NOAAs TEMPO Aerosol Products for Hourly Aerosol Monitoring over North America

0:01 Speaker 1: Okay, now let's begin with part one of this training, focusing on NOAA's TEMPO aerosol products for hourly aerosol monitoring over North America. The trainers for this part will be myself, Caterina Mogno. I'm an assistant research scientist at the University of Maryland, Baltimore County and NASA Goddard Space Flight Center. And today with me, there is Amy Huff, senior scientist at the Science and Technology Corporation at NOAA NET's aerosol science team. Our objectives for this part of the training are to be able to describe the capabilities of the TEMPO instrument and its advantages and disadvantages compared to the ADI on the GOES-R satellites.

0:47 Speaker 1: And secondly, to identify available TEMPO aerosol data products and recognize their characteristics, benefits, and limitations. This part is structured as follows. I will begin with an overview of geostationary instruments for air quality monitoring over North America, namely we will look at Tempo and Gozar ADI. Then I'll pass it on to Amy, who will go more in-depth on the new Tempo aerosol product. We will then summarize this part of the training, and finally we will open the floor.

1:17 Speaker 2: For Q & A. A bit of housekeeping for today.

1:22 Speaker 1: If you have any questions, please use the three dots in the bottom right of the training window to open the slide or Q & A option and put your question there. We'll try to answer all the questions together at the end of today's training, but the Q & A document with all the questions and our answer, along with all the slides and recordings of this training, will be posted as a document on the training website in about a week. And now let's start with an overview of geostationary instruments. As I said, we will review the two primary mission and instruments currently relevant for geostationary aerosolar quality over North America, namely TEMPO, the Tropospheric Emission Monitoring of Pollution, and GOES-R-ABI, the Geostationary Operational Environmental Satellites R-Series Advanced Baseline Imager.

2:16 Speaker 1: Okay, let's start with TEMPO. So TEMPO is really the first geostationary mission to provide hourly high resolution daylight monitoring of key trace gases and aerosols across the greater North America. The main objective of the TEMPO mission are to deliver revolutionary high quality and high resolution data to deepen our understanding of air quality impacts on health and the environment. and to improve air quality forecasts and alerts, as well as inform air quality policy and regulations. When we look at the global context, TEMPO serves as the North America component of an international geostationary constellation that tracks air quality every day light hour across the northern hemisphere.

3:04 Speaker 1: A bit of a Tempo history. Tempo was selected back in 2012 as a joint project between NASA and the Smithsonian Astrophysical Observatory. It was launched into geostationary orbit on April 7th, 2023. And the first light operation followed just a few months later on August 2023. And nominal operation really officially began on October of the same year. So the base mission of TEMPO concluded in June, 2025, but TEMPO's mission has since been extended through September, 2029, meaning that we'll continue to have these hourly air quality monitoring capabilities for a few more years.

3:46 Speaker 1: In terms of the TEMPO instrument, TEMPO is an hyperspectral spectrometer designed to capture detailed spectral information to identify trace gases and aerosol pollutants in the atmosphere. It operates in the UV to visible range from 293 to 741 nanometers with a spectral resolution of 0.6 nanometer and sampling at 0.2 nanometers. So this is fine enough to really distinguish the subtle spectral signature of different pollutants. In terms of area, TEMPO covers all of North America, from Mexico to Canada, and performs hourly daytime scans, achieving a spatial resolution of approximately 2 by 4.75 kilometers at the center of its field of regard.

4:40 Speaker 1: So the spectrum covered by TEMPO is really covered by two separate detectors. So one covering the UV range portion from about 293 to 494 nanometers, and the other one covering the visible range from about 538 to 741 nanometers. And so there is just a really small gap in between the two. I said like about roughly a bit deliberately because spectrometers like tempo tend to experience more wavelength drift over time than other type of instruments. So the exact wavelength range, the boundaries of the range can shift slightly as the instrument ages.

5:27 Speaker 1: The hyperspectral UV to visible bands that allow TEMPO to be able really to measure and identify key trace gases relevant to air quality, such as NO2, ozone, SO2, and formaldehyde, as well as aerosols. The UV bands in particular allows us to estimate aerosol layer height and enable the retrieval of the UV aerosol absorbing index. which is especially useful for identifying smoke and dust plumes. If you look now at tempo spatial resolution, the resolution is not uniform.

6:07 Speaker 2: It varies across the field of regard due to the viewing angle geometry.

6:12 Speaker 1: You can see in the image on the left showing the pixel size of Tempo on the east-west direction on the top panel and on the south-north direction in the bottom panel.

6:24 Speaker 2: The red and orange show where the Tempo pixels are the largest. This is towards the edge of the field of regard.

6:33 Speaker 1: The domain, while the blue tones near the center show the smallest high-resolution pixels. At the center of its field of regard, the pixel footprint size is about 2 by 4.8 kilometers. But as you move toward the edges, so like in places like Washington State, Alaska, the viewing angle increases and the pixel footprints grow substantially larger, up to 9 by 6 kilometers in the most oblique viewing locations. Here, for example, you can see the difference in spatial resolution across the field of regard.

7:12 Speaker 1: On the left, there is an image of Mexico City, which is near the center of TEMPO's field of regard, and where pixels of TEMPO are about 1.6 by 4.7 kilometers. This is really small enough to resolve individual neighborhoods of this large metropolitan city.

7:33 Speaker 2: On the right, there is the example of New.

7:35 Speaker 1: York City, which is farther from the center field of regard of TEMPO. And here, pixels go about to 2.4 by 5.2 kilometers. So how frequently TEMPO captures images? So TEMPO utilizes difference in scanning modes to really maximize the daylight coverage and respond to rapid environmental events. under nominal operation, which are basically the standard scans. Tempo performs hourly east-west scans of the entire domain, and it adjusts to 40-minute optimized scan when only a portion of the field of regard are sunlit.

8:18 Speaker 1: So, for example, during dawn and dusk. And also Tempo can perform a pre-sunrise scan of city lights and auroras known as twilight scans. So you can see the top image here on the right. So an example of nominal operation. This is from September 18th of this year.

8:37 Speaker 2: A standard, showing a standard full.

8:39 Speaker 1: Domain scan, you know, reporting aerosol optical depth. And the red outline marking the tempo field of regard.

8:51 Speaker 2: And you can see in the.

8:53 Speaker 1: Hour how the different, the scanning happens east to west. But Tempo has also another scanning mode, which is called special operations. So under special operations, which are high-frequency scans, Tempo can perform rapid regional scanning at 10 minutes or less to really monitor acute events like wildfires, dust storms, or volcanic eruptions, as well as to support other, for example, other NASA science missions upon request. The bottom image shows an example of this special operation mode from August 18th of last year.

9:31 Speaker 1: Notice that the scanning for that hour starts as a nominal operation, but then it stops, focusing on a small region. If you can see the advance of time on the top of the image, you can see there are really high frequency scans that allow tempo to capture this area with much higher definition on the temporal scale. So if you want to know when this special operation happens, data and times of the special operation are reported on the tempo operational logs that can.

10:12 Speaker 2: Be accessed through the link here in the slide.

10:18 Speaker 1: So now let's move on to the introduction to the GOES-R mission and the GOES-R ABI instrument. So the Geostationary Operational Environmental Satellite R series or known as GOES-R is the NOAA US largest generation of geostationary satellites that provides continuous high resolution observation of the Western hemisphere. The main objective of the GOES-R series mission are to detect and monitor environmental hazards, such as severe storms, hurricanes, wildfires, and dust storms, including space weather.

10:58 Speaker 1: Also to deliver advanced imagery and atmospheric measurements of Earth's weather, ocean, and the environment. And this imagery is aimed to support operational weather monitoring, forecasting, and early warning systems, and also to improve public safety and enable better protection of properties. So the current operational satellites in the GOES-ARS series fleet are GOES-East and GOES-West. GOES-East or GOES-19 has been operational since April 2025 and covers the Americas and the Atlantic Ocean to Africa.

11:35 Speaker 2: GOES-West on the other side, also known as.

11:37 Speaker 1: GOES-18, has been operational since January 2023 and covers the Western United States, Alaska, Hawaii, and the Pacific Ocean to New Zealand. The image on the right shows the schematic of the field of regard of GOES East and West. And also notice that the overlap in the field of regard over North America and part of South America. GOES-16 and GOES-17, that were the previous GOES East and West satellites, now they serve mainly as a primary backup.

12:14 Speaker 2: For the current GOES East and West.

12:18 Speaker 1: In a way that is enabled to ensure uninterrupted coverage if either one of the two current satellites in the fleet experiences any issue. As a quick note on the naming convention, because it was also confusing for me the first time I, you know, encountered that those are serious. So each satellite starts with a letter designation before launch. So like GOES-U. Then it becomes a number once it is in orbit. So like GOES-19. And finally takes on a position-based name like GOES-IST once it's placed into operational role.

13:02 Speaker 1: So now let's take a look at the Advanced Baseline Imager, or ABI. So ABI is the primary instrument for imaging Earth weather, oceans, and environment, and it's one of the six instruments carried on each of the GOES-R satellites. You can see it like ABI highlighted in the red box on the satellite image to the right, together with all the other instruments on board of the GOES-R satellites. In terms of instrument type, ABI is a multispectral imaging radiometer, and its primary targets are cloud, weather, fires, smoke, and aerosols.

13:41 Speaker 1: ABI covers the spectral range from visible to infrared across 16 spectral bands, and it provides coverage across different sector types, and we will go a bit more into detail in.

13:52 Speaker 2: The next slide.

13:54 Speaker 1: And it can perform by continuous daytime and nighttime scans at a few kilometers resolution every one to 10 minutes during normal operation. So in terms of spectral resolution, AVI has 16 spectral bands spanning 0.47 to 13.3 micrometers. Of the bands, two are in the visible, four in the near-infrared, and 10 in the infrared. You can see these bands laid out on the figure on the left. The top panel is showing the visible and near-infrared bands, and the bottom panel extends to the infrared. And in terms of partial resolution, the second band provides 0.5-kilometer resolution, while band 1, 3, and 5 provide 1-kilometer resolution, and the remaining bands provide 2-kilometer resolution.

14:44 Speaker 1: The multispectral nature of the bands enable for aerosol AOD retrievals, while the shortwave infrared band enables the detection of fire hotspots. AVI high-frequency imagery really enables rapid plume tracking, allowing us to track daytime smoke as well as monitor dust storms.

15:08 Speaker 2: And fire hotspots both day and night.

15:13 Speaker 1: In terms of scanning mode, as we've seen for TEMPO, also ABI use different scanning modes to balance coverage and temporal resolution. The current mode, this is mode six, known as M6, is the current default mode that scans the full disk every 10 minutes. You can see that illustrated in the figure on the right, where the outer red ring marks the full disk sweep. Within that, there is a con sector outline with a dash box, which is scanned every five minutes. And there are also two mesoscale boxes, the smaller boxes with the red arrow pointing to them.

15:52 Speaker 1: And each of those are scanned every one minute. The mesoscale domains really themselves are especially useful from an operational point of view. Each of them cover a thousand by thousand kilometer area. and they are repositionable anywhere within the disk. That really allows to rapidly target and track developing events like fires or storm and follow them along their track.

16:22 Speaker 2: Mode 4, by.

16:23 Speaker 1: Contrast, is continuous full disk only mode, scanning the entire disk every five minutes, but it is typically used mainly for calibration and testing. The colors you can see from the image across the disk, which range from blue to orange-red, represent the approximate pixel area. This is for the nominal one-kilometer band. Notice how the color shifts from blue near the center towards the red edges, meaning that the pixel size increases toward the limb of the disk as the viewing angle becomes more oblique and the resolution degrades.

17:07 Speaker 1: Okay, we've seen TEMPO and GOES-RABI separately, but let's now bring everything together with a side-by-side comparison of these two instruments. So in terms of core characteristics, TEMPO is an hyperspectral spectrometer which focuses on air quality and atmospheric chemistry operating in the UV and visible range with thousands of narrow bands at a roughly.

17:33 Speaker 2: Resolution of 2 by 4.75 kilometers. with hourly coverage over greater North America.

17:41 Speaker 1: On the other end goes ABI, is a multispectral radiometer focused on weather, hazards, and environmental monitoring, which has 16 broader bands across visible to infrared at a finer 0.5 by two kilometer resolution, sorry, from 0.5 to two kilometer resolution with a scanning from 10 to one minute frequency and it covers the full Western hemisphere. And if we translate this characteristic into air quality application, we can see that tempo strength relies in the atmospheric chemistry. So the ability to tracking trace gases and aerosol, the estimating the aerosol rate height and capturing daytime photochemical evolution of pollutants.

18:30 Speaker 1: But it is night blind and it has hourly temporal gaps and cannot really directly measure fire intensity. On the other hand, those ABI strengths lie in the ability to capture hazard onset and air transport through rapid plume tracking, 24-7 hazard monitoring via the infrared bands, and early fire and dust detection. But it lacks trace gap sensitivity, cannot track really smoke visibility at night, and can't really estimate aerosol height. So that's why we can see that those two instruments have complementary strengths and it is exactly the reason why combining TEMPO and AVI give us more complete picture of an air quality event than either instrument alone.

19:23 Speaker 1: And with that, I conclude my part. Here you can find references and additional material regarding the TEMPO and GOES AVI mission and instrument. And now I'll pass it on to Amy, who will give us more in-depth presentation about the new Tempo aerosol products.

19:47 Speaker 3: All right. Thank you, Katerina. So I'm Amy Huff, and I am going to go through this section on Tempo aerosol products. So this section covers Learning Objective 2, which is identify available Tempo aerosol products and recognize their characteristics, benefits, and limitations. So I'm also going to discuss the available ABI aerosol products in addition to the Tempo aerosol products. And I'm going to do that in the context of how Tempo enhances the existing ABI products. So what I'll do is I will discuss the Tempo and ABI synergy specifically for each type of aerosol product.

20:31 Speaker 3: And you can actually see here, I just wanted to start with an overview. So this table lists each. of the main types of aerosol products that I'm going to be discussing. So the ABI products are in white, the first two rows, and then the Tempo products are shaded with gray in the table. So those are the bottom three rows. So just note that the Tempo algorithms, I just wanted to mention this, they're in the process of being transitioned to operations. So what that means is that right now, Our group, the aerosol group at NOAA, we're running the Tempo aerosol product algorithms in research mode.

21:14 Speaker 3: And NOAA is in the process of transitioning the algorithms to operations. And that means once they are transitioned, then the Tempo aerosol products will become publicly available, and they'll be able to be downloaded from NASA Earth Data, just like the other existing Tempo trace gas products. And so we expect that process where the TEMPO products become operational to be completed by April of May of 2027. So we have, as I mentioned, we're running the algorithms in research mode. And so that means that we're able to display the TEMPO aerosol product imagery on the Aerosol Watch website.

21:58 Speaker 3: And so in part two, you'll get to see the TEMPO Aerosol Watch website and you'll be able to see the available Temple products that we have on the website. Okay, so in this section, I'm going to discuss each of the specific products in detail, and I'm also going to look, I'm going to show some examples, but we'll see more in-depth examples when we look at the smoke and dust cases in part two. And so, I don't want to go through this table because we're going to talk about most of these aspects as we go through each individual product.

22:31 Speaker 3: But I did want to mention, you'll notice that for several of both the ABI and the Tempo products, they're on what's called the ABI fixed grid. And I have just a little note here at the bottom. I just wanted to mention because most users are really familiar with the ABI fixed grid. It can cause some confusion. So this is... a grid that essentially is a map projection based on the viewing perspective of either the GOES East or GOES West satellite. So the grid is fixed, but it's irregular. So most of the time when users hear something like a file is gridded, they think it's a regular grid, so all the grid cells or the pixels have the same dimensions, but that is not the case for the ABI fixed grid.

23:18 Speaker 3: And if you actually, you look back, Katarina touched on this on slide 28, Because the ABI fixed grid, the actual grid or pixel sizes, they increase as the distance from nadir, which is essentially where the satellite is looking straight down at the Earth. So that's always going to be the equator. And then it's going to be 75 degrees west longitude for goes east and 137 degrees west longitude for goes west. So once you get away from that nadir point, the size of the ABI fixed grid pixel So I just want everyone to be aware of that as we move forward.

23:57 Speaker 2: Okay.

23:59 Speaker 3: So I'm going to discuss the tempo aerosol algorithms. And so just as a reminder, an algorithm is like a recipe or a formula. So for the aerosol products, the algorithm ingests level 1b radiances and other information. And then it calculates, or we call it retrieves, a specific level two product. So that would be, for example, aerosol optical depth, which we'll talk about today. And you might have heard it before. So just this is kind of a note on terminology. So to keep things as simple as possible for this training, when I refer to tempo aerosol products, that means the products that are coming from the NOAA algorithms that ingest the tempo level one B data.

24:49 Speaker 3: Okay, so even though those algorithms also, we just saw them on the table on the previous page, those algorithms, the TEMPO algorithms, they also inject AVI level 1b and some level 2 products. But just to keep it simple, instead of always referring to them as TEMPO AVI, we're just going to call them TEMPO aerosol algorithms and TEMPO aerosol products. Okay, so just another note I wanted to make, there's two different versions of the TEMPO level 1b radiance data. So there's a standard version and there's a near real-time version.

25:24 Speaker 3: So the TEMPO aerosol product algorithms use the near real-time version. And the reason why is because we want to have the shortest possible latency for our TEMPO aerosol products. In other words, we want them to be available as quickly as possible. So as I mentioned, the products that are in research mode right now and the latency is unfortunately can be, it's a little bit longer, it can be you know, four or six or sometimes eight hours. It really depends. But once the products become operational next spring, spring of 2027, the operational latency will be about two hours.

25:57 Speaker 3: So that's something to look forward to. And just keep that in mind. If you see the longer latency right now in aerosol watch, again, that's just because we're running the algorithms in research mode. And so things are not as fast as when they will be when they're operational. And then the last note I wanted to make a point is that The TEMPO aerosol algorithms, not only do they use the TEMPO level 1B data, some of the algorithms also ingest other TEMPO level 2 products. So the products, the aerosol products kind of build on each other.

26:31 Speaker 3: So, and I'm going to discuss all these as we go on, but just looking ahead to kind of give you an overview, the TEMPO smoke and death mask is used as an input to the TEMPO aerosol obstacle death algorithm. And the TEMPO aerosol optical depth product, in turn, is used as an input to the TEMPO estimated surface PM2. 5 outcome. Okay.

26:54 Speaker 3: So we're going to start with the qualitative aerosol product. So this is called aerosol detection product. It's abbreviated ADP. So it's a qualitative indicator of the presence of smoke or dust aerosols. It's also called the smoke and dust mask, and smoke and dust mask, that's the term I'm going to use in the training today. So you'll hear that quite a bit. So the idea with the smoke and dust mask, this is an example here on the right-hand side, these images. So this true color image here on the right is showing smoke, quite thick smoke, plumes from the Los Angeles wildfires in Southern California in January of 2025.

27:38 Speaker 3: So you can see those smoke plumes. And then on the left here, this is the smoke mask for those particular smoke plumes. And you can see the bright pink and magenta shading. It highlights and kind of draws your eye to the areas where there's smoke or, in the case of dust, where there are dust aerosols. And the aerosol detection product is based on spectral threshold tests. And they first identify absorbing aerosols, which include smoke and dust. And then additional tests will separate the aerosols from clouds and clear sky regions.

28:14 Speaker 3: An aerosol detection product, the smoke and dust mask, is available from both Tempo and from ABI. All right, so I'm going to, for each of the specific types of aerosol products that we're going to go through in this section, I'm going to give you an overview of the Tempo version of the algorithm. And that's because, of course, this training is focused on Tempo. And also the Tempo aerosol products, they're new. And furthermore, NOAA already has a lot of documentation that's been published about the ABI versions of these different algorithms.

28:52 Speaker 3: So, for example, there are what are called ATVDs, which are Algorithm Theoretical Basis Documents. And then also there's README User's Guides. that you can download and read that discuss the ABI version of the aerosol detection product algorithm. So for right now, I just want to talk about the TEMPO version. So the TEMPO aerosol detection product algorithm is what we call a hybrid algorithm. And that means it uses the level 1B data from both TEMPO and ABI. So essentially, it's using the UV and visible convolve bands from Tempo and the infrared bands from AVI.

29:35 Speaker 3: And there's a couple key products that I want to highlight. The first one is the UV absorbing aerosol index. And so in the algorithm, this is the first step. And so it's used to determine where there are absorbing aerosols, which include smoke and dusk. It can also be used on its own to highlight areas absorbing aerosols, okay? And then the UV absorbing aerosol index is used to generate the smoke and dust mask, which, as we just saw, that indicates the presence. It also gives information on the relative optical thickness or intensity, so thin versus thick smoke or dust aerosol.

30:12 Speaker 3: So this is a level T product, and the files are distributed as granules. And so on the right here, I have an example on the top. This is for just one individual tempo stance. This is showing an example of the UV-absorbing aerosol index, and you can see I've plotted the individual granules that make up the scan on the top, and then the bottom is just showing the full tempo scan, so each of those nine granules plotted over the tempo field of regard. And so the spatial resolution of the aerosol detection products are the same as the standard tempo Facial resolution, which is 2.0 by 4.75 kilometers at the center of the field of regard.

31:03 Speaker 3: All right, so this slide, actually for each of the products, I want to be able to kind of summarize the strengths and limitations of the products, so in this case, the smoke and dust masks that are being generated both by TEMPO and ABI. And so the reason why is because the two sensors, they're complementary. So, I have this, you know, the table here, and it lists kind of the details of the strength limitations for both the Tempo and then the ABI smoke and dust mask. But I'm only going to highlight kind of the main points.

31:42 Speaker 3: So, a couple of the things that I really want to highlight for smoke and dust mask is first that the primary advantage of the Tempo product is that it has the UV absorbing aerosol index. which is, as I just mentioned, that identifies smoke and dust. And so because absorbing aerosols, they have stronger spectrally dependent absorption of UV light compared to visible light. That means that the UV absorbing aerosol index, the TEMPO smoke and dust mass product, is going to be able to retrieve more smoke and dust aerosols across a larger geographic extent.

32:19 Speaker 3: Particularly optically thinner smoke that ABI misses. And that's the example I have here on the right. So on the top, I have GOES East ABI smoke and dust mask, and the bottom is the corresponding TEMPO smoke and dust mask. So this is for June 2nd of 2025. So this is when wildfire smoke from fires in Canada moved over the Eastern US. And hopefully you can see for the goes east, the ABI image, there's quite a bit of the smoke plume, these kind of gray colored smoke plumes that are being retrieved with the smoke and dust mask.

33:00 Speaker 3: So kind of just the thicker portion further north here is being retrieved, which is this kind of lavender shading. However, if you look at the corresponding tempo observation, you'll see all this kind of bright purple and magenta colors. And so that's corresponding to much more of the smoke plumes that the Tempo product is being able to retrieve compared to AVI. And again, that's by virtue of Tempo's UV spectral observation. Another kind of key difference between these two products is that Tempo, as I mentioned, gives us information.

33:39 Speaker 3: So in the case of smoke, the purple shading indicates more optically thick smoke, where the lighter pink or kind of magenta colors indicate optically thinner smoke. Whereas with ABI, we don't have that sort of information. It's just a binary flag. So it's just indicating the presence or the absence of smoke or dust. And let's see, one, I think I wanted to highlight, yeah, so I wanted to highlight one kind of benefit of the ABI product or strength of the ABI product over the Tempo product, So the ABI smoke and dust mask is much more accurate for dust detection over the ocean.

34:18 Speaker 3: Tempo tends to misclassify dust as smoke.

34:22 Speaker 2: Over the ocean.

34:23 Speaker 3: And I'll have in a couple of slides, I'll have an example of that. And then another consideration, we're going to see an example of this in part two. But another consideration, a kind of a strength of the Tempo product is that the ABI smoke and dust mask is not recommended at high goes east or goes west satellite view zenith angle, so view zenith angle greater than 60 degrees. We don't recommend using the AVI smoke or dust mask. And so what that translates into for a user is that the Tempo product will have a lot more geographic coverage basically in Canada and further in the western U.S. than the AVI product will.

35:10 Speaker 3: And so one other thing I wanted to highlight, it's kind of hidden here. Another advantage of having the UV spectrum range of TEMPO is that we have what's called a smoke callback, meaning that because the UV absorbing aerosol index can retrieve aerosols, absorbing aerosols that are on top of clouds, we can actually bring that back in the TEMPO smoke and dust products. So you'll see an example of this in part two. that with the Tempo Smoke and Dust Mask, we can see smoke that's kind of along, adjacent, and over clouds, whereas we don't have that capability with the ABI Smoke and Dust Mask product.

35:51 Speaker 3: Okay, so it's very important to note that most aerosol products, including Smoke and Dust Mask, have in their data files something called data confidence flags, or sometimes they're called quality flags. So these flags express the overall confidence in the retrieval of the smoke or dust. And so you really want to, if you're using the products, I mean, this is more kind of more important when you're working with the data files themselves. You just want to make sure that you're processing the aerosol products and using only the data quality or data confidence that's appropriate for your specific applications.

36:35 Speaker 3: And so we have the recommendations here for both the Tempo and ABI smoke and dust mass products. We recommend if you have a qualitative application, you can use all the confidences, high, medium, and low. But if you have a quantitative application where you really need to have higher confidence, we want you to use the high and medium confidence flag data. And then I just have some examples here on the left. These are ABI smoke and dust. confidence plots, but the tempo are very similar. So you can see in the top is the smoke confidence, and the dust is on the bottom.

37:13 Speaker 3: And for each, there's high, medium, and low confidence. And then also just be aware the smoke and dust mask is not retrieved over areas with clouds or snow or ice. I mentioned the smoke over clouds callback. That's just for the tempo product. And there's no smoke or dust mask available at nighttime. And so I just want to mention, because we'll come back to it multiple times today, is that clouds are an unfavorable condition for most satellite aerosol products, including smoke and dust baths. So that means anywhere there are clouds, we can't make a retrieval.

37:52 Speaker 3: And so a big part of most satellite aerosol product algorithms is identifying where the clouds are. And so one of the ways we do this in the NOAA algorithms is we use the ABI Cloud Mask product that essentially is a flag that tells us if a PICSO is mostly cloudy, partially cloudy, or cloud-free. So that'll become important as we continue. All right, so just to highlight a few things when you're using the Smoke and Dust Mask product, both the ABI and the Tempo Smoke and Dust Mask product, These are known issues that I just want people to be aware of, especially in part two when we start looking at the products for specific smoke and dust events.

38:41 Speaker 3: The smoke and dust mask is most accurate for thick smoke and dust, optically thick, over dark surfaces. Dark surfaces are things like water or vegetation. The products are less accurate over bright surfaces. So bright surfaces are arid and semi-arid regions, excuse me, like deserts or most of the western U.S. And so the second one, this is coming back to what I was just mentioning about the upstream, the ABI cloud mask, which is ingested into the smoke and dust, the aerosol algorithms. And the ABI Cloud Mask is used to identify where there are clouds because there are no smoke and dust retrievals there.

39:28 Speaker 3: But the problem is that for very thick smoke plumes and to a certain extent also very thick dust plumes, the cloud mask product itself identifies those plumes as clouds. And so even though they may be, for example, thick dust or thick smoke, The cloud mask says there are clouds, so there's no retrieval done there at all, even tribes detected. So you'll see that sometimes there are gaps of the smoke and death mask retrieval. And then the last one I wanted to highlight is under certain conditions, so particularly in the early morning and the late afternoon when the solar zenith angle is large, and areas where also the satellite view zenith angle is large, you'll see false smoke and dust detection.

40:21 Speaker 3: So I have examples from the AVI on the right, and then on the lower left is Tempo. So particularly with the AVI in the mornings, you'll see false smoke detection, morning and afternoon, but false smoke detection over clouds. You'll see false dust detection over bright surfaces, like the bottom right. And then tempo, you can get both. Again, when we have the areas that have the high satellite viewing angle. This example is from Maine and part of the Canadian Maritime.

40:56 Speaker 2: Okay, now the next slide.

40:58 Speaker 3: So here's just a couple of known issues that are specific to the Tempo product. So very bright surfaces. can have false smoke detections. And so this is the example on the upper right. So this is the White Sands, the arrows pointing to the White Sands Missile Range in southern New Mexico. It's very, very bright and reflective. And so you'll see that purple shading, that's a false smoke detection. Also, something similar can occur for thin dust plumes over vegetated surfaces. Those will be misclassified as smoke.

41:35 Speaker 3: And then the bottom example figure is showing for the Tempo product, Saharan dust over the ocean is often misclassified as smoke. And I just want you to be aware, this is actually a strength of the ABI product. So the ABI dust mask typically has very high accuracy for dust over the ocean, in contrast to the Tempo product.

42:04 Speaker 2: Okay.

42:04 Speaker 3: So now let's move on to the quantitative aerosol product. So this is just to review. Probably most people I would imagine taking this class or this training are familiar with aerosol obstacle depth, whereas aerosol detection is not quite as well-known of a product. But aerosol obstacle depth is a quantitative measure of aerosol loading. So mathematically, it's defined as the integrated extinction coefficient over a vertical column of unit cross-section. And so as a result, it's unitless. And values typically range from zero to one in North America, although the NOAA aerosol optical depth products can go, the AOE can go as high as five.

42:48 Speaker 3: And so here I have images. It's the same example that we saw for aerosol detection. So this is the January 9th, 2025 Los Angeles wildfire smoke. And so you can see on the right the aerosol obstacle depths, the very high values of AOD corresponding to the teal, yellow, orange, red, and dark red shading. So essentially AOD anywhere from about 0.6 or higher. So the larger values of AOD correspond to high concentrations of aerosol. And where we see, for example, the blue shading in this image, those are much lower values of AOD.

43:27 Speaker 3: that corresponds to either clean or clear atmosphere. So we mentioned that this is a column measurement, so it's important to remember there's no information from AOD about the altitude or the height of the aerosols in the atmosphere. And AOD, among other things, is useful for air quality applications because surface PM2. 5 concentrations can be estimated from aerosol optimal depth, and we'll see how that works in a bit. All right, so in terms of the TEMPO aerosol optical depth, aerosol layer height algorithm, so this is a combined algorithm.

44:06 Speaker 3: And again, I'm just going to focus on an overview of the TEMPO algorithm because NOAA has the ATBDs and the README guides, a lot of different existing documentation for the ADI AOD products.

44:22 Speaker 2: All right, so.

44:23 Speaker 3: In the TEMPO algorithm, it gives us both aerosol optical depth, which is reported for 550 nanometers, and then also the aerosol layer height. So the aerosol layer height is a new product. It's not available from the ABI version of the AOD product. And the aerosol layer height is the estimated altitude of the aerosols, and that's the aerosol layer that has the most extinctions. It's the estimated altitude of the aerosols in the troposphere, and that height or altitude is relative to the surface, not to sea level.

45:02 Speaker 3: And the algorithm uses a ratio of tempos convolved of 670 and 688 nanometer bands. So the ABI has a 640 nanometer band, but it doesn't have the 670 or 688 nanometer bands. That's why we can't get the aerosol layer height from the ABI. And also the last point there, the aerosol layer height is only retrieved when aerosols have an aerosol optical depth greater than or equal to two. In other words, we have to have some aerosols there in order to be able to calculate the aerosol layer height. And so in terms of the specifications for the product, they're the same as we saw with the aerosol detection products.

45:50 Speaker 3: The files are their level two products, so distributed as granules. And I have an example here of the individual granules at the top and the bottom, all the granules plotted for the full scan. All right. So, again, focusing on a summary of the strengths and limitations of the TEMPO and AVI aerosol obstacle depth products. The primary advantage of the TEMPO aerosol optical depth is that the algorithm, it ingests the TEMPO smoke and dust mask. So if you remember from that, I think it was the third slide I showed, the TEMPO AOD algorithm, it actually uses the smoke and dust mask to differentiate the aerosol type.

46:43 Speaker 3: So in other words, it improves the retrieval of smoke or dust because the smoke dust mask is actually telling the TEMPO AOD algorithm what the aerosol type is. And so that's really important, particularly for dust AOD retrievals because the, and we'll see an example of this in part two, the ABI AOD product tends to underestimate dust retrievals and sometimes misses them completely. particularly over land. So being able to accurately retrieve dust AOD is a strength of the TEMPO AOD product. Another point I wanted to make is that the strength, one of the strengths of the ABI AOD product is that it has frequent observations compared to TEMPO.

47:41 Speaker 3: So in one hour, basically in the time it takes TEMPO to make one observation, which is once per hour approximately, ABI AOD can make 12 observations. And so the fact that the ABI AOD has more frequent observations, we can make what are called AOD composites. So essentially this is an average of the AOD from ABI over a certain time period. And on the Aerosol Watch website, we have three hour ABI AOD composites. And I showed an example So this was for wildfires in the Pacific Northwest of the U.S. in June of this year, June 17th, 2026.

48:26 Speaker 3: And the top is a three-hour ABI composite. And you can see how a lot of the retrievals of those individual smoke plumes are filled in, whereas we have the corresponding tempo AUD observation. And hopefully you can see there's a lot of gap. A lot of the smoke plumes, the AOD isn't being retrieved. And again, that's due to the fact that the upstream cloud mask is indicating those thick smoke plumes are clouds, not smoke. But that issue, because of the much more frequent observations of the AVI AOD, we can smooth out some of those missing retrievals, both from the cloud mask issue and then actually from clouds themselves with the AVI process.

49:15 Speaker 3: Okay, all right, so just like we saw with the aerosol detection product, AOD product, the files include data quality flags, which express the overall confidence in the AOD retrievals. And you just want to make sure, particularly for AOD, if you ever are working with the data files themselves, that you use the correct data quality for your particular application. The recommendations for AOD are slightly different than they are for aerosol detection products. So for AOD, for tempo, you want to use high quality AOD for all applications.

49:54 Speaker 3: And for ABI, for qualitative applications, you want to use what we call the top two qualities, so that's high and medium quality. And then for quantitative applications, you want to use high quality only. And the reason why is because, illustrated here in these figures. So this is just examples of the same AOD retrievals, just showing the different data quality. So in the top here, we have high quality AOD. This is back in, when was this? 2020, December 2020. These were wildfires that were occurring just outside of Los Angeles, and the smoke plume extended out over the Pacific Ocean.

50:37 Speaker 3: And so you can see with the high quality, there are a lot of gaps, particularly along the coastlines, and then some of the actual smoke plume is missing. And then if we look at the high and medium quality here on the bottom, you can see much more of that smoke plume is filled in. There's still a few gaps, but most of the smoke plume over the water is complete. And there still are gaps along the coastlines, but they're less than they are compared with the high quality data. And then on the far right here, if we look at the all quality, so high plus medium plus low, we see that, you know, the good news is that all of that smoke plume is filled in and a lot of the AOD along the coastlines is filled in.

51:24 Speaker 3: But the drawback is you can see there's a lot of these anomalous red areas of high AOD. And these are anomalous. They don't actually correspond to aerosols. And so that's why we recommend that users avoid using low-quality AOD, both Tempo and ABI AOD, just because there's so much uncertainty in the low-quality retrieval. All right. And again, AOD isn't retrieved just like the aerosol detection product at nighttime.

51:56 Speaker 2: We need visible light.

51:58 Speaker 3: In addition, AOD isn't retrieved over areas, of course, with clouds, snow or ice, ephemeral water over land or shallow ocean waters, as well as areas with sun glints. And I should mention, I think I did talk about it in part two, but on the Aerosol Watch website, so the Tempo Aerosol Obstacle Depth, we use the qualitative recommendation. So the Tempo AOD is high quality. The ABI AOD is high plus medium, top two qualities. All right, so user awareness. There's a few known issues for TEMPO and ABI aerosol optical depth products.

52:42 Speaker 3: So the first one we've mentioned a couple of times, but it's just worth remembering because this is a pretty well-known issue with any satellite aerosol optical depth retrievals, the gaps that we get in thick smoke plumes in particular. because the upstream cloud mask, and by upstream, again, I mean the cloud mask that's ingested into the aerosol algorithms that are used to identify, one of the ways that are used to identify where clouds are, the cloud mask often mistakenly identifies thick smoke as clouds.

53:16 Speaker 3: So we don't have any AOD retrievals there. And you can see the examples here I have From the GOES-West ADI, you can see some of that gray shading coming through where there's no AOD retrieval, both for ADI and then similarly for TEMPO. Another issue that I don't have an image of, but this is something to be aware of, is something that's called cloud contamination. So basically what that means is that there can be anomalously high AOD retrievals adjacent to clouds, and this occurs for both TEMPO and ADI AOD retrievals.

53:51 Speaker 3: So, the AOD can look a little bit noisy. You just have to be aware that those are essentially false detections. So, again, it's important to stress that the AOD products that are vertical column retrievals, they by themselves do not give any information about where the aerosols are in the atmosphere. And we have the new Tempo Aerosol Layer height product that we can use. that we'll look at in a second that will give us information about where vertically the aerosols are.

54:21 Speaker 2: In the atmosphere.

54:22 Speaker 3: And then the last item I want to mention, it corresponds with this image here on the lower left, just for the AVI-only product. There periodically can be missing blocks of AOD, and this just happens randomly. This is something to be aware of if you're using this product and you see kind of these rectangles where there's missing AOD. That's just an anomaly that has to do with when the files were generated. Okay, so I mentioned the aerosol layer height product. Let's just look at an example. We're very excited about this.

54:57 Speaker 3: It's something that users have been asking for for a long time. So this is an example from June 1st of 2025. And this is when

55:08 Speaker 3: smoke from wildfires in Canada moved over the eastern U.S. So on the left here, we have the tempo aerosol obstacle depth retrieval for scan 9, which was from 1714 to 1807 UTC, so in the afternoon. And on the right, we have the corresponding aerosol layer height for that same scan. And so if we take a look, again, if we're just looking at the aerosol obstacle depth, we see that red, dark red shading that I've circled. So areas where there's very high aerosol obstacle depth indicating there's high concentrations of aerosols, or in this case, particles.

55:49 Speaker 3: But then if we now use the aerosol obstacle depth in conjunction with the aerosol layer height, we see that the aerosol layer height is telling us that that high AOD in the Midwest over, the Dakotas and Nebraska and part of Minnesota, that kind of teal-blue shading is indicating that those aerosols, that smoke, is in the boundary layer, so it's close to the surface. And we would infer that the smoke there is impacting surface air quality. In contrast, if we look over the mid-Atlantic region here, we see this kind of pink shading in the aerosol layer height.

56:31 Speaker 3: And that is indicating that those aerosols, even though there's high concentrations, high AOD, those aerosols are aloft. They're above the boundary layer. So we would not expect them to be impacting surface air quality in the mid-Atlantic region. So now if we actually check, so this plot now on the left is showing the corresponding one-hour average PMT. 5 concentrations from the air now regulatory monitors. And so we can see that, in fact, yes, in the upper Midwest, where we see the kind of teal shading in the aerosol layer height, we see the orange and red dots that indicate high concentrations of surface PM2.

57:14 Speaker 3: 5. And then in the mid-Atlantic region, where the aerosol layer height is telling us that smoke is a loss above the boundary layer, we're seeing, indeed, these green colored dots indicating low concentrations of surface PM2. 5. All right, so moving on now, our last product that I wanted to talk about is estimated surface PM2. 5. So there are both TEMPO and ABI PM2. 5 algorithms, and they both use what's called a geographically weighted regression or GWR model. And so the GWR uses satellite aerosol optical depth and surface PMT.

58:01 Speaker 3: 5 concentrations from the AirNow regulatory monitor network, and we just saw an example of that in the previous slide. And so the GWR, what it does is it captures the temporal and spatial variation in a relationship between AOD and PMT. 5. So essentially the model parameters, the geographically weighted regression model parameters, They're updated every hour when we get new observations of AOD and air now PM 2.5 as well. All right. So this, I just kind of wanted to talk about how the tempo estimated surface PM 2.5 values are calculated in the algorithm.

58:46 Speaker 3: So this figure at the bottom is kind of a schematic of how that process happens. All right. So there are two separate PMT. 5 estimates that are generated in the algorithm. So AVI-AOD is used in the geographically weighted regression model along with AirNow PMT. 5 to come up with with one estimate of PMT. 5. And then separately, Tempo Aerosol Obstacle Depth and Tempo Aerosol Layer Height are used to make, along with the AirNow PMT. 5 from the surface monitors,

59:28 Speaker 3: those are used to make a separate estimate of PMT. 5. And the advantage of the Tempo product is that Tempo, because we have the aerosol layer height, we can use that to make what's called an estimate of the boundary layer AOD. And so how that works is that we assume that the aerosol extinction profile is semi-Gaussian. And then the algorithm integrates the lowest one kilometer of that extinction profile, and that's assumed to be the aerosols in the boundary layer. And then also separately, the entire column is integrated.

1:00:11 Speaker 3: And then the ratio between these two values, the lowest one kilometer in the entire column, that's assumed to be the fraction of A of D in the boundary layer. Okay, and then we multiply the tempo AOD value by this ratio, this fraction, and that gives us the boundary layer AOD. And it's this boundary layer AOD that's used from tempo in the geographically weighted regression algorithm. All right, so then we'll have the estimate from the AVI and estimate from tempo, and then those two estimates are merged.

1:00:45 Speaker 3: into giving us our final estimated PM2. 5 product that's outputted from the algorithm.

1:00:52 Speaker 2: All right.

1:00:54 Speaker 3: So just a couple of details. So the surface PM2. 5, the estimated PM2. 5 is concentration in units of micrograms per cubic meter. And the one hour average, what that means, it's an average over a clock hour, meaning, for example, the 13 UTC observations would be the average of surface PM 2.5 from 13 to 1359 UTC. Okay, so it can be a little confusing because tempo scans typically will encompass, you know, those start in the middle of one hour and move to the next hour. But for service PM 2.5, we're doing it by the clock hour.

1:01:38 Speaker 3: Okay. So this is a level four product. It uses the ABI fixed grid, as I mentioned at the beginning of the presentation. And the domain is a little bit different from the Tempo field of regard. It's a little bit larger. So for this product, the domain includes the CONUS, Southern Canada, Northern Mexico, and the Caribbean. And those asterisks there are to remind me to let you know that for Mexico and the Caribbean, There aren't any air now monitor data that are inputted into the geographically weighted regression.

1:02:16 Speaker 3: So the accuracy of the estimated surface PMC-15 is less accurate in the Mexico and Caribbean. There's more uncertainty there. So when you're using the product, just keep that in mind. You just have to, you know, realize there's a little bit more uncertainty in those areas. And the spatial resolution of the product matches the spatial resolution of the ABI fixed grid, which is two kilometers in nature. And I also want to mention, so here, you know, we're talking about the Tempo PMC-15 product. There is also an ABI only, so essentially just this left part of the figure here.

1:02:53 Speaker 3: That product is also available, but it's not available publicly. It's on the AirNow Tech website. So, that's used by U.S. operational air quality forecasters and air quality managers. It's through a partnership with the U.S.

1:03:10 Speaker 2: EPA.

1:03:10 Speaker 3: So, I just wanted to let you know that product does exist. It's just not something that's publicly available.

1:03:17 Speaker 4: Okay.

1:03:19 Speaker 3: So, again, highlighting the synergy of the estimated surface PMC-55, both from PEMPO and ABI, looking at the strengths and limitations of both. So the strengths of the TEMPO product come from two places. So one, the strength comes from the fact that the TEMPO algorithm includes aerosol layer height, as we saw in the previous slide, and it's used to estimate AOD in the boundary layer. And so that increases the accuracy relative to the standalone AVI PM2. 5 product. particularly when aerosols are aloft.

1:03:57 Speaker 3: So they're greater than three kilometers aloft in the atmosphere. So in contrast, the ABI PMT. 5 product just relies on the AOD retrievals. It doesn't have the ALH. So it's not as accurate when aerosols are aloft. And then the other strength of the Tempo product is that if you recall, the TEMPO AOD product uses the TEMPO smoke and dust mask to identify or differentiate the aerosol type, which in turn means that the TEMPO AOD retrievals are more accurate and they have greater spatial coverage for dust, whereas the ABI AOD retrievals for dust tend to be underestimated or missed completely.

1:04:47 Speaker 3: The kind of the takeaway from that is that if you remember, we have the separate PM 2.5 estimates from AOD and from TEMPO and then they're merged. Well, the only time they're not merged, okay, or I should say only cases when they're not merged, when the TEMPO PM 2.5 estimate is always used is when the aerosol is aloft in the atmosphere. or the aerosol is identified as dust. And that's, again, just because the Tempo PM2. 05 product is much more accurate in those situations. And I just have an example of this here on the left.

1:05:22 Speaker 3: So this is for the dust situation. So this is a case from March 14th of 2025. So here on the top is the ABI geocolor, which is essentially a true color during the daytime. So you can kind of see the dust. And that's overlaid on the air now observed one-hour average PM2. 5 concentrations. You can see all those kind of red and orange dots corresponding to high surface PM2. 5 concentrations from the dust. And so on the bottom, we have on the left the tempo AOD retrieval. You can see they're much more accurate.

1:05:59 Speaker 3: So the tempo AOD is retrieving high AOD corresponding to that dust. Whereas on the right, you can see not only are there kind of more gaps in the ABI AOD retrievals, but the blue shading indicates that the AOD that is retrieved is quite low. It's very underestimated. And so again, that's why we rely on the tempo PM 2.5 estimates in the situation when we have dust. Okay. Yes, so again, user awareness. There's a few issues that we just want people to be aware of when they're using the one-hour estimated surface PM2.

1:06:43 Speaker 3: 5, both from Tempo and from ABI. So the first one is kind of the advantage of the product, is that kind of the whole point behind this product is that it helps fill gaps in the regulatory PM2. 5 monitor network. The monitors tend to be or located in urban and suburban areas, and there tend to be gaps in more rural areas. But oftentimes those rural or remote areas are where wildfires or dust storms can originate. So we have an example here. This is from May 31st of 2025. And again, this was some wildfire smoke transport into the upper Midwest.

1:07:31 Speaker 3: So I'm just focusing on North Dakota. But in Eastern North Dakota, there aren't any regulatory PM2. 5 monitors. So for this particular event, when the smoke was really impacting surface air quality, there wasn't any surface information from monitors. However, the estimated tempo PM2. 5 product gives us information, gives the public, it gives the emergency managers, it gives the air quality forecasters information about the surface air quality conditions in that region where there's no monitors. Another thing to note, there's no data quality or data competence flags with this product, so you don't have to worry about that sort of processing.

1:08:17 Speaker 3: You can just use it as is, so that's nice. And then kind of the last thing, just be aware, if you remember I mentioned that AOD retrieval is both from Tempo and ABI. They can be noisy due to cloud contaminations, well, that noise can be carried through into the estimated PM2. 5 product, and that will in turn make the PM2. 5 noisy and less accurate. And so I have an example here on the bottom left, just highlighting this was for another transported smoke event from June of 2026. And so you can see kind of all those like very pixelated orange and yellow and dark red PMT.

1:09:03 Speaker 3: 5 that's associated with cloud contamination.

1:09:07 Speaker 2: Of the AOD.

1:09:08 Speaker 3: And this issue tends to be worse for the Tempo product, the Tempo estimated PMT. 5 than it does for the ADI standalone estimated PMT. All right, so summing up, I just wanted to highlight a few things when you are thinking about do I want to use the ABI or the Tempo aerosol products for a particular event. So just remember there's differences in the frequency of observations, the latency, in other words, how soon are the data available after the actual satellite measurements, There's differences in spatial resolution.

1:09:52 Speaker 3: There's differences in the fact that the GOES-R satellites are operational missions, meaning there's 24-7 support, 24-7, 365 support for the satellites and the flow of data. So the products are almost always there when you need them. In contrast, Tempo is a research mission. So there will be and there are periodic interruptions in the near real-time observations in the data flow. Sometimes the products are missing. There's also, Katarina mentioned, the special operation scans. You know, where Temple will focus on just one region for, you know, multiple scans.

1:10:36 Speaker 3: So if you are not in a part of the field of regard where that special operation is happening, then there will be no tempo observations for your area. So the aerosol layer height, that's another big difference between the two. So of course with AVI, we don't have the required spectral bands for aerosol layer height, but we do for tempo and it's a great new product that users are really excited about. And then the last thing, just keep in mind that Tempo aerosol products, they're brand new, really, essentially.

1:11:11 Speaker 3: Their performance is still being evaluated. You know, every day we're looking at the products, we're making updates to the algorithms, whereas in contrast to EBI aerosol products, they have a long record of known proven performance. They've gone through validation processes at NOAA, that the Tempo products will go through those, but they just haven't yet because they're still relative. All right, and then the final slide, I just want to hopefully you'll take away, if nothing else, from this training that Tempo enhances the existing legacy ABI aerosol products.

1:11:46 Speaker 3: So the addition of Tempo's UV visible hyperspectral observations, they enhance the aerosol products from ABI. So we have new products that you saw, the UV absorbing aerosol index, the aerosol layer height, And then we have improved products as a result of Tempo. So the smoke dust mask, the AOD, the Surface PM 2.5. And I encourage you as users to use all the available geostationary products, both from Tempo and ABI, to get a full picture of an event that you're analyzing. And the reason why is because, as I mentioned on the last slide, the aerosol products, They're available at different latencies, different frequencies, different spatial resolutions.

1:12:35 Speaker 3: Each type of product has its strengths and limitations. And then as we saw, there can be, under certain situations, certain times of day, anomalous retrievals that can occur. So you always want to be looking at the various different available products and not relying just on one specific product. And here I just have examples. This is one of the cases that we look at in part two, but this is a dust event. And so we have the smoke dust mask, the UV absorbing aerosol index, aerosol optical depth, aerosol layer height, surface PM 2.5, and then also the AVI dust RGB composite.

1:13:15 Speaker 3: So for this particular case, there's all these different products that you can look at to get a really full picture of what's happening. And then the last slide, these are just here for a reference. References are for a reference. These are some links to some of the different documentation for the various products that we talked about. And I just want to call out the very last one. We have prepared a science team user's guide for the different aerosol products. And they're pretty short and succinct.

1:13:45 Speaker 3: So if you're interested in more details beyond what's in this training about the users, about the products, feel free to take a look at those. And you could always, of course, contact me if you have additional questions.

1:13:59 Speaker 1: Thank you, Amy, for this presentation. And now let's just summarize this first part of the training. So in summary, what we've learned from part one of this training, We have seen that TEMPO and GOES-ABI are the two geostationary instruments that provide air quality information over North America. When we compared those two instruments, TEMPO is an hyperspectral spectrometer covering the UV to visible wavelengths and provides information of trace gases and aerosol at a hourly resolution.

1:14:36 Speaker 2: On the other hand, and like ABI.

1:14:38 Speaker 1: Is a multi-channel visible to infrared radiometer, that for the quality application, it's really able to track smoke and dust sub-hourly to a minute scale. ABI, it's able to monitor rapid transport, rapid onset and transport of pollution event, while tempo reveals their chemical composition.

1:15:06 Speaker 2: ABI products are frequent.

1:15:08 Speaker 1: In the near real time, but lack the aerosol vertical structure. On the other hand, Tempo aerosol products deliver vertical structure and near-surface air quality information with two to three-hour latency when operational. So that's why it's important really to review all the available geostationary products, both from Tempo and ADI, to get a full picture of an air quality lens.

1:15:37 Speaker 1: And now as I'm looking ahead for the next part of the training, in part two, we will use the NOAA Aerosol Watch website to view near real time and archive Tempo and ABI aerosol product imagery.

1:15:51 Speaker 2: And then we will analyze Tempo.

1:15:53 Speaker 1: And ABI products and imagery on the Aerosol Watch website, going through a recent wildfire smoke event and a blowing dust event as case studies. And with that, thank you for attending this part one of the training, and we are now transitioning to the Q & A portion of this session. All right, as we transition to the Q & A portion, we're being a bit, few minutes over time for the presentation part. So probably we'll go around 10 minutes more for Q & A. But don't worry, as I mentioned at the beginning of the training, we have this document where we'll try to answer, where we have collected all like the question you had during this part, and we will try to answer.

1:16:53 Speaker 1: as much as possible in the next 15 minutes or so. But don't worry, all the questions that we cannot answer today, they are recorded in the document. We will answer them offline and all the questions with all the answers will be posted in the training webpage as soon as we answer all the questions. So I think we can start with the first question is, does TEMPO cover both day and night data for America. So TEMPO collects aerosol data during daylight hours because it is a passive sensor that measures the scattered UV and visible wavelength of sunlight.

1:17:35 Speaker 1: So it requires daylight to operate. So if you want to know more distinction between the passive and active sensor, this is covered in the recommended prerequisite for this course, the fundamental of remote sensing training. To mention, there are some certain Tempo Twilight data products that can use some emitted light during nighttime, but these are not related to air quality applications, so no aerosol or trace gases. Question number two, why will Tempo end in a few years? So tempo as all like a satellite mission, they have a lifespan.

1:18:18 Speaker 1: And so they're like ability to stay in orbit and collect data, need to go to a periodic review.

1:18:25 Speaker 2: So it has.

1:18:27 Speaker 1: Gone like review, I think in the last year and, you know, that has been extended until 2029. So this is a typical NASA mission review cycle process. Um, so technically the instrument is built to, uh, you know, the operational for a 15 year lifespan. Um, as long as like, uh, from another side, it continues to be a in operation via the review process. Question number three, do we still use TEMPO for Alaska data readings, given its larger pixel, and do we utilize other satellites for this region? Um, I think I can defer this question like, um, um, to Adam Nagel, who is part of our, um, um, but also is the tempo mission application leads.

1:19:18 Speaker 1: So he has a bit more, um, of like, uh, insight for, for this one. I don't know if you, uh, wanted to unmute and answer this question.

1:19:31 Speaker 2: Um, we can hear you. Okay.

1:19:36 Speaker 5: Yeah, so we go through our typical NASA mission.

1:19:41 Speaker 3: Reviews every few years.

1:19:43 Speaker 5: And, you know, Tempo has a lifespan and will be up there 15 years in space.

1:19:50 Speaker 3: So, yeah, we were.

1:19:52 Speaker 5: Recently actually reviewed and we had a very successful mission review and we're now extended beyond. We're now extended until September.

1:20:01 Speaker 4: 1, 2009, but.

1:20:08 Speaker 3: On our future and as the initiative.

1:20:10 Speaker 2: So keep that in mind. Yeah, okay, thanks.

1:20:14 Speaker 1: And also like in terms of like using the TEMPO data in the field of regards in Alaska.

1:20:25 Speaker 5: Oh, to me again, okay.

1:20:28 Speaker 3: Yeah, so TEMPO.

1:20:34 Speaker 5: Or Alaska data reading given the larger PIPs Yeah, I don't, again, we have limited coverage over Alaska. Tempo field regard doesn't cover the entire state of Alaska.

1:20:49 Speaker 3: We do have coverage over there.

1:20:52 Speaker 5: I haven't talked to many folks using data over Alaska from Tempo, but we do have forcer spatial resolution, but that being said, the data should.

1:21:06 Speaker 3: Good, as long.

1:21:07 Speaker 5: As, again, we're not in stratified flower cover, snow, ice.

1:21:13 Speaker 2: But, yeah, definitely, I don't, I would.

1:21:15 Speaker 5: Recommend central data here, so we'll see how well that's paired where possible.

1:21:22 Speaker 1: Yeah, and also, like, you know, there are other satellites, you know, polar orbiting satellites, like NOAA-21 or NOAA-20, and VIIRS that can you, you know, other satellites that can be used over that region as well.

1:21:37 Speaker 4: Hey, Katarina, this is Amy. I don't know, do you want me to show on Aerosol Watch how to access the VIIRS data?

1:21:47 Speaker 1: I think we can, like, since, like, you know, more coverage of the access of the data products is the.

1:21:55 Speaker 2: Next session, I'm happy, you know, maybe to do, yeah.

1:21:59 Speaker 1: Like we mentioned at the end, like the next section of the training will be more like how to access all this data we've talked about today. So I think we keep in mind and then it's something that maybe in the Q & A of that session we can address. Okay, question number four. how we can determine whether an apparent rapid increase in aerosol optical depth represent a genuine increase in aerosol loading rather than a change in aerosol vertical distribution, relative humidity driven particle growth or change in cloud contamination and how we can disentangle these effects when multiple processes occur simultaneously within the same pixel.

1:22:44 Speaker 1: I don't know, Amy, do you wanna take this question?

1:22:48 Speaker 4: Yeah, yeah, sure I can. There's a lot packed into this question. So we have kind of a long answer. I guess my first thought with this is aerosol obstacle depth. is it's quantitative. So it doesn't matter vertically where the aerosols are located in the atmosphere. I think maybe the person who's asked this question was conflating aerosol optical depth and the UV absorbing aerosol index because the UV absorbing aerosol index, those values do change depending vertically where the aerosols are in the atmosphere.

1:23:27 Speaker 4: So that's not an issue with aerosol optical depth. So in terms of cloud contamination, That's a very specific anomaly that occurs adjacent to cloud edges. So if you, you know, suspect that that is occurring, you should definitely take a look at whatever the corresponding true color or geocolor, if it's ABI, aerosol optimal depth, to see if those, you know, kind of noisy looking high AOD values, which would be adjacent to a cloud, And typically they occur when the surrounding AOD, you know, that's not where there's the cloud is very low.

1:24:07 Speaker 4: So usually those anomalous cloud containment nation is pretty easy to pick out. And then what was the other things? Oh, you know, humidity driven particle growth, things like that. We recommend definitely taking a look at some of the other products. Like, so for example, aerosol layer heights, Certainly the smoke and dust mask. If it's a dust plume, the ADI dust RGB. Remember, ideally you never want to look at these satellite products in isolation. You want to look at the different products that are available because that gives you a more complete picture of what is actually happening in the atmosphere.

1:24:50 Speaker 4: Is there something you want to add to that, Katarina?

1:24:52 Speaker 2: No, I think you've covered most of it.

1:24:55 Speaker 1: I just like adding that maybe You know, if you refer to the technical documentation of the product, you will find, you know, all like the flags and the threshold that's been applied and are recommended to get the best quality data.

1:25:12 Speaker 2: Yeah, thank you.

1:25:13 Speaker 4: That's a good point. You always want to make sure you're using the appropriate quality or confidence data for whatever your particular application is.

1:25:22 Speaker 2: Okay. So thank you, Amy. So question number five.

1:25:29 Speaker 1: Does Tempo continue to maintain the hourly full scan of the field of view during rapid observation mode?

1:25:36 Speaker 2: So during like.

1:25:37 Speaker 1: The special operation, so with high frequency observation, Tempo usually targets less than 10, it's below 10 minutes of like rapid scan, but it doesn't usually. Conduct special operation for the entire day. So usually the goal is to limit the special operation to more than 50% of tempo observing time. So like it can also provide some normal or standard observation over the complete field of regard. I don't know, Aaron, if you want to add anything on this one.

1:26:17 Speaker 5: Yeah, I mean, I just want to add like, you know, during that one hour, you know, we won't go for like two hours of SPECIAL OPERATIONS, RIGHT? SO THERE, LET'S SAY, THAT ONE-HOUR OR TWO-HOUR TIME FRAME OF THOSE SPECIAL OPERATIONS WOULD OBVIOUSLY BE TARGETING THAT ONE AREA, BUT THEN WE'LL SWAP BACK TO A MORE STANDARD OPERATION IN.

1:26:39 Speaker 3: THAT FALLING COLOR, RIGHT?

1:26:40 Speaker 5: SO WE WILL PROVIDE COVERAGE OVER THE SCHOOL FIELD REGARD EVERY SINGLE DAY, BUT THERE ARE SPECIAL OPERATIONS FOR THAT ONE SECTION.

1:26:53 Speaker 2: OKAY, THANK YOU, ADELINE.

1:26:56 Speaker 1: So we can go ahead and question number six. How does NOAA validate tempo satellite measurements against ground-based air quality monitoring station and how significant are the differences between them?

1:27:08 Speaker 2: So there is a.

1:27:08 Speaker 1: Paper describing the all methods. So we didn't go in detail in this training, but we refer to the paper if you wanna learn more about that. I don't know, Amy, if you have anything to add on this one.

1:27:21 Speaker 4: Yeah, we kind of ran out of time. We didn't have a chance to cover the details of the validation. But for the PM2. 5, it's a tenfold cross-validation. So like Katarina said, you can read all the details that are available in that paper that was published earlier this year.

1:27:41 Speaker 2: Okay.

1:27:42 Speaker 1: I think we'll probably try to answer maybe another couple of questions. And then we close the live session. And then, as I mentioned, whatever is not answered now, the question, all the questions you put in will be answered offline and we posted all the questions and all the answers in the training webpage.

1:28:03 Speaker 4: Oh, Katarina, sorry, could I just make a follow-up point? There were several questions I saw related to validation. And since we're only going to be able to answer one or two more, Just be aware, anybody who has questions about validation of any of the products, I forget what slide it was, but it was towards the end that listed the references. So this paper here by Zhang et al. about the PMC. 5 validation, and then there's also a paper on TEMPO, AOD, and ALH validation. So you can find all that information, the links to that on that particular slide.