NASA ARSET - LiDAR Profiling Satellite Observations for Air Quality Applications

All right. Hello, everyone. Welcome to part two of LIDAR profiling, satellite observations for air quality applications. I'm Ed Nowotnyk from NASA Goddard Space Flight Center, and I'll be your trainer for this second part. Okay, as I mentioned, this is part two of our training. Following our training, we'll have homework assigned, and that will be due two weeks from today. The homework will be posted on the training webpage. And following completion, Each attendee will be awarded a certificate certifying they've completed the training course.

So part two will be focused on how we translate our observations into data products. And we're going to walk through the theory as well as do some live real-world examples. So hopefully following this training here, you'll be able to understand how some of the LIDAR data products such as aerosol type and cloud phase are determined from the fundamental measurements of the LIDAR system.

So again, my name is Ed Nowotnik. I'm a research physical scientist at NASA Goddard Space Flight Center. I've worked with a lot of airborne and space-borne LIDAR systems, both in space as well as some that are currently being planned to launch in the early 2030s. So just a refresher of what we covered in part one. In part one, we identified past and currently available LIDAR missions and some of the differences between those different missions. We recognized the capabilities of LIDAR active remote sensing and measuring vertical profiles of aerosols and clouds and for how important they are for air quality applications.

And we also focused on understanding the strengths and limitations of LIDAR observations. Today we'll be focusing on number three and five. So today, as I mentioned, we're going to walk through the theory behind how some of these level two LIDAR products are determined for a given scene. And at the very end, I'll show you how to access LIDAR data both quick looks for case studies as well as where to go if you want to download and analyze larger volumes of data. And the focus of today's training will be using real data from two of NASA's airborne LIDAR systems, Calypso and CATS.

These two were primarily developed for aerosol and cloud profiling. In part one, we covered many other LIDARs that measured the atmosphere but necessarily weren't developed. for atmospheric applications. So, today we're going to be focusing on Calypso and CATS, and at the end we'll show you how to access data from those two LIDAR systems. So, again, if you have questions, please put them in the questions box, and we will address them at the end of the webinar. Feel free to enter them as we go. That way you may, you know, in case you forget, pop them in the question box.

We'll try to get through all of them during the Q & A. And then the remainder of the questions will be answered in the Q & A document, and we'll post that about a week after the training. Okay, so first let's do a quick review of part one as it pertains to part two. So as I mentioned, we're going to be talking about Calypso. Calypso was launched in 2006 as part of the NASA A-train, and it operated in space until 2023. Calypso operated at both 532 and 1064 nanometers, and it established a long record of space-borne LIDAR from NASA for data users to use.

It was useful for tracking aerosol plumes, like I'm showing in the bottom right here. This is a Saharan dust event transported from Africa into the Caribbean, and you can see over successive days, Calypso provided key vertical measurements of the vertical distribution of this dust event as it entered into the Caribbean. Now CATS, just a refresher on that, was a low-cost LIDAR that was developed for implementation on the International Space Station. This was a processing orbit, so it enabled diurnal sampling at different times of the day.

primarily operated at 1064 nanometers, but also had 532 nanometer measurements as well. And this instrument operated 33 months from 2015 until 2017. So before we move on, I wanted to pause and do a quick learning check. So given the choices below, what are two differences between Calypso and CAS? A, they had different orbits. B, they are both LIDARs. C, they both operated at 1064 nanometers. D, one was part of the A train, and the other was operating on the ISS. All right, hopefully you selected answers A and D. So, Calypso was in a polar orbit, and CATS operated in the ISS, the processing orbit.

And CLPSA was part of the A train. So, these are some of the key differences between these two LIDAR systems. They're very similar in a lot of respects. That's why I'll be showing observations from both datasets interchangeably throughout the remainder of this presentation.

All right. So, let's move on to the theory of how we get from LIDAR measurements of photons to level two data products such as aerosol extinction, aerosol type, things of that nature. So in part one, we went over the fundamentals of how LIDARs work, how they collect photons, different ranges from the telescope. So now we have that vertical information of backscattered photons. What do we do next? Well, the first step is we have to calibrate the signal for backscatter LIDARs. And to do that, basically what we do is we normalize the backscattered signal to a region of the atmosphere where we're relatively sure that it's free of aerosols.

That means higher in the atmosphere. And so we normalize to the molecular or Rayleigh signal. And here on the right, you can see an example of the different regions of the atmosphere where various LiDAR systems tend to calibrate their signal. Airborne systems are nominally operating at 20 kilometers or less, so they have to calibrate lower in the atmosphere. They have to be extra careful about avoiding clouds and aerosols in their calibration region. CATS calibrated a bit higher in the atmosphere, and calypso even higher.

And so as you go higher in the atmosphere, you actually have less molecules. So you're playing the game of you want to be free of aerosols and clouds, but you also want sufficient signal-to-noise ratio to be able to calibrate to the molecular signal. And so here's an example of that. This is real data from CATS and a ground-based LIDAR system called PolyXT in red. They're the two squiggly lines here. And you can see this dashed line is the theoretical molecular signal. And so that's what we're calibrating to operationally for cats and calypso.

And you can see higher in the atmosphere, this signal more or less follows that theoretical line. And then as you get down close to the surface, you see it deviates. And that's where we're getting particulates, whether it be aerosols or clouds, deviating from this theoretical molecular signal. Okay, so let's walk through some of the calibrated Level 1 products. This is a CATS example here. And this is the total attenuated backscatter at 532 nanometers. We call it attenuated because it includes signal attenuation We discussed this a bit in part one, but you can see it here where these clouds are fully attenuating the atmosphere below it.

We call it total because the backscattered signal includes both molecular and particulate contributions. So you're measuring molecules as well as clouds and aerosols. And this total attenuated backscatter, whether it be at 532 or 1064 nanometers, is the fundamental level one product that is used in LIDAR data. And this is used to go further down the road and generate level two data products. And so total attenuated backscatter is a measure of intensity. And, you know, I've had lots of experience staring at LIDAR curtains.

But looking at this, you can see that the clouds, these look to be clouds because they're higher intensity and backscattered signal. They're higher in the atmosphere. They're up around 15 to 20 kilometers. Whereas aerosols tend to have Protocol distribution is more common to what I'm showing here. They're closer to the surface. They're a bit more homogenous. And as you get more and more familiar with LiDAR data, these sorts of scenes will become more obvious. But we'll walk through all the steps that are used to actually determine clouds versus aerosols and assign their phase and type respectively.

So this is 532 nanometers. CATS also operated in the near IR at 1064 nanometers. And here you can see that this is the exact same scene, but you can see that it's very similar to the 532 scene. And I'll toggle back and forth. And one thing you may notice is there's a lot more signal in this region in 532 nanometers. And that's because at 532 nanometers, we have more molecular backscattering. And we use these two different wavelengths because, as you'll see in subsequent slides, we use the backscattered signal from these two wavelengths to help discriminate clouds from aerosols.

Another level one product is what we call the perpendicular attenuated backscatter. And as I mentioned in part one, LIDAR systems measure linear depolarization from particulates in the atmosphere. And non-spherical particles such as ice particles and clouds or dust particles that are aerosols will backscatter light in the perpendicular plane. So you'd have a perpendicular component to the backscattered signal. And so because we're getting perpendicular attenuated backscatter. You know, we thought that these were likely clouds.

We thought these were likely aerosols. Now we have some information about their shape. And because they're non-spherical, we can say, hey, this is likely cirrus, and this is likely dust. Okay, but how do we actually do that in algorithms? Well, we have our calibrated signal. And so the first step is to identify what we call features or layers in the atmosphere. And these are the deviations from the molecular signal that I was discussing earlier. These are the particulates. We don't know what they are yet, but we know that they are going to be either clouds or aerosols.

And so fundamentally how these algorithms work is you have your theoretical profile, molecular profile, And then when you see these deviances from that, you say, oh, there's a layer here between 15 and 17 kilometers. Well, we have to account for the attenuation in that layer. And so we have readjust the theoretical curve. We say, oh, here's another layer here. And we readjust the curve again. And so here in this case, in this single profile, we have four different layers. And so it's not uncommon to have multiple layers in a single column.

You can have multiple aerosol layers, clouds embedded in aerosols. But this is the general approach for identifying features or layers. And for tenuous layers, it's not uncommon for LIDAR algorithms to have to average horizontally to beat down noise and improve signal-to-noise ratio so that these features pop out from the theoretical curve. So that's sort of the traditional approach for feature detection. I wanted to mention some recent work that was done by Selmer et al. So very recently, several groups have been working on signal denoising.

And if you recall during the day, solar background contamination plagues the measured backscattered signal by the LIDAR. So that makes the data look extra noisy. So using machine learning techniques, groups have been developing methods to denoise the daytime signal, and it enables us to do feature detection at finer horizontal resolutions. Remember I said traditionally you have to average long distances using this feature signal denoising. You no longer have to do that. And so here's an example from the paper.

Here's a truth signal of attenuated total backscatter. Here's a noisy signal that would be representative of what was measured during the day. And then here's the denoise signal. You can see that looks much more comparable to the truth. And on the right are the feature locations. So here's the truth resolution. You can see that the The cloud feature detection was very fine. See lots of structure in the aerosol. When we do the standard feature detection, you can see it's much more blocky. And that's the side effect of averaging horizontally to do your feature detection.

And then in the bottom column here, you can see that the noise feature detection looks a lot more like the truth. So this is new capabilities that will hopefully be applied to past data sets, but also data sets going forward to improve daytime signal-to-noise and consequently feature detection. Okay, so we've talked about feature detection. Now let's talk about depolarization ratio. So we talked a little bit about the perpendicular attenuated backscatter. And so what the depolarization ratio is, is the ratio of the perpendicular return to the parallel return.

And recall that non-spherical particles will have a contribution to the perpendicular return. So that means spherical particles, such as sulfates, for example, or water clouds, will have depolarization ratios much closer to zero, whereas dust and cirrus clouds will have depolarization ratios that range from 0.2 to 0.6, 0.7. And we'll walk through an example here.

So here's a profile of depolarization ratio from the CAT slider.

And here are some real examples of what we just talked about. So as I mentioned, smoke tends to be spherical. You can see the depolarization ratio for that ranges from 0.2 down to zero. Dust is moderately polarizing. That's typically around 0.2 to 0.3. These yellows and oranges. And the ice within cirrus clouds is much more irregularly shaped. And you can see those depolarization ratios are higher. They're 0.4 to 0.6 and beyond that. So this is how you start to use your level one observables to help identify the various feature types.

And this is how the the algorithms are constructed. We use things such as the depolarization ratio to help identify cloud phase and aerosol type. Now that's depolarization ratio. Now we'll talk about color ratio. And so color ratio is the ratio of your attenuated total to backscatter at your two different wavelengths. So in this case, it's 1064 nanometers to 532 nanometers. And what this does, it tells you information about the particle size. So larger particles, such as clouds, exhibit color ratios that are 0.6 to 1 for cirrus clouds, 0.8 to 1.4 for water clouds.

Generally, they're closer to 1, meaning there isn't much spectral dependence in the visible and the near IR. Aerosols are smaller particles, and therefore they have a much stronger spectral dependence at these two wavelengths. And so what we do is we use this information to help identify clouds from aerosols. We call this cloud aerosol discrimination. And here's an example here where we have some Cirrus clouds. We can see that color ratio is closer to 1, and the aerosols are closer to 0.5 or so.

So we use this color information to help separate clouds from aerosols. Okay, so that's what it looks like in the data products. And then we start working towards what we call our vertical feature mask. And that includes information of aerosols versus clouds, as well as their phase and type. And so here's a level one attenuated total backscatter scene. And here's using the color ratio and depolarization. ratio. You can see how that translates into clouds and aerosols. You can see it's kind of binary, the blues versus the oranges.

But we first assign cloud versus aerosol before we get into phase and type. And again, we primarily use the color ratio to help discriminate clouds versus aerosols. Now once we've identified a feature or layer as a cloud, now we want to identify the phase. And this is where the depolarization information comes in. And here's a PDF of the depolarization ratio for ice clouds and water clouds. And as we discussed earlier, Ice clouds have a non-spherical component, which generates backscattered signal in the perpendicular channel.

And so we use the depolarization information to help identify cloud phase. Now, there is some ambiguity at times, like in this range here, where it could be mixed phase. There often we'll pull in temperature information from a reanalysis, for example, to help us with that choice. And so here's an example. Here's attenuated total backscatter from CATS, level one product. Here's the level two, cloud aerosol discrimination. And then going here, here's the cloud phase. So the white clouds are the cirrus clouds.

Blue are the water clouds. And you'll notice that where it was orange in this panel is not depicted here because those are aerosols. This is just a vertical profile of the clouds and their phase.

So that's clouds. Now let's talk about aerosol type. So now we have a feature identified as aerosols using the color ratio. Here's an example flow chart from the CATS aerosol typing algorithm for how we assigned aerosol type. And so there's several inputs into this. One is, as you can imagine, the depolarization ratio. That's very useful for identifying dust and dust mixtures. We use information on the feature thickness. Is it really thick? It's more likely to be a smoke plume. The underlying surface type.

Are we over the ocean, the remote ocean? It's more likely to be marine or sea salt. The elevation. Is it lofted high in the atmosphere? If it's very high, it might be UTLS aerosol. If it's lower, it may be a smoke event again. And then often we, for CATS at least, we pulled in information using MERA-2 to help assign an aerosol type where some of this information was kind of ambiguous. And so similar to the cloud phase map, here's a different case from CATS. We have CATS 1064 nanometers level one total attenuated backscatter.

Here's the feature mask. And then here's the aerosol type. And similarly, you see these light cyan colors, which are clouds, are not showing up in the feature type. And so I'm not showing it here, but this thick yellow plume is dust. This black is smoke. Here we most likely relied on the depolarization ratio information to help with that choice in the aerosol typing algorithm. Now, for both clouds and aerosols, if we want to get to optical property retrievals, such as extinction, for backscatter LIDARs, such as Cas and Calypso, you often have to assume an extinction to backscatter ratio known as the LIDAR ratio.

This is part of the LIDAR equation, which we're not going to get into. you have to make an assumption about the LIDAR ratio. And that's the purpose of this aerosol typing and cloud phase algorithm. Once we make that discrimination, we assign a LIDAR ratio so we can do our optical property retrievals. And these LIDAR ratios can range from 10 to 80 steradians. They're generally representative of different aerosol types. And what we do is once we have the aerosol type map, for example, we can go in and do optical property retrievals for the various contributions of the aerosol types to the extinction.

So this is a curtain here where we have an A, the continental urban pollution contribution to the extinction. Here's the dust contribution. Here's the marine. And so we can get more sophisticated in how we do our aerosol typing retrievals and our level two products. Okay, so we've gone through the theory of how we get to our level two products. We showed some case studies with a real world example, but here let's walk through this again here with Calypso. So here we're moving to Calypso. We have our total attenuated backscatter 532 nanometers in the top, perpendicular in the component.

And then hopefully looking at this, you can start to think about what sorts of features, phase, and type might pop out of this. So comparing box A in the total to the perpendicular, you can see there's a contribution there. So hopefully you're thinking, oh, this might be Cirrus Cloud. In circle B here, we have strong signal in the total, but not so much in the perpendicular. indication of spherical particles, so you might be thinking this could be something like sulfate or smoke, for example.

Now we fold in the color ratio information, and you can see, okay, you can see it with your eyes, the spectral dependence in region B, again, indicating that, hey, this is likely aerosol. In box A, those look pretty much the same. Remember I said that The color ratio for clouds generally hovers around one. And so piecing these pieces of information together, you can start to think about how the level two data products are going to shake out. So here's the feature mask for Calypso. You can see it looks very much like the cats.

As you may have guessed, box A is a cloud and circle B is aerosol. If we go a little bit further into the level two aerosol typing, you can see that this was actually a little bit of a dust mixture here. So there was a little bit of perpendicular return, but not strong enough to be a pure dust case. And so hopefully we're getting more comfortable with how the LiDAR algorithms go from photons measured to calibrated signal level one products to feature detection to cloud aerosol discrimination to aerosol type and cloud phase.

And so now we're going to walk through an example. You know, let's say there's an event of interest and you may want to go see if cath or calypso overflew that event. And if so, you know, walk through the data products and quick looks to see what the LIDAR saw. And so, let's first start with how to find an aerosol event and to see if Calypso, in this case, overflew it. All right. Now we're going to walk through a real-world example. First, we're going to see how or if Calypso overflew our event of interest.

And so I'm on the Worldview webpage here. I have, so I mean, MVP, true color. And I've already moved to my date of interest. And this is June 20th, 2020. Some may be aware, but there was a monstrous dust event that came off the coast of Africa. made its way into the Caribbean and into Texas. It actually had very adverse impacts for air quality. So you can see this huge dust event making its way across the tropical North Atlantic here on the 20th. So let's see if Calypso intersected this. So what you do is you go here, you click add layers, and you go up to the search bar, and I often this type Calypso.

And for this case, let's select both day and night tracks. Remember that Calypso sort of parses their data by day and night. So, let's select both of those, and then we'll go ahead and X out of this. And look at that. So we can see that there's a transect that goes right through this event. It's somewhere between 0450 and 0505 UTC. So now we're going to go over to the Calypso Quick Look website. and see what the data looks like. So now we're on the main Calypso web page. You can quickly get here by Googling Calypso NASA.

We'll also provide the direct link to the Quick Looks in our slides here. But here I'll show you how to navigate to the Quick Looks from the main web page. So we're on the main web page. We go to Products. and select LIDAR Browse Images. And for this case, we're going to look at the most recent data release, which is version 4.51. And here you see a calendar of data availability. So this event was June 20th of 2020. Here we are in 2020. Select the 20th. And it looks like there's a fair amount of data available on this day.

And let's go scroll on down to try to find our event of interest. So I mentioned it was between 04.50 UTC and 05.05. And here we are. And so you can see the timestamps here. This is the time Time start is 0445. Time end is 0531. And if we compare this to the worldview image, if you recall, that event was located, it looks to be along this pink curtain here. And so let's go ahead and select that. So this is a nighttime image. And here we go. So we get a nice map of the track. Here's the attenuated total backscatter at 532 nanometers.

See that this is, you know, very strong event. We have backscatter signals in this red portion of the color bar, which is very strong for aerosols. So that's the 532. We go down to the 532 perpendicular backscatter. And you can see that there's a significant component in the perpendicular part.

And recall that we used that information to construct the depolarization ratio. And you can see that this depole ratio is within the range that we discussed for Saharan dust, right, 0.2 to 0.4. So that all checks out. Now what about the 1064 nanometers? What about the color ratio? So we have lots of signal in the near IR here. Let's look at the color ratio. And we have some, it's mostly purple, some You know, 0.8 or so I say would be the average. So indicative of aerosols in this case. So this is all checking out.

Aerosol, it's dust. What does the feature mask look like? Okay, so the feature mask is coloring this mostly orange with some blue in there. There looks to be some embedded clouds within the Saharan dust event. And you can see, yeah, this is number three, which is orange tropospheric aerosol. I mentioned horizontal averaging during the day. If those are interested in that, how far along the curtain the signal was averaged to detect the features, that's a curtain of that. In this case, we're focused on aerosols and not so much on cloud phase.

And then here we are. What's the aerosol subtype? And this bright yellow is Saharan dust or desert dust in this case. And so, This is a real-world example of how you may want to go into and see if Calypso flew over an aerosol event or cloud event, convection event of interest, and how you can find the Quick Look data on their website. So we walked through a case study for how to find Calypso data for your event of interest. I mentioned that I would provide the direct links to The QuickLooks for both CATS and Calypso, they're provided here.

On the left, you can see that the CATS website for QuickLooks looks very similar to the Calypso website that was developed and designed intentionally to mimic that so that data users would be comfortable with both data products. On the right is the direct link to the browse images that we just walked through for this Dustzilla event. Now, this is for individual case studies. For CATS, for example, there is an option to download HDF5 files associated with that, we call it a granule, or, you know, this chunk of the orbit, so just a subsection of the orbit.

However, if you want to download, let's say, a month of data or a year of data and look at longer time scales of analysis, we want to go to one of the data acquisition centers. And for both Calypso and CAS, they're both housed at the NASA Atmospheric Science Data Center, or ASDC. And here you can see that they're sort of partitioned in various collections. And notionally, these are with the level of the data products. So level 1B would be your calibrated backscatter in this case. Your level 2 products would be your feature mask, feature detection.

So depending on what you're interested in, you can download a subset of these data products. You may download only daytime products. You may download only nighttime products. As I mentioned, the data products are partitioned accordingly. You can see there's a D in the file name for daytime data. and N for nighttime. But so depending on what you're interested in, if you want to download larger chunks of the data, I highly recommend using the ASDC for that. All right, so let's move on to the summary for this training.

So thank you again for attending part two of our LiDAR training. Today we covered how to interpret information from LIDAR curtains to help identify cloud phase, aerosol type, and aerosol plume information for a given scene. What I hope you really gained an understanding and appreciation for was how the different wavelengths and the polarization ratio information are used to classify clouds versus aerosols, their phase, and type. Following that, we went to Worldview and walked through how to search for whether cats or Calypso overflew an event of interest.

In the live demo, we used Calypso. If you want to see if cats overflew an event of interest or we entered Calypso in the search bar, just punch in ISS and you'll get the ISS orbit overlaid on top of your imagery. So following that, we went through a demonstration on how to access QuickLooks on the Calypso website. As I mentioned, if you go to the CATS website, it's structured very similarly. And that's really useful for the QuickLooks or a case study. But if you want to download longer chunks of data, we recommend that you go to the Langley DAC or ASDC.

So I mentioned homework at the beginning of this. We will have one homework assignment. It will open today. and it can be accessed from the training webpage, and use Google Forms for your answers, and it's due two weeks from now. And then following that, we will issue a certificate of completion. So if you're interested in getting in touch with me, if you have follow-up questions, feel free to reach me at the email address provided here. Here are links to the RSAT website and the RSAT YouTube page, And if interested, we want you to visit our sister programs, such as DEVELOP.

And thank you all for attending the training. At this point, we're going to transition to our Q & A. How can we tell that the cloud is ice and not water? That's a good catch. In the LIDAR community, we tend to refer to Liquid water clouds is water clouds. Of course, ice is water as well. That's just a, I guess, not well-defined term that we use in the LIDAR community. So when I was referring to water clouds in the presentation, I meant liquid water clouds. such as cumulus, for example. Which LIDAR product has available data for South America?

all of the space-based sensors that we went over last time and focused on this time will have coverage over South America. So that would be cats in the Calypso today, with the caveat that cats would only go down to 52 degrees south. Certainly, ICET-2 covers that, EarthCare, Aeolus, any of the polar orbiting centers will provide measurements over South America, but that would just be sort of a curtain or two overpassing them per day per instrument. On the ground, there are historically several micropulse LiDAR network, MPLnet stations located in South America.

I think there's an active one in Sao Paulo now. And I think there have been some previous field campaigns supported by NASA that have also deployed some ground-based systems there as well. And I'm just referring to NASA assets in this case. I'm there should be some other, I suspect there are other agencies or countries that have deployed LIDARs to this part of the world. Okay, why does EarthCare use 355 nanometers? So, the 355 nanometer laser transmitter has been developed for space by ESA in partnership with Leonardo in Italy.

So that's sort of the technology demonstration. And the reason for going to 355 is these are high spectral resolution LIDAR systems. So if you were in the first training, the HSRL system really takes advantage of strong molecular or Rayleigh returns to help separate that from the particulates, which are the aerosols and clouds. So you want a really strong molecular return for the HSRL system. And when you get into the 355, you'll have a stronger molecular return than you would at 532 and definitely 1064 nanometers.

Is there a way to find data based on location instead of the time of event? I'm not aware of an easy way to do that.

NASA has a website called Giovanni that you can do with the passive sensors, so MODIS, for example. The best way, especially for case studies, to check to see if the satellite of interest ever passed your event is to go into worldview and then overlay the orbit tracks. If you're going to do more sophisticated or more in-depth analysis, and if you go to the DAC and download some of the data, then you could just, at least in some of my previous experience, I just put up a box over my region of interest and just check to see

if the satellite overpassed. Okay. Can we access and analyze data in the GEE? And is there any data available to Ethiopia? Could you specify what GEE stands for? GEE is Google Earth Engine. Okay. I do not know. We

can get back to you on that one. But there is certainly data available. available to Ethiopia through the NASA DAX and the Quick Look websites that we went over in this presentation. So, I don't know if it made it in here. I did see a question come through the chat about single photon counting detectors, kind of more on the technical side. So, some of the LiDAR systems that we talked about utilizes single-photon counting detectors such as CATS and MP-LMED, the ones I think you were referring to.

I'm not aware of teams working on those, at least in this context, but that technology of SPCMs has been utilized for atmospheric LIDAR previously. Is there a hyperspectral future for LIDAR, or will it continue to focus on monospectral lasers?

I think near-term, just due to technology maturation, monospectral lasers will probably be the near-term future. I think there are some groups within NASA working on hyperspectral LIDAR systems, and I believe those are primarily being demonstrated on airborne platforms at the time. So that's typically the path to space. for a lot of these instruments is demonstrate your technology on an airplane, and then that's a stepping stone typically to flight implementation. Is it possible to use these satellites to measure gas emissions in a specific area?

Have any gas emission measurements or environmental alerts related to air quality have been created? Yes, so there are some ground-based, I don't know if you're referring to trace gases such as ozone, but there are trace gas ground-based systems to measure ozone, methane, NO2. There are methane lidars that have been deployed, NOs and lidars, I believe, that have been deployed on aircraft. I know ESA... believe I have that correct, is developing a space-based methane LIDAR called Merlin. And have any measurements or environmental alerts related to air quality been created?

Yeah, so these ground-based systems as well as some of these space-based systems that we discussed, these would be more along the lines of smoke hazards, things like that. There's near real-time capabilities they certainly have been using in the past to help warn the public or provide, you know, hazard outlooks or impact near-term forecasts. Can you do laser polarimetry, LIDAR from multiple angles? There are some groups, actually, the University of Wisconsin, they're They have a ground-based LIDAR system.

They've developed one that scans horizontally. There have been some past concepts for sort of multi-angle LIDAR systems, including CAS that demonstrated this for about six weeks in the first six weeks of its life. So it had two fields of view. looking left and right separated by seven and a half kilometers. And there the idea was to look at smaller scale spatial differences in clouds and aerosols. But there were some more sophisticated concepts floating around years ago that would have more view angles beyond just the two that Katz demonstrated.

Is it possible to obtain data for a specific cloud site? Yeah, so what you'd want to do there if you're interested in Cirrus Clouds, for example, is download your data file. And this would be in the level two data products. You would have your feature classification. So you would look for a flag, whatever it may be, depending on the space-based system or ground-based system that correlates to ice clouds. And then you would use that, and that flag would provide you the bounds of the feature, so the feature top and the feature bottom.

And then from there, you can look at things like the backscatter variability within the feature bounds. Different strategies use various spacetime co-location strategies. What is the most widely recommended approach for spacetime co-location when comparing calypso with aeronet observations. So for things like, yeah, aerosol-related research where they generally don't vary as quickly as, you know, convective environments, I think we typically use collocations with a three-hour window, you know, temporally, plus minus an hour and a half from the overpass time.

Then depending on your application, you would set your, you know, your range ring, something like 50 kilometers, 100 kilometers. Those are typically the scales we've used in the past. Again, that kind of depends on your application. If you're looking at, you know, small scale variability, you might want to set that ring smaller than broader analysis that you're hoping to perform.

And so for convection, just to follow up there, that's going to be a little bit tighter. So, when we fly our airborne systems under space-based systems, we typically use like a 30-minute window there just because in convective environments, the scenes evolve pretty rapidly compared to aerosols. So, can you talk about LIDAR applications for wildfire fighting? Yes. So,

Certainly for wildfire fighting, you want your data products available in near real time or real time if possible. So NASA has had a few field campaigns that have looked at fire emission and environmental controls. They've flown LIDARs on those aircraft. And those goals were kind of more focused on helping to constrain downwind effects and improving the forecasts downwind. More recently, I'm aware of some groups that are developing UAV technologies so that they can deploy small-scale LIDAR in these challenging environments where you wouldn't want to be operating a ground-based system or potentially even flying within these wildfires, where you would hover over the you know, the wildfire, get an estimate of the plume height and use that to help forecast downscaling or downwind effects.

That's more recent, this UAV development. And just another sort of application that I know was used for both CLPSO and CATS was helping to characterize thickness and tops of volcanic eruptions as well, because that has certainly important radiative impacts on the Earth, but it's also very important for the aviation industry to understand where those events are located in the vertical and the atmosphere so that they can avoid those regions. So as we mentioned, the homework will come out Following this second part, I hope everyone who attended today learned something and looks forward to working with LiDAR data going forward.

Certainly reach out to myself or Melanie if you have any follow-up questions that you didn't think of today. And at this point, I'll maybe turn it over to Melanie just in case I missed something or if you want to add anything. Nope, that all sounds great. Thanks, Ed, for a fantastic presentation. Thank you, and thank you to everyone who attended. Have a good day.

The Story