NASA ARSET Applications in Urban Heat Island Mapping

0:02 Speaker 1: Welcome back to our rset training series Introduction to Thermal

0:07 Speaker 1: Remote Sensing and Applications in Urban Heat Island Mapping. Today

0:12 Speaker 1: is part two of our training series Applications in Urban

0:16 Speaker 1: Heat Island Mapping. My name is Savannah Cooley. I'm a

0:20 Speaker 1: researcher at NASA AMES Research Center with the Bay Area

0:23 Speaker 1: Environmental Research Institute in Mountain View, California, and I'm a

0:27 Speaker 1: trainer with the URSET Ecological Conservation Team. We have two

0:35 Speaker 1: instructors for part two, myself and our guest instructor, Glenn Holly,

0:40 Speaker 1: who is a research scientist in the Earth Science Section

0:43 Speaker 1: of NASA Jet Propulsion Laboratory. For Part two, we have

0:50 Speaker 1: two objectives. First, participants will be able to filter and

0:57 Speaker 1: visualize ecostress land surface temperature data using provided our based

1:02 Speaker 1: data processing workflows. Second, participants will be able to downscale

1:09 Speaker 1: native seventy meter ecostress LST data to a fine ten

1:14 Speaker 1: meter spatial resolution using a random forest machine learning model

1:19 Speaker 1: implemented on an interactive Google Earth Engine interface to analyze

1:24 Speaker 1: neighborhood level urban heat patterns. If you have questions during

1:30 Speaker 1: today's training, you can put them in the Q and

1:33 Speaker 1: a box within WebEx at any time. We encourage you

1:37 Speaker 1: to put your questions into the Q and A chat

1:39 Speaker 1: as they arise. The earlier you share your questions in

1:43 Speaker 1: the chat, the more likely it is that we will

1:45 Speaker 1: be able to respond to them in real time during

1:48 Speaker 1: the Q and A session. If there are questions we

1:52 Speaker 1: do not have time to address during the live Q

1:54 Speaker 1: and A session, we will make sure to collect these

1:57 Speaker 1: and post the written answers to the training page within

1:59 Speaker 1: our week after the training. In other words, we will

2:03 Speaker 1: try to address as many questions as we can today.

2:06 Speaker 1: All of these questions and answers will be posted, along

2:09 Speaker 1: with the additional questions and written answers to the training

2:12 Speaker 1: page within a week.

2:13 Speaker 3: After the training.

2:14 Speaker 1: Before we delve into our next section, let's review some

2:19 Speaker 1: of what we learned in part one. Remember, satellite sensors

2:24 Speaker 1: measure radiance and convert it to brightness temperature, which is

2:28 Speaker 1: defined as a temperature A perfect black body would have

2:32 Speaker 1: to be to emit the radiance we observe. The problem

2:38 Speaker 1: is that real surfaces aren't perfect black bodies. They have

2:41 Speaker 1: emissivities that are less than one. If we assume everything

2:46 Speaker 1: has an emissivity of one, will systematically underestimate surface temperature.

2:52 Speaker 1: Let's do a knowledge check with a numerical example. What

2:57 Speaker 1: is the radiometric temperature of a sand desert surface with

3:01 Speaker 1: amissivity equal to point nine and kinetic temperature equal to

3:05 Speaker 1: three hundred and twenty kelvin. Take a minute or two

3:11 Speaker 1: to work this out on your own before I walk

3:13 Speaker 1: through the answer. With a sandy desert surface having an

3:24 Speaker 1: emissivity of zero point nine and a kinetic temperature at

3:29 Speaker 1: three hundred and twenty Calvin, we can calculate the radiometric

3:32 Speaker 1: temperature as follows. We would raise the emissivity to the

3:43 Speaker 1: power of zero point five as shown here, multiply that

3:49 Speaker 1: by three hundred and twenty Calvin, and then the result

3:57 Speaker 1: is three hundred and eleven point seven Calvin, which corresponds

4:03 Speaker 1: to answer D here in the set of answers. This

4:10 Speaker 1: implies that there is about an e Calvin underestimate of

4:15 Speaker 1: the surface temperature if amissivity is ignored. The key takeaway

4:20 Speaker 1: is that the lower the emissivity, the larger the gap

4:23 Speaker 1: between what a sensor measures and the service is true temperature.

4:28 Speaker 1: This is why accurate amissivity is so important and why

4:32 Speaker 1: we will spend time delving into the topic in this training.

4:39 Speaker 1: With that, I will turn it over to Glynn to

4:43 Speaker 1: tell us about Ecostress land service temperature data.

4:51 Speaker 2: Okay, good morning everyone. My name is Glenn Hally.

4:55 Speaker 3: I'm a research scientist at NASA's Jack Propulsion Lab. We

4:58 Speaker 3: have spent the past twenty also specializing in thermal infrared

5:03 Speaker 3: remote sensing and also spectroscopy applications and science. Today, I'll

5:08 Speaker 3: be walking through how we sharpen also sometimes called dancescale

5:13 Speaker 3: Ecostress land surface temperature from its native seventy meter resolution

5:17 Speaker 3: dwance ten meter resolution using a random forest machine learning approach,

5:21 Speaker 3: and we'll be doing this within Google Earth Engine API.

5:26 Speaker 3: This technique opens up a whole new range of applications

5:30 Speaker 3: for therm or remote sensing, specifically in urban settings, and

5:33 Speaker 3: by the end of the session you should be able

5:35 Speaker 3: to run the pipeline by yourself and create your own

5:37 Speaker 3: high resolution maps over any city in the US using

5:42 Speaker 3: data since twenty eighteen when Ecostress first started taking observations. Okay,

5:51 Speaker 3: so I'll be first going over some introduction on Ecostress,

5:56 Speaker 3: followed by urban heat islands, and then we'll go into

5:58 Speaker 3: some of the theory of the downscaling. So Ukostress was

6:03 Speaker 3: launched in twenty eighteen. It's stocked on the Japanese Experimental

6:08 Speaker 3: Module on the International Space Station and it's been delivering

6:12 Speaker 3: film I've read observation since about August of twenty eighteen.

6:18 Speaker 3: We're currently funded through about twenty twenty nine, which is

6:21 Speaker 3: really great considering that the initial mission was only supposed

6:25 Speaker 3: to last about one year, but we've been extended due

6:29 Speaker 3: to the credible success that we've had in science and

6:32 Speaker 3: applications from the data from the instrument. The native pixel

6:37 Speaker 3: resolution size is seventy x seventy meter resolution and we

6:40 Speaker 3: acquire data on average three days over most parts of

6:45 Speaker 3: the US and also the globe. And also one unique

6:49 Speaker 3: thing about Ecostress and then we'll coverver the next few slides,

6:51 Speaker 3: is that we can take these observations over the entire

6:55 Speaker 3: diurnal cycle. We also have spires five spectral bands in

6:59 Speaker 3: the filmal infrared, which is also unique amongst current spaceborn

7:03 Speaker 3: thermal instruments. So moving on to some of the actual

7:08 Speaker 3: characteristics of the instrument, what I've shown here is an

7:11 Speaker 3: actual example snapshot of what we call a scene or

7:16 Speaker 3: a granule that we observe from the space station. Each

7:20 Speaker 3: of these snapshots. Think of a camera snapshot is around

7:22 Speaker 3: four hundred by four hundred kilometers wide, so really wide snapshot,

7:26 Speaker 3: and within that we have seventy meter resolution pixels, which

7:29 Speaker 3: is about the size of a small football field. And

7:33 Speaker 3: this is currently the highest resolution thermal data from space

7:37 Speaker 3: now with the precessing orbit of the ISS meaning it

7:40 Speaker 3: doesn't come over the same place on Earth every single

7:43 Speaker 3: day like Lancet and ASTA for example. This image shows

7:48 Speaker 3: the number of observations we get over a three month

7:51 Speaker 3: period over the dinal cycle, which is shown in UTC

7:54 Speaker 3: time on the left y axis. Yet so you can

7:58 Speaker 3: see that echo Stress in blue captures a lot more

8:01 Speaker 3: observations than both lances at eight and nine and also

8:04 Speaker 3: throughout the dininal cycle, whereas Lansad only captures at a

8:07 Speaker 3: single time around the early morning overpass every day. The

8:13 Speaker 3: five thermal bands, which you've probably learned about in Savannah's

8:17 Speaker 3: tutorials are located between eight and twelve micron range. This

8:21 Speaker 3: is the classic long wave thermal infrared region where we

8:24 Speaker 3: can retrieve the surface temperature.

8:26 Speaker 2: From and these five bands.

8:29 Speaker 3: The more typically the more bands we have in this region,

8:31 Speaker 3: the more accurately we can derive the surface temperature and

8:34 Speaker 3: also the spectral emissivity from the thermal infrared measurement. So

8:40 Speaker 3: just some quick latest news from Ecostress. Currently, the Collection

8:44 Speaker 3: two products are available from NASA, but we are delivering

8:48 Speaker 3: Collection three science products, specifically the Level two Land Service

8:53 Speaker 3: Temperature product as an imminent release of the alp DC,

8:57 Speaker 3: and there's a number of improvements in that product products.

9:00 Speaker 2: Specifically, we have a new sea surface temperature.

9:02 Speaker 3: Algorithm and a lot of improvements to the data quality

9:05 Speaker 3: and of the cloud masking. Also very exciting for us

9:09 Speaker 3: this year is that, and I'll be showing you that.

9:12 Speaker 3: Of course, the demo in Google Earth Engine is that

9:14 Speaker 3: the entire Ecostress Land Surface Temperature archive is now being

9:19 Speaker 3: ingested into Google Earth Engine over the entire continental US.

9:22 Speaker 3: So this is fantastic news for us and opens up

9:25 Speaker 3: a lot of data and science and applications that we

9:28 Speaker 3: can use with other data within Google Earth Engine. We

9:31 Speaker 3: have more than one thousand per review publications so far,

9:35 Speaker 3: and since earlier this year, we've acquired more than six

9:38 Speaker 3: hundred and six thousand of these four hundred buffour hundred

9:41 Speaker 3: kilometers snapshots I showed you earlier. We also have a

9:44 Speaker 3: new applications page referenced here. You can see that we

9:49 Speaker 3: support a diverse discipline of different science applications. The one

9:56 Speaker 3: of course we'll be focusing on today is urban heat

9:58 Speaker 3: and urban resilience. So what makes ecostress unique. As I

10:03 Speaker 3: mentioned earlier, most thermal sensors in space have a sunsynchronous orbit,

10:09 Speaker 3: meaning that they cross overhead in the mid morning hours,

10:12 Speaker 3: in this case for LANSAT and Assets around ten or

10:14 Speaker 3: eleven AM.

10:16 Speaker 2: And you can see in this Los Angeles temperature.

10:18 Speaker 3: Profile that I've shown that the morning overpass, which is

10:22 Speaker 3: this white line going across the screen, typically misses the

10:25 Speaker 3: warmest times of the day and the hottest times of

10:27 Speaker 3: day in the peak afternoon hours. And this is critical,

10:31 Speaker 3: critical for urban heat mapping because you will miss the

10:34 Speaker 3: maximum temperatures during a typical urban heat urban heat extreme event.

10:41 Speaker 3: If we look at ecostress, we sample across the entire

10:45 Speaker 3: full dimnal cycle and seasonal cycle, which is exactly what

10:49 Speaker 3: makes it uniquely valuable for heat stress and urban applications.

10:56 Speaker 3: So then moving on to urban heat items, just I'll

10:59 Speaker 3: give some background and ourb and here is what we've

11:01 Speaker 3: done so far with ecostress and then we'll move on

11:03 Speaker 3: to the fun part, which is the downscaling.

11:08 Speaker 2: So why does this matter to us?

11:10 Speaker 3: Well, heat waves, a lot of people don't realize this,

11:13 Speaker 3: are the single biggest weather related killer in the United States.

11:16 Speaker 3: They more than double that of floods, torninators, and hurricanes combined.

11:21 Speaker 3: Even so, as you can see on the chart on

11:23 Speaker 3: the left, this is the No and National has its

11:25 Speaker 3: statistics chart. Heat fatalities are running by roughly two hundred

11:30 Speaker 3: deaths per year over the last decade, and a lot

11:33 Speaker 3: of people don't realize this because you can call them

11:35 Speaker 3: the silent killers. They don't cause as much physical destruction

11:39 Speaker 3: as other natural disasters, and so you don't care about

11:43 Speaker 3: them as much. And the downstream impacts go well beyond

11:47 Speaker 3: just mortality. For example, there's increased risk of heat stroke

11:51 Speaker 3: or hypothermia. These are the most common era visits and

11:54 Speaker 3: a lot of these are typically make a lot of sense.

11:59 Speaker 3: Increased electrical demand from the use of AC, higher wildfire risk,

12:03 Speaker 3: especially in the southern California region in the later parts

12:08 Speaker 3: of the year when we have Santa Ana events, and

12:10 Speaker 3: then of course degraded eir quality due to the increase

12:13 Speaker 3: in neo surface ozone, and these health effects also exacerbate,

12:20 Speaker 3: for example, other underlying conditions like cardiovascular and also mental

12:24 Speaker 3: health and cognitive function.

12:28 Speaker 2: Over the last few years we've spent a lot of

12:30 Speaker 2: time with a lot of very.

12:31 Speaker 3: Enthusiastic summer students at JPL where we've now mapped most

12:36 Speaker 3: of the extreme heat events with eucostress and cities across

12:40 Speaker 3: the world. So what you've see here are snapshots from

12:43 Speaker 3: these images in European and US cities from Prague to

12:46 Speaker 3: Loose all the way to Birmingham, Milan and Sharp and

12:50 Speaker 3: ecostress data have been featured in both BBC articles and

12:55 Speaker 3: also the European Space Agency typically features an article every

12:58 Speaker 3: year on how cities can adapt to heat waves, particularly

13:01 Speaker 3: during the nighttime, and it's also increasingly increasing by city planners,

13:07 Speaker 3: public health teams.

13:08 Speaker 2: Worldwide and a lot of the media within the US.

13:14 Speaker 3: So I'll go through some examples and why ecostress is

13:18 Speaker 3: unique in its ability to capture urban heat islands and

13:21 Speaker 3: urban heat in general. So this is a case study

13:24 Speaker 3: over the what it is called the Cherbis heat wave

13:26 Speaker 3: in July twenty twenty three broke wall kinds of records

13:29 Speaker 3: across Europe in terms of heat. This is over Athens

13:32 Speaker 3: in Greece. On the left, we have a daytime ecostress

13:36 Speaker 3: scene and the rider is a nighttime ecostress scene at

13:39 Speaker 3: eleven PM. And so in the daytime scene you can

13:43 Speaker 3: see an actual urban daytime cool island effect which most

13:47 Speaker 3: people don't realize is that during the daytime in certain cities,

13:51 Speaker 3: the urban area can actually be cool. Then the surrounding

13:54 Speaker 3: agricultural area that is driven by very hot, bare rural

13:57 Speaker 3: soils during the summertime we have generally less healthy vegetation.

14:04 Speaker 3: But at nighttime that picture flips. So the city retains

14:07 Speaker 3: its heat due to a high thermal heat capacity during

14:10 Speaker 3: the day, and then that stays significantly warmer than the

14:13 Speaker 3: surroundings at nighttime. And it's that nighttime urban heat island

14:17 Speaker 3: that is really drives up heat related illness in cities

14:21 Speaker 3: around the world. And it's only really eco stress that

14:24 Speaker 3: can capture these type of dynamics and changes in the

14:26 Speaker 3: urban heat island.

14:29 Speaker 2: This is another fantastic example from Macy Runkel at Chapman

14:34 Speaker 2: University who was with us last year looking at how

14:37 Speaker 2: urban heat island intensity changes of a full twenty four

14:40 Speaker 2: I period in Delhi India in this case, So each

14:44 Speaker 2: circle here represents the UHI, which is the difference in

14:48 Speaker 2: the rural and urban temperature every two hours throughout a

14:52 Speaker 2: typical summer cycle in Delhi. So you can see that

14:56 Speaker 2: the urban heat island is at maximum about ten pm

15:01 Speaker 2: at night, and then throughout the early morning hours it

15:04 Speaker 2: slowly diminishes as we go clockwise around the images here,

15:08 Speaker 2: and then during the daytime you can see that that

15:10 Speaker 2: flips around ten or twelve pm, that flips to a

15:14 Speaker 2: cool a cooling urban heat island effect, so the city

15:18 Speaker 2: is actually cooler than the surrounding dry urban rural landscape.

15:25 Speaker 2: Another real striking example here that got a ton of

15:28 Speaker 2: news media attention back in twenty twenty four was led

15:30 Speaker 2: by Ashley Agatet, also from Chapman and An a summer

15:33 Speaker 2: student of ours in twenty twenty four.

15:36 Speaker 3: Actually in twenty twenty three as well, there was suddenly

15:38 Speaker 3: a very sharp spike in hospitalizations from people in the

15:43 Speaker 3: Phoenix area from severe contact burns, and these were mostly

15:47 Speaker 3: due to several extreme heat events that occurred over those

15:51 Speaker 3: two years and actually resulted in almost fifty fatalities. For example,

15:55 Speaker 3: the gentleman here on the left actually lost his leg

15:59 Speaker 3: when he fell on the hot sidewalk walking out of

16:02 Speaker 3: a grocery store and wasn't able to get up because

16:04 Speaker 3: every time he put his hands on the sidewalk it'd

16:07 Speaker 3: be so hot that he wouldn't be able to push

16:09 Speaker 3: himself up, and he actually lost a leg from third

16:12 Speaker 3: degree burns as a result. So what we did here

16:17 Speaker 3: was we mapped second the second degree burn risk by

16:20 Speaker 3: downscaling ecostress temperatures down to ten meters resolution over the

16:24 Speaker 3: city and extracted it over the sidewalks and paved areas,

16:28 Speaker 3: and then we correlated that with the time and would

16:30 Speaker 3: take to get a second degree burn if someone came

16:33 Speaker 3: into contact with that surface. So as an example here

16:37 Speaker 3: you can see over most of the Phoenix area, specifically

16:40 Speaker 3: in the poorer regions in Maryvale, for example, where the

16:44 Speaker 3: surfaces succeed one hundred and fifty degrees fahrenheit. That's almost

16:48 Speaker 3: hot enough to cook an egg on the sidewalk. By

16:51 Speaker 3: the way, it would take less than three seconds to

16:53 Speaker 3: get a second reburn risk if you fell on the

16:55 Speaker 3: sidewalk or the paved area. And this work was what's

17:02 Speaker 3: really popular and was featured in Scientific American and La Times,

17:05 Speaker 3: New York Times. It even made the Nature Image of

17:08 Speaker 3: the Month for twenty twenty four and August I believe,

17:12 Speaker 3: And so it's this kind of public health communication that

17:16 Speaker 3: thermal remote sensing can directly support, especially for city health officials.

17:23 Speaker 3: So in the next part, we're gonna Savannah is going

17:27 Speaker 3: to lead you through a hands on exercise with land

17:29 Speaker 3: service temperature in LA County where she's going to step

17:32 Speaker 3: through how to visualize ecostress LST using a number of

17:36 Speaker 3: different software. We're going to use the peers to download

17:39 Speaker 3: the data and then visualize it with QJS. And also are.

17:48 Speaker 1: Thank you Glynn for that excellent foundation on urban heat,

17:51 Speaker 1: islands and ecostress land service temperature. We are now transitioning

17:57 Speaker 1: to a hands on exercise exploring nighttime land surface temperature

18:01 Speaker 1: in Los Angeles County. Nighttime LST is a metric that

18:06 Speaker 1: is often just as critical for public health as daytime

18:10 Speaker 1: peak temperatures, because sustained nighttime heat prevents physiological recovery from

18:17 Speaker 1: heat stress. By focusing on the nighttime, we can accurately

18:23 Speaker 1: quantify the thermal gap created as impervious urban surfaces slowly

18:28 Speaker 1: reradiate trapped heat while natural and irrigated spaces provide cooling

18:35 Speaker 1: through vappa transporation. In this demo, we will use R

18:41 Speaker 1: to define and determine if these natural cooling benefits actually

18:47 Speaker 1: persist as a heat wave intensifies, allowing us to identify

18:51 Speaker 1: the most effective locations for thermal interventions like tree canopy investments.

18:59 Speaker 1: The two objectives for this demo are to access, filter,

19:03 Speaker 1: and visualize ecostress LST data using appears, QGAS and R,

19:10 Speaker 1: and also to interpret LST characteristics for assessing urban thermal resilience.

19:18 Speaker 1: The activities for this demo include the following. There are

19:23 Speaker 1: some preparation steps, including cloning the GitHub repository or just

19:30 Speaker 1: simply downloading the R script. Second would be to make

19:35 Speaker 1: sure that you have R and qgis installed if they're

19:39 Speaker 1: not already, and then installing any required libraries that the

19:44 Speaker 1: R script lists. The script processes ecostress land surface temperature

19:55 Speaker 1: products by converting Calvin to celsius and applying mandatory QA

19:59 Speaker 1: and cloud masks to ensure only reliable surface temperature pixels

20:04 Speaker 1: are analyzed. The script then defines and isolates specific geographic

20:10 Speaker 1: boundaries for the Chatsworth Nature Preserve and an adjacent commercial

20:14 Speaker 1: area as an example, which allows for a direct comparison

20:19 Speaker 1: of different land covers. In particular, we will evaluate nighttime

20:26 Speaker 1: cooling dynamics by analyzing the differences in heat release between

20:31 Speaker 1: natural open spaces and impervious urban services during the September

20:38 Speaker 1: twenty twenty four Los Angeles heat wave. We will assess

20:44 Speaker 1: thermal resilience by determining if the cooling benefits of urban

20:48 Speaker 1: green spaces hold or erode as heat accumulates over multiple

20:52 Speaker 1: days of extreme temperatures. By extracting individual seventy meter pixel

20:58 Speaker 1: values into a tidy data format, the script generates mean

21:02 Speaker 1: temperatures and standard deviations to quantify the thermal behavior of

21:06 Speaker 1: each site. Finally, the analysis will produce a suite of

21:13 Speaker 1: sophisticated plots, including side by side LST maps, violin box

21:18 Speaker 1: plot distributions, and a thermal gap bar charts to visualize

21:24 Speaker 1: how the temperature difference between green and gray spaces evolves

21:28 Speaker 1: over a twenty three hour window. One thing I'd like

21:32 Speaker 1: to flag is that the nature preserve we focus on

21:36 Speaker 1: for this example comparison with the urban commercial area is

21:40 Speaker 1: in large part possible due to irrigation. So you will

21:45 Speaker 1: notice that in many of the unbuilt lands surrounding the

21:50 Speaker 1: urban area in the hills, for example, outside of Chatsworth,

21:56 Speaker 1: the temperatures aren't much lower than what we see in

21:59 Speaker 1: the urban air areas. This is because although there is

22:03 Speaker 1: vegetation in those areas, there is no irrigation, so the

22:09 Speaker 1: evaporative cooling capacity is much lower relative to the managed

22:14 Speaker 1: parks and other gardens that are well irrigated. So water

22:20 Speaker 1: helps plants stay cool, and if water is limited during

22:24 Speaker 1: a heat wave, plants will close their stamata to conserve

22:27 Speaker 1: water at the cost of having a higher temperature. I

22:32 Speaker 1: won't discuss this in the demo itself, so I wanted

22:35 Speaker 1: to just state that now as something to consider when

22:38 Speaker 1: we think about temperature impacts of green spaces. With all

22:44 Speaker 1: that said, let's start the demo. One thing I'd like

22:54 Speaker 1: to note is that these data from ecostress we're obtained

23:01 Speaker 1: through an appears request, and Appears is the application for

23:06 Speaker 1: extracting and exploring analysis ready samples. It's a tool that

23:11 Speaker 1: has a number of different data sets where through the

23:17 Speaker 1: extract feature you can get specific points or areas that

23:22 Speaker 1: are also able to be bounded, both specially and temporally.

23:25 Speaker 1: So I'm going to go ahead and log in here

23:28 Speaker 1: and just show briefly the request process. There is a

23:35 Speaker 1: more in depth instructional video for accessing ecostress data and

23:40 Speaker 1: other data sets with Appeers that will be linked to

23:43 Speaker 1: in this training resources page for the purposes of demonstrating here,

23:49 Speaker 1: I'm just going to show how we submit an appears request.

24:00 Speaker 1: So I'm going to give this request a title, and

24:05 Speaker 1: then i have an existing area of interest in Los

24:09 Speaker 1: Angeles County in California that will be the focus of

24:15 Speaker 1: this request, this data request, So I'm going to go

24:21 Speaker 1: ahead and upload that file here. I'm adding here a

24:34 Speaker 1: doojson file that is defining the area of interest over

24:38 Speaker 1: LA County for which the data will were requested for.

24:44 Speaker 1: And this demo will focus on one of the hottest

24:49 Speaker 1: summers on record, which was the summer of twenty twenty fourth.

24:55 Speaker 1: This in early September, and so the start date is

25:03 Speaker 1: going to be September third, twenty four and the end

25:16 Speaker 1: date will be September ninth. The product we're interested in

25:29 Speaker 1: is Ecostress land service temperature. The service temperature, this is

25:43 Speaker 1: the one that we will po us on. We will

25:46 Speaker 1: also want the quality Control data Layer QC file as

25:53 Speaker 1: well as the cloud mask. So those are the three

25:59 Speaker 1: data sets that will be requesting here and the file

26:04 Speaker 1: output file format will be GeoTIFF. Optionally we can use calendar,

26:10 Speaker 1: date and time for output file names, and then I'm

26:18 Speaker 1: just going to suggest include geographic, which is the WGS

26:24 Speaker 1: eighty four EPSG four three two six coordinate reference system.

26:31 Speaker 1: With that, we'll go ahead and submit the request. Appears requests,

26:38 Speaker 1: depending on the size of the requests and the amount

26:42 Speaker 1: of data that are being requested, can take some time

26:47 Speaker 1: to process, from hours to even a couple of days,

26:52 Speaker 1: and so to speed things along for the purposes of

26:55 Speaker 1: this demo, I already have an existing request that I

27:00 Speaker 1: submitted ahead of this demo and have downloaded that data

27:05 Speaker 1: from the email that is sent automatically. When the Appears

27:11 Speaker 1: request is ready to be viewed and downloaded the way

27:16 Speaker 1: that I will look for these data. If you're already

27:21 Speaker 1: logged in, you can click explore. So because this was

27:26 Speaker 1: a small area and number of files, the Appears was

27:33 Speaker 1: able to process it relatively quickly, and so we have

27:36 Speaker 1: the results of this request already to download. Here are

27:44 Speaker 1: the files that are included, where we will just go

27:52 Speaker 1: ahead and download all of them, and this may take

28:06 Speaker 1: a little while. And notice that for the purposes of

28:10 Speaker 1: this demo, I'm only using six out of the twenty

28:14 Speaker 1: two files here, which ended up being the ones that

28:17 Speaker 1: had the fewest number of clouds and kind of were

28:20 Speaker 1: the covered the most area in the study area. So

28:25 Speaker 1: again it's okay to just use the data directly from Zenodo,

28:30 Speaker 1: but I just wanted to illustrate how those files were obtained.

28:44 Speaker 1: And the download is now complete, So this is how

28:50 Speaker 1: the data were obtained from appears. Now we will go

28:58 Speaker 1: to QJA yes and inspect the data visually. Let's start

29:08 Speaker 1: by adding a basemap. I'm using the hcmgis lug in

29:16 Speaker 1: to ring in a satellite base map. Let's zoom into

29:28 Speaker 1: southern California and add our data. For now, we will

29:43 Speaker 1: focus just on the land surface temperature data. We'll do

29:50 Speaker 1: the filtering with the quality control flags and the cloud

29:55 Speaker 1: mask later on. Let's go ahead and visualize symbology that

30:10 Speaker 1: shows a color ramp. Notice that the values here are

30:16 Speaker 1: in kelvin, so we have yet to convert these into celsius,

30:22 Speaker 1: which we'll do later in our but for now, just

30:25 Speaker 1: to get a sense of where the hottest and coldest

30:28 Speaker 1: temperatures are. We can go ahead and apply this color ramp.

30:35 Speaker 1: Feel free to change if you have a different ramp

30:37 Speaker 1: that you'd like to look at. There's lots of different

30:39 Speaker 1: options here, and we'll do the same for this second

30:48 Speaker 1: day that we'll be looking at, which is September fifth.

31:02 Speaker 1: So notice that these two ecostress acquisitions within these two

31:08 Speaker 1: different days have some amount of spatial overlap, but are

31:12 Speaker 1: not fully spatially overlapped, and correspond to almost twenty four

31:20 Speaker 1: hours difference between these two acquisitions. So just by visually inspecting,

31:30 Speaker 1: we can see that this area of Los Angeles County

31:34 Speaker 1: has a mix of impervious services kind of built up

31:40 Speaker 1: areas as well as more vegetated areas and green spaces.

31:49 Speaker 1: And just through visual inspection, we can see differences in

31:54 Speaker 1: temperature in these urban areas where there are larger green spaces.

32:03 Speaker 1: So let's start by looking at the Chatsworth neighborhood in

32:11 Speaker 1: northwestern corner of LA County, about thirty miles from downtown,

32:19 Speaker 1: and it will zoom in here. What you're looking at

32:28 Speaker 1: is the Chatsworth Nature Preserve on the left, and then

32:34 Speaker 1: an adjacent commercial area which is immediately east of the

32:41 Speaker 1: Nature Preserve. These two areas sit almost side by side,

32:51 Speaker 1: so very similar weather and both experiencing the same heat wave.

32:57 Speaker 1: The question we're asking today is what other the surface

33:00 Speaker 1: underneath makes a measurable difference in nighttime temperature, and whether

33:04 Speaker 1: that difference holds up after days of sustained extreme heat.

33:10 Speaker 1: So after the first night of the heat wave on

33:14 Speaker 1: September third, Ecostress captured this image. And note that the

33:21 Speaker 1: naming convention here is in UTC, which in our case

33:27 Speaker 1: we're looking at September third, at about eleven pm Local times,

33:32 Speaker 1: a Pacific time. Even just at a glance, you can

33:36 Speaker 1: see cooler tones in over this nature preserve compared to

33:42 Speaker 1: the surrounding area, with the built areas showing warmer temperatures

33:50 Speaker 1: compared to the nature preserve. And then we can also

33:56 Speaker 1: be curious about the the second overpass, which is captured

34:03 Speaker 1: almost exactly twenty four hours later on September fourth, around

34:07 Speaker 1: ten thirty pm local time. So let's go ahead and

34:14 Speaker 1: check that and just see Note that there are some

34:17 Speaker 1: differences in temperature patterns here, but the overall kind of

34:24 Speaker 1: takeaway is that there seems to be a consistent difference

34:28 Speaker 1: even after another full day of the heat wave, where

34:32 Speaker 1: there are lower temperatures in this nature preserve compared to

34:36 Speaker 1: the surrounding built areas, and by this point La had

34:45 Speaker 1: baked through a full day of extreme heat. Burbank hit

34:49 Speaker 1: one hundred and eight degrees fahrenheit that afternoon, so these

34:52 Speaker 1: surfaces had been absorbing solar radiation for fourteen hours before

34:56 Speaker 1: the satellite passed Overhead said, the general spatial patterns do

35:02 Speaker 1: look similar. The preserve still reads cooler than the commercial area,

35:07 Speaker 1: but is the magnitude of that difference the same. Has

35:09 Speaker 1: the heat accumulated enough that even the preserve can't recover overnight?

35:14 Speaker 1: That's what we're going to quantify in our So first

35:19 Speaker 1: we will navigate to the GitHub repository for this training,

35:26 Speaker 1: where there is a read me file detailing an overview

35:33 Speaker 1: the prerequisites, data requirements, downloading process, and a few notes

35:41 Speaker 1: on setup. We will go ahead and download the r

35:47 Speaker 1: script here. Go ahead and save this file to a

35:54 Speaker 1: location on your local computer and open the file in

36:07 Speaker 1: our studio. Here we have the r script loaded. A

36:17 Speaker 1: written description of the focus of the script is here,

36:23 Speaker 1: which we also discussed in the slides leading up to this,

36:30 Speaker 1: so I'm gonna let participants review this in your own time.

36:38 Speaker 1: The first section here, Section one is to load all

36:42 Speaker 1: of the libraries. Go ahead and load them. If there

36:47 Speaker 1: are any libraries here that you don't already have installed,

36:49 Speaker 1: make sure to install them. And now we will go

36:57 Speaker 1: ahead and obtain the data for this demo, which is

37:03 Speaker 1: stored on Zenodo. So make sure to update this for

37:13 Speaker 1: repoder variable, which is where the data will be downloaded.

37:19 Speaker 1: Here I've made a folder twenty twenty six thermal rs

37:24 Speaker 1: Urban Heat Demo, so that's where I am keeping the

37:27 Speaker 1: data for this demonstration. If you're following along, go ahead

37:31 Speaker 1: and update this variable to wherever you'd like to store

37:34 Speaker 1: the data. This is the link to the data set

37:43 Speaker 1: which has all of the ecostress files. And then we

37:48 Speaker 1: will go ahead and in this portion of the script,

37:53 Speaker 1: it downloads and unzips the data as long as it's

37:55 Speaker 1: not already there in the first place, and I'll call

37:59 Speaker 1: you a t to the two files that were focused on.

38:05 Speaker 1: One is like I said, the first night September third,

38:09 Speaker 1: around eleven fourteen pm Pacific time, and then we have

38:15 Speaker 1: night two, which is September fourth, at ten twenty six

38:18 Speaker 1: pm Pacific time. So let's go ahead and run this

38:26 Speaker 1: section of the script. All right, that ran successfully. Section

38:38 Speaker 1: three focuses on is the processing function for processing ecostress data.

38:46 Speaker 1: One important note I want to flag here because it's

38:49 Speaker 1: common point of confusion and ecostress workflows. If you look

38:53 Speaker 1: at the level two algorithm theoretical basis document known as

38:58 Speaker 1: the ATBD or older tutorials, you'll see a point zero

39:03 Speaker 1: two scale factor applied before the kelvin to celsius conversion

39:09 Speaker 1: of the data. And that scale factor is correct, but

39:13 Speaker 1: it only applies for the native HDF five swath product,

39:17 Speaker 1: where integer storage requires that rescaling step. What we're working

39:23 Speaker 1: with here are the geotifh tiled data, which are analysis

39:29 Speaker 1: ready products from a PEERS where the scale factor has

39:33 Speaker 1: already been applied and values are stored as float thirty

39:37 Speaker 1: two in kelvin during a PEERS processing. So we will

39:42 Speaker 1: skip the point zero two multip multiplication entirely and go

39:46 Speaker 1: straight to subtracting two hundred and seventy three point one

39:49 Speaker 1: point five, which is what we do to convert from

39:55 Speaker 1: kelvin to degrees celsius. And this is the line that

40:02 Speaker 1: applies the quality control flag, where we're only extracting bits

40:08 Speaker 1: zero and one, and so this is the other thing

40:15 Speaker 1: that I wanted to highlights, just the QC masking, where

40:17 Speaker 1: the QC layer is a bit packed integer where multiple

40:22 Speaker 1: quality flags are encoded in different bit positions within the

40:26 Speaker 1: same value. So bits zero and one carry the mandatory

40:31 Speaker 1: QA flag, where zero means high quality and one means

40:36 Speaker 1: good quality. Bits two and three mean poor or cloudy.

40:44 Speaker 1: If we just wrote QC greater than one as a

40:48 Speaker 1: raw integer, comparison, higher order bits unrelated to quality could

40:52 Speaker 1: flip us into masking pixels that are actually fine. So

40:56 Speaker 1: the modulo four operation here extracts only the two least

41:03 Speaker 1: significant bits, which is the standard bit manipulation approach for

41:08 Speaker 1: this kind of packed flag field. And finally, the cloud mask.

41:18 Speaker 1: The L two user guide was updated in twenty twenty

41:22 Speaker 1: four to explicitly note that in collection two, cloud information

41:28 Speaker 1: is not propagated into the QC bitfield, So that means

41:33 Speaker 1: that QC filtering alone is not sufficient to filter out

41:37 Speaker 1: all clouds, so you must apply the separate cloud mask layer.

41:41 Speaker 1: Our function here does both in sequence. So I will

41:46 Speaker 1: now run section three of the code. Night one comes

42:02 Speaker 1: back at roughly thirteen to twenty nine degrees celsius, and

42:10 Speaker 1: night two goes from about twelve to thirty five degrees celsius.

42:18 Speaker 1: That upper bound climbing from twenty nine to thirty five

42:22 Speaker 1: already tells a story. After a full day of heat loading,

42:25 Speaker 1: the hottest urban surfaces are retaining substantially more heat by

42:29 Speaker 1: the second night. A quick note I'll say on the

42:37 Speaker 1: coordinate Reference system situation. So the file name contains eleven N,

42:43 Speaker 1: which refers to UTM zone eleven North, but the JUDITI

42:48 Speaker 1: files themselves are stored in geographic coordinates WGS eighty four

42:52 Speaker 1: longitude and latitude. The eleven N is part of the

42:56 Speaker 1: Ecostress tiling grid naming convention, not the store projection. So

43:02 Speaker 1: to avoid any CRS resolution issues, we keep everything in

43:06 Speaker 1: WGS eighty four, which is EPSG four three two six,

43:12 Speaker 1: and we define our aois directly in decimal degrees. So

43:17 Speaker 1: this is our default AOI that we're looking into. If

43:21 Speaker 1: you remember what we saw the map on qgis we

43:25 Speaker 1: have the latitude and longitude bounding box coordinates for the

43:31 Speaker 1: Nature Preserve area in Chatsworth, as well as a coordinates

43:36 Speaker 1: box around the extent of the commercial area that's adjacent

43:43 Speaker 1: to this nature preserve. So for the homework you all

43:55 Speaker 1: will be looking at this other area of interest to

43:59 Speaker 1: do a different analysis in a different part of La County.

44:03 Speaker 1: But for today, we're going to go ahead and so

44:06 Speaker 1: you can see this part is commented out, So I'm

44:10 Speaker 1: going to go ahead and run this section for here.

44:20 Speaker 1: Section five will produce a side by side map comparison

44:25 Speaker 1: where we can see these two Ecostress acquisition dates side

44:30 Speaker 1: by side with the two areas of interests highlighted. So

44:39 Speaker 1: I'm going to go ahead and run section five, where

44:42 Speaker 1: you can see we're using the gg plot package for visualization.

44:59 Speaker 1: You can see that. But these two maps are helping

45:05 Speaker 1: us visualize the data using an identical color scale, which

45:10 Speaker 1: is intentional and important because when you're comparing two images temporally,

45:16 Speaker 1: you need the colors to mean the same temperatures in

45:18 Speaker 1: both panels. If each panel were auto stretched to its

45:23 Speaker 1: own range warmer night two pixels would look the same

45:26 Speaker 1: color as night one pixels, and you'd lose the ability

45:29 Speaker 1: to make a visual comparison. So the green and orange

45:34 Speaker 1: outlines mark our two study areas. Even from the map alone,

45:40 Speaker 1: you can see the Preserve tends toward cooler end of

45:43 Speaker 1: the palette in both nights. What we can't see clearly

45:47 Speaker 1: from the map is the magnitude of this gap, or

45:51 Speaker 1: whether it changed between nights. That's what the next two

45:55 Speaker 1: figures will quantify. So in section six here of the code,

46:06 Speaker 1: we are pulling every valid seventy meter pixel that falls

46:12 Speaker 1: within each bounding box for each night. Each row in

46:17 Speaker 1: the resulting table is going to be is one pixel,

46:20 Speaker 1: so one actual satellite measurement, and this gives us the

46:25 Speaker 1: full distribution, not just a single summary statistic, which matters

46:29 Speaker 1: because urban services are spatially heterogeneous, a parking lot and

46:33 Speaker 1: a rooftop can behave quite differently even within the same

46:36 Speaker 1: bounding box area. So let's go ahead and run section

46:42 Speaker 1: six of this script. This provides a summary statistic table

46:55 Speaker 1: stored in the variable DF summary, where now we can

47:00 Speaker 1: quantitatively look at the mean land surface temperature in degrees celsius,

47:08 Speaker 1: along with the standard deviation and the number of seventy

47:11 Speaker 1: meter pixels included in each of these areas. So we

47:18 Speaker 1: can see from this numerically that the Nature Preserve on

47:23 Speaker 1: night one was cooler on average than it was on

47:29 Speaker 1: night two. Which is expected. The standard deviation remains pretty

47:37 Speaker 1: close on both nights, although slightly lower on night two,

47:40 Speaker 1: which is interesting. And then similarly in the commercial area

47:44 Speaker 1: we see an increase in temperature from twenty three point

47:49 Speaker 1: seven degrees celsius to twenty six point six degrees celsius

47:53 Speaker 1: on night two in the commercial area. But now let's

47:59 Speaker 1: look at this from a full pixel level distribution for

48:07 Speaker 1: each AOI, so that brings us into section seven here

48:14 Speaker 1: of this script, which is which will show a violin

48:18 Speaker 1: plot with raw pixels of the data values underneath. Let's

48:25 Speaker 1: go ahead and run this so the points you see

48:31 Speaker 1: are the individual seventy meter pixels and the violin shows

48:35 Speaker 1: the overall shape of the distribution, where the box shows

48:41 Speaker 1: the intercortile range and the raw points show you exactly

48:46 Speaker 1: where the data actually are. A few things to notice here. First,

48:52 Speaker 1: the preserve distribution is consistently cooler than the commercial area

48:57 Speaker 1: on both nights, which we saw in summary table as well.

49:02 Speaker 1: That gap is real and visible. Another thing to consider

49:09 Speaker 1: is the key comparison you're thinking about. Does the gap

49:13 Speaker 1: between the two aois change or areas of interest change

49:19 Speaker 1: from night one to night two? Hold that question while

49:22 Speaker 1: we look at the next figure here. So in section

49:34 Speaker 1: eight of the code, our core question is, after a

49:36 Speaker 1: full day of extreme heat on September fourth, do we

49:39 Speaker 1: still see the same thermal gap between the preserve and

49:42 Speaker 1: the commercial area on night two? So we'll calculate the

49:48 Speaker 1: mean LST per area of interest per night and plot

49:52 Speaker 1: the absolute temperature for each area of interest across both nights,

49:58 Speaker 1: which shows whether each land cover temperature warm from night

50:01 Speaker 1: one to night two, and then also the thermal gap,

50:04 Speaker 1: which is simply the subtraction of the temperature of the

50:08 Speaker 1: commercial area minus the temperature of the preserve on each night.

50:13 Speaker 1: If this gap shrinks, the cooling benefit of the preserve

50:16 Speaker 1: is being reduced by accumulated heat stress. But if it

50:21 Speaker 1: persists or grows, the preserve is maintaining its function even

50:25 Speaker 1: under sustained extreme conditions. So again we're using the library

50:36 Speaker 1: to make these plots, and I'm just going to go

50:38 Speaker 1: ahead and run section eight. Increase the size here for

50:51 Speaker 1: easier viewing. Panel A shows the mean LST trajectory for

50:59 Speaker 1: each land covered type across the two nights, with error

51:02 Speaker 1: bars showing one standard deviation. Panel B converts that into

51:07 Speaker 1: a single number. The thermal gap defined as the commercial

51:10 Speaker 1: area mean minus the preserve mean. So in this case

51:19 Speaker 1: we see the gap remains very similar on night two,

51:22 Speaker 1: with slight point two degrees celsius decrease from four point

51:26 Speaker 1: five degree c to four point three degree c. This

51:29 Speaker 1: suggests that the preserves of appotransporration and long wave emission

51:34 Speaker 1: are not overwhelmed by the heat event, at least not

51:37 Speaker 1: after two nights from the start of the heat wave.

51:40 Speaker 1: This supports the argument for urban green space as heat

51:42 Speaker 1: resilience infrastructure, not just as an amenity. Another possible outcome

51:51 Speaker 1: for subsequent analysis might be what to evaluate whether the

51:56 Speaker 1: gap substantially narrows on further nights into a heat wave.

52:02 Speaker 1: And for this particular heat wave, there weren't ecostress data

52:07 Speaker 1: available around similar times after these two days for comparison,

52:13 Speaker 1: but that's something to flag and to consider if you're

52:15 Speaker 1: interested in studying another heat wave in a different location.

52:18 Speaker 1: So it'll depend on what data are available, but it's

52:23 Speaker 1: something that could be interesting to follow up with looking

52:26 Speaker 1: at longer term multi day heat waves and whether even

52:33 Speaker 1: certain natural areas might have a threshold beyond which sustained

52:36 Speaker 1: heat stress suppresses their cooling function. So this would be

52:40 Speaker 1: interesting to test in future work in other places. Like

52:42 Speaker 1: I said, if a more prolonged period of the heat wave,

52:46 Speaker 1: the temperature difference between the two areas might actually decrease.

52:53 Speaker 1: So this is really one big urban heat island resilience question,

52:58 Speaker 1: and it's why night time ecostress data can be valuable

53:01 Speaker 1: for urban planning, not just because it captures the heat

53:04 Speaker 1: island effect, but because it lets us evaluate whether specific

53:08 Speaker 1: green infrastructure actually delivers measurable cooling when it's needed the most.

53:14 Speaker 1: These are the kinds of analyzes that urban heat emergency planners,

53:18 Speaker 1: parks departments, and climate adaptation offices could use for their

53:23 Speaker 1: local and regional needs.

53:26 Speaker 3: Data like this.

53:27 Speaker 1: Identify cooling refuges, evaluate wor tree canopy investments will have

53:31 Speaker 1: the greatest thermal return, and map residential neighborhoods where nighttime

53:36 Speaker 1: heat exposure is most dangerous, particularly for people without air conditioning.

53:41 Speaker 1: Now I will turn it over back to Glenn.

53:53 Speaker 3: All right, So moving on to section four now, which

53:57 Speaker 3: is land service temperature downscaling. This brings us to this

54:01 Speaker 3: central idea of today's talk, thermal sharpening, or it's also

54:06 Speaker 3: sometimes called downscaling.

54:08 Speaker 2: It's often used interchangeably.

54:10 Speaker 3: The basic concept here is to take course resolution thermal

54:14 Speaker 3: data will course here meaning seventy meters is actually that

54:18 Speaker 3: is the highest resolution available from space, but still it's

54:22 Speaker 3: all pretty coarse. If you look at the urban environment

54:25 Speaker 3: and then combine that with higher resolution data that is

54:28 Speaker 3: well correlated with temperature, and that's the key here, and

54:32 Speaker 3: by doing that we can recover the thermal detail in

54:34 Speaker 3: a much finer spatial scale. In this case, we're going

54:37 Speaker 3: to go all the way down to a ten meter resolution.

54:40 Speaker 3: So this is a nice example that illustrates that where

54:43 Speaker 3: we can go from what looks like gloves of cooler

54:48 Speaker 3: and warm temperature features in the environment. Here this is

54:51 Speaker 3: other LA by the way down to ten meters, where

54:55 Speaker 3: we can now start seeing very fine scale detail in

54:58 Speaker 3: the temperatures. We can see the differences between these hiking

55:01 Speaker 3: trails and Griffith Park, golf courses, freeways, building roofs and

55:06 Speaker 3: so forth. So this opens up just a huge capability

55:10 Speaker 3: for us over specifically urban areas, but also in other

55:13 Speaker 3: natural vegetation areas as well. So as you can see,

55:19 Speaker 3: these are two more comparisons over different parts of LA.

55:23 Speaker 3: The one I just showed Griffith Park on the top

55:24 Speaker 3: and the bottom is of the Long Beach Harbor area.

55:28 Speaker 2: But as I said before, we can start really.

55:31 Speaker 3: Discriminating different urban features from one another. We can see,

55:34 Speaker 3: for example, the hot costco roof versus the coolest cemetery.

55:39 Speaker 3: You can see the difference between the freeway and the

55:42 Speaker 3: much cooler fairways on the golf course, for example.

55:46 Speaker 2: So this animation, let's see if it works.

55:49 Speaker 3: Yep. This shows you the kind of detail that we'll

55:52 Speaker 3: we're revealing from the seventy meter down to the ten

55:54 Speaker 3: meter level. This is of a Pasadena and you can

55:57 Speaker 3: see the Rose Bowl highlight again the top left corner there,

56:01 Speaker 3: and after sharpening you can see that we can resolve

56:04 Speaker 3: the distinctive sort of cool areas north of the Rose

56:07 Speaker 3: Bowl and the golf course, and also the hottest parking

56:09 Speaker 3: lots and roads to the southern part of the Rose Bowl.

56:13 Speaker 3: And we'll be going over that in quite a bit

56:15 Speaker 3: more detail in the tutorial because this is actually going

56:18 Speaker 3: to be one of the venues for the upcoming Olympic

56:21 Speaker 3: Games in LA and so this kind of high resolution

56:24 Speaker 3: thermal information is critical for heat intervention planning and also

56:29 Speaker 3: planning for shade interventions in areas like this. This is

56:35 Speaker 3: I think just a fantastic image. This is also created

56:37 Speaker 3: by a summer student of ours last year, by Ashley Agatap,

56:42 Speaker 3: and it shows a dance scale image over Paris during

56:46 Speaker 3: extreme heat wave in June twenty twenty two, where they

56:49 Speaker 3: actually had to close down a lot of the public

56:51 Speaker 3: spaces and museums within the city because it was so hot.

56:56 Speaker 3: This is sharpened and then and overlaid on Google Earth,

57:01 Speaker 3: which is why you can see the structural features of

57:04 Speaker 3: the actual urban area come into play here. And I'll

57:07 Speaker 3: go through this a little bit. I'll show you how

57:09 Speaker 3: to do this in the tutorial as well. And also

57:11 Speaker 3: it has it has its caveats too, which I will

57:14 Speaker 3: go through. But you know, you can see individual neighborhoods, parks,

57:18 Speaker 3: you can clearly see the same river and other infrastructure

57:22 Speaker 3: standing out with the temperature. And it's this kind of

57:25 Speaker 3: intraurban thermal detail like you simply cannot resolve at the

57:29 Speaker 3: seventy meter or higher pixel scale as I mentioned before,

57:33 Speaker 3: and looking ahead, the Olympic Games is coming up in

57:36 Speaker 3: twenty twenty eight in La This is a sharp and

57:39 Speaker 3: echo stress temperature map similar to what we did over

57:42 Speaker 3: the Phoenix area where we pull out the temperatures at

57:44 Speaker 3: the street level.

57:47 Speaker 2: And what I've highlighted here.

57:49 Speaker 3: Is the actual venues of the Olympic Games, and there's

57:53 Speaker 3: just tremendous potential heir for designing, for example, heat mitigation approaches,

57:57 Speaker 3: planning where to add shade, both natural and man made

58:01 Speaker 3: shade structures, and also cooling it outdoor events and watch

58:04 Speaker 3: party locations during the Olympic Games, and even extending these

58:07 Speaker 3: techniques to the coastal zone coastal water temperature and water

58:10 Speaker 3: quality monitoring. So I will be going over some of

58:15 Speaker 3: these areas that we'll be focusing on also for the

58:17 Speaker 3: Olympic Games in the tutorial. So moving on to some

58:23 Speaker 3: theory of the actual downscaling. Statistical downscaling would call sharpening.

58:28 Speaker 3: So what we're doing is we're just aggregating a coarse

58:31 Speaker 3: scale thermal pixel into its subpixel components. And the trek

58:36 Speaker 3: here is to use auxiliary data that has a higher

58:38 Speaker 3: space resolution that is statistically correlated with the land service temperature,

58:43 Speaker 3: which is our target yet and so this illustration shows

58:47 Speaker 3: that quite nicely. It's from chenne Bel and colleagues and

58:49 Speaker 3: shows how a single course pixel A one in this

58:53 Speaker 3: case gets disaggregated into four subpixels B one through before

58:57 Speaker 3: using the relationships learned from the higher resolution and covariates.

59:02 Speaker 3: And we'll actually be doing this at a finer scales,

59:04 Speaker 3: so we'll be taking one ecostress pixel and dividing it

59:09 Speaker 3: up into fifty different individual pixels. So that's seventy meter

59:14 Speaker 3: to ten meter resolution. So at the highest level, the

59:18 Speaker 3: workflow looks like this. We start out with ecostress at

59:22 Speaker 3: seventy meters as our target variable, and then we assigned

59:25 Speaker 3: predictive variables consisting of Sentinel two. This is a European

59:31 Speaker 3: Space Agency sensor that is launched in twenty fifteen and

59:35 Speaker 3: generates ten and twenty meters high resolution visible shortwave infrared data, so.

59:40 Speaker 2: That is our primary predictor.

59:42 Speaker 3: Here is the reflected surface reflectants from Sentinel two. We

59:47 Speaker 3: also include land cover types built up fraction and also

59:51 Speaker 3: the elevation. The output then is a We run that

59:55 Speaker 3: through a random forest regression. We train the data at

59:58 Speaker 3: seventy meters and then we predict the final or sharpened

1:00:01 Speaker 3: LST at the ten meter scale. So you may ask, well,

1:00:06 Speaker 3: why do we use sentinel to the spectral reflectance bans. Firstly,

1:00:11 Speaker 3: we use we could use that there are some twenty

1:00:13 Speaker 3: meter bands as well that we have been using in

1:00:16 Speaker 3: the past, but in this tutorial we'll be focusing on

1:00:19 Speaker 3: the ten meter bands, which gets us down to that

1:00:21 Speaker 3: sort of street level resolution that we're after. So the

1:00:27 Speaker 3: bands are located in the visible part of the spectrum,

1:00:29 Speaker 3: so between four hundred and seven hundred nanometers, this is

1:00:32 Speaker 3: the part of the spectrum we actually see with our

1:00:34 Speaker 3: naked eyes. We have three bands there, bands two, three,

1:00:37 Speaker 3: and four, blue, green, and red. And so these bands,

1:00:42 Speaker 3: the surface reflectance of these bands correlate with temperature because

1:00:46 Speaker 3: the more reflective a surface, the less the surface will

1:00:49 Speaker 3: be heating and therefore the cooler the surface will be.

1:00:52 Speaker 3: And this is a bit of a generalization that doesn't

1:00:55 Speaker 3: always occur, but it's a good statistical correlation with the temperature.

1:01:00 Speaker 3: And we also then at band eight, which is the

1:01:01 Speaker 3: near infrared band. This is right on the edge of

1:01:04 Speaker 3: the infrared spectrum that entsed around nine hundred and animeters,

1:01:08 Speaker 3: and this band is actually the opposite information to the

1:01:11 Speaker 3: other three, so it's in it's sensitive to vegetation, health,

1:01:15 Speaker 3: chlorical content, water content, and it's one of the strongest

1:01:18 Speaker 3: predictors in fact of LSD special variability in urban rural

1:01:22 Speaker 3: gradients whenever you have vegetation involved. And we'll go over

1:01:25 Speaker 3: that a bit later as well. So why does this work?

1:01:30 Speaker 3: So what I've plotted here is land service temperature. Each

1:01:33 Speaker 3: pixel he has an actual temperature on the ground, and

1:01:37 Speaker 3: I've plotted it versus the NDVII, which is a normalized

1:01:40 Speaker 3: difference vegetation index. Some of you may have heard of this.

1:01:43 Speaker 3: It's a band ratio between banded eight and four that

1:01:46 Speaker 3: I just showed you. And generally high ENDBI values are

1:01:50 Speaker 3: greener vegetation and are generally cooler, and the opposite is

1:01:54 Speaker 3: true with low DVII values. So, because service temperature is

1:01:59 Speaker 3: to sartisically tied to these surface reflectance properties, the boat

1:02:04 Speaker 3: bear low reflectant services like roads for example, run hearts

1:02:09 Speaker 3: vegetated bhy NDBI services stay cooler. And then on the

1:02:13 Speaker 3: bottom end of the spectrum here we have high albedo

1:02:15 Speaker 3: services like white roofs for example, or white roads are

1:02:21 Speaker 3: also cool and so there's a sort of triangular nonlinear

1:02:25 Speaker 3: relationship between surface reflectants and land surface temperature. And these

1:02:30 Speaker 3: are the exact nonlinear relationships that our machine learning model

1:02:34 Speaker 3: is going to be learning at the seventy meter scale

1:02:36 Speaker 3: and then applying down to the seventy meter scale. So

1:02:41 Speaker 3: why sharpen ecostress down to these levels. What are the

1:02:45 Speaker 3: primary urban applications and what are our reasons for doing this? Well,

1:02:49 Speaker 3: there are six concrete reasons, heah. The firstly, we're able

1:02:52 Speaker 3: to of course resolve urban features from the seventy meter

1:02:56 Speaker 3: down to the ten meter scale. Streets, parks, rooftops blurd

1:02:59 Speaker 3: together at the seventy meter scale and be able to

1:03:01 Speaker 3: resolve those. It lets us valid our heat interventions at

1:03:05 Speaker 3: the scale at which they're actually deployed, so think of

1:03:08 Speaker 3: fool pavements being applied over city's cool roofs. It also

1:03:15 Speaker 3: dramatically reduces our mixed pixel error, especially when we're in

1:03:18 Speaker 3: sort of heterogeneous neighborhoods. It forthited lives at the parcel

1:03:24 Speaker 3: level planning data that the city agencies actually already use.

1:03:29 Speaker 3: It sharpens heat vulnerability mapping and undeserved poorer communities, which

1:03:33 Speaker 3: is critical for example downtown LA. And the last part

1:03:38 Speaker 3: here is really critical is that we apply this residual correction,

1:03:41 Speaker 3: which we will go over in detail, which means we

1:03:44 Speaker 3: actually preserve the radiometric accuracy by normalizing to that of

1:03:48 Speaker 3: the ecostress instruments. So we're adding spatial detail, not inventing temperatures,

1:03:53 Speaker 3: if you will, and I'll show you some of these.

1:03:56 Speaker 2: Examples in the next few sites.

1:04:00 Speaker 3: At seventy meters you can think of we have a

1:04:03 Speaker 3: like a small city buck scale temperature, and then after

1:04:06 Speaker 3: the sharpening, we end up with ten meters, which equivalent

1:04:09 Speaker 3: to forty nine pixels, and so we have a fifty

1:04:11 Speaker 3: times improvement in the space or resolution using this methodology.

1:04:17 Speaker 3: So why random forest for this task. Well, it's inherently

1:04:22 Speaker 3: available within Google Earth Engine, which is great. And the

1:04:25 Speaker 3: basis for which it works is you take an ensemble

1:04:29 Speaker 3: of these what we call decision trees. You have outlined

1:04:32 Speaker 3: m tree one through hand. Typically we start out with

1:04:35 Speaker 3: about one hundred trees, but you can increase or increase

1:04:38 Speaker 3: that value, and each of these is trained on a random.

1:04:41 Speaker 2: Subset of features.

1:04:43 Speaker 3: So what I mean here by features is that it's

1:04:46 Speaker 3: a subset of a different combinations of the predictive variables

1:04:50 Speaker 3: and also the different pixels within the scene. And so

1:04:54 Speaker 3: we then end up with a final prediction, which is

1:04:56 Speaker 3: an average across all of those trees. So it's a

1:05:01 Speaker 3: very powerful technique because it, as I said before, it

1:05:04 Speaker 3: handles these nonlinear LST and service relationships. It's robust outliers

1:05:09 Speaker 3: because it's able to generate an ensemble mean which reduces

1:05:14 Speaker 3: the variance in the data, and it's also transferable across scale,

1:05:18 Speaker 3: So we train at seventy meters and then we can

1:05:20 Speaker 3: predict a ten meters. Also, something that comes out of

1:05:24 Speaker 3: the random forest which is really interesting is a feature

1:05:27 Speaker 3: importance information which identifies which of our predictors were the

1:05:33 Speaker 3: most sensitive or the most important in arriving at our

1:05:37 Speaker 3: final result, and we'll go over that in a minute

1:05:40 Speaker 3: as well. So in from sort of the high level,

1:05:45 Speaker 3: this is what the sharpening and Google Earth Engine is

1:05:49 Speaker 3: going to involve. These are the different steps. We start

1:05:51 Speaker 3: with our training data and our target variable ecastress LST

1:05:55 Speaker 3: at seventy meter. We then train the random forest model

1:05:58 Speaker 3: within Google Earth Engine at the seventy meter scale, and

1:06:02 Speaker 3: then we predict the ten meter final temperature and then

1:06:07 Speaker 3: apply this residual correction, which means we upscale the ten

1:06:11 Speaker 3: meter prediction to seventy meters subtract from the original seventy

1:06:15 Speaker 3: meter observation from uka stress, and then we add that

1:06:18 Speaker 3: residual back to the ten meter result and I'll show

1:06:21 Speaker 3: you that on the next slide.

1:06:23 Speaker 2: So these are just some cutouts that are made from within.

1:06:26 Speaker 3: Google Life Engine at various steps along the way throughout

1:06:29 Speaker 3: the sharpening process. This is a cutout over an area

1:06:33 Speaker 3: heterogeneous urban area here in La So in the middle

1:06:38 Speaker 3: here you can see the original ecostress the LSD at

1:06:40 Speaker 3: seventy meters. The next step is the initial downscaled LSD.

1:06:45 Speaker 3: So this is after applying the model at ten meters,

1:06:47 Speaker 3: we end up with this result. So you can see

1:06:50 Speaker 3: immediately find spatial detail coming out within the urban environment

1:06:55 Speaker 3: that correlates nicely with the actual visible image. On the

1:06:58 Speaker 3: top left. You can also see that the actual absolute

1:07:02 Speaker 3: values of the temperature here are quite different to the

1:07:05 Speaker 3: actual observed data in the top we don't see these

1:07:08 Speaker 3: sort of hot spots in the lower left and bottom areas.

1:07:12 Speaker 3: So then after the residual correction I just explained, we

1:07:15 Speaker 3: upscale the ten meter to seventy subtracted from the seventy

1:07:18 Speaker 3: meter ecostress, and we end up with this residual corrected

1:07:21 Speaker 3: final LST at ten meters which preserves the actual radiometric

1:07:27 Speaker 3: information in the original seventy meter resolution image. And you

1:07:30 Speaker 3: can see that these correlate a lot better now, So

1:07:35 Speaker 3: quickly go through what these predictors look like, sort of

1:07:39 Speaker 3: at a much larger scale over the entire LA City

1:07:43 Speaker 3: and County area. This is the area that we'll be

1:07:46 Speaker 3: focusing on in the tutorial today by the work. By

1:07:48 Speaker 3: the way, so this is the sentinel to surface reflectance.

1:07:53 Speaker 3: This is for band eight. This is the near infrared

1:07:55 Speaker 3: band a ten meter resolution.

1:07:59 Speaker 2: This is.

1:08:01 Speaker 3: Provided by Coperticus and European Space Agency through Google Earth Engine,

1:08:07 Speaker 3: and so we'll be using the blue, green.

1:08:08 Speaker 2: Red, and ne infrared bands.

1:08:09 Speaker 3: This is the near infrared band. And you can see

1:08:13 Speaker 3: here immediately areas sort of features that pop out. For example,

1:08:18 Speaker 3: these wide areas that you see here with high reflectants

1:08:21 Speaker 3: are in fact green spaces.

1:08:22 Speaker 2: So these are pretty much the golf courses around the.

1:08:25 Speaker 3: LA region which have a high band aid reflectants because

1:08:28 Speaker 3: they're chlorofully active green types of irrigated vegetation. And conversely,

1:08:37 Speaker 3: we are very low reflectants over dense urban paved surfaces,

1:08:42 Speaker 3: mostly human made types of surfaces. This is the NDBIS

1:08:47 Speaker 3: I spoke about. It's a band ratio between band AID

1:08:49 Speaker 3: and Band four. High NDBI values correlate also to dance

1:08:55 Speaker 3: and green vegetation, and conversely varius soils and some nest

1:08:59 Speaker 3: vegetiation or urban materials have a very low NDDI And

1:09:04 Speaker 3: conversely with temperature. Of course, high NDBI values should have

1:09:08 Speaker 3: cool temperatures and low DBR values should have harder temperatures.

1:09:14 Speaker 3: The other.

1:09:16 Speaker 2: Predictor, yeah that I'll show is the built surface fraction.

1:09:19 Speaker 3: This is from the European Commission's Global Human Settlement Layer

1:09:23 Speaker 3: available within Google Earth Engine, and this data captures gradients

1:09:28 Speaker 3: and urbanization that other land cover land use categories alone

1:09:32 Speaker 3: are n't able to capture, and especially useful in mixed

1:09:35 Speaker 3: perio urban regions where you can see the fraction of

1:09:39 Speaker 3: the build up area changing very rapidly into more vegetator

1:09:43 Speaker 3: rural regions. And then our fourth predictor is actual elevation.

1:09:49 Speaker 3: This is from SRTMS that he need a product. We

1:09:53 Speaker 3: then resemble that ten meters for the training and elevation

1:09:57 Speaker 3: really matters in fact in this region because the lapse

1:10:00 Speaker 3: rate alone explains a couple of degrees of LSD variation,

1:10:04 Speaker 3: especially across these hilly areas of LA. You think of

1:10:06 Speaker 3: the Santa Monica Mountains, see are the Saint Gabriel's and

1:10:10 Speaker 3: we want the model to learn that effect explicitly rather

1:10:13 Speaker 3: than sort of baking it into the residuals. And then lastly,

1:10:18 Speaker 3: hear the star of the show. This is the Eker

1:10:19 Speaker 3: Stress Land Service temperature at seventy meter resolution.

1:10:24 Speaker 2: This is the actual image we'll be working with in

1:10:26 Speaker 2: the tutorial.

1:10:27 Speaker 3: Next. This was I believe in the fourteenth of August

1:10:31 Speaker 3: twenty twenty. This is an extreme heat wave events in LA.

1:10:35 Speaker 3: You can see the extremely hot temperatures here in the

1:10:37 Speaker 3: San Fernando Valley in the northern part here that is

1:10:40 Speaker 3: insulated from the cooler ocean breeze. You can see strong

1:10:44 Speaker 3: gradients here along the coastline having much cooler temperatures and

1:10:48 Speaker 3: then increasing too much harder temperatures.

1:10:52 Speaker 2: In the interior urban areas.

1:10:55 Speaker 3: This was quite a four PM in the afternoon, so

1:10:58 Speaker 3: one of the hottest parts of the day. I also

1:11:00 Speaker 3: notice that there's no cloud artifacts in this image, so

1:11:05 Speaker 3: there's a very nice clear sky day, which is why

1:11:07 Speaker 3: I chose this image.

1:11:08 Speaker 2: Also noticed that very distinctive cool.

1:11:11 Speaker 3: Islands here you can notice these sort of cool little

1:11:14 Speaker 3: areas within the urban environment. And again these are these

1:11:18 Speaker 3: mostly golf courses, parks, green spaces that are twenty thirty

1:11:23 Speaker 3: degrees cooler in fact than the surrounding urban landscape. And

1:11:28 Speaker 3: then I mentioned the feature importance outcome of the random forest,

1:11:33 Speaker 3: and this is really interesting and something that I will

1:11:36 Speaker 3: show you how to derive from the data and Google

1:11:39 Speaker 3: atventionine as well, and it's how the predictors actually rank

1:11:42 Speaker 3: in the train model in terms of feature importance. So

1:11:46 Speaker 3: in this case, elevation actually came in at the top

1:11:48 Speaker 3: at seventeen point six percent, followed by the four sentinel

1:11:52 Speaker 3: bands or plustered around seventeen percent, so that these four,

1:11:56 Speaker 3: these five predictors are driving most of the information that

1:12:00 Speaker 3: is provided in the random.

1:12:01 Speaker 2: Forest getting us to that ten meter scale.

1:12:04 Speaker 3: The built surface actually contributed about eleven point four percent,

1:12:08 Speaker 3: so pretty significant.

1:12:09 Speaker 2: But the actual land use land cover map.

1:12:13 Speaker 3: Which I included in this example had a very low

1:12:16 Speaker 3: information content, and that's actually because it's derived from the

1:12:19 Speaker 3: sentinel to reflectant data. So essentially you don't want to

1:12:23 Speaker 3: add in variables that are correlated with each other, so

1:12:26 Speaker 3: we should probably remove the land use land cover from

1:12:29 Speaker 3: this case. All right, So that's it on the theory.

1:12:35 Speaker 3: We'll now get onto the fun pat here, which will

1:12:37 Speaker 3: be the actual ads on exercise where we'll be diving

1:12:41 Speaker 3: into Google avengine and actually showing you how to produce

1:12:44 Speaker 3: these really nice striking maps of the environment. Before I

1:12:50 Speaker 3: get started here, I just want to preface this by

1:12:52 Speaker 3: saying that we do have everything I'm going to go

1:12:55 Speaker 3: through here. Literally every step and click I'm going to

1:12:58 Speaker 3: make is actually available on a tutorial that we created

1:13:03 Speaker 3: in Microsoft Work. So we'll make these available so that

1:13:08 Speaker 3: you don't have to, you know, look through the video

1:13:11 Speaker 3: tutorial again and listen to EGO going through it all.

1:13:15 Speaker 3: You can actually follow along step by step, every click

1:13:18 Speaker 3: and every move I make on this document tutorial that

1:13:22 Speaker 3: you can reference for later use. So starting up from

1:13:28 Speaker 3: the beginning, we're going to open a web browser. For example,

1:13:31 Speaker 3: it tells you here go to code dot earth Engine,

1:13:34 Speaker 3: Google dot com. You'll also need to sign up for

1:13:37 Speaker 3: a free Google account, and you can do that going

1:13:40 Speaker 3: by going to this website earth Engine Logogle dot com

1:13:43 Speaker 3: sign up, slash sign up. We also have a link

1:13:47 Speaker 3: to a YouTube video in our resource page which actually

1:13:51 Speaker 3: shows you step by step on how to actually sign

1:13:54 Speaker 3: up for a free account of Google Earth to start

1:13:57 Speaker 3: and be able to do this tutorial. So this is

1:14:01 Speaker 3: what you'll end up with if you start your Google

1:14:03 Speaker 3: Earth Engine page. On the left are your various scripts

1:14:08 Speaker 3: or pieces of code that you can stall. We then

1:14:12 Speaker 3: have a docs page. You know, I need to look

1:14:15 Speaker 3: at that. The assets are actually data that your code

1:14:18 Speaker 3: can operate on. We won't go into that either. So

1:14:23 Speaker 3: what you'll be doing is you'll be clicking new here

1:14:25 Speaker 3: it's a bigger red icon.

1:14:26 Speaker 2: Okay, click new. You'll go to file, give it a name.

1:14:32 Speaker 3: I'm just going to call it sharp that you can

1:14:34 Speaker 3: call it any name you want, and or create a

1:14:37 Speaker 3: new file which you then can paste directly the code

1:14:44 Speaker 3: that we're going to provide for you, which is the

1:14:46 Speaker 3: Eco sharp V one point one code which you're seeing

1:14:49 Speaker 3: right here. And so that's that's how to get started.

1:14:53 Speaker 3: So you'll click save and then you have your JavaScript

1:14:56 Speaker 3: Google Engine code ready to go. So a couple just

1:15:00 Speaker 3: orientation things here first and center. Of course, you can

1:15:04 Speaker 3: see the actual code. This is JavaScript code. You don't

1:15:06 Speaker 3: need to even look at this if you don't want to.

1:15:09 Speaker 3: But you'll see the buttons on the top you're showing

1:15:11 Speaker 3: how to run the code, reset the code. On the

1:15:14 Speaker 3: file right panel.

1:15:16 Speaker 2: You can see something called the console.

1:15:18 Speaker 3: This outputs various types of information that the code runs,

1:15:22 Speaker 3: and we'll see that in a minute. For example, some

1:15:25 Speaker 3: stats on accuracy, what data was used at what time.

1:15:29 Speaker 3: Tasks will also go over when we're going to finally

1:15:33 Speaker 3: export this data to a geot file and open it

1:15:36 Speaker 3: in qhis, you'll be able to run a task here

1:15:39 Speaker 3: to export it directly to your Google Drive and we'll

1:15:42 Speaker 3: go over that in a minute. Okay, so getting started,

1:15:47 Speaker 3: we're going to start over the Los Angeles region. Of course,

1:15:49 Speaker 3: I said, we're going to start with that image that

1:15:51 Speaker 3: I just showed you. If you grab this little toggle heah,

1:15:55 Speaker 3: you can actually move the map up and down to

1:15:59 Speaker 3: make it easier for you to actually see the data

1:16:01 Speaker 3: coming in. So I'm going to move it right up

1:16:03 Speaker 3: for now, and we're going to hit run to simply

1:16:06 Speaker 3: hit run, and what we end up with is a

1:16:09 Speaker 3: user interface panel on the left, and you use interface

1:16:12 Speaker 3: panel on the right. What you'll see on the left,

1:16:16 Speaker 3: let's start of the left. You'll see a date range here.

1:16:18 Speaker 3: So in this case, I've put in a single date,

1:16:21 Speaker 3: the fourteenth of August twenty twenty, where I knew we

1:16:24 Speaker 3: had this scene over Los Angeles during the heatwave events.

1:16:28 Speaker 3: If you had to change this and say, for example,

1:16:30 Speaker 3: go from June till July twenty twenty, what the code

1:16:33 Speaker 3: will do is it'll take all the ecostress observations over

1:16:36 Speaker 3: their time and generate a mean composite image over a

1:16:40 Speaker 3: date range. So we can either look at a single

1:16:43 Speaker 3: date for a single image or create a composite from

1:16:47 Speaker 3: a date range. You can also see additional summer years.

1:16:51 Speaker 3: If we clicked on those additional years, we can create

1:16:54 Speaker 3: a composite over consecutive years.

1:16:57 Speaker 2: Over a fixed date range.

1:16:58 Speaker 3: So if we want all this summertime average temperature across

1:17:01 Speaker 3: the entire mission period, can click on all of these

1:17:04 Speaker 3: years and generate a mean composite average. And so this

1:17:08 Speaker 3: is often useful for city planners and managers. They want

1:17:10 Speaker 3: to know what is the baseline average summertime temperature that

1:17:14 Speaker 3: we've experienced over the last five years. The next part

1:17:19 Speaker 3: here is the local time window, so this is in

1:17:21 Speaker 3: hours of day.

1:17:22 Speaker 2: What I've done is years.

1:17:23 Speaker 3: I've converted this you can convert to different time zones

1:17:28 Speaker 3: around the world, and that makes it easier where you

1:17:30 Speaker 3: don't have to deal with UTC times. So we're in

1:17:33 Speaker 3: Los Angeles, We're going to click Los Angeles time zone,

1:17:36 Speaker 3: and in this case we're going to look at all

1:17:39 Speaker 3: hours of the day. We know there's one scene here.

1:17:41 Speaker 3: You can change this to if you just want the

1:17:43 Speaker 3: afternoon hour, say twelve till eighteen. We'll get all the

1:17:47 Speaker 3: data from the hardest times of day twelve till six pm.

1:17:52 Speaker 3: But let's go back to zero twenty three. The next

1:17:56 Speaker 3: part here is the max cloud tolerance. So if you

1:17:59 Speaker 3: want less cloud on the scene, you can decrease this value.

1:18:04 Speaker 2: You can go down to ten.

1:18:05 Speaker 3: I would suggest for a long term composite that you

1:18:08 Speaker 3: go down to ten percent. So we don't want more

1:18:11 Speaker 3: than ten percent of cloud on the scene. The cloud

1:18:14 Speaker 3: is already screened by the way, so it won't be

1:18:16 Speaker 3: included in the data. But we don't want to combine

1:18:20 Speaker 3: images where the part of the image is missing, and

1:18:23 Speaker 3: then we combine it with other images and you get

1:18:26 Speaker 3: some artifacts coming through. So I would use ten for

1:18:30 Speaker 3: a composite and use thirty for a single images. Fine,

1:18:35 Speaker 3: then you have an option to draw custom rectangle. So

1:18:39 Speaker 3: if I click on that and click activate selection, and

1:18:42 Speaker 3: we just want to do San Fernando Valley, for example,

1:18:45 Speaker 3: we can drag and drop and run our sharpening over

1:18:48 Speaker 3: that specific area. So I'm going to clear that because

1:18:52 Speaker 3: what we're actually going to do here is go to

1:18:55 Speaker 3: an ecostress tile and we're going to run the code

1:18:58 Speaker 3: on an entire ecostress tile over this region. So if

1:19:01 Speaker 3: I activate that selection and I'll click on anywhere in

1:19:04 Speaker 3: the center of the LA map, it's going to give

1:19:08 Speaker 3: us the tile boundary for the specific area. And I

1:19:12 Speaker 3: won't go through the details of the tile boundaries, but

1:19:15 Speaker 3: the sentient we use the Sentinel two tiling structure here,

1:19:19 Speaker 3: which is based on the n g r S system.

1:19:21 Speaker 3: This is the Military Grid Reference system where the earth

1:19:24 Speaker 3: is divided up into these fixed tiles. They're roughly about

1:19:27 Speaker 3: one hundred kilometers wide, and you can go and Google

1:19:30 Speaker 3: this and learn more and that if you want. But

1:19:34 Speaker 3: this is the tile we're working with that covers the

1:19:37 Speaker 3: whole LA area quite nicely, as you can see.

1:19:40 Speaker 2: And then on the bottom left here you can run

1:19:43 Speaker 2: and export.

1:19:43 Speaker 3: But before we do that, let's go to the right

1:19:46 Speaker 3: here where we can see We have some other options

1:19:48 Speaker 3: here which I really useful, so we can choose exactly

1:19:52 Speaker 3: which sharpening predictors we want to use. The central two

1:19:55 Speaker 3: bounds here are required, so that's all is going to

1:19:58 Speaker 3: be clicked. But we can and click off the land use.

1:20:01 Speaker 3: As I said, it doesn't have a huge influence on

1:20:02 Speaker 3: the final results. We probably want to include the surface

1:20:05 Speaker 3: fraction and elevation those had significant feature importance values. I'm

1:20:10 Speaker 3: not going to include the NDBI just because that information

1:20:13 Speaker 3: is already baked into our reflectance spans. But you can,

1:20:16 Speaker 3: for example, use NDBI and the reflectant spans and you

1:20:20 Speaker 3: can see how what kind of data that provides, for

1:20:23 Speaker 3: example over ag areas that may be more beneficial. Next,

1:20:27 Speaker 3: we can choose the the temperature unit here, so we

1:20:32 Speaker 3: can go from calvin, which is the standard unit the

1:20:35 Speaker 3: scientific unit that stress comes in. We can switch to

1:20:38 Speaker 3: celsius and then pharranhyde here. I'm going to stick with

1:20:40 Speaker 3: pharranhid and then we have the export options. So this

1:20:45 Speaker 3: is you know, the actual data that we want to

1:20:48 Speaker 3: look at, and qgis eventually is going to be the

1:20:50 Speaker 3: LST serventy meter ten meter. Don't worry about this model

1:20:54 Speaker 3: asset as I mentioned before in the top left, if

1:20:58 Speaker 3: you did export the asset, you'd have this available in

1:21:00 Speaker 3: your assets here, and then you could apply that random

1:21:02 Speaker 3: forest model to another scene if you wanted. Later. You

1:21:07 Speaker 3: don't have to do the training on every scene, but

1:21:10 Speaker 3: that maybe you should be left to another tutoria.

1:21:15 Speaker 2: Then you can export also the predictor layers.

1:21:17 Speaker 3: If you want to look at, for example, the images

1:21:20 Speaker 3: I just showed you in the slides with the sentinel bands, Sandy,

1:21:23 Speaker 3: the island cover. You can export all of those that

1:21:25 Speaker 3: if you wish too. So I'm going to leave those alone.

1:21:29 Speaker 3: And if we click run and export, you can see

1:21:32 Speaker 3: the bottom left here that it's already found all the data.

1:21:35 Speaker 3: It's already on step two. It's not sharpening the data

1:21:37 Speaker 3: to ten meter resolution, and you can just see how

1:21:41 Speaker 3: quickly this operates.

1:21:42 Speaker 2: We already have the LST pulled in.

1:21:44 Speaker 3: This is the Ecostress Landservice temperature actually just showed you

1:21:48 Speaker 3: in the slides, and while it's running through its sharpening

1:21:52 Speaker 3: training process. Here, if you click on the layers here

1:21:56 Speaker 3: on the top right of the image, you can actually

1:21:59 Speaker 3: toggle between the different predictors and the LST that have

1:22:03 Speaker 3: been brought in by Google Earth Engine, and also on

1:22:07 Speaker 3: the right here as an opacity slider, which is kind

1:22:10 Speaker 3: of cool. You can see how the image fades in

1:22:12 Speaker 3: and out of the background Google Earth map. So you know,

1:22:16 Speaker 3: while this is running, I guess you can play around

1:22:18 Speaker 3: with the LSTs and the other data yet and look

1:22:22 Speaker 3: at what the data looks like. And then okay, let's

1:22:28 Speaker 3: go through the console here. There's actually a lot of

1:22:30 Speaker 3: information that gets printed out in the console on the right.

1:22:32 Speaker 3: Inside here, it just gives the summary of the actual

1:22:35 Speaker 3: options you use, the date range, the time, cloud tolerance.

1:22:40 Speaker 3: It'll show you the actual n g RS tile number

1:22:44 Speaker 3: that Echo stress relates to you. This is the eleven

1:22:48 Speaker 3: sl T tile, and then it will show you the

1:22:51 Speaker 3: actual Sentinel two scene that was used. So the code

1:22:55 Speaker 3: currently looks at Sentinel two data within a month of

1:22:59 Speaker 3: the observation. So in this case you can see we

1:23:02 Speaker 3: found a scene on the twelfth of July. You can

1:23:05 Speaker 3: stress scene was the fourteenth of August, so not exactly

1:23:09 Speaker 3: a month, but close enough. The summertime period vegetation is

1:23:14 Speaker 3: not going to change too much, for example, over LA,

1:23:16 Speaker 3: so it's fine that it's within a month.

1:23:18 Speaker 2: The code automatically looks for very clear Sentinel two scenes

1:23:22 Speaker 2: in this case is barely no cloud on the Sentinel

1:23:26 Speaker 2: two scene.

1:23:26 Speaker 3: That's pretty important. You should probably check that, but the

1:23:30 Speaker 3: code will automatically check for cloud pre data. There's some

1:23:34 Speaker 3: of the information. Yeah, it shows the data that you

1:23:37 Speaker 3: want to explot and then also actually shows the sharpening accuracy,

1:23:42 Speaker 3: which is the initial ten meter resolution image that showed earlier. Now,

1:23:49 Speaker 3: after normalization, this actually dramatically reduces in the bias and

1:23:54 Speaker 3: rms after we do the radiometrinmalization chief for stress. So

1:23:58 Speaker 3: these numbers may look pretty big, like eight degrees fahrenheit,

1:24:01 Speaker 3: but they end up being pretty small in the end.

1:24:03 Speaker 3: It actually might change this so that produces the stats

1:24:07 Speaker 3: after we've done that radiomentric organization. Anyway, let's go into tasks.

1:24:12 Speaker 3: This is the important partner. We see our three tasks.

1:24:15 Speaker 3: Here are the three data sets that we chose to

1:24:18 Speaker 3: be exported to our Google drive. So I've already done this,

1:24:22 Speaker 3: so I'm not going to We're not going to wait

1:24:23 Speaker 3: and run this in real time. It actually takes quite

1:24:26 Speaker 3: a bit of time because if you click on the

1:24:29 Speaker 3: ten meta here and head run, it will give you

1:24:31 Speaker 3: the task name. You can change that if you want,

1:24:35 Speaker 3: shows the coordinate system, the resolution, and the drive shows

1:24:39 Speaker 3: the drive folder that it will create for you if

1:24:42 Speaker 3: you haven't created one in your Google Drive out and

1:24:45 Speaker 3: then the file name and the file format of Yoursio

1:24:49 Speaker 3: is a geotep, so it has the special information included

1:24:53 Speaker 3: within the data.

1:24:54 Speaker 2: And so then head run.

1:24:56 Speaker 3: You'll see that there's a submitted task it we'll be running,

1:24:59 Speaker 3: and I'll show time and it's taken to run. You do

1:25:03 Speaker 3: that with the others and then you end up with

1:25:08 Speaker 3: your data and your Google Drive. So my case, this

1:25:11 Speaker 3: is the GEE you can stress folder and you can

1:25:14 Speaker 3: see a list of due to files here. You can

1:25:16 Speaker 3: see I have quite a few because I've obviously been

1:25:19 Speaker 3: doing a lot of testing on this in the last

1:25:20 Speaker 3: few months. And then you can click on the right side.

1:25:24 Speaker 3: You can see these three little dots on each file.

1:25:27 Speaker 3: Click that head download and it will download your file

1:25:30 Speaker 3: to your desktop. Okay, so let's move on to the

1:25:35 Speaker 3: next part here, which is actually visualizing this data in QGIS.

1:25:39 Speaker 3: So just to recap everything I'm just being through, is

1:25:45 Speaker 3: on this Google Earth Engine running ekshop version one point

1:25:48 Speaker 3: one tutorial, So you can go through that step by

1:25:52 Speaker 3: step at your own pace if you want, and then

1:25:55 Speaker 3: we'll be available if we're going to be available.

1:25:57 Speaker 2: For questions as well.

1:25:59 Speaker 3: So similarly, we have a full tutorial on the visualization

1:26:04 Speaker 3: in QGIS. So every click instep you see coming up

1:26:09 Speaker 3: now is laboriously documented in this file, so you can

1:26:13 Speaker 3: go through this at your own pace as well. There's

1:26:16 Speaker 3: little tip boxes, There's all kinds of nice little sympets

1:26:19 Speaker 3: of information here and it ends with having a quick

1:26:23 Speaker 3: reference and troubleshooting guide, so really nice reference to have

1:26:27 Speaker 3: handy while we go through the actual tutorial. So let's

1:26:33 Speaker 3: open up QGIS.

1:26:35 Speaker 2: I already have mine open.

1:26:36 Speaker 3: There's a link in the resources to how to download QGIS,

1:26:40 Speaker 3: and I'm pretty sure most of you will have it

1:26:42 Speaker 3: already from Savannah's tutorials. So let's.

1:26:50 Speaker 2: Let's begin.

1:26:51 Speaker 3: So if you go to the top left, had project

1:26:53 Speaker 3: and hit new, we start a new project going into

1:26:57 Speaker 3: project save as and call it something. So let's call

1:27:02 Speaker 3: it Eco stress sharpening. And I'm just going to say

1:27:08 Speaker 3: it in my documents folder. So we have it saved,

1:27:10 Speaker 3: we have our projects saved. We're I'm going to lose

1:27:11 Speaker 3: all the data that we're we're doing. So next step

1:27:16 Speaker 3: is we want to we want to this the Google basemp's.

1:27:20 Speaker 3: This is really important and very useful to have. So

1:27:23 Speaker 3: go up to right at the top wh is HC,

1:27:26 Speaker 3: M G, I S. Click on that, go to basemap

1:27:29 Speaker 3: and click on Google Satellite. So now we have our

1:27:31 Speaker 3: Google Nice satellite basemp. Next we go to a.

1:27:37 Speaker 2: Layer, add layer and raster layer and.

1:27:43 Speaker 3: On the source here you can see I already have

1:27:46 Speaker 3: mine that i've that I've uploaded. But if I click

1:27:49 Speaker 3: on the right here the three dots, it's going to

1:27:51 Speaker 3: take us to folder where mine is in our set,

1:27:55 Speaker 3: and I'm going to upload the firstly the ten meter

1:27:58 Speaker 3: lst add close and I'm going to do the same

1:28:03 Speaker 3: thing for the twenty meter sorry, with seventy meter layer.

1:28:09 Speaker 3: So we go yeah, and it's seventy meters ad. Okay,

1:28:13 Speaker 3: Now we have on the bottom left on our layers,

1:28:16 Speaker 3: you can see we have the seventy meter product ten

1:28:18 Speaker 3: meter product. Google said like that, Now we're not seeing

1:28:21 Speaker 3: anything because we actually have to zoom to the layer.

1:28:23 Speaker 3: So if we ride click and hit zoom to layer,

1:28:28 Speaker 3: we will now see the map center on our image

1:28:34 Speaker 3: of the Los Angeles and it's all default and gray scale.

1:28:38 Speaker 3: That's kind of boring. We don't want that. So what

1:28:40 Speaker 3: are we gonna do is were gonna double click or

1:28:43 Speaker 3: right click and hit properties on our file. We end

1:28:47 Speaker 3: up with this page. We start with symbology. Go to

1:28:51 Speaker 3: the top where it says render type hits single band

1:28:55 Speaker 3: pseudo color. We don't want any y scale. Then for

1:29:00 Speaker 3: the min max value settings, I'm going to click this errow.

1:29:03 Speaker 3: And what I typically like to use is I don't

1:29:06 Speaker 3: like to hard code the men in max because it's

1:29:08 Speaker 3: going to change for every scene. So I typically like

1:29:11 Speaker 3: to use between two to five of the lower percentile

1:29:15 Speaker 3: of data. And then on the hi side, I either

1:29:18 Speaker 3: go from ninety eight to one hundred, and that kind

1:29:20 Speaker 3: of varies on the area that I'm looking at. So

1:29:23 Speaker 3: let's use those values for the color ramp. I actually

1:29:29 Speaker 3: so this is really a personal preference. You can go

1:29:32 Speaker 3: Magma spectral works pretty well. If you look at all

1:29:39 Speaker 3: color ramps here, you can see there's a really wide

1:29:42 Speaker 3: variety of color ramps. I'll show you a little trick

1:29:46 Speaker 3: if the one that I actually like is if you

1:29:50 Speaker 3: go to create color ramp, go to catalog CBT city.

1:29:55 Speaker 3: I know this is a bit much, but qgi us

1:30:01 Speaker 3: and then all by author on the top rides. I

1:30:06 Speaker 3: already like this temperature, so let's let's use temperature all right.

1:30:12 Speaker 3: So if we click and we head apply, now we

1:30:14 Speaker 3: can see our image nicely scaled in color scale where

1:30:19 Speaker 3: the blues are the cooler colors red so are out

1:30:21 Speaker 3: of colors. So it's pretty intuitive, and we we're just

1:30:25 Speaker 3: going to go ahead and do the same thing quickly

1:30:27 Speaker 3: with our second band. I'll go through this pretty quick already,

1:30:34 Speaker 3: change our centerur thresholds by one hundred, and we'll go

1:30:44 Speaker 3: with our temperature color up, which again I have to

1:30:47 Speaker 3: add I actually need to save this once. I don't

1:30:49 Speaker 3: have to do this every time. I haven't temperature okay, apply, Okay,

1:30:57 Speaker 3: Now we have both data sets together at the same

1:31:02 Speaker 3: color range and more importantly, in the same color ramp.

1:31:07 Speaker 3: So firstly, what we can do here, Let's go in.

1:31:13 Speaker 3: Let's go into this area in the Sant Banana Valley.

1:31:15 Speaker 3: Let's look at this little blue island here and see

1:31:17 Speaker 3: what's going on. So we zoom in here. We can

1:31:21 Speaker 3: toggle all these on and north, which is really really useful.

1:31:25 Speaker 3: So you can look at the Google background image. Yep,

1:31:27 Speaker 3: it's a golf course. So this is a two golf

1:31:30 Speaker 3: courses in the Sant Banana Valley, and on the right

1:31:32 Speaker 3: side is an actual park. This is an actual area

1:31:36 Speaker 3: where people fly remote control aircraft. In fact, this image here,

1:31:41 Speaker 3: this must have been in the spring, because everything looks

1:31:43 Speaker 3: pretty green, and in actual fact, during the summer this

1:31:45 Speaker 3: is pretty barren and bare soil, very hot, and you

1:31:48 Speaker 3: can actually see that in the echo stressed data. You

1:31:52 Speaker 3: can see the very hot temperatures sere over the park

1:31:55 Speaker 3: and then much cooler over the irrigated golf course. So

1:31:59 Speaker 3: looking at ten meters, so we just see two big

1:32:01 Speaker 3: blobs right at seventy meters, there's not much detail there.

1:32:04 Speaker 3: If we're going to the ten meter scale, now we

1:32:07 Speaker 3: can really see the detail pop out. So we can

1:32:09 Speaker 3: start seeing the actual trees along the fairways of the

1:32:13 Speaker 3: golf courts that are cooler in the fairways, it's pretty cool.

1:32:16 Speaker 3: You can see this vegetated waterway going down the center

1:32:20 Speaker 3: of the golf course. You can see roads, you can

1:32:23 Speaker 3: see specific freeways that are hot. So it just opens

1:32:30 Speaker 3: up all this detail to what's actually going on in

1:32:33 Speaker 3: the urban environment. On the right here you can see

1:32:36 Speaker 3: building roofs and parking lots that are distinctive and have

1:32:40 Speaker 3: distinctive temperatures from each other. And if you zoom into

1:32:43 Speaker 3: this sort of suburban area, you can actually see the

1:32:47 Speaker 3: temperature of individual trees, so this is really cool. You

1:32:51 Speaker 3: can actually see the cooling impact of trees and green

1:32:56 Speaker 3: grass yards in this environment. You can see to the

1:33:00 Speaker 3: level of whose roof is the hardest, right, so you

1:33:03 Speaker 3: can see the dark roof right here you can see

1:33:05 Speaker 3: my person. This sort of bright red spot is actually

1:33:09 Speaker 3: someone's very dark black shingle roof. And this cooler area

1:33:13 Speaker 3: right here you can see it's actually a lighter shingle color,

1:33:16 Speaker 3: which is a lighter temperature. So really cool in that

1:33:20 Speaker 3: you can really zoom into very fine scale detail on

1:33:24 Speaker 3: the temperatures.

1:33:26 Speaker 2: Let's go to another area quickly that I really like.

1:33:28 Speaker 3: I actually used to live close to this area and

1:33:30 Speaker 3: actually play on this golf course quite a lot. This

1:33:33 Speaker 3: is the Griffith Park golf course on the edge of

1:33:36 Speaker 3: Griffith Park, and it's really nice because there's sharp gradients

1:33:40 Speaker 3: and temperature from the golf course to the freeway into

1:33:42 Speaker 3: this more industrial complex zone just off the western edge

1:33:47 Speaker 3: of Glendale.

1:33:50 Speaker 2: So it's seventy meters. Again, not much detail.

1:33:54 Speaker 3: You can see the broad outlines of where the hot

1:33:56 Speaker 3: spots are, but it's really only when we go to

1:33:59 Speaker 3: the ten meter scale that you start seeing the actual

1:34:02 Speaker 3: temperature of hiking trail. So these are bare gravel trails

1:34:06 Speaker 3: that are hotter than the surrounding landscape. Again, trees lining

1:34:11 Speaker 3: the fairways. And what's pretty interesting here is let me

1:34:15 Speaker 3: go to let me change the scale a little bit here.

1:34:21 Speaker 3: This is where as I mentioned, you want to change

1:34:25 Speaker 3: the fuminium max of the data depending on where you

1:34:30 Speaker 3: want to look at to increase the contrast. So this

1:34:32 Speaker 3: looks a lot better for this area. One thing I

1:34:35 Speaker 3: wanted to point out here, note that this is the

1:34:37 Speaker 3: one thirty four Freeway running here along the northern part

1:34:40 Speaker 3: of the image. Right now, you can see how much

1:34:43 Speaker 3: hotter that freeway is. It's running east west, so it's

1:34:46 Speaker 3: directly being illuminated by the afternoon sun. There's no intervening

1:34:51 Speaker 3: trees or shade on the freeway, whereas if you look

1:34:53 Speaker 3: at the five Freeway south right next to the golf

1:34:57 Speaker 3: course here, look how much cooler it is. It's actually

1:35:01 Speaker 3: ten to fifteen degrees cooler. And if we actually zoom

1:35:04 Speaker 3: into Google Earthlier. You can see that there's actually a tree,

1:35:07 Speaker 3: very tall trees lined up on the court on this side,

1:35:11 Speaker 3: on the western side of the freeway, which provides a

1:35:14 Speaker 3: lot of shading. Of course, they have Griffith Park here

1:35:17 Speaker 3: which provides also even more shading.

1:35:18 Speaker 2: During the afternoon hours, you have a much cooler freeway.

1:35:21 Speaker 3: So it just shows the level of detail that we

1:35:24 Speaker 3: can get to at this resolution by looking at the

1:35:27 Speaker 3: actual differences in freeway temperatures, which I think is really cool. Okay,

1:35:32 Speaker 3: then the last quick thing I want to show you

1:35:34 Speaker 3: here before we actually get to exporting the data as

1:35:38 Speaker 3: actual imagery. Let's go to the Rose Bowl. This is

1:35:43 Speaker 3: actually one of the venues for the World Cup and

1:35:49 Speaker 3: the Olympics coming up, so this is a pretty important

1:35:52 Speaker 3: area that they're going to be looking at to try

1:35:54 Speaker 3: and implement heat mitigation efforts. If we look at the

1:35:57 Speaker 3: Rose Bowl here, we have these huge, massive parking lots

1:36:00 Speaker 3: right These are n actually going to be very hot surfaces,

1:36:03 Speaker 3: and this large field is actually gets very hot during

1:36:06 Speaker 3: the summer too. It's mostly dry sinest grass. So looking

1:36:10 Speaker 3: at the seventimeter data again, we can see there's some

1:36:13 Speaker 3: hotspots around.

1:36:14 Speaker 2: We can't clearly see exactly what.

1:36:16 Speaker 3: They are, but it's only when we go to the

1:36:18 Speaker 3: ten meter scale that you can see yep, these are

1:36:20 Speaker 3: these large parking lots and roads surrounding the southern part

1:36:25 Speaker 3: of the stadium, which are you know, twenty thirty degrees

1:36:30 Speaker 3: warmer than the golf cooks for example, to the north.

1:36:34 Speaker 3: And this if you can imagine, if there's airflow and

1:36:37 Speaker 3: wind coming in from the southern, southwest and southern parts

1:36:42 Speaker 3: of this image, that you're going to get a lot

1:36:44 Speaker 3: of heat invection from these hot surfaces into the stadium.

1:36:48 Speaker 3: And so this is these are the kind of areas

1:36:50 Speaker 3: that we want to target for heat intervention during upcoming events,

1:36:53 Speaker 3: but also just for the future sustainability of heat in

1:36:58 Speaker 3: the ARLA region. Okay, so what if we wanted to

1:37:03 Speaker 3: export these images. Before I did that, I wanted to

1:37:06 Speaker 3: show you another cool trick that sometimes we use, but

1:37:12 Speaker 3: not very often because physically it doesn't stop making much sense.

1:37:16 Speaker 3: But if we go into symbology again, all the way

1:37:19 Speaker 3: down to layer rendering blending mode, it's currently normal if

1:37:23 Speaker 3: we head overlay, what it's going to do is actually

1:37:25 Speaker 3: overlay the data directly on Google at the engine background

1:37:29 Speaker 3: so you can.

1:37:30 Speaker 2: It's pretty cool.

1:37:31 Speaker 3: You can see the almost the three D structure coming

1:37:35 Speaker 3: through with the temperature overlaid on top. And this is

1:37:38 Speaker 3: how we generated that paras map that I showed you earlier.

1:37:41 Speaker 3: But the one caveat here is that you can see

1:37:44 Speaker 3: the building roofs here which are white, actually come through

1:37:47 Speaker 3: as looking white, whereas they lose their actual temperature color,

1:37:51 Speaker 3: which would be the dark red color. The same with

1:37:53 Speaker 3: green spaces that come out looking green, because blue on

1:37:56 Speaker 3: green gives us kind of a green. So if you

1:37:59 Speaker 3: can deal with that, that's fine, but it just provides

1:38:01 Speaker 3: a really nice, pretty image. Okay, let's go back to

1:38:05 Speaker 3: the normal and then we're gonna the last thing we're

1:38:10 Speaker 3: gonna do today is going to show you how to

1:38:11 Speaker 3: actually export a PNG image or an image in any

1:38:15 Speaker 3: format that you like. So let's let's put the image

1:38:20 Speaker 3: in this at this scale. You can change the magnifier

1:38:23 Speaker 3: on the bottom right by the way to any scale

1:38:25 Speaker 3: you want, and also you can just use your your

1:38:27 Speaker 3: mask to zoom in an out. I'm going to leave

1:38:30 Speaker 3: it at this scale, go up to projects. We're going

1:38:34 Speaker 3: to go to a new print layout or control P.

1:38:38 Speaker 3: Let's call it LA la shop and you can you

1:38:45 Speaker 3: can name the print layout anything you wants. Really, you

1:38:49 Speaker 3: end up with this sort of blank paddock here. What

1:38:51 Speaker 3: are you going to do is add item at the top,

1:38:54 Speaker 3: add map. Okay, now you're going to get across, and

1:38:57 Speaker 3: what you want to do is click and drag across

1:39:01 Speaker 3: to fill this entire blank canvas space. And what it's

1:39:04 Speaker 3: going to do, it's going to render the map that

1:39:06 Speaker 3: you see in qjis in this region. So now we

1:39:12 Speaker 3: can do a couple of things. We can go to

1:39:14 Speaker 3: add item, add a scale bar. For example, there's a

1:39:20 Speaker 3: scale bar on the bottom left. Let's add a north

1:39:23 Speaker 3: arrow is also kind of typical, shows the north up direction.

1:39:28 Speaker 3: You can also add what's pretty important actually is a legend,

1:39:33 Speaker 3: which shows us the legend of the color scale. I

1:39:37 Speaker 3: really don't like the legends that QJS comes with, to

1:39:41 Speaker 3: be honest, So usually I actually do this kind of

1:39:45 Speaker 3: work in our or math Lab or Python. But this

1:39:47 Speaker 3: is possible. So now you see twos because we have

1:39:51 Speaker 3: two layers in our QJI. So we click on that,

1:39:53 Speaker 3: go to the bottom right here where it says legend items,

1:39:57 Speaker 3: click auto update. Let's move this up, and let's click

1:40:03 Speaker 3: on the first one here because this is the ten

1:40:04 Speaker 3: meter map, and hit this minus red button so we

1:40:07 Speaker 3: delete that one. You can also delete the actual names

1:40:11 Speaker 3: I believe if you like. But anyway, that's what you

1:40:16 Speaker 3: can produce with your map as well. And then if

1:40:18 Speaker 3: we go to layout export as image, you can then

1:40:23 Speaker 3: export your image in Typically PNG is a pretty good

1:40:28 Speaker 3: format to do it and preserves the resolution of the image.

1:40:30 Speaker 3: You can also output in tiff or jpeg and then

1:40:34 Speaker 3: click save and then export resolution you want to go

1:40:38 Speaker 3: at least three hundred dpi for looking at for importing

1:40:43 Speaker 3: these into part points or anything else, hit save and

1:40:46 Speaker 3: you'll have your image set now. The other way we

1:40:49 Speaker 3: can do this without adding all this other fancy stuff

1:40:52 Speaker 3: to the image is if we just want to directly

1:40:55 Speaker 3: export this image right here. Go to project import export

1:40:59 Speaker 3: export map to image, change the dpi three hundred dpi.

1:41:05 Speaker 3: Let's save, and again I'll just name it something and

1:41:10 Speaker 3: you can directly say just this image right here. Sometimes

1:41:15 Speaker 3: what I do is, in fact, is I crop out

1:41:17 Speaker 3: the color scale on the on the left here, and

1:41:20 Speaker 3: I saved that and included with the image, which makes

1:41:24 Speaker 3: it a little easier if you want to do something

1:41:25 Speaker 3: in QGIS.

1:41:28 Speaker 2: I think we have a little bit of time left.

1:41:31 Speaker 3: The last thing I wanted to go through here was

1:41:36 Speaker 3: how to actually do some quantitative analysis within QGI. So

1:41:40 Speaker 3: this will take another minute or two. And this is

1:41:42 Speaker 3: all laid out in the tutorial again. So I say,

1:41:45 Speaker 3: I want to see what is it? What is the

1:41:46 Speaker 3: temperature difference between this golf course and this sort of

1:41:49 Speaker 3: bear park area. So what I'm going to do is

1:41:51 Speaker 3: it's a little bit elaborate, but if I go to layer, uh,

1:41:57 Speaker 3: let's see vector and layer and layer. Actually I've really

1:42:08 Speaker 3: forgot how to do this myself, so let's actually this

1:42:11 Speaker 3: is a good idea. Let's go to the QGIS tutorial

1:42:14 Speaker 3: and let's figure out how to how to do this.

1:42:18 Speaker 2: So create zone polygons. Okay, zone and statistics.

1:42:23 Speaker 3: Yah, we go. So layer, create layer, new shape file

1:42:27 Speaker 3: layer Okay, layer create layer, new shapefile layer. Let's call

1:42:34 Speaker 3: it zones geometry type. We want to do polygon and

1:42:43 Speaker 3: then we're gonna click. Okay, go to zones again. Right

1:42:47 Speaker 3: click on zones, head toggle editing. It's this with a

1:42:50 Speaker 3: little pencil mark, and then click this sort of green

1:42:54 Speaker 3: icon which is adding a polygon feature. Looks like a

1:42:57 Speaker 3: little cutting green, and then you can start drawing your

1:43:01 Speaker 3: polygon to any shape that you wish. Let's let's draw

1:43:05 Speaker 3: the entire golf course here. We then right click go ID.

1:43:08 Speaker 3: Let's call it one, and we have our first one.

1:43:11 Speaker 3: Let's click on the second one, which is this little

1:43:14 Speaker 3: really hot zone park right click, call it a different

1:43:18 Speaker 3: ID number two. Okay, then we have our two different zones.

1:43:24 Speaker 3: Now what we can do is go into processing toolbox,

1:43:29 Speaker 3: search for zonal statistics. Double click zonal statistics.

1:43:37 Speaker 2: We have our input.

1:43:38 Speaker 3: Layer as the zones that we just created rest a layer.

1:43:41 Speaker 3: We want the ten meter data and we want to

1:43:44 Speaker 3: see let's do the standard deviation the mean that will

1:43:50 Speaker 3: be important in the mid mat Okay, so run close that.

1:43:56 Speaker 3: And if we go to zonal statistics and open a

1:44:00 Speaker 3: tribute table, Okay, there we go.

1:44:03 Speaker 2: We finally have the actual quantitative results.

1:44:05 Speaker 3: Yes, so the ID one, which is the golf course

1:44:08 Speaker 3: mean temperature of about one hundred degrees faaranheit. The part

1:44:11 Speaker 3: next to it was yep, twenty degrees hotter on average

1:44:15 Speaker 3: hundred and twenty degrees, but up to one twenty eight

1:44:17 Speaker 3: degrees farahnite. So that's that's really hot and the minimum

1:44:21 Speaker 3: gold called temperature is about eighty seven, so you can

1:44:25 Speaker 3: see some zonal stats. I usually do this kind of

1:44:28 Speaker 3: thing in our and Matt lab and packon, but I

1:44:31 Speaker 3: just wanted to show you that you can actually do

1:44:33 Speaker 3: this in directly in QGS if you really want to

1:44:38 Speaker 3: do something to remove that layer, and remove that layer

1:44:43 Speaker 3: and then see if I can. All right, and then

1:44:53 Speaker 3: if the last thing here is to double click on

1:44:56 Speaker 3: your file again, if you go to transparency, you can

1:45:00 Speaker 3: change the transparency let's say seventy percent.

1:45:03 Speaker 2: This is a really cool trick to do as well.

1:45:05 Speaker 3: And I can start seeing some of the background Google

1:45:07 Speaker 3: of imagery coming in the background, and it just gives

1:45:10 Speaker 3: it a nicer look. So that's it. That's all I'm

1:45:15 Speaker 3: going to show you today on this tutorial. Again, you

1:45:17 Speaker 3: can go through the actual documents which we'll be sharing

1:45:20 Speaker 3: that will give you the full step by step on

1:45:23 Speaker 3: how to do what I just did in the QJAS

1:45:25 Speaker 3: tutorial and the shopping tutorial.

1:45:27 Speaker 2: So thank you very much everyone.

1:45:28 Speaker 3: I hope you enjoyed that, and we'll be here for

1:45:31 Speaker 3: questions when this is finally presented.

1:45:34 Speaker 1: Thank you, Thank you Glynn for leading us through the

1:45:43 Speaker 1: downscaling Ecostress LST Demo. I will now provide a recap

1:45:50 Speaker 1: of our training. Here's a summary covering all seven learning

1:45:58 Speaker 1: objectives from our thermal remote sensing training. First, the physics

1:46:06 Speaker 1: of thermal emission all matter above absolute zero AMIDST electromagnetic radiation.

1:46:14 Speaker 1: Plank's law, Bean's displacement law, and the Stefan Boltzman law

1:46:19 Speaker 1: govern how much energy is emitted at what wavelengths, and

1:46:24 Speaker 1: how sens sensitively it responds to temperature change. Earth surfaces

1:46:32 Speaker 1: at around three hundred Calvin peak near nine point seven micrometers,

1:46:39 Speaker 1: placing land surface temperature retrieval squarely in the eight to

1:46:44 Speaker 1: fourteen micrometer thermal infrared window, a range of purely emitted

1:46:51 Speaker 1: energy with no solar contamination. Real surfaces are not perfect

1:46:59 Speaker 1: black bodies. Emissivity being less than one encodes the composition

1:47:07 Speaker 1: of and the roughness as well as the mineralogy of

1:47:11 Speaker 1: these surfaces, and a one point five percent emissivity error

1:47:16 Speaker 1: propagates to about one kelvin of LST error. The temperature

1:47:22 Speaker 1: emissivity separation algorithm or the tests algorithm used by Ecostress

1:47:28 Speaker 1: Aster and Modus twenty one simultaneously retrieves both temperature and

1:47:35 Speaker 1: spectral emissivity, making it far more accurate than split window

1:47:41 Speaker 1: approaches over bare, arid or geologically complex surfaces, where split

1:47:45 Speaker 1: window can er by three to ten Calvin. Unlike optical

1:47:53 Speaker 1: sensors such as Sentinel two, lansat in the visible and

1:47:58 Speaker 1: neur infrared, which detect reflected sunlight and go dark at night,

1:48:04 Speaker 1: thermal sensors detect emitted surface radiation around the clock. This

1:48:10 Speaker 1: enables nighttime observations of heat retention, urban cooling failures, and

1:48:15 Speaker 1: ecosystem thermal stress signals that are invisible to any reflectant.

1:48:20 Speaker 1: Spased product from AFTERS sixteen day revisit and ninety mid

1:48:27 Speaker 1: meter pixels to ecostress's diurnal sampling from the International Space

1:48:33 Speaker 1: Station and Looking Ahead to the Eagles or formerly known

1:48:39 Speaker 1: as Surface Biology Geology Mission TRISHNA and LSTM. Each mission

1:48:46 Speaker 1: trades off spatial resolution, temporal frequency, and spectral coverage differently.

1:48:52 Speaker 1: Choosing the right one depends on whether your application demands

1:48:55 Speaker 1: long term record find spatial detail or multi time of

1:48:59 Speaker 1: day sampling. Thermal data underpin not just urban heat island characterization,

1:49:09 Speaker 1: but many other applications, including evapotranspiration estimation, agricultural drought monitoring,

1:49:16 Speaker 1: ecosystem thermal stress detection, long term climate trend analysis such

1:49:21 Speaker 1: as lakewarming, expanding hot nights, and surface energy balance modeling.

1:49:27 Speaker 1: Making accurate thermal and frared retrievals consequential not just for

1:49:31 Speaker 1: remote sensing, but for climate projections, ecological conservation, and public

1:49:36 Speaker 1: health decision making. In the hands on ecostress data processing

1:49:43 Speaker 1: in R, we walked through a complete R based workflow

1:49:48 Speaker 1: accessing ecostress LST and emissivity collection to geotips via a pears,

1:49:56 Speaker 1: applying quality control, bitmasking and cloud mass converting Calvin to Celsius,

1:50:02 Speaker 1: defining areas of interest, and producing violin box plots and

1:50:07 Speaker 1: thermal gap figures demonstrating how nighttime LST at the Chatsworth

1:50:13 Speaker 1: Nature Preserve remained measurably cooler than the adjacent commercial area

1:50:17 Speaker 1: even after a full day of extreme heat during the

1:50:20 Speaker 1: September twenty twenty four LA heat wave. For the component

1:50:28 Speaker 1: of the second demo, downscaling ecostress to neighborhood scale with

1:50:34 Speaker 1: machine learning, we saw that because the physics of thermal

1:50:38 Speaker 1: emission limits spaceborne thermal and for red resolution for instance

1:50:42 Speaker 1: seventy meters for ecostress versus ten meters for Sentinel two.

1:50:48 Speaker 1: Part two introduced a random force downscaling workflow in Google

1:50:52 Speaker 1: Earth Engine that leverages high resolution sentinel to spectral indices

1:50:58 Speaker 1: DBI and DBI DWI as predictors of land service temperature,

1:51:08 Speaker 1: exploiting the strong physical relationship between service cover type and

1:51:12 Speaker 1: heat retention to sharpen ecostress imagery to ten meters and

1:51:18 Speaker 1: reveal block level urban heat patterns invisible in the native product.

1:51:29 Speaker 1: Reminder that the homework for this training opens today on

1:51:33 Speaker 1: June second. It is due two weeks later on June sixteenth,

1:51:38 Speaker 1: twenty twenty six. Upon attendance of both sessions as well

1:51:45 Speaker 1: as completion of the homework by the due date, participants

1:51:49 Speaker 1: will receive a certificate of completion for this training. I

1:51:55 Speaker 1: would like to acknowledge my fellow colleagues in the Arset

1:51:58 Speaker 1: Ecological Conservation Team who have supported the creation of his training,

1:52:02 Speaker 1: Sativa Cruz, Juan Torres Berres and Justin Fain. I would

1:52:07 Speaker 1: also like to acknowledge the RSET program staff who made

1:52:10 Speaker 1: this training possible, including Selwyn Hudson, Odoy, Maria Maravito, Suzanne Monty,

1:52:16 Speaker 1: and Melanie Follette. We encourage you to check out our

1:52:24 Speaker 1: website as well as the rset YouTube channel linked here

1:52:29 Speaker 1: in this slide. You can also join our quarterly newsletter

1:52:33 Speaker 1: to stay up to date on our latest trainings by

1:52:36 Speaker 1: sending an email with no subject to our set dash

1:52:40 Speaker 1: join at lists dot NASA dot gov and then follow

1:52:44 Speaker 1: the instructions sent in the response. The contact information for

1:52:50 Speaker 1: myself and our other guest instructor, Glenn Holly, is also

1:52:53 Speaker 1: here if you'd like to email us directly with specific

1:52:57 Speaker 1: questions about this training. Here is a list of references

1:53:04 Speaker 1: for this training. These are all great places to start

1:53:09 Speaker 1: to dive deeper into the material that we have presented.

1:53:15 Speaker 1: Thank you all for your participation in this training. We

1:53:18 Speaker 1: will now transition into the Q and A. So let's

1:53:28 Speaker 1: address a few questions now and then the remainder of

1:53:32 Speaker 1: the questions we'll address later this week. So I want

1:53:36 Speaker 1: to first just thank doctor Glenn Holly who has joined

1:53:40 Speaker 1: us as for this training. Thank you so much Glenn

1:53:44 Speaker 1: for your time and expertise. And yeah, let's dive into

1:53:50 Speaker 1: the questions. Question one asks, given the possibility of a

1:53:58 Speaker 1: moderate to strong Almino year, how could al nino interact

1:54:02 Speaker 1: with long term climate warming to increase urban heat, island risks,

1:54:05 Speaker 1: and affect glacier mass balance. In particular, how might these

1:54:09 Speaker 1: impacts differ across regions such as tropical cities, mid latitude

1:54:14 Speaker 1: urban areas, mountain glaciers, and polar ice environments. When would

1:54:19 Speaker 1: you like to answer?

1:54:23 Speaker 3: Sure? Yeah, I mean, starting off, it's a it's a

1:54:25 Speaker 3: pretty broad question and interdisciplinary, you know, covering a lot

1:54:30 Speaker 3: of different science areas and disciplines. But you know, this

1:54:39 Speaker 3: is our best answer on the fly, is that you

1:54:41 Speaker 3: know we are getting in Elmina likely this summer it's

1:54:44 Speaker 3: going to be strong noise, predicting that as a result,

1:54:50 Speaker 3: we're going to have one of the woman's years on

1:54:53 Speaker 3: record coming up twenty twenty six, twenty seven. So in general,

1:55:00 Speaker 3: you know, the all ninios has a positive impact on

1:55:03 Speaker 3: temperature anomally, so we know expect warmer temperatures overlaying on

1:55:08 Speaker 3: an already warmer climate, and that is going to ramp

1:55:11 Speaker 3: up the UHI effects that that we typically experience in

1:55:15 Speaker 3: urban areas.

1:55:17 Speaker 2: I won't go into.

1:55:18 Speaker 3: Detail about how this affects different regions. I'm actually not

1:55:21 Speaker 3: a major expert on al Nino and other events like this,

1:55:26 Speaker 3: but I know, for example, in tropical cities they do,

1:55:30 Speaker 3: they should expect less cloud cover, so you can expect

1:55:34 Speaker 3: increased solar loading, increased temperatures, drought, spress vegetation, and combined

1:55:39 Speaker 3: with higher humidity. That's it's probably going to be subtropical

1:55:43 Speaker 3: areas they're going to experience the most detrimental effects. Mid

1:55:47 Speaker 3: latitude urban areas. Think of western coast of California, for

1:55:53 Speaker 3: example in La just from experience usually results in drier

1:55:57 Speaker 3: conditions and warmer conditions. It's rainfall, higher temperatures. So in general,

1:56:03 Speaker 3: the Anino events will ramp up and increase the Uhi effect.

1:56:11 Speaker 1: Thank you. Question two asks how does this urban cooling

1:56:16 Speaker 1: effect during the day differ depending on surrounding land type,

1:56:22 Speaker 1: for example arid versus agricultural versus forested.

1:56:30 Speaker 3: Yeah, so this is actually nicely demonstrated in the earlier

1:56:33 Speaker 3: presentation of two areas Delhi and Athens, two quite different cities.

1:56:41 Speaker 3: But you know Delhi, you think of a typical Mediterranean environment,

1:56:46 Speaker 3: and what happens there during the summertime is that the

1:56:48 Speaker 3: surrounding rural areas in Delhi, it's actually mostly agricultural regions,

1:56:55 Speaker 3: have a very dry, barren sinesse vegetation and an increased

1:57:01 Speaker 3: exposure to bear soil, and so typically we see a

1:57:05 Speaker 3: cool island effect during the daytime in cities like that,

1:57:10 Speaker 3: especially in Dailhi, we see a pronounced cool island effect.

1:57:14 Speaker 3: If you have a city in a tropical region, you know,

1:57:18 Speaker 3: surrounded by a dance forest, the reverse is going to

1:57:20 Speaker 3: be treated. The forest is going to be dance vegetation.

1:57:24 Speaker 3: Typically it's going to be ten to twenty degrees cooler

1:57:27 Speaker 3: than the urban surfaces, which heat up much faster than vegetation.

1:57:32 Speaker 3: So you'll see the classic urban heat island effect, which

1:57:34 Speaker 3: most people are familiar with.

1:57:39 Speaker 1: Great, let's do two more questions. So the question three,

1:57:44 Speaker 1: why does near surface ozone concentration increase in the summer.

1:57:49 Speaker 3: Yeah, that's that's a good question that most people don't realize.

1:57:54 Speaker 3: You know what the you know physically, what is happening then,

1:57:58 Speaker 3: And it's actually a photochemical meaning that it's actual voc

1:58:04 Speaker 3: so volatile organic compounds from car exhaust, just think pollution

1:58:08 Speaker 3: in general in the city that reacts with sunlight to

1:58:12 Speaker 3: produce ozone. And so the hotter it is, the sunnier

1:58:17 Speaker 3: it is, the morizone's gonna built up especially near the

1:58:21 Speaker 3: surface if you have high pollutants in the urban environment.

1:58:29 Speaker 3: Thank you.

1:58:31 Speaker 1: Question for which will be our final question for today,

1:58:36 Speaker 1: asks can ecostress land surface temperature data be integrated with

1:58:40 Speaker 1: hydrological or watershed models to jointly analyze urban heat of

1:58:44 Speaker 1: appotranspiration runoff and the benefits of green infrastructure. Also, are

1:58:49 Speaker 1: there any NASA related application case studies that demonstrate this

1:58:53 Speaker 1: type of integrated analysis.

1:58:57 Speaker 3: Okay, yes, another pretty loaded question. What I've written down

1:59:05 Speaker 3: here just some examples that came to the top of

1:59:08 Speaker 3: my head. You know, the ecostress we do provide a

1:59:12 Speaker 3: standard about a transporation product which is derived from hydrological

1:59:18 Speaker 3: models that we actually run at h apl work into

1:59:22 Speaker 3: the details of those models, but that that standard, those

1:59:26 Speaker 3: standard data products are available and they are in fact

1:59:30 Speaker 3: fed directly into a lot of order shared orter balanced

1:59:32 Speaker 3: models as the latent heat and you can do your

1:59:35 Speaker 3: own literature search to find those studies for the In

1:59:41 Speaker 3: terms of the benefits of green infrastructure. For the LA

1:59:44 Speaker 3: twenty eight Olympics, we've actually been working with LA Committee

1:59:47 Speaker 3: and also other groups from usc U c L A

1:59:50 Speaker 3: part of the Shade LA organization to see if we

1:59:55 Speaker 3: can help you use ecostress LST data specifically to try

1:59:59 Speaker 3: and quite to by which areas would be optimal for implementing,

2:00:05 Speaker 3: for example, heat mitigation measures, so planning more trees, implementing

2:00:09 Speaker 3: more shade structures and so forth. And so we've been

2:00:11 Speaker 3: working on this machine learning model. It's actually a double

2:00:15 Speaker 3: causal machine learning model that integrates ecostress LST with a

2:00:19 Speaker 3: number of other different variables to try and estimate what

2:00:22 Speaker 3: the cooling benefits are or would be of planning more

2:00:25 Speaker 3: trees in the city. So that is work that is ongoing.

2:00:29 Speaker 3: It's actually going to be sponsored by NASA pretty soon

2:00:31 Speaker 3: and we're going to be working more closely with the city.

2:00:36 Speaker 3: Two really good papers that look at the urban bias

2:00:39 Speaker 3: FERE influenced specifically on urban heat and carbon uptake in

2:00:43 Speaker 3: general in the urban environments focused on Los Angeles are

2:00:47 Speaker 3: these two papers right here written by two jpls. I

2:00:51 Speaker 3: would craze you to go and look more detail at

2:00:54 Speaker 3: those papers for what is done there.

2:01:00 Speaker 1: I would like to again thank you doctor Glenn Hully

2:01:04 Speaker 1: for joining us for this our set training series, and

2:01:10 Speaker 1: thank you to all of our participants for joining us.

2:01:20 Speaker 1: And we will again be posting all course training materials

2:01:26 Speaker 1: and the Q and A document within a week of

2:01:29 Speaker 1: this training