Multispectral Satellite Applications: Monitoring the Wildland Fire Cycle
Print Version
Produced by the COMET® Program

Quick links to pages:

Welcome

Section 1: Introduction
1.1  Role of Fire
1.2  Trends in Fires
1.3  Wildfire Management
1.4  Satellite Detection and Analysis of Fires
1.5  Fire Cycle Overview
1.6  Other Types of Fire
1.7  Nighttime Lights
1.8  About the Module

Section 2: Satellite Imagers, Present & Future
2.1  Satellite Detection of Fires
2.2  Geostationary and Polar-orbiting Satellites
2.3  Geo vs. Polar-orbiting Satellites: Geostationary
2.4  Geo vs. Polar-orbiting Satellites: Polar
2.5  Polar-orbiting Satellites: Improvements with NPOESS VIIRS
2.6  Satellite Application Table

Section 3: Pre-Fire Monitoring and Post-Fire Analysis
3.1  About the Section
3.2  Surface Observations for Fire Danger Ratings
3.3  Monitoring Vegetation with True Color Imagery
3.4  NOAA AVHRR NDVI-derived Greenness Products
3.5  Post-fire Scarring Using False Color Products
3.6  Section Summary

Section 4: Fire Detection and Monitoring
4.1  Detecting Fires
4.2  Shortwave vs. Longwave Imagery
4.3  Sub-pixel Effect, 1
4.4  Sub-pixel Effect, 2
4.5  Limitations in Fire Detection
4.6  DMSP OLS Nighttime Lights & Fires
4.7  MODIS True and False Color Products
4.8  False Color Products
4.9  False Color vs. Aircraft Data
4.10  Hazard Mapping System
4.11  WFABBA and MODIS Rapid Response System
4.12  MODIS USFS Active Fire Map
4.13  Section Summary

Section 5: Smoke & Aerosol Detection, Monitoring & Transport
5.1  Section Overview
5.2  Visible Animations
5.3  VIS vs. IR
5.4  Black & White vs. True Color Imagery
5.5  Smoke Detection at Night
5.6  Smoke Forecasting
5.7  Automatic Smoke Forecasting
5.8  Case Example
5.9  WFABBA and FLAMBE
5.10  Forecasting Smoke Advection and Aerosol Optical Thickness Using NAAPS
5.11  Section Summary

Section 6: Fire Product Suite
6.1  Introduction
6.2  Fire Danger Rating Product
6.3  NOAA AVHRR NDVI-derived Greenness Maps
6.4  MODIS Active Fire Map
6.5  Fire Detection Algorithms and Products
6.6  Geostationary and Polar-orbiting Animations
6.7  MODIS True Color Product
6.8  MODIS False Color Product
6.9  Smoke Forecasting Products

Section 7: Case Study
7.1  Introduction
7.2  Advantages of True Color Imagery
7.3  Relative Greenness
7.4  Relative Greenness Question
7.5  Hotspots
7.6  Position of Flames
7.7  Fires in GOES Animations
7.8  MODIS/Reflectivity Loop
7.9  MODIS Active Fire Maps
7.10  False Color Imagery, 13 July
7.11  False Color Imagery, 14 July
7.12  True and False Color Imagery
7.13  Comments

Section 8: Module Summary


Welcome

This module describes current and future satellites sensors and products used for monitoring the fire cycle, with an emphasis on polar-orbiting satellites. Product information is presented in the context of the stages of the fire cycle and is consolidated in the Fire Product Suite, which is also available as an independent resource. The module concludes with an interactive fire case study, supplemented with observations from a National Weather Service trainer/forecaster who experienced the fire. The module is intended for the wide range of users involved with wildfire detection and monitoring–from land use managers and hydrologists to weather forecasters and researchers.


1.0 Introduction

1.1 Role of Fire

Fire has been mankind's companion since prehistoric time. It enabled early peoples to travel from temperate regions to the colder regions of the mid and upper latitudes.

World map showing general patterns of human migration

Fire has long served as both a useful tool for civilization and an object of fear when out of control.

Photograph of Moran Fire

Fires drastically alter the landscape, bringing unwanted destruction as well as ecological renewal. In many parts of the world, fire is used to clear land for agriculture and other purposes.

picture of burned area at Mesa Verde

picture of regeneration after a forest fire

Fires occur everywhere - in forested, rural, and urban regions, at high latitudes, and in the tropics.

MODIS Rapid Response System Global Fire Map

To top


1.2 Trends in Fires

The pattern of wildfires follows the seasons, with fires breaking out in the northern midlatitudes with spring warming and disappearing when fall and winter weather brings cooler temperatures, increased moisture, and more widespread precipitation.

MODIS Rapid Response Fire Detections for 2006

(View animation)

Weather satellites are useful tools for monitoring seasonal fire trends. This monthly animation shows global fire detections in 2006 using the MODIS satellite, with each fire indicated by a red marker. Large blocks of red make it appear as though entire portions of the tropics are on fire, which is fortunately not the case. We will uncover the reason for this discrepancy when we study how satellites detect fires.

The prevalence of wildland fires also changes based on short- and long-term variations in climate, which satellites help to monitor. Some researchers predict that the warming associated with global climate change will cause the number of catastrophic fires to increase dramatically. As this chart shows, the number of acres burned annually in the United States has risen in the last ten years.

MODIS Rapid Response Fire Detections over a 10 year period

To top


1.3 Wildfire Management

Wildfire suppression costs the U.S. federal government in excess of $1 billion per year. Despite the use of radio communications, rapid helicopter transport, and new types of chemical firefighting apparatus, over four million acres of U.S. forests burn in a typical year.

collage of pictures of fire scenes

Although many wildland fires are started by lightning, most are caused by humans, whether intentionally or by accident.

pictures of lightning

To top


1.4 Satellite Detection and Analysis of Fires

Local and regional fire management agencies use a variety of systems to detect fires, such as aircraft observations.

Cedar Fire, San Diego County, CA 26 October 2003 as viewed by the FireMapper Thermal-Imaging Radiometer

For fire detection and monitoring at continental and global scales, only satellites can provide the required coverage and consistency.

Geostationary satellite coverage with Meteosat

Even at regional scales, satellites offer fire fighting agencies, weather forecasters, researchers, and the public a unique and timely perspective on fires.

True Color Image of fires in the Northern Rockies from MODIS Rapid Response aboard the Aqua satellite, 8/15/2005

To top


1.5 Fire Cycle Overview

Satellites allow us to observe the fire lifecycle—from the pre-fire environment to the onset and spread of fires, the transport of smoke, and the modifications to the landscape in the wake of fires. Let's take a brief look at the use of satellite data in monitoring each phase of the cycle.

collage of the phases of the fire cycle

Pre-fire conditions: Satellite products are used to observe pre-fire conditions and spot trends in fire susceptibility. This MODIS image shows southern California before the onset of a major fire episode. The brown chaparral hints at the tinder dry fire fuel over the hillsides. With the forecast of hot, dry Santa Ana winds, it will take only a spark to create a roaring fire.

Algal bloom in the Salton Sea, California, 2003/294 - 10/21 at 18 :20 UTC,  MODIS on Terra satellite

Fire danger: Dry vegetation over the hills is confirmed by the observed fire danger classification product produced from in situ observations on the ground. Reds and oranges indicate that the fire danger in southern California is very high or extreme. Later, we will see how satellite products can work together with in situ observations to assess the risk of fires.

Observed Fire Danger Class Product 21 Oct 2003

Burning fires: The MODIS false color image reveals the major fire fronts in orange. The MODIS true color image shows smoke plumes. The overlay of freeways and place names expands the usefulness of these products to fire fighters, media, nearby residents, travelers, and others. We will cover true and false color imagery in more detail in the module.

Terra MODIS false color image, Southern California, October 26, 2003

Terra MODIS true color image, Southern California, October 26, 2003

Smoke detection: This zoom shows how well suited true color imagery is for sensing the finer details of smoke. It can capture the texture, thickness, and extent of smoke plumes, especially in contrast to darker surfaces such as oceans.

Terra MODIS true color image, Southern California, October 26, 2003, zoom in on smoke

Fire aftermath: Satellite products also help us observe post-fire changes in vegetation, erosion potential, and the prospect for ecological renewal. Notice the burn scars in dark pink. Within them, vegetation is charred, the ground surface is often exposed, and the likelihood of erosion, including mudslides, is high.

MODIS False Color Product showing burn scars over So. CA, 26 Oct 2003

To top


1.6 Other Types of Fires

Not all fires fit into the concept of the fire cycle. During and after the 1991 Persian Gulf War, NOAA Polar-orbiting Operational Environmental Satellite System (POES) satellites were used to monitor wildfires, oil fires, and smoke.

This NOAA AVHRR image shows hotspots from crude oil burning in red and smoke from oil wells in black.

NOAA-10 AVHRR image showing hot spots and smoke from oil wells burning in Kuwait on 23 February 1991

To top


1.7 Nighttime Lights

Some satellites can detect light sources as well as fires. This sample product is compiled from the Defense Meteorological Satellite Program (DMSP) Optical Linescan Imager (OLS) from October 1994 through March 1995. Fires stand out in red, city lights in white, gas flares from natural gas burning in green, and fishing boats in aqua. We'll discuss this remarkable satellite sensor later.

Poster of Nighttime Lights of the World: 1994-95

To top


1.8 About the Module

This module is intended for the wide range of users involved with wildfire detection and monitoring–from land use managers and hydrologists to weather forecasters and researchers. Users are expected to have a basic understanding of satellite meteorology and remote sensing concepts.

photo of land managers
photo of hydrologists at work
photograph of a military weather forecaster in his office

The module is organized as follows:

  • Section 2 (Satellite Imagers, Present & Future) describes the satellites and imagers used for detecting and monitoring the fire cycle
  • Sections 3 (Pre-Fire Monitoring and Post-Fire Analysis), 4 (Fire Detection and Monitoring), and 5 (Smoke & Aerosol Detection, Monitoring & Transport) examine the phases of the fire cycle and the associated products
  • Section 6 (Fire Product Suite) consolidates the product information presented in the module and also provides additional information, such as product derivation; this optional section can be used as refresher or reference material and is also available as a PDF file
  • Section 7 (Case Study) is an interactive activity that helps you integrate and apply what you have learned. The material is supplemented with observations from a National Weather Service trainer/forecaster who experienced the fire
  • The module ends with a summary and quiz

To top


2.0: Satellite Imagers, Present & Future

2.1 Satellite Detection of Fires

The keys to detecting fires are imagers on board geostationary and polar-orbiting satellites that are sensitive to thermal emissions in the 3.5 to 4.0 micrometer shortwave infrared region. The sensors have channels in this range that can detect hotspots at varying spatial resolutions.

VIIRS/MODIS key wavelengths

Although the 2.2 micrometer channel, which is available on some satellites, can detect very hot fires, the 4.0 micrometer region is much more effective and widely used for the routine detection of small fires.

The graphic shows the current and future satellites involved with monitoring the fire cycle. The previous generation of weather satellite sensors was not launched with fire monitoring in mind. In some cases, the capacity to detect hotspots was discovered only after the satellite became operational, as in the case of NOAA Advanced Very High Resolution Radiometer (AVHRR). Now visible and infrared sensors are explicitly designed with fire detection capability.

Weather satellites with fire detection and monitoring capabilities

To top


2.2 Geostationary and Polar-orbiting Satellites

Both geostationary and polar satellites are powerful fire-monitoring tools, each with its own capabilities and limitations.

artistic rendition of geostationary and polor-orbiting satellite coverage of the Earth

Geostationary satellites provide excellent coverage over the tropics and midlatitudes (but not polar regions). The ability to image repeatedly over a fixed geographical region enables geostationary satellites to:

  • Monitor the thermal signatures associated with fires
  • Track smoke over thousands of kilometers
  • Observe diurnal processes that affect fire behavior over many parts of the globe
  • Through sequential observations, detect fires temporarily obscured by clouds

Most modern geostationary satellites have a shortwave infrared or ‘fire' channel that is sensitive to heat sources on the ground.

artist rendering of geostationary satellite coverage with Meteosat

In contrast, polar-orbiting satellites have:

  • Higher spatial resolution for more detailed images
  • In some cases (like that of the MODIS imager), more spectral channels for better products
  • The ability to monitor high-latitude regions that have a surprisingly large number of fires

artistic rendition showing the swaths of polar-orbiging satellites

To top


2.3 Geo vs. Polar-orbiting Satellites: Geostationary

Let's compare geostationary vs. polar capabilities for a particular fire. This GOES loop captured the development of the 2001 Chisholm fire in Alberta, Canada. The heat from the fire catalyzed a pyrocumulonimbus (pyroCb), which is a deep convective cloud system. The loop consists of visible data during the daytime and infrared data at night.

GOES loop of pyroCb from Chisholm fire (28 - 29 May 2001)

(View animation)

The white hotspots, overlaid using information from the GOES fire channel, show fire locations. The pyroCb erupted in two pulses before dark, marked by the switch from visible to infrared imagery. GOES loops cannot view the fires in great detail, especially at higher latitudes where geometric distortion is large. But GOES and other geostationary satellites have the advantage of frequent scanning, which enables us to view the feature in time lapse mode.

Notice the cold anvil cloud tops that blow off and persist for the entire night and the ‘billiard table' smooth gray cloud at daybreak. This is the smoke generated by the convection the evening before. Based on subsequent observations, we know that the pyroCb injected smoke into the stratosphere. Although this is not common with most wildfires, it does occur with pyroCbs in northern latitudes.

To top


2.4 Geo vs. Polar-orbiting Satellites: Polar

Now let's view polar imagery of the same fire. We'll look at four high-quality images from a two-hour period. This first image, from the DMSP OLS satellite, shows substantial low-level smoke from the growing Chisholm fire.

DMSP OLS image of Chisholm (Alberta) PyroCb, 28 May 2001

This NOAA AVHRR product was taken two hours later, with embedded hotspots overlaid in red on the visible image. The flat bubble in the center represents the massive pyroCB cloud that developed from the fire.

NOAA AVHRR image of Chisholm (Alberta) PyroCb, 28 May 2001

Two more DMSP images were taken only minutes later and show the advantage of frequent overpasses that comes from using polar-orbiting data at high latitudes.

DMSP OLS image of Chisholm (Alberta) PyroCb, 28 May 2001

DMSP OLS image of Chisholm (Alberta) PyroCb, 28 May 2001

To top


2.5 Polar-orbiting Satellites: Improvements with NPOESS VIIRS

In the NPOESS era, the Visible Infrared Imager Radiometer Suite (VIIRS) sensor will combine and improve upon the capabilities of the civilian NOAA AVHRR instrument and the military DMSP OLS sensor, with enhanced functionality.

NPOESS satellite with VIIRS highlighted and the NOAA AVHRR and DMSP satellites whose capabilities it will be replacing

VIIRS will continue monitoring all phases of the fire cycle. The new imager will have sensing channels and capabilities similar to MODIS, with a number of improvements and advantages.

  • The NPOESS Safety Net communications design will drastically decrease data latency (the delay between satellite imaging and delivery of data to users), resulting in more timely support; 95% of data will be delivered within 28 minutes
  • New scan geometry will lead to considerably improved image quality, especially near the edge of the swath
  • Finally, a Day/Night Band (DNB) will be added, providing nighttime detection capability for fire and smoke

VIIRS will have twenty two channels ranging from the visible to the longwave infrared, nearly all of which will have a role in monitoring the fire cycle.

  • Using channels at red, blue, and green wavelengths in the visible, VIIRS will continue to provide the true color capability previewed by MODIS, which forecasters are increasingly using
  • The Day/Night Band (0.7 micrometers) will be a vast improvement compared to the nighttime capability on the current DMSP OLS sensor
  • Three channels, one in the visible (0.6 micrometers) and two in the near and shortwave infrared (0.9 and 2.2 micrometers), will produce false color imagery to view burn scars and other vegetation features
  • Three channels (2.2 micrometers and two around 3.7 micrometers) will be able to detect the hotspot radiances directly associated with fires

VIIRS/MODIS key wavelengths

To top


2.6 Satellite Application Table

This table compares fire cycle monitoring capabilities of U.S. and European operational and near-operational satellites and sensors. Fire, smoke, and land applications are listed along the top, satellite sensors on the left. As you can see, the sensors on polar-orbiting satellites generally have more capability than those on geostationary satellites. But even among the polar-orbiting sensors, only the upcoming NPOESS VIIRS will be capable of performing all of the applications listed.

Satellite Application Table showing the capabilities of polar-orbiting and geostationary satellites in regard to the fire cycle

To top


3.0: Pre-Fire Monitoring and Post-Fire Analysis

3.1 About the Section

This section introduces various qualitative and quantitative products used for pre-fire monitoring and post-fire analysis. For pre-fire monitoring, we will examine fire danger ratings from surface observations, true color imagery from the MODIS instrument, and greenness maps from NOAA AVHRR data. For post-fire monitoring, we will look at false color imagery from MODIS.

If you are unfamiliar with the concept of true color and false color imagery, it is recommended that you review the background section before proceeding.

To top


3.2 Surface Observations for Fire Danger Ratings

The Fire Danger Rating level, shown here, is derived from surface observations. It takes into account current and antecedent weather, fuel types, and both live and dead fuel moisture. It can monitor trends in fire susceptibility for areas with reasonable data density, but in eastern Montana, where we'll be looking, reporting stations (shown by triangles) are sparse. Values between stations are interpolated, providing only a guess at actual vegetation moisture.

Observed Fire Danger Class Product, 05 July 2005

To top


3.3 Monitoring Vegetation with True Color Imagery

Fortunately, satellite observations can help us monitor broad areas to observe important changes in vegetation resulting from temperature and precipitation patterns (especially for grasses and shrubs). MODIS true color images convey vegetation browning qualitatively. The satellite's high spatial resolution makes the images an excellent resource in observing transitions in vegetation characteristics. In this true color image, the vegetation is still relatively green for the northern Rockies and Plains in the early summer of 2005.

Terra MODIS True Color Imagery, 05 July 2005

But after only several weeks of hot, dry weather, this next true color image shows that the vegetation has browned, suggesting significant drying and possible fire danger. In the case study for the module, we'll see how another drying episode led to the outbreak of severe fires in the same region.

Terra MODIS True Color Imagery, 28 July 2005

To top


3.4 NOAA AVHRR NDVI-derived Greenness Products

The Normalized Difference Vegetation Index (NDVI) is a simple calculation that describes the degree of lushness on the Earth's surface as observed by a satellite. NDVI products are composited weekly from NOAA AVHRR data and show vegetation trends more quantitatively than imagery. NDVI measures how dense and green plant leaves are, which indicate overall vegetative health and susceptibility to burning. Three types of products are available. In this section, we will examine the relative greenness product. The others are discussed in the Fire Product Suite.

Relative Greeness Product: 29 July 2005 to July 05 2005

This relative greenness product corresponds to the early summer true color image we saw previously and compares vegetation greenness to maximum greenness over the period from 1989 to 2003. A value near 100% indicates that an area is nearly as green as the maximum greenness value. A value near zero indicates that it is very brown in comparison. Notice the green over Montana in early summer, indicating greener than normal vegetation and relatively low fire danger.

Relative Greeness Product: 29 July 2005 to July 05 2005 with true color image overlaid

A few weeks later, the relative greenness map has turned from green to red and orange over much of the same region, suggesting a shift to browner and drier vegetation, which is similar to the trend we saw in the true color imagery.

Relative Greenness Product: 29 July 2005 to July 05 2005 with true color image overlay

To top


3.5 Post-fire Scarring Using False Color Products

We have seen how MODIS true color images can depict trends in vegetation and therefore fire risk.

Fires, smoke, and burn scars in Alaska and Yukon Territory, true color image, Aqua MODIS, 10 August 2005

Now we will see how MODIS false color images can help depict the aftermath of a fire.

Fires, smoke, and burn scars in Alaska and Yukon Territory, false color image, Aqua MODIS, 10 August 2005

These true and false color images show Alaskan forest fires in August 2005. Notice how the smoke obscures the surface on the true color image, but not on the false color image. It is much easier to see burn scars from past fires on false color imagery. They appear as dark red or brown in high contrast to the green vegetation.

Here's a zoomed-in view of the false color image. The burn scars from past and still-burning fires are scattered over the landscape. These scars often depict regions with enhanced susceptibility to erosion, alerting hydrologists to the potential for future mudslides.

Zoomed in view of burn scars from fires in Alaska and Yukon Territory, false color image, Aqua MODIS, 10 Aug 2005

To top


3.6 Section Summary

PRODUCTS USED FOR PRE-FIRE MONITORING

Fire danger ratings

  • Derived from surface observations
  • Account for current and antecedent weather, fuel types, and live and dead fuel moisture
  • Can monitor trends in fire susceptibility for areas with reasonable data density
  • Values between stations are interpolated, offering only a guess at actual fuel conditions

Terra and Aqua MODIS true color imagery

  • Convey vegetation lushness (greenness) qualitatively; the greener the scene, generally the lower the susceptibility to fires
  • High spatial resolution makes the product very useful in observing transitions in vegetation
  • For some products, fire locations are overlaid and appear in red; smoke plumes appear in gray
  • Smoke can obscure the surface, making it difficult to see burn scars and other surface features of interest

NOAA AVHRR NDVI-derived greenness maps

  • Derived from the Normalized Difference Vegetation Index (NDVI)
  • Measure how dense and green plant leaves are (indication of overall vegetative health and susceptibility to burning)

PRODUCTS USED FOR POST-FIRE MONITORING

Terra and Aqua MODIS false color imagery

  • Can ‘see through' smoke
  • Burn scars appear as dark red or brown
  • Burn scars alert officials about the potential for mudslides, plant succession, and invasive species

To top


4.0: Fire Detection and Monitoring

4.1 Detecting Fires

Perhaps the most vital role that satellites play in regard to the fire cycle is to detect the very hot, often tiny locations of fires.

Longwave infrared brightness temperatures, sensed between about 10.5 and 12 micrometers, usually provide fairly reliable estimates of, for example, cloud top temperature and sea surface temperature. However, longwave remote sensing methods are extremely limited when it comes to fires.

For that, we rely on shortwave infrared channels near 2 and especially 4 micrometers— spectral regions that are particularly sensitive to extremely hot, small fires. The channels around 4 micrometers are referred to as ‘fire’ channels.

VIIRS/MODIS key wavelengths.

In this section, we will discuss the use of longwave and shortwave imagery in more detail, describe how fire channels detect fires smaller than the size of a pixel, and examine the various products and systems used for detecting and monitoring fire as well as smoke.

To top


4.2 Shortwave vs. Longwave Imagery

This NOAA AVHRR shortwave infrared image covers the 2002 Rodeo-Chediski fire in Arizona. Warmer temperatures, including those of the fire, are white whereas colder features, such as high terrain, are dark.

NOAA AVHRR Ch 4 IR longwave image, 23 June 2006 1015 UTC over the Rodea-Chediski fire

Let's zoom in on the fire area and compare the shortwave infrared image with the longwave infrared image.

NOAA AVHRR Ch 3 near-IR shortwave image, 23 June 2006 1015 UTC, over the Rodea-Chediski fire

We've overlaid terrain contours in yellow showing the fire on the northeastern slopes of the Mogollon Rim. Due to the shortwave channel's heightened sensitivity to extreme heat, fires stand out vividly as white, whereas the longwave image shows hardly a trace of fire signatures. The red and blue lines are the positions of temperature transects through the fires.

NOAA AVHRR Ch 3 near-IR shortwave image, 23 June 2006 1015 UTC, over the Rodea-Chediski fire

This graphic shows the satellite-sensed brightness temperatures across the line segment shown in the previous images. The shortwave trace in red shows a much higher sensitivity to fire than the longwave trace in blue, with temperature difference as large as 40°C. A simple shortwave-longwave channel differencing technique is a key parameter often used by fire detection algorithms. While most algorithms are more sophisticated, all rely on the differential response to fires in the shortwave vs. longwave channels.

Multi-spectral Transects Across Fire, NOAA-16 AVHRR, 20060623 10:15 UTC

Once identified by algorithms, fire hotspot areas are often overlaid on composite imagery in red.

NOAA AVHRR Ch 3-4 23 June 2006 1015 UTC, hotspots of Rodea-Chediski fire overlaid

To top


4.3 Sub-pixel Effect, 1

As you know, satellite images are made up of tiny picture elements called pixels. Suppose for a moment that a pixel represents an area approximately 1x1 km.

graphic showing 1x1 km resolution grid over mountainous terrain

What is the smallest burning object that could show up in that pixel on a shortwave infrared image?

  1. A house
  2. A warehouse
  3. A brush fire covering several acres

Feedback: The correct answer is 1. A burning house or an even smaller object can be detected if the temperature is hot enough.

To top


4.4 Sub-pixel Effect, 2

Let's examine how satellites detect active fires and smoldering areas that are smaller in size than a pixel and some of the limitations. Imagine that this is a field of satellite pixels imaged with the infrared shortwave band. Assume that there are no active fires or smoldering areas and that the pixels have a uniform satellite temperature of 300 K, or 27°C.

A computer generated simulated image of a satellite view depicting subpixel-sized fires

These red pixels (1) indicate departures from the background temperature. In nearly all cases, this means that there is something very hot within the territory represented by the pixel. It could be a wildland fire, a house on fire, routine agricultural burning, or a battlefield fire. Or it could be a hot exhaust plume from a cooling tower or a burning-off of natural gas associated with an oil well. In general, it's impossible to tell just by looking at the image. What we can say is that the hot feature is often downright tiny compared to the pixel that contains it.

A number of hot, small fires may be contributing to the high temperature of a pixel (2). Or the satellite may not detect the fires in a pixel (3), assigning it the same satellite temperature as the surrounding pixels. This “missed detection” is caused by a fire that is too weak or small to register. Sometimes pixels have elevated satellite temperatures but no fires in them (4). This may be due to reflective clouds or sunglint off of water bodies mimicking hot features, which is called a "false alarm."

You can see the subpixel effect in real data later in this section, where a satellite depiction of fires is compared to an analysis from aircraft.

To top


4.5 Limitations in Fire Detection

There are fundamental limitations on the ability of environmental satellites to monitor fires. Which of the following are correct?

  1. The larger the satellite footprint, the less detailed the fire depictions
  2. Cloud cover can partially or completely obscure fires
  3. Satellites can only observe fires at night

Feedback: The correct answers are 1 and 2. The size of the satellite footprint (area on the ground observed within a satellite sensor's field of view) is an important limitation, with larger footprints resulting in less detailed depictions of individual fires.

artistic rendering of a satellite footprint

Cloud cover is another significant factor, since clouds can either partially or completely obscure a fire. Most satellites can observe fires both day and night.

DMSP OLS IR 23 June 06 over Rodeo-Chedeski fire

Another critical factor is the number of passes that a satellite makes over an area. Polar orbiters observe a fixed location twice daily, once during daytime and once at night, with more frequent coverage at higher latitudes.

artistic rendition showing the swaths of polar-orbiging satellites

In contrast, geostationary satellites have much better temporal coverage, on the order of 15 to 30 minutes. But geostationary satellites provide only limited coverage of polar regions.

artistic rendering of geostationary satellite coverage with Meteosat

To top


4.6 DMSP OLS Nighttime Lights & Fires

Although wildfires are most active during daytime, they can occur at any time of day. The Operational Linescan System (OLS) onboard the polar DMSP satellite can reveal important information about fires through the use of the nighttime visible channel. Although the DMSP OLS system is not used to detect fires operationally, the images preview the significant improvements that the NPOESS era will bring. In particular, the Day/Night band (DNB) will be vastly improved compared to the OLS nighttime visible channel.

This nighttime visible DMSP OLS image shows lights from flames raging out of control in southern California. But the city lights blend in with the fires in this raw image, making it impossible to distinguish fires from urban areas.

DMSP OLS Image of Nighttime Lights/Fires in So. California, 26 Oct 2003

By eliminating pixels known to be from artificial light sources, we obtain another fire identification product that isolates the fires in red. A terrain height overlay in white gives viewers additional context.

DMSP OLS Image of Nighttime Lights/Fires with channel differencing and terrain elevation added, over Southern California, 26 Oct 2003

Flames can often be distinguished from other lights through context without additional processing. For example, if you know that gas flares are common in the northern Persian Gulf, it is easy to identify them on DMSP OLS nighttime images. Additionally, if you are acquainted with the major roadways in a region, you may be able to identify the associated lighting system.

Gas Flares and City Lights over the Persian Gulf, DMSP OLS, 20 Nov 2006 1614 UTC

To top


4.7 MODIS True and False Color Products

We'll take a more operational focus now and examine products and systems used for detecting and monitoring fires, starting with true and false color products from MODIS data. These products are increasingly available online in near real time, and anticipate products that will come to forecasters even more promptly in the NPOESS era.

True and false color images are both created by combining multiple imager channels to arrive at products with more information than the individual channels would provide.

Terra MODIS false color image, Southern California, October 26, 2003

Terra MODIS true color image, Southern California, October 26, 2003

Comparing these true and false color images from the 2003 southern California fires, what general statements can you make? Select the word that correctly completes each sentence.

  1. Fires appear on (true/false) color products
  2. Smoke is (more/less) transparent in true color products
  3. Burn scars are (more/less) apparent in false color products

Feedback: The correct answers are 'false' for 1, 'less' for 2, and 'more' for 3. Fire hotspots can be detected in this false color image due to the inclusion of the fire-sensitive 2.2 micrometer channel. The hotspots appear as pink or orange dots. Smoke is much more transparent in the false color product, allowing viewers to more easily see terrain or ocean underneath. Burn scars show the history of recent fires and can be seen more distinctly on false color imagery. That's because the false color product algorithm uses channels sensitive to soil and vegetation characteristics.

Derivation of true color images and false color images

To top


4.8 False Color Products

Here are zooms of the true color and false color views from the same MODIS pass.

Terra MODIS false color product showing fires and smoke in southern California with overlaid hot spot perimeters. 26 Oct 2003 1840 UTC

Terra MODIS true color product showing fires and smoke in southern California with overlaid hot spot perimeters. 26 Oct 2003 1840 UTC

Fire perimeters (shortwave and longwave IR detected thermal anomalies) have been added in red using information from the 4 micrometer region. In the false color image, fires are shown in two ways:

  • By direct detection due to the 2.2 micrometer channel (orange dots)
  • Based on 3.5 to 4 micrometer data (red perimeters)

To top


4.9 False Color vs. Aircraft Data

Now let's compare the false color product to a composite of aircraft data imaged using 8 to 12 micrometers valid at about the same time.

Composite of 2 images: Cedar Fire, San Diego County, CA 26 October 2003 as viewed by the FireMapper Thermal-Imaging Radiometer; Terra MODIS False Color image from 26 Oct 2003

Notice that the pattern is the same but that the red fires shown by the aircraft data are much smaller in area than those shown by the satellite in orange. What might cause this discrepancy?  

  1. Aircraft data minimize the size of fires
  2. Aircraft data do not depict vegetation
  3. The subpixel effect is at work in the false color image

Feedback: The correct answer is 3. We're seeing the subpixel effect at work in the satellite fire depictions. Satellite depictions always overestimate the size of fires. Aircraft data are the preferred analysis tool for very intense established fires. That's because we can collect data at a very high resolution, 5 to 10 meters as in the FireMapper™ example shown.

To top


4.10 Hazard Mapping System

Now we will examine several methods and systems used to produce fire products:

  • Hazard Mapping System (HMS): An analyst-integrated fire and smoke product
  • Wildfire Automated Biomass Burning (WFABBA): Automated fire detections from GOES satellites developed by the Cooperative Institute for Meteorological Satellite Studies (CIMSS)
  • MODIS Rapid Response System: Fire detections from NASA's MODIS instrument
  • MODIS USFS Active Fire Map: Fire detections over many parts of the United States in near real time

The Hazard Mapping System (HMS) is an interactive multiplatform system in which satellite analysts at NOAA-NESDIS manually integrate output from automated fire detection algorithms with data from geostationary and polar-orbiting satellites. The result is a hybrid between automated and manual techniques, producing a quality controlled display of the locations of near real time fires and smoke plumes. The products are generated daily for the contiguous United States and Hawaii, and seasonally for Alaska, Canada, and Central America. The products are used by federal, state, and local agencies as well as private companies and universities.

To prepare HMS products, satellite analysts view loops such as this NOAA AHVRR shortwave infrared animation of the southern California fires of October 2003.

NOAA AVHRR 3.7 micrometer animation of southern California fires, Oct 2003

(View animation)

They also look at GOES visible imagery such as this animated sequence showing large smoke plumes from fires in southern California.

GOES VIS animation showing southern California fires, 26 Oct 2003

(View animation)

Here is a sample HMS product. The red dots outlined in black are fire locations based on GOES, AHVRR, and MODIS input. The gray-shaded areas are smoke plumes.

HMS analysis showing fire locations, burned regions, and smoke plumes over southern California on 10 Oct 2003

To top


4.11 WFABBA and MODIS Rapid Response System

Automated systems, such as the GOES Wildfire Automated Biomass Burning Algorithm (WFABBA), are used to monitor large areas and routinely detect hundreds of fires at a time. This animation centered on the Isthmus of Tehuamtepec depicts detected fires (mostly prescribed agricultural fires) over a 24-hour period against a NOAA AVHRR-derived land cover map. The loop starts at night when we can see red dots representing the fires, but not smoke. During the daytime, smoke appears. Geostationary monitoring of the diurnal cycle is a powerful tool in the tropics. Using GOES satellites, WFABBA products are available every 15 to 30 minutes.

Wildfire ABBA and GOES-8 regional view, 20 March 2003 0300 UTC

WFABBA currently generates fire data for the Western Hemisphere, which are very useful for forecasters. The algorithm's most important function, though, may be to provide the raw data for estimating smoke emissions, which enables smoke transport to be forecast.

The MODIS Rapid Response System provides fire detection maps across the globe in near real time. Here are some fire detections over Central America.

Terra MODIS 20 March 2003 1640 UTC, fires and smoke in Yucatan Peninsula

To top


4.12 MODIS USFS Active Fire Map

Finally, the Active Fire Map product provides an easy-to-read summary of satellite information in near real time. The product is derived from recent MODIS passes and various fire detection algorithms, and is produced by the USDA Forest Service Remote Sensing Applications Center (RSAC). Product archives are also available.

This animation corresponds to the 2003 southern California fires shown earlier. Red indicates fires burning within the last 12 hours; yellow indicates burned areas since the beginning of the year.

MODIS Active Fire Detections, 20 Oct 2003

(View animation)

Animating the scene shows the fires' progression over a nine-day period. Notice how the fires near San Diego race into populated coastal regions on 26 October under the influence of offshore winds.

To top


4.13 Section Summary

Shortwave vs. longwave imagery for fire detection: Shortwave imagery is particularly sensitive to extremely hot, small fires, whereas longwave imagery has an extremely limited ability to detect fires.

Sub-pixel effect: Causes fires to appear much larger on satellite products than they might really be.

Missed detection: A very small or relatively cool fire that goes undetected on satellite imagery.

False alarm: When a person or an algorithm detects a fire that is not really there (sunglint or clouds can contribute to the mistake).

Limitations on fire detection

  • Polar orbiters typically cover a set location twice daily; geostationary temporal coverage is continuous (every 15-30 minutes)
  • Geostationary satellites do not cover near-polar areas well
  • The larger the satellite footprint (area viewed), the less detailed the fire depictions
  • Cloud cover can completely or partially obscure fires

Systems for detecting and monitoring fire

  • False color products
    • Show burn scars in a brownish or brick red color
    • Often convey fire locations as pink (or other shades of red) or orange spots
    • Provide vivid depictions of landscapes and vegetation, helping users assess the availability of fresh fuel near an advancing fire
  • True color products
    • Reveal information about the landscape (mountains, chaparral, snow cover, and water bodies, etc.), which adds context for those forecasting or monitoring fires
    • Red spots or red perimeters are often added, marking the location of likely fires based on information from infrared shortwave ‘fire' channels
    • Are the optimal satellite product for the general interpretation of smoke and other aerosols; often reveal smoke where it cannot be seen on other products
  • MODIS Rapid Response System: Provides quick access to MODIS fire data globally
  • NOAA-NESDIS Hazard Mapping System (HMS): An analyst-integrated fire and smoke product
  • Wildfire Automated Biomass Burning Algorithm (WFABBA): Provides automated fire detections from GOES satellites

The ability to detect fires enables smoke to be analyzed and forecasted, which is the focus of the next section.

To top


5.0: Smoke & Aerosol Detection, Monitoring & Transport

5.1 Section Overview

The identification of smoke on satellite images has traditionally been a tricky process. Looking at this GOES image over southern Montana, can you tell what's smoke from a forest fire vs. cloud vs. land? It's easy to make a mistake.

GOES VIS image from 03 September 2001

Now examine this MODIS true color product over the eastern U.S and Canada. Smoke from a large area of fires appears in shades of white, making it easy to distinguish it from surrounding features.

MODIS True Color Image with fire overlays, 06 July 2002

In this section, we will illustrate some of the basics of smoke interpretation and investigate methods for forecasting smoke transport.

To top


5.2 Visible Animations

This visible animation over the 2002 Rodeo-Chediski fire tells a lot about smoke behavior in mountainous regions. Many fires, including this one, tend to die down at night and in the early morning, producing little new smoke. However, the smoke is often trapped in a very stable layer within mountain valleys (a nocturnal inversion), making smoke coverage greater in the morning. This is especially true in steeper valleys and makes it difficult to detect the location of the fire. As the day progresses, the atmosphere typically destabilizes, wind increases, and smoke plumes erupt convectively. General visibilities often improve, except in the vicinity of thick plumes associated with active fires.

GOES-10 VIS animation showing smoke plumes and fire complexes from the Rodeo-Chediski fire, AZ

(View animation)

To top


5.3 VIS vs. IR

Let's compare a pair of morning NOAA AVHRR images on the same day to illustrate how visible and infrared images differ.

On the visible image, the smoke from the Rodeo-Chediski fire can be seen in shades of white, clearly distinguishing it from the darker background.

MOAA AVHRR Ch 1 (VIS), 23 June 2006

But on the longwave infrared imagery on the right, the smoke is much less distinct because the land is still at a relatively cool temperature and the cool smoke disappears against it. In addition, smoke itself is highly transparent at infrared wavelengths.

MOAA AVHRR Ch 4 (IR), 23 June 2006

These factors make infrared imagery less useful than visible imagery for viewing smoke.

To top


5.4 Black & White vs. True Color Imagery

This single channel visible image of Iraq contains a lake and an oil smoke plume.

MODIS True Color 28 June 2003

Can you identify them?

  1. The oil plume is 2, and the lake is 1.
  2. The lake is 1, and the oil plume is 2.

Feedback: The correct answer is 1. The features are difficult to differentiate because both absorb solar radiation at the visible wavelength, giving them the same intense shade of black in black and white imagery.

True color imagery can help clarify this type of situation, presenting subtle variations of solar absorption and reflection. The water in the lake now appears blue, whereas the oil smoke appears black. The true color image also reveals an environmentally damaging feature that was hidden within the black and white image—a toxic sulfur plume from a factory accident. Due to disparate scattering and absorption properties, the white sulfur smoke contrasts sharply with the black oil smoke in true color imagery.

MODIS True Color 28 June 2003

To top


5.5 Smoke Detection at Night

Nighttime visible technology is one of the few means by which smoke can be detected at night. When there is sufficient moonlight to provide illumination (quarter moonlight or greater), nighttime visible imagery reveals smoke plumes, as this DMSP OLS image of the Rodeo-Chediski fire shows. The quality of the image is somewhat poor due to the sensing limitations of the nighttime imager. Therefore, the OLS is not widely used for fire and smoke detection. But the example previews the significant design improvements expected with NPOESS, when the Day/Night band will show smoke plumes in remarkable detail.

DMSP OLS VIS 23 June 06 over Rodeo-Chedeski fire

DMSP OLS IR 23 June 06 over Rodeo-Chedeski fire

To top


5.6 Smoke Forecasting

Now we'll examine several systems used to predict smoke events and forecast smoke advection. You will notice that many of these systems are the same ones used for detecting and monitoring fire. That's because fire locations must be identified before smoke effects can be anticipated.

We'll start by examining the NOAA-NESDIS Hazard Mapping System (HMS). This system has two main functions:

  • It determines the current locations of fire and smoke
  • It initializes the HYSPLIT model used for computing dispersion and transport of smoke; HYSPLIT is the HYbrid Single-Particle Lagrangian Integrated Trajectory model run by the National Weather Service as part of the Air Quality Forecast System

Analysts start with images and loops like these of agricultural fires over Kansas. Fires from the infrared fire channel flicker in the images on the top. The resulting smoke appears on the visible images on the bottom.

GOES 3.9 micrometer animation with flickering fires, April 12 2006

(View animation)

GOES VIS animation with flickering fires, April 12 2006

(View animation)

Analysts then review output from fire detection algorithms, such as WFABBA and the MODIS Rapid Response System, to identify fires that may be producing smoke. The graphic shows fire and smoke detections.

GOES VIS animation with flickering fires, April 12 2006

The locations of fires generating smoke plumes are then assimilated into the HYSPLIT Model, which forecasts smoke trajectories. This graphic shows the HYSPLIT projection for the next morning with smoke in blue. There is good agreement with the analyst-drawn smoke (shown in red) valid for the same time.

HYSPLIT Smoke - Day 1 Forecast, 13 April 2006

To top


5.7 Automatic Smoke Forecasting

For relatively small areas, the human-machine mix of the HMS makes sense. However, for broader or global areas, a fully automated scheme is needed to initialize a model that can generate smoke forecasts in a timely fashion. This requires the automatic detection of fires and accurate estimates of smoke emissions.

We will examine the U.S. Navy's scheme, which has four steps:

  1. Identifying fires using the WFABBA algorithm
  2. Estimating smoke output using the FLAMBE program (FLAMBE stands for Fire Locating and Modeling of Burning Emissions)
  3. Forecasting smoke advection downwind using the NAAPS Model (NAAPS stands for Navy Aerosol Analysis and Prediction System)
  4. Forecasting aerosol optical thickness

Note that optical thickness (or optical depth) is a dimensionless quantity that indicates the amount of depletion that a beam of radiation undergoes as it passes through a layer of the atmosphere. Optical thickness is a function of the density, composition, temperature, pressure, and volume of the layer.

To top


5.8 Case Example

We will use an eastern Canadian case that started on 6 July 2002, a day of intense fire activity. Numerous fires appear on this MODIS true color image, marked by red dots generated from shortwave and longwave infrared channels. The fires are generating intense smoke plumes, which have already impaired visibility hundreds of kilometers downwind. The broad impact of the smoke results from a combination of large, abnormally intense fires and an ideal meteorological transport scenario that keeps the smoke relatively concentrated while moving it quickly across the continent.

MODIS True Color Image showing fire detection locations and smoke, 06 July 2002

To top


5.9 WFABBA and FLAMBE

Here are the fires detected by the WFABBA algorithm based on GOES data. MODIS detections from the Rapid Response Team are very similar.

MODIS Fires, 06 July 2002, fire detects: 4746

The FLAMBE program is then used to determine how large the detected fires are, estimate their temperature, and determine how much fuel is available—all in an effort to predict the amount of smoke that will result. FLAMBE is continually being improved as more is learned about remote fires.

This sample FLAMBE six-hour fire summary is from a different day, 28 June 2007. Note that the fire indicators are larger than actual fire size.

FLAMBE fire summary for 28 June 2007

To top


5.10 Forecasting Smoke Advection and Aerosol Optical Thickness Using NAAPS

The NAAPS model forecasts the spread of smoke conditions using a specialized model of smoke transport and evolution. NAAPS runs in tandem with the NOGAPS global weather model and gets smoke source inputs from FLAMBE. NAAPS then calculates the transport of dust, smoke, sulfate, and sea salt aerosols.

This NAAPS smoke analysis was produced for 6 July 2002. It shows a large smoke region in blue heading right for the U.S. eastern seaboard. Three kinds of aerosols are represented in the color bar: sulfates in yellows and oranges; dust in shades of green; and smoke in shades of blue.

NAAPS Analysis, 06 July 2002

NAAPS forecast the plume to move into New England, the Mid-Atlantic States, and over the offshore waters over several time periods. This 48-hour forecast shows a thick pall of smoke over enormous areas of the U.S. East Coast. Even so, the system underforecast the smoke to some degree when compared to other satellite estimates.

NAAPS Analysis, 08 July 2002

This true color image from the same day shows the thickness and extent of the smoke from the Canadian fires.

MODIS True Color Image, 08 July 2002

To top


5.11 Section Summary

Satellite imagery

  • Visible imagery is much more useful for viewing smoke than infrared imagery
  • Nighttime visible technology enables smoke detection at night given sufficient lunar illumination
  • Smoke can be detected more easily in true color imagery than in false color imagery

Model-generated smoke forecasts

  • Produced based on the positions of known fires
  • Used to anticipate hazardous air quality with a lead time of hours to days, enabling the timely issuance of public warnings
  • Can forecast degraded visibilities globally
  • Forecast systems:
    • U.S. Navy Aerosol Analysis and Prediction System (NAAPS): Forecasts smoke transport and aerosol optical thickness
    • NOAA-NESDIS Hazard Mapping System (HMS): Determines locations of fire and smoke using an analyst/algorithm mix; data are used to initialize the HYSPLIT model
    • HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model: Forecasts smoke dispersion and trajectories
    • FLAMBE (U.S. Navy Fire Locating and Modeling of Burning Emissions) program: Estimates smoke output
    • WFABBA (WildFire Automated Biomass Burning Algorithm): Detects fires, which are assimilated into smoke forecasting schemes

To top


6.0: Fire Product Suite

6.1 Introduction

This section, the Fire Product Suite, consolidates the product information presented in the module and contains additional information, such as product derivation. Each product has a description, list of applications, one or more examples, information about the product's derivation and anticipated improvements with the upcoming NPOESS VIIRS instrument, and links.

The section is optional. It is intended for those who want to reinforce what they have learned, access product information directly (not in the context of the fire cycle), or learn more about the products. The Fire Product Suite is also available as a PDF file.

To top


6.2 Fire Danger Rating Product

Description: This widely used U.S. Forest Service product is an aid to predicting the potential for wildfires. Based on surface observations of fuels, it can be compared to satellite greenness maps to infer fire susceptibility.

Applications:

  • Is a primary product used to assess the risk of fire outbreaks
  • Can be supplemented by satellite greenness products in regions devoid of surface reports

Example: The triangles mark the locations of reporting stations used to analyze fire danger. Where they are sparse, for example over Montana, satellite products play an important supplementary role.

Observed Fire Danger Class Product, 05 July 2005

Derivation: A Fire Danger Rating level takes into account current and antecedent weather, fuel types, and both live and dead fuel moisture from station reports. It is less valid in regions with fewer reports.

NPOESS VIIRS Improvements: Not applicable

Link: USFS – Wildland Fire Assessment System: http://www.wfas.net/

To top


6.3 NOAA AVHRR NDVI-derived Greenness Maps

Description: There are three types of greenness maps, all of which are composited weekly from 1.1 km AVHRR data. The maps are derived from the Normalized Difference Vegetation Index (NDVI), which measures how dense and green plant leaves are. This indicates overall vegetative health and susceptibility to burning.

  1. Visual greenness maps
    • Portray vegetation greenness compared to a very green reference such as an alfalfa field
    • Are similar to what you would expect to see from the air
    • Normally dry areas are never as green as areas that are typically wetter
  2. Relative greenness maps
    • Portray how green vegetation is compared to how green it has been historically over a particular period of time
    • The browner (or less green) the vegetation, the higher the fire risk tends to be
    • Because each pixel is normalized to its own historical range, all areas (dry to wet) can appear fully green at some time during the growing season
    • Are useful indirect measures of fire risk in data-sparse areas
    • Are used for spotting local trends in fire susceptibility
  3. Departure from average greenness maps
    • Portray how green each pixel is compared to its average greenness for the current week based on historical weekly data

Applications:

  • Are used to detect areas with possible susceptibility to fires (areas appear red or yellow)
  • Are useful in regions without surface observations
  • Can supplement surface-derived products in data-sparse areas

Example: Relative greenness maps

Red shades indicate low levels of greenness compared to historical greenness, whereas greens indicate high levels. In this example, we see low values of relative greenness over Iowa, and high values over much of the Deep South.

Relative Greenness, 24 April - 30 April 2007, IOWA labeled

Derivation: The greenness products are based on the NDVI, which is derived from NOAA AVHRR satellite observations. NDVI is an indicator of the amount of photosynthetically active radiation being reflected by a plant's chlorophyll pigment within the green portion of the visible spectrum.

Conceptual illustration of how NDVI is calculated

NDVI is calculated as a difference between visible (red wavelength region) and near-infrared light reflected by vegetation. Healthy vegetation absorbs most of the visible light and reflects a large portion of the near-infrared light that it receives. In contrast, unhealthy and/or sparse vegetated surfaces reflect more visible light and less near-infrared light. Note that the data on the graphic are representative of conditions within a satellite footprint, but that real vegetation is, of course, far more varied.

NPOESS VIIRS Improvements: Compared to NOAA AVHRR, NPOESS VIIRS will have higher spatial resolution and more spectral channels for an improved characterization of vegetation greenness and perhaps vegetation moisture.

Links:

To top


6.4 MODIS Active Fire Map

Description: An easy-to-interpret map of the location of fires and burned areas along with highways, cities, and other geographical markers. The map represents the fire extent as detected by MODIS over the last 12 to 24 hours.

Applications:

  • Serves as a briefing tool for non-satellite experts who want an accurate summary of fire conditions
  • Is appropriate for inclusion in briefings and reports
  • An online archive makes past data accessible

Example: Product loops, such as this one from southern California in 2003, make it easy to see the progression of fires. Yellow indicates recently burned areas; red indicates active fires.

MODIS Active Fire Detections, 20 Oct 2003

(View animation)

Derivation: This product is compiled at the USDA Forest Service (USFS) Remote Sensing Applications Center in cooperation with NASA Goddard Space Flight Center, the University of Maryland, the National Interagency Fire Center, and the USFS Missoula Fire Sciences Lab. Each set of maps contains MODIS active fire detections for the 24-hour period preceeding the specified time. The maps are produced daily at 3:00 AM and 3:00 PM local time in the Mountain Time Zone.

NPOESS VIIRS Improvements: With decreased data latency, products will be available sooner in emergency situations.

Link:

To top


6.5 Fire Detection Algorithms and Products

Description: Several algorithms are available to identify fire locations from satellite observations. None are completely accurate; sometimes they fail to detect fires or detect too many fires. Most are fully automated, although some involve human input for better quality control.

  • Wildfire Automated Biomass Burning Algorithm (WFABBA): Automated fire detections from GOES satellites, developed by the Cooperative Institute for Meteorological Satellite Studies (CIMSS)
  • MODIS Rapid Response System: Provides rapid access to MODIS fire data globally
  • Fire Identification Mapping and Monitoring Algorithms (FIMMA): Automated fire detections from NOAA polar-orbiting weather satellites
  • NOAA-NESDIS Hazard Mapping System (HMS): Analyst quality controlled fire and smoke products that utilize the three automated fire detection algorithms listed above. Analysts can specify additional fires; the products are used by government agencies, private companies, and universities

Applications:

  • Fire locations are overlaid on images or maps, such as MODIS true and false color images, and distributed to government users, etc.
  • Fire detections are used to estimate smoke emissions for smoke transport forecasts

Example: WFABBA

Detections are shown as red dots in this loop of Central America.

Wildfire ABBA and GOES-8 regional view, 20 March 2003 0300 UTC

(View animation)

Example: MODIS Rapid Response System

Detections appear in red on this MODIS true color imagery over Australia.

MODIS True Color of fires and smoke in SE Australia, 11 Jan 2007 0345 UTC

Example: HMS fire detections

Detections are shown in red on this map of fires in southern California.

HMS analysis showing fire locations, burned regions, and smoke plumes over southern California on 10 Oct 2003

Derivation: The automated processing of digital multispectral satellite data is key to creating these products. Note that the NOAA-NESDIS HMS system output is quality controlled by a person.

NPOESS VIIRS Improvements: In addition to the traditional shortwave fire channels carried by NOAA AVHRR and MODIS, VIIRS will have the Day/Night Band (DNB), which will significantly enhance satellite detection of nighttime fires.

Links:

To top


6.6 Geostationary and Polar-orbiting Animations

Description: Sequences of images showing fire hotspots and smoke  

Applications:

  • Helping to locate new fires
  • Viewing trends in fire hotspots
  • Watching the evolution of smoke plumes

Example: Polar-orbiting loops

Satellite loops are often made from geostationary images, but polar-orbiting satellite data can provide effective loops as well, particularly at higher latitudes where there can be significant overlap between consecutive orbits. Here is an animation of polar NOAA AVHRR shortwave (‘fire' channel) images of thermal hotspots (shown in white) over southern California during a six-day period in October 2003, day and night. Even though there are significant time gaps between the images, the sequence tells an important story of how the fires spread. Notice how the strong offshore wind shapes the fires near San Diego into a bow-shaped firestorm on 26 and 27 October.

NOAA AVHRR 3.7 micrometer animation of southern California fires, Oct 2003

(View animation)

Example: Polar-orbiting and geostationary imagery loops

Images from geostationary satellites produce smoother animations than those from polar-orbiting satellites due to the shorter time interval between scans. The animation on the top is of GOES fire channel imagery from 12 April 2006. The flickering bright spots are fires, most or all of which are agricultural or prescribed burns. The animation on the bottom is of visible GOES imagery, which shows the smoke generated from these fires.

GOES 3.9 micrometer animation with flickering fires, April 12 2006

(View animation)

GOES VIS animation with flickering fires, April 12 2006

(View animation)

NPOESS VIIRS Improvements: With decreased data latency, VIIRS imagery will be available more quickly to forecasters, and loops can be constructed with channels of higher spatial resolution.

Links: WFABBA (GOES Fire Detections from the Wildfire Automated Biomass Burning Algorithm)

To top


6.7 MODIS True Color Product

Description: A color combination image that combines channels corresponding to the red, blue, and green wavelengths (MODIS channels 1, 4, 3). The realistic color combination on these products resembles what the human eye would see given the same scene.

Applications

  • Is the optimal satellite product for the general interpretation of smoke and other large aerosols
  • Often reveals smoke where it cannot be seen on other products
  • Reveals information about the landscape (mountains, chaparral, vegetation coverage, snow cover, water bodies, etc.), which adds important context for those forecasting or studying fires
  • Red spots or red perimeters are often added, marking the location of likely fires based on thermal anomaly information derived from shortwave and longwave infrared channels

Example: This true color image shows blowing dust and smoke from the east coast of Australia. It is easy to distinguish the smoke plume from the large area of blowing dust.

MODIS True Color Product with Hotspot Info Overlaid

In contrast, smoke from burning vegetation appears relatively white, reflecting the red, green, and blue components of visible light in relatively equal amounts. Clouds are easily separated from the suspended dust in true color imagery and appear white for much the same reason.

A small group of red pixels indicating hotspots or fires appears at the western end of the smoke plume. These were inserted after the channel compositing took place, using information from the thermally sensitive shortwave infrared channels on MODIS.

Derivation: Like the upcoming NPOESS VIIRS instrument, the MODIS imager has three visible channels that correspond closely to the red, green, and blue wavelengths within the visible spectrum.

MODIS True Color Enhancement Color, Conceptual Illustration

These channels are processed so that one color range is assigned to a particular wavelength region. Combining all of the color ranges produces what is known as true color imagery.

NPOESS VIIRS Improvements: NPOESS VIIRS true color products will arrive at data processing centers less than a half hour after overpass, unlike MODIS images, which often arrive several hours after imaging.

Links:

To top


6.8 MODIS False Color Product

Description: A color combination corresponding to the wavelengths represented by MODIS channels 7, 2, and 1. In this module, we refer to the combination as ‘the false color product' even though it is one of many possible false color products.

Applications

  • Shows burn scars in a brownish or brick red color; burn scars are often adjacent to a fire in progress, allowing forecasters to assess the availability of fresh fuel in the vicinity of an advancing fire
  • Often conveys fire locations as pink (or other shades of red) or orange spots
  • Provides vivid depictions of landscapes and vegetation

Example: Compare this true color image with a false color image during an outbreak of forest fires in Alaska in 2005. The true color composite gives an excellent view of smoke, whereas the false color composite enhances landscape features and individual fire elements.

MODIS True Color 1/4/3 from 23 June 2002 1804 UTC

MODIS False Color 7/2/1 from 23 June 2002 1804 UTC

Zooming in on the false color composite, we see burn scars from current and recent fires. In addition, fire perimeters, based on the shortwave fire channel, mark the outer boundaries of hot regions. Individual fires appear in pink or orange.

MODIS False Color 7/2/1 from 23 June 2002 1804 UTC, zoomed view to highlight perimeters, hotspots, and burn scars

Derivation: The MODIS false color composite is made by combining information from three channels:

  • Band 1 (0.65 micrometers assigned to the blue color gun), a visible channel
  • Band 2 (0.86 micrometers assigned to the green color gun), a near-infrared channel
  • Band 7 (2.2 micrometers assigned to the red color gun), a near-infrared channel capable of showing hot fires

Three MODIS images, VIS, band 1 (blue), near-IR band 2 (green), near-IR band 7 (red)

Notice the small white dots on this 2.2 micrometer image, which are hotspots. Although the 4 micrometer region is usually thought of as the fire detection region, fires can also be detected on daytime 2.2 micrometer images, which is a key to making this combination successful.

How do we make these three bands easier to use as a forecasting product? If we combine the images using the red, blue, and green color guns, salient features begin to pop out. Burn scars show up as brick red or brown, smoke appears bluish, and dense vegetation appears green. Fires appear bright red or orange.

MODIS False Color Image, Ch 7, 2, 1, 23 June 2002 1804 UTC

NPOESS VIIRS Improvements: Unlike MODIS images, which often arrive several hours after imaging, VIIRS false color products will typically be available less than a half hour after overpass.

Links:

To top


6.9 Smoke Forecasting Products

Description: Model-generated forecasts of smoke, which are produced based on the positions of known fires.

  • U.S. Navy Aerosol Analysis and Prediction System (NAAPS): Forecasts smoke transport and aerosol optical thickness
  • NOAA-NESDIS Hazard Mapping System (HMS): Determines the locations of fires and smoke using an analyst/algorithm mix; data are used to initialize the HYSPLIT model for smoke forecasting
  • HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model: Forecasts smoke dispersion and transport
  • FLAMBE (U.S. Navy Fire Locating and Modeling of Burning Emissions) program: Estimates smoke output
  • WFABBA (WildFire Automated Biomass Burning Algorithm): Detects fires, which are assimilated into smoke forecasting schemes
  • NOAA National Weather Service Air Quality Forecast Guidance: Provides ozone, particulate matter, and other pollutant forecasts

Applications:

  • Anticipating hazardous air quality from fires with a lead time of hours to days, enabling the timely issuance of public warnings
  • Forecasting degraded visibilities over North America (using the NOAA Interim Smoke Forecast Tool) or globally (using NAAPS)

Example: HYSPLIT/HMS

The blue-shaded areas are those forecast by the HYSPLIT model based on fires burning the day before. For comparison, the red-shaded area is the HMS analyst-drawn smoke analysis valid at the forecast time.

HYSPLIT Smoke - Day 1 Forecast, 13 April 2006

Example: NAAPS

This product shows the NAAPS forecast of a smoke plume over the northeastern United States and offshore waters. Smoke magnitudes appear in shades of blue. The units are aerosol optical thickness.

NAAPS Analysis, 06 July 2002

NPOESS VIIRS Improvements: Reduced data latency will mean that satellite detections will be available for assimilation sooner.

Links:

To top


7.0: Case Study

7.1 Introduction

In this section, we will examine a fire scenario that occurred in Billings, Montana in 2006. Included are comments from Don Moore, National Weather Service Science and Operations Officer in Billings, who experienced the fire. You can read Don's comments as a preview to the case, or at the end.

Map of the United States highlighting the location of Billings, MT

Don's commentary: “On July 12, 2006, the National Weather Service in Billings anticipated a critical fire weather day with expected dry and windy conditions ahead of and behind a cold front along with the possibility of new fire starts because of lightning. As the front began to move across the area, the storms and strong winds organized themselves into a solid line of wind in excess of 50 MPH. What made this event unusual was the duration of the strong winds; wind gusts of 50 to 70 mph were common just ahead of the storms but once the storms passed, the winds remained high. In fact, gusts of 50 mph or more were common for an hour. This created very dangerous conditions due to preexisting fires in addition to the newly started ones.

I went home at the end of my day as usual before the event unfolded. When the strong winds hit, they didn't die down, which left me concerned. I gave work a call and after talking to the crew, decided an extra body on shift would be a good idea. I wasn't at the office very long before it became apparent that this was one of those bigger events because many of the spotters we were calling to determine the strength of the wind were stating they were helping to fight fires. Even locally in Billings, the event was turning out to be significant. We learned of a wildfire that started on the northwest side of town due to the recent lightning. At the time, we weren't aware exactly of its location but knew it was threatening homes. The webcam on the tall tower near our office was clearly showing smoke blowing quickly downwind, which left us concerned.

The following day, we learned of the significance of the event, with numerous fires across our forecast area. Several of them grew into large fires in just a few hours of time. Even the fire on the outskirts of Billings remained active the following day. The fires continued to remain quite active for several days following the event.”

To top


7.2 Advantages of True Color Imagery

NAAPS Analysis, 06 July 2002

These true color images show how the area around Billings changed over a four and a half week period, from 6 June to 11 July 2006. Notice the distinct browning of vegetation, which often indicates an increased potential to burn.

In general, what are the advantages of using true color imagery to assess fire risk?

  1. Provides detailed views due to high spatial resolution
  2. Gives a sense of realism, mimicking what our eyes see
  3. Reveals seasonal trends in vegetation
  4. Conveys information about soil moisture

Feedback: The correct answers are 1, 2, and 3. True color imagery resembles what an astronaut would see from space. It gives a detailed high-resolution view of vegetation color (the amount of chlorophyll present in a plant), but provides only indirect information about vegetation moisture. It provides no information about soil moisture.

To top


7.3 Relative Greenness

Relative Greenness Product from 31 May - 06 June 2006

Relative Greenness Product from 05 July to 11 July 2006

Does the greenness trend in eastern Montana shown over a one-month period heighten or reduce your concern about the area's susceptibility to fire?

  1. Heighten
  2. Reduce

Feedback: The correct answer is 1. The maps verify what we identified on the true color imagery—that the vegetation is browning in the Montana area from spring to early summer.

To top


7.4 Relative Greenness Question

Relative Greenness Product from 05 July to 11 July 2006

Which of the following statements about the relative greenness product are correct?

  1. Compares current greenness to a historical data range
  2. Is a direct measure of vegetation moisture
  3. Is useful over regions with few or no conventional observations

Feedback: The correct answers are 1 and 3. The relative greenness product compares the greenness at a particular spot with the maximum or minimum observed during a multi-year period. It cannot directly measure vegetation moisture, but can supplement fire danger products in regions with few conventional observations.

To top


7.5 Hotspots

This MODIS false color image shows the first glimmers of the Bundy Railroad fire on 12 July 2006. We were only able to detect it by applying an extreme zoom. Even so, we can barely see the small pink dot of the hotspot, which is partially obscured by clouds. Searching for hotspots like this can be like searching for a needle in a haystack. In the NPOESS era, the greatly improved arrival times of VIIRS products will make it easier for forecasters to identify and forecast fires like this.

MODIS False Color Image, 12 July 2006 1730 UTC

To top


7.6 Position of Flames

The MODIS image from the Aqua satellite was taken several hours after the MODIS image from the Terra satellite. As you can see, the fires have grown and a red perimeter appears on the image.

MODIS False Color Image, 12 July 2006 1730 UTC

MODIS False Color Image, 12 July 2006 2030 UTC

Where on the ground would you expect to find flames in the Bundy Railroad Fire?

  1. In the entire area within the red perimeter
  2. In a small sub-region within the red perimeter
  3. Along the perimeter itself

Feedback: The correct answer is 2. Flames always cover a smaller area than shown by red perimeters. Notice the cloud cover, which often limits the ability to see hotspots in MODIS images.

To top


7.7 Fires in GOES Animations

3.9 micrometer GOES Imagery Animation, 12 July 2006

(View animation)

Fires (red dots) show up intermittently on this GOES shortwave fire animation of the same period. Why?

  1. The fires keep going out and re-igniting
  2. The burn intensity changes
  3. The fire lies within one pixel at first and then splits between several
  4. The fires are not large enough to show hotspots
  5. Clouds are obscuring the fires

Feedback: All of these are possible. In this particular situation, the fires were large enough to be detected but intermittent cloud cover obscures them from time to time. Note that hotspots are colored red for temperature values of 40°C and higher.

To top


7.8 MODIS/Reflectivity Loop

This MODIS Terra image is overlaid with a loop of enhanced radar reflectivity and includes the fire positions seen on the previous imagery. At around 7 PM on 12 July, a gust front associated with cold frontal convection moved through the area. Winds in excess of 60 miles per hour enabled the fires to make major advances over the next hour. Early in the animation, the streamers from the fire suggest a southwesterly wind, but later they suggest a shift to northwesterly. The echoes are probably due to clouds accompanying the gust front. It is possible that some of the radar signal is also due to ash from the fires.

MODIS Reflectivity Animation, 12 July 2006

(View animation)

To top


7.9 MODIS Active Fire Maps

Winds from convection and the passage of the front created very dangerous conditions. At 9:45 PM on 12 July, Yellowstone County Disaster and Emergency Services (DES) called, requesting that a Civil Emergency Message be transmitted, ordering evacuations for areas near the Bundy Railroad Fire. The next day, the Billings Gazette documented the rapid spread of the fire, with winds of 70 miles per hour in nearby Huntley. The fire was so severe that a 25-member team assumed management of it.

Here are two MODIS active fire maps from the U.S. Forest Service for 12 and 13 July. Notice the rapid spread of fires towards the southeast based on the northwesterly wind inferred earlier from the radar loop.

MODIS Active Fire Maps 12 and 13 July 2006 1900 MDT

What are the advantages of the MODIS Active Fire Maps product compared to true color and false color imagery?

  1. Has overlays of cities, terrain, and roads
  2. Shows smoke
  3. Is easier to interpret due to the color coding
  4. Is a value-added product, reflecting quality control
  5. Annotates positions of fires from past years

Feedback: The correct answers are 1, 3, and 4. MODIS Active Fire Maps show current and recent fires in the context of cities, terrain, and roads. They cannot show smoke. The products are quality-controlled and have color coding, which is useful for non-meteorologists.

To top


7.10 False Color Imagery, 13 July

This false color image from 13 July shows how the fires spread after the previous night's wind storm. You can see that by this time, they were burning actively on all sides. Forecasters were concerned that shifting winds would take the fires into unburned areas.

MODIS false color image showing Bundy Railroad and Pine Ridge fires and Billings MT, 13 July 2006

Note that although the photographs clearly document smoke on this day, the false color combination does not depict much smoke.

MODIS false color image showing Bundy Railroad and Pine Ridge fires and Billings MT, 13 July 2006 with pictures of smoke from the fire overlaid

To top


7.11 False Color Imagery, 14 July

MODIS false color image showing Bundy Railroad and Pine Ridge fires and Billings MT, 14 July 2006

What can you say about fires, smoke, and burn scars from this MODIS false color image on the next day, 14 July?

  1. No burn scars are evident
  2. A large burn scar appears for the Bundy Railroad fire
  3. Smoke partially obscures our view of the burn scars associated with the Pine Ridge fire
  4. The Bundy Railroad and Pine Ridge fires are made up of smaller fire elements
  5. Active fires are burning in the Billings area
  6. A burn scar appears in the Billings area

Feedback: The correct answers are 2, 3, 4, and 6. Notice the prominent burn scars and smoke signature. Pink dots mark fire sub-elements.

With the fires growing more serious, a Type 2 team was sent to the fires on 14 July 2006, with the NWS Weather Forecast Office turning over responsibility to local incident meteorologists early on 15 July 2006.

To top


7.12 True and False Color Imagery

MODIS false color image of Bundy Railroad and Pine Ridge fires with photo overlays showing smoke from both fires on 13 July 2006

Here are true color and false color images from 16 July.

Which of the following statements are accurate about true and false color imagery?

  1. False color depicts smoke better than true color
  2. True color depicts burn scars better than false color
  3. False color depicts vegetation more vividly than true color
  4. Both products overlay red perimeters based on shortwave data
  5. False color sometimes displays the hottest fire cores in pink

Feedback: The correct answers are 3, 4, and 5. True color imagery is better for identifying smoke, whereas false color shows burn scars and vegetation vividly. Red perimeters derived from shortwave infrared information can be overlaid on both true and false color imagery. The hot cores of fires (pink dots) are seen in false color imagery and come from channel 7 (2.2 micrometer) information.

To top


7.13 Comments

Smoke from a wildfire over Billings MT

(View animation)

Commentary from Don Moore, Science and Operations Officer, National Weather Service, Billings, MT:

“On July 12, 2006, the National Weather Service in Billings anticipated a critical fire weather day with expected dry and windy conditions ahead of and behind a cold front along with the possibility of new fire starts because of lightning. As the front began to move across the area, the storms and strong winds organized themselves into a solid line of wind in excess of 50 MPH. What made this event unusual was the duration of the strong winds; wind gusts of 50 to 70 mph were common just ahead of the storms but once the storms passed, the winds remained high. In fact, gusts of 50 mph or more were common for an hour. This created very dangerous conditions due to preexisting fires in addition to the newly started ones.

I went home at the end of my day as usual before the event unfolded. When the strong winds hit, they didn't die down, which left me concerned. I gave work a call and after talking to the crew, decided an extra body on shift would be a good idea. I wasn't at the office very long before it became apparent that this was one of those bigger events because many of the spotters we were calling to determine the strength of the wind were stating they were helping to fight fires. Even locally in Billings, the event was turning out to be significant. We learned of a wildfire that started on the northwest side of town due to the recent lightning. At the time, we weren't aware exactly of its location but knew it was threatening homes. The webcam on the tall tower near our office was clearly showing smoke blowing quickly downwind, which left us concerned.

The following day, we learned of the significance of the event, with numerous fires across our forecast area. Several of them grew into large fires in just a few hours of time. Even the fire on the outskirts of Billings remained active the following day. The fires continued to remain quite active for several days following the event.”

To top


Section 8: Module Summary

GENERAL RESOURCES

You can download the Fire Product Suite PDF, which contains detailed information about the products covered in the module. Each product has a description, information about its derivation, examples, list of applications, VIIRS improvements, and links.

Satellite Application Table showing the capabilities of polar-orbiting and geostationary satelliltes in regard to the fire cycle; includes products not covered in module

PRODUCTS FOR PRE-FIRE MONITORING

Fire danger ratings

  • Derived from surface observations
  • Account for current and antecedent weather, fuel types, and live and dead fuel moisture
  • Can monitor trends in fire susceptibility for areas with reasonable data density
  • Values between stations are interpolated, offering only a guess at actual fuel conditions

Terra and Aqua MODIS true color imagery

  • Convey vegetation lushness (greenness) qualitatively; the greener the scene, generally the lower the susceptibility to fires
  • High spatial resolution makes the product very useful in observing transitions in vegetation
  • For some products, fire locations are overlaid and appear in red, and smoke plumes appear in gray
  • Smoke can obscure the surface, making it difficult to see burn scars and other surface features of interest

NOAA AVHRR NDVI-derived greenness maps

  • Derived from the Normalized Difference Vegetation Index (NDVI)
  • Measures how dense and green plant leaves are, which indicates overall vegetative health and susceptibility to burning

PRODUCTS FOR POST-FIRE MONITORING

Terra and Aqua MODIS false color imagery

  • Can ‘see through' smoke
  • Burn scars appear as dark red or brown
  • Burn scars alert officials about the potential for future mudslides, plant succession, and invasive species

FIRE DETECTION AND MONITORING

Shortwave vs. longwave imagery for fire detection: Shortwave imagery is particularly sensitive to extremely hot, small fires, whereas longwave imagery has an extremely limited ability to detect fires.

Sub-pixel effect: Causes fires to appear much larger on satellite products than they might really be.

Missed detection: A very small or relatively cool fire that goes undetected on satellite imagery.

False alarm: When a person or an algorithm detects a fire that is not really there (sunglint or clouds can contribute to the mistake).

Limitations on fire detection

  • Polar orbiters typically cover a set location twice daily, whereas geostationary temporal coverage is continuous (every 15-30 minutes)
  • Geostationary satellites do not cover near-polar areas well
  • The larger the satellite footprint, the less detailed the fire depictions
  • Cloud cover can completely or partially obscure fires

Systems for detecting and monitoring fire

  • False color products
    • Show burn scars in a brownish or brick red color
    • Often convey fire locations as pink (or other shades of red) or orange spots
    • Provide vivid depictions of landscapes and vegetation, helping users assess the availability of fresh fuel near an advancing fire
  • True color products
    • Reveal information about the landscape (mountains, chaparral, snow cover, and water bodies, etc.), which adds context for those forecasting or monitoring fires
    • Red spots or red perimeters are often added, marking the location of likely fires based on information from infrared shortwave ‘fire' channels
    • Are the optimal satellite product for the general interpretation of smoke and other aerosols; often reveal smoke where it cannot be seen on other products
  • MODIS Rapid Response System: Provides quick access to MODIS fire data globally
  • NOAA-NESDIS Hazard Mapping System (HMS): An analyst-integrated fire and smoke product
  • Wildfire Automated Biomass Burning Algorithm (WFABBA): Provides automated fire detections from GOES satellites

PRODUCTS FOR DETECTING AND FORECASTING SMOKE AND AEROSOLS

Satellite imagery

  • Visible imagery is much more useful for viewing smoke than infrared imagery
  • Nighttime visible technology enables smoke detection at night given sufficient lunar illumination
  • Smoke can be detected more easily in true color imagery than in false color imagery

Model-generated smoke forecasts

  • Produced based on the positions of known fires
  • Used to anticipate hazardous air quality with a lead time of hours to days, enabling the timely issuance of public warnings
  • Can forecast degraded visibilities globally
  • Forecast systems:
    • U.S. Navy Aerosol Analysis and Prediction System (NAAPS): Forecasts smoke transport and aerosol optical thickness
    • NOAA-NESDIS Hazard Mapping System (HMS): Determines locations of fire and smoke using an analyst/algorithm mix; the data are used to initialize the HYSPLIT model
    • HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model: Forecasts smoke dispersion and trajectories
    • FLAMBE (U.S. Navy Fire Locating and Modeling of Burning Emissions) program: Estimates smoke output
    • WFABBA (WildFire Automated Biomass Burning Algorithm): Detects fires, which are assimilated into smoke forecasting schemes

To top


You may now take the Module Quiz. (If you are taking this lesson via your organization's learning management system (LMS), please navigate to the quiz from within that system, not via the link above to the MetEd registration system.)

We'd love to hear about your experiences with this module. Please send us your comments by taking a few minutes to complete the User Survey.