we are about DNA and is associative law a resistive determined, oh yah, oh yah. Oh.

A

Okay, so let's change Row one row, interchange down.

B

A

I'm going to make it and watch the curve watch what happens change it watch what happens.

B

In Gujarat down well Israel.

A

Okay, so here was here's our uniform earth and here was our predictive data so.

B

Totally unique on the drone coming in so I gotta be yeah, II, don't know come on ty. We do no come on. Take that so.

A

We want to make this point here in the prediction up here to match so.

B

Generally, Diego DV what you, what wasn't one in T sound turtle came out and.

A

Now we're going to do that by changing the resistivity of.

B

Resistivity to finding denominators see how you go to y'all go on look.

A

I raise it up to 295.

B

A

A

So that's good right now we we matched this point under the egg. Booty go Bobby.

B

We're not doing very well here. Amanda Nate sent me an Apple TV, but about the appointment, TV it's and they're, pointing into a taking, but.

A

We've got other other parameters, we've got the thickness of the first layer. We've got the resistivities from this at the bottom.

B

Comes on my head told you nothing Bobby, you know the resistivity Bobby great.

A

So what now should we do.

B

A

Change row: okay,.

A

We're going to change Row 2, which way should we, which way, should we change it? Do we want to be bigger or Lord.

A

Is way up here right, it's too resistant, so we got any bring it down, yeah! Okay! So let's bring it down.

A

So H what we decide to change right, I'm, going to bring it down, see.

B

What happens to changing it.

A

How much you want me bringing down lovely.

B

Little Charlie.

A

Fifty-Fifty: okay, that's 55, pretty close.

A

Okay, so at least we're going down right up that.

B

Lobby, that's good, oh, come on, but.

A

The curve is shifted so.

B

You eat like, what's.

A

Another parameter that we could.

B

C

A

Three: okay, so a to you make roll three.

B

A

Higher make Row three higher, so Row 3 is 2. A mm mm.

B

Did that change very much? No.

A

Right so this is interesting. You have parameters, but they affect the data differently. So we just change the bottom resistivity and it really didn't affect things very much. They.

B

Tell me with a machine.

B

A

To what H, okay so.

A

H 110, Wow, yeah I, think you've looked at these data before, let's, let's take it a little bit easier for people who don't know what the answer is right, it's always easy to know how to change things. When you know the answer, so you're not allowed any anybody. What are we going? I mean, let's think about what what happens what's H. One now is.

B

H one block she mentally like around him. Well.

A

How do we know how do we know that a little.

B

A

So h1 is an issue here, it's at 40, okay, what do you what is going to happen if I make h1 smaller h1 cool, yeah.

B

I miss it, but it's.

A

Going to happen to this curve, yeah it'll happen earlier right. It's right now. Everything that happens for large AP of pot too, is sort of costly we're seeing deeper right. If we want something to happen earlier, okay, then we have to reduce the space, so we've now make it shower. Watch what happens so this is now can.

B

I, okay, yep! So now the a be your debut in the neck. Nato this way up or might.

C

Not let me get p9 s really, why not at the after.

B

All, oh, my no, it goes. The engine has enough a be teenager. Yeah, yeah, yeah.

B

A

Let's watch I'm going to change this. It's currently about 40, so.

B

A

I gotta change, h1 to let's say 30: oh they.

B

Love Delta minute: okay,.

A

I got change it to 30.

C

A

What happening it's moving Oh ba.

B

La vie, do you not I can.

A

Change it to 20.

C

A

Say attainment.

B

But I'll order, my let's see like no wonder that was a lot takes a mom. You know twice.

A

It's better right, so we're missing.

B

Out on the nipple Neil IV, not.

A

Work we're closer here right, so it used to be that this line is over here and now we reduce that we.

B

Show like these are the developer in the lobby, so I I could even go smaller.

A

You I could go to ten, they are party lemon.

B

A

So, let's suppose, I go to champ just.

B

C

A minute because.

A

Somebody has look.

A

A

Look at that good.

B

Job and tang up Oh girls of today yeah.

E

B

Meto Halligan: oh thank you, Sheila, a more donate d diamond yeah.

A

So now I've got the okay I'm fitting these first few points here. So.

B

Long she did once naval new I know.

A

B

So good right, big job, no.

A

So what's what's happening here, so we're getting something that is much less resistant, so we're supposed to be.

B

Up here then uh phone that much yet Montserrat that man is this to be out job and.

A

It's also happening at a different location, so this is happening deeper. Somehow that means that it or model is not right.

B

A

So now, what what should we change there? Bob down money h2, so we're going to change so this is h1. Right goes.

B

A

So 500 ohm meters- this is h1 1010, and then we got this guy here and where we're going to make this shout Swami right.

B

A

Your limit, what should we make.

B

A

Make it smaller okay. So, let's see, let's see how that happens is right. Now h2 is 200 right. He's.

B

Really got me beat me time they on to the doctor, so.

A

Let's, let's maybe naked 100 the.

B

A

There's a hundred.

B

A

So look at this so now this is this is a twenty five, so you can see how much closer it is. The scales on the figure are changing if we had kept a fixed scale. The initial curve would be down here. Okay,.

A

B

Scalable Hedren teaching, you know yeah.

G

B

Most likely outcome: okay,.

A

So how are we doing.

A

A

So that's looking pretty good right, yeah nice.

B

That Bobby yeah.

A

That's where we're not quite there yet right, it's still too big, so.

B

I'm gonna know me I'll, be there not that people, let's.

A

B

It even smaller I nearly drowned in it.

A

So we could go a little bit smaller. Maybe let's go over. Let's go to small: let's go five I'm.

B

Not tossing oc5.

A

It's way above here right used to be ball, so mine is definitely too small I can't they long.

B

But that Dean is too big: fifteen gotta. Let you.

A

Try ten, what a good number can definitely make this way too easy sorry.

D

You'll find out what happens.

A

Okay, so now, let's let's review some things here.

B

So the verge- you don't even know luckily so suddenly anomaly, it won't involve the same thing. You go up Jana atomic time of the oven at door, TV directly mama. Ever though about the over net, but.

B

A

You that's the whole point so now, let's, let's recap a little bit about what we, what we've been doing.

B

So what Logan is a member of you don't see out so.

A

We started off with an earth model which was completely uniform, just a half space, so.

B

A half space cool, no one, Guinea farm sheep on his right, montenegro socket number. Now it means up down your name, two more space, 280.

A

No meters, but right through here, yeah.

B

That's that that I know but um yeah the Chuckle on it. It.

A

Was not a satisfactory.

B

Fit to the data so dude, oh yeah, daddy, donate no time or in that it's housing about our.

A

Fit to the data is quantified by how close the blue line is the predictions compared to the opposite.

B

On me, yeah I'm, teasing shooter block me so much. It is Rockaway generate more come down. Cerrada.

C

A

So this is dad. This is ro, that's what we got so this is. This is our earth model.

C

A

And then we have to go through a computer.

A

To do what's called the forward modelling.

A

To give us our data.

B

Okay, mother, our mother, she met Rebecca, an appropriate resistivity and, in the same vial of letter, D the same to motorable. What didn't love me.

A

So here's here's some jargon so the earth, the earth model we do what's called a forward modelling or sometimes we'll use the word simulating.

B

Simulate to make them that it forward more than he doesn't see me so I owe more than going out so.

A

The word simulating means forward modeling we're predicting what the data should be from.

B

This model a model under the same day, Daria palooka, Malaysia Raul, the saint is he crosses, live Davao simulate like that one. So.

A

We go from the earth model and here's all of the physics is in here the physics and the computing.

B

That my physics yet computing Yvonne memo so.

A

We don't we're not going to look at that in here, but for our case here we have a way of solving the D. Series is stiffening problem to compute the forward responses, so.

B

Shall respond to growl? Don't get an arrow use, a set of blue a minute I'm a tough shot so.

A

These were the predicted responses from here, but my observations suppose my observations for like this.

C

B

My uder be better off more though that's it all yet come on time. Model no come on time. Where is work? Massa not llamado. You're, like am artistically daddy, yeah, daddy, daddy, knows what amazing yeah you had a model neck. Look! What I bought.

A

So what I want to do is to have my forward model. My perch predicted data, that's the other word will sometimes use is predictive data.

B

The mutant re-balance wrong for modeling minute at the moon in Escalona, viva la oh, come and heed our commentator predicted yeah.

A

Yeah so simulated data predicted data forward model data. Unfortunately, we use all of these terms to mean exactly the same thing. So not.

B

The the global Mme I'm, not I'm, gonna forward modeling ladanyi, simulating de delito, they predicted dolly, come on dee da ba come on come on time. Me me, multiples, int, dialogue, no in.

A

Most of what we do in version we're actually probably going to use this word the most. We have some earth model we're going to predict what the observations should be.

B

Generally, update of the model of magnesium or don't love me, nobody nodded assent, you know, can manatee determine and.

A

Our goal is to adjust this model so that we make our predictions in this case the solid line in good agreement, with the observations.

B

Observe the observational time come, try litigated on it. There are commonalities and a guy go to the change.

A

So this this is not a good earth model to explain these data. We.

B

Did our shimmy a Buddhist it'll cause a mood to go bigger the after motive model got no ha no take accounting wanna, so.

A

These distances are too big. I was.

B

A

Remember yesterday, I introduced a misfit.

B

Malia sit down how about me.

A

And it had D, that's the data, the predicted data, minus D, I observed data and then I squared it.

A

So what that is it's the distance between the observations and the predictions so whatever that is okay and then I'm going to square it.

B

A

B

I, don't know the formula your a before me. Diet would not be a co-op. If you see a hobby predicted data predicted I am, you know, observes original covenant, eat I ain't, a couch on awake, your mom, not touching I got nothin I'd ever look like a Formula D the grandma Mina day when today I'd like honey. How about it.

A

But I want to look at everything at once, so I want to look at the misfit at this point. This point and this point and me.

B

A

Something that.

A

Accumulates the total squared misfit.

B

Mm, oh man, when they get harder yeah. How do know wha? No, no sir I, don't I saw Mitch I'm going to do something tomorrow. Finally, get the ball out of.

A

So this is the so we have end and data. So in this case, I've got one two three four.

A

So in this case and would be equal to eleven and not.

C

E

A

Yeah for each point, I look at the difference between the observed here and the predicted I find that value I square it and then I put it into the Sun. So.

B

China ah me not acting like that, be if you have a pocket.

B

A

And now I want to make this disagreement small and it's convenient to have a way of deciding okay M. Am I close enough so.

B

Don't be happy, the quality jacket better of al-azhar so don't mean. Maybe now take our neck. Commenting on it come tacky doesn't need something. She be last wrap your app so.

A

In general, any data that I measure so here's row of parent as a function of a be upon to any datum that I measure actually has some error to it, error an error, so the data is never it's never exact. So my observed data is equal to go the true data, what it should have been plus some error: noise.

B

The predictive d-darling, you know whatever kid in my way they talk. Thank you yeah Kenny. We should see my Sigma, so my impression is good about it. So.

A

Every day, job that we measure is inaccurate. Look every data every day took that we measure is wrong. It's inaccurate, so.

B

General, don't kind of did I mean in a modify some.

A

Data are really have high uncertainties, other have lower so in here for.

B

Instance of the diorama theta Rho did a prism days, ample Sigma americano.

A

So are people familiar with gaussian statistics, gaussian.

B

Statistic that.

A

For now, let's just think about error bars. Do you ever plot datanode, with some kind of an error bar on it and.

B

Robotic tail would it who they died, of taking the not it what.

A

This error bar does is indicate that the the value that I'm taking here could be up in here or down here, but the true value. Okay, it's someplace in Figure, so.

B

The guy over yonder, that's what I need! He didn't come to etomidate Johnny economy, that man there may not remember back when you're on buddy boy, hello, child on it d time up your name. Okay, so.

A

B

Of your data yeah.

A

So sometimes sometimes the error that is associated with the datum is really big, and sometimes it's really small. If it's really big, then we plot it with a big error bar on it. If it's really small big.

B

Quadrant, with a smaller than 800 IMU is ain't; no ID, they ever bottles a minute. That is enough. Okay,.

A

So how do we think about that? And what is what is the error bar so suppose that I've got a.

A

Day, jump! Ok, that's sitting out here. So this is my datum value.

C

A

Saudi I observed is equal to di true plus some error.

A

Then I can look at what the.

A

Yeah, what the probability is that, let's see, is there some connection here?

A

D

You talk about yeah.

A

How many people know what standard deviation is standard.

B

C

B

C

A

It's sort of variance. Yes, it's like this. So here's here's my true value, okay, but then, because I've got some error, then maybe the value I have could be. You know up here or up here or up here, and what the what this curve does, and it's actually called a Gaussian curve. Gaussian.

B

Curve the cost an inguinal Yoda, it.

A

Tells you what the chance is an upper lying close to the truth, value and there's two things here. One is called the standard deviation.

A

And that characterises sort of the width of curbs.

A

Yeah it sort of characterizes how likely it is. You are to find a value, that's away from your true value. So here's your true and.

A

If you're so suppose that you are rather than bouhanni.

B

That wound up watching it. What true better you look, the main number, whatever swap everybody aeroponic at sorrow, create how the custom craft Nadella Tech has a drop little bit. America come on idea with a lot. Would you name of achievement that man, hello, MA, your money, daddy's loopy, never know after being in our Western macro, served on Mathias to be knows of your nothing like it, I'll be mean way: yeah, we have an idea. Will you be talkative or not so.

A

Suppose that we have a true value of a number, so let's suppose that the true value is equal to seven point: zero. Okay,.

B

A

Then you've got a noise okay that madman as a standard deviation. Let's say one point: zero, okay,.

B

Then ever I got to say yeah.

A

Then you can imagine a hat okay in which you know are all these all the numbers that are kind of possible for you to draw to pick, and the number that you get is always going to be the true value, but plus some noise, that's associated with standard deviations, so.

B

I'm betting, my that's! How nice.

B

A

A

So suppose that the true value is equal to seven point.

B

Zero: two zero seven one: zero look at Muhammad and.

A

The standard deviation was equal to 1.0. Will.

B

Not be the standard deviation, not one point. Zero notable have my.

A

Observed value is equal to the truth, plus some no.

B

I'm sorry was that all through value net noise, Department and the.

A

Only thing I know about the noise is the standard deviation, nice.

B

A standard deviation so that deviations are nice to go out of it now.

A

I'm going to take different realizations of of this I'm, going to kind of pick a memo, no I'm going to make numerous experiments in which I try to obtain an observation. The.

B

Vision did I have with a dose-dependent after that, so.

A

My observations, okay might be so since this is a noise standard deviation of.

A

A

So then the numbers that I might hit that would just happen. I might get the exact number 7.0, but there's a good chance. I'd get something at 7.9, 8.2 6.5.

A

Maybe even 4.5 I get all of these numbers that have a range that looked like this, so there if I, wanted them on on a scale. I would see that here is the true value 7.0, but I get some things that look like this. Just keep doing this and I'm going to end up with a number of samples that are seven points, 17.5, etc.

A

Okay, so that's a standard deviation of one, and that means that I get a true value of seven, but it's fairly likely the number that I measure is going to be eight or six, because the standard deviation is is one plus all bogus one yeah I.

B

Extended deviation, a plus or minus.

C

B

Two values of email, a so.

A

Say yes, we never I want cinema.

B

A

The true value is, is seven but there's noise right, so maybe I take out. You know seven point five or maybe you pick out six point eight so but they're going to be around seven, but you know plus.

B

B

A

That's what the.

A

That's what this value is, so here's here's the data and here's seven points zero and we pluck plot the value. And then we put a error bar.

B

On it so Aramaki like a demo and so.

A

This so that's seven, this might be six and this might be eight. So that's sort of plus or minus okay.

E

A

Everybody kind of familiar with this.

B

A

Must have had this.

B

A

So the point about this.

A

The point about this is: if the standard deviation of the noise is large. Okay, then the error bars are big.

B

Right, it's none. Deviation of them, yeah, Anna, Beninati, missing a Ravana demon.

A

So suppose, suppose that the standard deviation was equal to zero point, one okay and I do my experiments. So now I get. You know, seven point zero at seven point, one six point: nine 6.85 7.13! You know I get all of these numbers, but they're really close to seven seven.

B

One he says, though she may look. Everyone I'm, sorry, nice guy. Let's do a point, one sort of either white love blah blah. If.

A

I had a standard deviation is equal to ten okay stay. Not my true number is seven. That's what I want so I might get seven so.

E

They might a bit you.

A

Know minus 1.2 I might.

B

Slip to know sixteen.

A

Point three yeah: it's big error right, Iran,.

B

Tod United, nobody I know so.

A

The size of the noise is really important, so.

B

Nice yet, but mine are very easy, so.

A

What we do is with each datum. Okay, we also supply some estimate of.

B

The noise so I doubled a time. Yeah imma drop comment on it come and you know table.

A

So now we have the following in this diagram here. Really all we cared about was trying to make this right on there wait. So we just want to make that misfit equal to.

A

Okay, so that's that's good! So if we look by eye, then that is as it is.

B

Really really good I'm in love in cement. I. Don't comment about that. Yeah he's on my go.

A

But imagine what would happen if you have some data in here that were just poor data. They were, they were really wrong. They had a big error bar so now I've got some data that I've got the datum that looks like this, like this I've got a datum that looks like like that, and I've got a datum. Let's put another one that has really small airports if I want to make it fit to the data. I might I'd like to try to fit everything within.

E

A

Standard error bar so.

B

Doc, the general Yamashita Muslim ya era, I got particle and robotics I, don't know Oh D job, but these are gonna. Let me do I ever a Paulina job. Yeah. Don't look at me with Arab army anyway Miami, oh yeah, that would ever bother my walk around myosin on their teammates.

A

So if I fit in my observed data or my predicted data, you'll kind of came along in here and fit each of these within the error bar that it's like I'm good to.

B

Go so everybody means out when it as an ally. Yes, indeed,.

A

So when we talked about a fit to a straight line yesterday, so we had points like this.

A

So we had points like that and then we tried to fit a straight line. You know we might end up with something that looks like that.

B

The only viewpoint years she doesn't know communion edu on a table if.

A

We take this same example, because if we look at that we say no, maybe that's not a good tip right, because this we don't hit these points.

B

Be pointing go Mattie, Tejada, t19, yellow magenta, dee da dee gum when you donate a puppy, but.

A

If I have error bars on here ever.

B

Bother shitty misleading so.

A

If the error bars were like this good I'd say a lot of buying, she.

B

Missed I know: that's eat a dog, that's a new idea. Okay,.

A

So there's some important points here beyond what we're just doing, and that is the concept of fitting the data study.

B

Well, your honors or the concept approval if it ended up dating sorry, so.

A

When you, when you work with data, every datum must have some error so.

B

Don't taste it now money did that I mommy ever did oh yeah I! Think that won't be done.

A

So every datum has got some error bar and our goal is to try to find something that fits. You know all of these through their.

B

Error bars wine, coolers, even I, ever volunteer in the eye and.

A

To prep, for one thing that we're going to do tomorrow,.

A

Instead, instead of looking at this point individually, ever I had this some, so we've got data that looks like this and they're going to have errors on them, and some of them might be big. Some of them might be small.

A

Our final misfit remember this. Some miss.

B

Big wish I'm missing I.

A

Equal one to n be my predicted. That's the simulated data minus D I observed.

A

B

A

Was our total or total misfit.

B

Daddy, the 15 matam, oh yeah, prom misfit get a small.

C

Time, Bobo.

B

Is similar to standard deviation, I'm.

A

Going to get there in just a second I haven't introduced the standard deviation. So this is this would be the total misfit if all of the errors were of all of the day to have the same standard deviation that doesn't matter, but if the some day to have different standard deviations, then we want to divide this by let's say epsilon, where epsilon is equal to the standard deviation standard.

A

Deviation of the data, so in this case this particular game would have a big value. This would have us a smaller value, and this is this. This is one of the most important equations that we will deal with in inversion.

B

So in Russian my little take that beanie I met equation because.

A

We we need to decide if we have a good fit to the data or not they're.

B

Bouncing oh yeah, there's a DJ.

E

A 15-time, it's also getting.

B

A

All right, uh let me just do one more thing and then maybe we'll take a short break.

A

So maybe they take a short break now, yeah.

A

10-Minute break okay. So what we'll do we'll come back and never going to load the field data set that we did yesterday.

C

C

A

C

A

Right so, let's begin again.

A

So, first of all, any questions.

C

A

Make one comment here.

G

A

We have a situation here. Here's our apparent reasons, dividends, that's a be upon to our observations. Are these X's and our predictions from some model are given by this line here? My question is: is this a good fit to the data?

A

A

You think it's not a good fit to the data, because you think you're supposed to be fitting the data exactly, but if each datum had an error bar okay. So let's suppose this one has an error bar like this, and this one.

A

If the error bars are like that, then the question is: is this a good fit to the data.

G

A

Okay, so that's the point, that's a very, very important point. We can't decide. We can't make a decision about whether something is a good fit to the data until we are have some idea about the uncertainty or the noise in the data. So.

B

The data and even the beads are moving up on deck I, totally I love. You.

A

So you remember yesterday you are great people in the field. Totally a plus field surveying people.

B

B

A

You you pounded electrodes in your father's right tables: here's em, here's em, a B move. It around got the parent resistivities and we collected two sets of data, but.

B

Yet I'm you will not see, does that missus yeah so.

A

The data set that you had from the SIS Cal system expanded the full length of the array or 200 meters from A to B.

B

A

With friends or Kevin system, it was somewhat shorter and we're going to compare the data between the two of them. So.

B

The cou you did what, with 19 0 minute, do you get man setting it to use this girl?

B

A

I first like to do, is: have you look at the data from the SIS Cal system and then we're going to talk about the data and we're going to then interpret those data in terms of a three-layer model. So.

B

Really and will donate interpret about it, go get it so.

A

What I want you to do is to go back to the three layer. Conversion app be there.

B

In boston applicative, a single buh-bye yeah.

A

And let's start again so go to kernels and click on.

B

Their money Bob, you look at the energies. Every stock, Nick, Lea output, spun it around.

A

Okay, okay and now you're going to do the first cell shift enter shift, enter shift, enter.

A

Okay, so come back up here now to step one step: one who lb load, observe data and then click on me down downward arrow and now Alisa.

B

A

Now, let's look at heart data winter ducks CSV.

B

A

Don't why nots.

B

The SVG bomb, okay,.

A

A

Everybody there who's got it ok and then press load, and now it tells you that the data are loaded.

B

Yeah what I mean so.

A

Now we can use the same app that we had before.

A

And that would go down to the manual fit so shift-enter, then.

B

You are Fitness, loving and.

A

Then you should be up to something that looks like this. Did.

B

I give you that development.

A

Everybody everybody here, okay, so remember the very first datum that was taken with the SIS cow. Everything was just like one meter, one meter apart and you collected data and the first datum was 297 ohm meters.

B

297, oh me that she'd have to myself. Why not? Okay.

C

A

The electrodes and the next day tomorrow.

B

A

Think it was 217 may.

B

Not wanna say yes, the co-op's, yeah and.

A

Then, as we went farther, we went up.

E

A

I had been dead, the way be charged I think this was like.

A

B

uh But what is 63 yeah.

A

So those are our data. There's not very many. There's only two, four six there's only eight data- that's not very many, but still it's got. It's got some nice aspects, there's certain things that are a little strange, maybe like this, because it looks like starting up at a high resistivity that sounded okay. But then this point down here, like that's pretty low, and then it goes back up that kind of goes really low. It.

B

Goes back up, but then this thing was not that bad mother, no society, no resistivity, VNA me too. On the side. Without of a body you, don't it tell what am I taught I got out of community, never to.

C

Send here, but.

E

No get that what.

A

You know you've got the entire. It's.

B

A

So now, if you're going to manually, fit these data you're going to have to make decisions about how accurate how much noise is on these data and what you regard as a good fit to the data.

B

So I'm a nice, not population, arm up to body booty video, my lovely 15 job.

A

So what I'm going to do and you're.

E

A