NASA ARSET_ Techniques for Plotting and Analyzing ATL24 Coastal Bathymetry Part 2

0:01 Speaker 1: Welcome to part two of the ARSET training NASA's satellite

0:04 Speaker 1: laser altimetry for coastal and nearshore bathymetry. My name is

0:08 Speaker 1: Sean McCartney from NASA's Goddard Space Flight Center in Maryland,

0:12 Speaker 1: serving as your ARSET instructor wherever you are joining from. Welcome.

0:18 Speaker 1: The following slides provide an overview of the two part

0:21 Speaker 1: webinar series. So why would somebody want to take this training?

0:26 Speaker 1: Nearly eighty percent of the world's oceans are unexplored and unmapped.

0:30 Speaker 1: Successful coastal management depends on monitoring and mapping. Coastal bathymetry

0:36 Speaker 1: is essential for navigational hazards for vessel operations, tidal modeling

0:41 Speaker 1: and prediction, tsunami risk assessment and forecasting, underwater cultural heritage preservation,

0:48 Speaker 1: and environmental change monitoring. Satellite drive pathymetry enhances capacity for

0:54 Speaker 1: collecting high resolution, accurate depth data supporting varied applications. By

1:02 Speaker 1: the end of this two part training, participants will be

1:05 Speaker 1: able to identify NASA bathymetry data used for global coastal

1:09 Speaker 1: and nearshore bathymetry mapping for risk reduction relating to shipping

1:13 Speaker 1: and navigation, identify the applications and limitations of IAT tube

1:19 Speaker 1: athymetry data or ATL twenty four for coastal and nearshore

1:23 Speaker 1: bathymetry mapping, Plot and download IAT tube athymmetry data using

1:28 Speaker 1: slide roll web client in analyze IAT tube bathymetry data

1:33 Speaker 1: using the slide roll Python client. The prerequisites for the

1:37 Speaker 1: training are two are SET courses, the Fundamentals of Remote

1:42 Speaker 1: Sensing Course and the Mapping and Monitoring Lakes and Reservoirs

1:46 Speaker 1: with satellite observations. From December two to December fourth, there

1:53 Speaker 1: will be two one and a half hour sessions which

1:55 Speaker 1: will include presentations, demonstrations, and question and answer sessions. All

2:01 Speaker 1: materials and recordings from each session will be available from

2:04 Speaker 1: the training web page. If you are not able to

2:07 Speaker 1: attend one part, a recording will be made available within

2:10 Speaker 1: forty eight hours on the RSET training web page. Homework

2:14 Speaker 1: opens today December fourth and will be due on December

2:18 Speaker 1: thirty first. A certificate of completion will be awarded to

2:22 Speaker 1: those who attend all live sessions and complete the homework

2:25 Speaker 1: assignment before the given due date. Part two of the

2:30 Speaker 1: training is focused on techniques for plotting and analyzing ATL

2:34 Speaker 1: twenty four coastal bathymetry. The objectives for part two of

2:40 Speaker 1: the training follows. By the end of today, participants will

2:43 Speaker 1: be able to identify the applications and limitations of i

2:47 Speaker 1: AT two pathymetry data for coastal and nearshore bathymetry mapping

2:53 Speaker 1: plot and download i AT two pathymetry data using slideral

2:56 Speaker 1: Web client in analyze i AT two pathymetry using slide

3:01 Speaker 1: rule Python client. Please put your questions in the questions

3:06 Speaker 1: box and we will address them at the end of

3:08 Speaker 1: the webinar. Feel free to answer your questions as we go.

3:12 Speaker 1: We will try to get to all the questions during

3:14 Speaker 1: the Q and A session. After the webinar, the remainder

3:17 Speaker 1: of the questions will be answered in the Q and

3:19 Speaker 1: A document, which we will be posted to the training

3:21 Speaker 1: website about one week from today's training. It is not

3:26 Speaker 1: my pleasure to introduce the guest trainers for today's webinar,

3:30 Speaker 1: Doctor Christopher Parrish, Joseph Paul Swinsky, Gretchen Imohori, and Keana Kief.

3:36 Speaker 1: Doctor Parrish is a professor of geomatics in the College

3:40 Speaker 1: of Engineering, School of Civil and Construction Engineering at Oregon

3:44 Speaker 1: State University. He also serves as director of the Geospatial

3:48 Speaker 1: Center for the Arctic and Pacific, a university led advancing

3:52 Speaker 1: research and education in geodesy, geomatics, and geospential geospatial engineering.

3:58 Speaker 1: He is past president of the America Society for Photogrammetry

4:02 Speaker 1: and Remote Sensing. Prior to joining Oregon State University, he

4:06 Speaker 1: held a position of lead Physical Scientists in the Remote

4:09 Speaker 1: Sensing Division of Noah's National Geodetic Survey. He also serves

4:14 Speaker 1: as an affiliate faculty member in the Center for Coastal

4:17 Speaker 1: and Ocean Mapping Joint Hydrographic Center at the University of

4:21 Speaker 1: New Hampshire. JP Swinsky is a twenty four year NASA

4:26 Speaker 1: Goddard software developer who spent most of his career creating

4:29 Speaker 1: flight software for spacecraft and science instruments. He led flight

4:34 Speaker 1: software development for the Atlas instrument from twenty ten until

4:38 Speaker 1: twenty sixteen. In twenty eighteen, he transitioned to cloud based

4:43 Speaker 1: data processing solutions, and in twenty twenty co developed slide

4:47 Speaker 1: role with the ice AT two Project scientist and University

4:50 Speaker 1: of Washington, a public web service providing fast cloud access

4:55 Speaker 1: to ice AT two data. Gretchen Imohori is estabbed scientist

5:01 Speaker 1: and Satpathy Project manager at Noah's National Geodetic Survey with

5:06 Speaker 1: twenty six years at NOAH, she supports hydrographic and remote

5:10 Speaker 1: sensing operations. She holds a Bachelor of Science and Chemistry

5:13 Speaker 1: from State University of New York, Buffalo and a Master

5:17 Speaker 1: of Science in Earth Science Ocean Mapping from University of

5:20 Speaker 1: New Hampshire. Keana Keith is an analyst programmer with two

5:25 Speaker 1: plus years experience working Chris Parish's research group at Oregon

5:29 Speaker 1: State University. She has a Bachelor of Science and Computer

5:32 Speaker 1: Science from Oregon State University. She is a lead developer

5:36 Speaker 1: for noah's Comprehensive Bathometric LIGHTAR Uncertainty Estimator, SATPATHY Photogrammatic Shoreline

5:43 Speaker 1: Uncertainty Estimator, and a contributing developer of NASA's ATL twenty four.

5:50 Speaker 1: Chris over to you.

6:00 Speaker 2: Thanks Sean, and hi everyone again. My name is Chris Parrish.

6:02 Speaker 2: I'm with Oregon State University and I'm really honored to

6:05 Speaker 2: give a demo on ATL twenty four and how we're

6:07 Speaker 2: going to use ATL twenty four and Sentinel two spectral

6:11 Speaker 2: imagery alongside each other for mathemetric analysis and also maybe

6:15 Speaker 2: looking at mathemmetric change. So to set the stage for this,

6:20 Speaker 2: if you're going to follow along with me. There are

6:22 Speaker 2: some things that you need to set up ahead of time.

6:24 Speaker 2: So first of all, there are a couple of files

6:26 Speaker 2: that you should have been able to find on the

6:28 Speaker 2: training website. There's a shape file which is called hyannas AOI,

6:33 Speaker 2: and then there's also a Python script which is called

6:36 Speaker 2: psdblinear aggression dot PI. So you want to go and

6:39 Speaker 2: download those two files if you haven't already done. So,

6:43 Speaker 2: you also need to set up an EU Copernicus account

6:47 Speaker 2: in order to be able to download Sentinel two imagery.

6:51 Speaker 2: You'll also need to download and install qgis also known

6:54 Speaker 2: as QGIS. That's free and open source GIA software that

6:58 Speaker 2: we're going to use in this demo. And there's three

7:00 Speaker 2: sites that you want to go ahead and bookmark. Once

7:03 Speaker 2: is the Copernicus Browser, the second is the slide Rural

7:06 Speaker 2: Earth web client, and then the third is the Noah

7:10 Speaker 2: Webmap Tile Service Electronic Navigational Chart. So here's our project site.

7:18 Speaker 2: This is where we're going to be working for this demo.

7:20 Speaker 2: This is Hyhana's Harbor in Cape Cod, Massachusetts. This is

7:24 Speaker 2: a site that is meaningful to me because it's actually

7:27 Speaker 2: where I started my career. At the very beginning of

7:29 Speaker 2: my career, I worked on a Noah Hidrographic Service ship

7:33 Speaker 2: and this is where I first met the ship and

7:34 Speaker 2: where we spent a few weeks working before transmitting down south.

7:39 Speaker 2: So just to kind of take a quick look at

7:42 Speaker 2: the site here, one thing that you might notice is

7:45 Speaker 2: that there's a lot going on in terms of where

7:48 Speaker 2: we are. So this is on the south shore of

7:50 Speaker 2: Cape Cod you can see in the in the image

7:52 Speaker 2: up in the upper left where we are relative to Boston.

7:55 Speaker 2: But in terms of things going on in the harbor,

7:58 Speaker 2: there are a lot of coastal engineering project you know,

8:01 Speaker 2: engineered structures, some of them going back nearly two hundred years,

8:06 Speaker 2: such as the breakwater, the original breakwater. We've also got

8:10 Speaker 2: a jetty grindfield, other shoreline protection structures. There's a dredged

8:15 Speaker 2: channel that's used by the ferry. And then there are

8:18 Speaker 2: also a lot of storms that impacts this area, including

8:21 Speaker 2: a lot of times large no o' easters. So really

8:25 Speaker 2: a lot of different reasons why there could be changes

8:29 Speaker 2: to the bathymetry and the site and why we might

8:31 Speaker 2: be interested in monitoring with bathemetric change over time, and

8:37 Speaker 2: just to make the point that bathymetry in this area

8:40 Speaker 2: has been of interest for a long long time. Here

8:42 Speaker 2: we're looking at an eighteen ninety four US Coast and

8:45 Speaker 2: Ganetics Survey chart of Hyannas Harbor. So quickly, this is

8:52 Speaker 2: the workflow that we're going to go through. We're going

8:54 Speaker 2: to start with just a visual analysis of our project site.

8:57 Speaker 2: We'll do that in QGIS. I'm just a kind of

9:00 Speaker 2: gain familiarity with the site. Then we're going to go

9:03 Speaker 2: into the slide role web client, and this is where

9:06 Speaker 2: we're going to be able to view and download ATL

9:10 Speaker 2: twenty four with imetry. So I set to withymetry from

9:13 Speaker 2: our new ATL twenty four data products that Laurie talked

9:17 Speaker 2: about in the earlier session, and then we're going to

9:20 Speaker 2: integrate that ATL twenty four with inmetry with satellite drive

9:24 Speaker 2: with imetry from sentinel to imagery and generate final bathometric

9:28 Speaker 2: data layers for visual assessment. Before I jump into actually

9:34 Speaker 2: doing the dewnload, there's one step that I just want

9:36 Speaker 2: to talk about briefly so that it doesn't confuse anyone

9:40 Speaker 2: when we get to this step. So there's going to

9:42 Speaker 2: be a step where we're going to do this integration

9:45 Speaker 2: of the ATL twenty four with inmetry with Stanel two imagery,

9:50 Speaker 2: and the way we're going to do that. We're going

9:51 Speaker 2: to run a widely known algorithm known as the Stumf algorithm.

9:55 Speaker 2: This is actually going to parallel in some ways, I

9:58 Speaker 2: think a workflow that Gretchen is going to show and

10:00 Speaker 2: demonstrating the no asatpathy tool. I'll just say the version

10:03 Speaker 2: that we're going to do, it's going to be a

10:06 Speaker 2: little bit basic compared to what Gretchen is going to show,

10:09 Speaker 2: So we're not going to make use of some of

10:11 Speaker 2: the new elements that just compositing in the switching model

10:15 Speaker 2: and acolyte and other things that Gretchen might mention, But

10:19 Speaker 2: for visual analysis, this should be a pretty good workflow.

10:21 Speaker 2: So we're going to start by generating what's called pseudobythymmetry

10:25 Speaker 2: or PSDB, and that's going to be the logarithms the

10:28 Speaker 2: ratio of logarithms of two different spectral bands. We're going

10:31 Speaker 2: to use blue and green, and then in the next

10:35 Speaker 2: step we're going to do a linear regression of that

10:37 Speaker 2: pseudobythymmetry on ATL twenty four or regression of ATL twenty

10:42 Speaker 2: four reference pythymetry on this pseudobythymmetry, and that's going to

10:45 Speaker 2: give us the parameters of the linear transformation that we

10:48 Speaker 2: can use to transform that pseudopythymmetry to actual depths. So

10:55 Speaker 2: with that, let me go ahead and at this point

10:58 Speaker 2: and just start in on the demo. And the first

11:02 Speaker 2: thing we want to do is, I'm just going to

11:04 Speaker 2: bring up qgis. Again, this is free and open source

11:08 Speaker 2: GIS software. The version I've got is three point four zero,

11:11 Speaker 2: but any recent version should work reasonably well. And I'm

11:16 Speaker 2: going to go ahead and just open a new empty project.

11:20 Speaker 2: And at this point, I want to navigate to those

11:23 Speaker 2: files that I mentioned were on the training website and

11:27 Speaker 2: so hopefully you had a chance to download those. If

11:30 Speaker 2: you did. Some of the files that you grabbed, what

11:34 Speaker 2: point were the ones called hs AOI shape file. And

11:39 Speaker 2: in this case, I can just kind of drag and drop.

11:41 Speaker 2: I'm going to take the one with the extension of

11:43 Speaker 2: dot SHP and I'm just gonna drag and drop that

11:47 Speaker 2: right into qgis. There you can see it opened up.

11:52 Speaker 2: The first thing I want to do is probably just

11:54 Speaker 2: change the symbology. I'm going to double click that to

11:57 Speaker 2: bring up the symbology, and I want to change it

11:59 Speaker 2: to just a red outline something like that and say, okay,

12:05 Speaker 2: so there's my project site. The next thing I want

12:08 Speaker 2: to do is, let's see if I can move that

12:11 Speaker 2: screen move that screen sharing box out of the way. Okay.

12:15 Speaker 2: The next thing I want to do is just like

12:17 Speaker 2: usual and a gis I want to bring up some

12:20 Speaker 2: base map imagery, and so to do this, I'm actually

12:25 Speaker 2: going to bring up some Google imagery. If you're if

12:29 Speaker 2: you haven't set this up already, what you want to

12:31 Speaker 2: do is you want to go into plugins manage and

12:34 Speaker 2: install plugins, and that should pop up this list of

12:39 Speaker 2: plugins that are both installed and not installed. So in

12:43 Speaker 2: my case you can see this one called quick Map

12:46 Speaker 2: Services that's already been installed. If you don't have that yet,

12:50 Speaker 2: you want to go not installed, search for it in

12:53 Speaker 2: the search bar and install it. Because I've already got it.

12:56 Speaker 2: I'm gonna go ahead and just close this and then

12:59 Speaker 2: hopefully at this point you'll see over here on the

13:02 Speaker 2: on the right hand side of the screen, the search

13:05 Speaker 2: next gis q MS. If you don't see that, the

13:10 Speaker 2: way to get that to pop up is you want

13:13 Speaker 2: to go into UH. You should be able to go

13:16 Speaker 2: into just Web Quick Map Services and search next gis QMS.

13:23 Speaker 2: So in this box, I'm just gonna type Google Satellite

13:29 Speaker 2: and the one I want to pick is the one

13:31 Speaker 2: called Google Satellite Hybrid, and I'm gonna.

13:33 Speaker 3: Say ad.

13:36 Speaker 2: So that one ahead and opened up the imagery for

13:39 Speaker 2: my site. And at this point I can just kind

13:42 Speaker 2: of explore a little bit. I'll turn off the the

13:46 Speaker 2: project side outline just so we can see what's going on.

13:48 Speaker 2: Here are some of those engineered structures that I've talked about,

13:51 Speaker 2: the jetty, the breakwater, and then over here we've got

13:57 Speaker 2: a grind field. There are also some private revetments that

14:01 Speaker 2: have been installed for shoreline protection, the dredged channel that's

14:07 Speaker 2: used by the ferry, and so, like I mentioned, just

14:10 Speaker 2: a lot going on in a lot of reasons why

14:13 Speaker 2: the symmetry and the site might be changing and might

14:16 Speaker 2: be of interest. So in terms of looking at existing

14:19 Speaker 2: with mmetry, one thing we want want to do at

14:21 Speaker 2: this point is go ahead and bring in some Noah

14:24 Speaker 2: Medical chart data e n C data, and so to

14:28 Speaker 2: do this, I'm going to go up here and click

14:30 Speaker 2: on Layer Data Source Manager, and I'm going to scroll

14:36 Speaker 2: down here until I see w MS that stands for

14:39 Speaker 2: Webmap Service w MTS Webmap Tile Service. I'm going to

14:43 Speaker 2: do that and just click on new and I think

14:47 Speaker 2: I've already got this set up, but it shouldn't hurt.

14:49 Speaker 2: I'll go ahead and do it anyway. I'll call this.

14:52 Speaker 2: I'll call this Noah Webmap Tiles Service e n C.

14:59 Speaker 2: And then I'm going to need to put the U

15:00 Speaker 2: r L in here and to do that again. This

15:04 Speaker 2: was one of the three links that should have been

15:07 Speaker 2: available on the on the trending website, so it's actually

15:12 Speaker 2: this one here that starts with GIS chart Tools. I'm

15:14 Speaker 2: just going to grab that, that whole U r L

15:20 Speaker 2: and go ahead and put that in here and say, okay.

15:25 Speaker 2: It looks like I had already set that up, so

15:26 Speaker 2: it says do you want to overwrite it? And I'll

15:28 Speaker 2: just go ahead and say sure, overwrite okay, and then

15:32 Speaker 2: I can I can close out of this box and

15:34 Speaker 2: then back over here. Now when I scroll down in

15:38 Speaker 2: this in this browser, I scroll down to where I

15:41 Speaker 2: see w MS w MTS. Expand that pick the one

15:46 Speaker 2: that I just added. I want to expand it and

15:49 Speaker 2: expand it again until I see something called default zero

15:53 Speaker 2: two eight milimeter, and I'm going to write click that

15:56 Speaker 2: and say add layer to projects. And that worked, and

16:01 Speaker 2: went ahead and popped up the elect electronic navigational chart

16:06 Speaker 2: for this area, so we can zoom in and look

16:09 Speaker 2: at what the chart soundings look like for this area.

16:13 Speaker 2: One thing we might see. I found this kind of

16:15 Speaker 2: interesting just in exploring the e and say, if we

16:18 Speaker 2: get close to shore and for example, close to this

16:22 Speaker 2: breakwater here, you see that there are a lot of

16:24 Speaker 2: these rock awash symbols. But actually it looks like in

16:29 Speaker 2: the very shallowest areas maybe not a great density of sounding.

16:33 Speaker 2: So hopefully we'll be able to use ATL twenty four

16:35 Speaker 2: and Sentinel too withthymmetry to get a good idea of

16:38 Speaker 2: what's going on in those areas. So within mind, let's

16:43 Speaker 2: go ahead and just jump into the next part of

16:46 Speaker 2: this where I'm going to show you how to get

16:48 Speaker 2: I set to ATL twenty four with ymmetry, and to

16:52 Speaker 2: do so, I'm going to go to the next of

16:56 Speaker 2: the links that we shared and the demo and the

17:01 Speaker 2: training website, and so this time I want to go

17:04 Speaker 2: to the one that's called slide Rural Earth. That's going

17:08 Speaker 2: to take me to this website here. So this is

17:11 Speaker 2: how we're going to be able to get NASA ATL

17:14 Speaker 2: twenty four data through through slide rule and this web client.

17:19 Speaker 2: So I want to click on the link that says

17:21 Speaker 2: web client it's going to pop up. I want to

17:27 Speaker 2: make sure this is something that I sometimes forget to do. Actually,

17:30 Speaker 2: I want to make sure that I've selected the table

17:32 Speaker 2: it's called I set to coastal mathymmetry, and then we're

17:35 Speaker 2: going to be able to just zoom into our area

17:38 Speaker 2: and take a look at pathymmetry in that site. Before

17:41 Speaker 2: we do that, I just wanted to make a quick

17:43 Speaker 2: point here. So in picking out the site for this demo,

17:47 Speaker 2: I very specifically chose a site that was someone on

17:50 Speaker 2: the challenging side. So at mid latitude site, a lot

17:53 Speaker 2: of times when you see people do Sili drigathymmetry or pathymmetry,

17:58 Speaker 2: a lot of times we're focused on the very clear

18:01 Speaker 2: tropical waters where we just know things are gonna work great.

18:05 Speaker 2: So I purposefully tried to pick a little bit of

18:07 Speaker 2: a more typical project site. But just while we're here,

18:12 Speaker 2: I do kind of want to take this opportunity just

18:15 Speaker 2: to show you what it looks like if we you know,

18:18 Speaker 2: if we were to select a site that was in

18:21 Speaker 2: a uh a really you know, sort of amazing tropical area. Here,

18:28 Speaker 2: it looks like I thought I had a site selected,

18:30 Speaker 2: so I might have to do this kind of on

18:32 Speaker 2: the fly. So I'm just assuming in here on Turks

18:35 Speaker 2: and Caicos, and I'm just going to draw a project box.

18:39 Speaker 2: We'll do one maybe that goes about like maybe something

18:44 Speaker 2: like that, and I'm gonna we'll click ahead, run slide

18:50 Speaker 2: while this process is again the reason I'm doing this,

18:52 Speaker 2: This is not our site. I just want to show

18:54 Speaker 2: what it looks like when we've got a site where

18:57 Speaker 2: uh here. What's gonna happen is when it gets processing,

19:01 Speaker 2: we're going to see that it's selected all of the

19:03 Speaker 2: different track lines of I said to you, bothymetry, and

19:07 Speaker 2: for a side like this, what we should see is

19:10 Speaker 2: that throughout the entire site we've got really amazing pathymetry.

19:16 Speaker 2: I'll give this a second to just finish processing. Okay,

19:23 Speaker 2: there we go. So that worked. That came up and

19:26 Speaker 2: we can see it found all these different tracks of bythymmetry,

19:29 Speaker 2: and in any given track, I'm going to change the

19:32 Speaker 2: color scheme. I'm going to change it to this coool scheme.

19:38 Speaker 2: So any track that's within this site, I'm going to

19:41 Speaker 2: see this really just nice bathmetric profile. I'll see the

19:44 Speaker 2: depths dropping off to greater than fifteen meters depending on

19:47 Speaker 2: where where I am in the profile. I can pick

19:50 Speaker 2: another line. Basically, any line I pick, I imagine it's just

19:54 Speaker 2: going to look like pretty amazing pythymmetry. And in this case,

19:58 Speaker 2: another thing that I can put the actually do here

20:01 Speaker 2: is I can go to this three D v and win.

20:05 Speaker 2: This's loads if I can get this to work. But

20:09 Speaker 2: it should do is oh yeah, there we go. I

20:15 Speaker 2: should be able to just sort of drag this up

20:18 Speaker 2: and kind of rotate this around and look at the

20:21 Speaker 2: bathymetry for that site in three D. And in a

20:24 Speaker 2: case like this, I couldn't even made my side a

20:27 Speaker 2: bit bigger. In a case like this, you can imagine

20:29 Speaker 2: that we could generate a pretty good bathymetric DM even

20:35 Speaker 2: without incorporating any additional data. Okay, with that in mind,

20:40 Speaker 2: let's go ahead and go back to requests. And now

20:43 Speaker 2: I'm going to go up to the site that we

20:45 Speaker 2: actually are interested in for this demo. So I'm gonna

20:48 Speaker 2: I'm gonna scroll all the way back up to New

20:51 Speaker 2: England and to Cape Cod and to our Highness Project site.

21:00 Speaker 2: So here we are, here's our here's our actual side

21:02 Speaker 2: of interest. I'm gonna do the same thing. I'm gonna

21:05 Speaker 2: draw a box. I can make it pretty big. I

21:09 Speaker 2: just basically want to make sure I'm covering the whole

21:11 Speaker 2: area so I can do something. Maybe I'll do something

21:16 Speaker 2: maybe about like that, and then same process before I'm

21:19 Speaker 2: gonna run slide rule. It's going to go through and

21:26 Speaker 2: process again. It's finding all the tracks of I SAT

21:28 Speaker 2: to bathymmetry. So this one in this case, if you're

21:33 Speaker 2: to spend some time investigating these different track lines, basically

21:37 Speaker 2: what you'll see is it's not as amazing as when

21:40 Speaker 2: we were just looking at turks and caicos, but we

21:42 Speaker 2: do seem to have some pretty good bathymetry going through

21:46 Speaker 2: the going through the harbor, and so I can, you know,

21:50 Speaker 2: just try try a different couple of different track lines,

21:53 Speaker 2: and so it looks like we've got some good pometry.

21:55 Speaker 2: This is all great at this point. Let's go ahead

21:57 Speaker 2: and and and download it. And there's a few different

22:01 Speaker 2: ways I can do this. One way, if I want

22:03 Speaker 2: a little bit more control, What I could do is

22:05 Speaker 2: I could go ahead and select these individual track lines

22:08 Speaker 2: one at a time, and I could click on this

22:11 Speaker 2: thing that says three D view. Oh excuse me, not

22:14 Speaker 2: the three D V sorry, that table. I could click

22:17 Speaker 2: on table, and when this pops up. At this point,

22:20 Speaker 2: I can actually edit the commands that's shown here, and

22:23 Speaker 2: I can do things like I can pick exactly what

22:25 Speaker 2: parameters I want to export. I can I can pick

22:28 Speaker 2: different you know, filters, or do things in different ways. However,

22:34 Speaker 2: for the sake of this demo, I'm going to do

22:35 Speaker 2: things in an easier way. I'm just going to go

22:37 Speaker 2: kind of right in the middle of my screen. If

22:40 Speaker 2: you see this button that says export, I'm just going

22:42 Speaker 2: to go ahead and click that, and I'm gonna choose

22:45 Speaker 2: an export format, and I want to make that CSV

22:48 Speaker 2: comma separated values and say export. And I'm gonna go

22:52 Speaker 2: ahead and put this into the folder that I created

22:54 Speaker 2: for this training. So I created this g Hyannas Harbor

22:58 Speaker 2: Aziel twenty four demo. I'll make a new subfolder that

23:01 Speaker 2: I will call ATL twenty four, and I will put

23:07 Speaker 2: this CSP file right in that folder. Something I should

23:12 Speaker 2: point out at this point because I was kind of

23:15 Speaker 2: I wasn't very discriminate in picking out exactly what tracks

23:18 Speaker 2: I want. I just grabbed them all. That was the

23:22 Speaker 2: easiest way to go and grab the data. I am

23:24 Speaker 2: going to pay a slight price for that later, and

23:26 Speaker 2: I'll talk about that when I get there. For now,

23:29 Speaker 2: we're actually done in slab role, though, so I can

23:31 Speaker 2: go ahead and close the web browser and get back

23:34 Speaker 2: over into QGIS and back to our project site. And

23:40 Speaker 2: at this point, since I just grabs that ATL twenty

23:44 Speaker 2: four data layer, why don't I just go ahead and

23:46 Speaker 2: bring that into QGIS. So the way I'm going to

23:51 Speaker 2: do that is, I'm going to click on layer and

23:54 Speaker 2: under layer, I'm going to do AD layer, and there

23:57 Speaker 2: should be one that's called ad to limit text layer,

24:01 Speaker 2: so I'll click that. Now I want to just browse

24:04 Speaker 2: to find that file, so it's the file that I

24:08 Speaker 2: just created. There, I'm going to go to AHL twenty

24:10 Speaker 2: four folder, select that CSP file and say, okay, if

24:15 Speaker 2: I want, I can scroll along down here in the

24:18 Speaker 2: bottom and see the individual fields that came in in

24:21 Speaker 2: that file, so all the different information refraction corrected orthometric height,

24:27 Speaker 2: so all of the pothometry has been refraction corrected. It's

24:30 Speaker 2: got a confidence interval. I think that's something that Lori

24:34 Speaker 2: covered in her talk. But basically the confidence and the

24:36 Speaker 2: and the bathymetry classifications, spot granule, et cetera. Everything basically

24:43 Speaker 2: auto populated. It said that my ex field is longitude

24:47 Speaker 2: phyfield is latitude pH. I can go ahead and set

24:50 Speaker 2: my Z field to ortho underscore H that's my Z

24:56 Speaker 2: and I can go ahead and just say AAD and

24:59 Speaker 2: at that point I you'd be able to close this window.

25:01 Speaker 2: And if you can see all these orangeish dots that

25:05 Speaker 2: just got added, that's the ATL twenty four data. If

25:09 Speaker 2: I want to can change it symbol. I kind of

25:11 Speaker 2: like doing this as a blue dot. I think that

25:13 Speaker 2: looks pretty good. There's my ATL twenty four data. So

25:18 Speaker 2: at this point we're really doing well. I would make

25:21 Speaker 2: an observation here, though, and that is that even though

25:24 Speaker 2: this brought in a lot of ATL twenty four data,

25:29 Speaker 2: it still is kind of sparse in terms of you know,

25:32 Speaker 2: there's areas of the harbor where if we're interested in

25:34 Speaker 2: looking at nearshore withythymmetry, all we've got is these individual

25:38 Speaker 2: discrete tracks. So this is the point where we want

25:41 Speaker 2: to see how we can combine ATL twenty four pythymmetry

25:46 Speaker 2: with specially drive pathymetry from SEVENEL two. So to do this,

25:51 Speaker 2: I'm going to go to the third of the weblinks

25:54 Speaker 2: that we're shared in the trending website, and the third

25:57 Speaker 2: link it's the one that's called Pernicus Browser. So I'm

26:02 Speaker 2: gonna open this up. You'll need to create an account

26:05 Speaker 2: for this. It's free, uh so free to create an account,

26:08 Speaker 2: and then you'll have to log in. I happen to

26:10 Speaker 2: already be logged in, and you can tell that because

26:12 Speaker 2: it's got my name listed up here. But go ahead

26:16 Speaker 2: and sign in if you need to. Okay, So the

26:20 Speaker 2: next thing I want to do is just go over

26:23 Speaker 2: to our project site. So I'm gonna scroll over to

26:28 Speaker 2: Cape cod and into our Hynas Harbor project site. At

26:36 Speaker 2: this point, I want to draw an AOI and so

26:40 Speaker 2: I'm going to use this little polygon tool, and then

26:42 Speaker 2: once I select that, I want to grab the tool

26:44 Speaker 2: next to it. It looks like the pencil icon. I'll

26:47 Speaker 2: select that and I might scroll it a little bit,

26:51 Speaker 2: and I'm just gonna I'm just gonna go ahead and

26:53 Speaker 2: and digitize something it doesn't. It doesn't have to be perfect,

26:56 Speaker 2: and it's okay if I make this a little bit

26:59 Speaker 2: big than what it actually needs to be, so I'll

27:03 Speaker 2: maybe do something and just finish it up off like that. Okay,

27:08 Speaker 2: Now we want to put in a date, and this

27:11 Speaker 2: is the point where you can actually spend a lot

27:14 Speaker 2: of time looking through the different dates trying to find

27:16 Speaker 2: imagery that looks good. Cloud free, water looks reasonably clear.

27:21 Speaker 2: I'm going to do a little bit of a shortcut.

27:23 Speaker 2: I've already done some analysis and i found a data

27:26 Speaker 2: that looks good, so I'm just going to go ahead

27:27 Speaker 2: and put that in. The day that I found is

27:29 Speaker 2: twenty twenty three, twenty twenty three oh three, twenty twenty

27:38 Speaker 2: three oh three, twenty all right there, it is right there,

27:45 Speaker 2: so it brought up that date. I want to make

27:48 Speaker 2: sure I'm in the visualize tab here, and at this

27:52 Speaker 2: point over on the right hand side of the screen,

27:55 Speaker 2: I've got this download icon, so I'll click that download icon,

28:01 Speaker 2: and then here I want to click on the tab

28:02 Speaker 2: that says analytical, and I want to be careful with

28:06 Speaker 2: the settings here. I want my image format to be

28:09 Speaker 2: tiff thirty two bit float. I want to change my

28:13 Speaker 2: resolution to high. I'm gonna go ahead and keep the

28:18 Speaker 2: court the default coordinate system. And then I'm going to

28:21 Speaker 2: select three different images to download. I'm going to do

28:24 Speaker 2: true color, so that's going to be the RGB, and

28:27 Speaker 2: then I'm gonna pick just two of the bands. I'm

28:29 Speaker 2: going to do BAN two, which is blue, and BAN three,

28:33 Speaker 2: which is green, and I can just click the click

28:37 Speaker 2: the download icon, so that should be saving them to

28:43 Speaker 2: my downloads folder. I can go ahead and open that

28:46 Speaker 2: up there it is, it just downloaded. I'm going to say,

28:53 Speaker 2: we'll just do cut and I'm going to paste this

28:56 Speaker 2: into the to the folder that I set up specifically

28:58 Speaker 2: for this demo. We'll do a subholder. My new subholder

29:02 Speaker 2: will just be called Sentinel two.

29:06 Speaker 4: No.

29:07 Speaker 2: Two and copy in that ZIP file. I'll go ahead

29:12 Speaker 2: and extract all of the different images that are in

29:17 Speaker 2: that in that. Okay, So that worked, it extracted all

29:26 Speaker 2: the images. I'm going to do one more thing, just

29:29 Speaker 2: because these are fairly long and cumbersome file names. I'm

29:33 Speaker 2: going to rename these the thing that says banned too.

29:36 Speaker 2: I'm just going to rename that to S two for

29:39 Speaker 2: Sentinel to blue. The one that says banned three, I'm

29:45 Speaker 2: going to rename that to be S two winterscore green.

29:51 Speaker 2: And then the final one that's the true color RGB,

29:55 Speaker 2: I'm going to rename that to be S two underscore RGB.

30:02 Speaker 2: So I've got those three three different images, all redn ends.

30:07 Speaker 2: At this point, I can get out of that Copernicus

30:11 Speaker 2: browser and I can actually just, uh, this is another

30:15 Speaker 2: case where I can just drag and drop this, So

30:16 Speaker 2: I'll just I'll just drag and drop those right in. Whoops,

30:21 Speaker 2: So it looks pretty good here. I am back in

30:23 Speaker 2: in QG. I S I've got my uh my RGB

30:28 Speaker 2: image up at the top. If I want, I can

30:30 Speaker 2: go ahead and take my ATL twenty four both imagery

30:33 Speaker 2: and drag that up to the up to the very

30:37 Speaker 2: top so that we can see where the where the

30:40 Speaker 2: I saw two pythymmetry is on top of the Sentinel

30:42 Speaker 2: two imagery. Okay, at this point, we're going to move

30:49 Speaker 2: on to the next step, which is, uh that some

30:52 Speaker 2: prestio that I showed you earlier on the slides, and

30:55 Speaker 2: we're going to create that pseudo bythymmetry layer. So this

30:58 Speaker 2: is going to be the ratio of the logarithms of

31:00 Speaker 2: the two different spectral bands, the blue band and the

31:02 Speaker 2: green band. So the way I'm going to do this

31:06 Speaker 2: is I'm going to click on raster and then raster calculator,

31:11 Speaker 2: so that brings me up here to raster a calculator.

31:14 Speaker 2: I'm going to create an output layer, and I want

31:16 Speaker 2: to give this a name, so I'm going to put

31:18 Speaker 2: it right back in my in my demo folder, I'll

31:21 Speaker 2: do a new subfolder called PSDB. Again, that stands for pseudopathymmetry,

31:27 Speaker 2: and that denotes the fact that we're not going to

31:29 Speaker 2: have actual depths. It'll just be sort of pseudodepths or pseudobythymmetry.

31:34 Speaker 2: So I'm going to go in there and I'm going

31:36 Speaker 2: to make my file name. I'm going to do stb

31:39 Speaker 2: dot tef save, and then the next thing I want

31:45 Speaker 2: to do is I just want to go ahead and

31:48 Speaker 2: create that stump pray show. So this is going to

31:52 Speaker 2: be natural log of one thousand times blue band. So

32:01 Speaker 2: S two blow in parentheses divided by natural log of

32:10 Speaker 2: one thousand. I'ms the green band. This might look a

32:19 Speaker 2: little funny, what what is this again? This is just

32:22 Speaker 2: this is just the stump rash show that that we

32:24 Speaker 2: saw earlier back in the in the slide. So we've

32:27 Speaker 2: done log of blow over log of green. That factor

32:30 Speaker 2: of a thousand, that's just a constant that is designed

32:32 Speaker 2: to keep the keep the logarithms positive and linear linearly

32:37 Speaker 2: related to depth. So I'm just gonna go ahead and

32:40 Speaker 2: run this. I'm gonna say, okay, and it worked. It

32:43 Speaker 2: already generated my pseudopythymmetry later. All right. The next step

32:51 Speaker 2: I want to do at this point is, like I

32:53 Speaker 2: mentioned earlier, our ultimate goal here is we want to

32:56 Speaker 2: use the ISAT tubeothymmetry from ATL twenty four to calibrate

33:00 Speaker 2: our pseudopythymmetry. And so to do this, we need to

33:04 Speaker 2: create a layer that has both values. We want to

33:09 Speaker 2: We want a layer that has the at L twenty

33:11 Speaker 2: four heights. I should have mentioned, by the way, uh

33:16 Speaker 2: in the at L twenty four the heights or orthometric

33:19 Speaker 2: heights best on the e g M O eight GID model.

33:22 Speaker 2: So those are you know, actual datam best heights, and

33:26 Speaker 2: then the pseudopythymmetry that's not those are just sort of

33:28 Speaker 2: relative values. But what I want to do is create

33:31 Speaker 2: a layer that has both of those together and so two. Uh,

33:36 Speaker 2: to do this, I'm going to go up to the

33:39 Speaker 2: processing toolbox. I'm gonna to go processing toolbox. I've already

33:43 Speaker 2: got this up. So in this search window for the

33:45 Speaker 2: processing toolbox, I'm going to search for clip raster by

33:50 Speaker 2: mass layer, clip raster by mass layer. And in this case, uh,

33:57 Speaker 2: my input layer is going to be let's see, I

34:04 Speaker 2: want to do uh, I want to do It looks

34:11 Speaker 2: like I made a mistake. I called the thing that

34:13 Speaker 2: I wanted to be set up with the matory actually

34:15 Speaker 2: called SDB. That should say PSDB. But that's okay. My

34:21 Speaker 2: input layer is going to be just my SDB layer,

34:27 Speaker 2: and my mask layer is going to be Wait, I'm sorry,

34:34 Speaker 2: I say what I've done. I've done the I've done

34:36 Speaker 2: the wrong tool here, the tool that I actually wanted. Sorry,

34:40 Speaker 2: back in this search box. The tool that I want

34:42 Speaker 2: here is the sample raster values. That's the one that

34:46 Speaker 2: I meant to bring up sample raster values. So my

34:49 Speaker 2: input layer is going to be my Hotel twenty four data.

34:53 Speaker 2: My raster letter layer is going to be SDB, and

34:55 Speaker 2: again I should have called that PSDB. I'm going to

34:59 Speaker 2: create an output column. I'm going to call that sample

35:04 Speaker 2: sample underscore PSDB and just go ahead and run that

35:10 Speaker 2: and I can close and that ran, and that created

35:15 Speaker 2: this layer called sampled, which now that has all the

35:19 Speaker 2: ATL twenty four heights and it also has the pseudopythymmetry.

35:24 Speaker 2: So at this point we're actually really close to being

35:27 Speaker 2: able to generate our final SDB layer. We just need

35:30 Speaker 2: to do a linear regression in order to calculate linear

35:34 Speaker 2: transformation coefficients that are going to allow us to transform

35:37 Speaker 2: or pseudo bythymmetry to actual withymetry. So to do this,

35:42 Speaker 2: there's a Python script. I'm going to go plug ins

35:46 Speaker 2: Python console and then open up the Python editor and

35:52 Speaker 2: there's the scripts that's for me already open. But if

35:55 Speaker 2: you don't have this open, you just want to click

35:57 Speaker 2: on the open script and want to browse and finds

36:01 Speaker 2: the scripts that was made available on the Trening website.

36:04 Speaker 2: It's the PSDB linear regression dot pile, and go ahead

36:07 Speaker 2: and open that up. That will open up the scripts.

36:10 Speaker 2: The only thing that you should actually have to change

36:13 Speaker 2: here is if you go down to this it's line seven,

36:16 Speaker 2: and you'll see this path to where the where that

36:19 Speaker 2: CSV file lives that you created. Again, that's the CSV

36:23 Speaker 2: file that we need, the one that has the ATL

36:25 Speaker 2: twenty four data in the pseudo bythymetry. You want to

36:28 Speaker 2: change that to your path and make sure you've got

36:32 Speaker 2: your slashes going in the right direction. And at that

36:35 Speaker 2: point you should just be able to hit this s

36:36 Speaker 2: green run script icon and hopefully this will run and okay.

36:49 Speaker 2: So that went ahead, and that finish running, and it

36:52 Speaker 2: gave me the scatter plot. So what I see here

36:55 Speaker 2: is on the x axis, those are my pseudo pythymmetry values.

36:59 Speaker 2: On the why I that's my I saw to bithymetry

37:02 Speaker 2: from ATL twenty four and then it fit a linear

37:05 Speaker 2: aggression line. The first thing I want to note in

37:08 Speaker 2: this case is that our our square value is actually

37:10 Speaker 2: fairly lost, so we've got an R squared of point

37:13 Speaker 2: five to four. Usually in my research group, we wouldn't

37:17 Speaker 2: move on with using this for bathymetry unless we had

37:21 Speaker 2: an R squared of about point seven or higher. The

37:25 Speaker 2: reason that I happen to know that this is okay

37:27 Speaker 2: in this case is if you'll remember that there was

37:29 Speaker 2: that step when we were in the slab real Web

37:32 Speaker 2: client and I said that we could have been a

37:34 Speaker 2: little bit more selective and just picking out exactly which

37:37 Speaker 2: tracks we wanted to download, but we didn't do that.

37:39 Speaker 2: We went ahead and just sort of grabbed everything. So

37:43 Speaker 2: I know, because I actually tried this, that if I'd

37:46 Speaker 2: spent a little more time in that step of grabbing

37:48 Speaker 2: just the tracks that I knew had the best bathymetry,

37:51 Speaker 2: and or if I'd done a little bit of cleanup afterwards,

37:54 Speaker 2: I could get a much better our square value here.

37:57 Speaker 2: But actually it doesn't really change the presameters of this

38:00 Speaker 2: linear transformation the sloping and intercept constant. So that's what

38:07 Speaker 2: we're really interested in here, So we're not going to

38:09 Speaker 2: worry too much about the kind of lowish R square.

38:12 Speaker 2: We just need these the values the sloping intercept constants.

38:20 Speaker 2: So at this point we are actually very close. Uh,

38:24 Speaker 2: this is a point where we're going to go back

38:26 Speaker 2: into raster calculator. Let me see if I can do

38:28 Speaker 2: it this way. I need to keep this scatter plot

38:30 Speaker 2: up because I need to be able to see these numbers.

38:34 Speaker 2: So I'm going to go back to raster calculator and

38:38 Speaker 2: what I'm basically going to do is I'm just going

38:40 Speaker 2: to apply a linear transformation that is going to transform

38:45 Speaker 2: my pseudo bythymmetry to actual byth inmetry. So I'm going

38:49 Speaker 2: to do an output layer. I will call this. I'll

38:53 Speaker 2: create a new folder here that's just going to be

38:55 Speaker 2: called sdb SDB, and I'll give this file a file name.

39:01 Speaker 2: And because I accidentally called the other one SDB, I'm

39:04 Speaker 2: going to call this one. I'll just do something like

39:07 Speaker 2: SDB final, SDB final dot tif and I'll say save.

39:16 Speaker 2: And now I'm going to create an equation to compute this,

39:20 Speaker 2: and it's going to be I'm just going to look

39:24 Speaker 2: over here. This is why I keet the scatter plot up.

39:26 Speaker 2: I want to look at uh at the two constants.

39:29 Speaker 2: So I'm going to do negative twenty five point eight

39:32 Speaker 2: five zero times SDB, which really again that should have

39:39 Speaker 2: been our pseudo athymetry minus oh, I'm sorry, not minus

39:43 Speaker 2: plus plus twenty one point five to three zero, where

39:48 Speaker 2: those two constants again are just from overhearing our scatter plot,

39:52 Speaker 2: I can say you okay, and it looked like that

39:57 Speaker 2: that one ahead ran. So I've got a layer that's

39:59 Speaker 2: now called sd be final. I do probably want to

40:03 Speaker 2: clip that to just our project site so we can

40:06 Speaker 2: and I can I can get rid of the Python

40:08 Speaker 2: script at this point, this is where I want to

40:10 Speaker 2: clip it. So this is the point where I actually

40:13 Speaker 2: want to do the tool that I started to show

40:15 Speaker 2: last time. I'm gonna do this clip raster by mass

40:18 Speaker 2: player and I'll do I'm going to clip SDB final

40:21 Speaker 2: and the mass Player is going to be the hynas

40:24 Speaker 2: AOI and just let that run. Oh oh, I guess

40:33 Speaker 2: it didn't select it. Try one more time. Run. Okay,

40:36 Speaker 2: that ran. Now I can get rid of this one

40:40 Speaker 2: and I can turn off that and that, and I've

40:43 Speaker 2: just got my final SDB layer and if I want

40:47 Speaker 2: to change the symbology, I can double click it. I

40:50 Speaker 2: can change from single band great to single band pseudo

40:54 Speaker 2: color and just pick out a color ramp. I kind

40:56 Speaker 2: of like this one. This one looks kind of like pytymetry.

41:00 Speaker 2: Go ahead and apply and close. And so now I've

41:03 Speaker 2: got a bathymetric DM for this project site. And that

41:06 Speaker 2: was created using the combination of I set two bathymetry

41:09 Speaker 2: from ATL twenty four and centaal two values. It looks

41:13 Speaker 2: pretty interesting. I can clearly see the dredge channel through there.

41:16 Speaker 2: I can see some of the deeper areas like in

41:18 Speaker 2: the anchor in the in the anchorage behind the breakwater.

41:21 Speaker 2: Over here, I can see some of the really shallow

41:24 Speaker 2: areas that are closer to shore. So everything looks pretty good.

41:28 Speaker 2: Uh So at this point, if I wanted to, if

41:31 Speaker 2: I were interested in in bathymetric change and kind of

41:33 Speaker 2: seeing how the pathymmetry in this site changes over time,

41:38 Speaker 2: I could just repeat that process using multiple dates, multiple

41:41 Speaker 2: epics of Sentinel two imagery and then I could do

41:45 Speaker 2: differences of the d ms or just sort of look

41:47 Speaker 2: visually at how pathymmetry in this area is changing over time.

41:53 Speaker 4: So that's it.

41:54 Speaker 2: That is the uh that's the end of the demo.

41:56 Speaker 2: But I just wanted to say thanks and if you

41:59 Speaker 2: want to reach me, my contact information is here, Christopher

42:02 Speaker 2: dot Parrish at Oregon State dot ADU.

42:05 Speaker 1: Thanks Chris, thank you for the excellent demonstration of ATL

42:09 Speaker 1: twenty four integration with Sentinel two and generation of mathymetric

42:14 Speaker 1: data layers. We will now transition to JP Swinsky, who

42:18 Speaker 1: will provide a demonstration of slide roll Python client. JP

42:23 Speaker 1: over to you.

42:27 Speaker 5: Great well, thank you.

42:28 Speaker 6: I'm going to be talking about slide rule worth again

42:31 Speaker 6: and just as a refresher from what we talked about

42:33 Speaker 6: a couple of days ago. Slide Rule is a public

42:36 Speaker 6: web service with rest like APIs for processing and accessing

42:40 Speaker 6: science data. And specifically what we're talking about here is

42:45 Speaker 6: accessing the ATL twenty four with inmetry data. Last time

42:50 Speaker 6: we talked about the web browser and I gave a

42:53 Speaker 6: demo of that web client. Today we're going to be

42:57 Speaker 6: talking about the Python client that typically runs inside a

43:02 Speaker 6: jubiterin notebook, and I'm going to be given an example

43:05 Speaker 6: of how to access to ATAIL twenty four data from

43:08 Speaker 6: a jubiteren notebook using the side rule Python client. Now

43:12 Speaker 6: before I do that, though, I want to just show

43:18 Speaker 6: our landing page of our website. So the Python client

43:23 Speaker 6: can be a little intimidating. It's not as easy to

43:26 Speaker 6: use as the web client, and so we provide a

43:29 Speaker 6: lot of documentation to go along with it to help

43:33 Speaker 6: to help you out, to help you get started. So

43:35 Speaker 6: if you navigate to side rule worth do io in

43:39 Speaker 6: your browser, find you'll come here. And the first thing

43:43 Speaker 6: I want to point out is this contact us page,

43:45 Speaker 6: so you should always feel free to reach out to us.

43:49 Speaker 6: Here's this email here goes directly to me. I'm always

43:53 Speaker 6: excited when users email us and we do our best

43:57 Speaker 6: to get back to you as soon as possible. You're

44:00 Speaker 6: also free for people a little bit more on the

44:04 Speaker 6: developer side, if you want to go to GitHub and

44:07 Speaker 6: open up an issue if you find anything wrong, or

44:09 Speaker 6: want to feature requests. And certainly we're an open source

44:12 Speaker 6: project and we invite collaborations, so you're also free to

44:16 Speaker 6: open pool requests and start discussions on features that you'd

44:19 Speaker 6: like to see. But now about the documentation. So clicking

44:26 Speaker 6: on the documentation link takes you to our documentation site

44:29 Speaker 6: and this is it here on the left. You can

44:33 Speaker 6: kind of walk through it in the order of which

44:36 Speaker 6: you would typically want to approach it if you're a

44:39 Speaker 6: new user. So getting started, the first thing you're going

44:42 Speaker 6: to want to do is know how to install the

44:46 Speaker 6: Python client, and this is instructions on how to install it.

44:49 Speaker 6: We support Conda and pie Pie. You can also directly

44:55 Speaker 6: install it from GitHub. Then we have a really short,

44:59 Speaker 6: the simple Hello World type Python program that can just

45:04 Speaker 6: get you started and make sure everything is installed correctly

45:06 Speaker 6: and you're running okay, and then we move into examples.

45:10 Speaker 5: I don't know about you, but I learned best from example.

45:12 Speaker 6: So here's some examples that we've put together, and the

45:15 Speaker 6: one we're actually gonna be going through today is this

45:17 Speaker 6: ATL twenty four example. But then as you progress you

45:23 Speaker 6: can go through the user's guide. You're gonna find out

45:27 Speaker 6: all the different parameters and APIs that are available. There's

45:32 Speaker 6: some more information for specific to ice at two right here,

45:39 Speaker 6: and then going down even further, we have our developer's guide.

45:45 Speaker 6: So if you're really interested in how this all works,

45:48 Speaker 6: we have a map of our project why we develop

45:51 Speaker 6: side rule. Here's an under the hood. So here's an

45:55 Speaker 6: eye chart on all the different components and an explanation

45:58 Speaker 6: of those. But just wanted to let you know this

46:02 Speaker 6: is where to go when you have questions about using

46:05 Speaker 6: the Python client. So now I'm going to switch over

46:08 Speaker 6: and actually look at our ATL twenty four example.

46:14 Speaker 5: So this is just check. I've checked out the repository.

46:19 Speaker 6: And I've opened it up in vs code, and so

46:21 Speaker 6: under the slide rule repo from GitHub, I go to

46:24 Speaker 6: clients Python examples, and then under the examples you'll find

46:30 Speaker 6: this notebook that I'm going to walk through right here.

46:33 Speaker 6: So I've set this all up, so I'm gonna just

46:35 Speaker 6: run through this out. So this is I'm importing the

46:40 Speaker 6: slide rule client into my kind of environment. Well, I

46:42 Speaker 6: should point out this file right here is all of

46:48 Speaker 6: the packages you need to run all of our examples.

46:52 Speaker 6: You don't have to use it, but if you're wondering, man,

46:55 Speaker 6: I wish I had a quick way to get all

46:57 Speaker 6: the dependencies.

46:59 Speaker 5: This is what you do.

47:01 Speaker 6: Okay, So now I'm going to initialize the client. You

47:04 Speaker 6: will find one difference here. I'm running against a private

47:07 Speaker 6: developers cluster. Just so I don't pound the public cluster.

47:12 Speaker 6: This organizational setting is absent in the in the version

47:18 Speaker 6: that is on our repository, and it would go right

47:21 Speaker 6: to the public cluster. Here are some helper functions. I'm

47:26 Speaker 6: not going to talk too much about those, but you'll

47:27 Speaker 6: see them being used. It's really just to plot the

47:30 Speaker 6: data that we get back from the server. So I

47:34 Speaker 6: have functions to plot ATAIL twenty four, ATAIL six, and

47:38 Speaker 6: atil oh three. And now I'm also going to define

47:42 Speaker 6: an area of interest. It's just on the north shore

47:44 Speaker 6: of the Dominican Republic. And this is There are many

47:48 Speaker 6: ways that you can find an area of interest using

47:50 Speaker 6: our Python client.

47:51 Speaker 5: This is the simplest.

47:52 Speaker 6: It's just creating a list of lat longs. You can

47:56 Speaker 6: also use Goojson and shape files. But I've done this

48:01 Speaker 6: this way here for clarity. So now if you just

48:04 Speaker 6: want I just you just want all the data ATAIL

48:07 Speaker 6: twenty four data in this area of interest. This is

48:10 Speaker 6: one line here you run it it's going to it

48:14 Speaker 6: says that there's one hundred and ten ATAIL twenty four

48:17 Speaker 6: granules that are intersect this area of interest.

48:21 Speaker 5: Ran in eight seconds.

48:22 Speaker 6: It processed all one hundred and ten of those granules

48:25 Speaker 6: in parallel on the server. So so just to talk

48:28 Speaker 6: through what's happening here, the Python client built a request

48:33 Speaker 6: when you hit this run, and it's sent that request

48:36 Speaker 6: to the server. The server is sending back status messages

48:40 Speaker 6: which we see displayed here, and then it sends back

48:43 Speaker 6: a data frame encoded as a geopark file that the

48:49 Speaker 6: Python client then opens up. And so that's all what

48:53 Speaker 6: happened right there. And now I'm just going to look

48:55 Speaker 6: at that data frame and has forty four thousand rows.

48:58 Speaker 6: That means it's got forty four thousandthymmetry photons. And these

49:02 Speaker 6: are the columns that come back by default when you

49:04 Speaker 6: just do a simple access of the ATL twenty four

49:07 Speaker 6: data so you'll see here that each photon has the

49:11 Speaker 6: same classification. It's all withymetry. This is an encoding for

49:14 Speaker 6: which groundtrack it is. This is the height of the surface,

49:18 Speaker 6: the rithmetric height of the surface. This is the reference groundtrack.

49:21 Speaker 6: If you know anything about ice A two, all of

49:24 Speaker 6: the tracks are labeled and so it's reference groundtracks seven

49:29 Speaker 6: to fifty eight. Here's the rithmetric height of each of

49:31 Speaker 6: these photons. So this has a depth of about twenty

49:33 Speaker 6: meters here, and here's the confidence level that the ATAIL

49:38 Speaker 6: twenty four algorithm assigned to it. This is which spot

49:42 Speaker 6: it was on the six spots of the ice AT

49:45 Speaker 6: two ATLAS instrument. Here's the along track coordinate, a cross

49:50 Speaker 6: track coordinate. And then this source ide is allows you

49:56 Speaker 6: to look up which granule the data came from. So

50:00 Speaker 6: embedded in the metadata of the geodata frame is a

50:03 Speaker 6: dictionary that links source IDs to granule names. And then lastly,

50:08 Speaker 6: this is which cycle the data was collected at. Oh,

50:12 Speaker 6: and we have the time and the lat long and

50:15 Speaker 6: if you're familiar with geodata frames, that this lat long.

50:18 Speaker 5: Isn't coorted as a point geometry.

50:21 Speaker 6: So let's plot this and see what it looks like. Okay,

50:28 Speaker 6: this is just a quick view of what it looks like.

50:31 Speaker 6: And so I can see here underneath where it's white,

50:33 Speaker 6: this is probably land. And then this is where we

50:36 Speaker 6: get a lot of good pathymetry near to the shore.

50:39 Speaker 6: And then the algorithm is doing its best to find

50:41 Speaker 6: bathymetry as it goes farther offshore.

50:44 Speaker 5: So this is a lot of data. Now let's say, you.

50:47 Speaker 6: Know what, there's actually a particular track that I'm interested in,

50:51 Speaker 6: and so I've preconstructed this. So now instead of pulling

50:55 Speaker 6: all of the data, I just want the ground track

50:59 Speaker 6: three the right beam. I want the reference groundtrack to

51:02 Speaker 6: be two two, and I want it to be from

51:05 Speaker 6: cycle twelve of the ICE two collection. So now let's

51:10 Speaker 6: run that. It's gonna the client's going to construct a request,

51:12 Speaker 6: send it out to the server. The server is going

51:14 Speaker 6: to process that request and send the data back in.

51:16 Speaker 6: It happened now in three seconds. So I'm gonna look

51:21 Speaker 6: at this data frame. And so now I've only got

51:22 Speaker 6: eight hundred and sixty one but THEMTRY photons. So this

51:26 Speaker 6: set of photons is a subset of this one up here,

51:30 Speaker 6: and the subset is defined by these parameters here the beam,

51:35 Speaker 6: the reference groundtrack, and the cycle.

51:37 Speaker 5: You can use others.

51:37 Speaker 6: You can include just one or include others that are

51:41 Speaker 6: in the data set, all of which are going to

51:44 Speaker 6: be explained in our documentation.

51:47 Speaker 5: Okay, so now let's plot this. Ah.

51:50 Speaker 6: Okay, so this is now a side view of the photons.

51:57 Speaker 5: Let's go back up here again.

51:59 Speaker 6: This is a top down view looking plotting them all

52:02 Speaker 6: of the bathymetry based on its lat long coordinate and

52:07 Speaker 6: kind of color coding it to show some depth. But

52:10 Speaker 6: now we're turning this on its side and looking at

52:14 Speaker 6: just one of the tracks, and we defined which one

52:16 Speaker 6: of those tracks are and we get this nice plot here.

52:18 Speaker 6: So this is a profile of the seafloor. Okay, So

52:25 Speaker 6: now I'm like, oh, I really like this. I want

52:28 Speaker 6: to I want to dig in and get more information

52:30 Speaker 6: about it. So now I'm going to do a detailed

52:32 Speaker 6: access of this single track, similar to what we talked

52:35 Speaker 6: about a couple of days ago with the web client.

52:38 Speaker 5: I'm turning the compact flag off so I get all

52:41 Speaker 5: of the data.

52:42 Speaker 6: I'm going to now say I don't want you to

52:45 Speaker 6: I don't want the server to threshold on confidence at all.

52:48 Speaker 6: Give me all of the bathymetry, even if it's very

52:51 Speaker 6: low confidence, and then give me all the different classifications.

52:56 Speaker 5: So let's run this.

52:57 Speaker 6: It's going to construct this request, it's going to send

52:59 Speaker 6: it to the server. Serve is going to respond with

53:02 Speaker 6: the results. And it just didn't. And now we have

53:06 Speaker 6: thirty five thousand photons. Again, the reason we have thirty

53:10 Speaker 6: five thousand photons. So now we have thirty five thousand photons.

53:17 Speaker 6: The reason we have thirty five thousand photons instead of

53:21 Speaker 6: the eight hundred some up here is because we're asking

53:26 Speaker 6: for all the photons, not just the bithymmetry. So we're

53:29 Speaker 6: getting a lot of sea surface and unclassified photons.

53:32 Speaker 5: So let's plot that.

53:35 Speaker 6: And here you see it, so you can see the

53:37 Speaker 6: gray is the unclassified. This looks like land, and then

53:40 Speaker 6: we come down there's a little drop to the ocean,

53:44 Speaker 6: and then we've got the sea surface in blue, and

53:47 Speaker 6: we've got that same track now in red in the

53:50 Speaker 6: context of the other photons.

53:53 Speaker 5: Let's look at this based on confidence.

53:56 Speaker 6: And you can see here we have a bunch of

54:00 Speaker 6: different confidence values spread throughout the bothymmetry being identified there. Okay,

54:08 Speaker 6: so again something that we talked about a couple of

54:11 Speaker 6: days ago was using the web client. We were able to,

54:15 Speaker 6: instead of going directly to a TAIL twenty four, go

54:18 Speaker 6: to ATL three and use ATL twenty four to classify it.

54:23 Speaker 6: So this is a more sophisticated request. The ATAIL twenty

54:27 Speaker 6: fours is run against the release six ATLO three, so

54:32 Speaker 6: I have to manually specify that in the request parameters.

54:36 Speaker 6: I'm going to ask for everything because this is an

54:39 Speaker 6: ATL three request. I'm gonna turn off any filtering based

54:44 Speaker 6: on its own confidence values, and I'm going to ask

54:47 Speaker 6: for the same beam reference, groundtrack, and cycle.

54:50 Speaker 5: So this is going to go out and grab that.

54:55 Speaker 5: And we got that there.

54:56 Speaker 6: And so now this is all of the ATLO three

55:00 Speaker 6: data classified with the.

55:06 Speaker 5: ATL twenty four.

55:07 Speaker 6: You notice it's a bigger area here, and there's more

55:11 Speaker 6: data data that wasn't used by the classifier because it

55:15 Speaker 6: was outside of a band considered outside of the band

55:21 Speaker 6: in which there'd be valid bathymetry. So that's why there's

55:24 Speaker 6: even more data. Now I'm ATLO three, But I also

55:27 Speaker 6: want to point out back up here, we're going to

55:30 Speaker 6: get different fields. So now we get what the spacecraft

55:33 Speaker 6: velocity was at each of these photons. We get what

55:35 Speaker 6: the solar elevation is. See, we get the background rate

55:40 Speaker 6: and we get the quality pH flags. These are all

55:42 Speaker 6: fields that come in the ATLO three product but do

55:46 Speaker 6: not come in the ATL twenty four product. So therefore,

55:49 Speaker 6: if you want those fields, this is an easy way

55:52 Speaker 6: to get that data and overlay.

55:55 Speaker 5: The ATL twenty four classifications. Okay, now.

56:02 Speaker 6: Let's combine some features of ATLO three and again combine

56:08 Speaker 6: it with ATL twenty four to do some filtering at

56:10 Speaker 6: the ATLO three level. So here I'm going to do

56:14 Speaker 6: the same request as before, except I'm going to do

56:16 Speaker 6: two new filters. I'm gonna say, only give me ATLO

56:21 Speaker 6: three that has a high confidence.

56:22 Speaker 5: According to the ATLO three algorithm.

56:25 Speaker 6: And the ATLO three has this classifier called yapsy, yet

56:30 Speaker 6: another photon classifier which gives us score based on the

56:33 Speaker 6: local density of the photon how many photons are near it.

56:38 Speaker 6: And so I'm gonna say, only give me photons that

56:40 Speaker 6: have a high density. So this is saying filter the

56:44 Speaker 6: ATIL of three with these parameters and then overlay the

56:49 Speaker 6: ATL twenty four classifications. And so we run that request

56:53 Speaker 6: and we get the data back. Took a little bit longer,

56:55 Speaker 6: A second longer because it's it had to do some

56:58 Speaker 6: more processing with the A and the confidence filtering, and

57:02 Speaker 6: now we have much less photons than before, and we see, interestingly,

57:10 Speaker 6: we see some gaps here now where there were photons

57:13 Speaker 6: that were filtered out by the ETL three filtering but

57:17 Speaker 6: were classified by ATL twenty four. And we do see

57:21 Speaker 6: though some more photons that were classified by ATL twenty four.

57:25 Speaker 6: And then it cleans up all of this, all of

57:28 Speaker 6: the photons if you look back up here, all of

57:31 Speaker 6: this stuff that was obviously not a surface reflection. Some

57:35 Speaker 6: of these filters very quickly cleaned up all of that.

57:39 Speaker 6: And here we're now taking the same plot and we're

57:42 Speaker 6: overlaying color coding it with the yapsy score, so you

57:45 Speaker 6: can see what the aapsy score is of some of

57:46 Speaker 6: these photons that were identified as bothymetry. Okay, so here

57:51 Speaker 6: is now a final step to this process of using

58:01 Speaker 6: ATLO three, and that is we're going to take ATLO

58:05 Speaker 6: three data, we're going to classify it with ATL twenty

58:08 Speaker 6: four data, then we're going to specify to only use

58:11 Speaker 6: the bathymetry photons, and then we're going to run the

58:14 Speaker 6: ATLO six surface surface finding algorithm on it to produce

58:20 Speaker 6: an aggregated surface elevation of the bothymmetry photons. And we're

58:24 Speaker 6: all going to do that on the server making this

58:26 Speaker 6: request here. So this request is constructed. You can see

58:30 Speaker 6: I talked through what it's doing, and this is the

58:34 Speaker 6: magic to tell the server to use the ATL three

58:37 Speaker 6: surface finding algorithm. And so it ran. It took three

58:42 Speaker 6: seconds to run. And now this is the this is

58:46 Speaker 6: the data. We're going to get different fields in this

58:49 Speaker 6: data set because it's this is really now like an

58:51 Speaker 6: ATLO six product, but we've fooled it by giving it

58:56 Speaker 6: bathymetry instead of land ice elevator photons. So we get

59:01 Speaker 6: you know, number of fit photons, we get the surface window,

59:04 Speaker 6: we get the misfit, the armist of the misfit, we

59:07 Speaker 6: get the slope here.

59:11 Speaker 5: And so forth.

59:12 Speaker 6: Okay, so now let's plot this and we get a

59:15 Speaker 6: very nice smooth return of this of the seafloor. Now

59:24 Speaker 6: again you may say, well, there's some things here that

59:27 Speaker 6: that aren't right. You can go back and tweak the

59:30 Speaker 6: parameters here and say, oh, you know what, let me

59:32 Speaker 6: get rid of low confidence photons, or let me do

59:35 Speaker 6: an atail twenty four confidence level.

59:38 Speaker 5: That's that's higher.

59:38 Speaker 6: I only want the really high confidence photons based on

59:42 Speaker 6: ATAIL twenty four and you and you can tweak that

59:46 Speaker 6: to get a very clean But even just with these parameters,

59:49 Speaker 6: we've gotten a very clean surface that's easy to work

59:52 Speaker 6: with and is only a handful right eighty one. Right,

59:56 Speaker 6: this this data set represents the same surface, but as

59:59 Speaker 6: much smaller and easier to work with.

1:00:04 Speaker 5: And then, lastly, for.

1:00:07 Speaker 6: Power users that you just want access to the raw data,

1:00:10 Speaker 6: here's a request that does some different filters, just based

1:00:13 Speaker 6: on ATAIL twenty four. So again, now we've moved past

1:00:16 Speaker 6: the talking about the ATAIL of three. We're back to

1:00:19 Speaker 6: talking about ATL twenty four and some of the more

1:00:21 Speaker 6: corner case features that we support this set. Here, I've

1:00:26 Speaker 6: just chosen to turn off the low confidence filter. You

1:00:28 Speaker 6: could turn on a night filter. Say you want, say

1:00:31 Speaker 6: you have an area of interest and you're just interested

1:00:34 Speaker 6: in pathymetry that was collected at night, Well you could

1:00:37 Speaker 6: turn this on and then we'll only return pothymetry accepted.

1:00:43 Speaker 5: That was collected at night.

1:00:44 Speaker 6: Or there's we have a calculation of what the maximum

1:00:48 Speaker 6: sensor depth is for the Atlas instrument when measuring bothymmetry.

1:00:54 Speaker 6: Sometimes we get we think the algorithm will find bothymmetry

1:00:58 Speaker 6: beyond that sensor death. But if you want to just

1:01:00 Speaker 6: reject all of that, you can turn this one and say,

1:01:04 Speaker 6: if the sensor depth is exceeded, turn it to false,

1:01:09 Speaker 6: and that way you'll only get photons in which the

1:01:13 Speaker 6: sensor depth is not exceeded. So let's take a look

1:01:16 Speaker 6: at this and you'll see now fewer photons because we've

1:01:22 Speaker 6: done some filtering here.

1:01:25 Speaker 5: Just for completeness, I'll plot that there. Now. Lastly, an exciting.

1:01:33 Speaker 6: Feature that we've just added is the ability to directly

1:01:37 Speaker 6: query for AHL twenty four resources from the client, just

1:01:42 Speaker 6: hitting the Python, just hitting.

1:01:47 Speaker 5: The slide rule servers.

1:01:48 Speaker 6: So these are requests that do not use CMR in

1:01:52 Speaker 6: order to querry ATL twenty four resources, and therefore we

1:01:55 Speaker 6: expose some additional query parameters.

1:02:01 Speaker 5: So here is an area of interest. I'm gonna build a.

1:02:05 Speaker 6: Polygon string from that, and now I'm gonna make a

1:02:07 Speaker 6: request for retail twenty four granules with this polygon, and

1:02:11 Speaker 6: I'm gonna say I want the maximum. I want to

1:02:15 Speaker 6: have a maximum mean depth of ten meters. So nothing

1:02:19 Speaker 6: should have a max a mean depth of greater than

1:02:23 Speaker 6: ten meters, so only mean depths for granules less than

1:02:26 Speaker 6: ten meters, and I only want granules that are collected

1:02:29 Speaker 6: in the fall. So there's an encoding for which season.

1:02:32 Speaker 6: Winter is zero, spring is one. Actually, this is the

1:02:39 Speaker 6: that wrong, this is the this is the summer. I

1:02:42 Speaker 6: only want to collect in the summer. Summer's two and

1:02:45 Speaker 6: the fall is three. So I'll run this request. It's

1:02:48 Speaker 6: gonna go out to the servers. It's gonna make that query.

1:02:51 Speaker 6: And now this is all of the granules, have found

1:02:53 Speaker 6: fifty three tracks, and it's returning the granule names with

1:02:57 Speaker 6: the beam that meets this criteria, have.

1:03:00 Speaker 5: Mean depth, mind depth, max depth season.

1:03:05 Speaker 6: We have the number of bathymetric photons and a few

1:03:09 Speaker 6: other things that you can look at. And then also

1:03:12 Speaker 6: now if you want more information about a granule itself,

1:03:19 Speaker 6: we have an API that allows you to say, give

1:03:22 Speaker 6: me information just about this granule. And so we'll run that.

1:03:26 Speaker 6: It went out very quickly, ran out. It made a

1:03:28 Speaker 6: request to the server of the server queried our own

1:03:32 Speaker 6: database and said, okay, for this granule here it has

1:03:36 Speaker 6: these beams. This beam was collected in the summer. Oh,

1:03:41 Speaker 6: there's no bothymetry in any in that beam okay, beam

1:03:45 Speaker 6: three it was collected in the summer, had four photons

1:03:49 Speaker 6: and this was the mean depth, mind depth and max

1:03:52 Speaker 6: depth for that and so forth.

1:03:57 Speaker 1: Thank you, JP. Great demonstration for participates wanting to explore

1:04:00 Speaker 1: the slide roll Python client. Next, we'll hear from Gretchen

1:04:04 Speaker 1: and Kana, who will present on Noah's Satpathy tool and

1:04:07 Speaker 1: integration of NASA's ATL twenty four product. Gretchen and Kiana

1:04:13 Speaker 1: over to you.

1:04:17 Speaker 7: Thank you pretty much.

1:04:19 Speaker 4: We're very excited to talk about our work today on

1:04:23 Speaker 4: noah Stat Value Tool and the integration of NASA's ATL

1:04:26 Speaker 4: twenty four. To begin with, we're going to give a

1:04:29 Speaker 4: little background talk about sat baby and what SDB is

1:04:34 Speaker 4: and its importance, and then give a few examples from

1:04:39 Speaker 4: our SAT value desktop tool. I'll run through a quick

1:04:42 Speaker 4: demo and then pass that off to Canna to talk

1:04:45 Speaker 4: about our ATL ATL twenty four integration and her demo,

1:04:50 Speaker 4: and then if we have time, we'll talk about the

1:04:52 Speaker 4: future of sapathy. So, Noah Stat BATHI Testktop Tool version

1:04:58 Speaker 4: three point one is now official. This is the interface.

1:05:02 Speaker 4: The whole purpose of the value tool is to automate

1:05:05 Speaker 4: the creation of satellite dry bythymmetry. And I should say

1:05:10 Speaker 4: that this desktop tool is for Noah internal use only

1:05:14 Speaker 4: at this point, but I will talk about the future.

1:05:18 Speaker 4: So for those that did not catch Chris Parrish's talk

1:05:21 Speaker 4: in the morning, Satellite dry pathymmetry or SDB is an

1:05:25 Speaker 4: indirect method of bathometric measurement. And while it is not

1:05:29 Speaker 4: the most direct compared to multibede sonar and vathametric lighter,

1:05:35 Speaker 4: SDB is more direct than the gravity based pythymmetry. But

1:05:41 Speaker 4: given that, why do we care about STAB? The example

1:05:45 Speaker 4: on the left or the right give a good example

1:05:49 Speaker 4: of why it's important. The top right image is the

1:05:53 Speaker 4: batometric lighter example from our group and wrote sunsing division

1:05:56 Speaker 4: in the National g JIC Survey, and this is a

1:05:59 Speaker 4: saint shows if you look at the lower SDB example

1:06:05 Speaker 4: that doctor Rick Stump created, it fills in the gaps

1:06:07 Speaker 4: of the bathymetric lighter. So there are some uses for him.

1:06:12 Speaker 4: But what is Sallee drieth inventary in detail? For our

1:06:16 Speaker 4: purposes today and sat pathy, we use the method of

1:06:21 Speaker 4: using sentinel two resolution of ten meters and updated research

1:06:26 Speaker 4: from doctor Rick Stump and doctor Isabel Camelero and I

1:06:31 Speaker 4: should note that all SDBA methods can be grouped into

1:06:34 Speaker 4: three method major character categories. The first one is spectral

1:06:39 Speaker 4: SDP special radiance for example, radio transfer equation or RTE optimization,

1:06:46 Speaker 4: RTE empirical and VAN ratio. The second is photochrometric and

1:06:52 Speaker 4: the wave and the third is wave kinematics. For our

1:06:55 Speaker 4: purposes today, we're going to focus on the band ratio approach,

1:06:58 Speaker 4: which is based on a relationship between reflectance and depth

1:07:02 Speaker 4: and it often requires but not always, which I'll explain

1:07:05 Speaker 4: a little bit when we talk about our simplified calibration

1:07:10 Speaker 4: requires a known dep depth data to calibrate the composed

1:07:15 Speaker 4: compositing approach for the data. The compositing approach for the

1:07:18 Speaker 4: Van ratio solution inherently removes a lot of the faulty data.

1:07:24 Speaker 4: So we're asked a lot of times is why do

1:07:26 Speaker 4: we need u s dB for our purposes at NAWA,

1:07:30 Speaker 4: it's for intermonautical chart updates, hydrographic survey planning and reconnaissance

1:07:35 Speaker 4: change analysis, and to fill in atometric glider caps in

1:07:40 Speaker 4: non navigationally stigific areas. And then image below as an

1:07:46 Speaker 4: example from our NOAH chart. These areas of unsurveyed locations

1:07:52 Speaker 4: still exist in areas like Alaska, but I'm going to

1:07:56 Speaker 4: focus on the hydrographic survey planning in Wisconsin. The reason

1:08:02 Speaker 4: i SDB is so important, as we've been using it

1:08:04 Speaker 4: quite a bit, is it's become a vicious cycle where

1:08:07 Speaker 4: hydrographic survey data sometimes doesn't exist. We need the data,

1:08:12 Speaker 4: then the hydro data. The hydro vessels need to collect

1:08:15 Speaker 4: the data. However, if they're an area where there isn't

1:08:18 Speaker 4: any data to operate safely, it becomes very dangerous. This

1:08:23 Speaker 4: is where SDB is very has become very important to

1:08:25 Speaker 4: provide some amount of data so the surveys can be

1:08:28 Speaker 4: as safe as possible. So my first example is of

1:08:34 Speaker 4: Saint Matthew's. Saint Matthews is off of in Alaska, off

1:08:39 Speaker 4: the coast, and the e n C or Electronic Navigation chart.

1:08:44 Speaker 4: While it does.

1:08:45 Speaker 7: Show some areas where there are rocks and areas of.

1:08:48 Speaker 4: Danger, where to the a survey of vessel like the

1:08:51 Speaker 4: Fairweather need to stay away from.

1:08:56 Speaker 7: We were asked when the nose ship.

1:08:58 Speaker 4: Fairweather had to survey this area, we could run the

1:09:03 Speaker 4: sad bady tool to provide reconnaissance data. The pseudopathymetry read

1:09:10 Speaker 4: examples on the left and it shows some areas that

1:09:13 Speaker 4: are not captured in the E n C current chart.

1:09:20 Speaker 4: This helps keep the noise ship fair weather safe to

1:09:23 Speaker 4: the best extent possible. Another example is in Kotsubu, Alaska.

1:09:29 Speaker 4: That's in the northern part of Alaska, and if you

1:09:32 Speaker 4: can take a look at the E and C on

1:09:34 Speaker 4: the left, there's a lot of gray areas that basically

1:09:37 Speaker 4: says there are no charted depths here, So you're basically

1:09:41 Speaker 4: working blind when you're serving in this area.

1:09:43 Speaker 7: Very dangerous.

1:09:45 Speaker 4: Again, we were asked to run the sad Bady tool

1:09:48 Speaker 4: and to provide data that would provide some information on

1:09:51 Speaker 4: the thethymmetry. What we found was not just the thethymmetry,

1:09:55 Speaker 4: but channels, multiple channels, two channels to show where they

1:10:00 Speaker 4: can navigate somewhat safely. So the next I'm going to

1:10:04 Speaker 4: run into the sat patty dust optimum. So this is

1:10:09 Speaker 4: the the sat patty desktop tool interface, and I'm going

1:10:18 Speaker 4: to zoom in on our area of interest, which should

1:10:22 Speaker 4: be around the area of Miami.

1:10:24 Speaker 7: Each of these.

1:10:25 Speaker 4: Squares are the set of two tiles, and I will

1:10:28 Speaker 4: select the tile that I'm interested in and zoom in

1:10:34 Speaker 4: further intro and the area of interest. You can also

1:10:43 Speaker 4: provide an AOI that's a ZIP shape pile, and then

1:10:48 Speaker 4: I'm going to select the project directory. Next, we're going

1:11:02 Speaker 4: to figure out what start date we want. So for

1:11:05 Speaker 4: the Miami area it's relatively clear in Florida most of

1:11:10 Speaker 4: the year, except when there's hurricanes. We're going to go

1:11:14 Speaker 4: up to January to January first, and we'll do it

1:11:23 Speaker 4: in data a couple of weeks earlier. For the cloud cover,

1:11:28 Speaker 4: we'll just leave it as default as twenty percent. You

1:11:31 Speaker 4: can move it to eight, ten percent or less cloud shadows.

1:11:36 Speaker 4: We'll talk a little bit about that as we're processing

1:11:39 Speaker 4: some of the data. It's no and nice you don't

1:11:41 Speaker 4: have to worry about. And I'll show example of the

1:11:43 Speaker 4: no data cover what I mean by that. It's now

1:11:46 Speaker 4: we're set. We're going to search for the imagery. The

1:11:50 Speaker 4: little dial shows that it's searching, and while we're doing that,

1:11:53 Speaker 4: I'm going to talk a little bit about well, it's

1:11:56 Speaker 4: already finished. Actually, let's look at our search results.

1:12:02 Speaker 7: It was pretty clear.

1:12:04 Speaker 4: So this is what it looks like after the imagery

1:12:07 Speaker 4: is run through, and we can go through and select

1:12:12 Speaker 4: which ones are the least turbid have the least amount

1:12:15 Speaker 4: of clouds. This is quite a bit turbid, so we

1:12:19 Speaker 4: don't want to use something like that. We don't want

1:12:21 Speaker 4: to use the stuff with all the clouds. I'll explain

1:12:25 Speaker 4: a little bit more about them. I'm going to go

1:12:31 Speaker 4: back to the first one. This looks pretty clear compared

1:12:34 Speaker 4: to this one looks pretty turbid still in this area, So.

1:12:37 Speaker 7: I hadn't selected.

1:12:39 Speaker 4: In a perfect world, if we had more time, we

1:12:44 Speaker 4: would select ideally six to eight images to create the composite.

1:12:47 Speaker 7: And I'll talk a little bit more about that.

1:12:49 Speaker 4: But the interesting time one was going to select one

1:12:52 Speaker 4: and then I'm going to submit the scenes. The SAT

1:12:56 Speaker 4: value workflow and composite product procedure for a given area

1:13:00 Speaker 4: interest or AOI Multiple sentinel two alwe scenes are collected

1:13:04 Speaker 4: and atmospherically corrected to remote sensing reflectance or RSS. Red, green,

1:13:09 Speaker 4: and blue bands are used as inputs to the band

1:13:11 Speaker 4: ratio models PSDB red and PSTV green which stands for

1:13:17 Speaker 4: pseudo STP. In addition, turbidity proxies are specified in the

1:13:22 Speaker 4: red edge seven O four band and the OC three

1:13:25 Speaker 4: based chlorophylla A product. So this is the PSD red

1:13:29 Speaker 4: in the psd green, so the red edge For each scene,

1:13:35 Speaker 4: individual pseudopithymmetry layers are generated and then composited along the

1:13:39 Speaker 4: temple dimension by selecting the maximum PSTV value per pixel

1:13:44 Speaker 4: representing the deepest least turbid observation. The temporal index of

1:13:50 Speaker 4: this maximum pseudo SDP is recorded and applied to extract

1:13:54 Speaker 4: corresponding values from the red edge and chlorophyl a letters,

1:13:57 Speaker 4: producing composite turbity turbidity peroxy maps. These composite products provide

1:14:03 Speaker 4: context for any residual tibidity and service inputs to the

1:14:07 Speaker 4: subsequent steps within the SAP value processing cheap. As they

1:14:14 Speaker 4: spoke about before, when you're selecting imagery, there are free

1:14:18 Speaker 4: things you need to take into consideration. We want low turbidity.

1:14:27 Speaker 4: When there's turbidity in the water column, the calculated DUFF

1:14:31 Speaker 4: is going to be more shallow than the actual value,

1:14:33 Speaker 4: so that's important to know. They also want little or

1:14:37 Speaker 4: no sunland sunglint, or the reflection of the sun off

1:14:40 Speaker 4: the water surface obscures the ability of the satellite sensor

1:14:44 Speaker 4: to properly capture reflectance and then see little or no

1:14:49 Speaker 4: little or no shadows due to clouds. Those are the

1:14:54 Speaker 4: stark areas here, shadows cast by clouds obscure the finalized

1:14:58 Speaker 4: DEBA product by goverestimating death values and then d little

1:15:04 Speaker 4: or no clouds. In the AOI we talked a little

1:15:06 Speaker 4: bit about that the presence of clouds also affects the

1:15:09 Speaker 4: final STV results by underestimating death values. Snow and ice.

1:15:15 Speaker 4: There's another one, little or no snow and ice. Snow

1:15:20 Speaker 4: and ice on or near the shore line affects the

1:15:22 Speaker 4: final STV results and also underestimates the death results. So

1:15:30 Speaker 4: I'm going to see if it's ready and it is

1:15:34 Speaker 4: process is incomplete. Take a look at our map results.

1:15:38 Speaker 4: As you can see there is our items take off.

1:15:43 Speaker 7: Some of these a little easier to see. So this

1:15:54 Speaker 7: is the s the.

1:15:55 Speaker 4: PSP red, and the PSDB green. We're going to focus

1:15:58 Speaker 4: on the PSDB red first. So now I'm going to

1:16:02 Speaker 4: walk into the walkay through the simplified calibration procedure. I

1:16:13 Speaker 4: think that there we go, so Stafady's simplified calibration procedure.

1:16:19 Speaker 4: As you recall I said, you normally need the charted

1:16:23 Speaker 4: soundings or reference survey to help tie the the PSDB

1:16:29 Speaker 4: down to to have actual UH true drops and meters.

1:16:35 Speaker 4: So we're going to convert the pseudo SDB drive from

1:16:38 Speaker 4: spectral ratios into true drops intometers. This option is beneficial

1:16:43 Speaker 4: for areas where we can't we don't have reconnaissance, or

1:16:48 Speaker 4: we lack recent high quality reference data. And the relationship

1:16:52 Speaker 4: is defined by PSDB red I'm sorry. SDB equals m one,

1:16:56 Speaker 4: which is the slope times PSDB e minus m not

1:17:01 Speaker 4: which is the offset here, where M one represents the

1:17:04 Speaker 4: slope and m not represents the offset. Okay, all right,

1:17:13 Speaker 4: so the next thing I'm going to talk about it

1:17:16 Speaker 4: is the offset parameter or I'm not just a side,

1:17:20 Speaker 4: I'm here. The offset parameter is selected by sampling a

1:17:23 Speaker 4: near ship pixel of the landwater interface, which serves as

1:17:26 Speaker 4: an approximate local data. It represents the PSD value at

1:17:30 Speaker 4: depth of zero and serves this initial reference point for

1:17:33 Speaker 4: the relationship between PSDB and water depth. When tidal information

1:17:38 Speaker 4: is available, the offset parameter can be further refined with

1:17:42 Speaker 4: a well with a known vertical reference such as local

1:17:46 Speaker 4: sea level. Okay, so we're going to take the magnifying

1:17:49 Speaker 4: glass to hover over these area along the shoreline. We'll

1:17:53 Speaker 4: see these values here will change through our research we

1:18:01 Speaker 4: know that the PSDB offset range is between point nine

1:18:06 Speaker 4: three and one point zero three ahead towards the interior,

1:18:19 Speaker 4: not the channel. See if we can find something, looks

1:18:33 Speaker 4: pretty good.

1:18:35 Speaker 7: Okay.

1:18:35 Speaker 4: So they now that I've clicked the offset, it fills

1:18:38 Speaker 4: in automatically. We're going to do the same thing for

1:18:40 Speaker 4: the PSDB green. Put this down to PSD green. You

1:18:46 Speaker 4: can usually pick the same from the same areas. The

1:18:49 Speaker 4: range the research for the PSB green offset is between

1:18:55 Speaker 4: point eight five.

1:18:56 Speaker 7: And point nine five.

1:19:00 Speaker 4: See that looks pretty good, okay. So now that's filled

1:19:08 Speaker 4: into the offset, so point nine to eight. We're going

1:19:12 Speaker 4: to go ahead and submit the parameters zum out a

1:19:18 Speaker 4: little bit.

1:19:20 Speaker 8: Okay, so a little bit more slow parameters M one.

1:19:27 Speaker 8: The slope M one value is roughily indicative of how

1:19:30 Speaker 8: strongly light it tenuates with depth in the water column.

1:19:34 Speaker 8: Higher tenuation, for example, more turbid or scattering waters requires

1:19:38 Speaker 8: a steeper slope, while lower tenuation and clear water corresponds

1:19:42 Speaker 8: to a short, shallower slope. An initial value is provided

1:19:47 Speaker 8: based on the mean of slope estimates reported by publications

1:19:50 Speaker 8: from doctor Rick Stump and doctor Isabel cabell Aro, which

1:19:53 Speaker 8: drive the calibration parameters to the regression with reference slid

1:19:56 Speaker 8: our data sets and or charted sellings.

1:20:00 Speaker 7: Go ahead and something.

1:20:04 Speaker 4: Because the optical properties and water clarity can read very regionally,

1:20:08 Speaker 4: users can fine tune the slope in plus or mine

1:20:11 Speaker 4: plus or minus five percent increments to better match local

1:20:14 Speaker 4: water conditions. Adjustment is constrained within specific ranges where the

1:20:18 Speaker 4: PSDB red is between four point eight eight and sixty

1:20:22 Speaker 4: point five and SDB green is between forty two point

1:20:25 Speaker 4: seven and eighty eight point four five. Supporting tubidity indicate

1:20:29 Speaker 4: caters such as the seven O four reddit good band,

1:20:34 Speaker 4: which is here in our those literary products.

1:20:39 Speaker 7: There we go. So that's the red edge. This is

1:20:45 Speaker 7: our chlorophyll. Oops, there we go. Red edge in chlorophyll.

1:20:54 Speaker 4: Can guide this adjustment by providing context on prevailing water quality,

1:20:58 Speaker 4: for example clear, moderately turbid or extremely turbid. Okay, so

1:21:04 Speaker 4: now that we have that, we're gonna we'll now have

1:21:08 Speaker 4: the We should now have the STB red and the

1:21:19 Speaker 4: SDV green players not available. Okay, so now you'd want

1:21:24 Speaker 4: to do a quick check of the SDB red make

1:21:27 Speaker 4: sure there's no negative values.

1:21:29 Speaker 7: It's pretty good.

1:21:30 Speaker 4: We're going to just stick to the shore between zero

1:21:32 Speaker 4: and five because that's really SDB red generally where the

1:21:38 Speaker 4: limits are and it works the best. And then the

1:21:42 Speaker 4: green we're going to do the same thing. Everything from

1:21:48 Speaker 4: probably about five meters to on should be in the

1:21:53 Speaker 4: positive reade. You just want to do a double check.

1:21:56 Speaker 4: That looks pretty good, Okay, So that's the firstness we'll

1:22:01 Speaker 4: check on that. So the area in Miami has basically

1:22:09 Speaker 4: a low tithe that is optimal first as satpathy. So

1:22:13 Speaker 4: in this side, I'm going to talk about the importance

1:22:17 Speaker 4: of tides and macro and macro tidal areas. These red

1:22:21 Speaker 4: points that we created represent the tie aid stations and

1:22:24 Speaker 4: their values are calculated through NOASV data tooled by subtracting

1:22:28 Speaker 4: the mean mobile water value from the mean high water

1:22:30 Speaker 4: value for that station. The difference between the two value

1:22:34 Speaker 4: represents the tide range of that location. If the tide

1:22:37 Speaker 4: range is a meter or less, it simplifies things for sappathy.

1:22:41 Speaker 4: This means and what is case dependent and general tide

1:22:44 Speaker 4: adjustment is not necessary because it falls within the vertical

1:22:48 Speaker 4: uncertainty of STB generated through satpathy. When the tide range

1:22:54 Speaker 4: is larger than one meter, locations require a tide adjustine

1:22:59 Speaker 4: because the potential difference between the title stages exceeds inherent

1:23:03 Speaker 4: uncertainty of SAP badie. We'll talk more about this when

1:23:06 Speaker 4: we discuss the time model.

1:23:09 Speaker 7: I'm going to go back to.

1:23:12 Speaker 3: Our tool.

1:23:14 Speaker 4: And to select the different options. I'm going to add

1:23:25 Speaker 4: the reference data. That's another way to kind of see

1:23:28 Speaker 4: if we're in the ballpark. So this is remote sensing,

1:23:35 Speaker 4: our Remote Senstensing Division Topo Batty data under the National

1:23:39 Speaker 4: Genetic Survey, and that.

1:23:43 Speaker 7: Will show up when I select.

1:23:47 Speaker 3: There we go.

1:23:48 Speaker 4: So let me take off our interim products so you

1:23:51 Speaker 4: can see this. So this is a little snippet of

1:23:54 Speaker 4: our data reference data.

1:24:01 Speaker 7: I'm gonna expand this.

1:24:02 Speaker 4: Because okay, so the SDB red, so the auxilliary grid

1:24:08 Speaker 4: the values are shown here. That's for the reference data.

1:24:11 Speaker 4: And I'm gonna flip this s t BE red and

1:24:14 Speaker 4: as they move across to see that the numbers change.

1:24:18 Speaker 4: So I'm gonna stick to probably close to the shore.

1:24:22 Speaker 4: I'm just gonna run around and see how close we are,

1:24:26 Speaker 4: so we don't anything like a half a meters between

1:24:31 Speaker 4: is okay that's roughly okay, Okay, we'll do the same

1:24:36 Speaker 4: thing for the SDB green. Oops again, we're gonna go

1:24:45 Speaker 4: a little farther out for the green. We do sometime

1:24:49 Speaker 4: after five meters. We're just doing a cursory glance here,

1:24:56 Speaker 4: all right, So those values look generally okay, We're go ahead,

1:25:00 Speaker 4: and you can always come back and adjust it again,

1:25:04 Speaker 4: but we're going to go ahead and submit those parameters.

1:25:08 Speaker 4: So before I hand this over to Kenna, I'm gonna

1:25:14 Speaker 4: talk about our motivation at Noah for integrating the ATL

1:25:17 Speaker 4: twenty four into SATPATHI first, also an entirely independent reference

1:25:22 Speaker 4: data source.

1:25:23 Speaker 7: It's near global coastal.

1:25:25 Speaker 4: Coverage is ideal for US well. It has a maximum

1:25:31 Speaker 4: depth ranges which are generally quite comparable to one seci

1:25:34 Speaker 4: up and occasionally slightly deeper and are The ATL twenty

1:25:39 Speaker 4: four actually tests confirm and stability for STB calibration and melidation. So,

1:25:45 Speaker 4: like I was talking about with the the future of

1:25:48 Speaker 4: SAP ABBY, the reason why we're interested in things like

1:25:52 Speaker 4: near global coastal coverage is because we were asked to

1:25:56 Speaker 4: expand the sat AALBI, to not just create a themetry

1:26:03 Speaker 4: just for United States, but to go globally, to work

1:26:05 Speaker 4: with our federal partners and with cost survey, so the

1:26:12 Speaker 4: global coverage is very advantageous, as well as the accuracy

1:26:16 Speaker 4: tests to make us assured that we're using reference data

1:26:22 Speaker 4: that's suitable. So that's a really incredible option and we

1:26:28 Speaker 4: are working on Although the desktop tool is just available

1:26:32 Speaker 4: for Noah at this point, there is a public site

1:26:37 Speaker 4: that's going to be that's in work now that's going

1:26:40 Speaker 4: to take advantage of the A TALE twenty four. So

1:26:43 Speaker 4: we'll talk about that hopefully in another presentation. Now i'd

1:26:47 Speaker 4: like to pass this off to Canon to talk about

1:26:50 Speaker 4: her ATL twenty four demo, and yeah, thank you Canon,

1:27:00 Speaker 4: Thank you Gretchen.

1:27:03 Speaker 9: So we get ATL twenty four data in side role

1:27:06 Speaker 9: using side rules Python API, which allows us to request

1:27:10 Speaker 9: granule data from October twenty eighteen to November twenty twenty four,

1:27:15 Speaker 9: when release one of ATL twenty four was published. In

1:27:19 Speaker 9: satbathy we use a greater confidence threshold than the default

1:27:23 Speaker 9: value of zero point six, and this is to prevent

1:27:26 Speaker 9: some false pathymetry that occurs, especially around the SA surface,

1:27:31 Speaker 9: where the bottom of the sea surface can get falsely

1:27:35 Speaker 9: identified as pathymetry. The ATL twenty four implementation takes the

1:27:45 Speaker 9: satellite derived bathymetry photon data and then it regrids that

1:27:50 Speaker 9: photon data into a ten meter resolution raster to match

1:27:54 Speaker 9: the resolution of satbathy. This is to make it easy

1:27:58 Speaker 9: for use with Brian Eaters valalidation code, which produces air

1:28:02 Speaker 9: analysis heat scatters and histogram plots for SDB red, SDB green,

1:28:09 Speaker 9: and SDB merged. These values can be used to validate

1:28:13 Speaker 9: results or can be used to recalibrate the shoreline. If

1:28:18 Speaker 9: we go to an example here, I have set up

1:28:23 Speaker 9: a sat bathy run that is very similar to what

1:28:26 Speaker 9: Gretchen showed us just a moment ago. I've got a

1:28:28 Speaker 9: similar area of interest selected. I've gone through the tile

1:28:33 Speaker 9: selection and calibration steps, and so now all you need

1:28:37 Speaker 9: to do to get ATL twenty four products is check

1:28:39 Speaker 9: this ATL twenty four comparison checkbox before you click create

1:28:43 Speaker 9: final Products. If you look at the terminal while this

1:28:47 Speaker 9: is happening, you'll see the request to satbathy start to generate.

1:28:51 Speaker 9: It's going to return photons in a random temporal order,

1:28:56 Speaker 9: just due to the nature of how side rule gathers

1:28:58 Speaker 9: the data, So we do a sort on that data

1:29:01 Speaker 9: from oldest to newest. That way, if a photon shares

1:29:05 Speaker 9: a pixel in the final raster. It'll use the newest

1:29:09 Speaker 9: photon to generate that pixel's value. This can take a

1:29:15 Speaker 9: moment to process, so I will just go and take

1:29:17 Speaker 9: us to an already completed folder. So to get to

1:29:22 Speaker 9: your data, you're going to go into your output folder.

1:29:25 Speaker 9: You're going to go into the prod folder and into

1:29:28 Speaker 9: the ATL twenty four sub folder. In here you're going

1:29:33 Speaker 9: to see the aristatistics, heat scatter and histogram results for

1:29:37 Speaker 9: your SDB green, SDB merged, and SDB red. You're also

1:29:42 Speaker 9: going to have a data subfolder, and in here will

1:29:45 Speaker 9: both be the ATL twenty four dot TIF raster that

1:29:48 Speaker 9: we use in satbathy, but also a CSB file that

1:29:51 Speaker 9: contains all the photon data retrieved for this request. That way,

1:29:55 Speaker 9: because it is slightly lossy to generate the raster, you

1:29:59 Speaker 9: can access all of the ATL twenty four data, including

1:30:02 Speaker 9: any of those photons that didn't make it into the

1:30:04 Speaker 9: final raster. This ATIL twenty four TIF, for example, can

1:30:12 Speaker 9: be used to pull it into qgis. I've already loaded

1:30:16 Speaker 9: the data just by taking the file and dragging and

1:30:19 Speaker 9: dropping it onto qgis with a base map already loaded in,

1:30:24 Speaker 9: and then I've set the colors to go from shallow,

1:30:27 Speaker 9: which is a yellow color, to a darker color in purple,

1:30:31 Speaker 9: and we can zoom in and see these lines. Here

1:30:34 Speaker 9: are our granule tracks that relate to the photon data

1:30:38 Speaker 9: that we've returned. Just an example of what this output

1:30:43 Speaker 9: file looks like. We'll go ahead and look at the

1:30:46 Speaker 9: merged values. So in merged we have air statistics, and

1:30:50 Speaker 9: so this will give you the median absolute air, the

1:30:53 Speaker 9: mean absolute air, the bias, and then the number of

1:30:56 Speaker 9: photons relating to that bin that we've done for the statistics.

1:31:02 Speaker 9: So from zero to ten meters, the bins are in

1:31:06 Speaker 9: two meter increments, and then from ten meters onward the

1:31:10 Speaker 9: bins are in five meter increments, just because in those

1:31:14 Speaker 9: deeper values you can see like ten six and one

1:31:17 Speaker 9: photon returns, the returns just get smaller. This is an

1:31:24 Speaker 9: example of what the heat scattering looks like. On the

1:31:28 Speaker 9: y axis, we have the reference data set, which in

1:31:30 Speaker 9: this case is ATL twenty four, and on the X

1:31:32 Speaker 9: axis we have SDB merged, which is the SDB merged

1:31:36 Speaker 9: results from our satbaty run. As we can see, there's

1:31:39 Speaker 9: fairly good correlation on this one to one axis, which

1:31:43 Speaker 9: generally means that the ATL twenty four data is validating

1:31:48 Speaker 9: the results from Sabbath and then we can look into

1:31:53 Speaker 9: these histograms that give us an idea of how the

1:31:57 Speaker 9: photon spread is within those bins. Alrighty, and then that's

1:32:11 Speaker 9: my person and I'll pass it back to Gretchen.

1:32:13 Speaker 4: All right, So thank you Kenna. Now I'm gonna give

1:32:18 Speaker 4: some major updates for tak a little bout our major

1:32:21 Speaker 4: updates for the SATPADI desk top version three point one. So,

1:32:26 Speaker 4: as Kana mentioned, we have the validation code that should

1:32:31 Speaker 4: be ready to show you and our actual data. So

1:32:38 Speaker 4: I'm gonna that's in our validation folder and we can

1:32:43 Speaker 4: see how we did. Let's take a look at the

1:32:47 Speaker 4: sv merged heat scatter PT this.

1:32:55 Speaker 7: Okay, so this looks pretty good.

1:32:59 Speaker 4: This is the generally from like I said, the zero

1:33:02 Speaker 4: to four is the SDB red so they get merged.

1:33:08 Speaker 4: The green STB, which is from about five meters to

1:33:12 Speaker 4: looks like it's about a little over eight. So this

1:33:16 Speaker 4: aligns pretty well. The reference data is not we used

1:33:19 Speaker 4: was again our remote sensing division to a badly lighter

1:33:24 Speaker 4: under the nest g n X survey. And Noah, we

1:33:29 Speaker 4: have a lot more data compared to the ATL twenty four,

1:33:32 Speaker 4: but so you can see a lot of more data points,

1:33:36 Speaker 4: which is nice.

1:33:36 Speaker 7: So it looks like it lined up pretty well.

1:33:38 Speaker 4: So the items, the the values we picked looked like

1:33:43 Speaker 4: it was on. But if it is a little off,

1:33:45 Speaker 4: you would do adjustment in the in the tool, you

1:33:52 Speaker 4: can use the multiplier. Let's say the greening was a

1:33:55 Speaker 4: little the dark spot was a little below that line,

1:33:59 Speaker 4: you would to the adjustment for the green here down here,

1:34:04 Speaker 4: and we probably do the minus and and you can

1:34:09 Speaker 4: always go back. That's the nice ability of the tools

1:34:12 Speaker 4: to be able to like adjust and recreate the products.

1:34:18 Speaker 7: It's a nice manifit.

1:34:21 Speaker 4: Okay, I go back to So that's and then can

1:34:27 Speaker 4: I already talked about the histogram and the aero statistics.

1:34:31 Speaker 4: I won't go over that, but it's very similar.

1:34:36 Speaker 7: And then.

1:34:39 Speaker 4: I'll talk a little bit about the tide module is

1:34:42 Speaker 4: I spoke about earlier. For areas like Miami and Hatteras,

1:34:46 Speaker 4: those are below one meter tied our water level. But

1:34:52 Speaker 4: for areas in Alaska that we've been working quite a bit,

1:34:55 Speaker 4: sometimes it can be five meters. So in these cases,

1:34:58 Speaker 4: in other areas where we don't have information on tide

1:35:03 Speaker 4: we can use the on datum SDB composite product using

1:35:08 Speaker 4: the Vizio tide model. So the workflows as follows. We

1:35:13 Speaker 4: extract a sentinel two scene gradual sensing time, query the

1:35:17 Speaker 4: tide model at each x y t point, create gridded

1:35:21 Speaker 4: water level rasters at local means level, and composite the

1:35:25 Speaker 4: water level of rasters according to maximum PSD RAD or green,

1:35:29 Speaker 4: and then finally we adjust the calibrated sd B red

1:35:32 Speaker 4: or SDB green products using the water level grid composite.

1:35:36 Speaker 4: The examples that we have to the right are for Hatterus.

1:35:38 Speaker 4: That's one of our test regular test locations to make sure.

1:35:41 Speaker 7: Everything was working correctly.

1:35:44 Speaker 4: Normally, we already know what the tides are for Hatteras,

1:35:47 Speaker 4: but for places outside that don't or uh maybe globally

1:35:53 Speaker 4: that we may not know is that will be very helpful.

1:35:57 Speaker 7: And there is plans to update to the most recent

1:36:01 Speaker 7: of the ZIO.

1:36:03 Speaker 4: So going back to where everything is, if you want

1:36:09 Speaker 4: to access that, that's in the SDB UH composites folder.

1:36:18 Speaker 4: So if you see the local mean sea level and

1:36:21 Speaker 4: at the beginning of the tiff and the water level,

1:36:24 Speaker 4: I have that already up along with the final project

1:36:29 Speaker 4: SDB merged, so this is the SDB merged this is

1:36:32 Speaker 4: indicates deeper water.

1:36:34 Speaker 7: Here and then.

1:36:37 Speaker 4: The tie module local mean Sea level adjusted. That is

1:36:44 Speaker 4: the product for my the MIAMI so resources going forward

1:36:51 Speaker 4: for folks that are looking for more information on sat

1:36:54 Speaker 4: pavvy and actually the research that all went into it.

1:36:58 Speaker 4: These are papers from doctor Cabellero and doctor Rick Stump,

1:37:06 Speaker 4: and I really like to do a special thanks to

1:37:09 Speaker 4: the SAUT Valley team, especially Brian Atter who has been

1:37:11 Speaker 4: a huge help for all their continued support, and thanks

1:37:20 Speaker 4: to NASA for allowing us to do this presentation and training.

1:37:30 Speaker 1: Thank you Gretchen and Keana for the terrific presentation and

1:37:33 Speaker 1: demonstration on satpathya. The following slides provide a summary of

1:37:39 Speaker 1: the concepts covered in part two of the webinar series.

1:37:43 Speaker 1: ATL twenty four can be integrated with optical bands blue

1:37:46 Speaker 1: and green green wavelengths for driving satellite drive pathymetry. Slide

1:37:51 Speaker 1: roll is a public web service with low latency access

1:37:54 Speaker 1: to on demand data products stored in S three. Slide

1:37:59 Speaker 1: role provides on demand processing next to the data for

1:38:02 Speaker 1: generating customized data products using parameters supplied in the user's request.

1:38:08 Speaker 1: Slide roll Python client can be used to access ATL

1:38:11 Speaker 1: twenty four. Slide roll provides documentation for the Python client,

1:38:15 Speaker 1: including example scripts, user guide, and developers Guide. SATPATHI is

1:38:21 Speaker 1: a hybrid web desktop tool created by Noah for automating

1:38:25 Speaker 1: automating satellite derived pathymmetry. Satpathy utilizes ten meter resolution satellite

1:38:31 Speaker 1: imagery from the Copernicus Sentinel two mission, Amazon Web service

1:38:36 Speaker 1: and Acolyte Atmospheric Correction processor. Satpathy integrated the ATL twenty

1:38:42 Speaker 1: four product as an independent reference data source. The following

1:38:47 Speaker 1: is a summary of topics covered over both days of

1:38:50 Speaker 1: the training. NASA's i AT two carries a photon counting

1:38:54 Speaker 1: laser altimeter named Atlas. Atlas provides near contiguous a long

1:38:59 Speaker 1: track sample using six individual beams of green light at

1:39:03 Speaker 1: five hundred and thirty two nanimeter wavelength, providing high vertical resolution.

1:39:08 Speaker 1: Every photon detected by Atlas has a latitude, longitude, and

1:39:12 Speaker 1: elevation associated with it. National Snow and Ice Data Center

1:39:16 Speaker 1: provides access to ATL twenty four data metadata and tools.

1:39:22 Speaker 1: Slide roll is a public web service with low latency

1:39:25 Speaker 1: access to on demand data products stored in S three.

1:39:29 Speaker 1: Slide roll Python Api can be used to access ATL

1:39:32 Speaker 1: twenty four. Satbathi is a hybrid web desktop tool created

1:39:38 Speaker 1: by Noah for automating satellite drive bathymetry. Satpathy utilizes ten

1:39:42 Speaker 1: meter resolution satellite imagery from the Copernicus Sentinal two mission

1:39:46 Speaker 1: Amazon Web service and Acolyte. Satpathy integrates the ATL twenty

1:39:53 Speaker 1: four product as an independent reference data source. Before we

1:39:59 Speaker 1: transition to the Q and A session, I want to

1:40:01 Speaker 1: remind you there will be one homework assignment accessible from

1:40:04 Speaker 1: the training page starting today. Answers must be submitted by

1:40:08 Speaker 1: Google Form with a due date of December thirty first.

1:40:12 Speaker 1: To receive a certificate of completion, you must attend all

1:40:16 Speaker 1: live webinars and complete the homework assignment by the deadline.

1:40:21 Speaker 1: You will receive a certificate via email approximately two months

1:40:24 Speaker 1: after the completion of the course. We want to thank

1:40:29 Speaker 1: once more doctor Christopher Parish from Oregon State University, JP

1:40:33 Speaker 1: Swinskey from NASA's Goddard Spaceflight Center, Gretchen Emilhore from Noah,

1:40:37 Speaker 1: and Keana Key from Oregon State University for their presentations

1:40:41 Speaker 1: and instructive demos. Below is the contact information for Chris, JP,

1:40:47 Speaker 1: Gretchen and Keana, along with links to the urset website

1:40:51 Speaker 1: and social media. If you enjoyed today's webinar, we hope

1:40:55 Speaker 1: you will sign up on the r set list serve

1:40:56 Speaker 1: to receive notifications of future trainings, and please follow us

1:40:59 Speaker 1: on social media for other relevant announcements pertaining to NASA's

1:41:03 Speaker 1: Earth Sciences. We will now transition to the question and

1:41:07 Speaker 1: answer portion of today's training.

1:41:17 Speaker 3: Okay, can you guys hear me?

1:41:19 Speaker 1: Okay, yes, Gretchen, Yes, we can hear you.

1:41:24 Speaker 3: Yes, thank you. I can start on the questions, starting

1:41:29 Speaker 3: with the question nineteen. Does that sound okay?

1:41:32 Speaker 1: Actually we start from the first, so we're going to.

1:41:34 Speaker 3: Jump Oh, I beg your pardon.

1:41:36 Speaker 2: So yeah.

1:41:37 Speaker 1: So question one, When integrating ATAIL twenty four data sets

1:41:41 Speaker 1: from ice AT two for shallow water but symmetry, do

1:41:43 Speaker 1: you typically use machine learning approaches? If yes, which techniques

1:41:47 Speaker 1: are most frequently applied example given random forest, convolutional neural

1:41:52 Speaker 1: network XG, BOOST, et cetera. And why are they preferred

1:41:55 Speaker 1: over traditional methods?

1:42:03 Speaker 2: I think that was my answer to that question, and

1:42:05 Speaker 2: then sorry, my answer ended up in a couple of

1:42:07 Speaker 2: different pluses on the on the screen there. That's a

1:42:12 Speaker 2: really active research topic. It seems like their paper is

1:42:14 Speaker 2: being published every month on looking at different machine learning

1:42:19 Speaker 2: methods for STB, So I would just point to a

1:42:24 Speaker 2: couple of recent review papers that were published on STB.

1:42:28 Speaker 2: And also we did a recent I st to bothymmetry

1:42:31 Speaker 2: review paper that was by John at All, just published

1:42:34 Speaker 2: this year. It contains a lot of references.

1:42:40 Speaker 1: Hey, Chris, thank you so much. Question number two, could

1:42:42 Speaker 1: you suggest some case studies or I sad to ETL

1:42:45 Speaker 1: twenty four imagery has been applied for ecosystem monitoring or

1:42:48 Speaker 1: restoration projects, for example, chloral refmapping, seagrass monitoring, or coastal

1:42:54 Speaker 1: habitat restoration.

1:42:59 Speaker 2: Yeah, I can take that one too. There was a

1:43:02 Speaker 2: paper bile Aqualic at All. Mike Sinskay of NASA was

1:43:06 Speaker 2: also a co arthor on that. Sorry, I was trying

1:43:08 Speaker 2: to pull up the exact reference for it, but that

1:43:11 Speaker 2: was a paper specifically looking at I said tube pothymetry

1:43:16 Speaker 2: for benic habitat mapping. And then our research group is

1:43:20 Speaker 2: actually currently looking at use of INTL twenty four for

1:43:25 Speaker 2: core refrustration site monitoring in the Florida case. That's a

1:43:29 Speaker 2: collaboration with gender Extra at the University of New Hampshire.

1:43:32 Speaker 2: Center for coastalin Ocean Mapping.

1:43:37 Speaker 1: Okay, Chris, thank you so much. Question number three. If

1:43:41 Speaker 1: we are integrating I said to Atail twenty four data

1:43:44 Speaker 1: sets with multi spectral remote sensing data such as sentinels

1:43:47 Speaker 1: the imagery, what are the best practices to improve accuracy,

1:43:51 Speaker 1: Which remote sensing data sets are most frequently used for

1:43:53 Speaker 1: this purpose? And why are there any tips for pre processing, filtering,

1:43:59 Speaker 1: or modeling that can enhance the symmetry estimates.

1:44:05 Speaker 2: I can keep going because I put the first part

1:44:07 Speaker 2: of the answer in the chat, but maybe Gretchen or

1:44:09 Speaker 2: others will want to chat in here as well, so

1:44:12 Speaker 2: and I think this was covered in part in Gretchen's presentation.

1:44:15 Speaker 2: But careful selection of the input scenes, trying to find

1:44:17 Speaker 2: scenes that avoid clouds, turbidity, artifacts, compositing of several scenes

1:44:24 Speaker 2: can improve the accuracy or blessed atmospheric correction, which was

1:44:28 Speaker 2: a step that Gretchen mentioned. And then another thing that

1:44:34 Speaker 2: we recommend is performing accuracy tests and using independent reference data.

1:44:37 Speaker 2: We're available as shown and Gretchen's and Cannas demos.

1:44:47 Speaker 1: Chris, thank you again. Question number four, Will this work

1:44:51 Speaker 1: for the purpose of archaeological diving?

1:44:59 Speaker 2: I can keep going, but obviously anybody else jump in

1:45:02 Speaker 2: at any time. Archaeological marine archaeology is potentially a really

1:45:08 Speaker 2: interesting application of issatude with tymmetry and ATL twenty four,

1:45:12 Speaker 2: and it's one that I've heard mentioned in the peer

1:45:16 Speaker 2: reviewed literature. To date, I haven't seen a published study

1:45:22 Speaker 2: dedicated to marine archaeology, but I know that again, it's

1:45:25 Speaker 2: an area that people are interested in and I think

1:45:28 Speaker 2: are maybe currently actively working on. Very cool.

1:45:34 Speaker 1: I look forward to reading some of his papers when

1:45:36 Speaker 1: they're published. Question number five, Are I too ATL twenty

1:45:39 Speaker 1: four data sets used for applications like natural resource exploration

1:45:43 Speaker 1: or disaster monitoring, such as tracking sediment deposition and can

1:45:47 Speaker 1: they be applied in hydrocarbon exploration as well?

1:45:57 Speaker 2: I can take just the first part of that. Are

1:45:59 Speaker 2: the second part I guess on disaster monitoring. So a

1:46:03 Speaker 2: former student, John Herman, did an excellent study on use

1:46:09 Speaker 2: of solely satellite based methods. So I set to you

1:46:12 Speaker 2: with tymmetry and specially drive pathymetry or looking at a

1:46:16 Speaker 2: hurricane change analysis. And in Joe's study it was shown

1:46:21 Speaker 2: that we could actually see sort of at least broad

1:46:25 Speaker 2: patterns of special change, for example, moving sandbars that occurred

1:46:31 Speaker 2: to the hurricane. I think the resource exploration again, that's

1:46:36 Speaker 2: another kind of topic that I think people are interested

1:46:41 Speaker 2: in right now. It's one that I've certainly heard people mention.

1:46:44 Speaker 2: I don't have any studies to point to off the

1:46:47 Speaker 2: top of my head.

1:46:52 Speaker 1: Chris, thank you so much again. Question number six, how

1:46:55 Speaker 1: does the vertical accuracy of HL twenty four compare to

1:46:59 Speaker 1: standard single beam hydrographic surveys in clear water? Can I

1:47:03 Speaker 1: use ATL twenty four to validate or even correct my

1:47:07 Speaker 1: existing field data?

1:47:11 Speaker 2: This one was also my response, and again anybody else,

1:47:14 Speaker 2: any of the other panelists this time, and I was

1:47:16 Speaker 2: trying to answer some of the questions in the in

1:47:18 Speaker 2: the chats as other four presenting my answer to this one,

1:47:24 Speaker 2: I said, too, is generally not as accurate as a

1:47:26 Speaker 2: single beam or multi beam ecosounding. We did a study

1:47:31 Speaker 2: of ATL twenty four accuracy and that was just published

1:47:35 Speaker 2: in a paper an Earthen Space Science earlier this year.

1:47:38 Speaker 2: What we found from the ATL twenty four pothymetry we

1:47:41 Speaker 2: had eight different test sites. The RMSS of the ATL

1:47:45 Speaker 2: twenty four pythymmetry was zero point six eight or sixty

1:47:49 Speaker 2: eight centimeters using all points, and if we filtered by

1:47:52 Speaker 2: just high confidence points. It was point four to three

1:47:55 Speaker 2: meters or forty three centimeters, So I think you know,

1:47:58 Speaker 2: half meter or better than half is really quite good

1:48:01 Speaker 2: for a lot of applicussions, but still not approaching what

1:48:05 Speaker 2: you can get from sonar.

1:48:10 Speaker 1: Great Chris, thank you again. Question number seven, did you

1:48:13 Speaker 1: use Sentinel two data instead of Lance hat for this

1:48:17 Speaker 1: for the spatial resolution for the model? Does this improve

1:48:21 Speaker 1: the model?

1:48:24 Speaker 2: We've used both. We've used both Lance eight nine and

1:48:29 Speaker 2: Sentinel two both can both can be great sources of

1:48:34 Speaker 2: input data for for SDB. The Sentinel two special resolution

1:48:39 Speaker 2: is a little bit better in the in the red,

1:48:41 Speaker 2: grin and blue spectral bands, which can be helpful.

1:48:47 Speaker 1: Hey, great Chris, thank you. Question number eight, My area

1:48:50 Speaker 1: of interest is a coral reef with high morphological complexity

1:48:54 Speaker 1: such as rough slopes, seven drop offs, and narrow values.

1:48:58 Speaker 1: Since the ice that two laser footprint is roughly eleven

1:49:01 Speaker 1: to seventeen meters, does it effectively capture these sharp changes

1:49:05 Speaker 1: or does it smooth out the the athymmetry too much

1:49:08 Speaker 1: for precise coastal mapping.

1:49:12 Speaker 2: Yeah, that one again was a great question, and that's correct.

1:49:16 Speaker 2: I said two footprint is around eleven meters, and that

1:49:20 Speaker 2: can be a limiting factor and extracting a really high

1:49:22 Speaker 2: frequency spatial detail. Something that's important is that the along

1:49:28 Speaker 2: track special resolution of I said two is quite good.

1:49:32 Speaker 2: Points in the along track direction and nominal points spacing

1:49:35 Speaker 2: is about seventy centimeters, so that can help and resolving features.

1:49:39 Speaker 2: But really my main suggestion here is just investigated and

1:49:42 Speaker 2: several of these talks by you know, JP and others,

1:49:47 Speaker 2: you saw examples of using that slide real web client.

1:49:50 Speaker 2: That's a really fast and easy way to go in

1:49:52 Speaker 2: and visually grab some ATL twenty four data for your site,

1:49:56 Speaker 2: and I would just start there in terms of looking

1:49:58 Speaker 2: at how good the resolution is and if it's going

1:50:00 Speaker 2: to support your needs.

1:50:05 Speaker 1: Awesome, Chris, thank you so much. Question number nine, when

1:50:07 Speaker 1: do you change when you do the change analysis when

1:50:11 Speaker 1: Sentinel two? Do you calibrate the reflectance with ATL again

1:50:14 Speaker 1: or use the regression coefficient from the original calculation and

1:50:18 Speaker 1: apply it to all the sentinel imagery.

1:50:23 Speaker 2: My recommendation is if you're doing multi temporal analysis, so

1:50:27 Speaker 2: change detection or something like that where you've got input

1:50:29 Speaker 2: scenes from different times, I would, if at all possible,

1:50:33 Speaker 2: I would do the calibration separately for each epic, but

1:50:37 Speaker 2: if you wanted to, you could start with the calibration

1:50:39 Speaker 2: parameters for the first from the first epic and and

1:50:42 Speaker 2: just rEFInd them as needed.

1:50:47 Speaker 1: Awesome, thanks Chris. Question number ten data collection model for

1:50:51 Speaker 1: water turbidity based on size, weight, composition, settling rate, overlay

1:50:57 Speaker 1: with world currents and closest coastal impact. Question mark does

1:51:02 Speaker 1: it need to include baseline TURBIDIV levels possible to make

1:51:05 Speaker 1: moving animated data overlay over seasonal title and trade wind

1:51:10 Speaker 1: change prediction models for cumulative impacts?

1:51:15 Speaker 2: My answer here was this sounds like an amazing tool.

1:51:17 Speaker 2: To the best of my knowledge, that's not something that

1:51:19 Speaker 2: currently exists, but if there are people who want to

1:51:23 Speaker 2: take the lead on developing something like that, I can

1:51:25 Speaker 2: see it being beneficial to a lot of different people.

1:51:31 Speaker 3: Great.

1:51:31 Speaker 1: Thank you again, Chris. Question number eleven, how can tide

1:51:35 Speaker 1: be applied or moved from pseudo satellite drive asymmetry.

1:51:42 Speaker 2: I'll give my answer here and Kretschener theres may want

1:51:45 Speaker 2: to jump in, so regarding tide correction, and if you're

1:51:49 Speaker 2: doing single scene STV, so you're not compositing multiple scenes,

1:51:53 Speaker 2: and if your calibration data are already reference to a

1:51:56 Speaker 2: vertical data, then your output STB are going to be

1:51:59 Speaker 2: in cparently a reference to that same vertical datam as

1:52:03 Speaker 2: the reference data that you're using for calibration. And so

1:52:08 Speaker 2: for that reason, in that specific case, you don't need

1:52:11 Speaker 2: a TI a dedicated tide correction step. And to kind

1:52:15 Speaker 2: of explain that if you're doing if you're doing a

1:52:17 Speaker 2: linear transformation from your pseudo SDB to SDB and the

1:52:23 Speaker 2: transformation parameters come from something that's already referenced to a

1:52:25 Speaker 2: vertical datam any kind of vertical shift, for example, digit

1:52:29 Speaker 2: tides is just going to be inherently accounted for the

1:52:32 Speaker 2: time that you do need a tide correction is if

1:52:34 Speaker 2: you're using if you're sort of compositing or pulling together

1:52:37 Speaker 2: multiple stands collected at multiple different stages of tide.

1:52:44 Speaker 1: Hey, Chris, thank you. Question twelve, Since the interest is

1:52:48 Speaker 1: in the symmetry, would you want to sample satellite data

1:52:51 Speaker 1: just after ice off to avoid photoplankton that might be

1:52:54 Speaker 1: growing in the water column in the later spring and

1:52:56 Speaker 1: summer which could mask the bethymmetry.

1:53:02 Speaker 2: Quite possibly, Yes, I would monitor global turbidity or you

1:53:07 Speaker 2: know kdie four ninety data sets to try to find

1:53:10 Speaker 2: optimal times for optimal water clarity. That is generally going

1:53:15 Speaker 2: to be the driving factor and whether you get good

1:53:17 Speaker 2: with ymmetry.

1:53:18 Speaker 5: Or not.

1:53:22 Speaker 1: Gay question number thirteen. Does the stump approach only work

1:53:25 Speaker 1: with the ratio of the log of blue and green

1:53:31 Speaker 1: or could other band ratios be applied?

1:53:35 Speaker 2: Absolutely, you can do other band ratios, and I think

1:53:38 Speaker 2: this was shown in Gretchen and Cannis talks for example,

1:53:41 Speaker 2: they're SDB green and SDB red. Those refer to, respectively,

1:53:48 Speaker 2: ratios of log blue to log green and then log

1:53:51 Speaker 2: blue to log red. So those are those are both

1:53:54 Speaker 2: combinations that can be used and really any any other

1:53:59 Speaker 2: combination essentially as well, if you're working with hyperspectrol imagery

1:54:03 Speaker 2: for example.

1:54:06 Speaker 1: All right, Chris, thank you again. Question fourteen. Have you

1:54:09 Speaker 1: tried to compare your depth data or the data source

1:54:11 Speaker 1: being discussed against real world bethymmetry data products such as

1:54:14 Speaker 1: the Noah Blue topo to assess its accuracy?

1:54:20 Speaker 2: We have in that accuracy test that I think I

1:54:24 Speaker 2: put in the in the chat that was a comparison

1:54:27 Speaker 2: against Noah data. A lot of it was a lot

1:54:31 Speaker 2: of it was airborne Tobo Bethy Ladder data available on

1:54:36 Speaker 2: Noah Digital Coast, and a lot of that same data

1:54:39 Speaker 2: is available in Blue Topo. For the next release, we're

1:54:43 Speaker 2: going to try to do an automated accuracy test using

1:54:47 Speaker 2: all of the Blue Topo, all the Blues, all the

1:54:51 Speaker 2: Noah Blue TEPA data, but the development of that is

1:54:54 Speaker 2: something that's still on the works.

1:54:59 Speaker 1: Okay, thank you, Chris. Question fifteen, what quality criteria in

1:55:03 Speaker 1: ATL twenty four would you specify to filter out and

1:55:06 Speaker 1: noisy data? Since I we'll be working with an inland

1:55:08 Speaker 1: water body, can you specify a comparable quality criteria for

1:55:12 Speaker 1: ATL thirteen data?

1:55:15 Speaker 2: For ATL twenty four, we've typically recommended a confidence threshold,

1:55:21 Speaker 2: So one of the parameters is that confidence parameter It

1:55:24 Speaker 2: was discussed in Laurie mcgroder's presentation that provides a confidence

1:55:28 Speaker 2: of the symmetry classification. We've used a threshold of point

1:55:32 Speaker 2: six as a good starting point for filtering out some

1:55:35 Speaker 2: of the lower confidence data, but that's kind of site specific,

1:55:39 Speaker 2: so you might want to sort of play around with

1:55:41 Speaker 2: that and fine tune it. The second part of the

1:55:45 Speaker 2: question is correct ATL thirteen is the is the corresponding

1:55:49 Speaker 2: products for inland waters thirteen provides a lot of other parameters,

1:55:57 Speaker 2: but there is a there is a pythymmetry parameter. To

1:56:00 Speaker 2: the best of my knowledge, there is not a similar

1:56:05 Speaker 2: confidence parameter. But I would definitely encourage rating the algorithm.

1:56:09 Speaker 2: Theoretical Basis document which provides all the technical specifications for

1:56:13 Speaker 2: ANTIL thirteen data product.

1:56:18 Speaker 1: Terrific. Question number sixteen, how would you integrate the SDB

1:56:22 Speaker 1: layer with the point the themetry data?

1:56:27 Speaker 2: This may have been where Gretchen was going to step

1:56:31 Speaker 2: step into this. Gretchen or somebody else want to want

1:56:34 Speaker 2: to pick up the questions from this point or would

1:56:38 Speaker 2: you all like me to keep going?

1:56:47 Speaker 5: Allright?

1:56:48 Speaker 3: Not hearing anyone else, I'm sorry because I was muted.

1:56:51 Speaker 3: Go ahead, I was going to pick up at nineteen.

1:56:54 Speaker 2: Okay, all right, I'll I'll turn it over. You integrate

1:56:57 Speaker 2: the STB layer with the point the theymmetry.

1:57:02 Speaker 4: I'm not.

1:57:03 Speaker 2: I'm not totally sure I understand the question. But if

1:57:05 Speaker 2: it's referring to how do you actually how do you

1:57:08 Speaker 2: take your pseudobythymmetry and then integrate that with the individual

1:57:12 Speaker 2: points from ETL twenty four, that was one of the

1:57:15 Speaker 2: steps that I should how to do that in my

1:57:18 Speaker 2: demo just using QGIS and and basically it was using

1:57:22 Speaker 2: that sample raster values tool to have the pseudo SDB

1:57:28 Speaker 2: and then just at each at each raster or each

1:57:31 Speaker 2: location of ATL twenty four point to sample the corresponding

1:57:35 Speaker 2: pseudo STB value.

1:57:40 Speaker 1: Okay, Question seventeen, what about the difference in pseudo STB

1:57:44 Speaker 1: per S two detector?

1:57:48 Speaker 2: Yeah, I think that's I think that's interesting. One of

1:57:52 Speaker 2: my recent students, Art mcculluugh, I looked at actually using

1:57:57 Speaker 2: small sets where you've got you know, each each individual

1:58:01 Speaker 2: satellite has its own imaging sensor on it, so you've

1:58:04 Speaker 2: got just differences between the different sensors. I think that's

1:58:09 Speaker 2: just a I don't have detailed information other than I

1:58:13 Speaker 2: think that's a great question and most definitely an area

1:58:16 Speaker 2: for ongoing research.

1:58:21 Speaker 1: Great question number eighteen. I've collected hydrographic survey data across

1:58:25 Speaker 1: my steady area. How should I combine field based hydrographic

1:58:28 Speaker 1: data with ice AT to a long track measurements to

1:58:31 Speaker 1: create a more robust validation data set. What are the

1:58:34 Speaker 1: trade offs between them?

1:58:36 Speaker 2: My answer to this one just in terms of things

1:58:38 Speaker 2: to consider, you know, need to pay attention to different

1:58:42 Speaker 2: special resolutions and accuracies. Different datums can definitely be a

1:58:47 Speaker 2: factor when you're trying to combine data. So for data

1:58:52 Speaker 2: collected in the US knows vertical datum transformation utility be

1:58:56 Speaker 2: datum can be a useful utility for that.

1:59:03 Speaker 1: Great question number nineteen, Where can I access the satpathy tool?

1:59:09 Speaker 3: Okay, yeah, Now the satvati tool is for noah internal use.

1:59:15 Speaker 3: Is a desktop tool, but we are working on a

1:59:18 Speaker 3: larger cloud effort, public cloud effort. This will incorporate lessons

1:59:23 Speaker 3: we have learned over the years as we've worked with

1:59:26 Speaker 3: Noah's Office of cod Survey for our reconnaissance projects with

1:59:30 Speaker 3: them and our other federal mapping partners. But I think

1:59:35 Speaker 3: you gave my sean as my information on the slide,

1:59:39 Speaker 3: so folks that are interested, you can feel free to

1:59:42 Speaker 3: contact me for updates on our status.

1:59:47 Speaker 1: Great Gretchen, thank you so much. Question twenty the example

1:59:50 Speaker 1: in Alaska was fascinating. Have there been any studies looking

1:59:52 Speaker 1: at the efficacy of these approaches in polar regions in

1:59:55 Speaker 1: particular given the challenge is presented by snow and ice.

2:00:00 Speaker 3: Yeah, that's a really good question. We've found that working

2:00:03 Speaker 3: in Alaska has been very challenging and we have been

2:00:08 Speaker 3: trying to do, where possible some parallel research along with

2:00:13 Speaker 3: the SDB reconnaissance operational work for NOS office A coast survey.

2:00:20 Speaker 3: Alaska alone is challenging areas, so there's more work needed

2:00:25 Speaker 3: there as well as the other northern areas and polar regions.

2:00:32 Speaker 1: Okay, christ thank you. Last question twenty one, I missed

2:00:36 Speaker 1: the definition of SDB underscore merged and SDB on the

2:00:39 Speaker 1: score read. Could you explain.

2:00:45 Speaker 5: H questions?

2:00:46 Speaker 2: Should I answer this?

2:00:48 Speaker 3: Yeah, Brian, thanks, that'd be great.

2:00:50 Speaker 2: Great.

2:00:52 Speaker 10: So, S tob red and SB green are the two

2:00:55 Speaker 10: products that are created from the or else SDB algorithm

2:00:59 Speaker 10: or s A coach. We found the S to b

2:01:02 Speaker 10: read typically works better in shallow regions, while SMB green

2:01:07 Speaker 10: is more suited for deeper water. So s to be

2:01:12 Speaker 10: merged us as a combination of those two algorithms, relying

2:01:18 Speaker 10: on s tob red for the shallow water and then

2:01:21 Speaker 10: transitions to s tob green for the deeper areas. More

2:01:26 Speaker 10: information about how that works could be found in some

2:01:28 Speaker 10: of the publications that were listed.

2:01:35 Speaker 1: Brian, thank you and thank you to everybody that asked questions,

2:01:38 Speaker 1: and I want to thank everybody that attended today's training,

2:01:41 Speaker 1: actually both parts of the training. Today was the last

2:01:43 Speaker 1: day of this two part webinar series, so for wherever

2:01:46 Speaker 1: you were joining from, thank you so much. We also

2:01:48 Speaker 1: want to thank once again our amazing presenters, Doctor Christopher

2:01:52 Speaker 1: Parish from Oregon State University, JP Swinsky from NASA's Goddard

2:01:56 Speaker 1: Space Light Center, Gretchen Imohorami from NOAH, and also Key

2:02:00 Speaker 1: from Oregon State University. We also want to thank subject

2:02:03 Speaker 1: matter expert Brian Edter from NOAH from joining today to

2:02:06 Speaker 1: help as a subject subject matter expert answering some of

2:02:09 Speaker 1: the questions, and also want to thank Amy Neely from

2:02:13 Speaker 1: NASA Goddard Space Flight Center as well for her vision

2:02:16 Speaker 1: in having this training take place. So thank you to

2:02:18 Speaker 1: all the wonderful presenters and we look forward to seeing

2:02:20 Speaker 1: you in the future at an upcoming our set training.

2:02:23 Speaker 1: Thank you