progressaudiocamera moved downcamera moved leftcamera moved rightcamera moved upfree exploration modeguided tour modehotspothourhoursUse the w, a, s, and d keys to move around the 360 degree image. Press the tab key to jump to interactive markers and hotspots.360 degree image interactionlabelmarkerminuteminutesnextpause audiopause videoplay audioplay videoprevioussecondseconds%count% of %total%step %count% of %total%tooltip%total% items%total% itemprogressvideovideo volume%count% of %total% items visited%count% of %total% item visited%count% of %total% visitedAccessible textBackground audioKeyboard shortcutsZoom to fitclose video settingsfull-screen videoVideo Playback SettingsOffOnCaptionsvideo settingsmute videoNormalpause videopicture-in-pictureplay videoPlayback Speedvideo captionsplayback speedvideo volumeCaptionsContinueNEXTPREVSUBMITPlayback SpeedNormalContinueGlossaryTermsActionAltCtrlKeyboard ShortcutsEnable keyboard shortcutsList keyboard shortcutsMute / unmuteNextPlay / pausePreviousReplayShiftShortcutSubmitToggle accessible textToggle background audioToggle captionsToggle full-screenToggle zoom to fitMenuNotesPlease rotate your deviceResourcesRestartResumeCloselockedNext (Ctrl+Alt+Period)Previous (Ctrl+Alt+Comma)searchSidebar ToggleBack to topslide progressStart CourseStart CoursevisitedClear and return to menuSearch in:FilterSearch...NotesSearch ResultsSlide TextLearn MoreTo watch this video, use the preview feature in Storyline 360, upload your published output to a server, publish to Review 360, or switch to static video quality.Streaming video is unavailable for local playback.Hide captions (Ctrl+Alt+C)Show captions (Ctrl+Alt+C)Enter full-screen (Ctrl+Alt+F)Exit full-screen (Ctrl+Alt+F)Mute (Ctrl+Alt+M)Next (Ctrl+Alt+.)Pause (Ctrl+Alt+P)Play (Ctrl+Alt+P)Playback speedPrevious (Ctrl+Alt+,)Replay (Ctrl+Alt+R)SearchSettingsSubmit (Ctrl+Alt+S)Unmute (Ctrl+Alt+M)AutoscrollCloseTranscriptMove windowOpenResize windowResume autoscrollSelect trackShow timestampsTranscriptsOptions<Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Welcome to this training session titled “Application of Data Fusion Techniques to Fire Detection”. I’m Dan Bikos from CIRA, Jim LaDue, Lexy Elizalde-Garcia, and Katy Christian from WDTD assisted me in putting this together. ">
<Style FontFamily="+minor" FontSize="12" />
</Span>
<Span Text="
" />
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="72" TrailingMargin="0" FirstLineMargin="36" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="">
<Style FontFamily="+minor" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="If you have not taken the WOC Severe course since the FY21 update we recommend this as prerequisite content to go through since we will build on concepts learned in the data fusion section of the course. This lesson will apply data fusion concepts introduced in the WOC Severe course to wildland fire detection. We’ll go through a case study to demonstrate initial detection of hotspots via satellite and webcams. Next, we’ll discuss fire detection considerations via satellite and radar. Finally, we’ll have a case study to demonstrate satellite-based fire detection when fires exceed decisional guidance thresholds in the context of fire warnings.">
<Style FontFamily="+minor" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Recall in the WOC Severe course we learned about the 4 major considerations for data fusion. They include anticipating data sources, data constraints, confidence in evaluation and data quality issues.">
<Style FontFamily="+major" FontSize="12" />
</Span>
<Span Text="
" />
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="72" TrailingMargin="0" FirstLineMargin="36" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="
" />
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="72" TrailingMargin="0" FirstLineMargin="36" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Learning objectives, after completing this lesson students will, number one, gain an introduction to GOES satellite and radar identification of fire hot spots and cold front / outflow boundary on fire intensity. Number 2, be able to describe strengths and limitations of satellite and radar data in detecting fires. Number three , learn how the four data fusion considerations impact identification of fires using satellite and radar. ">
<Style FontFamily="+major" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="During this training session we’ll be looking at time series plots like this of GOES maximum brightness temperature at 3.9 microns. This time series is for the Park Fire in California on July 24, 2024. The time of ignition is slightly before 22:00 UTC. Click on the links to view how I made this time series plot by visualizing the circle around the hotspot used in the AWIPS tracking meteogram tool to obtain the maximum brightness temperature and also the animation. Keep in mind that there is a saturation temperature for the GOES 3.9 micron band where higher temperatures may exist but will not be detected above the saturation temperature. For the GOES-R series the saturation temperature for the 3.9 micron band is 138.7 degrees Celsius. Note that the Park Fire does reach the saturation temperature at 0230 UTC. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s focus on detection of fire ignition by making use of the Mountain Fire which occurred on November 6, 2024 in southern California. Fire weather conditions were very favorable for fires on this day. NWS Oxnard issued a Particularly Dangerous Situation Red Flag Warning for this day, highlighting the risk of extreme and life threatening fire behavior. Wind gusts in the 50 to 80 mph range were forecast. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="The objective here is to look at the initial detection of a hotspot from a wildfire from the perspective of GOES satellite imagery with an unobstructed view, and compare it with web cam imagery that was focused on the location of the wildfire. The ALERTCalifornia program at UC San Diego has a network of many webcams across the state of California with the goal of monitoring for wildland fires. In this animation, I’ve time matched the webcam image (given in local time) with the time of the GOES-18 3.9 micron brightness temp given in UTC. This was made possible by the mesoscale domain sector for GOES-18 providing 1-minute imagery, along with webcam imagery available approximately every 1-minute. The time of initial detection of a hotspot from the GOES 3.9 micron imagery is 1650 UTC located just southeast of Santa Paula. One might say that the first detection of the hotspot was at 1649 UTC, but it becomes much more obvious by 1650 UTC. As for the webcam imagery, the first indication of a smoke plume was on the right side of the image at 8:48 AM local time which is 1648 UTC. The smoke plume becomes obvious by 8:50 AM. I recommend clicking on the animations below to step through the loops frame by frame to see for yourself when you first observe the hotspot from the satellite perspective and the smoke plume from the webcam perspective. I’ve made both available individually, as well as the combined animation shown here for closer inspection. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s zoom out a bit and look at other GOES satellite imagery for comparison, I’ll ask you to ignore the hotspot further south near Malibu for now. In the upper-left panel is the GOES 3.9 micron imagery with the default color table, exactly like I showed in the previous slide, in the upper-right panel is the 3.9 micron imagery with a different color table, I chose the color table used on the CIRA SLIDER webpage, in the lower-left panel is the GeoColor imagery and in the lower-right panel is the Fire Temperature RGB along with METARs to highlight the winds. First, for the 3.9 micron imagery with the CIRA SLIDER color table, the contrast between the hotspot and surrounding pixels is not quite as noticeable as the default color table in this case, but soon thereafter the hotspot becomes obvious. In the GeoColor imagery, the scene is complicated by high level cirrus clouds moving southwards, this makes it much more challenging to spot the smoke plume compared to a clear sky day without cirrus clouds. I recommend clicking on the view animation link so that you can carefully step through the animation and determine at what time you can discern the smoke plume. Also, ask yourself when you’d be able to discern the smoke plume without relying on the hotspot observed from 3.9 micron imagery. Finally, analyze the time of initial detection of the hotspot from the fire temperature RGB. Clearly, the time of initial detection of the hotspot is later than that of just using 3.9 micron imagery by itself. This makes sense since we are using other bands in the RGB that require higher magnitude hotspots to be detected. The fire temperature RGB is most useful for analyzing the magnitude of hotspots for existing fires, but is less useful for initial detection.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="The Next Generation Fire System commonly referred to as NGFS is an automated satellite fire detection algorithm. The NGFS uses satellite observations along with advanced spatial and temporal metrics, to detect fires in a manner consistent with human expert analysis of satellite imagery. For access to real-time data and additional information you may click on the NGFS link. The initial detection time of the Mountain Fire that we’ve been discussing was 1650 UTC. This time is consistent with initial detection of the hotspot in the GOES 3.9 micron imagery shown earlier. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Now let’s visualize the data types we’ve been looking at relative to a time series plot of maximum GOES 3.9 micron brightness temperatures over the Mountain Fire. I’ve placed the time of initial detection for each data type onto the plot. If we consider a webcam as “ground truth” data, we’d expect this data to provide the greatest lead time, particularly in this case as the scene was over the location of ignition. The detection from the NGFS and analysis of 3.9 micron imagery was within a couple minutes after the initial detection of the smoke plume from the webcam. The most important take away point here is that unobstructed satellite imagery of a hotspot due to wildland fire is comparable to “ground truth” initial visual detection of a smoke plume due to a wildland fire. In other words, satellite imagery can be trusted for the purpose of initial detection of a wildland fire, so long as obscuration is not an issue. Later, we’ll discuss sources of obscuration such as clouds and dust that may prevent or hinder detection. I also show the times of initial detection of the hotspot based on my own subjective analysis from the GOES Fire Temperature RGB and the smoke plume from GeoColor to illustrate that they are not as useful for initial fire detection relative to GOES 3.9 micron imagery or the NGFS.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Since lightning is one of the leading causes of wildland fires in the US, it’s important to monitor lightning for potential wildland fire ignition. While both satellite and ground-based lightning networks are helpful for lightning detection, we will specifically focus on the ground-based networks here due to their ability to differentiate between cloud to ground and intracloud flashes and provide precise geolocation information. However, if you'd like more information about using the GLM in relation to wildfire detection, you can check out the AMS Short Course on this topic linked under the "Resources" tab. We’ll consider the Gothic fire over southern Nevada from July of 2025 for this case. On the left, we have GOES 10.3 micron IR imagery along with cloud to ground lightning polarity data from 3 ground-based lightning networks, ENTLN in yellow, GLD in salmon and NLDN in green. On the right, we have the GOES 3.9 micron band over the same area about 2 days later. The white oval denotes the same location on the map, the lightning observed on the afternoon of July 2 was responsible for the fire that started on the evening of July 4. Note that lightning strikes within the white oval were positive polarity, past research has found positive cloud to ground strikes to be more likely to ignite wildland fires, however recent research has shown that negative strikes may also ignite wildland fires. Be sure to click on the links at the bottom to see animations of the above images which show multiple lightning strikes around the time shown above, the initial indications of the hotspot 2 days later along with the NGFS initial hotspot detection at 2340 UTC on July 4. This case is not unusual in the lag time of a couple days between lightning strike and wildland fire development, the strike can cause an area to smolder for a period of time before atmospheric conditions become favorable to increase the probability of a wildland fire. These type of fires are commonly known as Holdovers. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Monitoring for cold fronts is critical because firefighters need to know of the associated wind speed increase and change in wind direction that accompanies cold fronts. Let’s analyze the cold front that affected the Smokehouse Creek Fire in Texas on February 24, 2024. The 4 panel display shows GOES-16 imagery and products, upper-left panel is the 10.3 micron IR imagery, upper-right panel is the 3.9 micron imagery, lower-left panel is the Fire Radiative Power product, commonly referred to as FRP, and the lower-right panel is the visible imagery overlaid with METARs. First, let’s identify where the fires are. Look in the GOES 3.9 micron imagery and FRP product to identify the numerous hotspots in the northern Texas panhandle and also northwest Oklahoma later in the loop. By midway through the animation we can see a well defined cold front oriented east-west in Kansas and Colorado moving southwards towards the fires in Texas and Oklahoma. The cold front is very noticeable with the wind shift from southwest winds to north winds in the METARs, the cooler temperatures behind the front also show up in the GOES 10.3 and 3.9 micron imagery. The strong winds along and just behind the cold front as well as the change in wind direction is key information to communicate to the firefighters. Next we’ll look at outflow boundaries which can have equally important effects on fires, but typically with less lead time compared to cold fronts since they are a result of convection.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="View the YouTube video link, afterwards you may view the animations in more detail by clicking on the animation links.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="* video *">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
<Span Text="
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Monitoring for potential outflow boundaries is critical because firefighters need to know of the associated wind speed increase and change in wind direction that accompanies outflow boundaries. Let’s begin with the GOES-16 imagery, 3.9 micron band in the upper-left panel, the visible imagery in the upper-right panel, the fire temperature RGB in the lower-left panel and the IR imagery in the lower-right panel. First let’s identify the fire of interest, although we can see a hotspot located here , we’re going to focus on the hotspot further west located here, this is associated with the Lake Christine fire in western Colorado. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="In the visible and IR imagery we observe convection over this region which led to the development of an arc shaped line by later in the animation, this is the outflow boundary that developed from the aggregate of convection. You can see the outflow boundary moving westward, towards the Lake Christine fire. This is occurring near sunset, therefore we’re losing light in the visible imagery, the IR imagery is taking higher priority as we approach darkness. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Now let’s look at radar data from the nearest available radar located at Grand Junction, Colorado located here. The radar data shown is from the 0.5 degree tilt, reflectivity in the upper-left panel, velocity in the upper-right panel, correlation coefficient in the lower-left panel and for comparison purposes I show the CIMSS natural color RGB in the lower-right panel from GOES-16. The convection in the reflectivity is not that strong, however, in this dry environment it is sufficient to generate outflow boundaries. Multiple storms combine to form an outflow boundary that forms in a north-south oriented arc line and moves westward. The signature is somewhat subtle from the radar perspective, but when combined with the satellite data provides a clear indication of the location and movement of the outflow boundary. The outflow boundary becomes better defined in radar later in the animation, as it gets closer to the radar site, at the same time we are losing detection of it in the natural color RGB as we approach sunset. This is a great example of data fusion, with radar data increasing in priority as visible imagery from satellite is decreasing in priority. The power of the combined dataset offsets limitations associated with sunset and radar range. Surface observations are important as well, but keep in mind the elevation may vary considerably between sites in complex terrain regions such as here. The outflow boundary caused wind gusts in the 40 to 50 MPH range and led to rapid fire spread as well as unexpected fire spread direction. Active monitoring for potential outflow boundaries and communication with firefighters on scene is critical.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="View the YouTube video link, afterwards you may view the animations in more detail by clicking on the animation links.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="* video *
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Let’s look at another outflow boundary case that impacted a fire. We’ll focus on the late afternoon to evening hours of June 30, 2025 and focus on the Oak Ridge fire in eastern Arizona. This 4 panel display of GOES imagery shows the natural color RGB in the upper-left panel, the 3.9 micron imagery in the upper-right panel, the fire temperature RGB in the lower-left panel and the IR band at 10.3 microns in the lower-right panel. First, let’s identify the Oak Ridge fire being associated with the hot spot here. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="The hot spot appears intermittently since there are convective clouds which caused obscuration at times. Early in the loop we observe convection in western New Mexico, this led to the development of an outflow boundary that was approximately north to south oriented and moved westwards, into Arizona. I’ll stop the loop to annotate the outflow boundary.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Winds were strong just behind the outflow boundary, which caused rapid fire spread as can be seen with a pyrocumulus burst in the natural color RGB and an increase in magnitude of the hot spot temperature as detected in the 3.9 micron band. I’ll focus in on the pyrocumulus burst.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="If we go ahead and focus in on this region, we can see a pyrocumulus burst right here. There's the outflow boundary, so we have strong winds immediately behind the outflow boundary and that helped to cause rapid fire spread and associated with that this pyrocumulus burst and also we saw a flare up in the 3.9 micron brightness temperatures.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Next, we’ll consider the radar perspective of this outflow boundary. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Before we show output from radar let’s look at the MRMS Radar Quality Index from the nearest radars. To the west we have the radar in Flagstaff, Arizona and to the east we have the radar in Albuquerque, New Mexico. The location of the Oak Ridge fire is here, which is far from either radar site, resulting in low radar quality index values. Keep this in mind as we look at data from each radar site.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="We’ll start with data from the radar site in Albuquerque since this is closer to where the outflow boundary originated. All radar data shown here is from the 0.5 degree tilt, reflectivity in the upper-left panel, velocity in the upper-right panel, correlation coefficient in the lower-left panel and GOES natural color RGB in the lower-right panel for comparison purposes. Early in the loop, we can see the convection in the reflectivity field that led to the outflow boundary. Indications of the outflow boundary begin to appear near the end of the animation in the vicinity of the Arizona / New Mexico state border but is subtle, especially when compared to the natural color satellite imagery. Keep in mind the outflow boundary is moving away from the radar in time, making it less likely to be detected.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="We switch to the Flagstaff, Arizona radar looking at the same fields which depicts the outflow boundary in the later part of this animation. In this situation, the outflow boundary gets closer to the radar site in time, making it more likely to be detected. This case illustrates the challenge of detecting outflow boundaries by radar due to significant range from the nearest radar site, a typical issue in the western US. Be sure to use satellite and surface observations in tandem with radar observations for the best possible detection of outflow boundaries that can impact firefighting operations.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="View the YouTube video link, afterwards you may view the animations in more detail by clicking on the animation links.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="* video *
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="Infrared instruments will simply not be able to detect hotspots from wildland fires when clouds are optically thick enough. However, in a partial cloud coverage scenario such as shown here over southern California on December 11, 2024 we can still find some value from the 3.9 micron imagery. Early in the animation, a hotspot from a wildland fire can be observed at this location. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="However, optically thick mid to high level clouds advect over the hotspot, preventing detection. In these situations, existing fires can be detected by insuring a long enough loop duration to span during time periods where clear skies existed. However, new fires would still be challenging to detect.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="One actionable item for the mesoanalyst could be to predict likely changes in obscuration and help prepare for alternative ways to detect hotspots, perhaps even putting out a call to notify partners of increased difficulty in hotspot detection.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Radar may be used to detect smoke plumes from wildland fires, however, detection is hit or miss and generally less reliable at detecting fires relative to satellite data. Detection of a smoke plume by radar depends on radar range, potential beam blockage and depth of the smoke plume. Smoke plumes in conditions of very strong low-level winds tend to be not as deep as smoke plumes in weaker low-level winds. Shallow smoke plumes will be more difficult to detect by radar due to beam overshooting, particularly as range from radar increases. Because of these limitations, one cannot conclude that a wildland fire does not exist in the absence of a smoke plume detected by radar.">
<Style FontFamily="+minor" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="View the YouTube video link, afterwards you may view the animations in more detail by clicking on the animation links.
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="21" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="* video *
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" Color="#000000" BulletFont="Arial" />
</Style>
<Span Text="The case we’ll look at now is the Franklin fire that affected the area around Malibu, California in December 2024. Upper-left panel is the GOES-18 3.9 micron imagery, upper-right panel is the half degree reflectivity from the KVTX radar located northwest of the scene, lower-left panel is the same field except from the KSOX radar located southeast of the scene, lower-right panel is the GOES fire temperature RGB along with METARs. The time span of this animation shows ignition of the fire along with rapid spread and growth, detected in the GOES imagery since there was no cloud obscuration. Radar will typically detect the smoke plume later than initial detection of the hotspot from satellite. Since radar range and potential beam blockage are important considerations in detection by radar, we see differences in the appearance of the smoke plume between the two radars. In this case, KVTX shows the smoke plume better, likely due to less range relative to the other radar. 
">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s move to later in the day, the hotspot is not as robust as earlier from the GOES perspective, keep in mind fire mitigation efforts have been ongoing for a while by this time. As for detection of the smoke plume by radar, it is briefly discernible from the KVTX radar, but not readily seen from the KSOX radar. In this case, the smoke plume is relatively shallow due to strong low-level winds, the radar beam is likely overshooting the smoke plume most or all of the times shown depending on which radar you use. This illustrates an important point, under clear skies, satellite detection of hotspots associated with wildland fires is more reliable than radar detection of smoke plumes associated with wildland fires. If cloud cover exists which prevents detection by satellite, radar may be used to attempt to detect smoke plumes but keep in mind that if no smoke plume is observed one cannot conclude that no wildfire exists, limitations we’ve discussed may prevent the radar detection of the smoke plume.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="This is the hotspot from the Kruger Rock Fire near Estes Park, Colorado. The 2 panels on the left side are from GOES-East, while the 2 panels on the right are from GOES-West. The 3.9 micron imagery is shown on top while the visible imagery is shown at the bottom. The hotspot is detected by GOES-West but is not detected by GOES-East, so the question is why. Note the presence of an optically thick mountain wave cloud in the visible imagery, the cloud edge is near the hotspot location. The GOES-East perspective of the hotspot is obstructed by the mountain wave cloud, however, the GOES-West perspective offers an unobstructed view of the hotspot. Another potential factor is that the wildfire was burning on a west facing slope, this may also account for at least partial degradation of the GOES-East signal due to terrain blocking, while GOES-West did not experience this. This example illustrates why using both GOES-East and GOES-West is important. A more common situation is the availability 1-minute imagery from the mesoscale domain sector in one satellite but not with the other. You may click on the view animation link to view the animation on your own, there is also a link to a blog entry by Bill Line for more information on this case and topic.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Here's another scenario to illustrate why using both GOES-East and GOES-West is important. We’re looking at the GOES GeoColor imagery on the morning of 10 August 2025 centered over the Dakotas where smoke exists. The GOES West and GOES East GeoColor images are time matched, yet the smoke appears much thicker in the GOES-West imagery compared to the GOES-East imagery. The viewing perspective makes a difference. There is a preferred forward scattering of solar energy by smoke particles towards GOES West during the morning, which makes the smoke appear more readily visible and thicker compared to GOES East. Because of this effect, we recommend to use GOES-West during the first half of the day, and to use GOES East during the second half of the day when attempting to identify smoke.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Next we’ll look at the March 14, 2025 fire event over Oklahoma. Extreme fire weather conditions were forecast for this day with very strong winds. The discussion in this case with regards to the fire warning decision process is meant to show how data fusion considerations are involved in the hotspot detection. To understand this process requires us to show how the hotspot detection process fits inside the fire warning decision-making for WFO Norman. Their process may change should fire warning operations become operational throughout the NWS.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="The GOES GeoColor imagery shows a large region of blowing dust from the Texas panhandle advecting into Oklahoma. There’s another region of blowing dust behind a cold front from western Kansas into northwest Oklahoma later in the animation. Smoke plumes associated with wildland fires are observed in Oklahoma, note that these plumes are colored gray, while dust is brown and clouds are white.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s demonstrate how satellite-based fire detection is successfully used in the Southern Plains. This is part of a process known as the integrated warning team fire warning approach used in the Southern Plains. The motivation in discussing this is to show in what part of this process NWS meteorologists add value via satellite-based fire detection and monitoring. I’ll briefly summarize the IWT FRW workflow so that you can understand the context of where satellite-based fire detection lies within the bigger picture. The first step in this approach is a pre-fire environment coordination between the NWS and state officials on the predicted fire environment. The next step is the candidate fire identification so we’re at this step in the process. NWS meteorologists provide satellite-based fire detection and monitoring information via hotspot notifications. Updated hotspot notifications of potentially dangerous fire detected are sent to state officials when warning candidate fires exceed decisional guidance thresholds within the previously coordinated environmental conditions. I am going to spend some time going into more details of the decisional guidance thresholds on the next several slides since this is where NWS meteorologists add value to satellite-based fire detection. The third step in this approach is the state agencies corroborate on-the-ground fire behavior and public risks via on-site resources. If the fire is deemed an imminent threat to life and property, state officials (that is, fire analysts if you’re in Oklahoma) will request NWS dissemination of a fire warning.">
<Style FontFamily="+minor" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Past research has developed decisional guidance of hotspots based on simulations of fire activity by both meteorologists and fire analysts to establish environmental parameters and satellite-based remote sensing thresholds indicative of extreme fire behavior in the southern Plains. Fire weather in the Southern Plains is based on the Red Flag Threat Index, or RFTI, with categorical threats ranging from NIL at a value of zero to Historic at a value of 10. Fuels in the Southern Plains are based on the Energy Release Component, or ERC, Percentiles with drier fuels indicated by higher ERC percentile values. Since this was developed for the southern Plains the thresholds will vary geographically with different fuel types, for example, the numbers will likely be different in prairie grassland versus dense tall forests common in the western US. The motivation for introducing this approach is to keep in mind that the framework can be applied to anywhere, what will vary are the thresholds which will be governed by not only past research but future research and feedback. Perhaps future research will even find Fire Radiative Power to be a more useful metric, but it would still be applied within this framework of a combined fuel information, fire weather and satellite-based remote sensing thresholds.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="The thresholds in the lowest end of the decisional guidance are colored in purple. These are either low end fire weather conditions or low end fuels. There are no thresholds in the purple zone since fire warnings would only be issued at the request of the local emergency manager. For the boxes colored in yellow, candidate wildfires for an IWT fire warning would be brightness temperatures of at least 95 degrees Celsius for the wildland urban interface or where values are at risk, generally speaking, more highly populated regions. That threshold increases to brightness temperatures of at least 115 degrees Celsius for wildland intermix which is a rural, low population density region with sparsely dispersed infrastructure such as open rangeland.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="For the boxes colored in red, candidate wildfires for an IWT fire warning would be brightness temperatures of at least 55 degrees Celsius for the wildland urban interface or where values are at risk. That threshold increases to brightness temperatures of at least 70 degrees Celsius for wildland intermix. For demonstration purposes, I’m going to show how to make use of this decisional guidance during the March 14, 2025 fire event for the fire that impacted Stillwater, Oklahoma.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="First, we’ll look at the fuels, for the Energy Release Component (ERC) we bring up this map from the day before the event. Values in central Oklahoma were primarily in the 60 to 70 percentile range, however we do see a local maximum around Stillwater in the 70 to 80 percentile range suggesting percentiles above 70 are possible. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Recall earlier I showed that this event was forecast to be in the extreme category in terms of fire weather conditions, that’s based on RFTI values. For the ERC Percentiles, the larger scale was in the 60 to 70 percentile range, however we did see evidence of locally higher values above the 70th percentile around Stillwater. Combine the Extreme category and the ERC percentile category in the 70th to 80th percentile range and that puts us here, within the red category for threshold guidance. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s see what’s happening from a broader perspective by considering this zoomed out perspective of GOES 3.9 micron imagery over an extended time period. There are numerous hotspots associated with wildland fires across the state. Our focus will be on the fire that affected Stillwater, but I also wanted to point out the fires at Acadia and Blackburn that we’ll discuss in more detail later. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="This is the circle I used to construct the time series plot with the AWIPS tracking meteogram tool of 3.9 micron brightness temperatures I’ll show in the next slide. My goal is to capture time trends of the Stillwater fire, but in the animation shown in the link below you will see multiple hotspots at times which complicates analysis to some extent. In the animation, focus on the 3.9 micron brightness temperatures, I also show FRP values for reference, later I’ll discuss the dust and SO2 RGB imagery when we discuss the blowing dust in more detail. ">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s apply the thresholds we discussed to a time series of GOES 3.9 micron brightness temperatures. I show horizontal lines for both the Wildland Intermix and WUI thresholds, but for our case with the fire near Stillwater this is a wildland urban interface scenario so we will focus on the WUI threshold shown in the dotted horizontal line. The fire near Stillwater was first detected in the NGFS at 1816 UTC, the brightness temperatures reached the WUI threshold by about 1852 UTC, 36 minutes after initial detection. Soon thereafter, at 1902 UTC, a potentially dangerous hotspot (PDS) notification was sent. The IWT fire warning coordination occurs during this time period, the local fire official would then make a decision whether to request NWS dissemination of a fire warning. The fire warning was issued at 1918 UTC, you can check out the polygon associated with that by clicking on the link that says graphic at number 5. Soon after the initial fire warning, evacuation orders were issued at 1930 UTC followed by moving non-ambulatory patients from areas threatened at 1943 UTC. Later on, we see additional fire warnings and evacuation orders as the fire threatened the western portions of Stillwater. I also list the times of smoke plume detection from radar and satellite, keep in mind the lag between hotspot identification and initial indications of the smoke plume. This example demonstrates the framework of using a combined fuel information, fire weather and satellite-based remote sensing thresholds for a fire event in Oklahoma. One more thing I want to point out, note the trend in decreasing brightness temperatures during this time period I highlight with a green box. This doesn’t necessarily seem intuitive since the fire was significant at this time, with additional fire warnings and evacuations along with numerous structures being burned near Stillwater. Fire mitigation efforts were ongoing during this time but also during this time recall that we had thick dust moving in over the area, let’s look at that in more detail next.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="This 4 panel display shows different GOES RGB products that detect blowing dust. In the upper-left panel we have the color deficiency dust RGB which highlights higher dust concentration in yellow. In the upper-right panel we have the Dust RGB which highlights higher dust concentration in magenta. In the lower-left panel we have the SO2 RGB which highlights higher dust concentration in dark blue. In the lower-right panel we have the CIMSS Natural Color RGB which highlights higher dust concentration in brown. The imagery shows increasing dust concentration over the fires of interest in central Oklahoma with the highest concentration of dust arriving later in the animation. This is a strong signal of dust from these products which highlights the anomalously high concentration of dust for this area.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Here's the time series plot of GOES 3.9 micron brightness temperatures for 2 other fires in Oklahoma with locations shown previously on slide 4.9. Much like the time series plot for the Stillwater fire, we also see a period of decreasing brightness temperatures during the afternoon hours. I note the times on both plots when the dust signal was strongest based on the satellite imagery on the previous slide. While fire mitigation efforts were ongoing for all 3 fires, it is unlikely that this was the primary cause of the decreasing brightness temperatures for all 3 fires. The most likely primary cause of the decreasing 3.9 micron brightness temperatures is from the increased concentration of dust. This makes the point that for tracking hotspots from satellite imagery, not only do we have to be aware of potential cloud cover that leads to obscuration of the hotspot, but also for high concentrations of dust that can lead to a degradation of the hotspot signature of interest.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Recall when I introduced the hotspot decisional guidance that this will vary geographically by fuel type. I want to illustrate a case in California to emphasize how much variability exists across the US in terms of satellite detected hotpots which can influence how you make use of hotspot decisional guidance. This is the time series of GOES 3.9 micron brightness temperatures for the Hughes fire on January 22, 2025. Initial detection of the fire occurred about 1835 UTC, about 20 minutes later brightness temperatures already reached the saturation temperature. This case illustrates how quickly threshold brightness temperatures can be reached in high end extreme fire weather conditions. This fire occurred in very different fuel types compared to the event we just looked at in Oklahoma, also this occurred in complex terrain, not the relatively flat Plains. The Fire Behavior Triangle conveys that in order to assess wildland fire behavior, you need to account for the fuel, topography and weather. In a situation like this the early identification of the hotspot and communication with the appropriate agencies is the best approach to deal with wildland fires that will most likely grow and spread very quickly. The NWS WFO in Oxnard, California issued their first fire warning in conjunction with the California Governor’s Office of Emergency Services for this event.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Let’s review our 4 major data fusion considerations from this training session. First, for our anticipating data sources, initial hotspot detection is best accomplished through GOES 3.9 micron imagery for subjective analysis while objective analysis is available via the NGFS. For smoke plume detection, certainly webcams are the most useful but they’re not available everywhere so that’s where satellite data is useful, the GOES GeoColor and Natural Color RGBs in particular. If obscuration by cloud or dust is an issue, radar can be useful within a reasonable radar range. For cold front / outflow boundary monitoring, certainly surface observations are key, along with radar data and GOES visible, 3.9 or 10.4 micron imagery. Blowing dust can degrade the signal from the hotspot making it more challenging to analyze, we showed multiple GOES RGB products that can help with blowing dust identification. For data constraints, we considered the GOES East versus West perspective for the preferred forward scattering in detecting smoke. We discussed radar range limitations and we also mentioned that the GOES 3.9 micron band does saturate at 138.7 degrees Celsius, although it may not be operationally relevant to detect changes above that temperature anyway. For confidence in evaluation, we discussed cloud and dust obscuration and noting when it is present that webcams become higher priority and potentially radar so long as range is not an issue. We discussed identification of cold fronts and outflow boundaries for the hazards they present when interacting with a wildland fire, certainly sparse surface observations can challenge your confidence in evaluation. Radar can be useful except when range becomes sufficiently far, beam overshooting occurs. In general, outflow boundaries are more challenging to identify compared to cold fronts due to their transient nature resulting from convection. For data quality issues, the most important one we covered was GOES hotspot obscuration due to cloud cover and also dust, be aware of what other tools to use when obscuration becomes an issue.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Here’s an example of what data fusion in action looks like at NWS WFO Norman during the March 14, 2025 fire event we discussed earlier. Lead forecaster Kaitlin Schueth is making use of GOES 3.9 micron imagery for 2 sectors monitoring hotspots on the center screen, to the left we see the local obs monitor along with the NGFS dashboard, the hotspot notification tool and of course communication tools like NWS chat and NWS collab. In the background we see Meteorologist in Charge Mark Fox and Chris Mask from Oklahoma Forestry Services.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="Ultimately, all of what we’ve discussed can be used in making decisions for fire warnings. For example, we see a combination of GOES 3.9 micron imagery and fire spread model output which I’ll highlight with these arrows. The fire warning polygon being drawn up by the forecaster is highlighted with these arrows.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document><Document xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Content>
<Block>
<Style FlowDirection="LeftToRight" LeadingMargin="0" TrailingMargin="0" FirstLineMargin="0" Justification="Left" ListLevel="1" LineSpacingRule="Multiple" LineSpacing="23" SpacingBefore="0" SpacingAfter="8">
<ListStyle ListType="None" ListTypeFormat="Parentheses" BulletFont="Arial" />
</Style>
<Span Text="In conclusion, we made use of satellite imagery in tandem with webcam images serving as ground truth data to demonstrate that satellite imagery is a reliable way to identify new fires via hotspot monitoring. This can be done subjectively using GOES 3.9 micron imagery, or objectively like we demonstrated with the NGFS which is an automated algorithm. In either case we do have to be aware of potential obscuration which can prevent detection. Next, we went through some wildland fire considerations including awareness of obscuration from clouds or dust. We also compared satellite with radar and how they can be used in a complementary way, especially when limitations make identification more challenging in one of these. We also looked at cases with a cold front and outflow boundaries since these present significant hazards to firefighters present on an incident. Finally, we demonstrated satellite-based fire detection when fires exceed decisional guidance thresholds in the context of fire warnings for a case in Oklahoma. The motivation for introducing this approach is to keep in mind that the framework of using a combined fuel information, fire weather and satellite-based remote sensing thresholds can be applied to anywhere. However, keep in mind, that thresholds will evolve with future research and feedback and also vary geographically.">
<Style FontFamily="Calibri" FontSize="12" />
</Span>
</Block>
</Content>
</Document>